Datasets:
ccdv
/
arxiv-summarization

Dataset card Data Studio Files Community
7
article
stringlengths
0
745k
abstract
stringlengths
0
92.7k
one of the key issues in cosmological studies is to understand the physical processes related to the structure formation in the universe . in the cold dark matter scenario , gravitational effects play essential roles in amplifying small density fluctuations generated in the early universe to shape the large - scale structures seen today . being directly associated with galaxies and clusters of galaxies , virialized dark matter halos have been widely studied theoretically and observationally . their mass function , which describes statistically the formation and evolution of dark matter halos , is shown by numerical simulations to follow a functional form universally valid for a wide range of cosmological models ( e.g. , * ? ? ? * ; * ? ? ? such a universality can be largely explained in the context of halo model which links initial density fluctuations to nonlinear dark matter halos through gravitational collapse models ( e.g. , * ? ? ? * ; * ? ? ? considerable efforts have been made to improve the spherical collapse model to include more realistic characteristics in the modeling . it has long been realized that the anisotropic features contained in the initial density fluctuations can be magnified by nonlinear gravity @xcite . it is expected that the collapse of a region first happens along the direction with the largest eigenvalue of the linear deformation tensor , thus leading to a sheet - like structure . subsequent collapse along the direction of the second largest eigenvalue contracts the sheet structure to a filament . a halo can eventually form once further collapse occurs in the remaining direction . an ellipsoidal collapse model is developed to extend these considerations to the nonlinear regime ( e.g. , * ? ? ? * ; * ? ? ? the peak - patch scenario further includes the external tidal force self - consistently into consideration and improves the modeling of gravitational collapse around initial density peaks @xcite . @xcite and @xcite incorporate the peak - patch scenario into the excursion set approach in an averaged way . they first obtain statistically the averaged shape parameters of the initial tidal field . these averaged parameters are then used in the peak - patch ellipsoidal collapse model to derive the collapse criterion . it is noticed that on average , the halo formation is delayed due to the anisotropy of the gravitational effects . the predicted halo mass function ( mf hereafter ) is then in good agreements with that from numerical simulations . being very important in the hierarchy of large - scale structures , virialized dark matter halos of galaxy scale and above contain only @xmath6 of the total mass in the universe . majority of the mass is distributed outside these large halos . in the language of halo model , the dominant fraction of the mass in the universe is contained in numerous small halos down to very low mass depending on the physical properties of dark matter particles . these small halos present anisotropic clustering patterns in space , and form , together with the massive halos , cosmic web structures . from the view point of the large halos , their formation and evolution are affected mainly by the clustering properties of the surrounding small halos as a whole . thus to the zeroth order , the mass distribution around a large halo can be described by a smooth component without considering the individuality of small halos . this approach is clearly stated in the peak - patch scenario @xcite . in the framework of the excursion set theory , @xcite introduce filaments and sheets to model the large - scale mass distribution within which virialized halos are embedded . in their analyses , filaments are treated as an intermediate state of the ellipsoidal collapse when the collapse finishes along two directions . then these filaments represent the smoothed version of the anisotropic mass distribution around fully collapsed halos . various approaches have been proposed to geometrically define filamentary structures in cosmological simulations and observations . for example tessellation method is introduced to reconstruct the density field , and the edge between tessellations naturally constitutes a segment of filaments ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? the second order derivatives , namely , hessian matrix , of the tidal field or the density field , is also widely used to classify halos , filaments , sheets and voids according to the signs of the eigenvalues of the matrix ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? @xcite propose the so called candy model for filament finding , in which a marked point process with a set of chosen parameters is used to reject points at disfavored directions and to locate elongated filamentary segments . these geometrically defined filaments , however , can not be directly associated with the excursion - set - based filaments in @xcite . the dynamics of long geometrically defined filaments may not be dominated by the local field . therefore they may break at the saddle point and accrete into the two ends separately during the late evolution . furthermore , many geometrical definitions of filaments concentrate on the features of their spatial distribution rather than give rise to countable filamentary objects . to emphasize their dynamical structures and to compare with the results of the excursion set theory , in this paper , we mainly consider supercluster - like filamentary structures . we adopt a simple but natural definition of filaments by connectivity . specifically , we first obtain the density field on a set of grids from particle positions in a simulation . then the site percolation algorithm is applied to link cells together into groups by specifying a linking density threshold . at a high enough density threshold , only virialized halos are expected to be identified . at lower thresholds , filamentary superclusters surrounding virialized halos are located . the global percolation of the cosmic web occurs when the linking density threshold reaches a lower critical value . this is illustrated in figure 1 . from top left to bottom right , the linking density thresholds are @xmath7 , respectively . as we will discuss later , the groups identified at @xmath8 correspond to virialized halos . at @xmath5 , the individual structures seen in the plot are related to filaments defined in @xcite . at @xmath9 , we see the global percolation , and a large structure with a scale comparable to the size of the simulation box appears . at lower linking thresholds , the global cosmic web gets smoother . it should be noted however , that even for @xmath10 , the global web structure can still be seen clearly . the paper is organized as follows . 2 presents our method in detail . in 3 , we analyze the mass function and the occupation statistics of the identified groups with different linking density thresholds . in 4 , we compare our results from simulations with predictions of the excursion set theory . shape statistics are given in 5 . 6 contains summaries and discussions . to study the statistical properties of filamentary objects , we analyze sets of publicly available numerical simulations from gif project ( http://www.mpa-garching.mpg.de.gif ) and virgo project ( http://www.mpa-garching.mpg.de/virgo ) , which cover a range of cosmological models and simulation parameters . in addition , we also include in our analyses three @xmath11cdm simulations kindly provided by y.p.jing @xcite . the relevant simulation parameters are listed in table 1 and table 2 . there are overlaps between the simulations we use and those analyzed in @xcite to derive the mass function of dark matter halos , and thus comparisons can be made directly between the two studies . the simulations we use have relatively low resolutions . however , they are sufficient for our purpose of study , which aims to investigate large filamentary objects around halos of galactic scale and above without concerning the details of individual small subhalos . ccrrrrrrrrcrl js&js10&@xmath11cdm&100&0.039&@xmath12 + & js11&@xmath11cdm&100&0.039&@xmath12 + & js12&@xmath11cdm&100&0.039&@xmath12 + gif&gif_@xmath11cdm&@xmath11cdm&141.3&0.02&@xmath12 + & gif_ocdm&ocdm&141.3&0.03&@xmath12 + & gif_@xmath13cdm&@xmath13cdm & 84.5 & 0.036 & @xmath12 + & gif_scdm&scdm & 84.5 & 0.036&@xmath12 + virgo&virgo_@xmath11cdm&@xmath11cdm&239.5&0.025&@xmath12 + & virgo_ocdm&ocdm&239.5&0.03&@xmath12 + & virgo_@xmath13cdm&@xmath13cdm & 239.5 & 0.036 & @xmath12 + & virgo_scdm&scdm & 239.5 & 0.036&@xmath12 + we apply a percolation algorithm to identify connected groups . percolation techniques have been used as group finders ever since the first generation of cosmological simulations . the particle based percolation , the fof algorithm ( e.g. , * ? ? ? * ) , which links nearby particles by a given linking length , is widely used to locate halos . on the other hand , the site percolation ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , which links adjacent grids with density exceeding a threshold , is mainly used to analyze large - scale structures , such as superclusters and particularly the morphology of global percolative structures at the cosmic average density . in this paper , we focus mostly on relatively large structures between virialized halos and the global cosmic - average - density surface . we thus choose the latter algorithm , which has good enough accuracies for these large structures and can be much faster than the particle - based fof operations . the site percolation allows us to find different types of structures by specifying different linking densities . these structures are thus enveloped by different isodensity surfaces , from halos at very high density regions to the global cosmic web at densities approaching the average density of the universe ( see figure 1 ) . our specific procedures are as follows . for each simulation snapshot to be analyzed , we first obtain a density field on a set of regular grids by cic interpolation from particle positions . we set a density threshold and pick up only those cells with densities higher than the threshold . then a site percolation algorithm , in which cells with shared surfaces are linked together , is applied to connect these cells into groups . for more details of the algorithm , we refer to @xcite ( see also * ? ? ? * ; * ? ? ? * ; * ? ? ? if particles distribute uniformly within cells , the site percolation would be equivalent to the particle - based fof with a correspondence between the linking density threshold for grids and the linking length for particles . however , particles are not spatially uniform within cells , thus the resolution for the site percolation depends on the grid size . compact groups with size smaller than the grid size can be smoothed and merged artificially into larger groups . to test the grid effects , for different grid sizes , we compare halo groups identified with the site percolation to fof halos with the particle linking length of @xmath14 . in figure 2 , we show the corresponding scaled mass functions @xmath15 for js12 simulation , where @xmath16 and @xmath17[d\ln m/ d\ln\nu]$ ] . here @xmath18 is the mass function for mass @xmath19 and redshift @xmath20 . the quantity @xmath21 is the rms of the linear density fluctuations at the scale corresponding to the halo mass @xmath22 , and is calculated by @xmath23p(k),\ ] ] where @xmath24 $ ] is the fourier transformation of the top - hat window function with the characteristic scale @xmath25^{1/3}$ ] , and @xmath26 is the power spectrum of linear density fluctuations . the value of @xmath27 is taken to be @xmath28 . the redshift is @xmath29 . the symbols are the results for fof groups . the dotted lines are for the site - percolation groups with the density threshold @xmath8 , and from top to bottom , the number of grids is @xmath30 , @xmath31 and @xmath32 , respectively . for the upper most dotted line , we also attach the corresponding error bars . the shaded region represents the range of the mass function with the density threshold from @xmath33 to @xmath34 for the case of @xmath30 grids . it is seen that at @xmath35 , the mass function of the site - percolation groups agrees with that of the fof groups very well at the high mass end . thus we take @xmath36 as the fiducial threshold for virialized halos . on the other hand , the resolution effect is apparent for relatively low mass halos . for the case of @xmath30 grids , halos with @xmath37 can not be well resolved . however , we expect the resolution effect to be weaker for larger filamentary objects with lower densities , which are our main concerns in the paper . thus we take @xmath30 as our fiducial grid number in the following analyses . for the site percolation , the density threshold is the critical quantity to differentiate different groups . it specifies the overdensity level of the envelope of an identified group , and should be directly associated with the average density within the group . figure 3 shows such a relation , where the vertical axis @xmath38 is , in unit of the cosmic density of the universe , the average density within individual groups calculated by @xmath39 with @xmath40 the total volume of a group of mass @xmath22 . we present the results for three sets of groups with the density threshold @xmath8 , @xmath41 and @xmath42 from top to bottom , respectively . the separations of @xmath38 for the three sets of groups are clearly seen , with @xmath43 , @xmath44 and @xmath45 , respectively . it is noted that for halos , @xmath43 , rather than @xmath46 defined for spherical halos ( e.g. , * ? ? ? * ) . this is because the volume of a group calculated here is the actual volume occupied by the connected cells , and halos are known to be triaxial in shape with a typical value of @xmath47 for the long - to - short axial ratio @xcite . discussions in 4 show that groups with @xmath5 can be related to filaments defined in the excursion set theory @xcite . we see that they have a typical average overdensity of @xmath48 . for @xmath9 , the global percolation occurs , and such a cosmic web has a typical overdensity of @xmath49 . note that except the grid effect , we do not apply any additional smoothing for the density field in our analyses . ccrrrrrrrrcrl js10&0 & 0.3 & 0.7 & 0.21&1.0 & 100 & @xmath50 + js12&0 & 0.3 & 0.7 & 0.21&1.0 & 100 & @xmath50 + gif_@xmath11cdm & 0 & 0.3 & 0.7 & 0.21 & 0.9 & 141.3 & @xmath51 + & 0.5 & 0.3 & 0.7 & 0.21 & 0.9 & 141.3 & @xmath51 + & 1 & 0.3 & 0.7 & 0.21 & 0.9 & 141.3 & @xmath51 + & 5 & 0.3 & 0.7 & 0.21 & 0.9 & 141.3 & @xmath51 + gif_ocdm & 0 & 0.3 & 0.0 & 0.21 & 0.85 & 141.3 & @xmath51 + gif_scdm & 0 & 1.0 & 0.0 & 0.5 & 0.6 & 84.5 & @xmath52 + gif_@xmath13cdm & 0 & 1.0 & 0.0 & 0.21 & 0.6 & 84.5 & @xmath52 + virgo _ @xmath11cdm & 0 & 0.3 & 0.7 & 0.21 & 0.9 & 239.5 & @xmath53 + in this section , we analyze statistically the site - percolation groups , and present a generalized jenkins functional form that can describe well the mass function of different groups , from virialized halos to low - density supercluster - like groups . it has been shown that to a very high accuracy , the mass function of particle - based fof dark matter halos follows a universal functional form , which is largely independent of cosmological models and redshifts ( e.g. , * ? ? ? * ; * ? ? ? improved from the original press - schechter form @xcite , two fitting formulae are widely used to describe such a universal mass function , namely , the jenkins form and the sheth - tormen form . the jenkins mass function can be written as @xcite @xmath54 where @xmath16 , and @xmath55 , @xmath56 , and @xmath57 are three parameters with their fitting values @xmath58 , @xmath59 , and @xmath60 , respectively , for halos . considering the ellipsoidal collapse model @xcite , @xcite derive the mass function from the excursion set theory , which is given by @xmath61^{2}/2 } { } } \nonumber\\ & & { } \times\{1 + \frac{\beta}{(a\nu)^{\alpha}}[1-\alpha+\alpha(\alpha-1)/2+ ... ]\}.\end{aligned}\ ] ] for dark matter halos , @xmath62 , @xmath63 , and @xmath64 . thus as a test for our site - percolation - based group analyses , we first study the mass function of dark matter halos identified with our algorithm . as discussed in 2 , our percolation groups with the linking density threshold @xmath8 correspond well to fof dark matter halos with the linking length @xmath14 . we then define these groups as halos . figure 4 shows their scaled mass function for all the simulation outputs listed in table 2 . the solid , dotted and colored lines are the simulation results for @xmath11cdm at @xmath29 , @xmath11cdm at @xmath65 , and the other cosmological models , respectively . the heavy solid line is the result of the jenkins mass function , and the dashed line is for the sheth - tormen mass function . as expected , the universality of the mass function for the site - percolation halos is clearly seen , and both the jenkins and the sheth - tormen functional forms can fit the simulation results very well . we now turn to groups identified with lower linking density thresholds . when the density threshold decreases from the halo threshold @xmath8 , low density regions surrounding virialized halos are included in groups . nearby halos can also merge into larger filamentary - like objects . when the density threshold approaches the average matter density of the universe , the global cosmic web extending to the whole simulation box can be identified . a natural question raised here is whether a universal mass function also exists for these supercluster - like groups . to investigate this , for each simulation in table 2 , we construct different sets of group catalogs identified with different linking density thresholds , from the cosmic average density to the density threshold for virialized halos and to even higher thresholds . for each given threshold , we analyze and compare the mass functions of groups from different simulations . it is found that for a wide range of linking density thresholds , the universality of the mass function remains . in figure 5 , we show the mass functions for groups with the density threshold @xmath5 . the line styles for different simulations are the same as those in figure 4 . it is seen clearly that the mass function scaled with the quantity @xmath66 obeys a universal form to the level comparable to that of halos . to a certain extent , the universality of the mass function for groups beyond halos can be understood qualitatively in the same way as for virialized halos . for groups corresponding to relatively low linking density thresholds , although they have not been fully virialized , their average densities are already high enough ( see figure 3 ) so that their own gravity dominates their dynamical evolution . in other words , these groups can be regarded as isolated structures that are in the intermediate stages toward forming virialized halos . considering the process of gravitational collapse of an isolated region , as long as all the dependence on cosmology and redshift can be cast into the extrapolated linear density perturbations , just as in the spherical and ellipsoidal collapse models , the process can be described in a universal way . consequently , a universal mass function for these groups is expected . on the other hand , it is also expected that the universality of the mass function should break down when the linking density threshold reaches a low enough level for the occurrence of global cosmic web structures . we find that the global percolation occurs at @xmath67 , and indeed the mass function for groups with that linking density threshold and lower does not show a universal behavior anymore . the global percolation will be discussed in detail in 3.3 . the universality of mass functions for supercluster - like groups raises a possibility for us to find an analytical form for mass functions that is generalized from that of dark matter halos . we consider the jenkins form of eq . ( 2 ) and the sheth - tormen form of eq . ( 3 ) . as seen from figure 4 and figure 5 , the universal behavior of the mass function depends on the linking density threshold . thus when we fit the functional forms to the simulation results , we expect that the best fitted values for the parameters in eq . ( 2 ) and eq . ( 3 ) are functions of the linking density threshold . we first consider the sheth - tormen form [ eq . it involves three parameters @xmath55 , @xmath68 , and @xmath69 . from the excursion set theory ( e.g. , * ; * ? ? ? * ) , the parameters @xmath68 and @xmath69 are related to the shape of the collapse barrier with respect to @xmath66 , and @xmath55 reflects the overall height of the barrier . for dark matter halos , @xmath55 is usually taken to be @xmath64 ( e.g. , * ? ? ? * ) . in @xcite , they extend the ellipsoidal collapse model to obtain the respective collapse barriers for filamentary and sheet - like objects . their derived barriers for different types of objects are different only in parameters @xmath68 and @xmath69 . to be in accordance with their analyses , in our fitting here , we fix @xmath64 and vary @xmath68 and @xmath69 . in the next section , we will consider more general fitting to further discuss the relation between our results and the excursion set theory . the heavy dashed line in figure 5 shows our best fit result with eq . ( 3 ) for groups with @xmath5 . it is seen that the two - parameter ( @xmath68 and @xmath69 ) sheth - tormen functional form can fit the low - mass end of the mass function rather well . at high mass end , however , the model gives a poor fit to the simulation results . this indicates that the simple excursion set theory can not apply directly to low - density groups . as we will discuss in the next section , this should be related to the well known peak - exclusion effect @xcite . for low - density groups , such effect is stronger than that for high - density halos , and thus the deviation between the sheth - tormen fitting and the simulations is more apparently seen in figure 5 than that in figure 4 for halos . for jenkins form of eq . ( 2 ) , we regard it as an empirical form , and thus all the three parameters @xmath55 , @xmath56 and @xmath57 are treated as free parameters in our fitting . we then find that eq . ( 2 ) can fit the mass function of groups with different linking density thresholds very well . we further obtain a generalized fitting for the three parameters that is applicable to all the groups in consideration , from halos with the linking density @xmath8 to low - density groups with @xmath70 . this is given by @xmath71 @xmath72 @xmath73 for halos with @xmath8 , we have @xmath74 , @xmath75 and @xmath76 , in agreement with the original fitting of @xcite @xmath58 , @xmath59 and @xmath60 . the slightly lower value of @xmath55 is due to the grid effect in our site - percolation analyses . for @xmath5 , we have @xmath77 , @xmath78 and @xmath79 , and the corresponding fitting is shown by the heavy solid line in figure 5 . the generalized jenkins mass function obtained here allows us to perform statistically the abundance analyses not only for halos but also for more extended low - density groups . its cosmological applications will be explored in our future studies . several weak lensing measurements reveal the existence of massive dark clumps with unusually high mass - to - light ratios ( e.g , * ? ? ? * ; * ? ? ? one proposed explanation is that those dark clumps may arise from the projection effect of low - density filaments with their elongations happening to be near the line of sight . galaxies within these low density areas are thought to be less clustered than that in high density clusters of galaxies of similar mass . however , the proper question concerned in the dark clump problem should be whether a filament has significantly less projected number of galaxies than that of a cluster of the same projected mass . in other words , it is more or less the total number of galaxies contained in a filament or in a cluster that matters . numerical studies show that the gravitational effects determine dominantly the occupation statistics of galaxies in a halo ( e.g. , * ? ? ? thus the subhalo occupation statistics can give important information on the occupation distribution of galaxies . here we compare filament occupation distribution ( fod ) of subhalos with halo occupation distribution ( hod ) of subhalos . for fod , we define groups picked up with the linking density threshold @xmath5 as filaments , in accordance with the definition of @xcite ( see 4 ) . for each filament , we count the number of halos inside it . as discussed previously , halos with @xmath80 can not be well resolved with our site - percolation group finder due to the grid effects . thus here for occupation analyses , we use the particle fof method with the linking length parameter @xmath14 to find halos in filaments . for high - density virialized halos , we do not directly count the subhalos inside them because the simulations we used have limited dynamical resolutions . instead , we adopt the hod fitting result for the average number of subhalos with minimum mass @xmath81 from @xcite , which is given by @xmath82 where @xmath22 is the host halo mass , @xmath83 , @xmath84 and @xmath85 . the results of the occupation distribution are presented in figure 6 . the horizontal axis is for the mass @xmath22 of the host filament / halo in unit of @xmath81 . we limit our analyses to @xmath86 in accordance with the relatively low numerical resolutions of the simulations used . for the same reason and also concerning galactic - scale subhalos , we take @xmath87 . the fod results for individual filaments are shown by red dots . the red solid line shows the average value of @xmath88 from the red dots . the average hod from eq . ( 7 ) for virialized host halos is shown as the black solid line . naive comparison between the black and red solid lines indeed leads to the conclusion that the number of subhalos in a high - density virialized halo is statistically larger than that contained in a low - density filament of the same mass . it should be noted , however , that we find halos in a low - density filament by particle fof group finder with @xmath14 . these halos can have mass well above that of the typical galactic halo , and thus are expected to further contain subhalos of galactic scale in them . those subhalos can not be adequately identified in our simulation analyses due to the limited dynamical resolutions . on the other hand , in the studies of @xcite , they use high resolution simulations and their halo finder can resolve well subhalos and even sub - subhalos . thus the red solid line and the black solid line in figure 6 can not be directly comparable . to make a more meaningful comparison between the occupation statistics of virialized halos and that of filaments , we need to add subhalos into the halos in filaments . then for each halo found in a filament , we adopt eq . ( 7 ) as the average value to randomly assign a number of subhalos to it . the corresponding modified fod results are shown in blue dots in figure 6 . the blue solid line is the average of the blue dots . we see that the blue solid line lies above the black solid line , showing that after taking into account subhalos , the average fod result is actually larger by @xmath89 than that of hod of the same mass . such a difference may be understood as follows . considering two large halos in a low - density filament with one of them being the largest halo in the filament . when the filament evolves further to form a virialized halo , the less massive halo is very likely to merge into the largest one and loses its identity , thus reducing the number of occupation by @xmath90 . therefore if there is a proportional relation between the subhalo fod / hod and galaxy fod / hod , the mass - to - light ratio for a filament is comparable and can be even lower than that of a cluster of the same mass , leading to difficulties for the filament interpretation of dark clumps . on the other hand , although close relations between the occupation distribution of subhalos and that of galaxies are expected , differences between the two can exist . thus detailed analyses of galaxy occupation distribution are further needed concerning the quantitative interpretation of dark clumps with filaments . in figure 6 , the red and blue dashed lines are the second moments , defined as @xmath91 , of the distributions of the red and blue dots , respectively . the red dashed line lies above the red solid line , showing the super - poisson behavior for the fod in the case without adding subhalos into halos in filaments . considering subhalos in halos , the blue dashed line is nearly the same as the blue solid line , and the distribution of the blue dots is consistent with the poisson distribution . this is because we add in subhalos assuming the poisson distribution in accordance with the hod analyses of @xcite . the large - scale connected cosmic web is the most striking feature seen in numerical simulations as well as in large galaxy redshift surveys . both the power spectrum of initial density perturbations and late - time nonlinear gravitational interactions play important roles in shaping the cosmic - web structure ( e.g. , * ? ? ? * ; * ? ? ? * ) . from figure 1 , we can see that with the linking density threshold being the cosmic average density , i.e. , @xmath10 , the global web structure is clearly seen . as the linking density threshold increases , the cosmic web becomes sharper . at @xmath4 , the global web structure starts to break out into large tree structures with massive halos in their central regions . with further increased linking density thresholds , the large tree structures break into individual groups dominated mainly by their local gravity . thus at relatively low linking density thresholds , the percolation groups have two subclasses , relatively isolated ones in low - density regions , and the large connected groups that account for over @xmath92 of volume occupied by all the groups ( e.g. , * ? ? ? while the local gravity should still dominate the formation of isolated groups , nonlocal effects affect the formation of the global web structure significantly . therefore there should exist a critical linking density threshold below which the universality of the mass function of groups breaks down due to the nonlocal gravitational effects on those large tree - shaped groups . to understand the global percolation phenomenon , in figure 7 , we show the merging path with respect to the linking density threshold for halos of @xmath93 and @xmath94 , respectively . the results are for js12 simulation . it is seen that , for massive halos , they more or less stay as isolated ones until @xmath4 . after that , these large groups merge into the largest structure , i.e. , the cosmic web . this merging process is rather sharp . at @xmath95 , all the massive halos merge into the cosmic web , and no isolated ones are left . on the other hand , for relatively low mass halos , their merging path in @xmath96 space is extended . at @xmath70 , they gradually merge into larger individual groups with the decrease of the linking density thresholds . at @xmath97 , some of them merge into the cosmic web . however , there are still isolated halos left even at @xmath10 . the difference in the merging path between the massive and low mass halos reflects the fact that all massive halos locate at high density regions which eventually become parts of the global web . for low mass halos , while some of them are in high density regions , a considerable fraction of them are in low density void regions and can keep as individual groups at @xmath98 . figure 8 shows the scaled mass function for all the simulations listed in table 2 . the linking density threshold is @xmath9 . as expected , we can see that the mass function for different simulations behaves differently at high mass end , in contrary to those shown in figure 5 with @xmath5 . the differences are larger than the expected poisson fluctuations , showing that the universality of the mass function does not hold anymore due to the formation of large tree structures which later connect to form the global cosmic web . thus approximately , @xmath9 is a critical value for the linking density threshold below which the global web structure starts to be apparent , resulting the breakdown of the universality of the mass function for groups . originated from @xcite , the halo formation theory has linked virialized halos to linear density fluctuations through dynamical collapse models . the halo mass function can then be statistically determined by specifying a proper collapse barrier for linear density fluctuations smoothed over suitable scales . to overcome the cloud - in - cloud problem , @xcite propose the excursion set theory , which considers trajectories in @xmath99-domain of the linear density perturbations at random spatial positions . here @xmath99 is the smoothing scale applied to smooth the linear density perturbation field . by relating the mass fraction contained in halos with mass greater than @xmath22 to the volume fraction occupied by trajectories first crossing the specified barrier at scales larger than @xmath100 , the excursion set theory can give rise to the halo mass function with the correct `` fudge factor '' of @xmath101 @xcite . in this theory , all the nonlinear gravitational effects are encoded in the shape of the barrier , which in turn is determined by dynamical collapse models . the simple spherical collapse model leads to a constant barrier that is independent of the halo mass @xmath22 or equivalently the smoothing scale @xmath99 . taking into account the ellipsoidal collapse , @xcite obtain a @xmath22-dependent moving barrier for the halo formation , which leads to a better agreement between the derived mass function and that from n - body simulations . the moving barrier for trajectories of linear density perturbations can be written as @xmath102=\sqrt{a}\delta_{c}(z)[1+\beta(a\nu)^{-\alpha}],\ ] ] where @xmath20 is the redshift , @xmath27 is the barrier in the spherical collapse model , and @xmath62 and @xmath63 . the corresponding mass function from the excursion set theory is then approximately determined by eq . ( 3 ) @xcite . it is noted that from the ellipsoidal model , we should have @xmath103 . however , studies show that @xmath104 is often required to be in agreement with the mass function from numerical simulations ( e.g. , * ? ? ? * ) . @xcite further extend the ellipsoidal collapse model to consider the formation of sheet - like , filament - like and halo structures , defined as having collapsed along one , two and all three axes , respectively . the effective barriers for the three classes of objects can be generally described by eq . ( 8) with @xmath105 @xmath106 and @xmath107 it is seen that for halos , the effective barrier increases monotonically with @xmath21 , delaying the formation of low - mass halos due to the nonspherical collapse . for sheet - like objects , it is a decreasing function of @xmath21 . it is for filaments that the barrier is nearly a constant in @xmath21 . it has been extensively shown that the excursion set theory with the moving barrier and the adjusted @xmath55 parameter can give rise to the mass function of virialized halos that fits the simulation results very well ( e.g. , * ? ? ? * ; * ? ? ? * ) ( see also figure 4 ) . to a certain extent , it is expected that the excursion set theory with suitable barriers should also be able to model the mass function of low - density groups as long as their formation is dominated by their local gravity . the sheets and filaments discussed in @xcite are among these low - density groups . the universality of the mass function for low - density supercluster - like groups identified with the linking density threshold @xmath70 shown in 3 is indeed in accordance with the expectation of the excursion set theory . here we perform quantitative studies to investigate if the excursion set theory can be applicable to low - density supercluster - like groups . particularly , we analyze if the filaments defined in the excursion set theory of @xcite have good correspondences to the low - density groups found in our simulations . for each group catalog found in our site percolation analyses with the linking density threshold from @xmath8 for halos to @xmath9 below which the universality of the mass function breaks down , its mass function is calculated and compared with the functional form of eq . ( 3 ) derived from the excursion set theory . it should be emphasized that the parameters in eq . ( 3 ) , @xmath55 , @xmath68 and @xmath69 , are in principle , not free parameters , but determined by the barrier of eq . ( 8) given by dynamical collapse models . thus twofolds of test should be included in the comparison of the theory against simulations , namely the functional form of eq . ( 3 ) itself , and the values of the parameters therein . here we first apply eq . ( 3 ) , regarding the parameters as free parameters , to fit the mass function from simulations . then the barrier with the fitted values of the parameters is compared with that expected from the ellipsoidal collapse model . in figure 9 , we show the fitted barrier for groups with the linking density @xmath5 , labeled as fitting of original mf " . for comparison , we also show the sheth - tormen barrier for virialized halos . as expected , the barrier for low density groups is lower and its slope with respect to @xmath21 is shallower than those of halos . however , we find that down to @xmath9 , no group catalogs have mass functions with fitted barriers that resemble the barrier of filaments given by eq . ( 10 ) derived from the ellipsoidal collapse model in @xcite . the theoretical barrier is nearly flat in @xmath21 , while the fitted barriers all have significant slopes . the fitted positive slope seen in figure 9 reflects that for the simulation results , the suppression of the mass function at low mass end relative to that at high mass end is stronger than that predicted by the excursion set theory with the barrier from the ellipsoidal collapse model . the discrepancy is clearly shown in figure 10 , where the green solid lines are the mass functions with @xmath5 for all the simulations listed in table 2 , and the green dashed line is the mass function predicted by the excursion set theory with a flat barrier . the disagreement between the simulation results and the theoretical predictions can due either to the dynamical collapse model that gives rise to the barrier , or to the excursion approach itself . in the recent study of @xcite , they test the applicability of the excursion set theory for virialized halos against simulations . they obtain the collapse barrier directly from simulations by tracing the collapse of virialized regions . they conclude that while the barrier is consistent with that from the ellipsoidal collapse model , the mass function from the excursion set theory with the obtained barrier is not in a good agreement with that from simulations . this indicates the existence of some intrinsic shortcomings in the excursion approach itself . in @xcite and @xcite they re - emphasize the importance of the non - markovian corrections to the excursion set theory in predicting the halo mass function and halo bias for filters other than the sharp - k filter . within the spherical collapse model , an analytical formulation taking into account such corrections by introducing a @xmath108 parameter is presented in ma et al . maggiore and riotto ( 2010b ) and ma et al . ( 2010 ) also point out that the complicated halo formation process can be incorporated into a stochastic barrier to further improve the predictions of the excursion set theory . this results an additional parameter @xmath55 to change the barrier @xmath27 to @xmath109 and @xmath108 to @xmath110 in both the mass function and the bias for dark matter halos . another problem known to the excursion set theory is that the mass function is derived based on the statistics of the trajectories of random points in lagrangian space . on the other hand , the structure formation should happen mainly around peaks of the initial density fluctuation field . consider a peak region that eventually forms a group of mass @xmath22 . for a particle away from the central region of the peak , the gravitational interaction can drag the particle into the group . however , the average linear density fluctuation at the position of that particle obtained by applying a spherical smoothing centered on itself over the scale corresponding to @xmath22 can be lower than the collapse barrier . thus the particle is statistically assigned to lower mass groups in the excursion approach . this can lead to over predictions of low mass groups in comparison with those from simulations ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? it is expected that such an off - center problem can have more significant effects on low - density supercluster - like groups , such as filamentary groups , than those of high - density virialized halos . to overcome this , @xcite propose the peak - patch theory for halos . in this approach , density peaks in the linear density fluctuation field smoothed over different scales are found . those peaks with heights above the collapse barrier are potential halo centers . the differential mass function at mass @xmath22 is then related to the derivatives of the number of peaks with respect to the smoothing scale at @xmath22 . such an approach can be regarded as the excursion approach only on peak particles . the average excursion set theory on random particles in lagrangian space is widely used because it is believed that it should resemble statistically the peak excursion theory . however , to compare with the mass function from simulations , there is an important additional step in the peak excursion approach , namely , the peak exclusion , which is used to trim off overlapped peaks . to certain extent , it is expected that the peak exclusion should largely remove the off - center problem in the average excursion set theory discussed in the previous paragraph . to see if this is the case , we perform the following analyses . instead of trimming out peaks in the excursion approach , we add in small - scale structures back to groups found in our simulations . specifically , for each virialized halo identified with the linking density @xmath8 , we hierarchically increase @xmath96 from @xmath111 to @xmath112 . at each level , if a new percolation group occurs , and it is not the most massive subgroup of the parent group at the previous level , it is added into the original group catalog as an individual group . the same procedures are applied for supercluster - like groups identified with lower linking density thresholds . for example for groups with the linking density threshold @xmath5 , we hierarchically increase @xmath96 from @xmath41 to @xmath111 to add in subgroups back into the original group catalog . we then analyze the mass functions of the modified group catalogs and compare them with those from the excursion set theory of eq . ( 3 ) . in figure 10 , the mass functions for original halo catalogs and the modified halo catalogs for all the simulations listed in table 2 are shown in black and blue solid lines , respectively . while the blue lines are systematically higher than the corresponding black lines at low - mass ends , the two sets of mass functions are not significantly different . however , for low - density groups , the differences are large . the green and red lines in figure 10 show the mass functions for original and modified group catalogs with the original linking density threshold @xmath5 . the red lines are considerably higher than the green lines at low - mass ends . fitting the modified mass functions to eq . ( 3 ) , we find that for the original linking density threshold @xmath5 , the fitted barrier is consistent with a flat - shaped barrier expected from the ellipsoidal collapse model of eq . the @xmath55 parameter for the amplitude of the barrier is found to be @xmath113 . this value of @xmath114 may be explained by the stochastic barrier model proposed by @xcite and @xcite . the fitted barrier is shown in figure 9 labeled as `` fitting of modified mf '' , and the corresponding mass function from eq . ( 3 ) is shown by the red dashed line in figure 10 . the good agreement between the red dashed line and red solid lines shows that taking into account the peak exclusion effects ( in a reversed way here ) , the excursion set theory with the barrier given by the ellipsoidal collapse model of @xcite with adjusted @xmath113 can describe well the mass function of filaments . in this sense , it is appropriate for us to define groups identified with the linking density threshold @xmath5 as filaments . the analyses shown in this subsection demonstrate the importance of the peak exclusion effects in the excursion set theory , especially when it is applied to model the mass function of low - density filamentary groups . without considering such effects , the average excursion set theory predicts significantly more low - mass filaments than those identified in simulations . in the excursion set theory , the conditional mass function of objects can be analyzed by invoking two sets of barriers that are appropriate for the two classes of structures in consideration . the theory has been extensively applied to construct dark matter halo merging trees , where @xmath115 , the mass fraction of the main halo of mass @xmath22 at redshift @xmath116 that is contained in progenitor halos of mass @xmath19 at redshift @xmath117 , and @xmath118 , the probability that a halo of @xmath19 at redshift @xmath119 finds itself in halos of mass @xmath22 at @xmath120 , are often investigated . in this case , the two barriers correspond to the barriers of halo formation at @xmath119 and @xmath116 , respectively ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) . in the framework proposed by @xcite for the formation of different types of objects , halos , filaments and sheets , the conditional mass functions can also be studied , which can potentially reveal the environmental dependence of structure formation . here we analyze the halo - filament and filament - halo conditional mass functions from simulations and compare them with those predicted from the excursion set theory . we denote @xmath19 as the mass of halos , and @xmath22 as the mass of filaments . thus @xmath121 represents the halo - filament conditional mass function , i.e. , the mass fraction of a filament of mass @xmath22 that is contained in halos of mass @xmath19 . the filament - halo conditional mass function is written as @xmath122 , which gives the probability that a halo of mass @xmath19 locates at filaments of mass @xmath22 . as shown in 4.1 , the mass function of filaments directly from simulations with the linking density @xmath5 is not consistent with that from the nearly flat barrier derived by @xcite for filaments . a moving barrier with a significant slope shown as `` fitting of the original mf '' in figure 9 is needed . then to test the validity of the excursion approach itself , in our calculations of the conditional mass functions from the excursion set theory , we do not use the barriers given by @xcite for halos and filaments . instead , the barriers with the parameters @xmath55 , @xmath68 , and @xmath69 obtained by fitting eq . ( 3 ) to the mass function of halos with @xmath8 and the mass function of filaments with @xmath5 from simulations are adopted . for halos , the fitted barrier is very close to that of @xcite shown in figure 9 . for filaments , the fitted barrier is the one labeled as `` fitting of the original mf '' in figure 9 . figure 11 shows @xmath121 , the halo - filament conditional mass function , for different values of @xmath22 . the histograms are the results from simulations of js12 , and the solid lines are the sharp - k monte - carlo results of the excursion set theory with the fitted barriers for halos and filaments . it is seen that the excursion set theory can describe reasonably well the overall shape of @xmath121 . for relatively small @xmath22 , it overestimates @xmath121 at low @xmath123 . this difference should not be due to numerical resolutions of the simulations as we consider only halos with @xmath124 . similar difference is also seen in @xcite for halo - halo conditional mass function , where they compare the simulation results with those from the extended press - schechter theory with the barriers from the spherical collapse model . other studies indicate that taking into account the ellipsoidal collapse improves the agreement between the the excursion set theory and simulations for the halo - halo conditional mass function ( e.g. , * ? ? ? note that in our analyses , the solid lines in figure 11 are calculated by applying the fitted barriers from the unconditional mass functions of halos and of filaments , respectively . thus the differences seen in figure 11 should stem mainly from the excursion approach itself . the off - center problem discussed in 4.1 can lead to an underestimation of merging probability and thus an overestimation of the abundance of low - mass objects . although this problem is not apparent for @xmath125 , the average halo mass function , it can have significant effects on the conditional mass function @xmath121 due to the relatively high - density environment . furthermore , the underestimation of the merging probability for filaments from the excursion set theory can have larger effects , leading to an overestimation of @xmath121 at small @xmath123 especially for low - m filaments . in figure 12 , we present the results for the filament - halo conditional mass function @xmath122 . the simulation results shown as histograms appear to be rather extended . on the other hand , the results from the excursion set theory all have sharp peaks at @xmath126 . this should also be related to the off - center problem of the excursion approach . given the halo mass @xmath19 , filaments with @xmath127 most likely contain only single halos and the halo centers overlap with their host filaments . dynamically , those filaments are just the extension of the halos therein . in such a case , the excursion approach can describe well the conditional mass function @xmath122 as expected . however , for filaments with @xmath128 , they normally have multiple halos in them , and are likely the merging products of progenitor filaments each with a considered halo in it . thus to understand the behavior of @xmath122 , one needs to understand well the merging process of filaments . as we discussed earlier , the off - center problem in the excursion set approach is especially severe in explaining the formation of large filaments . particularly , in analyzing the filament - halo conditional mass function in which the existence of halos is pre - assumed , we are biased to emphasize regions with relatively high densities . these regions have high probabilities hosting large filaments . therefore even with the fitted barrier from the unconditional mass function of filaments , the excursion set theory can not predict well @xmath122 at large @xmath129 . such a trend is also seen , albeit to a less extent , in the halo - halo conditional mass function ( e.g. , * ? ? ? because we use the effective barriers obtained by fitting to the unconditional mass functions of halos and filaments from simulations , the differences between the excursion analyses and those from simulations for both @xmath121 and @xmath122 are directly related to the differences in the joint probability distribution @xmath130 . in simulations , merging process to form large filaments generates a valley in the @xmath131 plot . this valley leads to the broad double - peak behavior for the filament - halo conditional mass function @xmath122 that cuts @xmath130 at a given @xmath19 . for @xmath121 which cuts @xmath130 at a given @xmath22 , the valley leads to a relatively rapid decrease of @xmath121 at small @xmath123 . on the other hand , the two - barrier excursion set theory can not give rise to the valley , resulting the discrepancies seen both in figure 11 and figure 12 . in this section , we analyze the shape statistics of filament groups with @xmath5 to see if they indeed are filamentary like . we define the shape of a group through its grid - based inertia tensor , which is given by @xmath132 where the sum is over all grids occupied by a group and @xmath133 is the grid central position with @xmath134 . note that we do not apply density weights to grid positions in calculating @xmath135 . thus our measurement reflects the shape of the overall spatial extension of a group , and the contaminations from substructures are minimal . by modeling approximately the spatial distribution of a group as a triaxial object , we can link the axial ratios of the group to the eigenvalues @xmath136 of @xmath135 by @xmath137 , where @xmath138 are the three axes . following @xcite , we define the triaxiality as @xmath139 an object is classified as an oblate object if its @xmath140 , a traxial object if @xmath141 , and a prolate object if @xmath142 . figure 13 presents the shape statistics . the upper two panels and the lower left panel are for the axial ratios @xmath143 , @xmath144 and the triaxiality @xmath145 , respectively , for filament groups identified with the linking density threshold @xmath5 ( solid histograms ) and for halos with the linking density threshold @xmath8 ( dashed histograms ) . it is observed that filaments tend to have smaller @xmath143 and @xmath144 than those of halos . we have @xmath146 and @xmath147 for filaments , in comparison with @xmath148 and @xmath149 for halos . the shape differences between filaments and halos are best seen in the triaxiality @xmath145 statistics . for filaments , they are dominantly very prolate with @xmath150 , consistent with the expected configuration for filaments . for halos , the distribution of @xmath145 is rather wide , and most of them are traxial in shape . the lower right panel shows the distribution of @xmath143 for filaments of two mass ranges , @xmath151 ( solid ) and @xmath152 ( dotted ) , respectively . the differences are apparent with @xmath153 and @xmath154 for high and low mass filaments , respectively . from figure 1 , we can see that the massive filaments in the lower middle panel locate at the major nodes of the global cosmic web ( see for example , the upper right panel of figure 1 ) . they usually contain multiple large halos , and their spatial extension reflects the spatial orientation of the global web structure . dynamically , along the direction of the major axis of these massive filaments , collapse has not happened yet . thus to a large extent , these filaments should correspond closely to the two - axis collapsed filaments defined from linear density fluctuations . for relatively low mass filaments , most of them are extended structures of individual halos of similar mass . therefore their shape distribution should be in accordance with that of halos . on the other hand , because they include extended and dynamically unrelaxed regions , they tend to be somewhat more filamentary - like than halos therein . applying a grid - based site percolation method to numerical simulations , we study groups identified with different linking density thresholds @xmath96 . groups with @xmath8 correspond well to fof dark matter halos . lowering @xmath96 allows us to find supercluster - like groups beyond virialized dark matter halos . as the linking density threshold approaches the average density of the universe , the global cosmic web structure can be naturally found . in the studies presented in this paper , we focus on supercluster - like groups , which are expected to be dynamically bound , although not virialized yet . these groups provide immediate environments to dark matter halos therein . therefore understanding their properties is an important step towards understanding the environmental effects on the formation and evolution of galaxies . our analyses reveal that similar to dark matter halos , the mass functions of supercluster - like groups for different simulations listed in table 2 also follow a universal behavior . this universality is consistent with the consideration that these groups are gravitationally bound systems , and form mainly through their own gravitational interactions . in other words , the universality found for supercluster - like groups and that for dark matter halos should arise from the same origin . we further find that the jenkins functional form can describe well the mass functions for not only halos , but also supercluster - like groups . an extended jenkins mass function applicable to both halos and supercluster groups is then explicitly presented , in which the parameters @xmath55 , @xmath56 , and @xmath57 depend on the linking density threshold @xmath96 . as expected , the universality of the mass functions breaks down for groups with the linking density @xmath155 where the global web structures occur . we also compare the mass functions from simulations with those from the excursion set theory with effective barriers derived from the ellipsoidal collapse model . for halos with @xmath8 , consistent with other studies , the two agree very well with the parameter @xmath55 adjusted to be @xmath64 for the moving barrier . for supercluster - like objects , the ellipsoidal collapse model gives rise to a nearly flat barrier for filaments defined as two - axis collapse objects ( e.g. * ? ? ? however , incorporating this barrier into the excursion set theory predicts a mass function that can not fit to any mass function of supercluster - like groups identified in simulations with the linking density threshold @xmath156 . the off - center problem in the excursion set theory leads to a significant over prediction for the mass function at low mass end . taking into account this problem in the comparison , we find that the mass function of the groups identified with @xmath5 is in good agreement with that from the excursion set theory for two - axis collapse filaments . defining these groups as filaments , we further study the halo - filament and filament - halo conditional mass functions . deviations from the predictions of the two - barrier excursion set theory are seen , which are especially significant for filament - halo conditional mass function . the studies carried out in this paper can have important cosmological applications . the universality of the mass functions found for supercluster - like groups raises a possibility for us to probe cosmologies with supercluster abundances . it can also be applied to model statistically how the projection effects affect clusters weak - lensing signals . in the very recent paper by @xcite , they identify filamentary galaxy groups from the 2dfgrs survey using galaxy fof method , and compare the properties of the groups with those of mock surveys constructed from numerical simulations . this study shows that it is becoming feasible observationally to analyze filamentary galaxy groups statistically , and they are in turn can be used as cosmological probes . physically , we expect that these filamentary galaxy groups should be closely associated with the supercluster - like dark matter groups in our studies . to quantitatively understand the relation between the two , detailed fod modeling for galaxies in supercluster - like dark matter groups is necessary . we discuss the fod for subhalos in 3.2 . for galaxies , however , the fod can be much more complicated , and thorough investigations are highly desired . it is further noted that analyses on real galaxy groups can only be done in redshift space . redshift distortions from peculiar velocities of galaxies can affect group identifications , and further their mass functions and shape statistics considerably . for supercluster - like groups , their ambient member galaxies tend to be in the stage of coherent infall , and thus their distribution suffers oblate distortions in redshift space . on the other hand , for their virialized inner regions , the distortion can generate finger - of - god structures in redshift space . the detailed impacts of redshift distortions on entire supercluster - like groups will be explored in our future studies . we are very grateful for hou - jun mo for stimulating discussions , yipeng jing for kindly providing us his simulation data , and liang gao and jie wang for their help on virgo and gif data . virgo simulations and gif simulations used in this paper were carried out by the virgo supercomputing consortium using computers based at computing centre of the max - planck society in garching and at the edinburgh parallel computing centre . their data are publicly available at www.mpa-garching.mpg.de/numcos . part of our analyses is done on sgi altix 330 system at the department of astronomy , school of physics , peking university ( pku ) and ccse - i hp cluster of pku . this research is supported in part by the nsfc of china under grants 10373001 , 10533010 and 10773001 , and the 973 program 2007cb815401 . abel , t. , anninous , p. , & norman , m. l. 1997 , new atron . , 2 , 181 allgood , b. , flores , r. a. , & primack , j. r. 2006 , , 367 , 1781 aragn - calvo , m. a. , van de weygaert , r. , & jones , b. j. t. 2010 , , 408 , 2163 bardeen , j. m. , bond , j. r. , & kaiser , n. 1986 , , 304 , 15b berlind , a. a. , & weinberg , d. h. 2002 , , 575 , 587 berlind , a. a. , weinberg , d. h. , & benson , a. j. 2003 , , 593 , 1 bertschinger , e. , & jain , b. 1994 , , 431 , 486 bond , j. r. , cole , s. , & efstathiou , g. 1991 , 379 , 440 bond , j. r. , & myers , s. t. 1996 , apjs , 103 , 1 bond , j. r. , kofman , l. , & pogosyan , d. 1996 , nature , 380 , 603 bond , n. a. , strauss , m. a. , & cen , r. 2010 , , 406 , 1609 bond , n. a. , strauss , m. a. , & cen , r. 2010 , , in press , arxiv:1003.3237 brner , g. , & mo , h. 1989 , , 224 , 1 bullock , j. s. 2002 , astro - ph/0106380 cole , s. , helly , j. , & frenk , c. s. 2008 , , 383 , 546 colberg , j. m. , pearce , f. , & foster , c. 2008 , , 387 , 933 cooray , a. , & sheth , r. 2002 , phys . rep . , 372 , 1 davis , m. , frenk , c. s. , & white , s. d. m. , 1985 , , 292 , 371 dekel , a. , & birnboim , y. 2006 , , 368 , 2 doroshkevich a. g. 1970 , astrofizika , 3 , 175 eisenstein , d. , & loeb , a. 1995 , , 439 , 520 engineer , s. , kanekar , n. , & padmanabhan , t. 2000 , , 314 , 279 erben , t. , van waerbeke , l. & mellier , y. 2000 , , 355 , 23 franx , m. , illingworth , g. , & de zeeuw , t. 1991 , apj , 383 , 112 gao , l. , white , s. d. m. , & jenkins , a. 2004 , , 355 , 819 gao , l. , yoshida , n. , & abel , t. 2007 , , 378 , 449 giocoli , c. , moreno , j. , sheth , r. k. , & tormen , g. 2007 , , 376 , 977 gott , j.r . , iii , dickinson , m. , & melott , a. l. 1989 , 306 , 341 keres , d. , katz , n. , & weinberg , d. h. 2005 , , 363 , 2 hahn , o. , porciani , c. , carollo , c. m. , & dekel , a. 2007 , , 375 , 489 hahn , o. , carollo , c. m. , porciani , c. , & dekel , a. 2007 , , 381 , 41 hoekstra , h. , yee , h. k. c. , & gladders , m. d. 2004 , , 606 , 67 hirata , c.m . , seljak , u. 2004 , 70 , 063526 icke , v. 1973 , , 27 , 1 icke , v. , van de weygaert , r. 1987 , 184 , 16 jenkins , a. , frenk , c. s. , & white , s. d. m. 2001 , , 321 , 372 jing , y. p. , & suto , y. 1998 , , 494 , l5 jing , y. p. , & suto , y. 2002 , , 574 , 538 klypin , a. a. , & shandarin , s. f. 1993 , , 413 , 48 kravtsov , a. v. , berlind , a. a. , & wechsler , r. h. 2004 , 609 , 35 knebe , a. , & wiessner , v. 2006 , pasa , 23 , 125 kauffmann , g. , & white , s. d. m. 1993 , , 261,921 lacey , c. & cole , s. 1993 , , 262 , 627 lacey , c. & cole , s. 1994 , , 271 , 676 ma , c. p. , maggiore , m. , riotto , a. & zhang , j. 2010 , , in press , arxiv:1007.4201 maggiore , m. & riotto , a. 2010 , , 717,907 maggiore , m. & riotto , a. 2010 , , 717,515 mandelbaum , r. , seljak , u. , & cool , r. j. 2006 , , 372 , 758 mahdavi , a. , hoekstra , h. , & babul , a. 2007 , , 668 , 806 mo , h. j. , & white , s. d. m. 1996 , , 282 , 347 monaco , p. 1999 , aspc , 176,186 murphy , d. n. a. , eke , v. r. , & frenk . c. s. 2010 , astro - ph/1010.2202 navarro , j. f. , aaron , l. , & springel , v. 2010 , 402 , 21 newman , m. e. j. , ziff , r. m. 2001 cond - mat/0101295v2 navarro , j. , frenk , c. , & white , s. d. m. 1996 , , 462 , 563 pogosyan , d. , pichon , c. , gay , c. , prunet , s. , cardoso , j. f. , sousbie , t. , & colombi , s. 2009 , , 396 , 635 press , w. h. , & schechter , p. 1974 , , 187 , 425 platen , e. , van de weygaert , r. , & jones , b. 2007 , , 380 , 551 robertson , b. e. , kravtsov , a. v. , & tinker , j. 2009 , , 696,636 romano - diaz , e. , & van de weygaert , r. 2007 , , 382 , 2 snchez - conde , m. a. , betancort - rijo , j. , & prada , f. 2007 , , 378 , 339 schaap , w.e . , & van de weygaert , r. 2000 , , 369 , l29 shandarin , s. , sheth , j. v. , & sahni , v. 2004 , , 353 , 162 shandarin , s. , habib , s. , & heitmann , k. 2010 , , 81 , 103006 sousbie , t. pichon , c. , & colombi , s. 2008 , , 383 , 1655 seljak , u. , 2000 , , 318 , 203 shen , j. j. , abel , t. , mo , h. j. , & sheth , r. , 2006 , 645 , 783 sheth , r. , mo , h. , & tormen , g. 2001 , , 323 , 1 sheth , r . , & tormen , g. 1999 , , 308 , 119 sheth , r . , & tormen , g. 2002 , , 329 , 61 stoica , r. s. , martnez , v. j. , & mateu , j. 2005 , , 434 , 423 somerville , r. s. , & kolatt , t. s. 1999 , , 305 , 1 umetsu , k. , & futamase , t. 2000 , , 539 , l5 van den bosch , f. c. , yang , x. h. , & mo , h. j. 2007 , , 376 , 841 white , s. d. m. , & silk , j. , 1979 , , 231 , 1 yoshida , n. , abel , t. & hernquist , l. 2003 , , 592 , 645 zeldovich , ya . b. 1970 , a&a , 5 , 84 zeldovich , ya . b. , einasto , j. , & shandarin , s. f. 1982 , nature , 300 , 407 zhang , j. , fakhouri , o. & ma , c. p. 2008 , , 389 , 1521z
in this paper , we study large - scale structures from numerical simulations , paying particular attention to supercluster - like structures . a grid - density - contour based algorithm is adopted to locate connected groups . with the increase of the linking density threshold from the cosmic average density , the foam - like cosmic web is subsequently broken into individual supercluster - like groups and further halos . to be in accordance with normal fof halos with the linking length of @xmath0 in unit of the average separation of particles , halos in this paper are defined as groups with the linking density threshold @xmath1 , where @xmath2 is the grid density , @xmath3 is the average mass density of the universe . groups with lower linking densities are then generally referred to as supercluster - like groups . by analyzing sets of cosmological simulations with varying cosmological parameters , we find that a universal mass function exists not only for halos but also for low - density supercluster - like groups until the linking density threshold decreases to @xmath4 where the global percolation of large - scale structures occurs . we further show that the mass functions of different groups can be well described by the jenkins form with the parameters being dependent on the linking density threshold . on the other hand , these low - density supercluster - like groups can not be directly associated with the predictions from the excursion set theory with effective barriers obtained from dynamical collapse models , and the peak exclusion effect must be taken into account . including such an effect , the mass function of groups with the linking density threshold @xmath5 is in good agreements with that from the excursion set theory with a nearly flat effective barrier . a simplified analysis of the ellipsoidal collapse model indicates that the barrier for collapses along two axes to form filaments is approximately flat in scales . thus in our analyses , we define groups identified with @xmath5 as filaments . we then further study the halo - filament conditional mass function and the filament - halo conditional mass function , and compare them with the predictions from the two - barrier excursion set theory . the shape statistics for filaments are also presented .
the abelian 2-form gauge potential , usually referred to as kalb - ramond field ( k - r ) , exhibits a number of interesting features in four dimensional space - time ( 4d ) @xcite ; special emphasis is given to the fact that it mixes up with the electromagnetic gauge potential so as to yield a massive spin-1 excitation based on a @xmath0 ^2$]-symmetry , without the need for higgs scalars @xcite . more recently , the coupling between the k - r and maxwell fields has been reconsidered in connection with the issue of dirac - like monopoles with massive photons @xcite . also , interesting discussions on the possible non - abelian extension of the k - r field are set in the works of ref . @xcite . reassessing the k - r field in three dimensional space - time ( 3d ) brings about some peculiarities that show up as a by - product of three dimensional space - time . for example , the abelian gauge field dual to the k - r potential in 3d is subject to a particular gauge symmetry that selects its longitudinal component as the physical propagating mode , whereas its transverse part appears as a compensating mode , and so it can always be gauged away . once the mass - shell condition is imposed , no physical propagating degree of freedom survives and we have the peculiarity that the k - r gauge field carries _ no _ on - shell degrees of freedom in @xmath1d @xcite . we understand that this is a consequence of setting its dynamics on a flat space - time . so , our propose is to set out this paper in order to illustrate how the coupling of the k - r field to the 3d counterpart of the einstein - hilbert gravity excites a massive scalar mode in the spectrum of the theory . this is a result that already emerges at classical level . the main motivation of our work is to compute 1-loop corrections to the k - r field self - energy , so as to understand how these effects change the mass spectrum set up at tree - level . we pursue our investigation in the cases of ordinary ( torsionless ) gravity ( section 2 ) and , later on , we go over to the general case of gravity with torsion ( section 3 ) @xcite . we find some peculiarities at the end of our analysis and we shall comment on them in our conclusive comments ( section 4 ) . the model we contemplate accounts for the non - minimal coupling of @xmath1d - gravity to the k - r field according to the lagrangian density given below : @xmath2 where the first term is the usual einstein - hilbert lagrangian : @xmath3 notice the absence of the overall minus sign that appears in @xmath4d - gravity : in @xmath1d , this must be the choice in such a way to avoid that the graviton becomes a ghost . the second term describes the lagrangian for the antisymmetric gauge field @xmath5 , @xmath6 where @xmath7 is a @xmath1-form written in terms of the potential @xmath5 as follows : @xmath8 the last term accounts for the interaction between gravity and the gauge field : @xmath9 with @xmath10 , @xmath11 being the totally antisymmetric symbol , while @xmath12 denotes the covariant derivative . the coupling constant @xmath13 carries dimension of ( mass)@xmath14 . to analyse the espectrum of the model , we consider the linearised approximation of the full theory . this consists in the expansion of the metric field around the flat background as below : @xmath15 where @xmath16 is the perturbation associated to the graviton field . as for the @xmath17-form gauge potential , it is more convenient to parametrise it in terms of its dual , @xmath18 . doing so , the gauge transformation for @xmath5 can be rephrased as @xmath19 where @xmath20 is the @xmath1-vector gauge parameter . in order to invert the wave operator that mixes @xmath16 and @xmath21 in the bilinear piece of the action , we fix the de donder gauge for @xmath22 @xmath23 where @xmath24 on the other hand , the gauge transformations ( @xmath25 ) for @xmath21 suggests the following gauge - fixing term for the abelian symmetry associated to @xmath5 : @xmath26 with these elements , the bilinear piece of the action can be cast under the form : @xmath27 where @xmath28 and @xmath29 is the differential wave operator to be inverted to give us the graviton and the gauge field propagators . this operator can be decomposed in four sectors , namely , @xmath30 where @xmath31 @xmath32 @xmath33 @xmath34 where the @xmath35s are the barnes - rivers projector operators for symmetric second - rank tensors in d=3 ( see ref . @xcite for details ) , while @xmath36 and @xmath37 are respectively the transverse and longitudinal operators for vectors . the propagators are given by @xmath38 \right| 0\right\rangle = i\left ( \mathcal{o}^{-1}\right ) _ { \alpha \beta } \delta ^{3}\left ( x - y\right ) , \label{2.25}\ ] ] where wave operator is inverted by using of the multiplicative relations satisfied by the projectors @xcite . in so doing , the propagators in momentum space read as follows : @xmath39 } { \left ( p^{2}-m^{2}\right ) } p^{\left ( 0-w\right ) } \right . \\ & & -\left . \frac{m^{2}}{\left ( p^{2}-m^{2}\right ) } \left [ p^{\left ( 0-s\right ) } + \sqrt{2}\left ( p^{\left ( 0-sw\right ) } + p^{\left ( 0-ws\right ) } \right ) \right ] \right\ } , \end{aligned}\ ] ] @xmath40 @xmath41 where @xmath42 and we have suppressed the indices in @xmath16 and in the @xmath43 projectors . the above propagators display poles at @xmath44 and @xmath45 , so that tachyons do not show up . the next step is a investigation of the possibility that negative - norm state might be present . a necessary criterium for unitarity at tree - level is analysed by saturating the propagator with external currents @xmath46 and @xmath47 ( for the graviton and the gauge field , respectively ) , compatible with the gauge symmetries of the lagrangian . ghosts are absent in the model if the imaginary part of the residue of the current - current transition amplitude taken at the propagator poles is non - negative . for the graviton , this amplitude in momentum space is written as @xmath48 \right\rangle j^{\kappa \lambda } \left ( p\right ) , \label{2.29}\ ] ] where , by virtue of the transversality of @xmath49 , only the spin projectors @xmath50 and @xmath51 survive . this fact enable us to write the imaginary part of the residue of the transition amplitude at the pole @xmath44 as @xmath52 now , defining the following set of independent vectors in momentum space , @xmath53 that satisfy the conditions : @xmath54 the symmetric tensor current @xmath55can be decomposed according to : @xmath56 substituting ( [ 2.33 ] ) into ( [ 2.30 ] ) , and making use the relations ( [ 2.31 ] ) and ( [ 2.32 ] ) , we find that @xmath57 , showing that the massless pole does not propagate . for the pole at @xmath58 , one obtains @xmath59 that is always positive - definite . from this result , we find only one on - shell degree of freedom for the massive graviton . for the vector field , the transition amplitude is given by @xmath60 \right\rangle j^\nu \left ( p\right ) . \label{2.35}\ ] ] expanding the current @xmath61 with respect to the basis ( [ 2.31 ] ) , @xmath62 and by making use of its conservation law , @xmath63 we can show that the massless pole is again non - dynamical , while the residue of the amplitude at the massive pole give us : @xmath64 which ensures one physical degree of freedom . therefore , the non - minimal coupling of the k - r field to gravity results in a appearance of a dynamical massive pole in the longitudinal sector of @xmath5 . we now turn into the calculation of the 1-loop mass correction for the gauge field propagator . the relevant vertices to our discussion are shown in figure 1 , while all possible self - energy corrections are depicted in figure 2 . the vertices of the figures 1.a and 1.b come from the lagrangians ( [ 2.3 ] ) and ( [ 2.5 ] ) by expanding the metric up to order @xmath65 , while the one of figure 1.c is obtained from ( [ 2.3 ] ) up to order @xmath66 . from the feynman rules displayed in figure 1 , the contribution coming from the loop integrals can be evaluated . due to algebraic complexity of the vertices and propagators , the explicit evaluation of these diagrams only was possible by means of the software form . since we are concerned with the contribution to the vector field mass , and we checked that a chern - simons - like term is _ not _ generated , we can calculate the diagrams by setting all momenta appearing on the external legs to zero . one therefore , obtains the following contribution to the 1-loop shift in the gauge field mass : @xmath67 now , following the same procedure as the one in the previous section , we propose to extend our analysis to the case where torsion is non - vanishing . in three - dimensional space - time , the torsion field can be covariantly split according to its so(1,2 ) irreducible components : @xmath68 namely , a trace part @xmath69 @xmath70 , a totally antisymmetric part @xmath71 and a traceless rank-2 symmetric tensor @xmath72 @xcite . a possible gauge - invariant action may be so chosen that there is no coupling at all between the k - r field and torsion , i.e. , the former `` neither yields nor feels torsion '' . thus , let us consider the action @xmath73 where the tilde on @xmath74 means that we are considering only the riemaniann part of the covariant derivative . varying the total action ( [ 3.2 ] ) with respect to the independent fields , we get the following result for the trace part of the torsion : @xmath75 while the variations with respect to the fields @xmath76 and @xmath72 give us @xmath77 the torsion irreducible components act as mere auxiliary fields and one can eliminate them by means of their respective algebraic field equations . therefore , substituing this results back into the action , one obtain a lagrangian expressed exclusively in terms of @xmath21 and @xmath16 : @xmath78 from the bilinear sector of the lagrangian above supplemented by the gauge fixing terms ( [ 2.9 ] ) and ( [ 2.17 ] ) , we obtain the following propagators : @xmath79 p^{\left ( 0-w\right ) } \right . \nonumber \\ & & + \left . \left ( 1+\frac{p^2}{m^2}\right ) \left [ p^{\left ( 0-s\right ) } + \sqrt{2}\left ( p^{\left ( 0-sw\right ) } + p^{\left ( 0-ws\right ) } \right ) \right ] \right\ } , \label{3.6}\end{aligned}\ ] ] @xmath80 it is remarkable to notice now that the poles are located at @xmath44 , contrary to the previous case , where torsion was not present . it is just the last term of the action ( [ 3.5 ] ) the responsible for the suppression of the massive pole . the mixing between @xmath16 and @xmath21 arising from the mentioned term appears with the right coefficient that eliminates the massive pole . proceeding analogously to what we did in the previous section , the massless poles are shown to be non - dynamical for both fields . from this result , one concludes that the coupling of the k - r field to gravity with non - vanishing torsion may result in a theory without on - shell degrees of freedom at tree - level as long as the action ( [ 3.5 ] ) is concerned . moreover , by computing the feynman rules for this action and calculating the 1-loop self - energy graphs , we checked that no mass is dynamically generated by radiative corrections . as a final result , we can state that the coupling of the 3d k - r field with einstein - hilbert gravity excites a dynamical mode of the rank-2 gauge potential , whenever torsion is not considered . the inclusion of torsion as a non - dynamical field changes this result , as we checked in section 3 . however , had we considered non - minimal couplings with higher powers in the curvature , the situation would be completely changed , as torsion would become dynamical and its mixing with the graviton @xcite would for sure trigger a dynamical mass for the gauge field on the basis of the results worked out in section 2 . to conclude , we would like to mention that , if gravity were described by a pure chern - simons term , even in the torsionless case the k - r field would remain non - dynamical on - shell , since no mass would be generated neither at the classical nor at the 1-loop level . moura melo , j.a . helayl - neto and n. panza , int.j.mod.phys . * a14 * ( 1999 ) 3949 * * ; * * h.r . christiansen , m.s . cunha , j.a . helayl - neto , l.r.u . manssur and a.l.m.a . nogueira , int.j.mod.phys . * a14 * ( 1999 ) 147 , int.j.mod.phys . * a14 * ( 1999 ) 1721 ; c.n . ferreira , j.a . helayl - neto and m.b.d.s.m . porto , nucl.phys . * b620 * ( 2002 ) 181 ; j. barcelos - neto , a cabo , and m.b.d . silva , z. phys . * c72 * ( 1996 ) 345 ; a. lahiri , phys . rev . * d55 * ( 1997 ) 5045 ; m. henneaux , v.e.r . lemes , c.a.g . sasaki , s.p . sorella , o.s . ventura and l.c.q . vilar , phys.lett . * * b410 ( * * 1997 * ) * 195 ; f.w . hehl , p. von der heyde , g.d . kerlick , and j.m . nester , rev . mod.phys . * 48 * ( 1976 ) 3641 ; f.w . hehl , j.d . mccrea , e.w . mielke , and y. neeman , phys.rep . * 258 * ( 1995 ) 1 ; de sabbata and m. gasperini , `` introduction to gravitation '' , world scientific , singapore ( 1985 ) ;
the mechanism of dynamical mass generation for the gauge field is studied through 1-loop . we find out that torsion is an obstruction to the appearance of a 1-loop mass correction . contrary , if torsion is not present , a mass gap is generated for the 2-form field .
a liquid upon cooling undergoes first order phase transition and forms a crystal . however if the cooling rate is increased it can not crystallize and forms an amorphous glassy material @xcite . in addition to fast supercooling , there are other methods to favor glass formation over crystallization . in bulk metallic glass community the usual thumb rules are to at least have a two component mixture with negative enthalpy of mixing and a 12@xmath0 size ratio between the components @xcite . single component systems are known to crystallize in a fcc+hcp structure @xcite , thus multi - component systems are commonly used for making glasses . the negative enthalpy of mixing makes sure that the components remain in a mixed state and do not demix to form single component crystals , whereas the size ratio provides frustration in packing . although there is an array of experimental systems which form glasses , in computer simulation studies there is only a handful of systems known to be good glass former @xcite . note that most of the glass forming systems have global crystalline minima @xcite . thus depending on the barrier to crystallization it is just a matter of time for the systems to crystallize . with the increase in the available computational power some of the well known glass former like kob - anderson ( ka ) model and wahnstrom ( wa ) model are now found to crystallize @xcite . thus in order to design better glass formers we need to be able to estimate the glass forming ability ( gfa ) of these systems . in order to quantify gfa , first we need to understand the origin behind the stability against crystallization . this is an active field of research and different studies have attributed the gfa to different phenomena @xcite . the most popular among them is the theory of frustration first proposed by frank @xcite . according to him , the local liquid ordering is different from the crystalline order and this frustrates the system and decreases the rate of crystallization . it has also been argued that regions with locally favored structures ( lfs ) give rise to domains and are connected to the slow dynamics in the supercooled liquids @xcite . sometimes the lfs can also be related to the underlying crystalline structure @xcite . in some cases the lfs connected to crystal structure grows more than the one connected to the liquid structure @xcite . the locally favored structures can be different in different dimensions . there are lfs , like the icosahedral ordering , which can cause frustration in the euclidean space but tile the curved space @xcite . frustrations are not always structural but can also be energetic in nature @xcite . most binary equimolar mixtures form crystalline structures @xcite , where the crystal structure may vary according to the size ratio of the components . there are also some exceptions like the equimolar cuzr structure which is found to be a good glass former @xcite . however , when the composition of the mixtures are changed then it is usually found that close to the deep eutectic point many of them form glasses . one of the argument in favor of the deep eutectic point being a good glass forming zone is that the viscosity is highest at this point so kinetically it takes a longer time to form a crystal nucleus . however it has also been shown that the structural frustration between two different crystal structures can make this region a good glass former . this kind of phase diagram ( in temperature vs. composition space ) are often referred to as a v - shaped phase diagram where the bottom of the v is the glass forming region @xcite . in a recent work by some of us we have shown that even though all the systems at equimolar mixture undergo crystallization , as the composition of the larger size particles increases , the zone which forms cscl crystal at equimolar composition does not crystallize any more @xcite . it is already known from the study of energetics that the global free energy minima of these systems are cscl+fcc crystals @xcite . the well known ka glass former is one of the systems present in this more generic cscl zone and it has shown strong resistance towards crystallization even after being inserted with a cscl seed in the liquid @xcite . so far only in one study it has been reported to crystallize but in a structure which is different from that of the global minima @xcite . in the earlier study we have shown that in the cscl+fcc crystal structure the bigger a " particles need to have two different populations where there is a large difference in the order parameter ( coordination number and bond orientational order parameter ) of these two populations . according to us , this large difference in order parameter creates frustration . thus the stability against crystallization is attributed to the structural frustration between the cscl and the fcc crystal structure @xcite . in this present work we study a similar series of binary systems by changing the composition and also the inter species interaction length . many of the binary systems studied here are good glass formers and have a global minima which is cscl+fcc structure . thus according to our earlier study the structural frustration for these systems are similar . however , these different systems although share the same structural frustration are expected to have different glass forming ability . the goal of this work is to get a relative estimate of the gfa of different systems and then explore the origin behind their differences . our study shows that the free energy cost for cscl crystallization increases with the composition of the smaller particles . the system with lowest free energy cost also shows a pre - crystalline demixing in the liquid phase near the liquid / crystal interface . the demixing takes place due to the structural frustration between the cscl and fcc structures . upto a certain composition , the composition dependence of the free energy cost to create a crystal nucleus can be related to the composition dependence of this demixing entropy . our study of energetics shows that although in the whole range of composition the global minima is cscl+fcc crystal the driving force of crystallization in a certain region is the cscl crystal and in another region is fcc crystal . in the former region the system tends to demix and form cscl+fcc crystal and demixing frustrates the crystallization process . however , in the latter region we show that demixing does not play a crucial role . it is primarily the slow dynamics near eutectic point and lfs around the smaller b " particles which frustrate the crystallization process . the simulation details are given in the next section . in section iii we present the definition and method for evaluating different quantities , in section iv we have the results and discussion , and section v ends with a brief summary . the atomistic models which are simulated are two component mixtures of classical particles ( larger `` a '' and smaller `` b '' type ) , where particles of type _ i _ interact with those of type _ j _ with pair potential , @xmath1 , where r is the distance between the pair . @xmath1 is described by a shifted and truncated lennard - jones ( lj ) potential , as given by : @xmath2 where @xmath3 $ ] and @xmath4 . subsequently , we ll denote a and b types of particles by indices 1 and 2 , respectively . the different models are distinguished by different choices of lengths and composition parameters . length , temperature and time are given in units of @xmath5 , @xmath6 and @xmath7 , respectively . here we have simulated various binary mixtures with the interaction parameters @xmath5 = 1.0 , @xmath8 = 0.88 , @xmath9 = 1 , @xmath10 = 1.5 , @xmath11 = 0.5 , @xmath12 and the inter - species interaction length @xmath13 . we have simulated systems with different compositions , varying @xmath14 from 0.50 to 0.0 , where @xmath14 is the mole fraction of the smaller b type particles @xcite . the molecular dynamics ( md ) simulations have been carried out using the lammps package @xcite . we have performed md simulations in the isothermalisobaric ensemble ( npt ) using nos - hoover thermostat and nos - hoover barostat with integration timestep 0.005@xmath15 . the time constants for nos - hoover thermostat and barostat are taken to be 100 and 1000 timesteps , respectively . except for the liquid / crystal interface study where we use a rectangular box , all of the other studies are performed in a cubic box with periodic boundary condition . the free energy barrier calculations are done via biased monte carlo method . all the studies are performed at @xmath16 . bond orientational order ( boo ) parameter was first prescribed by steinhardt _ et al . _ to characterize specific crystalline structures @xcite . to characterize specific crystal structures we have calculated the locally averaged boo parameters ( @xmath17 ) of _ l_-fold symmetry as a 2__l__+1 vector,@xcite @xmath18 where @xmath19 here @xmath20 is the number of neighbours of the i - th particle and the particle i itself . @xmath21 is the local boo of the i - th particle . @xmath22 where @xmath23 are the spherical harmonics , @xmath24 and @xmath25 are spherical coordinates of a bond @xmath26 in a fixed reference frame , and @xmath27 is the number of neighbours of the _ i_-th particles . two particles are considered neighbours if @xmath28 , where @xmath29 is the first minimum of the radial distribution function ( rdf ) . for the liquids and the crystals the @xmath29 has been chosen as the first minima of the respective partial rdf of the a " type of particles . for the pure cscl crystal this comprises of 14 neighbours and for fcc 12 neighbours . in fig.[fig1 ] we plot the probability distribution of @xmath30 of the liquid at three different composition and also the same for pure cscl and fcc crystals . we note that at the level of this parameter all the three liquids can be clearly separated from the two different crystal forms . for the liquid at three different compositions @xmath31 at @xmath32 . we also plot the same for the cscl crystal made up of a " and b " type of particles and pure fcc made up of a " particles . ] we have calculated the relaxation times obtained from the decay of the overlap function @xmath33 , where @xmath34 . it is defined as @xmath35 the overlap function is a two - point time correlation function of local density @xmath36 . it has been used in many recent studies of slow relaxation @xcite . in this work , we consider only the self - part of the total overlap function ( i.e. neglecting the @xmath37 terms in the double summation ) . earlier it has been shown to be a good approximation to the full overlap function . so , the self part of the overlap function can be written as , @xmath38 again , the @xmath39 function is approximated by a window function @xmath40 which defines the condition of `` overlap '' between two particle positions separated by a time interval t : @xmath41 the time dependent overlap function thus depends on the choice of the cut - off parameter @xmath42 , which we choose to be 0.3 . this parameter is chosen such that particle positions separated due to small amplitude vibrational motion are treated as the same , or that @xmath43 is comparable to the value of the msd in the plateau between the ballistic and diffusive regimes . in order to calculate the crystallization rate and thus the glass forming ability we first determine the melting temperatures of the different crystals . the melting temperature is studied by calculating the temperature dependent growth / melting rate of the crystal and fitting them to a straight line . the temperature at which the growth rate cuts the temperature axis is the predicted melting temperature where the growth rate goes to zero @xcite . the simulations are done at @xmath16 . with the crystal at the center of the box and the crystal particles being pinned the liquid of 8000 particles is equilibrated at t=1.5 . the system is then quenched to the target lower temperatures and the crystal particles are unpinned . we then run a short equilibration of 1000 steps for the quenched system . depending on the temperature and the composition of the liquid the central seed either grows or melts . in the @xmath44 and @xmath45 systems we study the melting temperature of cscl crystal with an initial crystal seed of 432 particles . in the @xmath46 mixture we study the melting temperature of the pure fcc crystal comprising of 500 a " particles . the growth of the seed is monitored by cluster analysis where the @xmath30 is calculated for each particle and if the value of @xmath47 ( fig.[fig1 ] ) and it has a neighbour which is part of the existing cluster then it is included in the cluster . the cluster growth is monitored for about 100 - 500 @xmath48 , where @xmath48 is the temperature dependent @xmath49 relaxation time that varies across different systems . 5 - 10 independent runs are generated at each temperature by starting from the same initial configuration but randomized initial velocity . the growth rate is calculated by scaling the time w.r.t the corresponding @xmath48 . from the average growth / decay rate we approximate the melting temperature as the temperature where the predicted growth or decay rate is zero ( fig.[fig2 ] ) . the melting points obtained from fig.[fig2 ] is used to construct the composition dependent phase diagram reported in fig.[fig3 ] . system , the predicted melting temperature ( @xmath50 ) is 0.651 . ( b ) @xmath51 , @xmath52 . ( c ) @xmath46 , @xmath53 . ] ) . the melting temperature of the mixed cscl+fcc crystal and the distorted fcc crystal are obtained by step wise heating the system . ] we find that the @xmath44 mixture phase separates and forms a cscl+fcc crystal structure ( figs.[fig4]a ) . the @xmath51 mixture also shows similar tendency however the crystal growth rate is slower and within our simulation timescale the demixing is not complete . we also try to grow the cscl crystal in the @xmath46 mixture but we find that instead of cscl , fcc structure of a " particles grow around the initial seed ( fig.[fig4]b ) . this is similar to the observation reported earlier @xcite . when a fcc seed is inserted in the same mixture it continues to grow . in the above mentioned method it is not possible to calculate the melting temperature of the mixed cscl+fcc crystal as the growth of such crystal never happens . for this calculation at each composition ( @xmath54 ) we heat the mixed crystal ( cscl+fcc ) starting from temperature 0.2 - 0.3 and increase it up to 0.59 - 1.0 ( depending on the melting temperature of the crystal ) with temperature interval of 0.05 . closer to the melting temperature heating is done with 0.01 temperature interval . at each temperature equilibration is done for @xmath55 steps . the size of the initial crystal structure is in the range of 468 - 612 . the total number of particles are chosen in such a way that a perfect mixed crystal can be created . periodic boundary condition is applied in all directions . similar study is been done for the pure and distorted fcc crystal for @xmath56 systems . for @xmath57 we get pure fcc and for @xmath58 and @xmath59 the a " particles form fcc crystal but with distortion due to presence of the b " particles . in the @xmath46 system within our simulation run we could not form the fcc crystal . however as reported earlier in a mka2 model , if the interaction between the two species is reduced , then the system forms crystal @xcite . in a similar method by keeping the @xmath60 we first form a distorted fcc crystal of the @xmath46 system . once the crystal is formed we switch back to the larger inter species interaction of @xmath61 and study its melting . the melting of all the crystals happen instantaneously . the melting temperatures are reported in fig.[fig3 ] . in this section we perform a comparative study of the gibbs free energy ( potential of mean force ) of crystalline nucleation / growth in different binary mixtures using umbrella sampling technique with the reaction coordinate being the size of the largest crystalline cluster present in the system . the studies are performed at the same degree of undercooling at @xmath62 , where the melting temperatures used are those calculated by studying the temperature dependent growth / melting rate for the pure cscl and fcc crystals . a crystalline cluster is defined by a neighborhood criteria ( within a cut - off distance determined by the first minimum of the partial radial distribution of function of `` a''-type particles for respective systems ) of `` crystal - like '' particles ( with the criterion of @xmath63 ) . to grow the clusters we use a biased monte carlo approach , where we apply an external harmonic potential of the form @xmath64 , where @xmath65 is the force constant , @xmath66 is the number of particles in the largest cluster , and @xmath67 is the position of the bias window . we use @xmath68 for @xmath44 , and @xmath69 for @xmath70 and @xmath71 . we have used 5 - 7 umbrella windows ( depending on the system ) in the cluster size range of 15 - 35 . after equilibration , the data is collected for @xmath72 monte carlo steps per window and weighted histogram analysis method ( wham)@xcite is then used to compute the free energy as a function of the size of the largest cluster as reported in fig . [ fig5 ] . and @xmath45 we can grow the cscl cluster , whereas for @xmath46 we can only grow the fcc cluster . even with a initial small cscl seed the cluster that grows is of a " particles forming fcc lattice which is similar to that we find for melting study . ] , as obtained within each slab of width one @xmath5 as a function of the distance from the interface . the interface that has a " particles is taken and the plot is done for @xmath44 and @xmath51 . we find that for the former system where the initial rate of crystallization is higher the interface has higher concentration of b " particles compared to the bulk . thus there is pre - crystalline demixing in the liquid phase . ] while our calculations focus on the pre - critical region of the free energy surfaces , we can compare the relative free energy cost to form a crystalline nucleus of certain size as the composition of the system is varied . we observe that the free energy cost to grow a nucleus from 15 to 35 for all the systems are quite high ( in the range of 10 - 20 @xmath73 ) , which explains why all these systems are good glass formers . a comparative study of the cost of free energy shows that @xmath44 has a lower cost to grow a cscl crystal compared to @xmath51 . this explains the slow growth of the cscl crystal in the latter system which is observed during the melting study . we also try to grow cscl crystal for @xmath46 , which we do not observe during our simulation time . this implies that the free energy cost for cscl crystal growth in this system is even larger . however , similar to the melting study the crystal that grows around the initial cscl cluster in the @xmath46 system is made up of only `` a '' particles . next we study the free energy cost for fcc crystallization in @xmath46 system . we find that the free energy cost to grow a fcc crystal from 15 - 35 cluster size in @xmath46 system is lower than the free energy cost to grow a similar size range cscl crystal for @xmath51 . this implies that in the @xmath46 system the free energy cost for fcc crystallization is lower than the cscl crystallization . note that although we make this comparative statement we are unable to determine the free energy cost for growing a cscl crystal in the @xmath46 which leads us to believe that the cost must be very high . , that of the partially demixed pre - crystalline liquid @xmath74 and the difference between them @xmath75 plotted at different compositions . we also plot the @xmath76 where @xmath77 is the melting temperature of the mixed cscl+fcc crystal . the @xmath75 shows a non monotonic composition dependence with a maxima around @xmath46 . ] we next analyze the origin behind the difference in the free energy cost to grow a cscl crystal in different systems . in a recent study of crystallization in @xmath78 mixture it is found that the barrier to crystallization for the mixed system is about @xmath79 higher than the pure system @xcite . the @xmath80 and the @xmath81 have a small difference in their sizes and form fcc crystal structure . thus unlike structural frustration between the cscl and fcc crystal present in the systems studied here @xcite there exists no structural frustration in the @xmath78 system . however due to higher @xmath82 interaction the crystal nucleus for the @xmath78 system has a higher concentration of the @xmath80 molecules compared to that in the bulk . this leads to demixing in the system and the authors concluded that this demixing leads to higher barrier . in a separate study it is shown that the phase that nucleates easily is the one which has composition closer to the liquid @xcite . in this present study we note that in the cscl crystallization process , except for the equimolar mixture the composition of the nucleus is different from that of the liquid . the difference increases as we go towards smaller @xmath83 values . thus it is obvious that the growth of cscl crystal leads to demixing in the system . however , we would like to investigate if signature of demixing is present in the liquid which surrounds the crystal . in a recent study it has been shown that the liquid in the crystal / liquid interface shows some compositional ordering @xcite . in a similar spirit we now look at the crystal / liquid interface and investigate if the demixing takes place in the pre - crystalline liquid . for this study we perform @xmath84 calculation in a rectangular box where @xmath85 . initially the system consists of 432 cscl crystal particles ( equal amount of a " and b " particles ) and 864 liquid particles . the 010 layer of the crystal faces the liquid where the last layer of the crystal on one side has a " particles and on the other side has b " particles . the box length in the @xmath86 and @xmath87 direction is @xmath88 . the box length in the @xmath89 direction is @xmath90 . period boundary condition is applied in all directions . since we want to study the interface property it is important to not have a rugged interface thus the study is performed above the melting temperature of the pure cscl crystal in the respective liquid ( @xmath91 ) by pinning the crystal particles . although performed above the melting temperature while equilibrating the @xmath44 system we find the growth of a layer of particle on both sides of the crystal . in the analysis we consider these two layers , which are not pinned , to be part of the crystal . thus for this system after equilibration there are 504 crystal particles and 792 liquid particles . the liquid particles span over more than @xmath92 distance which makes it possible to study both the interfacial and bulk properties of the liquid . for the @xmath51 system an extra layer of a " particles grow on the surface which has b " particles facing the liquid . due to the scarcity of b " particles no extra b " layer grows on the other side . in this analysis we consider the surface where the extra layer of a " particle has grown . we calculate the fraction of b " particles , @xmath93 , within each slab of width @xmath94 , as a function of distance from the interface . the first point ( z=0 ) in this plot is taken within the crystal which for the both the systems show same value of @xmath93 . interestingly we find that for the @xmath44 system the concentration of the b " particles are higher at the interface and it gradually reaches the bulk value around z=4 . however for @xmath51 system the concentration of the b " particles are same at the interface and at the bulk . thus we show that the liquid which has a lower free energy cost for crystal growth also undergoes a pre - crystalline demixing in the liquid phase . similar to the earlier study @xcite we find that the liquid / crystal interface properties differ for apparently similar systems with different glass forming ability . thus we show that the process of crystallization requires demixing which takes place in the pre - crystalline liquid . we now analyze the role of demixing in the free energy barrier . note that the per particle mixing entropy in a liquid can be written as , @xmath95 where @xmath96 is the mole fraction of the components . to form cscl+fcc crystal the liquid needs to demix . we show here that the demixing takes place in a liquid state ( refer to fig . [ fig6 ] ) . although the demixing process happens step wise here we calculate the total effect of demixing . thus we consider that to form a cscl+fcc crystal , part of the liquid needs to form a equimolar mixture and the other part should have pure a " particles . thus the per particle mixing entropy in the pre - crystalline partially demixed liquid should be , @xmath97 the difference between these two entropies , @xmath98 , is the mixing entropy at per particle level that a liquid will loose in the process of partial demixing . @xmath75 as a function of @xmath83 is shown in fig . [ fig5 ] which shows a non - monotonic behaviour with a maximum around @xmath99 . note that this kind of non monotonic behaviour is obtained in the free energy barrier to crystallization for the @xmath78 mixture which as discussed earlier is attributed to the demixing process@xcite . thus our demixing entropy study can explain the increase in the free energy cost for cscl crystal growth with decrease in @xmath83 till it reaches a value of @xmath100 . however this study does not explain why in the @xmath46 system where cscl+fcc is the global minima the free energy cost for fcc crystllization is much lower than cost for the cscl crystallization , the latter being so high that an estimation of it is beyond the scope of the present study . in order to understand the origin behind lower free energy cost for fcc crystallization we analyze the role of different crystal structures in crystallization by studying the energetics . in fig.[fig8]a we plot the energy per particle of the liquid , the mixed crystal , the fcc crystal and the cscl crystal for different compositions at 0.8 times the melting temperature of their respective mixed crystals ( given in fig.[fig1 ] ) . this is the melting temperature which has been obtained by step wise heating the mixed crystal . we find that the energy of the mixed crystal is always lower than the supercooled liquid , which implies that the liquid is in a metastable state . the energy of the cscl crystal is always lower than the liquid . however for higher @xmath83 values the energy of the fcc crystal is above the liquid and at lower @xmath83 values although it becomes less than the liquid it is always higher than the cscl value . this would imply that the cscl crystal always drives the crystallization process . however this does not explain why both in the melting study and the free energy barrier calculation at @xmath46 although we can not grow cscl crystal we can grow fcc crystals . our subsequent analysis will explain this discrepancy . next we make an estimation of energy of the mixed crystal , @xmath101 , at the per particle level , at different compositions by assuming that 2@xmath83 of the crystal forms cscl and the rest forms fcc . @xmath102 here @xmath103 and @xmath104 are the energy of the cscl and fcc crystal respectively , at per particle level calculated for each system at their respective @xmath62 . @xmath105 and @xmath106 are the estimated contribution from the respective cscl and fcc crystal part of the mixed crystal again presented at the per particle level . note that the values of @xmath105 and @xmath106 take into account the fraction of the system which is in different crystal form . in this calculation we of course make some mistake by neglecting the surface energy . however we find that the value of energy per particle of the mixed crystal thus calculated is not too different from the value of the actual crystal ( fig.[fig8]b ) . these are again calculated at the same temperatures as reported in fig.[fig8]a . we now break up the contribution of the two components , the contribution from cscl and that from fcc and plot them separately . once we do that we find that although at higher @xmath83 values the cscl formation drives the crystallization at lower @xmath83 values it is the fcc crystallization which drives the crystallization . although the energy per particle of the fcc crystal is still lower than that of the cscl , the larger fraction of the fcc crystal wins over . a cross over happens just above @xmath46 . this explains why the system at @xmath46 , whose global minima is the mixed crystal , shows higher tendency towards fcc formation . however , this does not explain why the the crystallization process when driven by fcc formation has a lower free energy cost than when driven by cscl formation . in order to understand this , we study the coordination number between the b " particles , @xmath107 in the @xmath46 system , before and after crystallization . since we can not crystallize the @xmath46 system we study the crystallization of the mka2 model ( referred earlier in the melting temperature study ) which according to dyre and coworkers is similar in structure as the ka model but with a lower viscosity @xcite . confirming their conclusion we find that the lfs of the mka2 model appears quite similar to the ka model however the dynamics is orders of magnitude faster . we now analyze the @xmath107 as obtained in the mka2 system when it is in liquid form at @xmath108 and when it forms distorted fcc crystal around @xmath109 . these are plotted in fig . [ fig9 ] . for comparison we also plot the @xmath107 for the pure cscl+fcc crystal at @xmath108 . note that the probability distribution of @xmath107 in the cscl+fcc crystal should ideally have a peak at 6 but the peak is shifted to smaller value due to the presence of a large number of surface layer of b " particles . the study shows that to form distorted fcc crystal although there is an increase in the @xmath107 it is not as much as required for the cscl+fcc crystal . thus demixing in the distorted fcc is much weaker that cscl+fcc . analysis of the same kind for the @xmath110 system ( not shown here ) shows similar behaviour . thus on the right hand side of the crossover where cscl drives the crystallization there should be free energy barriers due to demixing . however on the left hand side the system can avoid or reduce the loss of mixing entropy by paying some energetic penalty to form distorted crystals . the lower energetic stability of the distorted fcc is evident from fig.[fig1 ] . we find that for @xmath46 the disordered fcc structure melts at a lower temperature compared to the mixed cscl+fcc structure . [ h ] as obtained in the mka2 liquid at @xmath108 , the distorted fcc crystal formed by the mka2 liquid at @xmath109 and cscl+fcc crystal at @xmath108 . the cscl+fcc is formed at the same composition as the mka2 liquid . the demixing required by the cscl+fcc liquid is much higher than the distorted fcc . ] in order to strengthen our argument that it is indeed the demixing that frustrates the cscl driven crystallization process and leads to high free energy cost , we present a study of a similar system . reported in a earlier study by some of us we have shown that for the nacl system ( @xmath111 ) the crystallization takes place not only at equimolar composition but also at smaller value of @xmath83 forming mixed nacl+fcc crystal @xcite . a similar energetic study of the @xmath111 system is shown in fig.[fig10]a . we find that energy of the nacl crystal is always lower than the fcc crystal ( fig.[fig9]a ) ) . a similar crossover is also obtained for this system where at higher @xmath83 values the crystallization is driven by nacl and at lower @xmath83 values it is driven by fcc ( fig . [ fig10]b ) . thus we should expect a similar crystallization problem in this system which appears not to be the case . although the nacl and cscl systems appear quite similar there are some basic differences . the cscl crystal is made up of two interpenetrating sc structures of a " and b " type of particles . thus in the cscl+fcc crystal the a " particles have two different population one which forms sc and the other which forms fcc structure . in an earlier work we had mentioned that this wide difference in the order parameter of the two population causes the frustration between the two structures@xcite . if we are away from the equimolar mixture the growth of a cscl will deplete the population of the b " particles in the neighbourhood which should promote the formation of fcc structure between the a " particles . however a unit cell of fcc is not compatible with the cscl structure thus to reduce the structural frustration the system sacrifices the mixing entropy and increase the concentration of the b " particles in the liquid near the cluster as seen in fig.[fig6 ] to form more cscl structures till finally it is devoid of any more `` b '' particles in the liquid . this is the reason we find ab " and a " rich zone separated in fig.[fig4]a . the nacl crystal on the other hand is compatible with a fcc crystal as both require the a " particles to form fcc structure with same lattice spacing . thus unlike cscl and fcc the nacl and fcc can grow in a seamless fashion and the system does not require any demixing which reduces the free energy barrier . a snapshot of the nacl+fcc structure is shown in fig.[fig11 ] which shows that there is no specific a " rich zone . [ h ] . the simulation is done at t=0.6 . the snapshot shows seamless formation of nacl and fcc structure with no demixing . ] the study of the energetics can also explain the glass forming ability of some systems which has been previously proposed by dyre and coworkers @xcite . in the above calculation if we decrease the interaction between the ab " particles then the contribution from the cscl in lowering the system energy will decrease and the crossover will happen at a higher @xmath83 value . thus the @xmath46 will show higher tendency of fcc crystallization as has been reported earlier @xcite . in the same system if we make the interaction between the a " particles repulsive then in a similar fashion the crossover will shift to lower @xmath83 values and this will imply that the @xmath46 system will still be driven by the cscl crystallization . since this will also require demixing thus the system will be a better glass former as reported earlier @xcite . we find that the loss of mixing entropy is maximum for @xmath46 . in the free energy study within the scope of our calculation we can not grow a cscl crystal and thus can not estimate the free energy cost to grow a cscl crystal in this system . which implies that free energy cost is high and w.r.t cscl formation the @xmath46 system is most frustrated and a better glass former . however in the free energy calculation and the study of energetics this system shows a tendency towards fcc crystallization . the fcc crystallization also has a free energy cost because although without clear demixing the system can crystallize in fcc structure the presence of the lfs centered around b " particles can frustrate this crystallization process . however the cost of free energy to form a fcc crystal in @xmath46 system is lower than the cost of free energy to form a cscl crystal in @xmath51 system . thus it is tempting to comment that the @xmath51 system is a better glass former . however the process of crystallization is not only dependent on the free energy barrier but also on the dynamics of the system . this is the reason the eutectic point is expected to be a better glass forming region and our study of dynamics shows that indeed the @xmath46 system is the slowest . for the study of the dynamics we calculate @xmath48 from overlap function at the respective 0.8@xmath77 . note these are the temperatures where the free energy calculations are done . we find that for the @xmath44 system @xmath112 at @xmath113 , for the @xmath51 system , @xmath114 at @xmath115 and for @xmath46 system @xmath116 at @xmath117 . thus according to the study of the dynamics the @xmath46 system is a better glass former . note that the mka2 model which undergoes crystallization differs from the ka model not in terms of the structure of the liquid but in terms of dynamics . the local structure around the smaller b " particles which actually frustrates the fcc crystallization is present even in the mka2 model . however the relaxation timescale of the mka2 model is orders of magnitude faster than the ka model . thus our study confirms that as stated earlier@xcite it is indeed the dynamics / viscosity of the system which makes ka model a good glass former . in this article we study the comparative glass forming ability of different binary systems . in an earlier study by some of us we have shown that binary systems which form cscl crystals in a equimolar mixture fails to crystallization if the mole fraction of the larger particles are increased @xcite . the well known ka model is one of the systems . thus the ka models stability against crystallization is more generic and is similar to systems which form equimolar cscl crystal . the global structure for these systems are a mixed form of cscl+fcc crystal @xcite . in the cscl+fcc crystal the bigger a " particles need to create two different population one which contributes towards the cscl formation and the other which contributes towards the fcc formation . the order parameters such a boo and coordination number of the a " particles are quite different in these two crystal form . thus the failure to crystallize has been attributed to the frustration between the cscl and fcc crystal structure . note that there is an array of systems which have similar frustration . however the glass forming ability of these systems although have not been calculated but is believed to be different . thus there should be more factors contributing to the glass forming ability . in this article we perform a comparative study of binary glass forming liquids all having good glass forming ability and similar global minima . the study has been performed by changing the composition . we find that the free energy cost to grow a cscl nucleus increases as we move away from an equimolar mixture . the study of the liquid at the liquid / crystal interface shows that the system which has lowest free energy cost to form a nucleus also shows a demixing near the crystal surface . we believe that the structural frustration between the cscl and fcc structure makes this demixing a prerequisite for crystallization . our calculation of the partial demixing entropy in the liquid state shows a non monotonic dependence on composition . it shows a maxima for @xmath46 system . we could show a connection between the change in free energy cost to create a crystal nucleus and the change in demixing entropy as a function of composition . our study shows that although the @xmath46 system is strongly frustrated against cscl crystallization , it has tendency towards fcc growth . we can justify this tendency of fcc growth from the study of the energetics . we show that in the composition range studied here there are two regions , one which is driven by the cscl crystallization and the other at lower @xmath83 values is driven by fcc crystallization . it is primarily in the former region that the structural frustration between the cscl and fcc structure leads to the requirement of demixing which eventually increases the free energy barrier and provides stability against crystallization . this point has been confirmed by studying a nacl+fcc system which naturally undergoes crystallization at all compositions . the study of the energetics of this system also shows two similar region . however unlike the cscl+fcc system , in the region where crystallization is driven by nacl , due to the compatibility of the nacl and fcc structure no demixing has been observed and the crystal grows in a seamless fashion . in the second region driven by fcc crystallization we show that demixing is not a stringent criteria and the stability against crystallization comes from the frustration caused by the presence of the b " particles with well defined lfs and also the systems proximity to eutectic point where the dynamics is slow . thus although we study three very similar glass former , which ideally belong to the same class of system and differ only in composition , we find that they do not share the same origin of stability against crystallization . we should also comment that our search of crystal structures is not exhaustive and the system which we claim to be a better glass former can crystallize in a different crystal form like the @xmath118 structure is found to be a low energy state of a system belonging to the same class where @xmath119@xcite . this system also is known to show resistance towards crystallization . the cuzr liquid which has a low energy cscl like structure is also a good glass former @xcite . note that in these two systems the composition of the crystal is identical to that of the liquid . thus even above @xmath120 it is not always demixing which provides stability against crystallization . 53 ifxundefined [ 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
we present a comparative study of the glass forming ability of binary systems with varying composition , where the systems have similar global crystalline structure ( cscl+fcc ) . biased monte carlo simulations using umbrella sampling technique shows that the free energy cost to create a cscl nucleus increases as the composition of the smaller particles are decreased . we find that the systems with comparatively lower free energy cost to form cscl nucleus exhibit more pronounced pre - crystalline demixing near the liquid / crystal interface . the structural frustration between the cscl and fcc crystal demands this demixing . we show that closer to the equimolar mixture the entropic penalty for demixing is lower and a glass forming system may crystallize spontaneously when seeded with a nucleus . this entropic penalty as a function of composition shows a non - monotonic behavior with a maximum at a composition similar to the well known kob - anderson ( ka ) model . although the ka model shows the maximum entropic penalty and thus maximum frustration against cscl formation , it also shows a strong tendency towards crystallization into fcc lattice of the larger a " particles which can be explained from the study of the energetics . thus for systems closer to the equimolar mixture although it is the requirement of demixing which provides their stability against crystallization , for ka model it is not demixing but slow dynamics and structural frustration caused by the locally favored structure around the smaller b " particles which make it a good glass former . although the glass forming binary systems studied here are quite similar , differing only in composition , we find that their glass forming ability can not be attributed to a single phenomena .
the number of supernova remnants ( snrs ) with convincing evidence for interaction with ambient molecular clouds has increased considerably in recent years . the evidence ranges from a simple morphological relation to the detection of broad and/or shock - excited emission lines from various molecules . although circumstantial evidence could be very suggestive , it is the molecular lines from the shocked gas that are essential for understanding the physical and chemical processes associated with the molecular shock . in this regard , there are still only a few snrs adequate for the study of molecular cloud - shock interaction ; perhaps w28 ( arikawa et al . 1999 ) , w44 ( seta et al . 1998 ) , w51c ( koo & moon 1997 ) , 3c391 ( reach & rho 1996 , 1999 ) , and the classical source ic 443 ( denoyer 1979 ; tauber et al . 1994 and references therein ) . in this paper , we report the discovery of broad emission lines from the shocked co gas in hb 21 . hb 21 ( g89.0 + 4.7 ) is one of those snrs with mixed morphology , e.g. , shell - like in radio and center - filled in the x - ray ( rho & petre 1998 ) , where the center - filled , thermal x - ray emission is suggested to be due to interaction with molecular clouds . it has a nearly complete radio - continuum shell with an angular extent of @xmath9 ( hill 1974 ; tatematsu et al . 1990 , hereafter t90 ) . the shell is elongated along the northwest - southeast direction . the brightness distribution of the shell is not uniform , but enhanced in scattered areas . particularly noticeable features are the v - shaped northern boundary , a @xmath4-sized loop structure in the south central area , and the one in the central region of the eastern boundary ( see fig . 1 ) . optical nebulosity associated with hb 21 has not been detected in h@xmath10 or [ sii ] plates ( van den bergh 1978 ) . x - ray emission from hb 21 was detected by leahy ( 1987 ) and studied in detail by leahy & aschenbach ( 1996 ) . the distance to the snr is uncertain . we adopt 0.8 kpc following t90 , which is the distance to the cyg ob 7 complex ( humphreys 1978 ) . hb 21 has been a suspect for interaction with molecular clouds based on its radio appearance and the ambient molecular clouds ( erkes & dickel 1969 ; huang & thaddeus 1986 ; t90 ) . erkes & dickel ( 1969 ) suggested that the distorted boundaries with enhanced radio brightness might be the places where the snr is interacting with dense ambient gas . huang & thaddeus ( 1986 ) found that the giant molecular cloud associated with cyg ob 7 appears to be partially surrounding hb 21 ( see also dobashi et al . t90 obtained a higher - resolution ( @xmath11 ) co map of the eastern part of the snr and found that the eastern boundary of the snr appeared to be in contact with molecular and atomic clouds . they also made a coarsely - sampled map of the regions with enhanced radio emission and detected molecular clumps . but no direct evidence for the shocked _ molecular _ gas has been detected for hb 21 . the search for oh masers , which are known to be an indicator for the interaction between a snr and molecular cloud , gave negative results too ( frail et al . on the other hand , koo & heiles ( 1991 ) detected shocked h i gas moving at 40 to 120 in the southern part of the snr , although the limited angular resolution ( @xmath12 ) hindered any detailed study of the shocked gas . we detected broad co emission lines in the northern and southern parts of hb 21 , but not in the eastern part . we have found no evidence suggesting that the molecular clouds in the eastern part are interacting with the snr , although they appear to be located along the boundary of the snr on the sky . we summarize the observations and the results of molecular line observations in 2 and 3 , respectively . in 4 . we discuss the implications of our results on the interaction between the snr and molecular clouds . 5 summarizes the main results of our paper . @xmath0co j=21 line observations were carried out using the 12 m telescope of the national radio astronomy observatory at kitt peak in 1999 june and 2000 january . the fwhm of the telescope at 230 ghz was 27@xmath13 . we mapped the eastern half ( @xmath1 ) of the snr almost completely using the otf ( on - the - fly ) observing technique . we used two 256 channel filter banks ; one with 500 khz , and the other with 1 mhz resolution . we split each filter bank into two sections and observed two linear polarizations simultaneously . the velocity resolution and coverage of the 500 khz filter bank were 0.65 and 83 , while those of the 1 mhz filter bank were two times greater . typical system temperatures were 350450 k. during the observing run in january 2000 , we had some trouble because of telluric co j=21 emission , which appeared at @xmath145 . the emission was not cancelled out completely by usual position - switching observation and produced a hill - and - valley feature in spectra , the strength of which depends on elevation . we were able to avoid the contamination from telluric co emission by averaging out the contaminated velocity channels because the telluric co emission line is narrow ( 0.8 ) and the absolute strength of the ` hill ' and ` valley ' features are equal . we also obtained sensitive spectra of @xmath6co j=21 , @xmath0co j=10 and j=21 lines toward several peak positions . for co j=10 observations , we used 1 and 2 mhz filter banks , so that they have the same velocity resolution and coverage with those of co j=21 line observations . we also used the millimeter autocorrelator ( mac ) for the @xmath0co observations , which provided high resolution ( 0.25 after smoothing ) spectra . we have converted the observed temperatures ( @xmath15 ) to the main - beam brightness temperatures ( @xmath16 ) using the corrected main - beam efficiency provided by the nrao . additional observations of @xmath0co and @xmath6co j=10 lines were performed using the taeduk radio astronomy observatory ( trao ) 13.7 m telescope ( hpbw=49@xmath13 at 115 ghz ) in 2000 january and march . an sis receiver equipped with a quasi - optical sideband filter was used along with a 250 khz , 256-channel filter bank . the main beam efficiency was 0.41 at 115 ghz ( roh & jung 1999 ) , and the pointing accuracy was better than 10@xmath13 . typical system temperatures were about 750 k at 115 ghz and 450 k at 110 ghz . figure 1 shows the distribution of the integrated intensity of co j=21 emission . the velocity range is between @xmath17 and @xmath18 , which covers most of the emission . the overlaied contour map shows the 1420 mhz brightness distribution of hb 21 ( t90 ) . the overall distribution of co gas in figure 1 is not very different from the low - resolution map of t90 . but figure 1 shows much detailed structure because of its high resolution , high sensitivity , and complete sampling . figure 1 immediately shows that molecular gas is distributed mainly along the boundary of the snr . ( we observed the central area in figure 1 , which had not been covered in our co j=21 observations , in co j=10 line emission , and have detected only several small [ @xmath19 clumps other than some faint extension associated with the clouds in southern and eastern parts of the remnant . ) but the overall distribution has little correlation either with the distortion of snr boundaries or with the distribution of radio brightness ( cf . 3.2 ) . for the purpose of discussion , we divide the remnant into three areas ( fig . 2 ) : ( 1 ) eastern area ( ra@xmath20 20@xmath21 47@xmath22 ) where three relatively large ( @xmath23 ) clouds and several filamentary clouds are present , ( 2 ) northern area centered at ( 20@xmath21 46@xmath22 , 51@xmath24 00@xmath25 ) , where a small ( @xmath2 ) , very bright u - shaped cloud is noticeable , and ( 3 ) southern area centered at ( 20@xmath21 44@xmath22 , 49@xmath24 50@xmath25 ) , where clumpy and filamentary clouds with complicated structures are present . we detected broad ( 2040 ) emission lines from the clouds in the northern and southern areas , which will be discussed in detail in the next sections . in the following , we summarize the results on the eastern area . in the eastern area , there are three clouds centered at declinations @xmath26 49@xmath25 , @xmath27 15@xmath25 , and @xmath28 @xmath29 . we call these three clouds by clouds a , b , and c following t90 ( see fig . 2 for the location of these clouds ) . figure 3 shows the channel maps of the eastern area . the velocity ranges of the channel maps were chosen to show the essential features clearly . cloud a appears at @xmath30 to @xmath31 and is composed of two velocity components centered at + 6 and @xmath32 , respectively . the former component ( 6 ) , which is seen in figure 3a , is extended and the emission peaks at the southern part ( @xmath33 45@xmath25 ) of the cloud , while the latter component ( @xmath32 ) , which is seen in figures 3b and 3c , is spatially confined and comprises the northern part of the cloud . their maximum brightnesses are @xmath34 and 7 k , respectively . cloud b appears at @xmath35 to @xmath36 and is seen in figures 3c and 3d . the south central part of the cloud , e.g. , the region between @xmath3717@xmath25 , is bright and appears to be connected to cloud c. cloud c has two components at very different velocities , e.g. , one at + 2 to @xmath38 , which is seen in figures 3c and 3d , and the other at + 17 to + 10 , which is not shown in figure 3 but has a distribution similar to the other velocity component . according to the result of t90 , cloud c extends to @xmath39 . an interesting feature in figure 3 is the semi - circular loop that appears above cloud b in figure 3b . the ratio of the minor , which is along the ns direction , to the major axis is 0.8 . if it is at 0.8 kpc , the linear size of the semimajor axis would be @xmath40 pc . the velocity increases systematically from both ends to the northern top of the loop , which is consistent with an expanding loop . the top portion is redshifted with respect to the both ends by @xmath41 . if we assume that the ellipticity is due to projection , then the expansion velocity would be @xmath42 , so that the dynamical age of the ring is probably shorter than @xmath43 yrs . this is much greater than the age ( @xmath44 yrs , koo & heiles 1991 , scaled to 0.8 kpc adopted in this paper ) of hb 21 and , therefore , the loop might not be associated with hb 21 . we suspect that the loop is originated from some energetic phenomena in cloud b. a faint v - shaped structure that connects cloud b and the ends of the loop in figure 3b seems to indicate the association of the two . the cloud in the northern area is composed of a small ( @xmath2 ) , very bright , u - shaped cloud and several clumps scattered around it ( fig . 1 ; see also fig . 9 for enlarged view ) . figure 4 shows its velocity structure . there are several points to be made from figure 4 : first the u - shaped cloud is composed of several clumps , whose central velocities shift systematically from + 3 to @xmath31 as we move from ne to nw along the structure . the integrated intensity attains a maximum at ( 20@xmath21 46@xmath22 [email protected] , 51@xmath24 00@xmath25 00@xmath13 ) , which we call hb21:bml - n1 ( broad molecular line northern position 1 ) , or simply n1 . second , there are several other clumps in the field . these clumps , except the one near the southeastern corner at @xmath46 , appear over a wide ( @xmath47 ) velocity range . among them , the one at ( 20@xmath21 45@xmath22 55.@xmath450 , 51@xmath24 03@xmath25 30@xmath13 ) , which we call hb21:bml - n2 ( or n2 ) , appears over the widest ( @xmath48 ) velocity interval . third , there is a diffuse emission at @xmath49 to the ne of the cloud . its line is narrow ( 23 ) and it is part of a large ( @xmath50 ) cloud that appears to be connected to cloud a. we consider that the clumps aligned along the ne - sw direction in figure 4 are associated and call them cloud n , i.e. , cloud n does not include the diffuse emission in the northeastern area and the clump in the southeastern corner . t90 detected in this area only the diffuse molecular gas at @xmath51 to 6 and called it cloud d. ) as can be expected from figure 4 , most clumps in cloud n have broad emission lines . as an example , we show the spectra of n1 and n2 in the top frames in figure 5 , where we see that the spectrum of n1 is box - shaped and its full width ( at zero intensity ) is 30 km / s , while that of n2 is asymmetric and extends from @xmath52 to + 11 km / s . for comparison , the spectrum of the diffuse , extended structure in the northeastern part of this area has narrow ( 23 ) emission lines centered at @xmath49 , a sample of which is shown in the right bottom frame in figure 5 . we have obtained sensitive j=10 and j=21 spectra of @xmath0co and @xmath6co molecules at the two peak positions , n1 and n2 , and figure 5 shows the spectra . the molecule and transition are marked in each spectrum . the second spectrum from the top is @xmath0co j=21 emission convolved to the j=10 beam size ( @xmath53 ) to be compared with the j=10 spectra . the difference between the top and convolved spectra indicates that some velocity components , e.g. , the narrow component centered at + 4 of n1 and the broad component at @xmath54 of n2 , are confined to small areas . by comparing the j=10 and the convolved j=21 spectra , we notice that the ratio of j=21 to j=10 intensities is high and that it varies over the profile : for n1 , the ratio is between 1.2 and 2.0 in the central parts of the spectrum , while it increases at the wings , e.g. , @xmath55 at + 5 and @xmath5613 . for n2 , the ratio varies between 0.8 and 2.8 , and it is higher between @xmath38 and 0 . the ratios of the @xmath0co j=21 and j=10 integrated intensities @xmath57 are 1.6 and 1.7 for n1 and n2 , respectively . the @xmath6co j=21 line is clearly detected toward n1 , while it is marginally detected toward n2 . for n1 , the line has double peaks centered at + 1 and @xmath36 , while the @xmath0co j=21 line profile toward n1 is composed of several narrow peaks . the narrow peaks might indicate that the emission is from several , unresolved subclumps . presumably , the @xmath6co emission might be from these subclumps too , which is not apparent in the profile in figure 5 because of low signal - to - noise ratio and low velocity resolution . ( note that the velocity resolutions of the @xmath0co j=2 - 1 and @xmath6co j=2 - 1 lines are 0.25 and 0.68 , respectively . ) the ratios of @xmath0co j=21 to @xmath6co j=21 integrated line intensities @xmath58 are @xmath59 and @xmath60 for n1 and n2 , respectively . ( the errors are statistical errors . ) table 1 summarizes the line parameters of the peak positions , i.e. , their coordinate , velocity range ( @xmath61 , @xmath62 ) , co j=21 peak brightness temperature @xmath63 , @xmath57 , @xmath58 , and the ratio of @xmath0co j=10 to @xmath6co j=10 integrated line intensities @xmath64 . the detailed line diagnostics based on the observed line parameters in table 1 is discussed in 3.4 . in the southern part of the snr , the emission is detected at @xmath65 to @xmath66 . the velocity structure is shown in figure 6 . at positive velocities , we see several clouds with narrow lines come and go , e.g. , a diffuse cloud that extends @xmath67 along the ns direction centered at ( 20@xmath21 [email protected] , 50@xmath24 00@xmath25 ) between @xmath68 and + 5 . the co distribution at negative velocities is fairly complicated : the distribution is filamentary , and small ( @xmath5 or 0.3 pc ) , bright clumps are seen along the filamentary structure . the filamentary structure , which we call cloud s , appears to form a loop of @xmath69 in extent , elongated along the ns direction . the eastern part of the loop is particularly clumpy and has a semicircular shape ( see the channel map centered at @xmath70 ) . the clumps generally have broad ( @xmath71 ) lines . among them , three clumps marked by crosses in figure 6 have broadest ( 3040 ) lines and we show the @xmath0co j=21 and j=10 spectra at their peak positions , which we call s1 , s2 , and s3 from east to west ( see fig . 2 ) , in figure 7 . again we show the convolved j=21 spectra together , although the line shapes do not change significantly by convolution toward these peak positions . we have obtained some sensitive j=10 and j=21 spectra of @xmath0co and @xmath6co molecules at these peak positions , and the line parameters are listed in table 1 . note that @xmath57(=1.72.3 ) and @xmath72 toward s1 ) are similar to those of northern positions , while @xmath64 s , although they have large uncertainties , appear to be much greater than that of n1 . the broad co lines are presumably emitted from the shocked gas , where physical parameters vary greatly over a short distance scale . but still it would be worthwhile to estimate their excitation parameters based on elementary considerations . first , the observed @xmath7340 are significantly less than either the average ratio ( @xmath74 ; langer 1997 ) of @xmath0c/@xmath6c in the solar neighborhood or the terrestrial value ( 89 ) , implying that the @xmath0co j=21 lines are _ not _ optically thin . if we adopt @xmath0c/@xmath6c=67 and assume that the emission is thermalized , then the optical depth for the @xmath0co j=21 line @xmath753.4 . on the other hand , the @xmath64(@xmath76 ) close to or greater than the terrestrial value imply that @xmath0co j=10 lines are optically thin , except at n1 where j=10 and j=21 lines appear to have comparable optical depths . this , however , is not conclusive because of large uncertainties associated with @xmath64 . second , the large values of @xmath772.3 imply that the broad - line emitting region is warm and dense . for typical molecular clouds , where the j=2 level is subthermally excited , the ratio is usually less than unity . for example , molecular clouds in the local arm exhibit ratios ranging from 0.53 ( taurus ) to 0.75 ( orion a ) ( sakamoto et al . 1994 , 1997 ) . our spectra of ambient gas also show this , e.g. , see the spectra toward s2 in fig . 7 where the narrow component at + 2 has @xmath78 . but the large ratio is common for the shocked molecular gas in snrs ( see 4.1 ) . third , the low ( @xmath79 k ) brightness temperature of j=21 lines , regardless of their moderate optical depths , imply that the emitting region must be clumpy , i.e. , composed of subclumps , and the emission is beam - diluted . from the above considerations , we may conclude that the broad emission lines are from warm , dense clumps with significant column densities so that the 21 lines are optically thick . we have applied the large - velocity - gradient ( lvg ) model ( scoville & solomon 1974 ; goldreich & kwan 1974 ) to our broad co lines in order to derive their excitation parameters . the model assumes an uniform , spherical cloud with a constant velocity gradient ( @xmath80 ) . if the lines are emitted from the shocked region where temperature and density vary greatly , the resulting parameters may be considered as ` average ' values . since we have found that the emission is beam - diluted , we have used the line ratios , @xmath57 and @xmath58 , instead of brightness temperatures to determine the excitation parameters . according to our lvg analysis , the observed ratios are possible for @xmath81 k. figure 8a shows the result of our model computations when @xmath82 k , where curves of constant @xmath57 and constant @xmath58 are drawn in ( @xmath83 , @xmath84 ) plane . @xmath83 is the fractional abundance of @xmath0co relative to h@xmath85 ( @xmath86 ) per unit velocity gradient interval . the asterisks ( * ) mark the observed ratios toward the peak positions where both ratios are obtained , i.e. , n1 , n2 , and s1 . according to figure 8a , @xmath87@xmath88 @xmath89 and @xmath90@xmath91 pc ( ) @xmath92 at the three peak positions . there are multiple choices for n1 , e.g. , the same ratios are obtainable when @xmath93 @xmath89 and @xmath94 pc ( ) @xmath92 . we adopt the lower density because it is comparable to the densities in the other peak positions and because the density of @xmath95 @xmath89 appears to be too high for the co emission to explore . if the temperature becomes higher , both @xmath84 and @xmath83 need to be greater . in figure 8b , we plot the _ expected _ co j=21 radiation temperature @xmath96 $ ] which is just the brightness temperature when @xmath97 and the expected @xmath64 from the same lvg model . note that the expected radiation temperatures are much greater than the observed main - beam brightness temperature . we have estimated beam filling factors of ( 7.78.8)@xmath98 from the ratio of these two brightnesses . we have estimated the co column densities @xmath99@xmath100 @xmath101 at these peak positions by @xmath102 \delta v$ ] , where @xmath10318 is the velocity width . the excitation parameters derived from the lvg analysis are listed in table 2 . note that the @xmath64 expected from the lvg model differ from the observed ones : at n1 , the observed value is small by a factor of 2 , while , at s1 , it is large by a factor of @xmath104 . considering the weakness of @xmath6co j=10 lines and various uncertainties associated with different telescopes , however , it is not obvious if this difference is critical . we have made a crude estimate of the mass of the broad - line clouds as follows . if the co j=10 line emission is optically thin , then @xmath105 can be obtained from the co j=10 luminosity @xmath106 by @xmath107 $ ] where @xmath108 is the fraction of co molecules at @xmath109 level and the other coefficients have their usual meanings . in our case , co j=10 emission has less optical depth than the j=21 emission , but is not very optically thin , so that the above formula might yield an underestimate . what we have is the luminosity of co j=21 emission @xmath110 , which has moderate optical depth . but , since @xmath772.3 at the peak positions , we may obtain @xmath106 by assuming that @xmath111 where @xmath112 and @xmath113 are co @xmath11410 and @xmath11421 line frequencies respectively . finally we assume @xmath115 , which is a mean value of those ( 0.150.26 ) at the three peak positions obtained from the lvg analysis . we have found that the h@xmath85 masses of clouds n and s are @xmath116 and @xmath117 , respectively . the mass of the central u - shaped part of cloud n is @xmath118 while the masses of the small clumps in cloud s are @xmath1191 . broad co emission lines with large @xmath57 in clouds n and s strongly suggest that they are being shocked . the observed velocity width is as large as @xmath120 . note that , toward this direction ( @xmath121 ) , the lsr velocity permitted by the galactic rotation is @xmath122 , so that it is not impossible for broad lines to be produced by molecular clouds accidentaly aligned along the line of sight . but it is highly improbable that such alignment ( over a few kpcs ) occurs in very small ( 12@xmath25 ) areas on the sky . we also searched for protostellar candidates around the broad line emitting regions using the _ infrared astronomical satellite _ ( _ iras _ ) point source catalog , because broad lines can be emitted from the high - velocity gas associated with protostellar object too . we did not find any suspicious sources except one , iras 20444 + 4954 , which is located close to the s2 clump , i.e. , at ( @xmath123 , @xmath124 ) from the peak position in table 1 . the source has been detected in two iras wavebands , i.e. , 60 and 100 @xmath125 , with flux densities of @xmath126 jy and @xmath127 jy . we have found that the source is located within a small ( @xmath128 ) , bright ( @xmath129 k ) co j=21 core at @xmath130 , so that it might be a young stellar object associated with the @xmath67-sized , diffuse cloud in the northern part of the + 2.9 map in figure 6 , not with the s2 clump . also the velocity of the s2 clump is similar to those of the other fast - moving clumps in this area , which suggests that they have the common origin . therefore , the broad lines that we detected are almost certainly from the fast - moving molecular gas swept - up by the snr shock in hb 21 . another indication that the broad co lines are from the shocked gas is their high @xmath772.3 . as we have shown in 3.4 , the high ratio implies that the emitting gas is warm and dense , which might be manifestation of shock . indeed high @xmath57 is a common property of the broad lines from the shocked molecular gas in snrs : all six snrs that are known to have broad molecular emission lines , i.e. , w28 ( arikawa et al . 1999 ) , 3c391 ( reach & rho 1999 ) , w44 ( seta et al . 1998 ) , w51c ( koo & moon 1997 ) , hb 21 ( this paper ) , and ic 443 ( e.g. , van dishoeck et al . 1993 ) , have ratios greater than 1 , which implies that the broad co lines in these snrs are all emitted from warm and dense , shocked gas . meanwhile , the maximum brightness temperature ( @xmath131 k ) of co j=21 line in hb 21 is significantly less than those of other snrs even if it was obtained with a higher spatial resolution ( 0.1 pc ) , e.g. , it is 33 k for w28 ( reach & rho 2000b ) and ic 443 ( van dishoeck et al . 1993 ) when observed with a resolution of 0.2 pc . the much smaller co j=21 brightness temperature with comparable @xmath57 imply that either the shocked gas in hb 21 is composed of much smaller clumps or is less dense . on a large scale , hb 21 appears to be in contact with a giant molecular cloud ( gmc ) along its eastern boundary ( huang & thaddeus 1986 ; tatematsu et al . the gmc is 130 pc@xmath13270 pc in extent ( dobashi et al . 1994 ) , and the structure that we call clouds a , b , and c defines the western boundary of the gmc . t90 inspected the correlation between these clouds and the snr in detail , and concluded that cloud a might be interacting with the snr because it is located where the radio continuum boundary of the snr is distorted . on the other hand , they concluded that clouds b and c might not , because there is no indication of the interaction in the distribution of radio brightness . according to our high - resolution observations , however , there is little relationship between the boundaries of cloud a and the snr . instead , since the velocity of the ambient molecular gas around hb 21 might be negative ( see 4.3 ) , we consider that clouds b and c have a better chance of interaction . but we have detected broad co lines toward none of these eastern clouds . even if a snr is interacting with a molecular cloud , the broad lines may be absent , however . 3c 391 ( g31.9 + 0.0 ) , which appears to be located at the edge of a large molecular cloud , for example , has no broad co lines along the interface ( wilner , reynolds , & moffett 1998 ; reach & rho 1999 ) . but strong [ oi ] 63 @xmath133 m emission has been detected near the interface , which indicates that the snr is interacting with the molecular cloud ( reach & rho 1996 ) . this would happen if the shock is dissociative and molecules have not reformed . figure 9 shows an enlarged view of the northern area overlaied with a 325 mhz radio continuum map of the snr . note that there is an enhanced radio continuum emission elongated along the ne - sw direction in the central area . its peak position falls exactly inside the u - shaped part of cloud n. the positional coincidence suggests that the enhancement is possibly associated with cloud n. we have derived a spectral index ( flux@xmath134 ) of @xmath135 for the radio emission associated with the u - shaped central part using the 325 mhz and 1420 mhz maps . the derived index has a large uncertainty because of the confusing background " level , but it appears to be flatter than the mean spectral index @xmath136 of hb 21 ( willis 1973 ) . for cloud s , there is no obvious correlation between the co emission and radio continuum brightness , although the radio continuum appears to be bright around the cloud in general . it is not obvious , observationally or theoretically , what determines whether or not radio synchrotron emission becomes enhanced when sn shock hits a dense cloud . observationally , we see very limited correlation between radio continuum brightness and shocked molecular gas in some snrs , i.e. , the shocked molecular gas is not usually associated with radio continuum enhancement and vice versa ( see also chevalier 1999 ) . in ic 443 , for example , shocked molecular gas is distributed in a fragmentary , flattened , ring , which partly overlaps with the radio continuum shell ( e.g. , see dickman et al . 1992 for the shocked molecular gas and green 1986 for the radio continuum ) . but the radio continuum is brightest in the northeastern part of the shell where there is no shocked molecular gas , and the radio continuum is not particularly bright toward the shocked molecular gas , perhaps except around the southern part of the molecular ring ( see next ) . in 3c 391 , there is a shocked molecular clump in the southern part of the snr , but there is only a faint , local radio continuum peak at @xmath137 pc apart from the shocked clump , whose association can not be confirmed ( reach & rho 1999 ) . in w28 , on the other hand , there is a ridge of radio continuum emission associated with the shocked molecular gas ( arikawa et al . theoretically , it could be either the dense cloud or the surrounding intercloud medium where synchrotron emission becomes enhanced . if the shock propagating through the dense cloud is radiative , there will be a large compression of cosmic rays and magnetic field , which would increase the synchrotron emissivity ( van der lann 1962 ; blandford & cowie 1982 ) . but this mechanism may not work because the molecular shocks in old snrs are not ionizing shocks and , therefore , high energy particles may escape from the shocked region ( draine & mckee 1993 ; chevalier 1999 ) . on the other hand , the shocked intercloud medium surrounding the cloud , particularly the medium behind the cloud , could have enhanced synchrotron emissivity because of the increased magnetic field strength there ( e.g. , jones & kang 1993 ; mac low et al . 1994 ) . in hb 21 , the peak of the enhanced radio emission is located behind the shocked cloud ( fig . it is also noteworthy that the cloud has a u - shape , which is similar to what we would expect when a small cloud is swept up by a strong shock ( e.g. , klein , mckee , & colella 1994 ; mac low et al . these morphological characteristics seem to suggest that the enhanced emission is not physically associated with the shocked cloud but with the shocked surrounding intercloud medium . in some respects , the region around cloud n in hb 21 is similar to the flat ( @xmath138 ) spectral region in ic 443 , which is also located behind a shocked molecular cloud ( green 1986 ; keohane et al . 1997 ) . in ic 443 , keohane ( 1997 ) found that the flat spectral region is particularly bright in hard x - rays , which are most likely due to synchrotron radiation . they concluded that the enhanced hard x - ray emission and the flat spectral index is due to shock acceleration of cosmic rays behind dense clouds . more observational study is certainly needed for hb 21 in order to reveal the relationship between the northern u - shaped cloud and the enhanced radio emission . cold molecular gas around hb 21 in general has central velocities between @xmath139 and + 6 , which must be the velocity range of the preshock gas . in the areas where broad lines have been detected , there is diffuse molecular gas at positive ( 0 to + 5 ) velocities ( see figures 4 and 6 ) . if this gas represents the preshock gas , then , since the broad lines are centered and spread out mostly at negative velocities , the gas should have been shocked and accelerated toward us _ systematically_. if we take + 3 as the velocity of the preshock gas and if we take either the central or peak positions of the broad lines as the systematic velocities of the shocked gas , then _ the line - of - sight _ velocities of the shock would be @xmath140 and @xmath8 for clouds n and s , respectively . and , if we take @xmath141 and @xmath142 ( @xmath143=the radius of the snr ) as their projected distances from the center of the snr , then their deprojected velocities are @xmath144 and @xmath145 , respectively . such coasting clumps are indeed theoretically expected for the old snrs such as hb 21 : molecular clumps swept up by a snr blast wave are accelerated by the shock propagating into the clumps and also by the ram pressure of interclump gas ( e.g. mckee 1988 ) . the characteristic timescale for acceleration is @xmath146 where @xmath147 is the radius of the clump , which becomes @xmath148 yrs ( @xmath149 yrs ) for the clumps in hb 21 using @xmath150 pc and @xmath15125 . the wings of the broad lines may be attributed to the gas accelerated by shocks propagating from sides . but one difficulty in this scenario would be that there is no clear correlation in the distributions of the broad lines between the preshock and postshock gases . in the southern area , for example , the preshock gas ( 23 component ) is distributed in a filamentary cloud extended along the ns direction , while the broad - line emitting clumps are distributed over a much wider area to the sw of this cloud . also it appears rather awkward that there are no broad lines centered near the velocity of the preshock gas . alternatively , it could be the molecular gas at negative velocities that is associated with the snr , and , in the northern and southern areas , most , if not all , of the ambient gas may have been shocked . in this picture , the shocked gas may be coasting too , but not necessarily at high speeds because the velocity of the preshock gas is somewhere within the velocity range of the broad lines , e.g. , near the velocity of the peak . the shock velocity determined from the line width is @xmath8 . between the two possible interpretations , we prefer the latter because of the difficulty with the former mentioned above . but , since the difference in the shock velocities between the two interpretations is @xmath140 , the following discussion remains valid basically even if the velocity of the preshock gas is 0 to + 5 . ( the derived properties of the shocked gas in 3.4 should remain valid too in either case because the emission from the preshock gas might be in narrow lines and its contribution to the integrated intensity of broad lines is expected to be small . ) the _ observed _ shock velocity is @xmath8 . this is less that the critical velocity for the dissociation of molecules , which is 2550 depending on preshock density and magnetic field strength ( hollenbach & mckee 1980 ; draine , roberge , & dalgarno 1983 ) . hence , the shock might be a _ non - dissociating c - shock_. the observed integrated intensity of the co j=21 emission is ( 39)@xmath152 ergs @xmath101 s@xmath92 sr@xmath92 at the peak positions . if we consider the beam dilution ( 3.4 ) , the actual surface brightness may be greater by an order of magnitude , e.g. , ( 410)@xmath153 ergs @xmath101 s@xmath92 sr@xmath92 . this is much larger than the _ angle - averaged _ surface brightness predicted from shock model computations . draine & roberge ( 1984 ) , for example , computed surface brightnesses expected for steady - state c - shocks propagating through molecular gas with different preshock conditions . according to their result , the angle - averaged surface brightness of co j=21 emission varies from @xmath154 to @xmath155 ergs @xmath101 s@xmath92 sr@xmath92 for a 1020 shock propagating through a molecular cloud with @xmath156 to @xmath157 @xmath89 . larger shock velocity does not raise the surface brightness while higher preshock density may yield @xmath158 ergs @xmath101 s@xmath92 sr@xmath92 . the much higher surface brightness toward the peak positions would be possible if these are directions where we are observing the shock tangentially . we want to briefly discuss the non - detection of oh 1720 mhz masers in hb 21 , because such masers are known to indicate interaction of snrs with molecular clouds ( frail et al . firstly , the oh masers may require a very specific set of physical conditions that might not be realized in hb 21 . according to lockett , gauthier , & elitzer ( 1999 ) , the 1720 mhz masers arise only in c - shocks when @xmath159 - 125 k , @xmath160 , and oh column density of @xmath161@xmath162 @xmath101 . according to our result in 3.4 , the density of the shocked molecular gas in hb 21 appears to be much lower than required . secondly , it is not impossible that the oh maser emission , even if present , had been missed in the survey by the frail et al . ( 1996 ) who mapped the snrs in rectangular grids with full - beam ( or @xmath163 full - beam ) grid spacing . it would be worthwhile to search for oh masers toward the shocked co gas in hb 21 . koo & heiles ( 1991 ) detected shocked h i gas associated with hb 21 . the shocked h i gas moves at @xmath164123 and is confined to the southern part ( @xmath165 0@xmath2550@xmath24 30@xmath25 ) of the snr . the highest velocity component coincides with cloud s , although the angular resolution ( @xmath12 ) of the hi observation is too large for a detailed comparison . koo & heiles ( 1991 ) assumed that the shocked h i gas represents a cap portion of a large expanding h i shell and derived a mean _ ambient _ h i density of 3.7 @xmath89 ( when scaled to 0.8 kpc adopted in this paper ) . if the molecular shock has been driven by this h i shell , then we can roughly estimate the density of the shell as follows : we assume that the shocked molecular gas is confined to a thin slab and that the radiative h i shell has an uniform density of @xmath166 . then it is straightforward to show that @xmath166 is related to the density of the molecular cloud @xmath167 by @xmath168 ( for @xmath169 ) where @xmath170 and @xmath171 are the velocities of radiative shell and shocked molecular slab , respectively ( e.g. , see chevalier 1999 ) . for hb 21 , @xmath172 and @xmath173 . and if we take the h@xmath85 density of the cloud @xmath174 @xmath89 , the density of the h i shell would be @xmath175 @xmath89 . by comparing with the mean ambient density ( 3.7 @xmath89 ) , this implies a compression factor @xmath176 for the h i shell . such moderate compression would be obtained if the ambient magnetic field strength tangential to the shell is @xmath177 @xmath133 g where @xmath178 is the density of the ambient medium . ( the equation is obtained by assuming that ambient magnetic field is uniform and magnetic pressure dominates the pressure in the shell . see chevalier 1974 for a discussion . ) alternatively , the fast - moving h i gas could be the gas originally associated with the molecular clouds , i.e. , the atomic and molecular shocks may be produced when the snr shock hits a large molecular cloud . in this case the fast - moving h i gas represents the swept - up interclump medium or h i envelope of molecular cloud . high - resolution h i observation is needed to reveal the relation between the shocked atomic and molecular gases . we used archival data from _ iras _ to search for infrared emission associated with the remnant . in his catalog of infrared emission from supernova remnants , arendt ( 1989 ) called hb 21 a `` probable '' infrared source at 12 and 60 @xmath133 m , with total fluxes of @xmath179 and @xmath180 jy , respectively . the main source of uncertainty is confusion with unrelated emission in the galactic plane , which can not be easily separated in infrared images . using the _ iras _ _ sky survey atlas _ ( _ issa _ ; wheelock et al . 1993 ) , we created an image covering the region around hb 21 at 60 and 100 @xmath133 m . there is extensive emission to the south , east , and west of the remnant , but with no clear correlation with the radio or co image . the region toward the center of the remnant is relatively fainter than these edges , but the northern part of the remnant is also faint . it is not possible to tell whether the emission is related to a partial shell around the remnant or just fluctuations in the background emission . to search for infrared emission associated with the remnant in more detail , we obtained a dedicated _ iras _ hires ( aumann , fowler , & melnyk 1990 ) image for a @xmath181 field centered on hb 21 . hb 21 is visible in the _ iras _ hires images at all four wavelengths ( 12 , 25 , 60 , and 100 @xmath133 m ) . the 60 @xmath133 m image , where hb 21 is most prominent , is shown in figure 10 . comparing the _ iras _ and co images , it is evident that the southern filamentary cloud , cloud s , is detected as a long arc , with very similar location , shape and width . the 60 @xmath133 m surface brightness of the filament is typically 7 mjy sr@xmath92 , and its structure is clumpy , like that of the co emission . but the peaks of co and infrared emission do not match in detail , suggesting that the 60 @xmath133 m emission does not arise from the exact same regions as the co. in the northern area , there is also a good correspondence between the infrared and co emission . the general correspondence in the south and north , and partial overlap in the east , show that many of the infrared features around the edge of hb 21 are related to the remnant , although the infrared emitting regions differ in detail from the co emitting regions . the nature of the infrared emission from hb 21 could be either dust grains surviving the shock , or from spectral lines from shock - excited gas , or both . we have estimated the mean surface brightness and color of the co clouds by using several faint regions in the field as background . the results are summarized in table 3 . the far - infrared color ratio , @xmath182 , for clouds a , b , and c is almost identical to that of diffuse cirrus clouds in the solar neighborhood ( boulanger & prault 1988 ) . this suggests that the infrared emission from clouds a , b , and c is most likely due to dust heated at the surface of the molecular clouds by the interstellar radiation field ; specifically , it suggests that the infrared emission is not related to shock fronts into the clouds . on the other hand , clouds n and s have a significantly higher color ratio @xmath183 . this enhanced color , and the morphological correspondence with the broad molecular line emitting regions , suggests that the infrared emission from clouds n and s is due to shocks in propagating into the clouds . conversely , the normal infrared color of clouds a , b , and c is consistent with their being due to ambient molecular clouds . if the infrared emission from clouds n and s is due to dust , then the relatively higher @xmath184 could be due to smaller or warmer dust grains . for dust heated by the average interstellar radiation field in the solar neighborhood , about half of the emission at 60 @xmath133 m is thought to be due to small , transiently heated grains ( draine & anderson 1985 ; dsert , boulanger , & puget 1990 ) . thus , if the enhanced 60 @xmath133 m emission is due to dust grains , then clouds n and s may contain a larger fraction of small grains . an enhanced abundance of small grains would be expected if a significant fraction of larger grains were shattered behind the shock front . observations of local cirrus clouds with significant velocities revealed that @xmath185 is typical for clouds with @xmath186 km s@xmath92 , suggesting that grain shattering was significant in the shocks that accelerated local clouds to intermediate velocities ( heiles , reach , & koo 1988 ) . theoretically , significant shattering is not predicted for slow shocks such as inferred from the widths of the co lines , but faster shocks through somewhat lower - density interclump gas , with @xmath187 km s@xmath92 , could produce the enhanced 60 @xmath133 m emission ( jones , tielens , & hollenbach 1996 ) . spectral lines could contribute significantly to the infrared emission from clouds n and s. the most important lines in the _ iras _ passbands , based on infrared spectra of similar supernova remnants ( oliva et al . 1999 , cesarsky et al . 1999 , reach & rho 2000a ) , are [ ] 88 @xmath133 m in _ iras _ band 4 , [ ] 63 @xmath133 m in band 3 , [ ] 26 @xmath133 m in band 2 , and [ ] 12.8 @xmath133 m and h@xmath85 lines in band 1 . if we were to interpret all of the _ iras _ emission from cloud s as due to the ionic lines listed above , then , using the system response and bandwidths ( beichman et al . 1988 ) , we find that the brightest line would be [ ] 88 @xmath133 m , with intensity @xmath188 erg s@xmath92 @xmath101 sr@xmath92 . relative to this line , the other bright lines would have ratios @xmath18963/88=0.27 , @xmath18926/88=0.13 , and @xmath18912.8/88=0.26 . the implied brightness of the [ ] 63 @xmath133 m line can be easily produced by shocks with velocities @xmath190 km s@xmath92 into moderate - density ( @xmath191 @xmath89 ) gas ( hollenbach & mckee 1989 ) ; however , such shocks do not produce as much [ ] emission as observed , because the column density of highly - ionized gas is insufficient . slower shocks into denser gas can also produce [ ] 63 @xmath133 m lines this bright ( draine , roberge , & dalgarno 1983 ) . however , slower ( c - type ) shocks produce essentially no ionic line emission especially not an ion such as . nor would the slow shocks destroy grains adequately to produce significant gas - phase . therefore , if the infrared emission is from slow shocks , the _ iras _ 100 @xmath133 m band emission is from dust , and the _ iras _ 60 @xmath133 m band emission is from a mix of dust and the [ ] 63 @xmath133 m line . the nature of the _ iras _ 12 and 25 @xmath133 m band emission is more difficult to constrain . if the infrared emission is from slower shocks , such as inferred from the co observations , then there is likely a contribution from h@xmath85 lines . rotational lines are the dominant coolant for a range of molecular shocks ( rho et al . 2000 , reach & rho 2000a ) . for now , it is not possible to clearly tell what fraction of the infrared emission is from gas or dust . the nature of the infrared emission from hb 21 ( and other supernova remnants ) can be determined in the future using spectroscopy or narrow - band imaging . we have mapped the eastern half ( @xmath192 ) of the snr hb 21 in @xmath0co j=21 line emission almost completely . our map , which has been completely sampled with @xmath193 resolution , shows the detailed structure of molecular clouds in this area . we have detected broad co lines with large @xmath57 in the northern and southern parts of hb 21 , which is direct evidence for the interaction between molecular clouds and the snr . in the following , we summarize the main results of this paper : \(1 ) we detected shocked molecular clouds , clouds n and s , with broad ( 2040 ) co lines in the northern and southern parts of the snr . cloud n is composed of a small ( @xmath2 or 0.5 pc ) , very bright , u - shaped , clumpy part and several clumps scattered around it . cloud s is filamentary and appears to form an elongated loop of @xmath4 in extent . small ( @xmath5 or 0.3 pc ) , bright clumps are seen along the filamentary structure . the h@xmath85 masses of clouds n and s are @xmath116 and @xmath117 , respectively.(2 ) we have obtained sensitive j=10 and j=21 spectra of @xmath0co and @xmath6co molecules toward several peak positions of clouds n and s. they have @xmath772.3 and @xmath7340 with co j=21 main - beam brightness temperature less than 7 k. according to our lvg analysis , @xmath81 k , and , for @xmath82 k , @xmath194(37)@xmath195 @xmath89 and @xmath196(2.411 ) @xmath197 @xmath101 . the emitting region appears to fill a small ( 0.0770.088 ) fraction of the beam . ( 3 ) there is an enhanced radio emission which attains a maximum exactly inside the central u - shaped part of cloud n. the emission has a spectral index ( @xmath198 ) flatter than that of the whole remnant . the association of this emission with cloud n needs to be explored.(4 ) clouds n and s are visible in the _ iras _ hires images at all four wavelengths ( 12 , 25 , 60 , and 100 @xmath133 m ) . they have the far - infrared color ratio @xmath183 , which is significantly greater than that ( 0.20 ) of the other clouds in this area . this enhanced color , and the morphological correspondence with the broad molecular line emitting regions , suggests that the infrared emission from clouds n and s is due to shocks in propagating into the clouds.(5 ) along the eastern boundary of the snr , three relatively large ( @xmath23 or 3.5 pc ) clouds and several filamentary clouds are present . no broad co emission or enhanced 60/100 @xmath133 m color were detected in any of these clouds , and there is little relationship between the boundaries of the clouds and the snr . therefore , there is no strong evidence for the interaction of the snr with molecular clouds along the eastern boundary . arendt , r. g. 1989 , apjs , 70 , 181 arikawa , y. , tatematsu , k. , sekimoto , y. , & takahashi , t. 1999 , pasj , 51 , l7 aumann , h. h. , fowler , j. w. , & melnyk , m. 1990 , aj , 99 , 1674 beichman , c. a. et al . 1988 , _ infrared astronomical satellite ( iras ) catalogs and atlases : volume 1 . explanatory supplement _ , nasa rp-1190 ( nasa : washington , dc ) blandford , r. d. , & cowie , l. l. 1982 , apj , 260 , 625 boulanger , f. , & prault , m. 1988 , apj , 330 , 964 cesarsky , d. , cox , p. , pineau des forts , g. , van dishoeck , e. f. , boulanger , f. , & wright , c. m. 1999 , a & a , 348 , 945 chevalier , r. a. 1974 , apj , 188 , 501 1999 , apj , 511 , 798 denoyer , l. k. 1979 , apjl , 232 , l165 dsert , f. x. , boulanger , f. , & puget , j .- l . 1990 , a&a , 237 , 215 dickman , r. l. , snell , r. l. , ziurys , l. m. , & huang , y .- l . 1992 , apj , 400 , 203 dobashi , k. , bernard , j .- yonekura , y. , & fukui , y. 1994 , apjs , 95 , 419 draine , b. t. , & anderson , n. 1985 , apj , 292 , 494 draine , b. t. , & mckee , c. f. 1993 , araa , 31 , 373 draine , b. t. , roberge , w. g. , & dalgarno , a. 1983 , apj , 264 , 485 draine , b. t. , & roberge , w. g. 1984 , apj , 282 , 491 erkes , j. w. , & dickel , j. r. 1969 , aj , 74 , 840 frail , d. a. , goss , w. m. , reynoso , e. m. , giacani , e. b. , green , a. j. , & otrupcek , r. 1996 , aj , 111 , 1651 goldreich , p. , & kwan , j. 1974 , apj , 189 , 441 green , d. a. 1986 , mnras , 221 , 473 heiles , c. , reach , w. t. , & koo , b .- c . 1988 , apj , 332 , 313 hill , i. e. 1974 , mnras , 169 , 59 hollenbach , d. , & mckee , c. f. 1980 , apjl , 241 , l47 1989 , apj , 342 , 306 huang , y .- l . , & thaddeus , p. 1986 , apj , 309 , 804 humphreys , r. m. 1978 , apjs , 38 , 309 jones , t. w. , & kang , h. 1993 , apj , 402 , 560 jones , a. p. , tielens , a. g. g. m. , & hollenbach , d. j. 1996 , apj , 469 , 740 keohane , j. w. , petre , r. , gotthelf , e. v. , ozaki , m. , & koyama , k. 1997 , apj , 484 , 350 klein , r. l. , mckee , c. f. , colella , p. 1994 , apj , 420 , 213 koo , b .- c . , & heiles , c. 1991 , apj , 382 , 204 koo , b .- c . & moon , d .- s . 1997 , apj , 485 , 263 langer , w. d. 1997 , in co : twenty - five years of millimeter - wave spectroscopy , iau symp . latter , w. b. et al . ( dordrecht : kluwer ) , 98 leahy , d. a. 1987 , mnras , 228 , 907 leahy , d. a. , & aschenbach , b. 1996 , a&a , 315 , 260 lockett , p. , gautheir , e. , elitzur , m. 1999 , apj , 511 , 235 mac low , m .- mckee , c. f. , klein , r. i. , stone , j. m. , & norman , m. l. 1994 , apj , 433 , 757 mckee , c. f. 1988 , in supernova remnants and the interstellar medium , proceedings of iau colloquium no . roger and t.l . landecker ( cambridge : cambridge ) , 473 oliva , e. , moorwood , a. f. m. , drapatz , s. , lutz , d. , & sturm , e. 1999 , a & a , 343 , 943 reach , w. t. , & rho , j. 1996 , a&a , 315 , l277 . 1999 , apj , 511 , 836 . 2000a , aph0007148 . 2000b , in preparation rengelink , r. b. , tang , y. , de bruyn , a. g. , miley , g. k. , bremer , m. n. , rttgering , h. j. a. , & bremer , m. a. r. 1997 , a&as , 124 , 259 rho , j. , & petre , r. 1998 , apjl , 503 , l167 rho , j. , jarrett , t. , cutri , r. m. , & reach , w. t. 2000 , apj , submitted roh , d .- jung , j. h. 1999 , publications of the korean astronomical society , 14 , 123 sakamoto , s. , hasegawa , t. , handa , t. , hayashi , m. , & oka , t. 1997 , apj , 486 , 276 sakamoto , s. , hayashi , m. , hasegawa , t. , handa , t. , & oka , t. 1994 , apj , 425 , 641 scoville , n. z. , & solomon , p. m. 1974 , apjl , 187 , l67 seta , m. , hasegawa , t. , dame , t. m. , sakamoto , s. , oka , t. , handa , t. , hayashi , m. , morino , j. , sorai , k. , & usuda , k. s. 1998 , apj , 505 , 286 tatematsu , k. , fukui , y. , landecker , t. l. , & roger , r. s. 1990 , a&a , 237 , 189 ( t90 ) tauber , j. a. , snell , r. a. , dickman , r. l. , & ziurys , l. m. 1994 , apj , 421 , 570 van den bergh , s. 1978 , apjs , 38 , 119 van der laan , h. 1962 , mnras , 124 , 179 van dishoeck , e. f. , jansen , d. j. , & phillips , t. g. 1993 , a&a , 279 , 541 wheelock s. l. , gautier , t. n. , chillemi , j. , kester , d. , mccallon , h. , oken , c. , white , j. , gregorich , d. , boulanger , f. , and j. good 1994 , _ iras sky survey atlas : explanatory supplement_. ( jpl / caltech : pasadena ) willis , a. g. , 1973 , a&a , 26 , 237 wilner , d. j. , reynolds , s. p. , & moffett , d. a. 1998 , aj , 115 , 247 hb21:bml - n1 & ( 20 46 03.2 , 51 00 00 ) & @xmath199 & 3.6 & 1.6&@xmath59&@xmath200hb21:bml - n2 & ( 20 45 55.0 , 51 03 30 ) & @xmath201 & 3.1 & 1.7&@xmath202& ... hb21:bml - s1 & ( 20 44 37.2 , 49 47 10 ) & @xmath203 & 2.7 & 1.9&@xmath204&@xmath205hb21:bml - s2 & ( 20 44 31.0 , 49 55 20 ) & @xmath206 & 2.9 & 1.7& ... &@xmath207hb21:bml - s3 & ( 20 42 45.2 , 49 56 50 ) & @xmath208 & 6.9 & 2.3& ... &@xmath209 a & 1.1(0.2 ) & 1.1(0.1 ) & 4.2(0.2 ) & 20.8(2.3 ) & 0.20(0.03)b & 1.7(0.1 ) & 1.8(0.1 ) & 5.9(1.1 ) & 29.5(2.4 ) & 0.20(0.04)c & 1.7(0.1 ) & 1.7(0.1 ) & 8.0(1.1 ) & 38.0(2.4 ) & 0.21(0.03)n & 1.0(0.2 ) & 0.8(0.1 ) & 4.8(0.2 ) & 17.0(2.3 ) & 0.28(0.04)s & 1.2(0.1 ) & 1.2(0.1 ) & 6.8(1.1 ) & 26.3(2.4 ) & 0.26(0.05 )
we have carried out @xmath0co j=21 line observations of the supernova remnant ( snr ) hb 21 in order to search for evidence of interaction with molecular clouds . we mapped the eastern half ( @xmath1 ) of the snr almost completely . molecular gas appears to be distributed mainly along the boundary of the snr , but the overall distribution has little correlation either with the distortion of the snr boundary or with the distribution of radio brightness . along the eastern boundary , where the snr was considered to be interacting with molecular clouds in previous studies , we have not found any strong evidence for the interaction . instead we detected broad ( 2040 ) co emission lines in the northern and southern parts of the snr . in the northern area , the broad - line emitting cloud is composed of a small ( @xmath2 or 0.5 pc ) , very bright , u - shaped part and several clumps scattered around it . there is a significant enhancement of radio emission with flat ( @xmath3 ) spectral index possibly associated with this cloud . in the southern area , the broad - line emitting cloud is filamentary and appears to form an elongated loop of @xmath4 in extent . small ( @xmath5 or 0.3 pc ) , bright clumps are seen along the filamentary structure . we have obtained sensitive j=10 and j=21 spectra of @xmath0co and @xmath6co molecules toward several peak positions . the intensity of @xmath0co j=21 emission is low ( @xmath7 k ) and the ratio of @xmath0co j=21 to j=10 integrated intensities is high ( 1.62.3 ) , which suggests that the emission is from warm , dense , and clumpy gas . we have applied an lvg analysis to derive their physical parameters . the detected broad co lines are believed to be emitted from the fast - moving molecular gas swept - up by the snr shock . the small ( @xmath8 ) shock velocity suggests that the shock is a non - dissociating c - shock . we discuss the correlation of the shocked molecular gas with the previously detected , shocked atomic gas and the associated infrared emission .
consider the piecewise linear periodic function @xmath12 \text { for some } n\in{\mathbb z } , \\ n - t & \text{if } t\in[n-\frac{1}{2},n ] \text { for some } n\in{\mathbb z } , \end{cases}\ ] ] whose graph looks as follows ( 300,80)(-30,-10 ) ( -20,0)(1,0)300 ( 275,-10)@xmath13 ( 0,-10)(0,1)70 ( -20,20)(1,-1)20 ( 0,0)(1,1)40 ( 40,40)(1,-1)40 ( 80,0)(1,1)40 ( 120,40)(1,-1)40 ( 160,0)(1,1)40 ( 200,40)(1,-1)40 ( 240,0)(1,1)20 for every @xmath14 consider the function @xmath15 which is a homothetic copy of the function @xmath16 . spaces called _ shark teeth _ are constructed in @xcite and are parametrized by an infinite non - decreasing sequence @xmath17 . let @xmath18\times\{0\}$ ] be the _ bone _ of shark teeth , and for every @xmath19 let @xmath20\big\}$ ] be the @xmath21th _ row _ of teeth . the space shark teeth is given by the following formula @xmath22 in @xcite is shown that the shark teeth constructed in the plane @xmath23 with the non - decreasing sequence @xmath24 where @xmath25 is the integer part of @xmath26 , is not homeomorphic to an ifs - attractor ( see figure [ shark ] ) . in other words it is not an ifs - attractor in any metric . ] we show that the space @xmath27 from @xcite is a topological ifs - attractor . for @xmath19 and the sets @xmath28 , @xmath29 and @xmath27 , by the same names we denote the functions : @xmath30\ni t\to \big(t,\frac{1}k\varphi_{n_k}(t)\big)\in m_k,\ ] ] @xmath31\ni t\to ( t,0)\in i \text { and}\ ] ] @xmath32\ni t\to i(t)\cup\bigcup_{k=1}^\infty m_k(t).\ ] ] note that for every @xmath33 there exists unique @xmath34 $ ] , such that @xmath35 or @xmath36 for some @xmath21 . therefore we can represent every point of the space @xmath27 as an element from the unit interval and perhaps with positive parameter @xmath21 . note that for @xmath37 and for every @xmath38 we have @xmath39 , because then @xmath26 belongs to @xmath29 . in three steps we will present the construction of topological ifs and prove that @xmath27 is its attractor . * let @xmath40 be the collection of continuous functions on @xmath27 to itself such that for every @xmath33 @xmath41 @xmath42 @xmath43 @xmath44 thus the union of images of @xmath27 under every function @xmath45 fills up the first row of the teeth @xmath46 . analogously we construct functions @xmath47 which fill up the second row @xmath48 . now we are going to construct functions @xmath49 and @xmath50 which cover left and right side of the rest of rows . define @xmath51 , so it only shifts values of function @xmath49 . for every @xmath52 let us define @xmath53 as @xmath54-th _ generation _ of shark teeth . we can also treat it like a function @xmath55\ni t\to \bigcup\{m_k(t ) : n_k = i\}\in g_i$ ] . note that every row in one generation contains the same number of teeth ( @xmath56 ) . by @xmath57 we denote the number of first row of teeth in @xmath58 , and by @xmath59 we denote the number of rows in @xmath58 . function @xmath49 has to transform every generation into the left part of next generation , so let @xmath60 be the number of rows from @xmath61 filled by one row from @xmath58 . in our case @xmath62 and @xmath63 for every @xmath52 . we want the function @xmath49 to transform whole row from @xmath58 into @xmath64 rows from @xmath65)$ ] . therefore , points @xmath66 for @xmath67 and some positive @xmath21 , must have distinct values @xmath68 in the same order on @xmath29 . to obtain this , every tooth from @xmath58 must be divided into @xmath69 pieces , which each of them covers one tooth from @xmath61 and the last one fills small part of bone @xmath29 . in other words for @xmath70 a tooth from @xmath71\big)$ ] is transformed by @xmath49 into @xmath64 teeth from @xmath72\big)$ ] and bone @xmath73\big)$ ] ( see figure [ tooth ] ) . is transform to @xmath64 teeth from @xmath61 and small part of bone @xmath29 . ] note that for @xmath74 and for similarity @xmath75 , we can write @xmath76=p_{i , j}([0,1])$ ] . moreover , define @xmath77 $ ] for @xmath78 . now we can present the formula for the function @xmath49 : @xmath79 and for @xmath52 , @xmath80 and @xmath70 we have @xmath81 we can write that @xmath82 . indeed @xmath83 and easy calculations can show that for every @xmath52 we have @xmath84)\cup i([0,\frac{1}2])$ ] and @xmath85)\cup i([\frac{1}2,1])$ ] , so @xmath86 * step 2 . * according to the definition of functions @xmath45 and @xmath47 we have the following property for @xmath87 @xmath88 so for every positive @xmath8 and connected set @xmath89 we have @xmath90 where @xmath91 and analogously for functions @xmath47 . we know also the similar thing about functions @xmath92 . for any positive @xmath8 @xmath93 where @xmath94 . this arose due to the fact that for every @xmath52 and @xmath70 @xmath95\big)\big ) = g_{i+1}\big(\big[\frac{j}{2^{i+1 } } , \frac{j+1}{2^{i+1}}\big]\big)\cup i\big(\big[\frac{j}{2^{i+1 } } , \frac{j+1}{2^{i+1}}\big]\big).\ ] ] * step 3 . * let @xmath7 be an open cover of @xmath27 . in the last step we are going to find a positive number @xmath96 , such hat the diameter of @xmath97 is less than the lebesgue number @xmath98 of @xmath7 , where @xmath99 . let us consider every possible compositions of functions from @xmath100 . we will study the diameter of image of the space @xmath27 under this composition . from step 2 we know that composition of functions only from @xmath101 , from @xmath102 or from @xmath103 makes half the size of the space @xmath27 ( see equations ( [ zlozenia_g ] ) and ( [ zlozenia_f ] ) ) . note also that for every connected set @xmath89 its images @xmath104 , @xmath105 and @xmath106 are contained in @xmath107 , @xmath108 and @xmath109 respectively , so @xmath110 @xmath111 because they are all singletons . this means that if the functions @xmath45 , @xmath47 and @xmath92 appear in composition in the above order , the diameter of the image will be 0 . it only remains for us to consider the compositions of the form @xmath112 and analogously @xmath113 , where @xmath114 and @xmath115 . let @xmath116 be the lipschitz constant of function @xmath49 and @xmath50 restricted to @xmath21-th generation . it is finite because of the definition of @xmath49 . note that the set @xmath112 is contained in generation @xmath117 , so we obtain @xmath118 on the other hand @xmath119 now fix @xmath120 such that @xmath121 and fix @xmath122 such that @xmath123 then we claim the thesis holds for @xmath124 . indeed , all images of @xmath27 under compositions only from @xmath101 , from @xmath102 or from @xmath103 have diameters less than @xmath98 , because of the definition of @xmath125 . moreover @xmath126 for @xmath114 and @xmath115 because 1 . if @xmath127 then @xmath128 2 . if @xmath129 then @xmath130 analogously we show that @xmath131 . the others compositions transform whole space @xmath27 into the point so the diameter of the image of @xmath27 is @xmath132 . this ends the proof . in fact the construction above can be extended to all shark teeth . if we try to construct a topological ifs for shark teeth with an arbitrary sequence @xmath17 , we can meet the following problems 1 . some @xmath58 are empty . + then we have to renumber the sequence @xmath58 such that the empty sets are omitted . 2 . @xmath133 . + then define @xmath134 , where @xmath135 is a minimal integer grater or equal to @xmath26 . consequently , the formula for the function @xmath49 slightly changes . the last row of teeth from every @xmath54-th generation has to be transformed into less than @xmath64 rows from @xmath61 . it can be done by covering some rows form @xmath61 once again . @xmath64 is odd . + then we do not have to cover a small part of bone under every tooth , so we divide every tooth from @xmath58 into @xmath64 pieces , like in the figure [ tooth2 ] .
we show that the space called shark teeth is a topological ifs - attractor , that is for every open cover of @xmath0 , its image under every suitable large composition from the family of continuous functions @xmath1 lies in some set from the cover . in particular , there exists a space which is not homeomorphic to any ifs - attractor but is a topological ifs - attractor . iterated function systems ( ifs ) are one of the most popular and simple method of constructing fractal structures , which has wide applications to data compression , computer graphics , medicine , economics , earthquake and weather prediction and many others . a compact metric space @xmath2 is called an _ ifs - attractor _ if @xmath0 for some contractions @xmath3 . in this case the family @xmath4 is called an _ iterated function system_. we recall that a map @xmath5 is a _ contraction _ if its lipschitz constant @xmath6 is less than 1 . the notion of an iterated function system was introduced by john hutchinson in 1981 @xcite and popularized by michael barnsley @xcite . topological properties of ifs - attractors were studied in @xcite , @xcite and @xcite . in particular the definition of topological ifs - attractor was proposed in the last paper : compact topological space @xmath2 is a _ topological ifs - attractor _ if @xmath0 for some continuous maps @xmath3 with the property that for any open cover @xmath7 of @xmath2 there is @xmath8 , such that for any functions @xmath9 the set @xmath10 lies in some set @xmath11 . note that every compact , metric space @xmath2 is a topological ifs - attractor if for its any open cover @xmath7 the diameter of the set @xmath10 is less than the lebesgue number of @xmath7 , for some @xmath8 and every @xmath9 . it is easy to see that each ifs - attractor is a topological ifs - attractor but not the other way around . moreover , we show that a space called shark teeth , constructed in @xcite , which is not homeomorphic to attractor of any iterated function system is a topological ifs - attractor .
accretion of material from the circumstellar disk onto the central star is a common process occurring during the pre - main sequence ( pms ) phase of stellar evolution @xcite . disk accretion is responsible for the build up of a significant fraction of the final stellar mass , thus affecting in part the shape of the initial mass function ( imf ) . on - going accretion is also direct evidence of the presence of gaseous circumstellar disc material close to the star and therefore the time dependence of the mass accretion rate ( @xmath5 ) traces disc evolution . moreover , accretion influences the disk structure and thus contributes to the conditions for planet formation . understanding the temporal evolution of mass accretion rates is crucial for modelling both planet formation and the early evolution of stars . the dependence of @xmath0 on age @xmath6 and stellar mass @xmath7 has been investigated both theoretically and observationally ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? a general finding is that @xmath0 decreases with time indicating that the inner circumstellar disk is cleared within a few myr of pms evolution and scales with stellar mass . the large scatter in the data does not allow an accurate constraint on the form of this relation , which is commonly assumed to be a power law : @xmath8 measurements in a number of nearby star forming regions generally find @xmath9-@xmath10 and @xmath11@xmath12 @xcite . in the orion nebula cluster ( onc ) @xcite found a typical trend for the entire stellar sample ( @xmath3 ) , from a uniquely large sample of precise @xmath0 estimates ( over 700 sources ) . these measurements were derived combining hubble space telescope photometry of stellar flux excesses with the available spectroscopically determined stellar photospheric parameters in the onc from @xcite . despite the typical value of @xmath2 for the overall population , @xcite also found that stars of about a solar mass showed a systematic slower decay of accretion , with @xmath13 . similar results for intermediate masses ( @xmath14 m@xmath15 ) were obtained in several detection - limited photometric studies of @xmath0 in the magellanic clouds ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) as well as in some galactic star forming regions @xcite . these findings suggest that solar - type stars may sustain mass accretion from long - lived gaseous disks on timescales @xmath16 myr . the measurement of accretion rates , as well as that of the stellar parameters of young stars , are affected by large uncertainties . as we will show , such uncertainties , together with observational obstacles such as incompleteness due to detection limits , influence the measured dependence of @xmath0 on stellar properties in this paper we focus on the biases in the measured correlations between @xmath0 and stellar age , and the estimated accretion timescales . in what follows we will characterize this bias by comparing the true value of @xmath2 in equation [ equation : mdot_propto ] , which we refer to as @xmath17 , with that obtained from fitting the @xmath0 versus @xmath18 relationship to observational or simulated data , which we refer to as @xmath19 . in section [ section : the_problem ] we introduce the major underlying problems with measuring @xmath0 , stellar parameters , and thus @xmath2 ; in section [ section : simulations ] we set out our modelling approach and the underlying assumptions . in section [ section : results ] we present the results of simulations to study the systematic biases in @xmath2 introduced by a combination of uncertainties and selection effects . a more detailed individual treatment of each of these effects is presented in appendices [ appendix : simulations_simplespreads ] and [ section : selection_effects ] . finally we discuss implications and caveats of our findings ( section [ section : discussion ] ) and summarize ( section [ section : summary ] ) our conclusions . the stellar parameters of young stars usually can not be assigned with good accuracy , due to both observational uncertainties and the physical nature of pms stars ( see the review by * ? ? ? young clusters are often partially embedded in their parental clouds , causing differential extinction between their members ; @xmath20 and @xmath21 must be estimated for individual stars based on either spectroscopic or photometric techniques . pms stars are also known to exhibit substantial variability @xcite , due to rotation of heavily spotted photospheres and non steady accretion . moreover , the continuum emission produced by accretion is often not properly characterized @xcite and it may present peculiar spectral energy distributions @xcite . unresolved stellar multiplicity also biases the observed luminosities and inferred ages @xcite . furthermore , the history of protostellar accretion is thought to influence the subsequent stellar structure for the first few myrs of pms evolution @xcite . if different stars of a population undergo a diversity of protostellar buildup histories during the embedded phase ( intensity of accretion bursts , different initial conditions ) , this can result in a significant luminosity spread even for coeval stars . the effect is most prominent for `` cold '' protostellar accretion - when little or no energy from the accretion flow is stored in the stellar interiors , and may result in systematically smaller radii and lower luminosities for stars of @xmath22 even several myr after they emerge as t - tauri stars @xcite . stellar radii may also expand in response to `` hot '' protostellar accretion , but the contraction timescales will then be so short ( @xmath23 myr ) that such objects are unlikely to contribute significantly to the luminosity spread seen in a population of pms objects . whereas both stellar masses and ages are uncertain due to these issues , ages are far more problematic : first , protostellar accretion produces spreads mostly in @xmath24 , which correlates with @xmath6 and biases @xmath7 only marginally ; second , the typical range of masses observed in pms clusters is large ( @xmath25 dex ) , and much larger than the uncertainties on this quantity . all these effects broaden the measured age distribution in young clusters and lead to overestimates of the real intrinsic age spread the actual extent of this effect is partially unknown , but some studies carried out in the onc , have tried to constrain it . the population of the onc , when stars are carefully positioned in the hrd using spectroscopic spectral types complemented with photometry to disentangle dust extinction for individual sources , shows an isochronal spread in log age of @xmath26 dex around a mean , model dependent , age of @xmath27 myr @xcite . independent studies based on the analysis of uncertainties @xcite and projected stellar radii @xcite indicate that most of the luminosity spread must be attributed to a spread in stellar radii , and the observational uncertainties have a marginal role , accounting for an apparent age spread of only @xmath28 . nevertheless , @xcite points out that not all the spread in radii can be attributed to an intrinsic age spread , since an expected correlation between isochronal ages and disk fractions is not observed . through statistical simulations , they estimate an upper limit on the real age spread in the onc @xmath29 dex . this result , combined with evidence for a true spread in radii and the small impact of observational uncertainties , suggests that a `` non negligible '' luminosity spread may be induced by protostellar accretion histories . based only on spatial and kinematic considerations , a real age spread must exist at some level ( e.g. , * ? ? ? * ) and extreme perturbations in the hrd due to protostellar accretion histories are not supported by measurements of lithium depletion @xcite . the gravitational energy released by infalling material produces a characteristic flux excess over that expected from the stellar photosphere @xcite . this comprises a strong uv excess produced by the shock at the photosphere , wide optically thin recombination lines from the infalling column , and optical continuum from the heated photosphere , often referred to as veiling . practically , the bolometric accretion luminosity @xmath30 is estimated by measuring the intensity of a portion of the excess emission , usually hydrogen ricombination lines especially h@xmath31 in the optical or the balmer jump excess in the @xmath32-band , from either spectroscopy @xcite or photometry @xcite . practically , the measurement @xmath30 is some cases imprecise . separating @xmath30 from the total stellar flux requires a good knowledge of the photospheric flux , but this is often imperfect ( see section [ section : the_problem_age ] ) . in reality , the system of star , disk and accretion have a more complex interplay between the radiative and hydrodynamical aspects ( e.g. , * ? ? ? * ; * ? ? ? also , the bolometric @xmath30 is extrapolated from portions of its continuum or line emission based on empirical relations @xcite , which are affected by a significant dispersion , and in some cases the spectrum of accretion excess can present large anomalies @xcite . even when the measurements of @xmath30 are precise , their conversion into mass accretion rates requires additional assumptions . since @xmath30 originates from the release of gravitational energy , @xmath33 where @xmath34 , @xmath35 and @xmath7 are the stellar radius , luminosity and mass , and @xmath36 the inner disk radius . thus , since it is commonly assumed @xmath37 : @xmath38 in reality @xmath39 is probably closer to the co - rotation radius @xcite , which varies star by star as the typical pms rotation periods @xmath40 range from less than a day to @xmath41 days . this produces uncertainties of a factor 2 or more in estimates of @xmath5 . fortunately , the uncertainties in estimates of @xmath0 should be a minor cause of bias in @xmath2 compared to uncertainties in the stellar parameters ( section [ section : the_problem_age ] and [ section : the_problem_correlated_uncertainties ] ) . first , estimates of @xmath2 are generally based on a large number of sources . for example , the work of @xcite determined @xmath0 in the onc for over 700 sources , using hst photometry to measure both h@xmath31 and @xmath32-band excess . the large number statistics mitigate against the inevitable presence of peculiar sources departing from the assumptions of equation [ equation : mdot1 ] , ( although this does not preclude systematic problems with this equation ) . second , in the majority of the studies cited in section [ section : intro ] , @xmath0 , and hence @xmath2 , was estimated for sources whose stellar parameters have been assigned from spectroscopy ; thus large anti - correlations between @xmath35 and @xmath30 from a lack of knowledge on what fraction of the photospheric flux is to be attributed to accretion should not be present . if the errors in @xmath0 are symmetric and independent of the stellar parameters , which is largely the case for several of the issues described above , only a vertical scatter is introduced in the measured @xmath42@xmath6 relation , with no effect on @xmath2 . however , equation [ equation : mdot1 ] shows that @xmath0 is strongly dependent on the assigned stellar parameters @xmath34,@xmath35 and @xmath7 . since , at a given @xmath20 , @xmath43 , equation [ equation : mdot2 ] implies that @xmath44 . moreover , for low - mass pms stars stellar evolutionary models predict that @xmath45 , and thus : @xmath46 during the pms contraction towards the main sequence , @xmath35 decreases roughly as @xmath47 , while @xmath7 depends on both @xmath35 and @xmath20 . therefore , if @xmath35 or @xmath20 are uncertain , the consequent errors in age and mass will correlate with those in @xmath0 . the bias in @xmath0 for a given star depends on the actual interplay between uncertainties in stellar parameters and in @xmath30 , which is in turn dictated by the physical or observational effects altering the estimate of these quantities , and can vary from star to star . we can however treat this problem in a simplified way and describe 3 limiting cases : * * _ type a _ * : the ratio @xmath48 is not affected by the uncertainty in the stellar parameters . this assumption is adequate for a number of scenarios where the presence of an error in the estimated stellar luminosity affects @xmath30 in the same way ; e.g. : 1 ) wrong or uncertain distance to a particular source ; 2 ) when a fraction of the stellar light is blocked , by , e.g , a circumstellar disk seen nearly edge - on ; 3 ) when @xmath20 is correctly estimated , but the extinction in the band used to obtain @xmath35 is incorrect because of a peculiar reddening law ; 4 ) in the case of unresolved binarity and both components are ( equally ) accreting . also dust extinction does not significantly affect the ratio @xmath48 , especially if @xmath0 is measured from the @xmath49 excess @xcite . + from equation [ equation : mdot2 ] we then have : @xmath50 * * _ type b _ * : the accretion luminosity @xmath30 is correctly measured , even if the @xmath35 is wrong . this is the case for errors in the stellar luminosity such as : 1 ) rotational variability due to dark spots ; 2 ) incorrect accounting for optical continuum excess ( veiling ) , 3 ) unresolved multiplicity with only one companion responsible for the measured @xmath30 . from equation [ equation : mdot1 ] we have : @xmath51 * * _ type c _ * : a variety of protostellar accretion histories @xcite produces a range of possible @xmath34 for the same @xmath7 and age , leading to an observed luminosity spread . here we assume that @xmath20 is unperturbed . unlike types a and b , this is not an error in the position of the star in the hrd , as @xmath20 , @xmath35 , @xmath34 are correctly determined ; rather , @xmath52 and @xmath7 are wrong , since they are assigned using a unique , incorrect grid of evolutionary models with a single birthline . thus , the biased estimate of mass accretion rates will be : @xmath53 where @xmath54 is the mass assigned to the new position in the hrd using non - accreting isochrones . in reality , when a real stellar population is considered , the overall effect of uncertainties will be due to a combination of biases of types a , b and c. in principle , following the definition of the type a and b biases , an additional type of uncertainty is present ; this is when @xmath35 is correct but @xmath30 is not . this scenario , which encompasses most of the difficulties with the measurement of @xmath30 we discussed in section [ section : the_problem_mdot ] , is however irrelevant for our analysis : if the photospheric parameters are not perturbed , the ages are unaffected , and the result is a vertical ( i.e. uncorrelated ) spread in the @xmath42 versus @xmath6 plane , which does not alter @xmath19 . this will occur when @xmath30 uncertainties are uncorrelated with the stellar parameters ; for instance , where the `` non universality '' or contamination of the accretion spectrum @xcite or the temporal variability of @xmath0 alter the estimated @xmath30 @xcite . figure [ fig : hrd_example ] illustrates an example of how uncertain stellar parameters bias the measured @xmath0@xmath6 correlation . we consider 3 stars on a 3 myr @xcite isochrone , with masses @xmath55 , 1 and 2.5 m@xmath15 , and assign them an arbitrary value of @xmath56 in units of @xmath57yr@xmath58 . we then perturb @xmath20 and @xmath24 , shifting their values along circles in the hrd , mimicking the effect of an uncertainty of 0.2 dex in @xmath24 and 0.02 dex in @xmath59 . for each of these points , we re - assign @xmath7 and @xmath6 by isochrone interpolation , and derive the correction to @xmath0 due to the resulting change in the stellar parameters @xmath60 from equation [ equation : mdot2 ] , assuming a bias of either type a or b. the result shows the highly correlated shifts in @xmath18 and @xmath42 produced by errors in @xmath59 and @xmath24 , with a slope ( on logarithmic axes ) ranging from 0 to @xmath61 . the examples shown in figure [ fig : hrd_example ] suggest that when an entire stellar population is considered , the data in the @xmath42@xmath18 plane are shifted by uncertainties along some preferential direction , affecting the measured slope @xmath19 and introducing an apparent correlation between these two quantities even in the limiting case of a coeval population . since , as we discussed earlier , uncertainties in the isochronal ages in young clusters are large , perhaps accounting for half or more of the width of the apparent age distribution , this bias in @xmath2 may be significant . also , given that ages are generally placed on the @xmath62-axis to determine @xmath2 from the @xmath42@xmath18 plane , regression dilution ( @xcite ) becomes important . this is a general phenomenon affecting linear regression and a consequence of it being a non symmetric approach in the two variables . statistical variability , measurement error or random noise in the @xmath63 variable cause uncertainty in the estimated slope , but not bias ; on the other hand any noise , errors or scatter in the independent variable @xmath62 does bias the estimated slope , and the greater the variance in the @xmath62 measurement ( our stellar age ) , the closer the estimated slope ( our @xmath19 ) approaches zero instead of the true value . this bias could in principle be removed , but only if error bars in both axes are well known and accounted for when fitting the data . this is not generally performed for stellar ages . in the analysis of this paper ( see section [ section : results ] ) we will assume that the bias in @xmath0 affecting @xmath2 originates from the errors in the stellar parameters , and that any uncertainty from the @xmath30 estimate is independent of the stellar parameters . however , in some cases the interplay between the uncertainties is more complex and may produce additional biases in @xmath2 . when the data are insufficient to properly separate the photospheric and accretion contribution to the total flux e.g. , when results are based solely on broad - band photometry , and for intrinsically faint sources like brown dwarfs an underestimate of @xmath30 leads to an overestimate of the stellar @xmath20 and @xmath35 , with a consequent simultaneous error in both age and @xmath64 @xcite . additionally , as stellar rotation periods in young stars vary with isochronal age @xcite , some correlation between age and inner ( co - rotation ) disk radius may be present , introducing an age - dependent variation of the term @xmath65 in equation [ equation : mdot1 ] . similarly , stellar rotation in young clusters often appears slower for disk - bearing stars ( e.g. * ? ? ? * ; * ? ? ? * ) , and the presence of disks is both directly related to the accretion process , and a contributor to the assignment of inaccurate stellar parameters @xcite . all of these factors are yet to be sufficiently characterized or understood well enough to study their possible biasing impact on the measured time decay of @xmath0 . because of this , and as in this work we aim to look for systematic biases in the published values of @xmath19 , which usually assume equations [ equation : mdot_propto ] and [ equation : mdot1 ] , we will focus on the bias produced by uncertain stellar parameters , and leave additional comments addressing the possiblity of other effects to the discussion in section [ section : discussion ] . selection effects due to accretion detection thresholds may also bias @xmath19 . small @xmath0 is detected more easily in faint stars than in bright stars , since the contrast , @xmath48 , is larger . this issue has been pointed out as critical , especially for measuring the mass dependence of @xmath0 through the parameter @xmath66 in equation [ equation : mdot_propto ] @xcite . the measured @xmath0@xmath6 correlation is also affected since @xmath35 decreases with apparent age . alternatively , if the stellar sample is incomplete at faint luminosities , older stars with small @xmath0 could be systematically missed . a flat detection limit in stellar luminosity ( or magnitude ) and a uniform threshold in the quantity used to determine accretion excess ( e.g. , h@xmath31 equivalent width , or @xmath67 band excess in magnitudes ) will result in an inclined detection threshold in the @xmath68@xmath6 plane , skewing the relationship between these quantities deduced from samples with detected accretion . this type of selection effect is dominant in studies of distant clusters ( e.g. , the magellanic clouds , or extincted galactic clusters ) , where sources can not be detected in the very low - mass range , and in older @xmath69 myr clusters , where the fraction of sources with detectable mass accretion is minimal . we perform simulations of synthetic populations in the hrd to characterize the biases in the measured exponent @xmath19 , due to inaccurate assignment of stellar parameters and selection effects . the general method is straightforward : we generate a large set of pms stars assuming an age distribution ; we assign values of @xmath70 to each star according to an assumed value of @xmath71 and @xmath66 . then we displace the sources in the hrd mimicking the effect of different types of uncertainties . for each source , its perturbed position in the hrd results in the assignment of a new measured age from isochrone interpolation , as well as a new measured @xmath0 using the perturbed stellar parameters , @xmath34 , @xmath35 , @xmath7 , in equation [ equation : mdot2 ] ) . lastly , @xmath19 is found by linear regression and compared to the assumed @xmath17 . throughout our simulations , we assume a cluster population with a mean age of 2.5 myr , similar to that of the onc when @xcite isochrones are used to assign ages @xcite . in fact the absolute age of orion , as that of any young cluster , is uncertain by a factor of several and model dependent . however , the onc has a relative age comparable , within a factor of 2 , to the majority of star forming regions where accretion timescales are studied . our assumed mean cluster age of 2.5 myr is purely dictated by our intention to generate an observed hrd comparable with the onc , since we will adopt @xcite models , but its absolute value has little to do with the phenomena we aim to study . since pms isochrones in the hrd are almost parallel and spaced by nearly identical offsets in @xmath18 , changing the mean age for a population produces mostly a horizontal shift in the @xmath0@xmath6 plane with little effect on @xmath19 . in section [ appendix : evolmodels ] we demonstrate this , and show that the behavior of the bias in @xmath2 remains nearly identical if another set of evolutionary models with a large difference in predicted absolute age scale is adopted instead of the @xcite isochrones . although @xmath19 is insensitive to changes in the absolute age scale , the measured accretion timescale @xmath72 is proportional to the absolute age of the system . this is a general problem for the study of young cluster evolution , with consequences extending far beyond the scope of this paper @xcite . in several parts of this work , e.g. , when studying the effect of luminosity spreads alone ( appendix [ section : simulations - simplespread ] ) , and when multiple combined uncertainties are considered ( section [ section : combined effects ] ) , the apparent broadening of the hrd is tuned to reproduce the apparent age spread @xmath73 observed in the onc , as well as its observed mean age , and the value of @xmath19 in the region . none of these quantities are peculiar to the onc , and other galactic star forming regions share similar values , so our treatment and results should be of general applicability . we assume that the observed spreads in apparent age are produced by a combination of a real age spread and spreads due to observational uncertainties that are formed by displacing sources in the hrd along the @xmath24 axis , the @xmath59 axis , and along both axes due to uncertain @xmath21 . we let the real spread in log age range between @xmath74 dex and 0.3 dex , tuning the observational uncertainties to produce the observed apparent log age spread of 0.4 dex . this leaves open the possibilities that the additional spread in apparent age , beyond that attributable to the _ known _ observational uncertainties ( e.g. , * ? ? ? * ) , is due to ( a ) physical effects that lead to different hrd positions for stars of the same age ( e.g. the protostellar accretion history ) or ( b ) further unrecognised sources of uncertainty in @xmath35 and @xmath20 ( e.g. , * ? ? ? lastly , we initially assume that @xmath0 depends solely on the stellar mass and age through equation [ equation : mdot_propto ] . as anticipated in section [ section : the_problem_correlated_uncertainties ] , other physical and observational effects may add further complications if the error in deriving @xmath30 has some correlation with the assigned stellar parameters . however , we deliberately base our work on the onc to mitigate such scenarios . the onc , is perhaps the most extensively studied young cluster , where stellar parameters and accretion luminosities have been assigned with the highest achievable precision and completeness , and are subject to the minimum of observational bias . in section [ section : discussion ] we discuss the implications of relaxing this assumption . we generate 50,000 random stars with ages drawn from the assumed age distribution and masses drawn from the @xcite initial mass function . we then assign their @xmath20 , @xmath35 , @xmath34 by interpolation of the @xcite evolutionary models . the assigned @xmath0 , following a chosen @xmath17 ( generally between 0 and 4 ) and @xmath66 ( generally either 0 and 2 ) , requires an arbitrary normalization : we impose @xmath56 for @xmath75 m@xmath15 and @xmath76 myr . a different normalization would only produce a rigid vertical shift in the @xmath42@xmath18 plane , with no effect on @xmath19 we then randomly displace the positions of these sources in the hrd to reproduce the effect of a given observational uncertainty ( e.g. , an assumed apparent luminosity spread , or a differential extinction ) , or physical mechanism ( e.g. , a distribution of @xmath34 from protostellar accretion history ) . in any case , the resulting ( broader ) hrd represents the observed one , thus , for each star , the measured stellar parameters ( @xmath54 , @xmath77 , @xmath78 ) , altered from the true ones , are assigned by interpolation of the siess evolutionary models in the perturbed hrd . as for the measured value of @xmath0 for each source ( see equation [ equation : mdot1 ] ) , the way it changes compared to the true value assigned to each star depends on the nature of the particular uncertainties affecting the stellar parameters , and is computed from equations [ equation : typea ] , [ equation : typeb ] and [ equation : typec ] . figure [ fig : simplespread_mdot_age ] illustrates an example of our simulations . this assumes to begin with a real age spread @xmath79 dex ( the upper limit in the real age spread for the onc from @xcite ) , @xmath80 , @xmath81 ; a gaussian random luminosity spread with @xmath82 dex is added , and the uncertainties in @xmath0 are of type a. the resulting age distribution , from isochrone interpolation after applying the luminosity spread , has a width @xmath83 dex , similar to that measured in the onc . the best linear fit to the population in the @xmath42@xmath18 has a slope @xmath84 , considerably shallower than @xmath17 . we start by separately considering the bias in @xmath2 from individual types of uncertainty in the hrd . these are respectively , uncertainties in luminosity ( including those due to variations in protostellar accretion history ) , uncertain @xmath20 or reddening , and systematic mass - age correlations introduced by evolutionary models . the full description of our methods , assumptions , and findings are in appendix [ appendix : simulations_simplespreads ] ; here we just summarize . * luminosity spreads in the hrd strongly affect the measured correlation between @xmath0 and age , resulting in @xmath85 . * if we assume an intrinsic age spread @xmath86 ( half of the total apparent one ) , then @xmath87 . if the observed luminosity dispersion has a smaller contribution from a real age spread , then this bias becomes larger , and vice - versa . * if the luminosity spreads originate from changes in radius ( at a given age ) due to varying protostellar accretion histories rather than observational uncertainties , then @xmath88 becomes even larger . * uncertainties in @xmath20 and extinction @xmath21 also result in @xmath85 ; however , for realistic values , the effects are more modest than those due to luminosity uncertainties . * changing family of stellar evolutionary models has very little effect on the bias in @xmath2 , even when large systematic differences in the absolute age scale are present . however , a large tilt in the isochrones , as needed to correct them for the typically observed mass - age correlation in their prediction , further increases the bias in the same direction as produced by uncertainties . in appendix [ section : selection_effects ] we study in detail the bias in the @xmath0@xmath6 correlation introduced by source detection incompleteness and a threshold for the measurement of @xmath70 . these selection effects generate apparent correlations in the mass - age and @xmath42@xmath18 planes for those sources which are detected and for which accretion can be measured . our simulations show that as a result , @xmath19 is displaced towards a value of 1 . since @xmath19 in young clusters is commonly in the range @xmath89 , a larger @xmath17 is required to match observations when these selection effects are present . following the analyses in appendices [ appendix : simulations_simplespreads ] and [ section : selection_effects ] where individual biasing effects on @xmath2 are studied , we now analyse the overall effect of realistic simultaneous combinations of different sources of uncertainty . we start by considering the simultaneous effect on @xmath2 of both uncertainties in @xmath20 or @xmath21 and luminosity spreads , but ignore incompleteness due to selection effects . we assume a real age spread @xmath79 dex around a mean age of 2.5 myr and perform simulations of synthetic populations as described in section [ section : simulation_howtheywork ] , proceeding in two steps . first we randomly displace the stellar parameters by adding an uncertainty in @xmath20 of half a spectral subtype , or , separately , an uncertainty in @xmath21 with a flat distribution @xmath90 mag ( which can also be thought of as neglecting to take account of differential extinction ranging between 0 and 1 mag ) . in appendices [ section : simulations - teff ] and [ section : simulations - teffav ] these effects are treated in isolation , but here we assume that these uncertainties will inevitably lead to correlated uncertainties in the luminosity estimate . in other words , an error in @xmath21 leads to an error in @xmath24 , and an error in @xmath20 will result in an incorrectly estimated reddening and extinction , and hence an error in @xmath24 . both of these shift the simulated stars in the same direction in the hrd but the extent of the shifts depends on the respective @xmath21 or @xmath20 errors . in a second step , we add a further luminosity uncertainty , by applying a gaussian shift in @xmath24 for each star , with a standard deviation tailored to obtain an apparent age spread @xmath91 dex . we consider a combination of uncertainties of types a , b and c for this additional luminosity spread , taking the onc as a guide . @xcite showed that only @xmath92 dex of the 0.4 dex apparent age spread could be explained with observational uncertainties of types a or b. the rest ( about 80% ) we choose to attribute to protostellar accretion histories , a case c uncertainty . our assumed combination of the 3 cases has only a minor effect on the result . as we show in appendix [ appendix : simulations_simplespreads ] , the three uncertanty types produce similar biases in @xmath19 , differing only by systematic offets of a few tenths . lastly , in a separate set of simulations , we also include a correction to the isochrones to remove the small mass - age correlation introduced by using the siess models to interpret the observed hrd ( see appendix [ section : acchistory ] ) , as this turns out to have a far larger effect on @xmath2 than even the adoption of a different family of evolutionary models ( appendix [ appendix : evolmodels ] ) . when the siess isochrones ( and most of other evolutionary models ) are used to assign stellar parameters , intermediate mass stars appear older , on average , than low mass stars . we make the assumption that this is a systematic error and tilt the simulated , broadened population in the hrd by the amount that would remove any correlation between mass and age in the observed onc population . equation [ equation : typec ] is then used to calculate a new @xmath0 from the resulting perturbed measured stellar parameters . results comparing the recovered @xmath19 versus the input @xmath17 are shown in figure [ fig : tautau_lumi_teff_underl ] , separating the assumptions of an uncertainty in @xmath20 or of differential @xmath21 , and including or not the correction of the isochrones for the mass - age correlation . each simulation is performed for two input mass dependences of @xmath0 : @xmath81 ( no mass dependence ) and @xmath93 ( large mass dependence ) . results are presented for the entire mass range and also for three subsets of @xmath7 . we find that @xmath94 and that when considering only the @xmath20 or @xmath21 uncertainties ( upper panels of fig . [ fig : tautau_lumi_teff_underl ] ) , this bias is not sensitive to @xmath66 and only weakly dependent on stellar mass . thus , for the level of uncertainties we have assumed , a value of @xmath95 that is typically found in the onc and other young galactic clusters ( see section [ section : intro ] ) can only be reproduced if @xmath96 , which indicates a much faster decay of @xmath0 than commonly assumed . it is noteworthy that the simulations with spreads in @xmath20 or @xmath21 _ and _ @xmath24 ( upper panels of figure [ fig : tautau_lumi_teff_underl ] ) yield a similar bias to those found from simulations with only a luminosity spread of the same total extent ( see figure [ fig : simplespread_tautau02 ] in appendix a ) . the reason for this is presumably that uncertainties in @xmath20 or @xmath21 at the level we have assumed do not lead to @xmath24 errors that contribute in a significant way to the observed @xmath24 dispersion , which is therefore dominated by the additional uncertainty in @xmath24 that we have inserted in the simulations . when we correct the isochrones to eliminate a possible mass - age correlation in intermediate mass stars ( see section [ section : acchistory ] ) , the bias in @xmath2 remains similar ( fig . [ fig : tautau_lumi_teff_underl ] , bottom panels ) . however , the small mass dependence seen in the upper panels is completely removed . this phenomenon is a coincidence , but simplifies any quantitative estimate of the bias in @xmath2 . on the other hand the bias does now depend on @xmath66 , getting stronger as @xmath66 increases . we have performed similar simulations changing the assumed real age spread between @xmath97 dex . we have considered the same uncertainties in @xmath20 and @xmath21 , but varied those in @xmath24 in order to reach the same total apparent age spread @xmath73 . as before , we find the results almost indistinguishable from those derived in appendix [ section : simulations - simplespread - varyingspread ] when considering solely the effect of a luminosity spread ; the bias in @xmath2 is very sensitive to the balance between any real age spread and the apparent age spread due to uncertainties in @xmath24 . our results predict that the bias in @xmath2 varies from @xmath98 for @xmath74 dex , to @xmath99 for @xmath100 dex . as anticipated , observations in a number of regions have shown systematically lower values of @xmath19 ( @xmath101 ) when the stellar samples are restricted to the solar - mass range ( e.g. , * ? ? ? * ; * ? ? ? our simulations accounting for all uncertainties ( figure [ fig : tautau_lumi_teff_underl ] bottom panels ) on the other hand predict a mass - independent bias in @xmath2 , so that a slower decay of @xmath0 for intermediate masses compared to low - mass stars may be genuine . however , our modelling assumes uniform uncertainties at all masses and this may not be realistic . observations often show systematically larger apparent age spreads for solar - type stars ( e.g. , * ? ? ? * ; * ? ? ? * ) , which could indicate that this mass range is more affected by uncertainties , thus experiencing a larger bias in @xmath2 . luminosity spreads caused by accretion history may also be more important for solar - mass objects @xcite and since this type c uncertainty produces the largest bias ( see appendix [ section : simulations - simplespread - onc ] ) , a smaller @xmath102 might be found for intermediate - mass objects even if @xmath17 is mass - independent . we now add selection effects due to detection incompleteness to our simulations . a full description of the origin of the bias in @xmath2 due to this effect alone is given in appendix [ section : selection_effects ] . we first run a synthetic population simulation as in section [ section : combined effects ] , assuming an intrinsic age spread @xmath79 and the same combination of uncertainties to reach a total apparent age spread @xmath103 . for simplicity , since the effects of uncertainties in @xmath20 and @xmath21 produced similar effects , we only include the @xmath20 uncertainty . the simulations include a correction to the isochrones that removes the apparent correlation between mass and age . starting from the resulting _ observed _ @xmath20 and @xmath24 , we introduce sample incompleteness following the same method as in appendix [ section : selection_effects ] . this assumes that @xmath0 is derived from h@xmath31 photometry , which has a parameterized detection limit and photometric error distribution . simulated sources that are a ) fainter than the detection limit , or b ) have h@xmath31 excess smaller than the photometric uncertainties , are removed from the sample and excluded from the fit in the @xmath0-@xmath6 plane used to derive @xmath19 . two detection limits are simulated : one relatively deep ( @xmath104 ) and one shallow ( @xmath105 ) . we only consider one value for the mass dependence of @xmath0 . the normalization of @xmath0 in the synthetic population modelling ( see section [ section : simulation_howtheywork ] ) leads to different fractions of excluded sources for different values of @xmath66 at a given detection limit . this would provide an inconsistent comparison of results for the same @xmath17 . instead we consider only @xmath106 , which is intermediate to the cases studied in the last section . as in appendix [ section : selection_effects ] , we introduce an additional dispersion of 0.65 dex in the individual values of @xmath0 , which represents star - to - star variations of accretion at a given mass and age . this value has been tailored so that , after the addition of observational uncertainties and the mass dependence of @xmath68 , the overall dispersion from the fitted power law in the @xmath0@xmath6 plane is 0.85 dex , as found in the onc @xcite . figure [ fig : tautau_lumi_teff_underl_seleffects ] compares @xmath102 and @xmath17 for these scenarios . by comparing these plots with those without the inclusion of selection effects ( figure [ fig : tautau_lumi_teff_underl ] ) it appears that incompleteness further displaces the @xmath19 towards unity . this is because the photometric detection threshold in the @xmath107@xmath18 plane is generally inclined with a ( negative ) slope @xmath108 ( see appendix [ section : selection_effects ] ) and @xmath19 is biased towards larger values if the real @xmath109 , and towards smaller values if @xmath110 . the typical measured values @xmath111 ( in the onc and elsewhere ) would require @xmath112 . however the comparison between figure [ fig : tautau_lumi_teff_underl_seleffects ] and figure [ fig : tautau_lumi_teff_underl ] also shows that this latter result is still dominated by the bias caused by uncertainties in the stellar parameters alone ( figure [ fig : tautau_lumi_teff_underl ] ) and hence by the extent of the real age spread @xmath86 in this case . we have also checked the behavior of the results starting from a smaller or larger @xmath113 . as we found in appendix [ section : simulations - simplespread - varyingspread ] , a smaller @xmath113 leads to larger @xmath114 and vice - versa . our modeling has shown that uncertainties in the parameters of pms stars and sample incompleteness strongly bias the measured correlation between accretion and age , parameterised as @xmath115 , leading to smaller values of @xmath102 or equivalently , overestimated accretion timescales . the size of the bias in @xmath2 depends chiefly on the extent to which the apparent age spread in a region / cluster is explained by a real age spread within the region rather than various observational uncertainties and physical mechanisms that broaden the luminosity distribution in the hrd . this has two general implications : 1 . the real decay of mass accretion is faster than that inferred from any measured correlation between @xmath0 and isochronal ages . 2 . if @xmath17 is constrained , either on a theoretical basis , or from observational evidence that is not based on isochronal ages in a single cluster ( e.g. , from the measurement of the mean @xmath0 or the fraction of accretors in clusters of different mean age ) , then comparison between @xmath17 and @xmath19 places a constraint on the real age spread in a cluster . we have shown that assuming a real age spread @xmath116 dex combined with luminosity uncertainties , which yield a total apparent age spread of 0.4 dex ( e.g. , * ? ? ? * ) , leads to @xmath117 . this indicates that the decay of @xmath0 is @xmath118 times faster than that inferred from a simple fit of observed @xmath0 versus isochronal ages . such observations generally suggest @xmath119@xmath10 and hence @xmath120 . this implies an accretion timescale ( @xmath121 myr ) , shorter than that inferred from the measured fraction of accreting sources in clusters of different age ( @xmath122 myr , @xcite ) . this interpretation however is not without caveats : first , precise comparison of these methods is not straightforward ; the fraction of accretors is defined above a given threshold ( e.g. , in @xcite @xmath123 m@xmath15yr@xmath58 ) , and when this fraction is small , at cluster ages of 510 myr , it carries little information about the _ average _ values of @xmath0 in the entire population . second , whereas @xmath2 is weakly affected by the absolute age scale ( see section [ section : simulations ] and appendix [ appendix : evolmodels ] ) which is highly dependent on the evolutionary models one adopts @xmath124 is proportional to the assumed mean cluster age . still , if the current age scales were systematically wrong @xcite , both disk and accretion lifetimes would have to be corrected in a similar way @xcite . if observational and physical uncertainties dominate the luminosity spread observed in young clusters , such that the real ( gaussian ) dispersion in @xmath125 age is smaller than 0.2 dex , then or simulations show that @xmath17 would need to be even larger ( @xmath126 ) to reproduce the observed @xmath42 vs @xmath18 relationship . such a rapid , almost instantaneous , decay of @xmath127 would be incompatible with both observational evidence and the theory of disc evolution . if accretion ceased this rapidly within the first myrs of evolution , clusters of older mean age ( @xmath41 myr ) should not host @xmath128 accreting stars , but a small fraction of accreting older stars are observed in such clusters ( e.g. , * ? ? ? * ; * ? ? ? also , @xmath129 implies a decrease of the average @xmath0 of more than 2 orders of magnitude between 1 and 3 myr , and clusters with mean ages in this age range do not exhibit systematic differences in their average accretion rates to this extent ( e.g. * ? ? ? these observations together suggest that a very fast cluster formation scenario producing a real age spread @xmath130 dex is not favoured . on the other hand , if most of the observed spread in age were real ( more than 0.2 dex dispersion in log age ) , the bias in @xmath2 is smaller , allowing longer accretion timescales . it is difficult to reconcile this , at least in the case of the onc , with the lack of correlation between disc fractions and isochronal ages @xcite . thus it appears most likely that the spread in log age for the onc should not be far from 0.2 dex to satisfy these constraints . assuming a mean age for the onc of 2.5 myr and a lognormal age distribution , this implies that 95% of stars were born between 1.3 and 4 myr ago . this is of order 5 - 10 crossing times in the region ( e.g. , * ? ? ? * ) , indicative of a relatively slow star formation process . considering therefore @xmath86 , and the typical values of @xmath19 , our simulations suggest @xmath131 . the fast decay of @xmath0 implied by our results is not predicted by models that assume viscous disk evolution , ( @xmath132 , * ? ? ? this indicates that other factors such as photoevaporation and high energy photons from the central star @xcite or planet formation may either accelerate the clearing of material in the inner disk during the first few myrs of pms evolution , or inhibit its accretion onto the stellar surface . the measured fraction of sources showing warm dust emission excess ( e.g. , * ? ? ? * ; * ? ? ? * ) in different young clusters may point to longer timescales for inner disk dissipation than the accretion timescale our finding suggest . there is also observational evidence for fast inside - out dispersal @xcite , which supports the notion that the accretion timescales is shorter than the disk dispersal timescale . our analysis shows that the apparent flattening in @xmath2 is driven mainly by the fraction of the observed age spread that can be attributed to luminosity dispersion caused by either observational or physical uncertainties . the simplest solution to mitigate the requirement for a very fast decay of @xmath0 is to allow this fraction to be small and hence a large real age spread . however , there are other possibilities . we have made the general assumption that @xmath0 depends solely on @xmath6 and @xmath7 , with only an additional possible scatter ( e.g. , section [ section : combined effects_uncertanties_seleffects ] ) . in fact this is not the case : @xmath0 is also affected by other parameters , such as the temperature , mass and structure of the disc . let us consider a change in stellar radius produced by protostellar history ; our type c uncertainties . we assume that the temperature of the inner disk is dominated by irradiation from the central object , which is appropriate for low ( present - day ) @xmath0 . if a star shrinks rapidly due to an intense previous accretion phase , the disk temperature @xmath133 decreases @xmath134 . according to models of viscous disk evolution ( e.g. , * ? ? ? * ) , this affects the disk viscosity , and @xmath135 . since ( at constant @xmath20 ) the isochronal age @xmath136 , we have that @xmath137 , where the age @xmath6 here is not the real age , but that inferred from the hrd . thus if a star appears more compact so that its age is overestimated by 1 dex , its @xmath127 is expected to be lower by @xmath138 dex . however , this is a relatively small effect compared to the major biases we have found in this paper . in particular it seems incapable of mitigating the conclusion that age spreads @xmath130 are unlikely . for a coeval population where the apparent age spread is entirely due to the protostellar accretion history @xmath19 would be @xmath139 ( see appendix [ section : simulations - simplespread - coeval ] for type c ) , and hence much smaller than observed in young clusters . another possibility is that stars which have undergone vigorous protostellar accretion have current disk properties that inhibit further accretion . for example , their discs could have a lower inner temperature , be less massive , be more stable , or be truncated at a larger inner radius . in the scenario of @xcite , pms stars may accrete their mass through short - lived intense accretion bursts , reaching up to @xmath140 m@xmath141yr@xmath58 . during the burst phase , the inner disk heats up to very high temperatures ( @xmath142 k , @xcite ) , driving intense accretion through mri . if most of the stellar mass is built up by this process in a relatively short time ( @xmath143 myr ) , it is possible that at later stages @xmath0 remains on average lower than a star with a more steady and quiescent accretion history . thus , a fraction of `` bursting '' stars appear underluminous hence , older and with lower @xmath0 , mimicking an apparent decay of accretion even though they are coeval with the rest of the population . this speculative scenario may allow young populations to be more coeval than inferred from the hrd and for their accretion timescales to be longer . however it is completely unclear to what extent present - day and historical accretion rates are anti - correlated ( or correlated ) . it would take a large systematic difference in the present day @xmath0 , of 2 or more orders of magnitude between stars with different protostellar accretion histories to compensate for our finding that @xmath144 if @xmath145 dex in orion . last , as we anticipated in section [ section : the_problem_mdot ] , the standard methods to derive @xmath64 from @xmath30 , i.e. , equation [ equation : mdot1 ] , are a simplification of a more complex process . in particular the assumption @xmath146 is not realistic , as the inner disk radius is close to the co - rotation radius . if rotation rates are independent of isochronal ages , @xmath147 increases with isochronal age as the star shrinks . in addition , in the onc @xcite measured systematically slower rotation rates for older stars , so that their @xmath39 would be even larger . for the same measured @xmath30 , a larger @xmath39 implies a smaller @xmath0 than that derived from equation [ equation : mdot1 ] . thus , correcting for this effect would result in a larger @xmath19 even _ before _ correcting for the biases treated in this paper , making the main conclusions of this work stronger . also , as we discussed in section [ section : the_problem ] and [ section : simulations ] , the interplay between stellar properties and evolution , accretion process , and disk properties may involve more complicated correlations which could further modify the actual bias in @xmath2 ; however , we find it unlikely that additional unknown effects could forcibly invalidate our main conclusions . this work is based on simulations that aim to represent observational results in the onc . as we discussed in section 3.1 , the spread of luminosities at a given @xmath20 in the onc is quite typical of other star forming regions . our simulations adopt a particular family of evolutionary models , the @xcite isochrones . @xcite showed that assuming 3 different families of evolutionary models to assign stellar parameters and @xmath0@xmath6 values , @xmath2 and @xmath66 remain nearly the same even though the average age of the cluster is model dependent by up to a factor 2 @xcite . in appendix [ appendix : evolmodels ] we further demonstrate that also the bias in @xmath2 , as the ratio @xmath88 is very weakly affected by the choice of evolutionary models , as it turns out to be nearly identical the simulations adopt @xcite isochrones rather than siess models , even though the two grids differ by more than a factor 2 in their absolute age scale . thus our particular choice of evolutionary models is largely irrelevant to our results . our analysis is based on a single imf @xcite . although large deviations from standard imfs are generally not observed for young clusters , at least in the stellar mass range @xcite , in principle a different imf might affect our results . however , our simulations show , that the bias in @xmath2 is very similar even when analysis is restricted to specific mass ranges . thus reasonable imf variations would not affect our results . there are no reasons to suppose that the accretion timescales in the onc are unrepresentative of typical pms evolution : @xmath19 in the onc is comparable with values measured with similar techniques ( stellar parameters derived with spectroscopy , @xmath0 from measuring flux excesses and equation [ equation : mdot1 ] ) in other regions @xcite . since these analyses share similar methods , and given that most of the uncertainties affecting young stars are either intrinsic to the nature of these objects , or due to the simplifying assumptions used to derive @xmath0 from observed quantities , we see no reasons to assume that uncertainties in the onc have a peculiar amplitude . there remains the possibility that something in the onc environment could affect the evolution of accretion rates or the extent of any real age spread . o - stars in the central part of the onc are observed to be externally photoevaporating disks , yet the fraction of members with disks and detectable accretion ( @xmath148 , @xcite ) is similar to lower mass clusters at the same age @xcite . dynamical interactions in the densest parts of the onc may have slightly affected the disk radii distribution @xcite , but this also has little impact on the inner disks . as for the real age spread , it is generally assumed to depend on the cloud density @xcite , through the dynamical collapse time . the onc can be considered an `` average density '' star forming regions , between compact starburst clusters and sparse young populations . our analysis has made the assumptions that @xmath2 is obtained from the measured @xmath0 and isochronal @xmath6 in one cluster ; in some cases the stellar samples considered may include multiple separate stellar populations . under these circumstances , we distinguish two cases . if the mean ages of the populations are similar , differing by say less than their age spreads , then they can be treated as a single cluster and the bias in @xmath2 will be similar . if there is a larger difference in age , the sample could be considered as a single population with a larger real age spread ( and possibly also a larger overall apparent age spread than that measured in individual clusters ) . the quantitative effect on @xmath19 will depend on the exact values of these age spreads ; as the ratio between the overall real age spread ( from multiple populations ) and the total apparent observed age spread increases , the bias in @xmath2 is smaller , and vice - versa . in this work we have investigated how uncertainties in estimated stellar parameters ( mass , radius , age ) and observational selection effects bias estimates of the temporal decay of mass accretion rates , parameterised as @xmath149 . these uncertainties may be due to observational limitations ( e.g. photometric errors , differential extinction , accretion veiling etc . ) or to physical mechanisms such as variations in their protostellar accretion histories @xcite . the overall extent of this broadening is not well known , but many studies have suggested that these uncertainties are not negligible . as a consequence the bias in the measured value of @xmath2 originates from a combination of attenuation bias in the linear regression of @xmath150 const since @xmath18 has large relative errors and correlated uncertainties between the estimated stellar parameters and @xmath0 . our analysis assumes @xmath0 depends only on age and mass . we have performed monte carlo simulations that introduce uncertainties into the hertzsprung - russell diagram and observational selection effects and realistic combinations of these . in all cases the recovered @xmath19 inferred from the observational data is smaller than @xmath17 . if multiple types of uncertainties are simultaneously added , the relative contribution of each is largely irrelevant ; the result is dictated by the overall extent of uncertainties , which in turn depend on the _ apparent _ age spread determined from placing objects in the hrd and the fraction of this that is due to a _ real _ spread in age . assuming a typical apparent age spread observed in young clusters of @xmath91 dex but that only 0.2 dex of this is attributable to a real age spread ( for example in the orion nebula cluster , ( e.g. , * ? ? ? * ) ) , then the @xmath15111.5 that is typically measured in the onc and other young galactic clusters , can only be reproduced if @xmath152 . if the real age spread were smaller ( @xmath153 dex ) , as advocated by fast star formation scenarios , the bias becomes extreme , implying an almost instantaneous decay of @xmath127 , which is incompatible with the detection of accretors in young clusters over a wide range of mean ages ( 1 - 10 myr ) . we conclude that clusters like the onc can not be coeval and must have an age spread @xmath154 dex . conversely , assuming the minimum levels of observational uncertainty , the corresponding real age spread be must be @xmath155 dex , and means @xmath17 would still be at least twice @xmath19 . the present uncertainties in the absolute age scale of pms stars and young clusters do not allow us to obtain an accretion decay timescale ; however the bias in @xmath2 is only weakly affected by the assumed age scale . observational selection effects , leading to incomplete samples as a function of mass or age introduce an additional bias to @xmath19 . depending on the detection limit and intrinsic scatter in the accretion rates , @xmath19 is always biased towards @xmath1561 . thus , a combination of selection effects together with uncertainties in the stellar parameters will worsen the apparent flattening of @xmath2 when @xmath110 , strengthening our conclusion that the decay of @xmath0 is much faster than the observed relationship between @xmath42 and @xmath18 . this result , and the suggestion that the real age spread in a young cluster can not be very small , could be mitigated if there were a strong anti - correlation between present - day @xmath0 and any previous protostellar accretion that gave the star a smaller radius and the _ appearance _ of being older . conversely , correcting the simplistic assumption that the inner disk radius is proportional to the stellar radius , by allowing for the influence of stellar rotation on disk truncation , could exacerbate the bias . additional scenarios originating from a more complex interplay between stellar , disk and accretion properties as well as observational biases in their estimation may add further bias in @xmath2 or partially mitigate our results . however , the magnitude of the bias in @xmath2 we have found is so large that unknown additional effects would have to introduce very strong correlations to alter our main conclusions . lastly , our analysis can not directly explain the observation that the measured @xmath0@xmath6 relation can appear flatter for solar - mass pms stars than those of lower mass @xcite . one possibility is that any radius and hence apparent age perturbation due to varying protostellar accretion histories , plays a more important role , and introduces a greater bias in the measured @xmath2 , in intermediate mass stars , as suggested by @xcite . _ acknowledgments _ we thank the anonymous referee for helpful comments that improved the manuscript , and guido de marchi for helpful discussions on the interpretation of the results . alcal , j. m. , natta , a. , manara , c. f. , et al . 2014 , a&a , 561 , a2 alencar , s. h. p. , basri , g. , hartmann , l. , & calvet , n. 2005 , a&a , 440 , 595 allard , f. , homeier , d. , & freytag , b. 2011 , 16th cambridge workshop on cool stars , stellar systems , and the sun , 448 , 91 antoniucci , s. , garca lpez , r. , nisini , b. , et al . 2011 , a&a , 534 , a32 baraffe , i. , vorobyov , e. , & chabrier , g. 2012 , apj , 756 , 118 baraffe , i. , chabrier , g. , & gallardo , j. 2009 , apjl , 702 , l27 bastian , n. , covey , k. r. , & meyer , m. r. 2010 , ara&a , 48 , 339 beccari , g. , spezzi , l. , de marchi , g. , et al . 2010 , apj , 720 , 1108 bell , c. p. m. , naylor , t. , mayne , n. j. , jeffries , r. d. , & littlefair , s. p. 2013 , mnras , 434 , 806 biazzo , k. , alcal , j. m. , covino , e. , et al . 2012 , a&a , 547 , a104 biazzo , k. , melo , c. h. f. , pasquini , l. , et al . 2009 , a&a , 508 , 1301 bouvier , j. , grankin , k. n. , alencar , s. h. p. , et al . 2003 , a&a , 409 , 169 calvet , n. , & gullbring , e. 1998 , apj , 509 , 802 calvet , n. , muzerolle , j. , briceo , c. , et al . 2004 , aj , 128 , 1294 cardelli , j. a. , clayton , g. c. , & mathis , j. s. 1989 , apj , 345 , 245 clarke , c. j. , & pringle , j. e. 2006 , mnras , 370 , l10 costigan , g. , scholz , a. , stelzer , b. , et al . 2012 , mnras , 427 , 1344 dantona f. , mazzitelli i. , 1998 , in rebolo r. , martin e. l. , zapatero osorio m. r. , eds , asp conf . ser . 134 : brown dwarfs and extrasolar planets a role for superadiabatic convection in low mass structures . pp 442 da rio , n. , & robberto , m. 2012 , aj , 144 , 176 da rio , n. , robberto , m. , hillenbrand , l. a. , henning , t. , & stassun , k. g. 2012 , apj , 748 , 14 da rio , n. , gouliermis , d. a. , & gennaro , m. 2010 , apj , 723 , 166 da rio , n. , robberto , m. , soderblom , d. r. , et al . 2010 , apj , 722 , 1092 da rio , n. , robberto , m. , soderblom , d. r. , et al . 2009 , apjs , 183 , 261 dahm , s. e. 2008 , aj , 136 , 521 de juan ovelar , m. , kruijssen , j. m. d. , bressert , e. , et al . 2012 , a&a , 546 , l1 de marchi , g. , panagia , n. , guarcello , m. g. , & bonito , r. 2012 , american astronomical society meeting abstracts # 219 , 2 # 337.06 de marchi , g. , panagia , n. , romaniello , m. , et al . 2011 , apj , 740 , 11 de marchi , g. , paresce , f. , panagia , n. , et al . 2011 , apj , 739 , 27 de marchi , g. , panagia , n. , & romaniello , m. 2010 , apj , 715 , 1 edwards , s. , fischer , w. , hillenbrand , l. , & kwan , j. 2006 , apj , 646 , 319 ercolano , b. , mayr , d. , owen , j. e. , rosotti , g. , & manara , c. f. 2013 , mnras , _ in press _ ( arxiv:1312.3154 ) fang , m. , kim , j. s. , van boekel , r. , et al . 2013 , apjs , 207 , 5 fang , m. , van boekel , r. , wang , w. , et al . 2009 , a&a , 504 , 461 fedele , d. , van den ancker , m. e. , henning , t. , jayawardhana , r. , & oliveira , j. m. 2010 , a&a , 510 , a72 fischer , w. , edwards , s. , hillenbrand , l. , & kwan , j. 2011 , apj , 730 , 73 frost , c. , & thompson , s. g. 2000 , journal of the royal statistical society , series a , 163 , 173 - 18 garcia lopez , r. , natta , a. , testi , l. , & habart , e. 2006 , a&a , 459 , 837 gatti , t. , natta , a. , randich , s. , testi , l. , & sacco , g. 2008 , a&a , 481 , 423 guarcello , m. g. , damiani , f. , micela , g. , et al . 2010 , a&a , 521 , a18 gullbring , e. , hartmann , l. , briceno , c. , & calvet , n. 1998 , apj , 492 , 323 gullbring , e. , barwig , h. , & schmitt , j. h. m. m. 1997 , a&a , 324 , 155 haisch , k. e. , jr . , lada , e. a. , & lada , c. j. 2001 , apjl , 553 , l153 hartmann , l. , cassen , p. , & kenyon , s. j. 1997 , apj , 475 , 770 hartmann , l. , calvet , n. , gullbring , e. , & dalessio , p. 1998 , apj , 495 , 385 hartmann , l. 2001 , aj , 121 , 1030 hartmann , l. 2003 , apj , 585 , 398 hartmann , l. , zhu , z. , & calvet , n. 2011 , arxiv:1106.3343 herbst , w. , bailer - jones , c. a. l. , mundt , r. , meisenheimer , k. , & wackermann , r. 2002 , a&a , 396 , 513 herczeg , g. j. , & hillenbrand , l. a. 2008 , apj , 681 , 594 hernndez , j. , hartmann , l. , calvet , n. , et al . 2008 , apj , 686 , 1195 hillenbrand , l. a. , bauermeister , a. , & white , r. j. 2008 , 14th cambridge workshop on cool stars , stellar systems , and the sun , 384 , 200 hillenbrand , l. a. , & hartmann , l. w. 1998 , apj , 492 , 540 hillenbrand , l. a. 1997 , aj , 113 , 1733 hosokawa , t. , offner , s. s. r. , & krumholz , m. r. 2011 , apj , 738 , 140 ingleby , l. , calvet , n. , herczeg , g. , et al . 2013 , apj , 767 , 112 jayawardhana , r. , coffey , j. , scholz , a. , brandeker , a. , & van kerkwijk , m. h. 2006 , apj , 648 , 1206 jeffries , r. d. , littlefair , s. p. , naylor , t. , & mayne , n. j. 2011 , mnras , 418 , 1948 jeffries , r. d. 2007 , mnras , 381 , 1169 koenigl a. , 1991 , apjl , 370 , l39 koepferl , c. m. , ercolano , b. , dale , j. , et al . 2012 , mnras , 213 kraus , a. l. , & hillenbrand , l. a. 2009 , apj , 704 , 531 kroupa , p. 2001 , mnras , 322 , 231 krumholz , m. r. , & tan , j. c. 2007 , apj , 654 , 304 kudryavtseva , n. , brandner , w. , gennaro , m. , et al . 2012 , apjl , 750 , l44 kurosawa , r. , romanova , m. m. , & harries , t. j. 2008 , mnras , 385 , 1931 littlefair , s. p. , naylor , t. , mayne , n. j. , saunders , e. , & jeffries , r. d. 2011 , mnras , 413 , l56 luhman , k. l. 1999 , apj , 525 , 466 manara , c. f. , beccari , g. , da rio , n. , et al . 2013 , a&a , _ in press _ ( arxiv:1307.8118 ) manara , c. f. , robberto , m. , da rio , n. , et al . 2012 , apj , 755 , 154 mayne , n. j. , & harries , t. j. 2010 , mnras , 409 , 1307 mayne , n. j. , naylor , t. , littlefair , s. p. , saunders , e. s. , & jeffries , r. d. 2007 , mnras , 375 , 1220 mohanty , s. , jayawardhana , r. , & basri , g. 2005 , apj , 626 , 498 muzerolle , j. , luhman , k. l. , briceo , c. , hartmann , l. , & calvet , n. 2005 , apj , 625 , 906 muzerolle , j. , hillenbrand , l. , calvet , n. , briceo , c. , & hartmann , l. 2003 , apj , 592 , 266 natta , a. , testi , l. , & randich , s. 2006 , a&a , 452 , 245 naylor , t. 2009 , mnras , 399 , 432 nguyen , d. c. , scholz , a. , van kerkwijk , m. h. , jayawardhana , r. , & brandeker , a. 2009 , apjl , 694 , l153 owen , j. e. , hudoba de badyn , m. , clarke , c. j. , & robins , l. 2013 , mnras , 436 , 1430 parmentier , g. , & pfalzner , s. 2013 , a&a , 549 , a132 pascucci , i. , gorti , u. , & hollenbach , d. 2012 , apjl , 751 , l42 pascucci , i. , & sterzik , m. 2009 , apj , 702 , 724 preibisch , t. 2012 , research in astronomy and astrophysics , 12 , 1 rebull , l. m. , stauffer , j. r. , megeath , s. t. , hora , j. l. , & hartmann , l. 2006 , apj , 646 , 297 reggiani , m. , robberto , m. , da rio , n. , et al . 2011 , a&a , 534 , a83 rigliaco , e. , natta , a. , testi , l. , et al . 2012 , a&a , 548 , a56 rigliaco , e. , natta , a. , randich , s. , testi , l. , & biazzo , k. 2011 , a&a , 525 , a47 robberto , m. , song , j. , mora carrillo , g. , et al . 2004 , apj , 606 , 952 romaniello , m. , robberto , m. , & panagia , n. 2004 , apj , 608 , 220 sergison , d. j. , mayne , n. j. , naylor , t. , jeffries , r. d. , & bell , c. p. m. 2013 , arxiv:1306.2282 sicilia - aguilar , a. , henning , t. , & hartmann , l. w. 2010 , apj , 710 , 597 sicilia - aguilar , a. , hartmann , l. w. , frsz , g. , et al . 2006 , aj , 132 , 2135 siess , l. , dufour , e. , & forestini , m. 2000 , a&a , 358 , 593 shu , f. , najita , j. , ostriker , e. , et al . 1994 , apj , 429 , 781 spezzi , l. , de marchi , g. , panagia , n. , sicilia - aguilar , a. , & ercolano , b. 2012 , mnras , 421 , 78 tan , j. c. , krumholz , m. r. , & mckee , c. f. 2006 , apjl , 641 , l121 tannirkulam , a. , monnier , j. d. , millan - gabet , r. , et al . 2008 , apjl , 677 , l51 tout , c. a. , livio , m. , & bonnell , i. a. 1999 , mnras , 310 , 360 tognelli , e. , prada moroni , p. g. , & deglinnocenti , s. 2011 , a&a , 533 , a109 tognelli , e. , prada moroni , p. g. , & da rio , n. 2011 , _ in prep . _ valenti , j. a. , basri , g. , & johns , c. m. 1993 , aj , 106 , 2024 zhu , z. , hartmann , l. , gammie , c. , & mckinney , j. c. 2009 , apj , 701 , 620 this appendix describes in the detail the simulations of synthetic populations to analyse the bias in the @xmath0 temporal decay exponent @xmath2 due to uncertainties in the stellar parameters . we separately treat the effect of luminosity spread , @xmath20 and @xmath21 errors , and systematic mass - age correlations due to inaccurate stellar evolutionary models . we study the bias in @xmath2 introduced by an apparent luminosity spread . we start with a simulated population with an intrinsic age spread @xmath157 . this corresponds to a lognormal age distribution peaked at 2.5 myr and having 95% of the stars between 1 and 6 myr old . this is the upper limit on the age spread in the onc deduced by @xcite from the absence of correlations between disk fraction and apparent isochronal age . we then apply a luminosity spread @xmath158 to the simulated population ; this value is chosen to produce an overall apparent age spread @xmath159 as observed in the onc ( e.g. , da rio et al . 2010a ) . we iterate our simulations for different assumed @xmath17 and @xmath66 ; the results for biases of types a , b and c ( see section [ section : simulation_howtheywork ] are shown in figure [ fig : simplespread_tautau02 ] in terms of @xmath19 versus @xmath17 . this shows that the assumed luminosity spread leads to significant underestimation of @xmath17 in most cases . the value @xmath160 derived in the onc @xcite is recovered if @xmath1613 for uncertainties of type a , and up to 4 or more for types b and c. thus , assuming that only a spread in luminosities affects the observed hrd , then @xmath0 decreases by 3 - 4 orders of magnitude between 0.5 myr and 5 myr . figure [ fig : simplespread_tautau02 ] also shows that this result is only weakly dependent on the mass dependence of @xmath0 , as @xmath19 changes by @xmath162 between @xmath81 to @xmath93 . we find that @xmath19 is slightly smaller for very low mass stars than for higher stellar masses . the best fit for the entire population ( solid lines ) generally follows the results for the lowest masses , as these represent the dominant population because of the imf . if we assume @xmath163 , which indicates a constant @xmath164 over age , the luminosity spread leads to @xmath165 ; it is positive ( a decay of @xmath166 ) , @xmath167 for biases of type a and @xmath168 for type b , but negative ( an apparent increase of @xmath0 with age ) for type c. although assuming @xmath169 is unphysical , provided that the population is not coeval and since @xmath0 must decrease ( in a population ) over time , this offset reveals the effect of the correlation of uncertainties we discussed in section [ section : intro ] . we now test how the bias in @xmath19 is influenced by the assumed real age spread @xmath170 . specifically , we test smaller and larger values than previously assumed , namely , @xmath171 dex and @xmath172 dex . once again , the age distribution peaks at 2.5 myr . the simulations are as in section [ section : simulations - simplespread - onc ] , except that , in order to reach the same apparent observed @xmath73 , we add luminosity spreads @xmath173 = 0.27 dex and 0.15 dex for the cases of @xmath174 dex and 0.3 dex respectively . the results are shown in figure [ fig : simplespread_tautau013 ] . if the assumed real age spread is smaller , the bias in @xmath2 increases , producing both a more prominent flattening of the @xmath19 versus @xmath17 correlation , and a larger offset from correlated uncertainties ( the intercept @xmath175 ) . such increased bias is due to the larger amount of luminosity scatter needed to obtain the same observed @xmath176 starting from a smaller @xmath177 . the opposite occurs when the real age spread increases and the relative contribution from apparent luminosity spreads is smaller . in the presence of only luminosity spreads , the assumption of @xmath174 leads to a very severe bias : considering the full mass range ( solid lines ) , @xmath19 is always @xmath178 , even for @xmath179 , @xmath93 and bias type a. since in reality a combination of cases a , b , and c is expected , a value @xmath180 or more would be needed to reproduce the slope observed in the onc and other young regions ( * ? ? ? * , @xmath3 ) . with such an instantaneous decay of @xmath0 , no accreting sources would be detected in clusters older than a few myr . we also explore the limiting case of a fully coeval young stellar population , whose apparent age spread from the measured hrd is solely due to uncertainties in @xmath24 ( bias of types a and b ) , or a luminosity spread induced by differences in the protostellar accretion history ( bias of type c ) . we start with a @xmath181 myr population and add a luminosity spread as in the previous subsections . the input value @xmath17 is now meaningless , as all simulated stars share the same age ; instead , we vary the amount of luminosity spread and each time we fit the measured @xmath42 @xmath6 relation determining @xmath19 . again , we simulate @xmath81 or @xmath93 , and study separately the 3 types of bias a , b and c. the results are shown in figure [ fig : simplespread_coeval ] , in terms of @xmath19 as a function of the apparent isochronal age spread . the general trend is qualitatively close to that obtained for a small intrinsic age spread ( e.g. , @xmath182 , see figure [ fig : simplespread_tautau013 ] upper panel ) and @xmath183 . for bias of type a , correlated uncertainties generate a positive value of @xmath19 , i.e. , an apparent temporal decrease of @xmath0 , with @xmath184@xmath185 for 0.4 dex of apparent age spread and depending on the assumed @xmath66 . on the contrary , bias of type c produces a negative @xmath19 , down to @xmath186 , and type b lies in between . as the luminosity spread will likely be some combination of the 3 cases , then accretion rates in a coeval population should show an apparent correlation with age with @xmath187 . this result , leaving aside the additional caveats reported in section [ section : discussion ] , provides a test for the coevality of any young population : it is difficult to advocate coevality for a cluster in which @xmath188 . we analyze now the bias in @xmath2 from uncertainties in @xmath20 alone , using the same technique we adopted for spreads in luminosity ( section [ section : simulations - simplespread ] ) . we start with two values for the real age spread , @xmath189 and @xmath190 . the @xmath20 of the simulated stars are displaced with gaussian shifts with standard deviation of either half or one spectral subtypes ; we assume the temperature scale of @xcite for the conversion between spectral types and @xmath20 . in principle an error in an assigned spectral types may propagate to the estimated @xmath24 as well , in two cases : a ) when it leads to an error in the estimated @xmath21 , this case will be explored in section [ section : simulations - teffav ] ; b ) when the bolometric corrections used to obtain @xmath35 are quite sensitive to @xmath20 . the latter is generally not a significant effect , when @xmath35 is derived from apparent magnitudes close to the peak of the stellar seds , e.g. , photometry in the red optical range for low - mass stars . we consider that the bias in @xmath166 from @xmath20 errors always follows type a ( see section [ section : simulation_howtheywork ] ) . this is a fair approximation , since , generally , errors in @xmath20 do not significantly affect the accretion luminosity contrast @xmath48 . strictly speaking , this holds if the bolometric correction of the photospheric flux , at the wavelength of the indicator used to estimate @xmath30 ( e.g. , h@xmath31 ) , is weakly dependent on @xmath20 . figure [ fig : teff_tautau ] shows the results : as for the luminosity spreads , @xmath191 . the bias is significant even for the modest error of 1/2 a spectral subtype , which is practically the smallest uncertainty achievable with spectroscopy : the flattening is at least @xmath192 when the entire mass range is considered ( solid lines ) and starting from a large initial age spread . for larger age spreads the bias is smaller , because the stellar population is already well spread in the hrd , and thus an additional uncertainty in @xmath20 introduces a smaller relative broadening of the width of the age distribution . when the parameters of pms stars are derived from optical data in the presence of differential @xmath21 , an error in @xmath20 also to leads to a correlated error in @xmath21 , and thus in @xmath35 . this is because , given an observed color for a star , if @xmath20 is overestimated ( e.g. , from inaccurate spectral type assignment ) , the intrinsic photospheric color will be underestimated , the reddening overestimated , and the intrinsic luminosity also overestimated . thus , @xmath20 errors tend to produce diagonal shifts in the hrd , whose inclination depends on the photometric bands used to derive stellar parameters , and typically will be steeper for shorter wavelengths . this type of error will also apply to any analyses where a uniform reddening is assumed for a population where some level of differential extinction exists . we run another set of simulations as before , and apply a spread in the hrd due to extinction uncertainties . to define the correlation between @xmath59 and @xmath24 , we assume that the hrd is derived from an optical @xmath193 versus @xmath194 cmd . first , considering the galactic reddening law of @xcite , we derive the @xmath194 and @xmath193 shifts due to @xmath195 . these are then converted into @xmath20 and @xmath24 shifts , as a function of @xmath20 , based on synthetic photometry using the code ta - da @xcite and assuming the bt - settl synthetic spectra from @xcite . we randomly displace each source in @xmath20 and @xmath24 from a uniform probability distribution in the range @xmath196@xmath197 . we run simulations for two amounts of differential extinction : a low ( @xmath198 mag ) and an intermediate one ( @xmath199 mag ) , and , as before , for two intrinsic age spreads : @xmath200=0.2 and 0.4 dex . as for errors in @xmath20 ( section [ section : simulations - teff ] ) , this uncertainty follows a type a bias ( section [ section : simulation_howtheywork ] ) , since the relative excess @xmath48 is only weakly dependant on @xmath21 . this holds in particular when @xmath49 excess together with optical fluxes are used to measure accretion @xcite . the results are shown in figure [ fig : teffav_tautau ] . once again , the @xmath201 ; however , the relative bias is smaller than that of simple @xmath20 errors ( figure [ fig : teff_tautau ] ) . this is because the reddening vector , in @xmath202 and @xmath203 bands , is almost parallel to the low - mass pms isochrones , and therefore extinction does not significantly alter the isochronal ages . this is not the case when restricting to intermediate - mass stars ( @xmath204 m@xmath15 ) , for which we find @xmath19 is 2030% smaller than the @xmath17 . estimates of @xmath0 and stellar parameters rely on stellar evolutionary models ; hence , the derived @xmath2 remain model dependent . there are open problems with current pms isochrones , as the different families of models differ in their predictions @xcite and often deviate systematically from the data @xcite . in what follows we investigate if the adopted set of models , or an empirical calibration of them in the hrd , have important effects on the biases in @xmath2 . throughout our simulations , we have used @xcite isochrones ; here we test if the bias in @xmath2 is heavily dependent on the assumed evolutionary models . to this end , we consider isochrones from @xcite , as they present some of the largest systematic differences compared to siess . in particular , their models predict a much younger age for pms stars , for instance , a mean age of 1 myr for the onc , less than half of that predicted by siess isochrones . thus , to generate a synthetic population in the hrd comparable with our previous simulations , we simulate cluster populations with a mean age of 1 myr from @xcite isochrones . for simplicity , we test only the case of a luminosity spread in the hrd , and use the same assumptions as in appendix [ section : simulations - simplespread - onc ] , i.e. , a real age spread @xmath86 and additional scatter in @xmath24 to reach an apparent measured @xmath205 . the resulting @xmath19 versus @xmath17 is shown in figure [ fig : simplespread_tautau02_dantona ] . the bias in @xmath2 for the 3 types of uncertainties is almost identical to that we obtained for the analogous simulations from siess models ( figure [ fig : simplespread_tautau02 ] ) , despite the large systematic difference in the predicted absolute ages between these two grids of evolutionary models . this is due to the fact that the spacing in @xmath24 between contiguous isochrones ( in @xmath18 ) remains similar even when different families of models show large offsets in their absolute age scales . our result shows that the ratio @xmath206 is largely unaffected by the choice of evolutionary models . furthermore , for a given cluster the value @xmath19 does not change much if one uses different evolutionary models ; this was shown e.g. , by @xcite for the onc , and also supported by the fact that measurements of @xmath2 in the literature , on a multitude of young clusters , and based on a heterogeneity of assumptions and models do not show noticeable differences in their results . hence we argue that the true value of @xmath2 and the biases therein suggested by the modeling in this paper are not induced by our particular choice of adopting the @xcite isochrones . -dependent luminosity shift sufficient match the observed population . [ fig : acchistory_cmd ] ] in the previous subsection we analyzed the effect of discrepancies between different evolutionary models . here we focus on their general deviation from the observations . one of these discrepancies is the apparent correlation between mass and average age for pms clusters @xcite , as in a young system intermediate - mass stars appear generally older than low - mass stars , by several myr ( see figure [ fig : acchistory_cmd ] . if this systematic difference were real , it implies a significant temporal delay of a few myr in the formation of low - mass stars in a cluster . this seems unlikely and the trend probably indicates an inaccurate prediction by the evolutionary models . there are many possible origins for this inconsistency . several assumptions in the modeling of pms mass tracks affect the predicted stellar parameters in different ways depending on the mass @xcite : for example , reducing the mixing length could remove the mass - age correlation @xcite . alternatively , whereas standard pms tracks are computed assuming an evolution without accretion , the inclusion of mass accretion in the modeling also affects the predicted parameters , in particular leads to smaller radii - therefore fainter luminosities @xcite . this is expected to have a negligible effect for low - mass stars @xcite , but could be significant at masses @xmath207 . we consider this second scenario , and assume that the standard isochrones overestimate the predicted @xmath24 for intermediate masses . we quantify the required correction empirically , based on the measured hrd of the onc from @xcite . this is shown in figure [ fig : acchistory_cmd ] , together with @xcite evolutionary models . the dashed red line is the isochrone corresponding to the mean isochronal age of the cluster ( 2.5 myr ) . we have computed the required shift in @xmath24 from this isochrone , as a function of @xmath20 , to match the mean observed luminosities ; the resulting locus is the blue dashed line in figure [ fig : acchistory_cmd ] . thus , the two lines correspond , respectively , to the parameters that these models _ do _ predict , and those they _ should _ predict . since observations , as well as our simulations in this paper , are based on the standard isochrones , this is a further bias we investigate . for our purposes , it is practical to treat this systematic effect in reverse : we assume that the siess isochrones are correct , and that some other effect shrinks the stellar radii of an amount equal to the difference between the two lines in figure [ fig : acchistory_cmd ] . we run a simulation as in the previous section , where in this case the positions of the simulated sources in the hrd are displaced with a rigid , @xmath20-dependent shift in @xmath34 , instead of with a random scatter . this case now follows a bias of type c ( see section [ section : simulation_howtheywork ] , since effectively the measured stellar @xmath34 @xmath24 are correct , whereas @xmath7 and @xmath6 are wrongly inferred from the models . the results are shown in figure [ fig : tautau_acchistory ] . considering the entire sample ( solid lines ) , @xmath19 is generally lower than @xmath17 . when we separate the mass ranges , it is evident that this is due mainly to the contribution of intermediate mass stars ( dash - dotted lines ) , for which @xmath17 is underestimated by 1525% . as anticipated in section [ section : intro ] , sample incompleteness , from either lack of detection of stellar members , or lack of measurement of @xmath0 for weak accretors , introduces selection effects in the @xmath0@xmath6 plane . figure [ fig : selection_effects ] shows a quantitative example of these effects . we have considered 3 stellar masses , 0.2 , 0.6 and 2 m@xmath15 , and computed the variation of the stellar luminosity and accretion excess with age and @xmath0 . we assumed that the h@xmath31 excess is the observational indicator to measure @xmath0 , and the specific technique is narrow band photometry with the @xmath208 wide field camera 3 filter @xmath209 . we computed , for every @xmath6 and @xmath0 , both the measured excess in magnitudes in @xmath209 and the total absolute magnitude ( photosphere and excess ) . this was done as follows : first , we applied equation [ equation : mdot2 ] to translate @xmath0 into @xmath48 . from the latter we derived @xmath210 assuming the relation from @xcite @xmath211 next , we derived the equivalent width of the @xmath49 line excess from : @xmath212 where @xmath213 is the photospheric spectral energy distribution of the star , for which we assumed bt - settl synthetic spectra from @xcite . last , we derived the associated ( positive ) excess in magnitudes from : @xmath214 where @xmath215 is the equivalent width of the filter . we also computed , for every mass and age , the photospheric magnitude @xmath216 by means of synthetic photometry using the tool ta - da @xcite , again using bt - settl models . finally , the observed magnitude is the sum of photosphere and excess : @xmath217 . the solid contours in figure [ fig : selection_effects ] indicate constant absolute magnitude : at very low @xmath0 the excess is negligible , thus @xmath218 is dominated by the photosphere and depends only on age , leading to vertical contours . at higher @xmath0 , however , the excess which correlates with the stellar parameters , leads to shallow correlations between @xmath0 and age at a fixed magnitude . if the detection limit is shallow enough ( for instance , @xmath219 ) , in the low mass range ( e.g. 0.2 m@xmath15 ) the detection threshold will be inclined in @xmath0 - @xmath6 plane , as the contours show . depending on the depth of the sample , such a cut alters the underlying correlation between these two quantities . the dotted contours in figure [ fig : selection_effects ] , on the other hand , indicate constant magnitude excess @xmath220 , respectively 0.1 mag ( lower lines ) and 1 mag ( upper line ) . again , these loci are inclined in the plane , so when the measurement of @xmath0 is limited by the ability to isolate a small accretion - related excess from noises or other uncertainties , the sample of measured @xmath0 will suffer from a correlated selection effect with age . as shown , in this case the problem dominates at large masses , where high values of @xmath0 produce small flux excesses . we model a more realistic representation of these effects to understand their overall influence on @xmath19 for a stellar population . we simulate a young population as in section [ section : simulations ] , assuming a mean age @xmath76 myr , an age spread @xmath221 dex , and assign values of @xmath0 assuming with b=2 and variable @xmath17 . for each generated star , considering its stellar parameters and @xmath0 , we compute both the magnitude excess in @xmath49 , @xmath220 , and the total ( absolute ) magnitude @xmath218 as we did before for figure [ fig : selection_effects ] . we introduce a detection limit @xmath222 , for which the assumed signal to noise ratio is 3 and the completeness abruptly drops from 100% to 0 . this simplified choice allows us to isolate the general bias due selection effects without adding unnecessary scatter due to the stochasticity of source detection . we assign to every magnitude @xmath218 a photometric error , assuming that it is only due to shot noise from photon counting , thus : @xmath223 as another general assumption , we assume that h@xmath31 photometry is responsible for all the incompleteness and photometric uncertainty affecting the sample . since in practice photometric estimates of @xmath0 generally require additional broad band photometry to isolate the excess ( e.g. , @xmath224 and @xmath193 , @xcite ) , this implies the latter should be significantly deeper than h@xmath31 . figure [ fig : selection_effects_population ] shows the result for 3 combinations of @xmath17 and @xmath222 producing a significant incompleteness in the pms mass range . depending on the magnitude limit , some stars are too faint to be detected . also , another fraction are detected , but their accretion excess is smaller than the photometric error , leading to an average estimate of @xmath0 compatible with 0 and the removal of these sources from the @xmath107@xmath18 plane . the remaining detections show a correlation of @xmath0 with age , determined through linear fit , which differs from the real one , and is much shallower for the 3 cases shown in figure [ fig : selection_effects_population ] . we have inspected a broad range of assumed @xmath17 , @xmath66 and @xmath222 ; the examples in figure [ fig : selection_effects_population ] appear representative of the general behavior , except if @xmath225 , in which case the diagonal threshold in the @xmath0 increases @xmath19 by a few tenths . the scatter of the @xmath0 values at a given age shown in figure [ fig : selection_effects_population ] is solely due to the mass dependence introduced by @xmath93 . in reality an additional scatter at any @xmath7 and @xmath6 is expected , as the measured @xmath0 depends not only on @xmath7 and @xmath6 , but can also vary star by star due to differences in the disk properties , temporal variability of @xmath0 , or due to poorly understood correlations with the stellar parameters ( see e.g. , figure [ fig : simplespread_mdot_age ] ) . we test if this modifies the bias due to selection effects . we repeat the same simulation adding a random gaussian scatter to @xmath42 before computing the photospheric and excess magnitudes . we assume a @xmath226 width of 0.65 dex : this is the extent of the random star - to - star variations of @xmath0 needed to obtain a total standard deviation of the data points in the @xmath107@xmath18 plane that matches the 0.85 dex found in the onc by @xcite . we also add 0.1 mag in quadrature to the assumed photometric errors @xmath227 from equation [ equation : photerrha ] . this is to avoid the results being altered by unrealistically small photometric errors at bright magnitudes . the results are shown in figure [ fig : selection_effects_population_scatter ] ; data corresponding to good detections and non detections appear scattered , but the general trends are not very different from those of figure [ fig : selection_effects_population ] . we can not derive a general correction for these types of selection effects , as in reality they depend on multiple factors : e.g. , the detection completeness functions and the photometric error distribution in each band , as well as the other overall uncertainties affecting both magnitudes and stellar parameters , and the possible inaccuracy of the atmosphere models we assumed in our modeling . however , if the studied stellar population is characterized to a much fainter luminosity than those for which @xmath166 is assessable ( e.g. , through broad band photometry ) , it is useful to study the fraction of accretors as a function of age ( and mass ) . this is because , if the age spread within a region is sufficiently small , one would not expect any significant variation of the fraction of detected accretors with age or mass . if on the contrary , as evident from the right panels of figures [ fig : selection_effects_population ] and [ fig : selection_effects_population_scatter ] , at low masses the detected accretors tend to be all younger than the cluster average age , strong selection effects in the samples may be present and bias @xmath19 .
the temporal decay of mass accretion in young stars is a fundamental tracer of the early evolution of circumstellar disks . through population syntheses , we study how correlated uncertainties between the estimated parameters of young stars ( luminosity , temperature , mass , age ) and mass accretion rates @xmath0 , as well as observational selection effects , can bias the temporal decay of mass accretion rates ( @xmath1 ) inferred from a comparison of measured @xmath0 with isochronal ages in young stellar clusters . we find that the presence of realistic uncertainties reduces the measured value of @xmath2 by up to a factor of 3 , leading to the inference of shallower decays than the true value . this suggests a much faster temporal decay of @xmath0 than generally assumed . when considering the minimum uncertainties in ages affecting the orion nebula cluster , the observed value @xmath3 , typical of galactic star forming regions , can only be reproduced if the real decay exponent is @xmath4 . this effect becomes more severe if one assumes that observational uncertainties are larger , as required by some fast star formation scenarios . our analysis shows that while selection effects due to sample incompleteness do bias @xmath2 , they can not alter this main result and strengthen it in many cases . a remaining uncertainty in our work is that it applies to the most commonly used and simple relationship between @xmath0 , the accretion luminosity and the stellar parameters . we briefly explore how a more complex interplay between these quantities might change the results . [ firstpage ] stars : formation , pre - main - sequence , variables : t - tauri , fundamental parameters , circumstellar matter ; hertzsprung - russell and colour - magnitude diagrams
the spectra of bright blazars obtained by _ fermi gamma - ray space telescope _ ( _ fermi _ ) large area telescope ( lat ) showed clear deviations from a power - law shape @xcite . there spectra could not be described by smooth functions such as exponentially cutoff power - law or a log - parabola ( log - normal distribution ) , but were found to be better described by a broken power - law . the derived break energies lying in the 110 gev energy range @xcite were rather stable @xcite . those breaks seemed puzzling : the hypothesis that the break is caused by photon - photon annihilation through @xmath2 pair production was considered and rejected by @xcite , who argued that `` to produce a break in the 110 gev , the photon field should have an energy peaking in the 0.050.5 kev range , which excludes the broad - line region peaking in the uv . '' the conclusion that such large energies of the target photons were required was based on an erroneous assumption that the break energy correspond to the maximum cross - section for pair production . @xcite suggested that the breaks should actually appear at the energies close to the threshold for a corresponding reaction , where the opacity has a sharp rise . the observed breaks at a few gev correspond then well ( correcting for the redshift ) to the lyman recombination continuum ( lyc ) and ly@xmath3 emission of ionized he . they also showed that the inner part of the blr can provide sufficient flux of lyman lines and lyc to provide enough opacity for gev photons and to produce a spectral break . @xcite further demonstrated that the data for a number of bright blazars are well described by a power - law spectrum modified by the absorption within the blr . the fits with this model were acceptable and the reduction in @xmath4 ( compared to a simple power - law model ) was very significant . similar @xmath4 could also be achieved with the broken power - law model which , however , does not have any physical basis . a high significance of the spectral breaks partially results from a null hypothesis for the underlying spectrum , which is assumed to be a power - law . however , a typical blazar has a curved spectrum extending over many orders in energy and peaking in the mev gev range . this implies that the spectrum in the _ fermi _ energy band should be slightly concave and the power - law null hypothesis gives an overestimated significance of the break . as a more realistic null hypothesis one can take a lognormal distribution ( log - parabola ) , which is the simplest way to introduce a curvature in logarithmic coordinates . @xcite studied in details the spectrum of the exceptionally bright flat - spectrum radio quasar ( fsrq ) 3c 454.3 and indeed found that the spectrum below the absorption break significantly differs from the power - law and can be well described by a log - parabola . with this null hypothesis the statistical significance of any break is lower than with a power - law hypothesis . however , the absorption break in the time - integrated spectrum of 3c 454.3 was still highly significant and its energy coincided with the predicted one from lyc absorption . while the main attention of the cited papers has been payed to the gev breaks , @xcite and @xcite also revealed hydrogen lyc breaks at @xmath020 gev , which were less significant and less impressive because of a lower photon statistics in that energy range actually , it is obvious that in most cases h ly radiation should produce a stronger absorption feature that ly emission . that is why it is important to revisit the spectral analysis of bright blazars with a higher statistics that was accumulated since previous works . another , more important reason to revisit previous results is a slight but significant difference in the blazar spectra obtained with the new pass 7 version of the _ fermi_/lat data and detector response and those obtained with the older pass 6 version . the new spectra look smoother and this is a reason to suspect that the sharp breaks at a few gev were the artifacts of the pass 6 response function partially or completely . we use _ fermi_/lat photon data for 1740 days ( from 2008 august 6 till 2013 may 12 ) using the new pass 7 event classification , selecting photons of clean class , and imposing the cut on the zenith angle at @xmath5 . we used ` p7_clean_v6 ` response function . the diffuse background was calculated using the background model elaborated by the _ fermi _ team . it should be noted that the pass 7 response function significantly differs from the pass 6 one in at least two aspects : 1 . it has a wider point spread function , 2 . it has a different effective area function ( see fig . [ fig : acceptance ] ) . the second fact is of crucial importance for the analysis of the gev energy breaks because the pass 6 response has a hump starting at 45 gev , which introduces a break in the photon spectrum at energies close to the lyc absorption ( see fig . [ fig:3c454 ] for comparison of the spectra of 3c 454.3 obtained with passes 6 and 7 ) . the break is not very strong : it changes the index of the power - law by @xmath00.1 . if one then fits the resulting spectrum with a broken power law , one obtains a significant break . we believe that the break in the pass 6 effective area is not real , because there is no clear reason for existence of such a feature at this energy . moreover , it looks strange that the hump appears only in the effective area of the front detector . the pass 7 version of the effective area does not have such a feature and looks more natural . the energy - dependent exposure function was calculated using the spacecraft pointing history : @xmath6 where @xmath7 is the time interval of interest , @xmath8 is the angle between the detector axis and the direction to the object and @xmath9 is the detector effective area at energy @xmath10 . we accumulated counts in the circle centered at the source location with the energy - dependent radius @xmath11 , where @xmath12 is the radius of 90% event containment , which was calculated with monte - carlo integration of the point - spread function assuming the isotropic distribution of the exposure angle . then number of counts in each energy bin was corrected to the containment factor for @xmath13 . in order to reveal the absorption features in the blazar emission , we analyzed the spectra of individual bright fsrqs as well as the stacked spectra of various samples . we have selected a sample of 15 brightest objects from the 2nd fermi catalog @xcite using the following criteria : 1 . the total number of counts above 1 gev after background subtraction is above 1800 , 2 . the signal / background ratio exceeds 2 , 3 . known redshift ( according to the 2nd fermi catalog ) , 4 . classification as a fsrq , or as a low - synchrotron peak bl lac if its redshift exceeds 0.5 ( which would mean that the latter is probably a misclassified fsrq ) , 5 . there is no strong source confusion . the brightest blazar 3c 454.3 was excluded from the stacking analysis and studied only individually because of its exceptional brightness , which is comparable to the total signal from other selected objects . the remaining 14 fsrqs bright above 1 gev constitute group 1 , where we hoped to reveal the absorption break at @xmath05 gev associated with lyc emission @xcite . we also selected group 2 of eight blazars with the highest number of counts above 5 gev ( more than 200 after background subtraction ) . this sample was more promising to study h lyc absorption break at @xmath020 gev . all such blazars also belong to group 1 despite the selection was independent . llcr 3c 454.3 & & 0.859 & 01740 + 4c + 55.17 & 1 , 2 & 0.896 & 01740 + pks 0537@xmath14441 & 1 , 2 & 0.892 & 01740 + pks 2326@xmath14502 & 1 & 0.518 & 6001740 + 4c + 21.35 ( pks 1222 + 21 ) & 1 , 2 & 0.433 & 3501100 + pks b1424@xmath14418 & 1 , 2 & 1.522 & 12001740 + pks 0426@xmath14380 & 1 , 2 & 1.111 & 01740 + pks 0454@xmath14234 & 1 , 2 & 1.003 & 01740 + pks 0727@xmath1411 & 1 & 1.591 & 0 800 + pks 1510@xmath1408 & 1 , 2 & 0.360 & 01740 + 3c 279 & 1 & 0.536 & 01300 + pks 1502 + 106 & 1 , 2 & 1.893 & 0 500 + b2 1520 + 31&1 & 1.484 & 01400 + pks 0235 + 164 & 1 & 0.940 & 0 400 + 4c + 38.41 & 1 & 1.813 & 01740 + + mrk 421 & 3 & 0.030 & 01740 + 3c 66a & 3 & 0.444 & 01740 + s5 0716 + 714 & 3 & 0.310 & 01740 + pks 2155@xmath14304 & 3 & 0.117 & 01740 + to optimize the signal - to - noise ratio , we have selected for each blazar a time interval , when its flux substantially exceeded the background . limits of these time intervals are given in table [ tab : groups ] . in order to prepare the stacked spectra , we first derived individual spectra using the energy bins with the width of 0.1 in decimal logarithm . the bin edges were adjusted in such a way that they are the same in the object rest frame . each spectrum was blue - shifted by a factor @xmath15 . we summed up the obtained spectra with their absolute normalization , so that the spectra of the brighter objects have a larger contribution . such approach optimizes relative statistical errors . the blazar spectra were modeled with a lognormal function with the superimposed absorption by the blr emission ( see sect . [ sec : opacity ] ) . we use the opacity computed for different ionization parameters @xmath16 of the blr as described in @xcite . we also checked a simpler monochromatic absorber model ( h and lyc ) , which is less realistic but helps to estimate separate contributions of h and he emission . we have also constructed two comparison spectra , where breaks are not expected , to make sure that the detection of the breaks is not an artifact of the detector response . the first one is the stacked redshift - corrected spectrum of four bright bl lacs ( group 3 in table [ tab : groups ] ) . the second one is the `` empty '' sky spectrum . the sky was sliced into 72000 bins of equal area ( @xmath17 str ) and the photons were collected from the bins where number of photons above 100 mev is less that 300 , these bins constitute about 0.226 fraction of the sky . the main contribution to this spectrum is from the high - energy protons producing pions and a smaller contribution of unresolved bl lacs . statistical errors were treated as gaussian , except in a few bins at higher energies , where the number of photons is low , we use poisson likelihood adding @xmath18 to @xmath4 ( here @xmath19 is the number of counts in the bin and @xmath20 is the prediction of the model ) . the number of such bins is small and the meaning of @xmath4 is not significantly affected . for the minimization we use the standard code minuit from the cern library . gamma - ray photons emitted presumably by the relativistic jet emanating from the black hole are strongly beamed . they propagate through the radiation field made by the accretion disk , the blr and the dusty torus are potentially can be absorbed by photon - photon pair production . the disk radiation moves nearly parallel to the photon beam and therefore does not interact efficiently . the infrared photons from the dust absorb radiation mostly in the tev range . thus , the most important source of opacity is the optical / uv ( nearly ) isotropic radiation from the blr . the opacity depends on the blr spectrum , which is computed using spectral synthesis code xstar ( version 2.2 , * ? ? ? * ) as described by @xcite . we repeat here the basic assumption for completeness . the ionizing spectrum of a quasar is taken as a sum of the standard multi - color accretion disk plus a power law of total luminosity 10% extending to 100 kev @xcite . the blr clouds are assumed to be simple slabs of constant gas density and a clear view to the ionizing source . the hydrogen column density was fixed at @xmath21 @xmath22 and the blr spectra were computed for different ionization parameters @xmath23 varying from 10@xmath24 to 10@xmath25 . if one assumes a dependence of the cloud density on distance from the black hole @xmath26 , then @xmath27 and our ionization range would correspond to the distance interval between 1 and 0.01 pc to the ionizing source for a quasar luminosity @xmath28 erg s@xmath29 . in cgs units . ] the scaling and the estimated distances are very approximate and are model - dependent . if the blr were to emit only one line at energy @xmath30 , the optical depth for a @xmath1-ray photon of energy @xmath10 through the region of size @xmath31 filled with isotropic soft photon field of column density @xmath32 would be @xmath33 , where @xmath34 @xmath35 is the thomson cross - section , @xmath36 is the line luminosity , @xmath37 and @xmath38 is the angle - averaged cross - section for photon - photon pair production ( see , e.g. * ? ? ? * ; * ? ? ? note , that @xmath38 has threshold @xmath39 ( i.e. at @xmath40 ) and has a peak of about @xmath41 at @xmath42 . the blr spectrum , of course , contains many lines and recombination continua . one can introduce the cross - section weighted with the photon distribution : @xmath43 where @xmath44 . the spectrum transmitted through the blr is attenuated as @xmath45 , where @xmath46 , and @xmath47 is computed using equation ( [ eq : taut ] ) replacing @xmath36 by @xmath48 and @xmath30 by the mean photon energy of the blr : @xmath49 as an illustration we present the results of absorption of a power - law spectrum by the blr of different ionizations in fig . [ fig : blr_abs ] fixing the total blr photon column density at @xmath50 @xmath22 , which corresponds to @xmath51 . for the considered @xmath47 , the flux drops at most by a factor of 34.5 depending on @xmath16 , corresponding to the maximum optical depth of about 1.11.5 . note that the transmitted spectrum in the range from 30 gev to 1 tev has nearly the same slope as the intrinsic one at larger @xmath16 , because the opacity is nearly constant in this range . the opacity drops at energies above 1 tev and the spectrum recovers . we see that the lyc breaks at 5 gev are more pronounced at high ionizations @xmath52 , while the h lyc breaks are seen at any @xmath53 . this allowed @xcite to introduce a simpler double - absorber model for @xmath1-ray opacity , where the blr spectrum is replaced by the strongest emission features of h and lyc . for low ionization , one can even consider only a single absorber due to the h lyc . llccccccccc 3c 454.3 & 55.0/21 & 38.3/20 & [email protected] & 28.1/19 & [email protected] & 29.6/20 & [email protected] & 25.8/19 & 8.8 @xmath541.7 & 5.5@xmath55 + pks b1424@xmath14418 & 23.0/23 & 19.0/22 & [email protected] & 19.0/19 & @xmath560.3 & 18.0/20 & [email protected] & 23.8/20 & @xmath564.2 & + pks 0426@xmath14380 & 42.3/23 & 27.5/22 & [email protected] & 27.5/21 & @xmath560.4 & 22.9/22 & [email protected] & 36.4/22 & 6.3@xmath54 2.6 & 4.5@xmath55 + pks 1502 + 106 & 30.5/21 & 22.5/20 & [email protected] & 22.5/19 & @xmath560.17 & 21.1/20 & [email protected] & 30.2/20 & [email protected] & 3@xmath55 + pks 0537@xmath14441 & 46.0/23 & 34.1/22 & 3.2@xmath57 & 34.1/21 & @xmath561.4 & 29.3/22 & [email protected] & 40.6/22 & [email protected] & 4@xmath55 + pks 0454@xmath14234 & 35.7/23 & 28.7/22 & [email protected] & 27.5/21 & @xmath560.4 & 28.2/22 & [email protected] & 34.1/22 & [email protected] & 2.5@xmath55 + 4c + 21.35 & 35.7/22 & 33.1/21 & [email protected] & 25.5/20 & 1.8@xmath58 & 33.4/21 & 25@xmath59 & 23.8/21 & [email protected] & 3.5@xmath55 + pks 1510@xmath1408 & 20.0/21 & 20.0/20 & @xmath560.5 & 20.0/19 & @xmath560.5 & 20.0/20 & @xmath561.4 & 20.0/20 & @xmath560.9 & + 4c + 55.17 & 66.0/21 & 57.9/20 & [email protected] & 57.0/19 & @xmath560.8 & 57.0/20 & [email protected] & 58.0/20 & [email protected] & + group 1 & 44.0/23 & 36.8/22 & [email protected] & 36.4/21 & @xmath560.05 & 30.2/22 & [email protected] & 44.0/22 & @xmath561.2 & 3.5@xmath55 + group 2 & 65.6/23 & 42.6/22 & [email protected] & 42.6/21 & @xmath560.1 & 31.6/22 & [email protected] & 52.9/22 & [email protected] & 6@xmath55 + group 3 ( bl lacs ) & 35.4/22 & 33.7/21 & [email protected] & 33.7/20 & @xmath560.1 & & & & & the results of the spectral fits for 3c 454.3 , all objects of group 2 and for the stacked spectra are presented in table [ tab : fits ] and some of them are shown in figs . [ fig : spe_individ ] and [ fig : spe_stacked ] . the best - fit model for all objects except 3c 454.3 and 4c + 21.35 is that of the blr emission with lower ionization degree @xmath60 = 1.5 . in this ionization state , the contribution by absorption is small and one can see from table [ tab : fits ] that the double absorber model h+ lyc does not improve the fits with respect to the single h lyc absorber . this means that in most spectra there is no sign of lyc absorption . the exceptions are 3c 454.3 and 4c + 21.35 ( pks 1222 + 21 ) where the presence of absorption is detected at @xmath61 level . the best - fit model for absorber in these sources is blr emission with @xmath62 . in the stacked spectra of both groups , the situation is similar : the addition of does not change @xmath4 significantly . the typical optical depth , @xmath47 , for the best - fit blr emission model ( mostly with @xmath63 ) was measured to between 4 and 20 . this corresponds to the maximum optical depth of about 0.42.2 ( see blue dashed line in fig . [ fig : blr_abs ] ) and the flux reduction at @xmath0100 gev by a factor of 1.59 . to estimate the absorption optical depth that is contributed by h and emission only , one can consider corresponding optical depths from single- or double - absorber models ( see table [ tab : fits ] , columns 4 and 6 ) . spectra of the six from the nine brightest ( above 5 gev ) blazars demonstrate clear absorption breaks dominated by the h lyc absorption . the significance of these breaks ranges from 2.5@xmath55 to 5.5 @xmath55 . the typical optical depth by h lyc only is @xmath6424 , which can be converted directly to the column density of lyc ( plus ly@xmath3 ) photons on the line of sight to the @xmath1-ray emitting region @xmath65 . the stacked spectrum of group 2 ( which does not include 3c 454.3 ) has 6@xmath55 significance of the break . this is the most confident and the most conservative demonstration that the h lyc absorption in bright blazars is a very typical phenomenon . previously @xcite , @xcite , and recently @xcite revealed this absorption in individual sources . however , in first and in the last papers a less conservative assumption of a power law null hypothesis was used . as for the gev breaks associated with absorption ( now with a more accurate pass 7 effective area ) , they get a status of a rare phenomenon . there are indications of such breaks in two objects : 3c 454.3 ( 3@xmath55 ) and 4c + 21.35 ( almost @xmath66 ) , see fig . [ fig : spe_individ ] . detection of the breaks due to h lyc absorption has high significance . the amplitude of the spectral deviation from the null hypothesis is factor of 2 in the case of the stacked spectrum ( group 2 ) and factor of 3 in the case of pks 0426@xmath14380 . this is much above possible systematic errors like uncertainties in the response function . the breaks due to lyc are much weaker than was claimed before @xcite mostly because of the difference between pass 6 and pass 7 detector response functions . a sharp rise in the pass 6 effective area ( see fig . [ fig : acceptance ] ) made the spectral break observed at 23 gev sharper and more significant . however , we believe that the breaks observed in two objects , 3c 454.3 and 4c + 21.35 , are real . first , now there is no any feature in the lat response function at the corresponding energy . second , deviation of the spectra in the rest - frame range 520 gev from the extrapolation of the lognormal function fitted to the data below 5 gev reach at least 50% ( see fig . [ fig : spe_individ ] ) , much above the uncertainty in the lat response . it should be noted that the 3@xmath55 significance level is quite serious in this case , because the detection of the effect does not include `` hidden trials '' like thresholds adjustments and sample manipulation . this significance is also very conservative as it was measured for the lognormal null hypothesis , not a power - law . for 3c 454.3 , we do not expect systematical errors associated with the background subtraction or source confusion due to exceptionally high @xmath1-ray brightness . 4c + 21.35 is weaker and there could be some systematics , e.g. , an underestimated soft background : the soft part of the spectrum could be lower and the spectrum could then be fitted with a narrower lognormal distribution without absorption . therefore , the absorption break in 4c + 21.35 needs further studies . in order to prove the presence of the gev breaks in fsrq , we checked whether the breaks appear also in the comparison spectra ( see fig . [ fig : sky_bllac ] ) , where they are not expected . we see that the stacked bl lac spectrum ( group 3 ) is well described by a power - law ( with some deviations because of imperfectness of the response function ) without any breaks ( except possibly at 100 gev ) . the upper limits on the opacity due to h or he lyc are significantly below the detected opacity in the bright blazars ( see table [ tab : fits ] ) . the `` empty '' sky spectrum also does not show any signs of the breaks in the 220 gev range . thus , inaccuracies in the detector response are unlikely to affect our conclusions . we also made a simple test whether a different null hypothesis for the underlying spectrum would remove the necessity of the breaks . instead of the log - normal function , we assumed , following the referee s suggestion , a biquadratic function @xmath67 $ ] , which has the same number of free parameters as our log - normal distribution with blr absorption . this ad hoc function gives a slightly better fit with @xmath68 for 3c 454.3 , but a much worse fit with @xmath69 for the group 2 spectrum ( see table [ tab : fits ] ) . interestingly , a fit for 3c 454.3 gives @xmath70 , so that the spectral curvature is fully determined by the quartic term ; what is the physical meaning of such a model is a mystery to us . our main results can be formulated as follows . * we find that the 20 gev breaks due to h lyc are ubiquitous . they are statistically significant in the majority of the bright blazars as well as in the stacked redshift - corrected spectra . * the 5 gev breaks due to lyc remain significant only in two objects . * a more complicated function describing the underlying spectrum can change the significance of the break existence . an ad hoc biquadratic function of @xmath71 gives a slightly better fit that the physically justified lognormal function with the blr absorption for 3c 454.3 , but a much worse fit for the stacked blazar spectrum . thus this model does not eliminate the need for the break . * breaks are not seen in the stacked redshift - corrected bl lac spectrum or the spectrum of the `` empty '' sky . * the presence of the breaks associated with absorption by uv photons implies that at least some fraction of the @xmath1-rays are produced within the blr . the presence of lyc absorption in 3c 454.3 is not surprising . this object is exceptional in all its components : the @xmath1-ray emission from the jet reaches luminosities in excess of @xmath72 erg s@xmath29 @xcite , the accretion disk emits @xmath73 erg s@xmath29 @xcite , the blr @xmath74 erg s@xmath29 @xcite , and luminosity in ly@xmath3 only is @xmath75 erg s@xmath29 @xcite . here we can expect that the ionization degree is high and the photon - photon optical depth is substantial . the second object that shows lyc absorption is 4c + 21.35 . that case is important because this fsrq has been detected during a flare in the 70400 gev range by magic @xcite and the coexistence of the absorption break and vhe emission is difficult to understand in a single emission zone scenario , because the multi - hundred gev photons would have trouble escaping from the blr . the @xmath1-ray luminosity of 4c + 21.35 is smaller than in 3c 454.3 , @xmath76 erg s@xmath29 during flares @xcite , but the accretion disk is almost as luminous with @xmath77 erg s@xmath29 @xcite . on the other hand , the blr luminosity is significantly smaller @xmath78 erg s@xmath29 @xcite with ly@xmath3 producing @xmath79 erg s@xmath29 , i.e. ten times less than in 3c 454.3 . a much lower blr luminosity implies that the opacity for 100 gev photons is smaller , because blr size scales roughly as @xmath80 cm @xcite and the opacity as @xmath81 ( see equation ( [ eq : taut ] ) and * ? ? ? thus in 4c + 21.35 , the optical depth through the blr is expected to be 57 times smaller than in 3c 454.3 . this might be the reason why vhe emission is detected in 4c + 21.35 , but not in a much stronger source 3c 454.3 . for a given object having both the measurement of the line luminosity and the estimation of the blr size , we can obtain an expected value for the opacity @xmath47 using equation ( [ eq : taut ] ) , i.e. assuming that the blr emission is isotropic and @xmath1-rays have to penetrate through the whole blr . for 3c 454.3 with @xmath82 and @xmath83 , we get the opacity from h lyc / ly@xmath3 of @xmath84 , while for 4c + 21.35 we have @xmath85 . we see , however , for both objects the observed value is @xmath6425 ( table [ tab : fits ] ) , much below the expectation . what are the possible solutions for this discrepancy ? we can propose at least two solutions : * the size of the h lyc / ly@xmath3 emission region is larger than that measured from reverberation mapping using civ @xcite and h@xmath86 lines @xcite . we note that no ly@xmath3 variability was detected in any of the quasars analyzed by @xcite supporting this picture . this would immediately reduce the expected @xmath1-ray opacity , which scales inversely proportionally with the size . * alternatively , if the blr is flat @xcite , i.e. elongated alone the accretion disk , the @xmath1-ray opacity is much reduced ( e.g. * ? ? ? * ) . in this case , the threshold energy , which depends on the maximum interaction angle between the blr photons and the @xmath1-rays , should be a factor of two larger ( i.e. @xmath040 gev instead of @xmath020 gev ) contradicting the data . however , if in addition a few blr clouds are situated along the jet axis outside the @xmath1-ray emitting region , they would produce enough photons to collide head - on with the @xmath1-rays to make a break at 20 gev . * finally , the @xmath1-ray emitting region can be extended and its location can change depending on the luminosity , with less gev absorption during the strong flares @xcite . the vhe emission does not need to be produced in exactly the same place as the gev emission . this can explain the fact that we see both gev breaks as well as @xmath87100 gev emission in 4c + 21.35 ( but not necessarily at the same time ) . how then these models compares with the results on absorption ? photoionization models predict that luminosity in lyc / ly@xmath3 is typically @xmath010% of the hydrogen one @xcite . thus , taking the same blr size and geometry and using four times larger photon energy gives @xmath88 . in that case , the absorption would be negligible . on the other hand , reverberation mapping shows that lines are produced closer to the black hole that e.g. h@xmath86 @xcite , implying a higher photon density _ inside _ the zone of complete he ionization and a larger @xmath89 . however , if the @xmath1-ray emitting region is slightly outside of that region the opacity is reduced . again in this situation the break energy should be higher , but the quality of the data do not allow us to reject a hypothesis that the break energy is actually two times larger than the fiducial ( rest - frame ) 5 gev value . thus we do not see an obvious contradiction between the fact that the @xmath1-rays are produced outside ( or at the edge ) of lyc emitting region and the presence of the break . in general , the situation with the h lyc absorption strongly dominating over lyc absorption seems more natural than the picture presented by @xcite and @xcite , where a few gev breaks look more prominent than those at @xmath020 gev . now the statement of @xcite that the jet emission takes place in the inner ( higher ionization ) regions of blrs should be modified : the gamma - ray emission site lies within the normal blr region . this fact does not change the main astrophysical implications of photon - photon absorption of the jet gamma - ray emission , this particularly implies that the jet is already accelerated within a parsec distance from the black hole and therefore the blandford - znajek mechanism @xcite is responsible for the jet launching and the blr dense photon field can provide conditions for energy dissipation via photon breeding @xcite .
we reanalyze _ fermi_/lat gamma - ray spectra of bright blazars with a higher photon statistics than in previous works and with new pass 7 data representation . in the spectra of the brightest blazar 3c 454.3 and possibly of 4c + 21.35 we detect breaks at @xmath05 gev ( in the rest frame ) associated with the photon - photon pair production absorption by lyman continuum ( lyc ) . we also detect confident breaks at @xmath020 gev associated with hydrogen lyc both in the individual spectra and in the stacked redshift - corrected spectrum of several bright blazars . the detected breaks in the stacked spectra univocally prove that they are associated with atomic ultraviolet emission features of the quasar broad - line region ( blr ) . the dominance of the absorption by hydrogen ly complex over , rather small detected optical depth , and the break energy consistent with the head - on collisions with lyc photons imply that the gamma - ray emission site is located within the blr , but most of the blr emission comes from a flat disk - like structure producing little opacity . alternatively , the lyc emission region size might be larger than the blr size measured from reverberation mapping , and/or the @xmath1-ray emitting region is extended . these solutions would resolve a long - standing issue how the multi - hundred gev photons can escape from the emission zone without being absorbed by softer photons .
single - pion photoproduction has been used extensively to explore the electromagnetic properties of nucleon resonances , and most determinations of the @xmath5 resonance couplings @xcite have been obtained through multipole analyses of this reaction . the study of @xmath6 properties , in general , has enjoyed a resurgence , driven by the growth of new facilities worldwide . many precise new measurements have focused on the photoproduction of pions and other pseudoscalar mesons , in the hope that they would reveal states not seen previously in the photoproduction and elastic scattering of pions . associated studies have suggested both new states and couplings to existing states which contradict those found in analyses of the full pion production database . these results demonstrate the influence of measurements sensitive to new quantities and suggest that older analyses may have been based on insufficient or flawed data sets . a survey of existing data in the 1 gev region shows most polarization measurements to have only one or two angular points at a given beam energy , often with rather large uncertainties . this allows a great deal of freedom in multipole analyses , and is clearly insufficient to pick out more than the strongest resonance signals . one quantity which has consistently influenced these new analyses is the beam polarization quantity @xmath0 . examples include the e2/m1 ratio @xcite , where new measurements of the differential cross section and @xmath0 changed this ratio by nearly a factor of two . measurements of @xmath0 in @xmath7 photoproduction were shown @xcite to be sensitive to the @xmath8 ratio of photo - decay amplitudes for the @xmath9 , implying a ratio very different from the pion photoproduction value . the behavior of @xmath0 in kaon photoproduction has also proved @xcite to be crucial in determining the character of a bump seen in precise new total cross section measurements . in @xmath10 photoproduction , results for @xmath0 from graal have been used to argue for a change in the @xmath11 amplitude for the @xmath12 @xcite . in this experiment , we have performed systematic measurements of the cross section asymmetry @xmath0 for the reaction @xmath13 @xcite simultaneously with measurements of compton scattering @xmath14 @xcite , over the energy range @xmath15 = 0.5 - 1.1gev and at pion cm angles @xmath16 . the experimental data ( 158 data points ) from @xmath1-meson photoproduction constitute the first systematic and high statistics measurement of @xmath0 for this reaction channel over the present kinematic region . preliminary results of the measurements have been presented in @xcite . in sections 2 and 3 , we describe the methods employed in this experiment , with special emphasis on the control of systematic errors . in section 4 , we describe how these measurements , and the recent measurements from graal @xcite , have effected the gw multipole analyses . comparisons to the mainz fits @xcite are also made . finally , in section 5 , we summarize our results . the experiment was carried out on the linearly polarized photon beam of the 4.5gev yerevan synchrotron . the same setup has been used for the parallel study of two reactions : compton scattering and photoproduction of @xmath17-meson off protons . the experimental setup is shown in fig.[f1 ] . the linearly polarized photon beam is generated by coherent bremsstrahlung ( cb ) of 3.03.5gev electrons on the internal @xmath18 m diamond crystal target . the beam has 23msec long pulses with a 50hz repetition rate and an intensity of about @xmath19 eq . photons / sec . the beam , shaped by a system of collimators and sweeping magnets , has a 10@xmath2010 mm size at the position ( @xmath21 ) of the liquid hydrogen target ( @xmath21 ) ( 9 cm in length ) . it is transported in a vacuum pipe to the wilson - type quantameter ( q ) . the energy spectrum is measured and controlled by a 30-channel pair spectrometer ps-30 ( energy resolution 12% ) @xcite . the whole range of bremsstrahlung energies are covered by 5 scans with respective current settings on the psm analyzing magnet . the full energy spectra of bremsstrahlung on amorphous and crystal targets , measured by the pair spectrometer ps-30 , are presented in fig . [ f2 ] . the recoil protons are detected by the magnetic spectrometer ( ms ) consisting of a doublet of quadrupole lenses ( @xmath22 , @xmath23 ) , a bending magnet ( bm ) , a telescope of four scintillation trigger counters ( @xmath24 ) and a system of coordinate detectors , including a scintillation hodoscope ( @xmath25 ) and seven multiwire proportional chambers ( @xmath26 ) , allowing the reconstruction of momentum , azimuthal and polar angles of the outgoing proton . time - of - flight measurements were carried out between @xmath27 and @xmath28 counters on a flight base of 9 m for particle identification ( p,@xmath29 ) . the ms covered a solid angle of @xmath30 , and its angular and momentum resolutions were @xmath31 , @xmath32 and @xmath33 , respectively . the incident photon energy and the cm scattering angle were reconstructed on average with an accuracy of @xmath34 and @xmath35 , with corresponding acceptances of @xmath36 and @xmath37 , respectively . the @xmath1 decay photons were detected by the @xmath38erenkov spectrometer ( @xmath39 ) consisting of a veto counter ( v ) , a lead converter ( pb ) , scintillation hodoscope ( @xmath40 ) for @xmath41 and @xmath42 coordinate analysis , trigger counter ( t ) , and a lead glass @xmath43erenkov counter of full absorption ( @xmath44 ) . the energy resolution of the @xmath43erenkov counter could be parameterized as @xmath45 without converter and @xmath46 with converter ( e being the photon energy in gev ) . the kinematics of the analyzed process was completely determined by defining the kinematic parameters of the proton in the magnetic spectrometer . for identification of the process @xmath47 , the protons in the magnetic spectrometer were detected in coincidence with the photon detection branch . the coordinate detectors of the @xmath43erenkov spectrometer were in addition used for correlation analysis between two detecting branches to identify the low rate yield of the compton scattering process @xcite . an independent experimental study was carried out to check the reconstruction accuracy in the magnetic spectrometer ( ms ) @xcite . monte carlo calculations , intended to estimate the influence of different possible factors on the reconstruction precision , indicate that inaccuracies in setting the positions , angles and fields of magnetic elements , poor quality of track reconstruction in @xmath48 , additional multiple scattering in materials not taken into account , influence equally the reconstruction accuracy of both the interaction vertex and the scattering angles in ms . therefore , an experimental study of the vertex position reconstruction accuracy was defined as an effective test of the characteristics of the ms , and was easily carried out by means of a point - like target . good agreement between the experimental results for vertex reconstruction and monte carlo calculations ( fig.[f3 ] ) provides reliable support for the performance of our proton detection branch within designed precisions . finally , the possibility to treat compton scattering data @xcite , despite significant difficulties caused by the physical background process of @xmath1 photoproduction with approximately two orders of magnitude higher cross section , and with practically the same kinematic parameters , also points to a reliable operation of the detection system within designed accuracies . during the measurements , several factors were considered to minimize the influence of systematic uncertainties . a particular emphasis has been placed on maintaining the stability of the coherent peak position . in the case of its deviation , the event registration was paused , until respective correction procedure was made . such checks of the coherent peak were carried out every 4050sec . in fig.[f4 ] , the measured peak region is presented , superimposed on the full coherent bremsstrahlung spectrum , for the cases with correct and distorted peak position . a possible influence on the asymmetry results from instabilities in subsystems was compensated by periodical alternating the orientation of the photon beam polarization during the measurement of the same kinematic point . also , during data acquisition the trigger rates , efficiencies of the coordinate detectors , distributions of kinematic variables were controlled . hardware and software of all these control procedures was realized through the camac standard . fiber optic lines were used to manage the current of the pair spectrometer magnet ( psm ) and the electronics of the ms trigger , placed immediately on the ms . during the off - line analysis , prior to the reconstruction step , the time - of - flight measurements between @xmath27 and @xmath28 were analyzed , and the events with a proton candidate in ms were selected ( fig.[f5 ] ) . a soft cut was made , mainly at low energies , on the energy response of the @xmath38erenkov detector , to reduce the contribution of accidental coincidences and multiparticle events resulting from the energetic part of the coherent bremsstrahlung spectrum @xcite . then , the selected events were sent for reconstruction of the kinematic parameters of proton and the kinematics of the event . within further selection , the interaction point in the target was reconstructed and events falling out of the target limits by more than one standard deviation are cut . the yield of the @xmath49 reaction was defined from the number of selected events in the time spectrum of start - stop measurements between two detection branches ( fig.[f6 ] ) , by fitting and subtracting the respective number of background events in the peak region . several corrections were applied to compensate for the contribution of various factors . the results from the analysis of compton scattering process were used to correct for the influence of that source . further corrections were applied , compensating for the rate reductions due to the dead time of data acquisition and efficiency of different subdetectors . the maximum trigger rates was about 5event / sec , and the dead time did not exceed 10% . another insignificant contribution , resulting from the empty target cell , was measured separately and amounted to @xmath501% . the background @xmath51 rate is mainly dominated with double pion production processes generated by the high energy tail of cb spectrum . this background contribution has been determined in additional measurements , such as : * @xmath51rate dependence on the coherent peak energy of the cb spectrum at fixed kinematics of the setup , allowing the significant variation in the relative yield of single and multiple pion production processes . * the @xmath51rate dependence on the @xmath52photon decay angle . the data obtained were then compared with results of monte carlo calculations . the contribution of background processes was estimated to be less than 5% . calculations of the polarization of the cb spectra were realized by a solution of integral equations for intensity and polarization of the coherent spectra @xcite . in the energy range of the measurements , the photon beam , polarization ranged from 50% to 70% . the cross section asymmetry @xmath0 of the reaction @xmath53 is defined as : @xmath54 where @xmath55 is the polarization of photon beam and @xmath56 are the normalized yields for the photons polarized perpendicular and parallel to the reaction plane . the systematic errors of the measurements were dominated mainly by the accuracy of the linear polarization calculations . the uncertainties in @xmath57 have been evaluated to be within @xmath58 in full energy range and are included in the errors of the experimental data . the possible shift in the reconstructed photon energies was calculated from the uncertainty in the ms optics and did not exceed 0.4% . finally , overall systematic uncertainty for each energy set did not exceed 3% . it is also important to note , that most kinematic bins were formed from at least two independent data samples collected at different kinematic settings of the detector . the final asymmetry was calculated from the corresponding results in such overlapping bins . on the other hand , comparison of the results on such equivalent bins , produced from different event samples , with different acceptance conditions and polarized incident photons from different parts of the cb spectrum , provided a good check for systematic uncertainties in the obtained results . our data for the asymmetry @xmath0 , in the kinematic range @xmath15 = 0.5 @xmath59 1.1gev and @xmath60 , are presented in table1 . some of the measured energy and angular distributions are shown in figs.[f7 ] @xmath59 [ f9 ] together with the experimental data of other groups @xcite and predictions of different phenomenological analyses @xcite . the most systematic measurements of @xmath0 for @xmath49 , over the resonance region , exist for the angle @xmath61 ( fig.[f7]a ) . from fig.[f7]a , one can see , as a whole , a good agreement between the values of prior measurements and our present results . near 900 mev , our data disagree with data from mit @xcite and some early slac @xcite measurements , but are in good agreement with later measurements from another slac group @xcite . our points are also consistent with the results of this group for the angle ( fig . [ f8]b ) . from fig.[f9 ] , it is seen that our results for the angular distributions at @xmath15 = 700 , 750 , and 800 mev are in agreement with the existing data as well . in figs.[f7 ] @xmath59 [ f9 ] , the experimental data are also compared with the predictions of phenomenological analyses from the mainz @xcite and gw @xcite groups . while both analyses reproduce the qualitative behavior of @xmath0 , a shape discrepancy is noticeable in the gw fit . the main difference between multipole analyses and experiment is observed above @xmath15 = 700 mev where there have been few previous measurements . inclusion of our data in the gw fit results in an improved description , but shape differences remain . the mainz fit appears to have a shape more consistent with the data . ( it should be noted that some of the older fits , in particular the fit of ref . @xcite , predict a shape consistent with the mainz result . ) it is interesting to compare these results to those recently published by the graal collaboration @xcite . in that work , a similar comparison was made for @xmath0 measurements in the reaction @xmath62 . there a deviation from gw predictions was noted in the 800 @xmath59 1000 mev range , at backward angles . it was suggested that this discrepancy could be removed by a change in the n(1650 ) photo - decay amplitude ( @xmath11 ) . we have examined the graal angular distribution at 950 mev ( @xmath63 = 1630 mev ) and agree that a change ( reduction ) in the @xmath64 multipole can account for the shape difference . here too , for neutral pion production and energies near 900 mev , a reduction of this multipole in fa00 by about 20% ( in modulus ) results in an improved description . a reduction of the same amount is found in comparing the energy - dependent and single - energy solutions . the effect is displayed in fig.[f10 ] . note that this modification to the gw fit , while improving the agreement with data , actually worsens the agreement with the mainz prediction at the most backward angles . the obtained data on cross section asymmetry ( @xmath0 ) are in good agreement with existing experimental data and in at least qualitative agreement with phenomenological predictions @xcite . we have found evidence in support of the graal claim that some discrepancies with previous gw analyses could be linked to the @xmath64 multipole and possibly the n(1650 ) resonance contribution . it will be useful to have a mainz fit including these data in order to isolate differences with the gw fits . a set of fits employing identical data sets would also be useful in determining whether differences are due to the chosen formalism or data constraints . work along this line is in progress @xcite . in summary , results of present experiment on @xmath0 asymmetry have significantly improved the existing data base in the second and third resonance regions . forthcoming measurements of cross section and polarization observables in the photoproduction of mesons ( jlab @xcite , graal @xcite ) , may lead to more successful determinations of the underlying scattering amplitudes and a more precise determination of resonance photocouplings . the yerphi group is indebted to the synchrotron staff and cryogenic service for reliable operation during the experiment . this work was supported in part by the armenian ministry of science ( grant-933 ) and the u. s. department of energy grant de the gw group gratefully acknowledges a contract from jefferson lab under which this work was done . jefferson lab is operated by the southeastern universities research association under the u. s. department of energy contract de ac0584er40150 . v. kouznetsov _ , to be published in proceedings of ix moscow international seminar on electromagnetic interactions off nuclei at low and medium energies , institute for nuclear research , moscow , russia , 2123 sept . 2000 . @xmath65(mev ) & @xmath0 & @xmath66 & @xmath65(mev ) & @xmath0 & @xmath66 & @xmath65(mev ) & @xmath0 & @xmath66 + 500 @xmath67 12.5 & 0.564 & 0.081 & 500 @xmath67 12.5 & 0.643 & 0.060 & 500 @xmath67 12.5 & 0.643 & 0.060 + 525 & 0.683 & 0.052 & 525 & 0.636 & 0.039 & 525 & 0.617 & 0.039 + 550 & 0.572 & 0.039 & 550 & 0.600 & 0.031 & 550 & 0.636 & 0.031 + 575 & 0.638 & 0.034 & 575 & 0.640 & 0.028 & 575 & 0.627 & 0.029 + 600 & 0.686 & 0.032 & 600 & 0.681 & 0.027 & 600 & 0.693 & 0.030 + 625 & 0.720 & 0.031 & 625 & 0.718 & 0.027 & 625 & 0.715 & 0.033 + 650 & 0.764 & 0.031 & 650 & 0.776 & 0.028 & 650 & 0.702 & 0.041 + 675 & 0.828 & 0.031 & 675 & 0.802 & 0.030 & 675 & 0.756 & 0.063 + 700 & 0.864 & 0.032 & 700 & 0.845 & 0.035 & 700 & 0.878 & 0.032 + 725 & 0.876 & 0.033 & 725 & 0.843 & 0.025 & 725 & 0.864 & 0.027 + 750 & 0.854 & 0.032 & 750 & 0.842 & 0.025 & 750 & 0.823 & 0.025 + 775 & 0.812 & 0.031 & 775 & 0.746 & 0.022 & 775 & 0.711 & 0.023 + 800 & 0.747 & 0.030 & 800 & 0.679 & 0.021 & 800 & 0.640 & 0.023 + 825 & 0.675 & 0.031 & 825 & 0.671 & 0.021 & 825 & 0.601 & 0.025 + 850 & 0.577 & 0.024 & 850 & 0.541 & 0.022 & 850 & 0.497 & 0.026 + 875 & 0.515 & 0.026 & 875 & 0.475 & 0.024 & 875 & 0.394 & 0.025 + 900 & 0.397 & 0.028 & 900 & 0.353 & 0.025 & 900 & 0.304 & 0.026 + 925 & 0.356 & 0.030 & 925 & 0.223 & 0.023 & 925 & 0.144 & 0.028 + 950 & 0.196 & 0.031 & 950 & 0.076 & 0.024 & 950 & -0.045 & 0.031 + 975 & 0.168 & 0.032 & 975 & -0.002 & 0.026 & 975 & -0.108 & 0.047 + 1000 & 0.105 & 0.034 & 1000 & -0.054 & 0.028 & 1000 & -0.144 & 0.084 + 1025 & 0.097 & 0.032 & 1025 & -0.108 & 0.039 & & & + 1050 & 0.103 & 0.042 & 1050 & -0.159 & 0.064 & & & + 1075 & -0.032 & 0.061 & & & & & & + @xmath65(mev ) & @xmath0 & @xmath66 & @xmath65(mev ) & @xmath0 & @xmath66 & @xmath65(mev ) & @xmath0 & @xmath66 + 700 @xmath67 12.5 & 0.867 & 0.029 & 700 @xmath67 12.5 & 0.851 & 0.029 & 700 @xmath67 12.5 & 0.803 & 0.027 + 725 & 0.824 & 0.029 & 725 & 0.794 & 0.027 & 725 & 0.778 & 0.026 + 750 & 0.772 & 0.027 & 750 & 0.720 & 0.024 & 750 & 0.708 & 0.024 + 775 & 0.724 & 0.026 & 775 & 0.692 & 0.024 & 775 & 0.649 & 0.024 + 800 & 0.648 & 0.025 & 800 & 0.645 & 0.023 & 800 & 0.587 & 0.028 + 825 & 0.627 & 0.026 & 825 & 0.585 & 0.025 & 825 & 0.478 & 0.038 + 850 & 0.595 & 0.029 & 850 & 0.524 & 0.032 & 850 & 0.360 & 0.085 + 875 & 0.482 & 0.038 & 875 & 0.382 & 0.052 & & & + 900 & 0.211 & 0.058 & 900 & 0.204 & 0.089 & & & + @xmath65(mev ) & @xmath0 & @xmath66 & @xmath65(mev ) & @xmath0 & @xmath66 & @xmath65(mev ) & @xmath0 & @xmath66 + 500 @xmath67 12.5 & 0.637 & 0.071 & 500 @xmath67 12.5 & 0.717 & 0.070 & & & + 525 & 0.728 & 0.049 & 525 & 0.677 & 0.043 & 525 @xmath67 12.5 & 0.739 & 0.051 + 550 & 0.750 & 0.040 & 550 & 0.724 & 0.033 & 550 & 0.744 & 0.034 + 575 & 0.792 & 0.036 & 575 & 0.686 & 0.028 & 575 & 0.718 & 0.028 + 600 & 0.742 & 0.033 & 600 & 0.718 & 0.029 & 600 & 0.700 & 0.027 + 625 & 0.744 & 0.032 & 625 & 0.746 & 0.029 & 625 & 0.649 & 0.032 + 650 & 0.770 & 0.032 & 650 & 0.792 & 0.027 & 650 & 0.744 & 0.052 + 675 & 0.862 & 0.034 & 675 & 0.831 & 0.027 & 675 & 0.731 & 0.034 + 700 & 0.809 & 0.025 & 700 & 0.801 & 0.028 & 700 & 0.727 & 0.027 + 725 & 0.795 & 0.026 & 725 & 0.736 & 0.028 & 725 & 0.690 & 0.026 + 750 & 0.723 & 0.028 & 750 & 0.683 & 0.025 & 750 & 0.641 & 0.024 + 775 & 0.664 & 0.027 & 775 & 0.588 & 0.023 & 775 & 0.557 & 0.024 + 800 & 0.594 & 0.026 & 800 & 0.541 & 0.024 & 800 & 0.537 & 0.030 + 825 & 0.498 & 0.027 & 825 & 0.464 & 0.020 & 825 & 0.384 & 0.024 + 850 & 0.439 & 0.022 & 850 & 0.349 & 0.025 & 850 & 0.315 & 0.022 + 875 & 0.329 & 0.024 & 875 & 0.261 & 0.022 & 875 & 0.257 & 0.019 + 900 & 0.171 & 0.024 & 900 & 0.166 & 0.019 & 900 & 0.135 & 0.018 + 925 & 0.017 & 0.022 & 925 & 0.048 & 0.018 & 925 & 0.047 & 0.020 + 950 & -0.176 & 0.022 & 950 & -0.110 & 0.018 & 950 & -0.133 & 0.021 + 975 & -0.278 & 0.023 & 975 & -0.272 & 0.019 & 975 & -0.253 & 0.032 + 1000 & -0.397 & 0.023 & 1000 & -0.384 & 0.025 & & & + 1025 & -0.494 & 0.029 & 1025 & -0.444 & 0.038 & & & + 1050 & -0.651 & 0.042 & 1050 & -0.541 & 0.071 & & & + experimental setup . d is the diamond target ; @xmath68 are collimators ; @xmath69 are sweeping magnets ; @xmath70 is the liquid hydrogen target ; @xmath71 is the wilson quantameter ; @xmath72 is the fast monitor ; @xmath73 are quadrupole lenses ; @xmath74 and @xmath75 are bending magnets ; @xmath76 is the thin converters ; @xmath77 is the lead converter ; @xmath78 are scintillation counters ; @xmath79 are hodoscopes ; @xmath80 is the @xmath80erenkov counter ; @xmath81 is the pair spectrometer ; @xmath82 is the magnetic spectrometer ; @xmath83 is the @xmath80erenkov spectrometer . [ f1 ] energy spectra of bremsstrahlung on amorphous and crystal targets for different coherent peak ( @xmath84 ) and electron ( @xmath85 ) energies : spectrum with amorphous target ; coherent spectrum with @xmath86 and @xmath87 ; coherent spectrum with @xmath88 and @xmath89 . [ f2 ] results of the test experiment with the point - like target ( solid histogram ) , compared with the monte carlo predictions ( dashed histogram ) , for different proton momenta ( @xmath90 ) and target locations ( @xmath91 ) . [ f3 ] the peak region of the coherent bremsstrahlung ( shown with arrows ) , measured periodically during the experiment , and used to check the peak position . the peak position is in the correct place ( a ) ; it is shifted to the side of lower energies ( pair spectrometer channels ) ( b ) . [ f4 ] energy dependence of @xmath1 photoproduction asymmetry @xmath0 at @xmath92 ( a ) and @xmath93 ( b ) , respectively . experimental data are from yerevan , present experiment ( filled circles ) , slac @xcite ( open triangles ) , mit @xcite ( open circles ) , slac @xcite ( open squares ) , kharkov @xcite ( filled triangles ) , and yerevan , previous measurements @xcite ( filled squares ) . solid ( dash - dotted ) curves give wi00 ( fa00 ) results by gw @xcite versus maid2000 results by the mainz group @xcite ( dashed curves ) . [ f7 ] angular dependence of @xmath1 ( a ) and @xmath99 ( b ) photoproduction asymmetry @xmath0 at @xmath100 . @xmath1 data are present measurements . @xmath99 data are from graal @xcite ( filled asterisk ) and dnpl @xcite ( open diamond ) . plotted are fits fa00 ( dash - dotted ) , fa00 with a modified @xmath101 multipole ( see text ) ( solid ) , and the maid2000 result@xcite ( dashed ) .
the cross section asymmetry @xmath0 has been measured for the photoproduction of @xmath1-mesons off protons , using polarized photons in the energy range @xmath2gev . the cm angular coverage is @xmath3 with energy and angle steps of 25mev and @xmath4 , respectively . the obtained @xmath0 data , which cover the second and third resonance regions , are compared with existing experimental data and recent phenomenological analyses . the influence of these measurements on such analyses is also considered . + pacs numbers : 13.60.le , 25.20.lj +
the shapley supercluster ( ssc ) has for long been known as a large region of a high overdensity of galaxies ( shapley 1930 ) . located in in the direction of hydra - centaurus at @xmath4 it is one of the densest large scale concentrations of matter ( fabian 1991 ) in the universe . if clusters of galaxies are taken as mass tracers ( scaramella et al . 1989 ) , the shapley concentration accounts for at least 10 - 20% of the acceleration of the local group towards the great attractor ( lynden - bell et al . 1988 , scaramella et al . 1991 , drinkwater et al . 1998 ) . optical ( vettolani et al . 1990 , raychaudhury et al . 1991 , bardelli et al . 1994 , quintana et al . 1997 ) and x - ray observations ( raychaudhury et al . 1991 , day et al . 1991 , breen et al . 1994 , bardelli et al . 1996 , ettori et al . 1997 ) reveal that the ssc has a remarkable structure dominated by a central high - density core . centered on a3558 ( shapley 8) this central region of the ssc includes five abell clusters ( a3558 , a3556 , a3559 , a3560 , a3562 ) and two rich groups of galaxies ( sc1327 - 312 and sc1329 - 313 ) . this corresponds to an overdensity of @xmath5 with respect to the average density of aco clusters ( vettolani et al . 1990 ) . a comprehensive study of the mass distribution in the ssc inferred from x - ray observations has been reported by ettori et al . their analysis was based on a mosaic of rosat pscp and einstein observatory ipc data enclosed in a sky area of @xmath6 and centered on a3558 . the individual mass estimates for 14 clusters and two groups of galaxies amount to an estimated total mass of the ssc of @xmath7m@xmath8 . assuming a cdm - like power spectrum , ettori et al . ( 1997 ) find that the core of shapley is an overdensity on the level of @xmath9 ( @xmath10 ) to @xmath11 ( @xmath12 ) . in this letter we report on the x - ray observation of the core of the ssc by rosat combining pointed observations and data from the rosat all - sky survey . the region of our study is centered at @xmath13 ; @xmath14 and covers a field of @xmath0 . it contains the central clusters a3562 , a3558 and a3556 and the two rich groups of galaxies sc1327 - 312 and sc1329 - 313 . this part of the core of the ssc is of particular interest since there are strong indications ( bardelli et al . 1994 ) that it forms a single elongated structure on the scale of @xmath15 mpc at a redshift of @xmath16 . our analysis shows that the core of ssc is traced by an elongated x - ray emission on similar scale . this confirms the existence of a physically connected , filamentary structure in the core of shapley . while rosat pspc pointed data is limited by the sky coverage , rosat all - sky survey ( rass ) ( trmper 1993 , voges et al . 1996 ) data offers a virtually unlimited field of view . on the other hand , rass data is limited by the short exposure time ( i.e. @xmath17 sec vs. typical @xmath18 sec of pspc pointings ) . as only a part of the core of the ssc is covered by pointed observations , a completion of this area by rass data offers a possibility to combine both advantages . fig . [ fig_region ] shows the field of our study . the clusters a3562 , a3558 and a3556 are shown as solid circles the radius of which correspond to 1 abell radius ( @xmath19 mpc ) . the rosat pointed pspc observations covering the region are indicated by line - shaded circles of @xmath20 diameter each . table [ tab_pointings ] lists the center and total observation time for each observation . the rass and pointed pspc data has been processed following the standard procedure for extended x - ray sources . in order to obtain the best signal - to - noise - ratio , only the 0.5 - 2 kev energy band ( the hard rosat band ) is used ( snowden et al . 1994 ) . this reduces both the foreground x - ray emission of the galactic interstellar medium and the detector background . the absorption of the x - ray emission due to neutral hydrogen in the line of sight varies only little . the neutral hydrogen column density ranges form @xmath21 @xmath22 towards a3558 to @xmath23 @xmath22 ( a3562 ) and @xmath24 @xmath22 towards a3556 ( stark et al . we adopted in the following the mean value @xmath25 @xmath22 . we began our analysis by merging the data of the pointed rosat pspc observations . we then binned both the rass and the pointed data into bins of size 36@xmath2636 arcsec@xmath27 each and calculated the corresponding exposure maps using exsas . [ fig_contour ] presents the contour plot of the merged count rates in the rosat hard band ( rosat band b , channels 52 - 201 , @xmath28 0.5 - 2 kev ) obtained from the rass and the pointed data . the contour plot is based on data smoothed with a gaussian of fwmh 12 arcmin . this choice compensates for the different fwmh of the point spread functions ( psf ) of pointed and rass observations as well as for the varying psf within the aperture of single pointings . the contour plot is overlaid on the binned flux . due to the shorter exposure time , the region outside the pointed observations ( i.e. the rass data ) exhibits a much lower signal - to - noise ratio resulting in a patchy appearance . the figure shows an overall picture of the region . neither the background nor point sources have been subtracted . local maxima of the count rate clearly trace the clusters a3562 , a3558 and a3556 as well as the two groups sc1327 - 312 and sc1329 - 313 . note also the region in the vicinity of a3556 . the peak of the count rate centered on a3556 is surrounded by faint x - ray emission with an extension of about @xmath29 in the east - west direction and @xmath30 in the north - south direction . we have checked for x - ray sources in this region . clearly visible is the point source 1rxs j132129.4 - 314054 of the rosat all - sky survey bright source catalogue . its position roughly coincidents with the western edge of the emission around a3556 . there are about five other point - like sources in this field , presumably associated with emission of single galaxies ( c.f . section [ sec_discuss ] ) . in analogy with a more comprehensive study of the x - ray emission of a3558 ( bardelli et al . 1996 ) , we find the point - like sources to contribute @xmath31 to the observed count rate . to check the significance of the elongated emission , we rebinned the central region into 31 bins as shown in fig . [ fig_bins ] . the count rates in the various bins as obtained from the unfiltered data is plotted in fig . [ fig_flux ] . the error bars are the @xmath32 ( poisson ) errors . the background shown as dotted line ( @xmath33 cts sec@xmath3 arcmin@xmath34 ) is estimated from the average of the six bins with lowest flux . its value is in good agreement with the background estimated from the smaller northern and south - western fields ( @xmath35 and @xmath36 cts sec@xmath3 arcmin@xmath34 , respectively ) . since the emission found for the background bins shows some structure , we quote here conservative gaussian errors . while the region used for the first background estimation is covered by pointed observations , the two other fields lie mostly within the area of the rass . because of the poor photon statistics , the data from these two fields has not been included in the subsequent analysis . the peaks of the flux seen in fig . [ fig_bins ] are clearly associated with the clusters and rich groups in the field . from fig . [ fig_flux ] it is also seen that the emission between the clusters a3556 and a3558 lies well above the background . thus the x - ray emission traces the the core of the ssc as a filamentary - like structure with an extension of @xmath37 corresponding to @xmath38 @xmath39 mpc . as a final test we checked for the effect of different bin sizes and applied the same analysis to the rosat a band ( channels 11 - 41 , @xmath28 0.1 - 0.4 kev ) , c band ( channels 52 - 90 @xmath28 0.5 - 0.9 kev ) and d band ( channels 91 - 201 @xmath28 0.9 - 2 kev ) . while no corresponding emission could be found in the a band , the emission observed in the c and d band traces the same structure . therefore we can exclude any contamination by galactic foreground emission miming the elongated emission . table [ tab_lum ] shows background subtracted fluxes and luminosities for various combinations of bins . for the conversion of the count rates in the rosat 0.5 - 2.0 kev band to the flux and luminosity in the 0.1 - 2.4 kev band we used a raymond - smith code with metallicity 0.35 ( solar units ) . where available , the plasma temperature for the different regions has been taken from the literature . the plasma temperature of sc1327 - 312 is estimated to be equal to sc1329 - 313 . for the faint emission between the clusters a3558 and a3556 as well as for the elongated emission in the vicinity of a3556 we assumed a plasma temperature of 1 kev , corresponding roughly to the temperature expected for ( poor ) groups of galaxies or for faint filamentary emission ( c.f . also section [ sec_discuss ] ) . the ( background subtracted ) flux and luminosity obtained for a3558 ( i.e. bins 13 - 17 ) are in good agreement with the corresponding values @xmath40 erg @xmath22 s@xmath3 and @xmath41 @xmath42 erg s@xmath3 ( rescaled to the energy range 0.1 - 2.4 kev ) published by bardelli et al . note however , that the values given in table [ tab_lum ] are of limited relevance since the binning used only crudely reflects the actual spatial structure of the different sources . in analogy , the total flux and luminosity found for the whole structure by summing up the respective values for the bins 3 - 27 has to be considered as a lower limit . the combination of flux and temperature data with a spatial model for the emitting region allows for an estimation of the electron density and gas mass in the respective region . in order to get upper limits , we assume here the emitting gas to be uniformly distributed in a cylinder whose projection corresponds to the binned region ( i.e. the axis of the cylinder is perpendicular to the line of sight ) . the volume of a bin then is @xmath43 @xmath44 mpc@xmath45 . the resulting electron density and gas mass estimations are quoted in table [ tab_lum ] . note that the approximation we applied overestimates the actual values since it assumes uniform distributions of the flux and emitting gas . thus the quoted values have to be considered as upper limits . the value @xmath46 we find for a3558 ( i.e. bins 13 - 17 ) is in rough agreement with @xmath47 within @xmath48 mpc calculated by bardelli et al . extrapolating to the whole structure , we find as an estimation of the total gas mass @xmath49 . an more precise determination of the gas mass as well as the estimation of the total mass should involve a more accurate model of both the gas and the flux distribution . based on rosat data , bardelli et al . ( 1998 ) performed a comprehensive analysis of spectral properties of the emission of the central cluster a3558 ( for a comparison with asca data see markevitch & vikhlinin 1997 ) . because of comparable exposure times of a3558 and the rest of the region covered by pointed pspc observations , a corresponding analysis of the whole elongated emission should be possible . however this is beyond the scope of this letter . as a first step towards a spectral analysis of the region we determine here the hardness ratio ( i.e. x - ray color ) related to the rosat c and d energy band @xmath50 where @xmath51 and @xmath52 denote the flux in the rosat bands c and d , respectively . in fig . [ fig_hardness1 ] the hardness ratio of the background corrected flux is shown as a function of the bin number . the errors are the @xmath32 errors . as it is the case for the flux , the hardness ratio traces the structure of the core of shapely . a clearly higher hardness ratio is found for the emission of clusters and groups than for intermediate , connecting emission . this is evidence for higher gas temperatures associated with density peaks , i.e. a result one would expect . for temperatures between 2 and 10 kev , and a neutral hydrogen column density of @xmath25 @xmath22 the theoretical values for the hardness ratio range form 0.33 to 0.355 . these values are in rough agreement with the hardness ratio found for the clusters and groups in the field . the corresponding hardness ratio of the background ( i.e. the hardness ratio of six bins with lowest flux ) is -0.22 . for a gas temperature of 1 kev , the theoretical value for the hardness ratio is @xmath53 . as is clear form fig . [ fig_hardness1 ] , the hardness ratio between a3556 and a3558 lies bellow this value , indicating a temperature bellow 1 kev . the oscillating behavior of the hardness ratio westwards of a3556 ( i.e. for bins @xmath54 ) is due to the background subtraction procedure . by subtracting the average flux of the six bins with lowest flux , the background corrected net flux in some of these bins may become negative giving rise to observed oscillatory feature . the analysis of the rosat rass and pointed data of the core of the ssc shows clear evidence for x - ray emission related to a filamentary superstructure . optical observations ( bardelli et al . 1994 , 1998 ) reveal a striking similarity of isodensity contour lines of galaxy counts and x - ray contours . this strongly suggests a tight connection between the galaxy distribution and the x - ray emission . it is intriguing that galaxy counts and x - ray emission not only correlate on the scale of the single clusters in the field but also on the scale of the whole filamentary structure , including the region of faint emission between a3558 and a3556 and westwards of a3556 . the angular and redshift distribution of galaxies in this field provides strong evidence for a filamentary superstructure connecting the clusters a3562 , a3558 and a3556 ( bardelli et al . 1994 ) . we thus conclude the elongated x - ray emission of the core of the ssc to trace this filamentary superstructure . in principle the elongated x - ray emission could arise from projection effects . however , there is little doubt that this is nt the case . first of all , redshift measurements indicate the structure to lie perpendicular to the line of sight which excludes projection effects . additional evidence comes from the flat , elongated x - ray emission of a3556 which is not comparable to the x - ray emission of compact , isolated clusters . it is interesting to note that there is additional independent evidence for an elongated gas distribution westwards of a3556 . the radio survey of a3556 ( venturi et al . 1997 ) reveals a wide - angle tailed ( wat ) radio galaxy at a distance of @xmath55 mpc westwards from the center of a3556 . normally , wat galaxies are found at the center of groups or clusters and it is assumed that the tails are bent by gas flows due to a merging processes ( gomez et al . 1997 ) . the original suggestion of venturi et al . ( 1997 ) that the bent morphology of the wat galaxy is due to interaction with surrounding relatively cold gas ( i.e. @xmath56 kev ) is confirmed by the present study . x - ray emission from filamentary large scale structures is predicted by models ( e.g. bond et al . 1996 ) and recent n - body / hydro simulations ( e.g. cen & ostriker 1996 ) of structure formation . it is expected to originate from the hot phase of the intergalactic medium ( igm ) of temperature @xmath57 kev , either found in association with filaments themselves or with single cd galaxies or groups of galaxies tracing the structure . there have been only few claims of detection of x - ray emission on scales extending the scale of clusters of galaxies ( i.e. wang et al . 1997 , soltan et al . 1997 , but see also briel & henry 1995 ) . the fact that the present study found clear evidence for x - ray emission on the scale of @xmath1 mpc is reconciled by the extraordinary dense nature of the core of shapley . while the emission centered on the clusters a3558 , a3562 and a3556 as well as the one centered on the poor groups sc1329 - 313 and sc1327 - 312 is due to hot intercluster medium , the nature of the fainter emission connecting the whole structure is less obvious . to some extent , it remains debatable if the x - ray emission connecting the clusters a3558 and a3556 originates from poor groups of galaxies , from igm distributed between the clusters or from overlapping gas distributions . in general , the x - ray luminosity of groups of galaxies in the 0.1 - 2.4 kev band is typically @xmath58 erg s@xmath3 ( henry et al . 1995 , mulchaey et al . 1996 ) . comparing this value to the x - ray luminosity found in the bins 18 - 20 leads to @xmath59 groups in this region , a number which seems rather high . in addition , at least compact groups should be resolved as individual sources . we thus exclude this possibility . likewise , the background subtracted x - ray surface brightness of the bins 18 - 20 is @xmath60 erg @xmath22 s@xmath3 arcmin@xmath34 ( in the 0.1 - 2.4 kev energy band ) corresponding to about 2.5 times the upper limit for emission from filamentary large scale structure ( briel and henry , 1995 ) . while the extraordinary dense nature of the region could explain this difference , the overall elongated structure may also suggest that the emission seen in the bins 18 - 20 is due to overlapping gas distributions of a3558 and a3556 . in this case , the low hardness ratio of the intermediate region could be due to a temperature decrease in the outskirts of the clusters . yet another possibility is that we are observing the beginning of the merging of a3558 and a3556 . a more definite answer to this question should involve the determination of the temperature distribution . if the low temperature ( i.e. @xmath57 kev ) assumed for the gas in the bins 18 - 20 is confirmed , this would be a strong indication for x - ray emission originating from igm distributed on intrasupercluster scale . the low hardness ratio obtained for the questionable bins as well as the fact that we are dealing with an overall exceptional dense region indicates that this actually may be the case . in summary , our analysis shows that the central part of the shapley supercluster is traced by elongated x - ray emission connecting the three clusters of galaxies a3562 , a3558 , a3556 and the two groups sc1327 , sc1329 . the filamentary - like x - ray emission extends over @xmath37 corresponding to @xmath38 @xmath39 mpc . this is strong evidence for the original claim of bardelli et al . ( 1994 ) that the three clusters a3562 , a3558 and a3556 form a physically connected , single structure . while the emission between a3558 and a3556 seems not to originate from groups of galaxies it remains debatable if it is due to overlapping gas distributions of a3558 and a3556 or if its of intrasupercluster origin . a more definite answer to this question as well as to questions related to the dynamical state of the structure could be provided by the temperature distribution of the region . its determination is thus of great interest . bardelli s. , zucca e. , vettolani g. , zamorani g. , scarmella r. , scaramella r. , collins c. a. , macgillivray t. , 1994 , mnras , 267 , 665 bardelli s. , zucca e. , malizia a. , zamorani g. , scarmella r. , vettolani g. 1996 , a&a , 305 , 435 bardelli , s. , zucca , e. , zamorani , g. , vettolani , g. , & scaramella , r. 1998 , mnras , 296 , 599 bond j. r. , kofman l. , pogosyan d. , 1996 , nature , 380 , 603 breen j. , raychaudhury s. , forman w. , jones c. , 1994 , apj , 424 , 59 briel u. g. , henry j. p. 1995 , a&a , 302 , l9 cen r. , ostriker j. p. , 1996 , apj , 464 , 27o dickey j. m. , lockman f. j. , 1990 , ara&a , 28 , 215 drinkwater , m.j . , parker , q. a. , proust , d. , quintana , h. , slezak , e. 1998 , astro - ph 9807118 day c. s. r. , fabian a. c. , edge a. c. , raychaudhury s. , 1991 , 252 , 394 fabian a. c. , 1991 , mnras , 253 , 29 gomez p. l. , pinkney j. , burns j. o. , wang q. , owen f. n. , & voges w. 1997 , apj , 474 , 580 henry j. p. , et al . 1995 , apj , 449 , 422 lynden - bell d. , faber s. m. , burstein d. , davies r. l. , dressler a. , terlevich r. j. , & wegner g. 1988 , apj , 326 , 19 ettori s. , fabian a. c. , white d. a. 1997 , mnras , 289 , 787 markevitch m. , vikhlinin a. 1997 , apj , 474 , 84 mulchaey j. s. , davis d. s. , mushotzky r. f. , burstein d. 1996 , apj , 456 , 80 postman m. , lauer t. r. , 1995 , apj 440 , 28 quintana h. , melnick j. , proust d. , infante l. , 1997 , a&as , 125 , 247 raychaudhury s. , fabian a. c. , edge a. c. , jones , c. , forman w. , 1991 , mnras , 248 , 101 shapley h. , 1930 , harvard obs . bull . 874 , 9 snowden s. c. , mcgammon d. , burrows d. n. , mendenhall j. a. , 1994 , apj , 424 , 714 scaramella r. , baiesi - pillastrini g. , chincarini g. , vettolani g. , zamorani g. , 1989 , natur , 338 , 562 soltan , a. m. , hasinger , g. , egger , r. , snowden , s. , & trmper , j. , 1997 , a&a , 320 , 705 trmper , j. 1993 , science , 260 , 1769 venturi , t. , bardelli , s. , morganti , r. , & hunstead , r. w. 1997 , mnras , 285 , 898 vettolani g. , chingarini g. , scaramella r. , zamorani g. , 1990 , aj 99 , 1709 voges , w. , et al . 1996 , mpe report 263 , 637 wang q. d. , connolly a. j. , brunner , r. j. 1997 , apj , 487 , l13 white d. a. , jones c. , forman w. 1997 , mnras , 292 , 419 @xmath62 } & \mbox{[cts s$^{-1}$ ] } & \mbox{[kev ] } & \mbox{[erg cm$^{-2}$ s$^{-1}$ ] } & \mbox{[$10^{44}$ erg s$^{-1}$ ] } & \mbox{[cm$^{-3}$ ] } & \mbox{[$10^{14 } m_\odot$]}\\ \hline \noalign{\smallskip } a3562 & [ 3 - 7 ] & 1.73 & 3.8^{\rm c } & 3.6\times10^{-11 } & 3.70 & 1.33 & 2.5\\ sc1329 & [ 8 - 9 ] & 0.43 & 3.0^{\rm c } & 9.0\times10^{-12 } & 0.91 & 1.48 & 0.8\\ sc1327 & [ 11 - 12 ] & 0.80 & 3.0^{\rm d } & 1.7\times10^{-11 } & 1.70 & 2.01 & 1.1\\ a3558 & [ 13 - 17 ] & 3.98 & 3.8^{\rm c } & 8.3\times10^{-11 } & 8.51 & 2.82 & 3.8\\ & [ 18 - 20 ] & 0.15 & 1.0^{\rm d } & 2.4\times10^{-12 } & 0.24 & 0.52 & 0.4\\ a3556 & [ 21 - 23 ] & 0.31 & 2.1^{\rm c } & 6.3\times10^{-12 } & 0.64 & 1.02 & 0.8\\ & [ 24 - 26 ] & 0.13 & 1.0^{\rm d } & 2.2\times10^{-12 } & 0.22 & 0.50 & 0.4\\ & [ 3 - 27 ] & 7.53 & - & 1.6\times10^{-10 } & 15.95 & - & 9.8\\ \hline \end{array}\ ] ]
we report on x - ray observations of the core of the shapley supercluster . combining data from pointed observations of the rosat pspc detector and data from the rosat all - sky survey , the observed region covers an area of @xmath0 . it contains the central clusters a3562 , a3558 and a3556 . we find clear evidence for x - ray emission connecting the three clusters . this confirms the existence of a filamentary , physical structure embedding the three clusters a3562 , a3558 and a3556 . we also find evidence for faint emission westwards of a3556 . in total , the extension of the filamentary x - ray emission of the core of the shapley supercluster amounts up to at least @xmath1 mpc . the total luminosity in the 0.1 - 2.4 kev energy band is @xmath2 erg s@xmath3 .
hd 23514 ( hii 1132 ) is an f6 star in the pleiades . it is one of only a handful of youthful stars , ages 35 to @xmath2100 myr , that are surrounded by warm dust particles located in the terrestrial planet zone and which absorb a percent or so of the bolometric luminosity of their central star @xcite . we have observed hd 23514 with an adaptive optics system at the keck observatory using the nirc2 camera to check the binarity of this unusually dusty star . to the best of our knowledge , to date no substellar companion to a stellar primary has been found in the pleiades . however , while this paper was being reviewed , @xcite announced the discovery of a substellar companion to the pleiad hii 1348 . studies of substellar objects in the pleiades focus on free floating brown dwarfs or companions to low mass stars @xcite . with an adaptive optics survey of 144 solar type stars in the pleiades , @xcite was sensitive to objects with masses down to @xmath20.03 m@xmath1 at separations of @xmath21 . however , their lowest mass companion ( @xmath20.15 m@xmath1 ) is still above the substellar boundary ( they did not observe hd 23514 ) . our work has revealed one such substellar companion in the pleiades . we observed hd 23514 with the adaptive optics ( ao ) system and the nirc2 narrow camera at the 10-m keck telescope at mauna kea observatory . observations were performed in natural guide star mode . some images were obtained in angular differential imaging mode ( adi ) while others had the primary behind a 600 mas or 800 mas diameter coronographic spot . the rest of the images were taken with dithered observations in position angle ( pa ) mode where the field of view is rotated so as to keep the sky fixed on the nirc2 camera . all data were reduced in the standard manner ( flat fielding , dark subtraction , sky subtraction ) with iraf routines . the data was linearized prior to processing using an idl program . we have applied the nirc2 distortion solution presented in @xcite to refine our astrometry . we adopt a pixel scale of @xmath3 mas / pix and a @xmath4 offset from north for the camera columns @xcite . we did not apply any advanced loci processing to analyze the data . our observations are summarized in table [ tab : data ] . relative photometry was performed with iraf s * apphot * tasks using aperture sizes that maximized the signal - to - noise of the companion . generally , this was 11.4 times the fwhm of a point source . the apparent magnitude for the companion at k@xmath5 is estimated by correcting the 2mass ks magnitude of @xmath6 for the primary as described by @xcite , namely k@xmath5k @xmath7(h k ) and using the relationship between k and ks described by @xcite . hd 23514 then has a k@xmath5 magnitude of @xmath8 , not too different from the ks magnitude . for the coronographic data , the primary is attenuated by a factor of 28.57 and 27.80 for the 600 and 800 mas diameter spot , respectively ( r. galicher , private communication , 2011 ) . multiple short exposures were obtained with integration times range from 10 - 60 seconds , as listed in table [ tab : data ] , and the companion is clearly visible in most of the individual frames . the primary is saturated in the 2006 , 2007 , and 2008 ks and k@xmath5 long exposures . the individual frames were registered to shorter , unsaturated frames using iraf s * xregister * task . photometry for these paired observations is performed with the same aperture size , but takes into account the different exposure times . we measure the candidate companion at a separation of @xmath9 with position angle @xmath10 in our 2010 coronographic observations . the object has remained at the same location between epoch 2006 and 2010 ( see table [ tab : meas ] and figure [ fig : image ] ) . had this been a more distant background object , then , based on the primary s proper and parallactic motion ( @xmath11 mas / yr , @xmath12 mas / yr ; * ? ? ? * ) , after 5 years this object should have been separated by @xmath13 and had a pa of @xmath14 ( see figure [ fig : location ] ) . as mentioned in section [ results ] , the 2006 adi data is somewhat unreliable as evidenced by the large uncertainties in table [ tab : meas ] . comparison from 2007 yields an expected position of a background source of @xmath15 with pa of @xmath16 . over this shorter 4-year baseline , a background object can be ruled out at the @xmath212-@xmath17 level . the consistent values for separation and pa for the object from 2006 to 2010 imply this is a companion to the system or is a nearby co - moving pleiades member , a possibility we address in section [ pleiad ] . the quoted photometry and astrometry in table [ tab : meas ] is the average of all the frames and the uncertainties are standard deviations of the individual measurements . the more precise astrometry present in the latter epochs is possible since both the primary and secondary are detected and are unsaturated in the same frame . a companion 6.8 magnitudes fainter than the primary f6 star at ks would be a late - m dwarf ( see section [ results ] ) . the j ks and h ks colors of this object are also consistent with being late - m ( @xmath2m79 ; see @xcite ) . the j h color is somewhat discrepant being bluer than expected ; however , as shown in figure [ fig : cmd ] , the measurement is consistent given the uncertainties . evolutionary models for a 100-myr old f6 star like hd 23514 ( 1.35 m@xmath1 and @xmath18 k ; * ? ? ? * ) predict an absolute ks magnitude of about 2.5 @xcite giving a distance modulus of 5.65 and a distance of 135 pc . an additional distance estimate can be obtained by considering that this star is part of the pleiades , which lies 133.5 pc away ( @xcite ; see also table 3 in @xcite ) . the pleiades extends nearly 10 in diameter ( * ? ? ? * and references therein ) ; if we assume a spherical distribution , this amounts to a possible radial extent of 11.7 pc . hence , the distance to hd 23514 would be somewhere between 122 pc and 145 pc ; we adopt a distance of @xmath19 pc . considering our ks photometry , the companion has @xmath20ks = @xmath21 , which when comparing with the 2mass photometry for hd 23514 gives an apparent magnitude of ks = @xmath22 . for our adopted distance , the absolute ks magnitude of the companion is @xmath23 . to estimate the mass of the companion we make use of low - mass / substellar evolutionary models . the dusty evolutionary models of @xcite take into account the formation of condensates in the atmosphere and are appropriate for low - temperature objects down to @xmath21300 k. with these models , a 100 myr object with absolute k magnitude 9.27 would have a mass somewhere 0.06 and 0.07 solar masses and effective temperature @xmath22700 k. the evolutionary models of @xcite use the nextgen dust - free atmospheres and are appropriate for objects with @xmath24 k. these models similarly predict @xmath25 k and mass @xmath20.050.06 solar masses when compared to the ks - band photometry . the j and h - band photometry suggest somewhat lower masses ( down to 0.05 solar masses ) and cooler temperatures ( @xmath22500 k ) . combining all results together , we can infer that the companion has a mass of @xmath26 solar masses . this is below the stellar / substellar boundary making the object a brown dwarf . the temperature , likewise , appears to be @xmath27 k , placing it at spectral type @xmath2m7 ( see figure 2 in @xcite ) , consistent with the near - infared colors @xcite . no other object is detected in the @xmath28 field of view . at the distance to hd 23514 , this amounts to a radius of 675 au . figure [ fig : limit ] presents the 5@xmath17 magnitude limit as a function of distance from the primary for the two coronographic ks observations ( 2009 and 2010 ) . these are our best limits for hd 23514 . limits are obtained by performing aperture photometry over blank regions of the sky at annuli surrounding the primary . these limits are averaged over the multiple frames . no companion is detected with a limit of @xmath21112 magnitudes beyond 1.5 ( @xmath2200 au ) . this suggests an absolute magnitude limit of @xmath214 at ks , or a mass of nearly @xmath211 jupiter masses when compared to evolutionary models @xcite . the average ks magnitude for the secondary , 14.92 , differs from the k@xmath5 magnitude , 15.40 , by half a magnitude . a late - m star is expected to have h [email protected] , which suggests k@[email protected] , not enough to account the difference . however , the uncertainty in our k@xmath5 photometry is the highest of our measurements , at nearly 0.4 magnitudes . for the 2006 data , the uncertainty is a result of the larger aperture size ( @xmath29 fwhm ) needed as the m - type companion is smeared out in the adi exposures . for 2007 and 2008 , the primary is saturated and the secondary s flux ( and location ) is compared to separate , shorter exposures as described in section [ observations ] . the high uncertainty , combined with the expected difference in the k@xmath5 and ks filters is enough to account for the discrepancy . while the common motion with hd 23514a is suggestive , there exists the possibility that the observed object may be a more distant ( or closer ) free - floating pleiades brown dwarf . as a member of the pleiades , it would have about the same proper motion and thus appear to be co - moving with hd 23514 . the most robust way to ascertain this would be to observe the system over many years in order to detect ( or rule out ) orbital motion . however , statistical arguments suggest the brown dwarf is indeed bound to hd 23514 . @xcite used a group of surveys to estimate the spatial distribution of brown dwarfs in the pleiades . hd 23514 is located about 1.7 degrees from the center ( @xmath30 , @xmath31 ; @xcite ) . at that distance , @xcite suggest a brown dwarf ( mass @xmath32 m@xmath33 ) density of 2.7 per square degree . however , more recent studies performed by @xcite and @xcite suggest average brown dwarf ( mass @xmath34 m@xmath33 ) densities of 7.8 and 6.3 per square degree , respectively . @xcite surveyed the pleiades in fields located 0.75 to 3.5 degrees away from the field center , @xcite incorporated that survey and in addition also covered the center of the pleiades with ukidss . thus , the area covered by these deeper surveys is similar in radial extent as that in @xcite ( @xcite ; out to 2.2 degrees from the center ) , but the density of brown dwarfs is larger by a factor of about 2 than the average brown dwarf density of 4.2 per square degree found by @xcite . to account for these deeper surveys ( mass @xmath34 m@xmath33 ) , we scale the density provided by @xcite and find that , at hd 23514 s distance from the center of the pleiades , the density of brown dwarfs is about @xmath25 per square degree . the nirc2 field of view is @xmath35 , or an area of @xmath36 square degrees . the likelihood of finding a free - floating pleiades brown dwarf with mass @xmath37m@xmath33 in the field of view is thus about @xmath38 . the small likelihood derived lends weight to the idea this is a bound companion . there are few brown dwarf companions known to stellar primaries ( the well - known brown dwarf desert " phenomenon ; @xcite ) . even though the frequency of brown dwarf companions at separations @xmath3930 au is small estimated to range from 0.5% to nearly 20% ( see discussion in @xcite ) hd 23514b is more likely to be a bound companion than an isolated pleiades brown dwarf . if we assume the orbit is circular and face - on , we can use the measured projected separation to estimate the yearly orbital motion of the system . given a mass of 1.35m@xmath33 , we estimate a period of @xmath26000 years . this yields an angular motion of 0.06 degrees per year or a linear motion of @xmath22.8 mas per year . this is comparable to our astrometric precision in our later epochs ( see table [ tab : meas ] ) , suggesting a baseline of at least 5 more years would be needed to confirm orbital motion . however , the yearly motion is likely to be smaller as the orbit may be inclined and/or eccentric . it is instructive to consider whether or not a substellar object at that separation will remain bound over the 100 myr lifetime of the pleiades . figure [ fig : binde ] presents the binding energy of resolved stellar and substellar binaries in the pleiades ( from @xcite ) . note that since inclination and eccentricity are typically not available for these systems ( and certainly not for hd 23514 ) , the true binding energy may be somewhat different than plotted here . despite not being the most widely separated binary in the pleiades ( @xcite present systems at @xmath2900 au ) , hd 23514 does have the lowest binding energy among the stellar mass binaries . however , substellar pleiades binaries with separations @xmath4020 au have comparable or lower binding energies . the similar binding energies suggests that stellar interactions within the pleiades , while more frequent than in the field , may not be enough to disrupt the hd 23514 system in timescales much shorter than 100 myr . we note that hip 78530 @xcite and 1rxs j160929.1210524 @xcite both host substellar companions at wide separations in the 5 myr - old upper scorpius region . these are considered physically bound based on common proper motion , similar to the case of hd 23514 . however , both systems have estimated binding energies lower than that of hd 23514 . previous searches for brown dwarf companions in the pleaides have focused on low - mass stars or free - floating brown dwarfs ( e.g. , * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? * ; * ? ? ? @xcite used adaptive optics to observe 144 g- and k - type stars in the pleiades . the lowest mass companion they found was 0.15 m@xmath1 , though they were sensitive to objects with masses down to @xmath20.03 m@xmath1 at separations of @xmath21 . a deeper study conducted by @xcite also looked at solar - type stars , but found no substellar companions among @xmath220 pleaides stars . hence , while there may be hundreds of free - floating brown dwarfs in the pleaides @xcite , this is the first brown dwarf detected as a companion to a sun - like star . the scarcity of companion brown dwarfs to the @xmath2300 solar - type stars in the pleiades @xcite is consistent with the findings of a brown dwarf desert among solar - type stars out to separations of @xmath21000 au @xcite . hd 23514 is surrounded by a substantial amount of warm ( @xmath2700 k ) dust in a thin ring located @xmath20.25 au from the star if the emission comes from large blackbody grains or at most a few au if the grains are smaller and radiate less efficiently @xcite . at a distance of 135 pc , the companion in figure [ fig : image ] has a projected separation of about 360 au , far larger than the dust semi - major axis . only a handful of systems are known to contain substantial quantities ( @xmath41 ) of warm dust : cha - near member ef cha @xcite , @xmath42 pic member hd 172555 @xcite , lower centaurus crux member hd 113766 @xcite , ab dor member hd 15407 @xcite , m47 member p1121 @xcite , and ngc2547 member id8 @xcite . bd+20 307 is another main - sequence star known to contain a large amount of warm dust @xcite , however it is much older ( @xmath21 gyr ) than these other systems @xcite . of these 8 systems , hd 23514 , hd 15407 , hd 172555 , bd+20 307 , and hd 113766 are known to be binary or triple systems . all are wide separation systems ( @xmath43100 au ) with the exception of bd+20 307 , whose separation is only about 0.05 au ( though with a high probability of a third more distant , and as yet unseen , object ; see @xcite ) . interestingly , hd 23514 , hd 15407 , and hd 172555 all exhibit signs of silica emission suggesting high - velocity impacts produced the dust in the system @xcite . binaries with separations of a few tens of au are less likely to contain substantial amounts of dust when compared with single stars or with wide separation binaries @xcite . however , if hd 23514 hosts a tertiary companion , say a close - in planet , whose initial inclination is different than that of the substellar companion , the kozai mechanism @xcite can work to disrupt its orbit by continuously changing the inclination and eccentricity . this can potentially stir up planetesimals in the system and thus enhance collisions in the terrestrial zone ( for example , see @xcite ) . @xcite suggests a catastrophic or even planet - planet collision to account for the large amount of dust in the system . such an event could take place if the eccentricity of a planet or large planetesimal is sufficiently increased through the kozai mechanism . a large eccentricity can lead to high - velocity impacts near periastron , which can account for the silica emission observed in the system . equation 36 in @xcite can be used to estimate the period of the eccentricity oscillation of this tertiary body . for a planet or planetesimal at 5 au , and assuming hd 23514b s orbit is circular and its projected separation is the semi - major axis , the period is @xmath280 myr . while there are many unknowns , it is interesting to note that this is comparable to the age of the system . further monitoring of the system , as well as deeper searches for any additional companions closer to the star , will be key in determining the relationship , if any , between the secondary companion and the dust around the primary . using adaptive optics imaging at keck , we have discovered a substellar companion to the dusty pleiades star hd 23514 . based on comparison with evolutionary models , we estimate the companion to have a mass of about @xmath26m@xmath33 and temperature @xmath27 k. to our knowledge , hii 1348 and hd 23514 are the only pleiades stars known to have brown dwarf companions @xcite . hd 23514 is one of the dustiest main - sequence stars known to date with @xmath0 . the warm dust present in the system is located within a few au of the primary . in contrast , the companion has a projected separation of @xmath2360 au . at that distance , the companion s gravity is expected to have little influence over the evolution of the warm dust seen around the primary . however , if a tertiary object were present in the system very close to the star , the kozai mechanism can affect its orbit and possibly enhance collisions that generate dust . we thank j. stauffer for useful discussions on the pleiades , r. galicher for the nirc2 coronographic attenuation , and s. yelda for assistance in nirc2 reduction . we thank our anonymous referee for a prompt review and constructive suggestions . the data presented herein were obtained at the w.m . keck observatory , which is operated as a scientific partnership among the california institute of technology , the university of california and the national aeronautics and space administration . the observatory was made possible by the generous financial support of the w.m . keck foundation . the authors wish to recognize and acknowledge the very significant cultural role and reverence that the summit of mauna kea has always had within the indigenous hawaiian community . we are most fortunate to have the opportunity to conduct observations from this mountain . this research was supported in part by nasa grants to ucla . lafrenire , d. , jayawardhana , r. , & van kerkwijk , m. h. 2008 , , 689 , l153 lafrenire , d. , jayawardhana , r. , & van kerkwijk , m. h. 2010 , , 719 , 497 lafrenire , d. , jayawardhana , r. , janson , m. , et al . 2011 , , 730 , 42 2006 - 12 - 10 & k@xmath5 & 30 & 2 & 60 & 18 & adi mode , primary saturated + 2006 - 12 - 10 & k@xmath5 & 0.2 & 100 & 20 & 9 & adi mode , secondary not detected + 2006 - 12 - 10 & k@xmath5 & 0.005 & 50 & 0.25 & 12 & adi mode , secondary not detected + 2007 - 10 - 25 & k@xmath5 & 30 & 2 & 60 & 14 & primary saturated + 2007 - 10 - 25 & k@xmath5 & 0.018 & 20 & 0.36 & 8 & secondary not detected + 2007 - 10 - 25 & k@xmath5 & 0.005 & 50 & 0.25 & 6 & secondary not detected + 2008 - 11 - 04 & ks & 20 & 1 & 20 & 8 & primary saturated + 2008 - 11 - 04 & ks & 0.5 & 20 & 10 & 15 & secondary detected in 9/15 frames + 2009 - 11 - 01 & h & 1 & 60 & 60 & 7 & + 2009 - 11 - 02 & ks c600 & 30 & 1 & 30 & 14 & coronographic ; 600 mas diameter spot + 2010 - 10 - 30 & j & 0.053 & 200 & 10.6 & 5 & + 2010 - 10 - 30 & h & 0.053 & 200 & 10.6 & 10 & + 2010 - 10 - 30 & ks & 0.1 & 200 & 20 & 5 & + 2010 - 10 - 30 & ks c800 & 6 & 2 & 12 & 10 & coronographic ; 800 mas diameter spot + 2006 - 12 - 10 & k@xmath5 & 45 & @xmath44 & @xmath45 & @xmath46 & @xmath47 + 2007 - 10 - 25 & k@xmath5 & 25 & @xmath48 & @xmath49 & @xmath50 & @xmath51 + 2008 - 11 - 04 & ks & 20 & @xmath52 & @xmath53 & @xmath54 & @xmath55 + 2009 - 11 - 01 & h & 25 & @xmath56 & @xmath57 & @xmath58 & @xmath59 + 2009 - 11 - 02 & ks c600 & 20 & @xmath60 & @xmath61 & @xmath62 & @xmath63 + 2010 - 10 - 30 & j & 25 & @xmath64 & @xmath65 & @xmath66 & @xmath67 + 2010 - 10 - 30 & h & 25 & @xmath68 & @xmath69 & @xmath70 & @xmath71 + 2010 - 10 - 30 & ks & 25 & @xmath72 & @xmath73 & @xmath74 & @xmath75 + 2010 - 10 - 30 & ks c800 & 25 & @xmath76 & @xmath77 & @xmath78 & @xmath79 + . the uncertainty in these two measurements are comparable to the symbol size and they effectively lie on top of each other in the figure . the black line indicates the relative motion a stationary background source would have throughout these observations with circles denoting the expected location at the time of our observations . we adopt a distance of @xmath19 pc ( see section [ results ] ) . , width=529 ] ) as a function of separation for our 2010 and 2009 coronographic data . limits are computed using the same aperture size by performing aperture photometry on blank regions of the sky in annuli around the primary . the limits are averaged over all frames . companions @xmath211 - 12 magnitudes fainter at ks than the primary could have been detected at separations larger than 1.5 . mass limits in solar masses from the dusty evolutionary model of @xcite are indicated . , width=529 ]
with adaptive optics imaging at keck observatory , we have discovered a substellar companion to the f6 pleiades star hd 23514 , one of the dustiest main - sequence stars known to date ( @xmath0 ) . this is one of the first brown dwarfs discovered as a companion to a star in the pleiades . the 0.06 m@xmath1 late - m secondary has a projected separation of @xmath2360 au . the scarcity of substellar companions to stellar primaries in the pleiades combined with the extremely dusty environment make this a unique system to study .
clusters of galaxies are the largest gravitationally bound and nearly virialized systems in the universe . high resolution x - ray surveys have revealed that the icm ( with temperatures in the range @xmath1 kev ) is not fully relaxed . the substructures in the surface brightness and temperature are observed both on large scales and in the cluster cores indicating that the gas is not at rest . at the same time strong shocks are rarely observed ( see markevitch & vikhlinin , 2007 for a review ) suggesting that the cluster gas motions are predominantly subsonic . the shape of the emission lines as a source of information on the icm velocity distribution has been discussed in detail in sunyaev , norman & bryan ( 2003 ) and inogamov & sunyaev ( 2003 ) . these papers mostly consider the gas motions induced by cluster mergers . in the present paper we focus on the linewidth as a diagnostic of gas motions in the cores of the galaxy clusters . in the cores of most clusters and many groups ( e.g. stewart et al 1984 , nulsen et al 1984 , edge & stewart 1991 , mulchaey et al 1993 , fabian 1994 ) , the radiative cooling time of the gas is short compared to the age of the cluster and an external source of energy is needed to avoid catastrophic cooling ( e.g. david et al . 2001 , matsushita et al . 2002 , peterson et al . 2003 , kaastra et al . the dissipation of turbulent gas motions is one of the highly plausible sources of the required energy ( e.g. pedlar et al . 1990 , churazov et al . 2002 , fabian et al . 2003a , chandran 2005 ) . these motions can have distinctive effects on the properties of the x - ray lines ( e.g. brggen , hoeft & ruszkowski 2005 ) . there are several indirect ways of probing the gas velocities , i ) using the resonant scattering of the brightest emission lines ( e.g. gilfanov , sunyaev & churazov , 1987 , churazov et al . , 2004 ) ; ii ) considering the spreading of metals ejected by a central galaxy through the icm ( rebusco et al . , 2005,2006 , graham et al . , 2006 , roediger et al . , 2007 ) ; iii ) using h@xmath2 emitting filaments as tracers of the gas motions ( e.g. fabian et al . , 2003b , hatch et al . , in particular in rebusco et al . ( 2005,2006 ) the characteristic spatial and velocity scales of the turbulent eddies were estimated assuming that the dissipation of turbulent motions can generate enough heat to prevent the gas from extreme cooling and that the same motions provide moderate mixing of metals through the icm . using these constraints we calculated the expected width of the brightest x - ray lines for the brightest clusters ( e.g. , a426 - perseus ) . a micro - calorimeter with energy resolution of less than @xmath3 ev and effective area of @xmath4 @xmath5 has been built for the suzaku mission ( kelley et al . , 2007 ) . the effective areas of the upcoming calorimeters may reach thousands and even tens of thousands of square cm in the next 10 - 15 years and will have a spatial resolution between a few arcseconds and a few arcminutes ( see e.g. ` http://www.astro.isas.ac.jp/future/next ` , ` http://www.rssd.esa.int/index.php?project=xeus ` , ` http://constellation.gsfc.nasa.gov/ ` ) . the spektr - rg calorimeter ( sxc ) ( mccammon et al . , 2007 ; mitsuda et al . , 2007 ) , proposed for the spektr - rg mission , has a peak effective area of 300 @xmath5 at @xmath6 kev , decreasing to about @xmath7 @xmath5 at @xmath8 kev and an energy resolution ( goal ) of @xmath9 ev . as we show in section @xmath10 , an instrument with these characteristics is well suited to study the line broadening associated with turbulent gas motions in the cores of the brightest galaxy clusters . the structure of the paper is the following . in section [ sec : turb ] we describe the line broadening under different assumptions for the character of the gas motions ; in section 3 we consider the specific case of the perseus cluster as an example and we investigate the capabilities of sxc in this context . our findings are discussed and summarized in sections 4 and 5 . throughout the paper when we mention turbulence , we mean all types of macroscopic gas motions . there are several bright emission lines in the x - ray range , that are characteristic of the icm in galaxy clusters ( table [ tab : lines ] ) . the he - like iron line at 6.7 kev is the most prominent feature above a few kev : it is especially bright in high temperature ( @xmath11 kev ) clusters . for the lower temperature clusters , and in particular for the cooling flow regions , l lines of iron at @xmath01 kev , lines of si , mg and o can be substantially brighter than the 6.7 kev line . in table [ tab : lines ] we list a sub - sample ( incomplete ) of these lines and provide an estimate of their fluxes from the central @xmath12 ( radius ) region of the perseus cluster . all line parameters were taken from the atomdb data base ( smith et al . 2001a , b ) . [ cols="^,^,^,^,^,^,^,^ " , ] it is obvious from equation [ eq : sigmas ] that the heaviest elements are the best probes of gas motions , since more massive nuclei have smaller thermal line broadening and , therefore , the turbulent broadening can dominate the thermal broadening . indeed the ratio of the e.g. iron thermal velocity to the proton thermal velocity is small ( @xmath13 ) . thus even for small velocities the turbulent broadening will exceed the thermal broadening ( table [ tab : fwhm ] , figure @xmath14 ) . for an instrument with a given spectral resolution the energy of the line also plays an important role , since the width of the line simply scales with the line energy . this makes the iron 6.7 kev line the best probe of the turbulent icm , as long as the instrument has a substantial effective area at 6 - 7 kev . the above assumption of a gaussian line shape due to turbulent motions is of course only an approximation ( see inogamov & sunyaev 2003 for a discussion ) . however it is fully sufficient to assess the detectability of the turbulent broadening . in fully developed turbulence an entire range of scales is produced through an energy cascade ( e.g. richardson 1922 , tennekes & lumley 1972 , lesieur 1997 , mathieu 2000 , davidson 2004 ) , which is characterized by large scale energy - containing motions and a small dissipation scale . in our simplified treatment we assume that the @xmath15 entering all our equations can be obtained from the kinetic energy of the turbulent motions : @xmath16 , where @xmath17 is the gas density . since the integration of equation [ eq : int ] may result in a line shape different from a pure gaussian we use below an `` effective '' full width half maximum ( efwhm ) , which is defined as the interval of energies centered at the line energy and containing @xmath18 of the line flux . there is indirect evidence that the gas in cluster centers is involved in some sort of motion . h@xmath2 emitting filaments , that are thought to be drawn behind buoyant gas bubbles , have been used to trace the icm velocity field ( e.g. hatch et al . 2006 , 2007 , salom et al . 2007 ) . however neither the characteristic velocity scales nor the characteristic patterns of the hot gas motions have been measured in a more direct way so far . one can identify two major sources of turbulent gas motions in the cores : cluster mergers and the action of a central active galactic nucleus ( agn ) . each process may produce a velocity field different in intensity and directionality . let us consider the extreme cases of isotropic , pure radial and pure tangential velocity fields as illustrated in figure [ fig : sketch ] . if one fixes the total kinetic energy in the motions , then the line of sight component of the velocity can be written as : @xmath19 the first case ( isotropic gas motions ) can occur for a wide variety of driving mechanisms , ranging from mergers to the motions caused by rising buoyant bubbles of relativistic plasma or to convection driven by a mixture of thermal plasma and cosmic rays ( chandran 2005 ) . pure radial gas motions would naturally appear if energy generated by a central agn goes into weak shocks and sound waves ( e.g. fabian et al . 2003 , forman et al . 2005 , 2006 , fabian et al . 2006 ) which propagate through the icm almost radially from the central source . the case of pure tangential motions seems to be less natural and it is included mostly for completeness . we note however that in the stratified atmosphere of clusters the characteristic frequencies of the turbulent motions can be smaller than the brunt - visl frequency : @xmath20 where @xmath21 is the gravitational acceleration due to the dark matter potential , @xmath17 is the gas density , @xmath22 is its pressure and @xmath23 is the adiabatic index . in figure [ fig : bv ] we plot the brunt - visl frequency , evaluated for the perseus cluster , and compare it with the characteristic frequency of turbulent motions @xmath24 for @xmath25 km s@xmath26 and @xmath27 kpc . in the range where the turbulent frequency is smaller than the local brunt - visl frequency , the gas motions may excite gravity waves ( e.g. churazov et al . , 2001 , 2002 , omma et al . , 2004 ) . these waves can propagate and transport energy from a localized patch of turbulent icm within this region ( e.g. in the case of perseus within a radius of @xmath28 arcmin ) . such a process is well known in atmospheric science and oceanography and it may lead to the formation of non - propagating pancake - shaped vortices ( e.g. riley & lelong 2000 ) . therefore , one can imagine that patches of predominantly 2-dimensional ( tangential ) vortices can be created with this mechanism . we now calculate ( figure [ fig : beta ] ) the expected width of the iron 6.7 kev line for each of the three limiting cases given in equation [ eq : vel ] , for a model cluster with the density parameterized by a simple @xmath29 model : @xmath30^{-3/2 \beta}$ ] , where @xmath31 is the core radius of the cluster and @xmath32 . the gas temperature and the abundance of iron are assumed to be constant . the velocity @xmath15 is fixed to 300 @xmath33 . the thermal broadening ( first term in equation [ eq : sigma ] ) is omitted in the calculation of the linewidth , but the level of thermal broadening for a 6 kev plasma is shown for comparison ( dashed line ) . the three plots in figure [ fig : beta ] correspond to @xmath32 from left to right , respectively . as expected from equation [ eq : vel ] , isotropic turbulence produces a linewidth independent of the projected distance from the cluster center . pure radial gas motions would result in a line whose width is peaked towards the center of the cluster , while pure tangential motions produce the broadest line outside the cluster core . for @xmath34 , the curve approaches asymptotic values ( different for each velocity pattern ) , that are functions of @xmath29 only . for large @xmath29 the cluster is more `` compact '' and the gradients of the width of the line with radius become stronger , while for low @xmath29 the change of the linewidth is more gradual . it is quite plausible that the characteristic turbulent velocity scale varies with the distance from the cluster center . if gas motions are driven by the outflows of a central agn , then their amplitude will likely decrease for large radii , since the energy would be spread over larger masses of gas and it would be partly dissipated at smaller radii . for example , the characteristic amplitude of spherical sound waves propagating through a declining density profile decreases as @xmath35 . for @xmath36 the velocity varies as @xmath37 . if , furthermore , dissipation is taking place then the amplitude will decline more quickly , as suggested by fabian et al . ( 2003 ) and studied in three - dimensional viscous simulations by ruszkowski et al . ( 2004 ) . for rising buoyant bubbles of relativistic plasma the terminal velocity scales as @xmath38 ( e.g. churazov et al . , 2001 ) , where @xmath39 is the keplerian velocity at radius @xmath40 and @xmath41 is the characteristic size of the bubble . the adiabatic expansion of the bubble leads to a slow change of the bubble size @xmath42 ( assuming that the adiabatic index of the medium inside the bubble is 4/3 ) . thus the characteristic bubble velocity is @xmath43 - a decreasing function of the radius for plausible @xmath44 and @xmath45 . if the bubble breaks down into smaller bubbles , then the terminal velocity will decrease even further with the radius . alternatively , if the turbulence is driven by shocks originating from minor mergers , then it is possible that the velocity amplitude will instead decline towards the center , where the gas density and the thermal gas pressure are the highest . for example , for a plane sound wave propagating into a region of increasing density the characteristic velocity is @xmath46 , i.e. @xmath47 if @xmath48 . here we have ignored both the possible reflection of the sound waves by a steep density gradient ( important for long wavelength perturbation ) and the focusing of sound waves due to decreasing sound speed towards the center ( pringle , 1989 ) . in reality the effect of mergers on the cluster core is much more complicated and the above estimate can at best be considered as indicative . for illustration we model two possibilities ( decreasing or increasing turbulence ) by considering @xmath49 , with @xmath50 and @xmath51 km / s . we further assume that when @xmath52 km / s , the velocity saturates at @xmath53 km / s . the resulting linewidth profiles are shown in figure [ fig : raising ] . clearly the radial dependence of the velocity amplitude can strongly affect the behavior of the linewidth as a function of projected distance . we now calculate the expected width of the 6.7 kev line by using the perseus cluster as an example . perseus ( a 426 ) is the brightest nearby x - ray cluster and it is one of the best - studied cool core clusters , together with m87 and centaurus . it hosts in its core a luminous elliptical galaxy ngc 1275 , containing a bright radio source ( 3c 84 ) . in the core region a complex substructure is seen in x - ray temperature , x - ray surface brightness and optical light distributions . such substructure includes holes in the x - ray images due to bubbles of relativistic plasma ( bhringer et al . , 1993 , fabian et al . 2000 ) , quasi - spherical ripples ( fabian et al . 2003a ) and optical h@xmath2 filaments ( fabian et al . 2003b , hatch et al . 2005 ) . in our estimates of the expected linewidth in perseus , we assume that the gas cooling losses are compensated by the dissipation of the turbulent motions at all radii . this is of course a strong and not fully justified assumption , but it provides clear predictions if the turbulent heating is indeed important in cluster cores . thus one can write : @xmath54 where @xmath55 is the gas cooling function and @xmath56 is a dimensionless constant of the order of unity ( see e.g. dennis & chandran , 2005 ) , which depends on the character of the turbulent motions . since the electron density and temperature are known from observations ( e.g. churazov et al . , 2003 ) one can estimate @xmath57 from equation [ eq : bal ] . as the linewidth depends explicitly on @xmath15 one has to determine ( or assume ) some constraints on @xmath58 . for instance , rebusco et al . ( 2005 , 2006 ) and graham et al . ( 2006 ) considered the turbulent spreading through the icm of metals produced by the brightest cluster galaxy . treating this process in a diffusion approximation one can estimate the effective diffusion coefficient @xmath59 . in the perseus cluster @xmath60 ( rebusco et al . one can then cast @xmath59 in the form @xmath61 , where @xmath62 is a dimensionless constant of the order of unity ( see dennis & chandran 2005 for a compilation of values of @xmath56 and @xmath62 ) . combining equation [ eq : bal ] ( at some radius ) and the expression for the diffusion coefficient , both @xmath15 and @xmath58 can be measured . in perseus the characteristic values were found to be @xmath63 20 kpc and @xmath64 410 km / s ( rebusco et al . if we adopt @xmath65410 km / s as the characteristic velocity of turbulent motions at every radius in perseus then , depending on the directionality of the turbulence ( see figure [ fig : pdir ] ) , the expected width of the 6.7 kev line in perseus should be in the range 10 - 20 ev . of course the above values of @xmath58 and @xmath15 are only order of magnitude estimates . from equation [ eq : bal ] it is clear that @xmath66 . if @xmath67 then @xmath68 ( when @xmath69 ) . in figure [ fig : heat ] we calculate the linewidth for several illustrative cases : three solid lines with different thickness correspond to @xmath702 , 20 , 200 kpc ( at all radii ) . this spread of two orders of magnitude in @xmath58 results in a factor of 3 change of the linewidth . if instead we follow the arguments of dennis & chandran ( 2005 ) and set @xmath71 , where @xmath72 , then @xmath73 const when @xmath69 . the corresponding curve is shown in figure [ fig : heat ] by the dashed line . in order to make some estimates we consider the proposed sxc with an effective area which can be provided by the current generation of x - ray mirrors with a rather short focal length of less than @xmath74 meters . at @xmath8 kev the effective area is @xmath75 @xmath5 . we adopt a spectral and spatial resolution of @xmath76 ev and @xmath14 arcmin respectively . the design goal for sxc is 4 ev resolution , a field of view of @xmath77 arcmin square and pixels of @xmath78 arcmin square . we want to check how accurately a micro - calorimeter with such parameters will measure the turbulent broadening in the perseus cluster . the expected number of photon counts in the iron complex in a 1 msec observation is plotted in figure [ fig : flux ] . note that , although we focus on the line broadening of the @xmath8 kev line , in what follows we use all the photon counts in the 0.5 - 8.0 kev band , in order to obtain better fits ( for reference the number of counts due to the @xmath8 kev line accounts for @xmath79 of the total number of counts in the 0.5 - 8.0 kev band ) . we simulated the observed spectra using xspec , specifically the _ bapec _ model , which takes into account isotropic broadening . for @xmath80 km / s , the component of the velocity along the line of sight is about @xmath81 km / s . an example of the simulated spectrum ( for a 1 msec long observation ) is shown in figure [ fig : xspec ] . the uncertainty in the measured 6.7 kev line broadening as a function of the number of counts is plotted in figure [ fig : fake ] . the solid line shows the true broadening ( @xmath81 km / s ) , while the long - dashed and dotted lines indicate the level of @xmath14 sigma confidence for the fitted velocity . the comparison with figure 8 suggests that already within @xmath82 arcmin there are enough photons to measure the turbulent velocity with an uncertainty of @xmath83 km / s . figure [ fig : fake ] assumes a plasma temperature of @xmath76 kev . we repeated the simulations also for higher temperatures and assuming a 2-temperature plasma . in both cases , the turbulent velocity can be recovered with an uncertainty below @xmath84 km / s . in order to test the capabilities of our fiducial micro - calorimeter , we also calculated the accuracy of the fits within each single square of @xmath85 arcmin . even in the outer parts , the resolution is high enough to measure the turbulent velocity with @xmath14 sigma confidence below @xmath3 km / s . therefore the spatial dependence of the turbulent velocity could be directly measured and this would give clear indications about the origin of the turbulent gas motions ( see section @xmath74 ) . all the results obtained in the previous sections are based on the assumption that the characteristic size of the turbulent eddies @xmath58 is much smaller than the characteristic length of the line of sight . in practice there are at least two complementary ways of deriving information on the turbulent gas motions with micro - calorimeters . one can use i ) the variations of the line centroid with position and/or ii ) the width / shape of the line . assuming that the resolution of the instrument is substantially better than the width of the observed line ( e.g. see fig . [ fig : heat ] ) the accuracy of determination of the line characteristics ( in the case of a strong single line with negligible continuum , which corresponds to the most optimistic scenario ) is : @xmath86 where @xmath87 is the width of the line , @xmath88 is the number of counts in the line , @xmath89 and @xmath90 are the 1@xmath91 errors in the centroid and width of the line . + the centroid variation method works best only if the characteristic size of the eddies is not too small . indeed , in practice one observes the spectrum coming from a selected region of the cluster with characteristic sizes @xmath92 ( in projection ) and @xmath93 ( along the line of sight ) . if @xmath94 and @xmath95 then @xmath96 independent eddies are seen by the instrument at the same time . assuming that each eddy produces a centroid shift of the order of @xmath97 , then the observed centroid shift will be @xmath98 . increasing @xmath92 would increase the number of line photons the instrument detects during observations , but this would simultaneously decrease the amplitude of the centroid shift . obviously these two effects cancel each other and increasing @xmath92 beyond the characteristic eddy size @xmath58 does not improve the detectability of the centroid variations . making @xmath92 smaller than @xmath58 would just make the number of line photons smaller : the amplitude of the centroid variation should not vary much . of course studying the centroid variations on spatial scales larger and smaller than @xmath58 will be a very useful test of the distribution of characteristic eddy sizes , but for the principal detection of the gas turbulent motions , the choice of @xmath99 is obviously the optimal one . thus the question of prime importance is whether the instrument has an effective area large enough to collect photons from a region with size @xmath100 ( e.g. sufficient to detect the centroid variations ) . of course , averaging along the line of sight remains an unavoidable effect independent of the choice of the region size . in figure [ fig : el ] we plot the effective length @xmath93 of the region along the line of sight which provides 75% of the total surface brightness in the 6.7 kev line , at a given projected distance from the center of the perseus cluster . it is clear from this figure that , in the core of the perseus cluster , @xmath93 is of the order of 130 kpc and it increases linearly outside the core . the non - monotonic behavior of the curve at 20 - 30 kpc from the center is caused by the complicated iron abundance profile ( see e.g. schmidt et al . 2002 , churazov et al . thus , for the optimal size of the observed region , the typical amplitude of the centroid variations will be @xmath101 . for estimating the centroid variations , we adopt @xmath64 400 km / s ( i.e. in the isotropic case @xmath102 237 km / s ) and @xmath63 20 kpc ( rebusco et al . , 2006 ) . in the perseus cluster a single @xmath103 sxc pixel , corresponds to a @xmath104 kpc region pixel , and an effective length along the line of sight of @xmath105 kpc ( see fig.[fig : el ] ) . thus one can expect @xmath106 eddies ( assuming a @xmath107 volume per eddy ) and the typical centroid variations on these spatial scales are of the order of 50 km / s . such a shift in velocity corresponds to a @xmath108 1.1 ev shift in the line energy from pixel to pixel for the 6.7 kev line . from eq . 8 it follows that such variations should be detectable with @xmath109 significance in sxc pixels with @xmath105 counts ( for the 6.7 kev line with fwhm of 13 ev - as in fig . 6 ) . as one can see in fig . 8 the number of photons in the central pixel for a 1 msec observation of the perseus cluster is much higher ( of the order of @xmath110 ) : hence the centroid variation can be measured with great accuracy . + the detection of the line broadening has no limitation from the point of view of the region size and the whole cooling flow region can be probed at once . according to eq . 9 the 1@xmath91 error on the width of the line is @xmath111 ev ( for the 6.7 kev line with fwhm of 13 ev and number of detected photons @xmath112 ) . this corresponds to a change in the characteristic velocity scale of @xmath113 km / s . in a 1 msec observation such broadening can be measured in each of the @xmath103 pixels of sxc up to a distance of @xmath114 arcminutes from the cluster center ( note that sxc s field of view is @xmath115 arcmin ) . from fig . 10 and from eq . 9 it follows that during a single 1 msec observation of the perseus cluster with sxc , the turbulent broadening at the level of a few hundred km / s can be measured in a central @xmath103 pixel and a map of the linewidth with the accuracy of the order of @xmath3 km / s covering the area @xmath116 can be obtained . given the expected dependencies of the line broadening shown in fig.6 and 7 , such a dataset would allow one to firmly establish the level of microturbulence in the perseus cluster and test the hypothesis that turbulent dissipation acts as a mechanism for heating the icm . the dissipation of turbulent gas motions driven by agn activity is a plausible source of heat for the cooling gas in cluster cores . in this case , we show that the expected width of the iron 6.7 kev line is well above the thermal broadening for most plausible values of the turbulent velocity . in our fiducial model of the perseus cluster , the 6.7 kev linewidth is larger than 10 ev . hence the proposed sxc micro - calorimeter is well suited for firmly establishing the level of microturbulence in the brightest galaxy clusters . + as discussed in sections 2.1 and 2.2 the linewidth is sensitive to both the radial dependence of the velocity amplitude and the `` directionality '' of the stochastic motions . this may complicate an unambiguous determination of the properties of the gas motions . however this degeneracy can be removed if the width of several lines , characteristic for different gas temperatures , can be measured . assume for example that in addition to the 6.7 kev iron line , the width of the iron l - shell line is detected . this line is bright at lower temperatures , that are present only in the central region ( cool core ) . therefore the combined measurement of the width of two or more lines should be sufficient to discriminate among the simplest models with different `` directionalities '' . we acknowledge the support by the dfg grants ch389/3 - 2 . we would like to thank dan mccammon , jan - willem den herder , richard kelley and kazu mitsuda for the data and discussions about sxc . we wish to thank the anonimous referee for his / her useful comments . p.r . is grateful to the international max planck research school for its support , to andreas bauswein , marcus brggen and rasmus voss for the helpful discussions and to the `` velisti per caso '' of the adriatica for their enthusiasm . 99 bhringer h. , voges w. , fabian a.c . , edge a.c . , neumann d.m . 1993 , mnras , 264 , l25 brggen m. , hoeft m. , ruszkowski m. 2005 , apj , 628 , 153 chandran b.d.g . 2005 , apj , 632 , 809 churazov e. , brggen m. , kaiser c.r . , bringer h. , forman w. 2001 , apj , 554 , 261 churazov e. , sunyaev r.a . , forman w. , bringer h. 2002 , mnras , 332 , 729 churazov e.,forman w. , jones c. , bringer h. 2003 , apj , 590 , 225 churazov e.,forman w. , jones c. , sunyaev r. , bringer h. 2004 , mnras,347 , 29 david l. , nulsen p. , mcnamara b. , forman w. , jones c. , ponman t. , robertson n. , wise m. 2001 , apj , 557 , 546 davidson p.a . 2004 , _ turbulence _ , oxford univ . press ( new york ) dennis t. , chandran b. 2005 , apj , 622 , 205 edge a.c . , stewart g.c . 1991 , mnras , 252 , 414 fabian a.c . 1994 , ara&a , 32 , 277 fabian a. c. et al . , 2000 , mnras , 318 , l65 fabian a. c. , sanders j.s . , allen s.w . , crawford c.s . , iwasawa k. , johnstone r.m . , schmidt r.w . 2003a , mnras , 344 , l43 fabian a.c . et al . 2003b , mnras , 344 , l48 fabian a.c . , sanders j.s . , taylor g.b . , allen s.w . , crawford c.s . , johnstone r.m . , iwasawa k. 2006 , mnras , 366 , 417 forman w. et al . 2005 , apj , 635 , 894 forman w. et al . 2006 , astro - ph/0604583 gilfanov m.r . , sunyaev r.a . , churazov e. 1987 , sval , 13 , 233 graham j. , fabian a. c. , sanders j. s. , morris r. g. , 2006 , mnras , 368 , 1369 hatch n.a . , crawford c.s . , fabian a.c . 2007 , mnras , 380 , 33 hatch n.a . , crawford c.s . , johnstone r.m . , fabian a.c . 2006 , mnras , 367 , 433 inogamov n.a . , sunyaev r.a . 2003 , astron . lett . , 29,12 kaastra j.s . et al . 2004,a&a , 413 , 415 lesieur m. 1997 , _ turbulence in fluids _ , reidel ed . ( dordrecht ) kelley r. l. , et al . , 2007 , pasj , 59 , 77 markevitch m. , vikhlinin a. 2007 mathieu j. 2000 , _ an introduction to turbulent flow _ , cambridge univ . press ( cambridge ) matsumoto h. , koyama k. , awaki h. , tomida h. , tsuru t. , mushotzky r. , hatsukade i. , 1996 , pasj , 48 , 201 matsushita k. , belsole e. , finoguenov a. , bhringer h. 2002 , a&a , 386 , 77 mccammon d. et al 2007 , in preparation mitsuda k. , kelley r. , mccammon d. , den herder j - w , ohashi t. 2007 , journal of low temperature physics , in preparation mulchaey j.s . , davis d.s . , mushotzky r.f . , burstein d. 1993 , apj , 404 , l9 nulsen p.e.j . , stewart g.c . , fabian a.c . 1984 , mnras , 208 , 185 omma h. , binney j. , bryan g. , slyz a. 2004 , mnras , 348 , 1105 pedlar a. et al . 1990 , mnras , 246 , 477 peterson j.r . , kahn s.m . , paerels f.b.s . , kaastra j.s . , tamura t. , bleeker j.a.m . , ferrigno c. , jernigan j.g . 2003 , apj , 590 , 207 pringle j.e . 1989 , mnras , 239 , 479 rebusco p. , churazov e. , bhringer h. , forman w. , 2005 , mnras , 359 , 1041 rebusco p. , churazov e. , bhringer h. , forman w. , 2006 , mnras , 372 , 1840 richardson , l.f . 1922 , weather prediction by numerical processes , cambridge univ . press riley j.j . , lelong m. 2000 , annu.rev.fluid mech . , 32 , 613 roediger e. , brggen m. , rebusco p. , bhringer h. , churazov e. 2007 , mnras , 375 , 15 ruszkowski m. , brggen m. , begelman m. c. 2004 , apj , 615 , 675 salom p. et al . 2006 , a & a 454 , 437 schmidt r.w . , fabian a.c . , sanders j.s . 2002 , mnras , 337 , 71 smith r.k . , brickhouse n.s . , liedahl d.a . , raymond j.c . 2001a , apj , 556 , l91 smith r.k . , brickhouse n.s . , liedahl d.a . , raymond j.c . 2001b , spectroscopic challenges of photoionized plasmas , asp conference series vol . 247 , 159 . wolf savin stewart g.c . , fabian a.c . , jones c. , forman w. 1984 , apj , 285 , 1 sunyaev r.a . , norman m.l . , bryan g.l . 2003 , astl , 29 , 783 tennekes h. , lumley j.l . 1972 , _ a first course in turbulence _ , mit press ( cambridge )
the dissipation of turbulent gas motions is one of the likely mechanisms that has been proposed to heat the intracluster medium ( icm ) in the cores of clusters and groups of galaxies . we consider the impact of gas motions on the width of the most prominent x - ray emission lines . for heavy elements ( like iron ) the expected linewidth is much larger than the width due to pure thermal broadening and the contribution due to turbulent gas motions should be easily detected with the new generation of x - ray micro - calorimeters , such as the spektr - rg calorimeter ( sxc ) . for instance in the perseus cluster the turbulent velocity required to balance radiative cooling ( as derived by rebusco et al . 2006 ) , would imply a width of the 6.7 kev fe line of 10 - 20 ev , while the pure thermal broadening is @xmath04 ev . the radial dependence of the linewidth is sensitive to i ) the radial dependence of the velocity amplitude and ii ) the `` directionality '' of the stochastic motions ( e.g. isotropic turbulence or predominantly radial gas motions ) . if the width of several lines , characteristic for different gas temperatures , can be measured , then it should be possible to probe both the `` directionality '' and the amplitude of the gas motions . moreover a measurement of the width would put a lower limit on the amount of the kinetic energy available for dissipation , giving a constraint on the icm models . [ firstpage ] turbulence - line : profiles - cooling flows - clusters : individual : perseus
low - dimensional , correlated systems have attracted much attention in recent years due to a large variety of unconventional electronic properties directly connected with electronic correlations . in this context one - dimensional ( 1d ) systems are of special interest , essentially stimulated by theoretical predictions like charge - spin separation for the 1d hubbard model @xcite . while the 1d hubbard model @xcite is the prototype of an exactly solvable model for correlated electrons , such a charge - spin separation corresponds to the description of the energy eigenstates in terms of occupancy configurations of holons and spinons . recently the charge - spin separation of that model was shown to occur for all values of excitation energy , the corresponding holon and spinon description being extended to the complete set of energy eigenstates which span the hilbert space @xcite . on the other hand , quasi-1d materials are ideal model systems which allow the study of basic physical concepts in one dimension . in the last years there has been a renewed interest on the unconventional spectral properties of these materials @xcite . moreover , recent angle - resolved ultraviolet photoemission spectroscopy revealed very similar spectral fingerprints from both high-@xmath0 superconductors and quasi-1d compounds @xcite . the similarity of the ultraviolet data for these two different systems could be evidence of the occurrence of the above mentioned charge - spin separation associated with holons and spinons . the anomalous temperature dependence of the spectral function could also indicate a dimensional crossover @xcite . the results of refs . @xcite also suggest that the unconventional spectral properties observed in two - dimensional ( 2d ) materials could have a 1d origin . thus the holons and spinons could play an important role in spectral properties of both 1d and 2d low - dimensional materials . recently it was found that for the 1d hubbard model there are exact selection rules which limit the number of holons and spinons created by application of rotated electron operators onto any eigenstate of the spin @xmath1 electron number @xcite . the concept of rotated electron is associated with a unitary transformation introduced in ref . for such rotated electrons double occupation is a good quantum number for all values of the on - site coulombian repulsion @xmath2 . in order to distinguish the rotated - electron description from the electron representation , the concept of _ effective electronic lattice _ was introduced @xcite . it is such that rotated electrons occupy the effective electronic lattice , whereas electrons occupy the _ real - space lattice_. however , these lattices are identical , having the same lattice constant @xmath3 and length @xmath4 , where @xmath5 designates the number of sites . the occurrence of rotated electrons leads to clear finger prints in the few - electron spectral properties . indeed , the few - electron spectral functions include several upper hubbard bands corresponding to different and separated energy scales . the @xmath6 band is spanned by excited states of rotated - electron double occupation @xmath7 . however , in general only the first few upper bands have a significant amount of spectral weight . although the above - mentioned selection rules refer to rotated - electron operators , their existence induces restrictions in the numbers of holons , spinons , and pseudoparticles contained in few - electron excitations . for instance , according to the results found in ref . @xcite , about 99% of the upper hubbard band spectral weight of the one - electron addition spectral function corresponds to the first upper hubbard band . interestingly , rotated electron double occupation @xmath7 equals the number of holons of @xmath8 spin projection @xmath9 and charge @xmath10 . here @xmath11 denotes the electronic charge and the @xmath8 spin is a @xmath12 algebra associated with the charge degrees of freedom @xcite . the charge @xmath10 and @xmath13 holons refer to all values of the on - site coulomb repulsion @xmath2 and are related to the large @xmath2 dublons and holons respectively , studied in ref . @xcite . the related to the electron - rotated electron unitary transformation is part of the non - perturbative diagonalization of the quantum problem . through out this paper the designation rotated - electron site distribution configuration ( and electron site distribution configuration ) refers to the effective electronic lattice ( and real - space lattice ) . a property of deep physical meaning is that the energy eigenstates are described by the the same rotated - electron site distribution configurations for all values of @xmath14 , where @xmath15 is the first - neighbor transfer integral . this is in contrast to the complex electron site distribution configurations of these states , which change upon variations of the value of the on - site repulsion . fortunately , the electron - rotated electron unitary transformation becomes the unit transformation as @xmath16 . thus one can reach the @xmath14 independent rotated - electron site distribution configurations which describe the energy eigenstates by studying the corresponding electron site distribution configurations for the 1d hubbard model in the limit @xmath16 . in the present paper we study such @xmath14 independent rotated electron configurations . in part , this is achieved by considering the corresponding problem of the electron real - space lattice charge ( and spin ) sequences which describe the energy eigenstates of the 1d hubbard model in the limit @xmath16 . the problem of the 1d hubbard model in the limit of @xmath16 has been previously studied by many authors @xcite . in such a limit there is a huge degeneracy of @xmath8-spin and spin occupancy configurations . thus there are several choices for complete sets of energy eigenstates with the same energy and momentum spectra . however , only one of these choices is associated with the rotated - electron site distribution configurations reached by the electron - rotated electron canonical unitary transformation . it refers to the complete set of energy eigenstates generated from the corresponding energy eigenstates of the 1d hubbard model for finite values of @xmath14 by turning off adiabatically the parameter @xmath17 . we call the obtained set of states of the model _ band - momentum energy eigenstates_. the electron occupancy configurations which describe these states have not been studied so far . in addition to study this problem , in this paper we discuss the relation of the obtained states to an alternative choice of energy eigenstates whose electron site distribution configurations were studied in ref . @xcite . concerning the open questions whose clarification our results are useful for , one should mention the evaluation of few - electron spectral functions for finite values of excitation energy . except in the limit @xmath16 @xcite , this is an important open problem of interest for the further understanding of the unconventional spectral properties observed in low - dimensional materials . the relationship of the pseudoparticles , holons , and spinons to the electrons is a complex problem of crucial importance for the evaluation of these spectral functions . as discussed in later sections , the concepts of local pseudoparticle and effective pseudoparticle lattices introduced in this paper , as well as the relationship of the energy - eigenstate pseudoparticle occupancy configurations in these effective lattices to the rotated - electron occupancy configurations , provide useful information about such a relationship . as open questions are concerned , we emphasize that the studies of this paper are a necessary step and a valuable contribution for the evaluation of the overlap between few - electron excitations and energy eigenstates , as further discussed in sec . elsewhere the tools and concepts introduced in this paper are used in the study of few - electron spectral distributions for finite values of momentum and excitation energy @xcite . importantly , predictions obtained recently by application of a preliminary version of the theory introduced in this paper and in refs . @xcite seem to describe both qualitatively and quantitatively the one - electron removal spectral lines observed in real quasi-1d materials by photoemission experiments for finite values of the excitation energy @xcite . the paper is organized as follows : in sec . ii we introduce the 1d hubbard model and summarize the concept of rotated electron as well as the description of the model in terms of holons , spinons , and pseudoparticles . in sec . iii we introduce the complete set of energy eigenstates of the 1d hubbard model in the limit @xmath16 which refers to the electron - rotated electron canonical unitary transformation . this includes the introduction of a set of basic properties which turn out to be useful for finding the internal structure of the local @xmath18 pseudoparticle , which is one of the subjects of sec . iv . in sec . iv we introduce a complete basis of local states which we express in terms of rotated - electron local charge , spin , and @xmath19 pseudoparticle sequences . in that section we also find the rotated - electron site distribution configurations which describe the internal structure of the local @xmath18 pseudoparticle . the local @xmath19 pseudoparticle and @xmath20 pseudoparticle effective lattices are introduced in sec . v. in that section we also express the energy eigenstates in terms of the fourier - transform superpositions of local charge , spin , and @xmath19 pseudoparticle sequences introduced in the previous section . a pratical application of the derived local pseudoparticle representation is given in sec . vi . finally , in sec . vii we present the concluding remarks . in this section we introduce the model used in our studies . in addition , we discuss the concept of rotated electron and summarize some basic information about the pseudoparticle , holon , and spinon description which is useful for the studies of this paper . in a chemical potential @xmath21 and magnetic field @xmath22 the 1d hubbard hamiltonian can be written as , @xmath23 where the hamiltonian , @xmath24 has @xmath25 symmetry @xcite and @xmath26 is the simple hubbard model . the operators @xmath27 and @xmath28 on the right - hand side of eq . ( [ hh ] ) are the kinetic energy operator and the electron double occupation operator respectively . the operator @xmath29 on the right - hand side of eq . ( [ dop ] ) counts the number on spin @xmath1 electrons at real - space lattice site @xmath30 , where the number of lattice sites @xmath5 is even and large , @xmath31 is odd , and we consider periodic boundary conditions . the associated spin @xmath1 electron number operator reads , @xmath32 the operators @xmath33 and @xmath34 which appear in the above equations are the spin @xmath35 electron creation and annihilation operators at site @xmath36 respectively . moreover , on the right - hand side of eq . ( [ h ] ) , @xmath37 , @xmath38 , @xmath39 is the bohr magneton , and @xmath40 $ ] and @xmath41 $ ] are the diagonal generators of the @xmath12 @xmath8-spin @xmath42 and spin @xmath43 algebras @xcite respectively . in the latter expressions @xmath44 is the electron number operator and the operator @xmath45 counts the number of spin @xmath1 electrons and is given in eq . ( [ nsi ] ) . we denote by @xmath46 and @xmath47 the number of spin - up electrons and spin - down electrons respectively , and by @xmath48 the number of electrons . the lattice constant is denoted by @xmath3 and thus the length of the system is @xmath49 . we consider electronic densities and the spin densities given by @xmath50 and @xmath51 respectively , where @xmath52 and @xmath53 . these densities and spin densities belong to the domains defined by the following inequalities @xmath54 ; @xmath55 and @xmath56 ; @xmath57 respectively . the hamiltonian @xmath58 defined in eq . ( [ hso4 ] ) commutes with the six generators of the @xmath8-spin @xmath59 and spin @xmath60 algebras @xcite . while the expressions of the diagonal generators were provided above , the off - diagonal generators of these two algebras read , @xmath61 and @xmath62 respectively . we note that the bethe - ansatz solution of the 1d hubbard model refers to the hilbert subspace spanned by the lowest - weight states ( lwss ) of the @xmath8-spin and spin algebras , _ i.e. _ such that @xmath63 @xcite . finally , the momentum operator is given by , @xmath64 \ , , \label{popel}\ ] ] and commutes with the hamiltonians introduced in eqs . ( [ h ] ) and ( [ hso4 ] ) . rotated electrons are associated with a unitary transformation first introduced by harris and lange @xcite . such a transformation plays a key role in the expression of the holon and spinon number operators in terms of electronic operators for the whole parameter space of the model @xcite . the electron operators that occur in the 1d hubbard model ( [ h ] ) and ( [ hh ] ) are defined by @xmath65 , while the rotated electron operator @xmath66 is given by , @xmath67 where @xmath68 is the harris and lange unitary operator . as mentioned in sec . i , such rotated electrons conserve double occupation for all finite values of @xmath14 . the operators @xmath69 and @xmath68 can be written as , @xmath70 these operators are uniquely defined by the following two equations , @xmath71 + { 1\over 2}\,[[{\hat{h}}_h,\,{\hat{s}}\,],\,{\hat{s}}\ , ] + ... \ , , \label{hhtil}\ ] ] and @xmath72 = [ { \hat{h}}_h,\,\tilde{d } ] = 0 \ , . \label{hhdtil}\ ] ] in these equations the hamiltonian @xmath73 is given in eq . ( [ hh ] ) and and the rotated - electron double occupation operator @xmath74 reads , @xmath75 where @xmath76 is the electron double occupation operator given in eq . ( [ dop ] ) . note that @xmath65 and @xmath66 are only identical in the @xmath16 limit where electron double occupation becomes a good quantum number . the expression of any rotated operator @xmath77 such that , @xmath78 in terms of the rotated electron operators @xmath66 and @xmath79 is the same as the expression of the corresponding general operator @xmath80 in terms of the electron operators @xmath65 and @xmath34 respectively . equations ( [ hhtil ] ) and ( [ hhdtil ] ) can be used to derive an expression for the unitary operator @xmath68 in terms of electronic elementary operators order by order in @xmath17 . the authors of ref . @xcite carried out this expansion up to eighth order ( see foot note [ 12 ] of that reference ) . operators which commute with the harris and langue unitary operator have the same expressions in terms of both electronic and rotated - electronic operators . examples of such operators are the six generators of the @xmath12 @xmath8-spin and spin algebras . thus it follows that the commutators @xmath81 $ ] and @xmath82 $ ] vanish , where @xmath83 and @xmath84 are the square @xmath8-spin and spin operators respectively , and @xmath45 is the spin @xmath1 electron number operator ( [ nsi ] ) . the harris and langue unitary operator also commutes with the momentum operator ( [ popel ] ) . let us summarized the aspects of the holon , spinon , and band - momentum pseudoparticle description introduced in ref . @xcite and further investigated in refs . @xcite which are useful for the studies of this paper . such a description is extracted directly from combination of the bethe - ansatz takahashi s thermodynamic equations with the @xmath25 symmetry of the model and the symmetries associated with the concept of rotated electron . indeed , holons , spinons , and band - momentum pseudoparticles emerge naturally from the non - perturbative diagonalization of the model . the first step of such a diagonalization is precisely the hilbert - space unitary rotation which maps electrons onto rotated electrons and is such that rotated - electron double occupation is a good quantum number . let us denote by @xmath85 the number of rotated - electron singly occupied sites . thus the numbers @xmath85 , @xmath86/2 $ ] , and @xmath87/2 $ ] equal the number of rotated - electron singly occupied sites , doubly occupied sites , and empty sites respectively . ( here @xmath88 $ ] . ) importantly , once rotated electron double occupation is a good quantum number also all these numbers are good quantum numbers for all values of @xmath14 . there are precisely @xmath85 @xmath19 pseudoparticles and @xmath85 spinons . these @xmath19 pseudoparticles ( and spinons ) describe the charge part ( and spin part ) of the @xmath85 rotated electrons which singly occupy lattice sites . thus the charge and spin degrees of freedom of these @xmath85 rotated electrons separate . for each @xmath19 pseudoparticle there is a chargeon and a rotated - electronic hole . the chargeon corresponds to the charge part of the rotated electron which singly occupies the site . the non - perturbative organization of the electronic degrees of freedom associates the @xmath86 $ ] rotated electrons which doubly occupy @xmath86/2 $ ] sites with @xmath86/2 $ ] holons of @xmath8-spin projection @xmath9 . indeed each rotated - electron doubly occupied site corresponds to a spin singlet rotated electron pair which is nothing but a @xmath9 holon . finally , the @xmath87/2 $ ] rotated - electron empty sites are nothing but the @xmath87/2 $ ] holons of @xmath8-spin projection @xmath89 . such an analysis confirms the importance played by the unitary operator @xmath90 in the description of the quantum objects whose occupancy configurations describe all energy eigenstates of the model . let us summarize some of the properties of the quantum objects which we have briefly related to rotated - electrons above . one should distinguish the total @xmath8 spin ( and spin ) of the system , which we denote by @xmath59 ( and @xmath60 ) and the corresponding @xmath8-spin ( and spin ) projection , which we denote by @xmath91 ( and @xmath92 ) from the @xmath8 spin ( and spin ) carried by the elementary quantum objects . we call @xmath93 ( and @xmath94 ) the @xmath8 spin ( and spin ) carried by the holons , spinons , and pseudoparticles and @xmath95 ( and @xmath96 ) their @xmath8-spin ( and spin ) projection . the holons are such that @xmath97 , @xmath98 , and @xmath99 whereas the spinons have @xmath100 and @xmath101 . through out this paper we denote the holons and spinons according to their @xmath102 @xmath8-spin and spin projections respectively . the @xmath9 and @xmath89 holons have charge @xmath10 and @xmath13 respectively . it is shown in later sections that the band - momentum @xmath19 pseudoparticle is associated with the local @xmath19 pseudoparticle . the latter local quantum object describes the chargeon and the corresponding rotated electronic hole of each rotated - electron singly occupied site . there are also @xmath103-holon ( and @xmath103-spinon ) composite @xmath104 pseudoparticles ( and @xmath105 pseudoparticles ) . the @xmath104 pseudoparticle has @xmath106 and @xmath98 ( and the @xmath105 pseudoparticle has @xmath98 and no charge degrees of freedom ) and contains an equal number @xmath107 of @xmath9 and @xmath89 @xmath8-spin projection holons ( and spin projection spinons ) . the @xmath19 pseudoparticle has @xmath106 and no spin degrees of freedom . the @xmath102 holons ( and @xmath102 spinons ) which are not part of @xmath103-holon composite @xmath104 pseudoparticles ( and @xmath103-spinon composite @xmath105 pseudoparticles ) are called @xmath108 yang holons ( and @xmath102 hl spinons ) . in the designations _ hl spinon _ and _ yang holon _ , hl stands for heilmann and lieb and yang refers to c. n. yang respectively , who are the authors of refs . @xcite . all energy eigenstates can be specified by the @xmath19 pseudoparticle band - momentum distribution function @xmath109 , by the set of @xmath20 pseudoparticle band - momentum distribution functions @xmath110 where @xmath111 and @xmath112 , and by the set of numbers numbers @xmath113 of @xmath97 , @xmath98 , and @xmath9 yang holons ( @xmath114 ) and @xmath100 and @xmath9 hl spinons ( @xmath115 ) . an important concept is that of cphs ensemble subspace where @xmath116 stands for _ @xmath19 pseudoparticle , holon , and spinon_. this is a hilbert subspace spanned by all states with fixed values for the @xmath9 yang holon number @xmath117 , @xmath9 hl spinon number @xmath118 , @xmath19 pseudoparticle number @xmath85 , and for the sets of @xmath20 pseudoparticle numbers @xmath119 and @xmath120 corresponding to the @xmath121 branches . the variable @xmath122 appearing in the above distribution functions @xmath109 and @xmath110 where @xmath111 and @xmath112 is the continuum band momentum associated with the discrete band - momentum values @xmath123 which are of the following form , @xmath124 the @xmath125 numbers are integers ( half - odd integers ) , if @xmath126 is odd ( even ) and the @xmath127 numbers are integers ( half - odd integers ) , if @xmath128 is odd ( even ) . the discrete index @xmath36 can have the following values @xmath30 and @xmath129 for the @xmath19 and @xmath18 pseudoparticles respectively , where the number of lattice sites @xmath5 and the number @xmath130 where @xmath131 give the total number of corresponding discrete band - momentum values @xmath123 . on the right - hand side of eq . ( [ n*sum ] ) @xmath132 and @xmath133 are the number of @xmath20 pseudoparticles and @xmath20 pseudoparticle holes respectively . on the right - hand side of eq . ( [ nhag ] ) @xmath134 is the number of yang holons ( @xmath114 ) and hl spinons ( @xmath115 ) such that @xmath135 where @xmath136 is the @xmath8-spin ( @xmath114 ) and spin ( @xmath137 ) respectively @xcite . on the other hand , the number of @xmath19 pseudoparticle holes is @xmath138 $ ] , as given below in eq . ( [ mcs ] ) . the pseudoparticles obey a pauli principle respecting the band - momentum occupancies , _ i.e. _ a discrete band - momentum value @xmath123 can either be unoccupied or singly occupied by a pseudoparticle . it follows from eq . ( [ qj ] ) that the discrete band - momentum values @xmath123 are such that , @xmath139 the band - momentum distribution functions @xmath140 and @xmath110 where @xmath111 and @xmath112 read @xmath141 and @xmath142 for occupied values of the discrete band momentum @xmath123 and @xmath143 and @xmath144 for unoccupied values of that band momentum . the limiting band - momentum value which defines the continuum band - momentum @xmath122 range @xmath145 of the @xmath20 pseudoparticle branch is given by , @xmath146 \approx { \pi n^*_{\alpha , \,\nu}\over l } \ , . \label{qag}\ ] ] the @xmath19-pseudoparticle continuum band - momentum limiting values read @xmath147 for @xmath148 even and @xmath149 for @xmath148 odd . here @xmath150 \approx { \pi\over a } \ , . \label{qc}\ ] ] the numbers @xmath151 of @xmath95 yang holons ( @xmath114 ) and of @xmath96 hl spinons ( @xmath115 ) such that @xmath152 are given by , @xmath153 where the values @xmath136 of @xmath8 spin ( @xmath114 ) or spin ( @xmath115 ) of the system can be written as , @xmath154 = { 1\over 2}m_{\alpha } - \sum_{\nu = 1}^{\infty}\nu\,n_{\alpha , \,\nu } \ , ; \hspace{1 cm } \alpha = c , s \ , . \label{scs}\ ] ] here @xmath155 denotes the total number of holons @xmath156 and of spinons @xmath157 and @xmath158 gives the number of @xmath18 pseudoparticles . the @xmath8-spin ( @xmath114 ) and spin ( @xmath115 ) projections read , @xmath159 = -{1\over 2}[m_{c,\,+1/2}- m_{c,\,-1/2 } ] \ , ; \nonumber \\ s_s^z & = & -{1\over 2}[l_{s,\,+1/2}- l_{s,\,-1/2 } ] = -{1\over 2}[m_{s,\,+1/2}- m_{s,\,-1/2 } ] \ , , \label{szhs}\end{aligned}\ ] ] respectively , where @xmath160 and @xmath161 give the total number of @xmath95 holons and of @xmath96 spinons respectively . these numbers can be written as follows , @xmath162 the holon and spinon numbers @xmath156 and @xmath157 respectively , can be expressed as , @xmath163 \ , ; \hspace{1 cm } m_s = \sum_{\sigma_s = \pm 1/2}\,m_{s,\,\sigma_s } = n_c \ , , \label{mcs}\ ] ] where @xmath85 and @xmath164 are the numbers of @xmath19 pseudoparticles and @xmath19 pseudoparticle holes respectively . the above equations are valid for all energy eigenstates . let us now consider the particular case of a ground state . in the case of electronic densities @xmath165 and spin densities @xmath166 such that @xmath167 and @xmath168 respectively , a ground state belongs to a cphs ensemble subspace with the values for the pseudoparticle , @xmath102 holon , and @xmath102 spinon numbers given by @xmath169 , @xmath170 , @xmath171 , and @xmath172 in the charge sector . in the spin sector the numbers are @xmath173 , @xmath174 , @xmath175 , @xmath176 , and @xmath177 . in these expressions the electron numbers @xmath178 , @xmath179 , and @xmath180 refer to the ground state . in the case of the ground - state cphs ensemble subspace the expression of the number @xmath181 given in eqs . ( [ n*sum ] ) and ( [ nhag ] ) also simplifies @xcite and reads , @xmath182 whereas @xmath183 is given by @xmath184 for all energy eigenstates . it follows that also the expressions of the limiting band - momentum values ( [ qag])-([qc ] ) simplify in the case of the ground - state cphs ensemble subspace . disregarding corrections of the order of @xmath185 , in the case of the ground state these limiting band - momentum values can be written as @xmath186 and @xmath187 where , @xmath188 \ , ; \hspace{1 cm } q^0_{s,\,\nu } = [ k_{f\uparrow}-k_{f\downarrow } ] \ , , \hspace{0.25 cm } \nu > 1 \ , . \label{qcangs}\end{aligned}\ ] ] on the right - hand side of eq . ( [ qcangs ] ) the fermi momenta are such that @xmath189 , @xmath190 , and @xmath191 . in most situations one can disregard the @xmath185 corrections and use the band - momentum limiting values given in eq . ( [ qcangs ] ) . as mentioned in secs . i and ii , the energy eigenstates of the 1d hubbard model are described by the the same rotated - electron site distribution configurations for all values of @xmath14 . fortunately , since the electron - rotated electron unitary transformation becomes the unit transformation as @xmath16 , one can reach the @xmath14 independent rotated - electron site distribution configurations which describe the energy eigenstates by studying the corresponding electron site distribution configurations for the 1d hubbard model in the limit @xmath16 . thus the main goal of this section is to extract from the physics associated with the @xmath16 limit of the model useful information for the construction of the rotated - electron occupancy configurations that describe the energy eigenstates for finite values of @xmath14 . in the limit @xmath16 there is a huge degeneracy of @xmath8-spin and spin occupancy configurations . such a degeneracy results from the simple form that the kinetic energy @xmath192 and potential energy @xmath193 spectra associated with the operators ( [ top ] ) and ( [ dop ] ) respectively , have in the limit @xmath16 , @xmath194 \ , ; \hspace{1 cm } v / u = d\ , , \label{ehuinf}\ ] ] where @xmath195 is the energy spectrum of the hamiltonian ( [ hh ] ) and the electron double occupation @xmath196 is a good quantum number which equals rotated - electron double occupation in that limit . the kinetic energy equals that of a system of free spin - less fermions @xcite . the corresponding momentum spectrum is given in eq . ( [ ppuinf ] ) of appendix a. a short discussion of the relation of our results to some of the well known concepts of the @xmath16 physics is presented in that appendix . since the energy spectrum ( [ ehuinf ] ) is independent of the @xmath8-spin and spin occupancy configurations , there are several choices for complete sets of energy eigenstates with the same energy and momentum spectra . however , only one of these choices is associated with the rotated - electron site distribution configurations which describe the energy eigenstates for all values of @xmath14 . the discussion presented in appendix a and in this section about the relation of the energy - eigenstate rotated - electron description introduced in this paper to other choices of energy eigenstates and to the @xmath197 physics associated with the leading - order @xmath17 corrections to the energy spectrum ( [ ehuinf ] ) contextualizes the quantum problem studied in this paper and contributes to its deeper understanding . the pseudoparticle band momentum obeys well defined boundary conditions which are a necessary condition for the fulfilment of the periodic boundary conditions for the original electrons . however , such a pseudoparticle band - momentum boundary conditions are not a sufficient condition to ensure the electronic periodic boundary conditions . a second condition imposes that the internal structure of the local @xmath18 pseudoparticles introduced in later sections must be of a specific form . in this section we introduce a set of properties which are rather useful both for the construction of the specific rotated - electron site distribution configurations which describe the internal structure of such local @xmath18 pseudoparticles and the definition of the corresponding pseudoparticle _ empty sites _ in terms of occupancy configurations of rotated - electron sites . as further discussed in sec . vi , the construction of the rotated - electron site distribution configurations which describe the energy eigenstates and the introduction of the associated concepts of local pseudoparticle and effective pseudoparticle lattice are necessary steps for the evaluation of few - electron spectral functions at finite values of excitation energy . our choice of energy eigenstates corresponds to the band - momentum energy eigenstates associated with the thermodynamic bethe - ansatz equations introduced by takahashi @xcite . as discussed in appendix a , for @xmath16 there are other choices for complete sets of energy eigenstates . one of the alternative complete set of energy and momentum eigenstates for the model in the limit @xmath16 can be explicitly constructed for the harris - lange model mentioned in appendix a by the method used in ref . both the band - momentum energy eigenstates and the _ symmetrized energy eigenstates _ used in that reference are superpositions of charge ( and spin ) sequences formed by local electron distribution configurations of doubly occupied and empty sites ( and spin - down and spin - up singly occupied sites ) . the expression of both these two sets of energy eigenstates ensures the periodic boundary conditions for the original electronic problem . through out this paper we denote the rotated - electron doubly occupied and empty sites by @xmath198 and @xmath199 ( and the spin - down and spin - up rotated - electron singly occupied sites by @xmath200 and @xmath201 ) respectively . in the present @xmath16 limit such a concept also refers to electrons . often we add an index to these symbols which defines the position of the doubly occupied site or empty site ( and spin - down singly occupied site or spin - up singly occupied site ) . in the case of the _ symmetrized _ energy eigenstates the charge and spin sequences are properly symmetrized owing to the periodic boundary conditions of the original problem . ( this justifies the designation of _ symmetrized energy eigenstates_. ) the procedure used in such a symmetrization involves powers of suitable charge and spin operators @xmath202 and @xmath203 respectively , for cyclic permutation of the electron site distribution configurations of the local charge and spin sequences @xcite . for instance , let @xmath204 and @xmath205 be a charge and a spin sequence respectively . then the operators @xmath202 and @xmath203 are such that , @xmath206 and @xmath207 respectively . the powers @xmath208^{k_c}$ ] and @xmath209^{k_s}$ ] were called in ref . @xcite @xmath210 and @xmath211 respectively . this introduces the _ charge momentum _ @xmath212 and _ spin momentum _ @xmath213 respectively , such that , @xmath214 the symmetrized energy eigenstates are classified according to their charge and spin sequence , their charge momentum @xmath212 and spin momentum @xmath213 , and a number @xmath215 of momenta which in the notation of ref . @xcite equals the number of charges . the latter discrete momenta are closely related to the discrete band - momentum values occupied by @xmath19 pseudoparticle holes in the pseudoparticle representation of the band - momentum energy eigenstates studied in later sections . moreover , the number @xmath215 of charges equals both the number @xmath164 of @xmath19 pseudoparticle holes and the number @xmath156 of holons . thus this is a good quantum number for both the band - momentum and symmetrized energy eigenstates . however , the charge momentum @xmath212 and spin momentum @xmath213 are eigenvalues of the charge momentum operator @xmath216 and spin momentum operator @xmath217 respectively , which in general do not commute with the set of operators @xmath218 of the band - momentum basis . thus in general the band - momentum energy eigenstates are not eigenstates of the charge momentum operator @xmath216 and spin momentum operator @xmath217 . exceptions are the band - momentum energy eigenstates with occupancy of a single @xmath104 pseudoparticle ( and a single @xmath105 pseudoparticle ) and with no finite occupancy of yang holons ( and hl spinons ) and of @xmath219 pseudoparticles ( and @xmath220 pseudoparticles ) belonging to other branches such that @xmath221 . we find below that in this case the corresponding band - momentum energy eigenstate is also an eigenstate of the charge ( and spin ) momentum operator of eigenvalue @xmath222 ( and @xmath223 ) . according to eq . ( [ qj1j ] ) the discrete values of the band momentum @xmath123 are such that @xmath224 and @xmath225\,i^c_j$ ] or @xmath225\,i^{\alpha,\,\nu}_j$ ] where the numbers @xmath226 and @xmath227 with @xmath228 and @xmath229 respectively , are integers or half - odd integers as a result of the following boundary conditions , @xmath230 } \ , , \label{pbccp}\ ] ] in the case of the @xmath19 pseudoparticle branch and , @xmath231}=(e^{i\pi})^{[1+l_{\alpha}+ n_{\alpha,\,\nu}]}=(e^{i\pi})^{[1+n_c + n_{\alpha,\,\nu } ] } \ , ; \hspace{1 cm } \alpha = c , s \ , , \hspace{0.5 cm } \nu = 1,2 , ... \ , , \label{pbcanp}\ ] ] in the case of the @xmath18 pseudoparticle branches . here @xmath158 and @xmath134 denote the number of @xmath232 pseudoparticles and of yang holons ( @xmath114 ) or hl spinons ( @xmath115 ) respectively , and the number @xmath181 is given in eq . ( [ n*sum ] ) . according to eq . ( [ qag ] ) , in the case of the @xmath18 pseudoparticle branches the @xmath233 minimum and @xmath234 maximum index of the band - momentum values @xmath123 are such that , @xmath235 \ , . \label{limits}\ ] ] in the case of the @xmath19 pseudoparticles the limiting values are @xmath236 and @xmath237 where the band momenta @xmath238 are defined in eqs . ( [ qcev])-([qc ] ) . there is a holon , spinon , @xmath19 pseudoparticle separation for the whole parameter space of the 1d hubbard model @xcite . in the case of the @xmath239 limit the description of the @xmath19 pseudoparticle excitation sector is very similar for both the representations in terms of band - momentum energy eigenstates and of symmetrized energy eigenstates . the separation of the charge and spin excitation sectors occurs for these two alternative representations . the construction of the band - momentum energy eigenstates also involves superpositions of charge ( and spin ) sequences associated with rotated - electron distribution configurations of doubly occupied and empty sites ( and spin - down and spin - up singly occupied sites ) . such superpositions are also due to the periodic boundary conditions but are not in general generated by the above symmetrization procedure . in spite of these general similarities , the symmetrized and band - momentum energy eigenstates of the 1d hubbard model in the limit of @xmath239 are different states . the difference between the symmetrized and band - momentum energy eigenstates refers in general both to the form of the local electron distribution configurations of doubly occupied and empty sites ( and spin - down and spin - up singly occupied sites ) which describe the charge ( and spin ) sequences and to the form of the superposition of these local sequences which describes the energy eigenstates . let us introduce eight properties which correspond to a first step of the introduction of the concepts of local pseudoparticle and effective pseudoparticle lattices . however , the precise definition of these concepts involves clarification of several issues and is only fulfilled in sec . v. these properties are used in the ensuing section in finding the electron site distribution configurations of the local charge and spin sequences whose fourier - transform superpositions describe the band - momentum energy eigenstates of the model in the limit @xmath239 . these properties follow in part from symmetries and features of the pseudoparticle , holon , and spin description and related rotated electron representation studied in refs . @xcite and from well known properties associated with the @xmath239 physics . we emphasize that the properties given below also apply to finite values of @xmath17 provided that electrons are replaced by rotated electrons . these useful properties read : 1-iii the numbers of electron doubly occupied sites , empty sites , spin - down singly occupied sites , and spin - up singly occupied sites are good quantum numbers whose values are equal to the total numbers of @xmath9 holons , @xmath89 holons , @xmath9 spinons , and @xmath89 spinons respectively , of the band - momentum energy eigenstates . 2-iii the kinetic energy @xmath192 given in eq . ( [ ehuinf ] ) arises from the movements of the singly occupied sites relative to the doubly occupied and empty sides which do not change double occupation . these movements are fully described by the @xmath19 pseudoparticles which are associated with the charge degrees of freedom of these sites . in the limit @xmath239 these quantum objects acquire a spin - less fermion spectrum . on the other hand , the electron distribution configurations of doubly occupied and empty sites and of spin - down and spin - up singly occupied sites do not contribute to the kinetic energy @xmath192 . moreover these electron site distribution configurations must remain unchanged in spite of the movements of the @xmath19 pseudoparticles . alternatively , we can consider that the electron doubly occupied and empty sites move relative to the singly occupied sites . in this case one describes the movements of the @xmath19 pseudoparticles in terms of the movements of the corresponding @xmath19 pseudoparticle holes . this is the choice of ref . @xcite for the case of the symmetrized energy eigenstates . we recall that the numbers @xmath85 of @xmath19 pseudoparticles , @xmath164 of @xmath19 pseudoparticle holes , @xmath157 of spinons , and @xmath156 of holons are such that @xmath240 , @xmath241 , and @xmath242 and thus these two alternative descriptions are fully equivalent . 3-iii we call _ local charge sequences _ and _ local spin sequences _ the occupancy configurations of the @xmath102 holons ( and corresponding electron distribution configurations of doubly occupied and empty sites ) and the occupancy configurations of the @xmath102 spinons ( and corresponding electron distribution configurations of singly occupied sites of spin projection @xmath108 ) respectively . these local charge ( and spin ) sequences can also be expressed in terms of occupancy configurations of @xmath108 yang holons and local @xmath104 pseudoparticles ( and @xmath102 hl spinons and local @xmath105 pseudoparticles ) belonging to @xmath243 branches . from the general properties introduced in ref . @xcite one finds that the electron description of these quantum objects in terms of distribution configurations of electron doubly occupied and empty sites ( and spin - down and spin - up singly occupied sites ) is as follows : a @xmath9 yang holon ( and @xmath9 hl spinon ) is described by a doubly occupied site ( and a spin - down singly occupied site ) ; a @xmath89 yang holon ( and @xmath89 hl spinon ) is described by an empty site ( and a spin - up singly occupied site ) ; a local @xmath104 pseudoparticle ( and a local @xmath105 pseudoparticle ) is described by a number @xmath107 of doubly occupied sites and a number @xmath107 of empty sites ( and a number @xmath107 of spin - down singly occupied sites and a number @xmath107 of spin - up singly occupied sites ) . the electron distribution configurations of doubly occupied and empty sites ( and of spin - down and spin - up singly occupied sites ) of any local charge ( and spin ) sequence can be expressed in terms of a corresponding occupancy configuration of @xmath102 yang holons and local @xmath104 pseudoparticles ( and @xmath102 hl spinons and local @xmath105 pseudoparticles ) belonging to the @xmath243 branches . the form of eqs . ( [ pp ] ) and ( [ ppuinf ] ) of appendix a reveals that a @xmath9 holon carries momentum @xmath244 and that in the limit @xmath239 each doubly occupied site must also be associated with a momentum @xmath244 respectively . therefore , the expressions of the local charge sequences must include a phase factor operator @xmath245 such that the @xmath36 summation runs over the sites doubly occupied and empty and the local double occupation operator has eigenvalues @xmath246 and @xmath247 if the site @xmath36 is doubly occupied and empty respectively . 4-iii the number of different possible discrete locations of the electron distribution configurations of doubly occupied and empty sites ( @xmath114 ) or of spin - down and spin - up singly occupied sites ( @xmath248 describing a local @xmath20 pseudoparticle equals the number @xmath181 given in eq . ( [ n*sum ] ) . each of these possible positions corresponds to a different local charge ( @xmath114 ) or spin ( @xmath115 ) sequence . the number @xmath181 is directly provided by the bethe - ansatz solution @xcite since it also equals the number of different discrete band momentum values @xmath123 , where @xmath229 , of the band - momentum @xmath20 pseudoparticle band . 5-iii in case of energy eigenstates with finite occupancy of @xmath9 and @xmath89 yang holons ( and @xmath9 and @xmath89 hl spinons ) the electron distribution configurations of doubly occupied and empty sites ( and spin - down and spin - up singly occupied sites ) describing the local @xmath104 pseudoparticle ( and local @xmath105 pseudoparticle ) must remain unchanged under the application of the off - diagonal generators of the @xmath12 @xmath8-spin algebra given in eq . ( [ sc ] ) ( and off - diagonal generators of the @xmath12 spin algebra given in eq . ( [ ss ] ) ) . moreover , in case of states with no @xmath89 or @xmath9 yang holons ( and no @xmath89 or @xmath9 hl spinons ) application of the operators @xmath249 or @xmath250 ( and @xmath251 or @xmath252 ) onto these states must give zero . these requirements result from the value of the @xmath8 spin of the @xmath104 pseudoparticles ( and spin of the @xmath105 pseudoparticles ) which is given by @xmath106 ( and @xmath98 ) @xcite . on the other hand , the transformations generated by application of these off - diagonal generators onto the electron distribution configurations of doubly occupied and empty sites ( and spin - down and spin - up singly occupied sites ) describing the @xmath102 yang holons ( and @xmath102 hl spinons ) must be the ones defined by the @xmath8-spin ( and spin ) algebra @xcite . 6-iii let us consider a local charge ( and spin ) sequence with no yang holons ( and no hl spinons ) and consisting of @xmath107 electron doubly occupied sites and @xmath107 electron empty sites ( and @xmath107 electron spin - down singly occupied sites and @xmath107 electron spin - up singly occupied sites ) . if such a sequence describes a single local @xmath104 pseudoparticle ( and local @xmath105 pseudoparticle ) it is properly symmetrized in such way that the distribution configurations associated with the internal structure of that quantum object remain unchanged under cyclic permutations . furthermore , property 5-iii imposes that a band - momentum energy eigenstate with the above yang holon and @xmath104 pseudoparticle ( and hl spinon and @xmath105 pseudoparticle ) numbers is an eigenstate of the charge ( and spin ) momentum operator @xmath216 ( and @xmath217 ) of eigenvalue @xmath222 ( and @xmath223 ) . 7-iii the electron distribution configurations of the @xmath107 doubly occupied sites and @xmath107 empty sites ( and @xmath107 spin - down singly occupied sites and @xmath107 spin - up singly occupied sites ) which describe the internal structure a local @xmath104 pseudoparticle ( and @xmath105 pseudoparticle ) should be the same for all local charge ( @xmath114 ) or spin ( @xmath115 ) sequences involved in the description of the @xmath253 band - momentum energy eigenstates . this is a necessary condition for the indiscernible character of local @xmath18 pseudoparticles with the same lattice position but involved in occupancy configurations describing different local charge ( @xmath114 ) or spin ( @xmath115 ) sequences . this indiscernible character of the local @xmath18 pseudoparticles follows from the corresponding indiscernible character of the @xmath18 pseudoparticles of band - momentum @xmath123 , which are indistinguishable quantum objects . 8-iii since in the limit @xmath239 there is nearest - neighbor hopping only and it does not change double occupation , the charge , spin , @xmath19 pseudoparticle separation studied in ref . @xcite implies that both the local charge and spin sequences of the energy eigenstates must be separately conserved . moreover , the periodic boundary conditions of the original electronic problem are ensured if both the pseudoparticle band - momentum discrete values obey eqs . ( [ pbccp ] ) and ( [ pbcanp ] ) and the requirements of the basic properties 6-iii and 7-iii are fulfilled . the band - momentum energy eigenstates are fourier transform superpositions of local charge sequences , spin sequences , and @xmath19 pseudoparticle sequences corresponding to slater determinants involving the pseudoparticle band momentum @xmath123 and the spatial coordinate @xmath254 of these quantum objects . both in the case of the @xmath19 pseudoparticle branch and of the @xmath18 pseudoparticle branches with finite occupancy in a given state , such spatial coordinate is the conjugate of the band momentum @xmath123 of the above fourier transforms . the spatial coordinate @xmath254 of the @xmath19 pseudoparticles ( and @xmath18 pseudoparticles ) is associated with an effective pseudoparticle lattice of length @xmath255 , number of lattice sites @xmath5 ( and @xmath181 given in eq . ( [ n*sum ] ) ) and lattice constant @xmath3 ( and @xmath256 ) . the possible values of these spatial coordinates are @xmath257 where @xmath30 ( and @xmath258 where @xmath259 ) . the effective pseudoparticle lattices arise because the occupancy configurations of the different pseudoparticle branches are separately conserved . these eight basic properties play an important role in the mechanisms which in the @xmath239 limit determine the choice of the electron site distribution configurations of the local charge , spin , and @xmath19 pseudoparticle sequences whose fourier transform superpositions describe the band - momentum energy eigenstates . property 1-iii follows from well known properties of the 1d hubbard model in the limit @xmath239 . property 2-iii is consistent with the finding of ref . @xcite that in the limit @xmath239 only the @xmath19 pseudoparticles move and _ carry _ kinetic energy , whereas the @xmath102 holons and @xmath108 spinons correspond to unchanged occupancy configurations in that limit . such a property results from well established features of the bethe - ansatz solution @xcite . property 3-iii is a consequence of the combination of property 1-iii with the relation of yang holons , hl spinons , and @xmath20 pseudoparticles to @xmath102 holons and @xmath102 spinons . in property 4-iii the number @xmath181 given in eq . ( [ n*sum ] ) plays a central role . the expression of that number is valid for all values of @xmath14 and is provided by the bethe ansatz solution . this is consistent with the electron site distribution configurations which describe the local @xmath20 pseudoparticles in the limit of @xmath239 being the same as the rotated - electron configurations which describe these local @xmath20 pseudoparticles for finite values of @xmath17 . property 5-iii follows from general symmetries of the 1d hubbard model which are also valid for all values of @xmath14 . this basic property is also consistent with the above equivalence of the electron site distribution configurations in the limit of @xmath239 and of the rotated - electron configurations for finite values of @xmath17 . property 6-iii is a consequence of the periodic boundary conditions in the particular case when a local charge or spin sequence corresponds to a single @xmath104 or @xmath105 pseudoparticle respectively . property 7-iii results from the indiscernible character of the @xmath18 pseudoparticles . in the case of band - momentum pseudoparticles such an indiscernible character is implicit in the description of the band - momentum energy eigenstates in terms of pseudoparticle occupancy configurations @xcite . finally , property 8-iii is related to the periodic boundary conditions of the original electronic problem and to the existence in this limit of nearest - neighbor hopping only which does not change double occupation . such a property is also related to the charge , spin , @xmath19 pseudoparticle separation studied in ref . @xcite for all energy scales and for the whole parameter space of the 1d hubbard model . in addition , in the case of the @xmath18 pseudoparticle branches , we find below that the occurrence of the effective pseudoparticle lattices mentioned in this property is related to the separation of the internal and translational degrees of freedom of the local @xmath18 pseudoparticles . since in the @xmath239 limit the rotated electrons become the electrons , often below we refer to the rotated - electron site distribution configurations only . however , whenever referring to rotated electrons we mean implicitly that in the limit @xmath239 the electron site distribution configurations which describe the band - momentum energy eigenstates are the same . in this section we find the rotated - electron site distribution configurations which describe the local charge , spin , and @xmath19 pseudoparticle sequences . this requires the study of the internal structure of the local @xmath18 pseudoparticles . in addition , such an analysis requires the definition of the positions in the effective electronic lattice of the rotated - electron site distribution configurations which describe both the local @xmath19 and @xmath18 pseudoparticles and their _ empty sites_. we find that each specific site distribution configurations of a local charge , spin , and @xmath19 pseudoparticle sequence defines a local state . the set of all different possible such states constitutes a complete set of local states . in the ensuing section we express the energy eigenstates of the model as a fourier - transform superposition of these local states . such a study reveals the connection of the energy eigenstates to the rotated electron distribution configurations of doubly occupied sites , empty sites , spin - down singly occupied sites , and spin - up singly occupied sites which describe the local states . in the following we use in part the basic properties 1-iii to 8-iii to study the rotated - electron site distribution configurations which describe the local pseudoparticles introduced in this paper . we start by restricting our study to rotated - electron site distribution configurations which describe the lowest - weight states of both the @xmath12 @xmath8-spin and spin algebras . these states have no @xmath9 yang holons and no @xmath9 hl spinons . thus in this case the number of yang holons ( and of hl spinons ) is such that @xmath260 ( and @xmath261 ) . in the limit @xmath16 the @xmath85 @xmath19 pseudoparticles behave as spin - less fermions in a lattice of @xmath262 sites . such a lattice is nothing but the effective @xmath19 pseudoparticle lattice mentioned in property 8-iii . as in the case of the band - momentum @xmath123 , such an effective @xmath19 pseudoparticle lattice and its site coordinates @xmath254 remain the same for the whole parameter space of the model . also the @xmath19 pseudoparticle occupancy configurations of such a lattice describing a given energy eigenstate are the same for all values of @xmath14 , density @xmath165 , and spin density @xmath166 . at a fixed value @xmath263 of the numbers @xmath85 of @xmath19 pseudoparticles and @xmath157 of spinons the number of occupancy configurations of the local @xmath19 pseudoparticles in such an effective lattice is given by , @xmath264 each of these @xmath19 pseudoparticle occupancy configurations of the effective @xmath19 pseudoparticle lattice defines a _ local @xmath19 pseudoparticle sequence_. let @xmath265 where @xmath266 be the actual set of occupied coordinates of the effective @xmath19 pseudoparticle lattice out of the available @xmath257 coordinate sites where @xmath267 . we note that these @xmath5 sites have the same coordinates and lattice constant @xmath3 as the @xmath5 sites of the effective electronic lattice . moreover , the set of coordinates @xmath265 where @xmath266 corresponding to the occupied sites of the effective @xmath19 pseudoparticle lattice are precisely the same as the coordinates of the rotated electron singly occupied sites . thus we can define each of the occupancy configurations whose total number is given in eq . ( [ ncnhc ] ) by the coordinates in units of the lattice constant @xmath3 of the rotated - electron singly occupied sites , @xmath268 the positions occupied by the rotated - electron doubly occupied and empty sites are then these left over by the rotated - electron singly occupied sites . since the sequence ( [ lcp ] ) also gives the positions of the local @xmath19 pseudoparticles in their effective lattice , the sites left over by these pseudoparticles define the positions of the local @xmath19 pseudoparticle holes . thus the relation of the numbers @xmath85 and @xmath269 of local @xmath19 pseudoparticles and local @xmath19 pseudoparticle holes respectively , to the rotated - electron site distribution configurations of a given energy eigenstate confirms that there are @xmath85 rotated - electron singly occupied sites and @xmath138 $ ] rotated - electron doubly - occupied and empty sites , as mentioned in sec . ii . from the use of eqs . ( [ mas ] ) and ( [ mcs ] ) we find that these numbers can be expressed as follows , @xmath270 the local charge sequences ( and spin sequences ) introduced in property 3-iii involve only the rotated - electron doubly - occupied and empty site distribution configurations ( and spin - down and spin - up rotated - electron singly occupied site distribution configurations ) of @xmath138 $ ] sites ( and @xmath85 sites ) out of a total number @xmath5 of sites . as a result of the independent conservation of the charge and spin sequences , the rotated - electron site distribution configurations of a charge ( and spin ) sequence is for any energy eigenstate obtained simply by omitting the @xmath85 singly occupied sites ( and @xmath138 $ ] doubly - occupied and empty sites ) . thus a local charge ( and spin ) sequence corresponds to the rotated - electron occupancy configurations of a site domain of a number @xmath138 $ ] of sites ( and @xmath85 of sites ) . it follows that the number of sites of such a domain depends on the specific state under consideration . in the case of the ground state numbers provided in sec . ii , the number of sites of the effective electronic lattice which belong to the charge and spin sequences is @xmath271 $ ] and @xmath272 respectively . also the effective @xmath18 pseudoparticle lattices mentioned in property 8-iii and introduced below , result from independent conservation laws associated with each @xmath18 pseudoparticle branch such that @xmath273 and @xmath274 branch . these effective @xmath104 pseudoparticle ( and @xmath105 pseudoparticle ) lattices are generated below by omission of a number @xmath275 n_{c,\,\nu'}-n_{c,\,\nu}$ ] ( and @xmath275 n_{s,\,\nu'}-n_{s,\,\nu}$ ] ) of sites out of the total number @xmath138 $ ] ( and @xmath85 ) of sites of the local charge sequence ( and spin sequence ) . again the number of sites of the local charge ( and spin ) sequence which contribute to an effective @xmath104 pseudoparticle ( and @xmath105 pseudoparticle ) lattice depends on the specific state under consideration . for instance , in the case of a ground state the number of sites of the local charge sequence ( and spin sequence ) which contribute to an effective @xmath104 pseudoparticle ( and @xmath105 pseudoparticle ) lattice is @xmath271 $ ] ( and @xmath276 for the @xmath277 pseudoparticle branch and @xmath278 $ ] for the @xmath105 pseudoparticle branches such that @xmath279 ) out of the @xmath271 $ ] sites ( and @xmath280 $ ] sites ) of such a sequence . on the other hand , the number of sites of the effective @xmath19 pseudoparticle lattice equals the number @xmath5 of sites of the effective electronic lattice and is the same for all energy eigenstates . the expressions of eq . ( [ lcslws ] ) show that the value of the number @xmath281 of yang holons ( and @xmath282 of hl spinons ) of the local charge ( and spin ) sequence is uniquely determined by the values of the numbers of @xmath19 pseudoparticles and @xmath104 pseudoparticles ( @xmath19 pseudoparticles and @xmath105 pseudoparticles ) of the same sequence . thus for given number @xmath283 $ ] of rotated - electron doubly - occupied and empty sites ( and @xmath85 of rotated - electron singly occupied sites ) we can uniquely define the rotated - electron site distribution configurations of the local charge ( and spin ) sequence by providing the sites occupied by local @xmath104 pseudoparticles ( and @xmath105 pseudoparticles ) belonging to the branches @xmath274 with finite occupancy . the charge - sequence ( and spin - sequence ) sites of the effective electronic lattice left over by these local pseudoparticles define the positions of the yang holons ( and hl spinons ) . it is useful to introduce the charge - sequence site index @xmath284 and the spin - sequence site index @xmath285 such that , @xmath286 \ , ; \hspace{1 cm } l=1,2, ... ,n_c \ , . \label{hl}\ ] ] the ordering of the charge - sequence ( and spin - sequence ) index corresponds to the location order from the left to the right - hand side of the corresponding site in the effective electronic lattice . the position in the effective electronic lattice of a spin - sequence site of index @xmath285 and of a charge - sequence site of index @xmath284 are given by , @xmath287 \ , , \label{chcl}\ ] ] where @xmath288 are the indices of eq . ( [ lcp ] ) which define the location of the rotated - electron singly occupied sites and @xmath289 are the indices which define the location of the rotated - electron doubly occupied and empty sites . the location of the latter sites corresponds to the sites left over by the rotated - electron singly occupied sites . from both the above analysis and property 3-iii we conclude that one can uniquely specify a given rotated - electron site distribution configuration by providing the position of the @xmath85 sites occupied by local @xmath19 pseudoparticles , @xmath290 sites occupied by local @xmath104 pseudoparticles , and @xmath291 sites occupied by local @xmath105 pseudoparticles . the @xmath85 sites occupied by @xmath19 pseudoparticles define the location of the rotated - electron singly occupied sites . the @xmath292 $ ] rotated - electron doubly - occupied and empty sites are the sites left over by the rotated - electron singly occupied sites . the @xmath293 sites occupied by yang holons and the @xmath294 sites occupied by hl spinons are the sites left over in the charge and spin sequences respectively , by the sites occupied by local @xmath104 pseudoparticles and local @xmath105 pseudoparticles respectively . as mentioned above , to start with we consider that all yang holons ( and hl spinons ) have @xmath8-spin projection ( and spin projection ) @xmath89 and thus correspond to rotated - electron empty sites ( and rotated - electron spin - up singly occupied sites ) . the generalization to rotated - electron site distribution occupancies associated with states containing both @xmath102 yang holons ( and @xmath102 hl spinons ) is straightforward and is introduced later in this section . the yang holons ( and hl spinons ) have no internal structure and are the simplest of the quantum objects which occupy the charge ( and spin ) sequence of the effective electronic lattice . according to property 1-iii , the @xmath89 yang holons ( and @xmath89 hl spinons ) correspond to rotated - electron empty sites ( and spin - up rotated - electron singly occupied sites ) of these sequences . on the other hand , a local @xmath18 pseudoparticle has internal structure and thus is a more involved quantum object than a yang holon or a hl spinon . this internal structure corresponds to the rotated - electron distribution configurations of the @xmath107 doubly occupied sites and @xmath107 empty sites ( @xmath114 ) or @xmath107 spin - down singly occupied sites and @xmath107 spin - up singly occupied sites ( @xmath115 ) which according to property 3-iii describe such a local @xmath18 pseudoparticle . if in the rotated - electron site distribution configurations which describe the local @xmath104 pseudoparticles we replace doubly occupied sites @xmath198 and empty sites @xmath199 by spin - down singly occupied sites @xmath200 and spin - up singly occupied sites @xmath201 respectively , and omit the phase factors generate by the operator @xmath295 mentioned in property 3-iii , we obtain the corresponding distribution configurations of the @xmath105 pseudoparticles . therefore , often we consider the rotated - electron site distribution configurations which describe the local @xmath104 pseudoparticles only . it is useful to classify the @xmath103 lattice sites of the local charge sequence involved in the description of a local @xmath104 pseudoparticle into two sets of @xmath107 sites each . let us denote the index of these two sets of @xmath107 sites by @xmath296 and @xmath297 respectively . here @xmath298 and the index @xmath299 refers to the position @xmath300 $ ] of the local @xmath104 pseudoparticle in the effective electronic lattice where , @xmath301 and the numbers @xmath302 are the indices @xmath289 of eq . ( [ lcp ] ) which define the location of the rotated - electron doubly occupied / empty sites in such a lattice . combination of the values of the indices @xmath299 and @xmath298 fully defines the position of the above @xmath103 sites . the index @xmath303 is such that @xmath304 and @xmath305 respectively , where @xmath306 . equivalently , often we denote these @xmath103 indices simply by @xmath307 , where @xmath308 and @xmath309 . the set of @xmath103 indices @xmath310 is in general a sub - set of the @xmath292 $ ] charge - sequence indices @xmath284 given in eq . ( [ hl ] ) . the latter indices define the location of the charge - sequence rotated - electron doubly occupied and empty sites . we can also define the _ charge - sequence position _ of the local @xmath104 pseudoparticle which is defined as , @xmath311 the same definitions hold for the local @xmath105 pseudoparticles with the indices @xmath296 and @xmath297 replaced by the indices @xmath312 and @xmath313 respectively , and thus the equivalent indices @xmath307 replaced by @xmath314 . the position of the local @xmath105 pseudoparticle in the effective electronic lattice is defined as @xmath315 $ ] where , @xmath316 @xmath317 , and @xmath288 are the indices of eq . ( [ lcp ] ) which define the location of the rotated - electron singly occupied sites . on the other hand , the _ spin - sequence position _ of the local @xmath105 pseudoparticle is defined as , @xmath318 where again @xmath317 . we note that the indices @xmath319 ( and @xmath320 ) given in eq . ( [ bstringc ] ) ( and eq . ( [ bstrings ] ) ) which define the charge - sequence ( and spin - sequence ) position of the local @xmath104 pseudoparticle ( and local @xmath105 pseudoparticle ) are always positive integer numbers @xmath321 . according to eq . ( [ stringc ] ) ( and eq . ( [ strings ] ) ) the position @xmath300 $ ] ( and @xmath315 $ ] ) of the local @xmath104 pseudoparticle ( and local @xmath105 pseudoparticle ) in the @xmath5-site effective electronic lattice refers to a single point inside the @xmath103-site domain associated with such a quantum object . on the other hand , the charge - sequence position ( and spin - sequence position ) of the local @xmath104 pseudoparticles ( and local @xmath105 pseudoparticles ) defined by the index of eq . ( [ bstringc ] ) ( and eq . ( [ bstrings ] ) ) refers to the location of that quantum object relative to the @xmath292 $ ] sites ( and @xmath85 sites ) of the charge ( and spin ) sequence only . below we clarify the following two issues : first , we find the rotated - electron distribution occupancies of the @xmath103 sites which describe a local @xmath18 pseudoparticle ; second , we find the rotated - electron site distribution occupancies which define the @xmath322 charge - sequence ( @xmath323 ) or spin - sequence ( @xmath324 ) _ empty sites _ corresponding to the local @xmath18 pseudoparticle branch . this study reveals that the position in the effective electronic lattice of the sites associated with the @xmath322 charge - sequence or spin - sequence _ empty sites _ of each local @xmath18 pseudoparticle branch is uniquely determined by the position of the local @xmath19 pseudoparticles , local @xmath104 pseudoparticles , and local @xmath105 pseudoparticles . the same holds for the locations of the yang holons and hl spinons , as discussed above . below we also confirm that for fixed values of the numbers @xmath158 and @xmath181 the number of occupancy configurations of the local @xmath18 pseudoparticles is given by , @xmath325 we classify each of the local @xmath18 pseudoparticle occupancy configurations by providing the indices @xmath326 or @xmath327 given in eqs . ( [ stringc ] ) or ( [ strings ] ) respectively , corresponding to the @xmath328 and @xmath329 pseudoparticle locations in the effective electronic lattice , @xmath330 and @xmath331 respectively . from the combination of eqs . ( [ ncnhc ] ) and ( [ nanns ] ) it follows that the number of occupancy configurations of the local @xmath19 and @xmath18 pseudoparticles of a cphs ensemble subspace is given by , @xmath332 let us denote each local state representing a specific local charge , spin , and @xmath19 pseudoparticle sequence whose number for a given cphs ensemble subspace is given in eq . ( [ ncphs ] ) by , @xmath333 here @xmath334 gives the positions in the effective electronic lattice of the local @xmath104 pseudoparticles belonging to branches with finite occupancy in the state under consideration and @xmath335 provides the positions in the same lattice of the local @xmath105 pseudoparticles belonging to branches with finite occupancy in the same state . we emphasize that the number ( [ ncphs ] ) of such local states which equals the number of different local pseudoparticle occupancy configurations of a cphs ensemble subspace indeed equals the dimension of such a subspace . it is straightforward to confirm from the results of ref . @xcite that the number of band - momentum energy eigenstates that span such a cphs ensemble subspace also has the same value . this suggests that the set of local states of form ( [ localst ] ) is complete in that subspace , as is confirmed below . we find later on in this section that the generalization of the present analysis to rotated - electron site distribution configurations describing energy eigenstates with finite occupancies of @xmath102 yang holons and @xmath102 hl spinons requires the introduction of the numbers @xmath117 of @xmath9 yang holons and @xmath118 of @xmath9 hl spinons in the labelling of the local states ( [ localst ] ) . in order to find the rotated - electron site distribution configurations which describe the internal structure of a local @xmath18 pseudoparticle we use mainly the basic properties 3-iii , 5-iii , 6-iii , and 7-iii . often we consider the specific case of a local @xmath104 pseudoparticle . the internal structure of such a quantum object is studied by considering first a charge sequence constituted by @xmath107 rotated - electron doubly occupied sites and @xmath107 rotated - electron empty sites . the use of properties 5-iii and 6-iii leads to the finding of the specific rotated - electron site distribution configurations which describe the local @xmath104 pseudoparticle . property 7-iii then states that the obtained rotated - electron site distribution configurations describe a local @xmath104 pseudoparticle in any charge sequence . according to property 3-iii , the internal structure of a local @xmath104 pseudoparticle involves a number @xmath107 of rotated - electron doubly occupied sites and an equal number of rotated - electron empty sites . there is a number @xmath336 of different distribution configurations of these @xmath107 rotated - electron doubly occupied sites and @xmath107 rotated - electron empty sites . for simplicity let us restrict the present preliminary analysis to charge sequences with @xmath107 rotated - electron doubly occupied sites and an equal number of rotated - electron empty sites and without yang holons . below we find that a local @xmath104 pseudoparticle is a superposition of a number @xmath337 of the above rotated - electron site distribution configurations . for @xmath338 the local @xmath339 pseudoparticle is a properly symmetrized superposition of the two available rotated - electron site distribution configurations . for @xmath340 the local @xmath341 pseudoparticle is a properly symmetrized superposition of four out of the six available rotated - electron site distribution configurations . thus states including all possible six different distribution configurations are a superposition of two states including two local @xmath339 pseudoparticles and a local @xmath341 pseudoparticle respectively . for @xmath342 the local @xmath343 pseudoparticle is a properly symmetrized superposition of eight out of the twenty available rotated - electron site distribution configurations . states including all possible twenty different distribution configurations are a superposition of three states including ( i ) three local @xmath339 pseudoparticles , ( ii ) a local @xmath339 pseudoparticle and a local @xmath341 pseudoparticle , and ( iii ) a local @xmath343 pseudoparticle . in the general case one has that the local @xmath104 pseudoparticle is a properly symmetrized superposition of @xmath337 out of the @xmath336 available rotated - electron site distribution configurations . states including all possible @xmath336 different distribution configurations are a superposition of several states including different numbers @xmath344 of @xmath219 pseudoparticles such that @xmath345 and @xmath346 . among the @xmath337 rotated - electron site distribution configurations of a @xmath104 pseudoparticle there is always a distribution configuration where the first @xmath107 sites of index @xmath296 and @xmath298 are doubly occupied by rotated electrons and the last @xmath107 sites of index @xmath297 and @xmath298 are free of rotated electrons . we represent such a local @xmath104 pseudoparticle rotated - electron site distribution configuration of a charge sequence by , @xmath347 the same applies to a local @xmath105 pseudoparticle , the rotated - electron site distribution configuration of a spin sequence corresponding to the one given in eq . ( [ lcnp ] ) being , @xmath348 we confirm below that the @xmath104 pseudoparticle located in the effective electronic lattice at position @xmath300 $ ] and the @xmath105 pseudoparticle located in the same lattice at position @xmath315 $ ] are described by the following properly symmetrized superposition of @xmath337 rotated - electron site distribution configurations , @xmath349\ , \bigl[\prod_{g=1}^{\nu}(1-{\hat{\cal{t}}}_{c,\,\nu,\,j,\,g})\bigr ] \,(\bullet_{h_{j,\,1}}, ... ,\bullet_{h_{j,\,\nu } } , \,\circ_{h_{j,\,1+\nu } } , ... , \circ_{h_{j,\,2\nu } } ) \ , ; \hspace{0.5 cm } j=1,2, ... ,n_{c,\,\nu } \ , , \label{cnp}\ ] ] and @xmath350 \ , ( \downarrow_{l_{j,\,1 } } , ... , \downarrow_{l_{j,\,\nu}},\,\uparrow_{l_{j,\,1+\nu } } , ... , \uparrow_{l_{j,\,2\nu } } ) \ , ; \hspace{0.5 cm } j=1,2, ... ,n_{s,\,\nu } \ , , \label{snp}\ ] ] respectively . here and according to property 3-iii , the site-@xmath307 rotated - electron double occupation operator @xmath351 has eigenvalue @xmath246 and @xmath247 when that site is doubly occupied by rotated electrons and free of rotated electrons respectively , and the operator @xmath352 ( and @xmath353 ) acts on the pair of sites of indices @xmath296 and @xmath354 ( and @xmath312 and @xmath355 ) only . this operator always acts onto rotated - electron site distribution configurations of the particular form illustrated in eq . ( [ lcnp ] ) ( and in eq . ( [ lsnp ] ) ) . from the application of this operator onto such a rotated - electron site distribution configuration a new distribution configuration is generated where the site of index @xmath296 is free of rotated electrons , the site of index @xmath354 is doubly occupied by rotated electrons , and the occupancy of the other @xmath356 sites remains unchanged , _ @xmath357 the same transformation law , @xmath358 is associated with the application of the operator @xmath353 onto an rotated - electron site distribution configuration of the type illustrated in eq . ( [ lsnp ] ) . the @xmath337 internal rotated - electron site distribution configurations on the right - hand side of eq . ( [ cnp ] ) ( and eq . ( [ snp ] ) ) are generated by considering that in each of the @xmath298 pairs of sites of indices @xmath359 and @xmath354 , the site of index @xmath359 is doubly occupied by rotated electrons and the site of index @xmath354 is free of rotated electrons and vice versa . in general we call @xmath360 _ site pair _ a superposition of two rotated - electron distribution configurations of a rotated - electron doubly occupied site and a rotated - electron empty site where @xmath307 and @xmath361 , such that @xmath362 and @xmath363 , are the indices of the two lattice sites involved . in the first rotated - electron site distribution configuration the sites of indices @xmath307 and @xmath361 are doubly occupied by rotated electrons and free of rotated electrons respectively . in the second rotated - electron site distribution configuration the sites of indices @xmath307 and @xmath361 are free of rotated electrons and doubly occupied by rotated electrons respectively . the rotated - electron distribution configurations of all the remaining sites of the charge sequence except these two are identical . the @xmath364 site pair is defined as a superposition of these two rotated - electron site distributions configurations where the first and the second distributions are multiplied by a phase factor of @xmath246 and @xmath365 respectively . in figure 1 we represent a @xmath366 site pair by two vertical lines connected by a horizontal line . the two vertical lines have the same height and connect the corresponding site of the charge sequence to the horizontal line . thus the two sites of a charge sequence connected by three such lines are assumed to be involved in a superposition of two rotated - electron site distribution configurations multiplied by the phase factors of @xmath246 and @xmath365 respectively , where according to the above recipe one of these sites is doubly occupied by rotated electrons and the other one is free of rotated electrons and vice versa . similarly , we call @xmath367 site pair an equivalent superposition of two reotated - electron site distribution configurations of a spin sequence . the definitions are the same provided that we replace rotated - electron doubly occupied sites and rotated - electron empty sites by spin - down rotated - electron singly occupied sites and spin - up rotated - electron singly occupied sites respectively . such a rotated - electron singly occupied site pair is also graphically represented by the two vertical lines connected by a horizontal line plotted in fig . 1 . the concept of a site pair plays an important role in the description of the rotated - electron site distribution configurations of a local @xmath18 pseudoparticle . it is confirmed below that the rotated - electron site distribution configurations which describe a local @xmath18 pseudoparticle always involve a number @xmath107 of site pairs such that @xmath368 and @xmath369 where @xmath370 . since there are two possible rotated - electron site occupancies for each pair and the number of pairs of each local @xmath18 pseudoparticle is @xmath107 , the total number of different internal rotated - electron site distribution configurations is indeed @xmath337 . for example , in the case of a @xmath104 pseudoparticle located in the effective electronic lattice at the position @xmath300 $ ] , the @xmath337 rotated - electron site distribution configurations superposed in expression ( [ cnp ] ) are the following , @xmath371\,(1-{\hat{\cal{t}}}_{c,\,1,\,j,\,g } ) \,(\bullet_{h_{j,\,1}},\ , \circ_{h_{j,\,2 } } ) = - e^{i\pi h_{j,\,1}}\,(\bullet_{h_{j,\,1}},\ , \circ_{h_{j,\,2 } } ) + e^{i\pi h_{j,\,2}}\,(\circ_{h_{j,\,1}},\ , \bullet_{h_{j,\,2 } } ) \label{c1p}\ ] ] in the case of the @xmath339 pseudoparticle , @xmath372 \,\bigl[\prod_{g=1}^{2}(1-{\hat{\cal{t}}}_{c,\,2,\,j,\,g})\bigr ] \,(\bullet_{h_{j,\,1}},\,\bullet_{h_{j,\,2}},\ , \circ_{h_{j,\,3}},\,\circ_{h_{j,\,4 } } ) = \nonumber \\ & + & e^{i\pi [ h_{j,\,1}+h_{j,\,2}]}\ , ( \bullet_{h_{j,\,1}},\,\bullet_{h_{j,\,2}},\ , \circ_{h_{j,\,3}},\,\circ_{h_{j,\,4 } } ) - e^{i\pi[h_{j,\,2}+h_{j,\,3 } ] } \,(\circ_{h_{j,\,1}},\,\bullet_{h_{j,\,2}},\ , \bullet_{h_{j,\,3}},\,\circ_{h_{j,\,4 } } ) \nonumber \\ & + & e^{i\pi [ h_{j,\,3}+h_{j,\,4}]}\,(\circ_{h_{j,\,1}},\,\circ_{h_{j,\,2}},\ , \bullet_{h_{j,\,3}},\,\bullet_{h_{j,\,4 } } ) - e^{i\pi [ h_{j,\,1}+h_{j,\,4}]}\ , ( \bullet_{h_{j,\,1}},\,\circ_{h_{j,\,2}},\ , \circ_{h_{j,\,3}},\,\bullet_{h_{j,\,4 } } ) \ , , \label{c2p}\end{aligned}\ ] ] for the @xmath341 pseudoparticle , and @xmath373 \,\bigl[\prod_{g=1}^{3}(1-{\hat{\cal{t}}}_{c,\,3,\,j,\,g})\bigr]\ , ( \bullet_{h_{j,\,1}},\,\bullet_{h_{j,\,2}},\ , \bullet_{h_{j,\,3}},\,\circ_{h_{j,\,4}},\ , \circ_{h_{j,\,5}},\,\circ_{h_{j,\,6 } } ) = \nonumber \\ & - & e^{i\pi[h_{j,\,1}+h_{j,\,2}+h_{j,\,3}]}\,(\bullet_{h_{j,\,1}},\,\bullet_{h_{j,\,2}},\ , \bullet_{h_{j,\,3}},\,\circ_{h_{j,\,4}},\ , \circ_{h_{j,\,5}},\,\circ_{h_{j,\,6 } } ) \nonumber \\ & + & e^{i\pi[h_{j,\,2}+h_{j,\,3}+h_{j,\,4}]}\,(\circ_{h_{j,\,1}},\,\bullet_{h_{j,\,2}},\ , \bullet_{h_{j,\,3}},\,\bullet_{h_{j,\,4}},\ , \circ_{h_{j,\,5}},\,\circ_{h_{j,\,6 } } ) \nonumber \\ & - & e^{i\pi[h_{j,\,3}+h_{j,\,4}+h_{j,\,5}]}\,(\circ_{h_{j,\,1}},\,\circ_{h_{j,\,2}},\ , \bullet_{h_{j,\,3}},\,\bullet_{h_{j,\,4}},\ , \bullet_{h_{j,\,5}},\,\circ_{h_{j,\,6 } } ) \nonumber \\ & + & e^{i\pi[h_{j,\,1}+h_{j,\,3}+h_{j,\,5}]}\,(\bullet_{h_{j,\,1}},\,\circ_{h_{j,\,2}},\ , \bullet_{h_{j,\,3}},\,\circ_{h_{j,\,4}},\ , \bullet_{h_{j,\,5}},\,\circ_{h_{j,\,6 } } ) \nonumber \\ & - & e^{i\pi[h_{j,\,1}+h_{j,\,5}+h_{j,\,6}]}\,(\bullet_{h_{j,\,1}},\,\circ_{h_{j,\,2}},\ , \circ_{h_{j,\,3}},\,\circ_{h_{j,\,4}},\ , \bullet_{h_{j,\,5}},\,\bullet_{h_{j,\,6 } } ) \nonumber \\ & + & e^{i\pi[h_{j,\,1}+h_{j,\,2}+h_{j,\,6}]}\,(\bullet_{h_{j,\,1}},\,\bullet_{h_{j,\,2}},\ , \circ_{h_{j,\,3}},\,\circ_{h_{j,\,4}},\ , \circ_{h_{j,\,5}},\,\bullet_{h_{j,\,6 } } ) \nonumber \\ & - & e^{i\pi[h_{j,\,2}+h_{j,\,4}+h_{j,\,6}]}\,(\circ_{h_{j,\,1}},\,\bullet_{h_{j,\,2}},\ , \circ_{h_{j,\,3}},\,\bullet_{h_{j,\,4}},\ , \circ_{h_{j,\,5}},\,\bullet_{h_{j,\,6 } } ) \nonumber \\ & + & e^{i\pi[h_{j,\,4}+h_{j,\,5}+h_{j,\,6}]}\,(\circ_{h_{j,\,1}},\,\circ_{h_{j,\,2}},\ , \circ_{h_{j,\,3}},\,\bullet_{h_{j,\,4}},\ , \bullet_{h_{j,\,5}},\,\bullet_{h_{j,\,6 } } ) \ , , \label{c3p}\end{aligned}\ ] ] in the case of a @xmath343 pseudoparticle . the @xmath374 rotated - electron site distribution configurations of a @xmath375 pseudoparticle located in the effective electronic lattice at position @xmath300 $ ] are shown and discussed in appendix b. similar expressions for local @xmath277 , @xmath376 , @xmath377 , or @xmath378 pseudoparticles are obtained by replacing in eqs . ( [ c1p])-([c3p ] ) and in eq . ( [ c4p ] ) of appendix b respectively , the @xmath198 and @xmath199 sites by @xmath200 and @xmath201 sites respectively , the @xmath379 indices by @xmath380 indices , and the local rotated - electron double occupation operator phase factors by one . before using properties 3-iii , 6-iii , and 7-iii to confirm the validity of the local pseudoparticle expressions ( [ cnp])-([c3p ] ) , let us check whether these expressions conform to the requirement of property 5-iii . according to that requirement , application of the off - diagonal generators of the @xmath12 @xmath8-spin algebra given in eq . ( [ sc ] ) onto the superposition of the @xmath337 internal rotated - electron site distribution configurations of a @xmath104 pseudoparticle located in the effective electronic lattice at position @xmath381 $ ] must give zero . it is straightforward to confirm from the analysis of the changes in the rotated - electron site distribution configurations generated by these operators that this requirement is fulfilled and then , @xmath382\ , \bigl[\prod_{g=1}^{\nu}(1-{\hat{\cal{t}}}_{c,\,\nu,\,j,\,g})\bigr ] \nonumber \\ & \times & ( \bullet_{h_{j,\,1}}, ... ,\bullet_{h_{j,\,\nu } } , \,\circ_{h_{j,\,1+\nu } } , ... , \circ_{h_{j,\,2\nu } } ) = 0 \ , ; \hspace{0.5 cm } j=1,2, ... ,n_{c,\,\nu } \ , . \label{cnpsu2}\end{aligned}\ ] ] the same is valid for application of the off - diagonal generators of the @xmath12 spin algebra given in eq . ( [ ss ] ) onto the superposition of the @xmath337 internal rotated - electron site distribution configurations of a @xmath105 pseudoparticle located in the effective electronic lattice at position @xmath383 $ ] , @xmath384 \ , ( \downarrow_{l_{j,\,1 } } , ... , \downarrow_{l_{j,\,\nu}},\,\uparrow_{l_{j,\,1+\nu } } , ... , \uparrow_{l_{j,\,2\nu } } ) = 0 \ , ; \hspace{0.5 cm } j=1,2, ... ,n_{s,\,\nu } \ , . \label{snpsu2}\ ] ] in order to confirm the validity of the pseudoparticle expressions ( [ cnp])-([c3p ] ) , let us follow property 5-iii and consider again a local charge sequence with no yang holons and constituted by a number @xmath107 of rotated - electron doubly occupied sites and a number @xmath107 of rotated - electron empty sites . if we request that the rotated - electron distribution configurations of these @xmath103 sites describe a single @xmath104 pseudoparticle , according to the restrictions imposed by the bethe - ansatz solution @xcite , the associated local charge sequence must be characterized by the following numbers , @xmath385 such a local charge sequence has a single @xmath104 pseudoparticle and no @xmath104 pseudoparticle _ empty sites_. this corresponds to an _ effective @xmath104 pseudoparticle lattice _ constituted by a single _ site _ located in the effective electronic lattice at @xmath386 $ ] . that _ site _ is occupied by a @xmath104 pseudoparticle . in this limiting case in order to ensure periodic boundary conditions for the original electronic problem it is required that the charge sequence should be properly symmetrized by the method used in ref . @xcite . such a symmetrization involves the operator @xmath202 of eq . ( [ tcseq ] ) . a result of key importance is that the fulfillment of the requirement of property 5-iii imposes that the superposition of rotated - electron distribution configurations of the @xmath107 doubly occupied sites and @xmath107 empty sites ( and @xmath107 spin - down singly occupied sites and @xmath107 spin - up singly occupied sites ) of such a @xmath104 pseudoparticle ( and @xmath105 pseudoparticle ) must be expressed in the form of a number @xmath107 of @xmath387 site pairs ( and @xmath388 site pairs ) where @xmath389 and @xmath390 are such that @xmath391 and have the values @xmath392 . there is a number @xmath393 of possible different choices for these @xmath107 @xmath394 ( and @xmath388 ) site pairs . fortunately , the use of requirement 6-iii reveals that only one of these @xmath393 choices corresponds to the @xmath104 pseudoparticle ( and @xmath105 pseudoparticle ) . the remaining @xmath395 choices can be expressed as a superposition of several states described by different numbers @xmath344 of @xmath219 pseudoparticles such that @xmath346 ( and @xmath396 of @xmath220 pseudoparticles such that @xmath397 ) . moreover , the fact that the symmetrized rotated - electron site distribution configuration which describes the @xmath104 pseudoparticle ( and @xmath105 pseudoparticle ) involves a number @xmath107 of site pairs implies that it corresponds to a state with charge ( and spin ) momentum @xmath222 ( and @xmath223 ) . the possible values of such a charge momentum are given in eq . ( [ kcks ] ) . for other values of charge momentum that state can not be described in terms of a number @xmath107 of site pairs and thus the requirement introduced in property 5-iii is not fulfilled . for instance , when @xmath338 application of the basic properties 3-iii , 5-iii , and 6-iii leads uniquely to the following description of the @xmath339 pseudoparticle , @xmath398\,\sum_{\nu_c=0}^1\,e^{i\pi\nu_c}\ , \bigl({\hat{\cal{t}}}_c\bigr)^{\nu_c } \,(\bullet_{h_{1,\,1}},\,\circ_{h_{1,\,2 } } ) \ , . \label{c1scs}\ ] ] this is the product of a rotated - electron double - occupation operator phase factor with a symmetrized charge sequence . the charge sequence defined by eq . ( [ c1scs ] ) can also be written as , @xmath398\,(1-{\hat{\cal{t}}}_{c,\,1,\,1,\,1})\ , ( \bullet_{h_{1,\,1}},\,\circ_{h_{1,\,2 } } ) \ , , \label{c1alone}\ ] ] in agreement with the general expression ( [ cnp ] ) and the local @xmath339 pseudoparticle expression ( [ c1p ] ) . for @xmath233 , @xmath399 , and @xmath400 fig . 1 describes the @xmath401 site pair associated with this @xmath339 pseudoparticle . next we consider the case @xmath340 . the fulfillment of the requirement introduced in property 5-iii alone imposes in this case that the superposition of rotated - electron distribution configurations of the two doubly occupied sites and two empty sites representative of a @xmath341 pseudoparticle must be expressed in the form of two @xmath394 site pairs where @xmath389 and @xmath390 are such that @xmath402 and have the values @xmath403 . there are three possible different choices for the superposition of rotated - electron distribution configurations formed by two site pairs . these choices correspond to the following sets : ( a ) @xmath404 , ( b ) @xmath405 , and ( c ) @xmath406 . according to property 7-iii , the charge sequence ( c ) is excluded because it describes two @xmath339 pseudoparticles and thus there remain two possible choices for the @xmath341 pseudoparticle . figure 2 represents the charge sequences corresponding to the choices ( a ) and ( b ) . each of the @xmath394 site pairs are represented as in fig . 1 with @xmath233 . the figure shows how the rotated - electron site distribution configurations of these two charge sequences change as a result of a cyclic permutation which transforms the first site of the charge sequence to the last site and multiplies the final state by a phase factor of @xmath365 . this phase factor arises from the required @xmath222 charge momentum . we note that the single @xmath407 site pair descriptive of the @xmath339 pseudoparticle given in expressions ( [ c1scs ] ) and ( [ c1alone ] ) transforms into itself under such a transformation . on the other hand , in the present case the rotated - electron site distribution configuration ( a ) transforms onto two @xmath339 pseudoparticles whereas the rotated - electron site distribution configuration ( b ) transforms onto itself . thus according to property 6-iii only the rotated - electron site distribution configuration ( b ) is properly symmetrized and describes the @xmath341 pseudoparticle . moreover , only this sequence can be written as a symmetrized charge sequence of charge momentum @xmath222 . again , such a symmetrized charge sequence can also be written in the form given in the general eq . ( [ cnp ] ) and in the local @xmath341 pseudoparticle expression ( [ c2p ] ) . the two equivalent expressions read , @xmath408\,\sum_{\nu_c=0}^3\,e^{i\pi\nu_c}\ , \bigl({\hat{\cal{t}}}_c\bigr)^{\nu_c } ( \bullet_{h_{1,\,1}},\,\bullet_{h_{1,\,2}},\ , \circ_{h_{1,\,3}},\,\circ_{h_{1,\,4}})\nonumber \\ & = & \bigl[\prod_{x=1}^{4}e^{i\pi h_{1,\,x}{\hat{d}}_{j,\,x}}\bigr ] \,\bigl[\prod_{g=1}^{2}(1-{\hat{\cal{t}}}_{c,\,2,\,1,\,g})\bigr]\ , ( \bullet_{h_{1,\,1}},\,\bullet_{h_{1,\,2}},\ , \circ_{h_{1,\,3}},\,\circ_{h_{1,\,4 } } ) \ , . \label{c2alone}\end{aligned}\ ] ] let us now consider the case @xmath342 . again the fulfillment of the requirement introduced in property 5-iii imposes that the superposition of rotated - electron distribution configurations of the three doubly occupied sites and three empty sites representative of a @xmath343 pseudoparticle should be expressed in the form of three @xmath394 site pairs where @xmath389 and @xmath390 are such that @xmath402 and have the values @xmath409 . there are fifteen different possible choices for superposition of rotated - electron site distribution configurations formed by three site pairs . however , only one of these charge sequences can be written in the form of symmetrized charge sequence of charge momentum @xmath222 . for example , in fig . 3 we represent the charge sequences ( a ) @xmath410 and ( b ) @xmath411 . the figure shows how the rotated - electron site distribution configurations of these two charge sequences change as a result of a cyclic permutation which transforms the first site of the charge sequence to the last site and multiplies the final state by a phase factor of @xmath365 . both the rotated - electron site distribution configuration ( a ) and all remaining possible rotated - electron site distribution configurations constituted by three @xmath412 site pairs except the rotated - electron site distribution configuration ( b ) do not transform onto themselves . the latter rotated - electron site distribution configuration transforms onto itself , as confirmed by analysis of the figure . thus according to property 6-iii only the rotated - electron site distribution configuration ( b ) is properly symmetrized and describes the local @xmath343 pseudoparticle . only this sequence can be written in terms of a symmetrized charge sequence of charge momentum @xmath222 . again , such symmetrized charge sequence can also be written in the form given in the general eq . ( [ cnp ] ) and in the local @xmath343 expression ( [ c3p ] ) . the two equivalent expressions are the following , @xmath413\,\{\sum_{\nu_c=0}^5\,e^{i\pi\nu_c}\ , \bigl({\hat{\cal{t}}}_c\bigr)^{\nu_c}\ , ( \bullet_{h_{1,\,1}},\,\bullet_{h_{1,\,2}},\ , \bullet_{h_{1,\,3}},\,\circ_{h_{1,\,4}},\,\circ_{h_{1,\,5}},\ , \circ_{h_{1,\,6 } } ) \nonumber \\ & + & \sum_{\nu_c=0}^1\,e^{i\pi\nu_c}\,\bigl({\hat{\cal{t}}}_c\bigr)^{\nu_c}\ , ( \circ_{h_{1,\,1}},\,\bullet_{h_{1,\,2}},\ , \circ_{h_{1,\,3}},\,\bullet_{h_{1,\,4}},\,\circ_{h_{1,\,5}},\ , \bullet_{h_{1,\,6}})\ } \nonumber \\ & = & \bigl[\prod_{x=1}^{6}e^{i\pi h_{1,\,x}{\hat{d}}_{j,\,x}}\bigr]\ , \bigl[\prod_{g=1}^{3}(1-{\hat{\cal{t}}}_{c,\,3,\,1,\,g})\bigr]\ , ( \bullet_{h_{1,\,1}},\,\bullet_{h_{1,\,2}},\ , \bullet_{h_{1,\,3}},\,\circ_{h_{1,\,4}},\,\circ_{h_{1,\,5}},\ , \circ_{h_{1,\,6 } } ) \ , . \label{c3alone}\end{aligned}\ ] ] note that in this case one needs two symmetrized charge sequences to reach the @xmath343 expression ( [ c3p ] ) upon application of the operator @xmath202 onto these charge sequences two and six times respectively . however , in agreement with properties 5-iii and 6-iii , the state obtained from the superposition of these two charge sequences is an eigenstate of the charge momentum operator with eigenvalue @xmath222 . also the @xmath104 pseudoparticles belonging to branches such that @xmath414 are a superposition of several such symmetrized charge sequences . however , always the state obtained from the superposition of several charge sequences is an eigenstate of the charge momentum operator with eigenvalue @xmath222 . that state can also be expressed in the general form given in eq . ( [ cnp ] ) . as a last example we provide the expression in terms of symmetrized charge sequences of the rotated - electron site distribution configurations representative of a @xmath375 pseudoparticle . this can again be written both ( i ) in the form given in eqs . ( [ cnp ] ) and in eq . ( [ c4p ] ) of appendix b and ( ii ) as a superposition of symmetrized charge sequences . the two equivalent expressions are in this case given by , @xmath415\,\{\sum_{\nu_c=0}^7\,e^{i\pi\nu_c}\ , \bigl({\hat{\cal{t}}}_c\bigr)^{\nu_c}\ , ( \bullet_{h_{1,\,1}},\,\bullet_{h_{1,\,2}},\ , \bullet_{h_{1,\,3}},\,\bullet_{h_{1,\,4}},\,\circ_{h_{1,\,5}},\ , \circ_{h_{1,\,6}},\,\circ_{h_{1,\,7}},\ , \circ_{h_{1,\,8 } } ) \nonumber \\ & + & \sum_{\nu_c=0}^7\,e^{i\pi\nu_c}\,\bigl({\hat{\cal{t}}}_c\bigr)^{\nu_c}\ , ( \bullet_{h_{1,\,1}},\,\circ_{h_{1,\,2}},\ , \bullet_{h_{1,\,3}},\,\circ_{h_{1,\,4}},\,\circ_{h_{1,\,5}},\ , \bullet_{h_{1,\,6}},\,\circ_{h_{1,\,7}},\ , \bullet_{h_{1,\,8}})\ } \nonumber \\ & = & \bigl[\prod_{x=1}^{8}e^{i\pi h_{1,\,x}{\hat{d}}_{j,\,x}}\bigr]\ , \bigl[\prod_{g=1}^{4}(1-{\hat{\cal{t}}}_{c,\,4,\,1,\,g})\bigr]\ , ( \bullet_{h_{1,\,1}},\,\bullet_{h_{1,\,2}},\ , \bullet_{h_{1,\,3}},\,\bullet_{h_{1,\,4}},\,\circ_{h_{1,\,5}},\ , \circ_{h_{1,\,6}},\,\circ_{h_{1,\,7}},\ , \circ_{h_{1,\,8 } } ) \ , . \label{c4alone}\end{aligned}\ ] ] also in this case one needs more than one symmetrized charge sequence to reach the @xmath375 expression ( [ c4p ] ) given in appendix a. again , the state obtained from the superposition of the involved charge sequences is an eigenstate of the charge momentum operator with eigenvalue @xmath222 . for the case of charge ( and spin ) sequences constituted by several local @xmath104 pseudoparticles ( and local @xmath105 pseudoparticles ) belonging to one or several @xmath274 branches and by yang holons ( and hl spinons ) the internal structure of the local @xmath104 pseudoparticles ( and local @xmath105 pseudoparticles ) is given by eq . ( [ cnp ] ) ( and eq . ( [ snp ] ) ) . this result follows from the basic property 7-iii . according to that property the rotated - electron distribution configurations of the @xmath107 doubly occupied sites and @xmath107 empty sites ( and @xmath107 spin - down singly occupied sites and @xmath107 spin - up singly occupied sites ) which describe the internal structure , a @xmath104 pseudoparticle ( and @xmath105 pseudoparticle ) located in the effective electronic lattice at position @xmath381 $ ] ( and @xmath315 $ ] ) are the same as for the above limiting charge ( and spin ) sequence constituted by a single pseudoparticle . we will in the ensuing section confirm that in agreement with property 8-iii , this ensures the periodic boundary conditions of the original electronic problem provided that the band - momentum energy eigenstates are suitable fourier - transform superpositions of the local charge sequences with band - momentum values obeying eqs . ( [ pbccp ] ) and ( [ pbcanp ] ) . however , whether inside the domain of @xmath103 sites of a local @xmath104 pseudoparticle there can be sites with index @xmath289 such that @xmath416 and which sites are occupied by yang holons or other local @xmath219 pseudoparticles remains an open question . this problem is absent in the limiting case we considered here where the charge sequence is constituted by a number @xmath103 of sites of the effective electronic lattice . this issue is closely related to the definition of the local @xmath104 pseudoparticle charge - sequence _ empty sites _ which is discussed in the following . a similar problem occurs for the spin sequences . in the definition of the @xmath417 charge - sequence _ empty sites _ of a local @xmath104 pseudoparticle it is useful to consider different charge sequences where all remaining local @xmath219 pseudoparticles except that specific local @xmath104 pseudoparticle keep their charge - sequence positions @xmath319 defined in eq . ( [ bstringc ] ) . below we call _ steady _ local @xmath219 pseudoparticles the former local pseudoparticles whose charge - sequence positions @xmath319 remain unchanged . obviously , any local pseudoparticle can be chosen to be the _ moving _ pseudoparticle . this just refers to a set of different charge sequences where the charge - sequence position of that quantum object changes whereas the charge - sequence positions of the remaining local pseudoparticles remain unchanged . we note that according to eq . ( [ bstringc ] ) the charge - sequence position of a local @xmath104 pseudoparticle can remain unchanged in spite of the _ movements _ of some of its @xmath103 sites , as we confirm below . concerning the definition of the @xmath417 charge - sequence _ empty sites _ of a local @xmath104 pseudoparticle , from the use of properties 1-iii to 8-iii and of other properties of the 1d hubbard model we find the following distribution configuration properties : 1-iv let us consider a set of rotated - electron doubly occupied and/or rotated - electron empty sites located at positions @xmath418 $ ] where according to eqs . ( [ hl ] ) and ( [ chcl ] ) the index @xmath284 has the values @xmath419 $ ] . moreover , we consider that all of these sites are located inside the domain of sites of a local @xmath104 pseudoparticle and thus their indices @xmath289 are such that @xmath416 and @xmath420 where @xmath308 . this property states that the set of rotated - electron doubly occupied / empty sites of index @xmath289 obeying these inequalities can not correspond to sites occupied by yang holons or to any of the @xmath421 sites of the charge sequence associated with local @xmath219 pseudoparticles belonging to branches such that @xmath422 . thus such a set of rotated - electron doubly occupied or empty sites can only correspond to sub - sets of @xmath421 sites of the charge sequence associated with local @xmath219 pseudoparticles belonging to branches such that @xmath423 . 2-iv let us consider a local @xmath104 pseudoparticle and a local @xmath219 pseudoparticle belonging to the same charge sequence and to branches such that @xmath424 . moreover , we consider that there are no charge - sequence local pseudoparticles inside the @xmath421-site domain of the @xmath219 pseudoparticle other than the @xmath104 pseudoparticle . the description of cases involving more than two local pseudoparticles can be easily reached by generalization of the two - pseudoparticle situation considered here . then the @xmath103 sites of the charge sequence associated with the local @xmath104 pseudoparticle are located either outside the domain of @xmath421 sites of the local @xmath219 pseudoparticle limited in the left and right ends by the sites of index @xmath425 and @xmath426 respectively , or are all located inside that domain . if the @xmath103 sites are located inside the domain of these @xmath421 sites there is a number @xmath427 of permitted positions for the @xmath104 pseudoparticle . in each of these permitted positions all of the @xmath103 sites of the local @xmath104 pseudoparticle are located between two first neighboring sites of indices @xmath428 and @xmath429 of the local @xmath219 pseudoparticle . these permitted positions correspond to the neighboring sites of indices @xmath428 and @xmath429 such that @xmath430 . 3-iv let us consider again the local @xmath104 pseudoparticle and the local @xmath219 pseudoparticle referred to in property 2-iv . the charge - sequence position @xmath319 given in eq . ( [ bstringc ] ) of such a steady local @xmath219 pseudoparticle remains unchanged in all possible @xmath427 positions of the local @xmath104 pseudoparticle while passing its @xmath421-site domain . the inverse statement is even stronger . it is required that when a local @xmath219 pseudoparticle belonging to a branch such that @xmath431 passes a steady local @xmath104 pseudoparticle , the position in the effective electronic lattice of all the @xmath103 sites of the latter pseudoparticle must remain unchanged . moreover , the @xmath427 possible different positions of the @xmath103 sites of the local @xmath104 pseudoparticle inside the site domain of the local @xmath219 pseudoparticle must be the same when the local @xmath104 pseudoparticle passes the steady local @xmath219 pseudoparticle from its left to its right - hand side and vice versa . also the relative positions of the two quantum objects when the local @xmath104 pseudoparticle passes the steady local @xmath219 pseudoparticle and when the local @xmath219 pseudoparticle passes the steady local @xmath104 pseudoparticle must be the same . however , note that although these relative positions are the same , when a local @xmath219 pseudoparticle belonging to a branch such that @xmath431 passes a steady local @xmath104 pseudoparticle it does not occupy any of the @xmath103 sites of the effective electronic lattice occupied by the latter local pseudoparticle . thus the positions in the lattice of these @xmath103 sites remain indeed unchanged . a last requirement is that the positions of all @xmath421 sites of a steady local @xmath219 pseudoparticle belonging to a branch such that @xmath431 must be the same in the rotated - electron site distribution configurations before and after it is passed by the moving local @xmath104 pseudoparticle . if the local @xmath104 pseudoparticle is moving from the left to the right - hand side of the steady local @xmath219 pseudoparticle , _ before _ and _ after _ is meant here as the rotated - electron site distribution configurations where the @xmath103 sites of the former quantum object are located outside the domain of @xmath421 sites of the steady local pseudoparticle and on its right and left hand sides respectively . 4-iv corresponding rules are also valid for the case of the spin sequences and local @xmath105 pseudoparticles . a simple example of a domain constituted by fourteen sites of a permitted charge sequence is represented in fig . 4 . in such a figure the @xmath432 site pairs of the local @xmath104 pseudoparticles are represented as in figs . the six sites with no vertical lines correspond to empty sites associated with @xmath89 yang holons . there are two local @xmath339 pseudoparticles and a local @xmath341 pseudoparticle in the charge - sequence domain represented in the figure . in order to confirm the validity of the counting of rotated - electron site distribution occupancies leading to expression ( [ nanns ] ) , let us use the properties 1-iv to 4-iv in the classification of the different @xmath433 possible positions of a local @xmath104 pseudoparticle in the effective electronic lattice . our analysis can be easily generalized to the case of a @xmath105 pseudoparticle . our study includes the introduction of the concept of _ @xmath103-leg caterpillar step _ , where the @xmath103-leg caterpillar is nothing but the @xmath103-site pseudoparticle block associated with the moving @xmath104 pseudoparticle . for simplicity we consider a charge sequence such that @xmath434 for the @xmath107 branch . however , the generalization of our results to branches with finite occupancy of several local pseudoparticles is straightforward . as mentioned above , the possible positions of such a pseudoparticle correspond to a number @xmath433 of different charge sequences where the charge - sequence positions of all remaining quantum objects remain unchanged . thus these @xmath433 charge sequences differ in the position of the @xmath104 pseudoparticle relative to the other quantum objects . the above property 3-iv states that the charge - sequence position @xmath319 given in eq . ( [ bstringc ] ) of all quantum objects except that of the _ moving _ @xmath104 pseudoparticle must remain unchanged in all these @xmath433 charge sequences . let us confirm that the charge - sequence position of a steady local @xmath219 pseudoparticle belonging to a branch such that @xmath422 indeed remains unchanged when it is passed by a local @xmath104 pseudoparticle , in spite of changes of the positions in the effective electronic lattice of some of its @xmath421 sites . we start by considering the case of a local @xmath435 pseudoparticle where @xmath436 denotes the largest @xmath107 value of the @xmath104 pseudoparticle branches with finite occupancy in the charge sequence . according to eq . ( [ n*sum ] ) , in this case the number of local @xmath435 pseudoparticle _ empty sites _ is given by @xmath437 and equals the number of yang holons . the different possible positions of a @xmath435 pseudoparticle can be achieved by elementary steps in the charge sequence where each of its @xmath438 sites moves forward by a single site of the charge sequence . if the movement is from the left to the right - hand side , the @xmath439 pseudoparticle site of index @xmath440 moves into the site previously occupied by a yang holon . the remaining @xmath441 pseudoparticle sites of index @xmath307 move into the site previously associated with the pseudoparticle site of index @xmath442 where @xmath443 . finally , the yang holon removed from the site newly associated with the pseudoparticle site of index @xmath440 moves into the site previously associated with the first pseudoparticle site of index @xmath425 . we call such an elementary collective step of the @xmath438 pseudoparticle sites _ @xmath438-leg caterpillar step_. according to such an analogy the @xmath438 _ legs _ refer to these sites and the _ caterpillar _ refers to the local pseudoparticle . while each of these legs moves forward by a charge - sequence site , the whole caterpillar itself also moves forward by a single charge - sequence site . the net result of such a @xmath438-leg caterpillar step is that the compact domain of @xmath438 pseudoparticle sites interchanges position with a yang holon . when such a moving @xmath435 pseudoparticle passes a steady local @xmath104 pseudoparticle , each of its @xmath438 sites jump the @xmath103 sites of the latter local pseudoparticle . thus the @xmath438-leg caterpillar moves forward like the @xmath103 sites of such a steady local @xmath104 pseudoparticle do not exist in the charge sequence . as we discuss below , the construction of the corresponding effective @xmath435 pseudoparticle lattice mentioned in property 8-iii involves omission of the sites of the effective electronic lattice belonging to the @xmath103-site domains of such a local @xmath104 pseudoparticles . moreover , for such an effective @xmath435 pseudoparticle lattice the @xmath438-site domain of the @xmath435 pseudoparticles is _ seen _ as a point - like occupied site . thus in the particular case of the @xmath444 branch such an effective lattice has @xmath445 sites and a number @xmath446 free of @xmath435 pseudoparticles , as we confirm below . let us consider now the case of a general local @xmath104 pseudoparticle belonging to a branch such that @xmath447 . for simplicity we keep the assumption that @xmath434 for the @xmath107 branch . we consider a given charge sequence where the @xmath104 pseudoparticle is located in the effective electronic lattice at position @xmath300 $ ] . it follows from property 4-iii that the @xmath417 remaining possible positions of the local @xmath104 pseudoparticle in the charge sequence define the positions of the local @xmath104 pseudoparticle charge - sequence _ empty sites_. out of these @xmath448 possible positions , @xmath281 is the number of positions associated with yang holons . again , the different positions of the moving local @xmath104 pseudoparticle can be achieved by @xmath103-leg caterpillar steps . the net result of such a @xmath103-leg caterpillar step is that the compact domain constituted by the @xmath103 @xmath104 pseudoparticle sites interchanges position with a yang holon or with a site of a @xmath449 pair of a steady local @xmath219 pseudoparticle belonging to a branch such that @xmath422 . a @xmath103-leg caterpillar step is defined here as in the case of the above moving @xmath435 pseudoparticle . if the movement is from the left to the right - hand side , the local @xmath104 pseudoparticle site of index @xmath450 moves into the site previously occupied by a yang holon or into one of the two sites previously associated with a local @xmath219 pseudoparticle @xmath451 site pair . the remaining @xmath452 sites of the local @xmath104 pseudoparticle of index @xmath307 such that @xmath453 , move into the site previously associated with the local @xmath104 pseudoparticle site of index @xmath442 where again @xmath453 . finally , the yang holon or @xmath219 pseudoparticle site associated with the local @xmath454 site pair which was removed from the site newly associated with the local @xmath104 pseudoparticle site of index @xmath455 , moves into the site previously associated with the local @xmath104 pseudoparticle site of index @xmath425 . again , the net result of such a @xmath103-leg caterpillar step is that the local @xmath104 pseudoparticle moves forward by a single charge - sequence site . let us consider that in the initial rotated - electron site distribution configuration the moving @xmath104 pseudoparticle is located on the left - hand side of the steady local @xmath219 pseudoparticle . thus in the final state just after passing the latter pseudoparticle , the @xmath104 pseudoparticle will occupy in the charge sequence a domain of @xmath103 sites located on the right - hand side of the @xmath219 pseudoparticle site domain . according to property 3-iv , in both the initial and final rotated - electron site distribution configurations all @xmath421 sites of the latter quantum object remain the same . for simplicity we consider that in the initial rotated - electron site distribution configuration , the @xmath103 charge - sequence sites located just on the right - hand side of the steady local @xmath219 pseudoparticle are occupied by yang holons . however , our results can be generalized to rotated - electron site distribution configurations where these sites are occupied by a third pseudoparticle . another case which we could consider is when the moving local pseudoparticle passes two steady local @xmath219 and @xmath456 pseudoparticles such that one of them is located within the site domain of the other one . from the systematic use of the properties 1-iv to 4-iv one can describe the movements of a local @xmath104 pseudoparticle when it passes any rotated - electron site distribution configuration involving several steady local pseudoparticles . according to property 2-iv the @xmath103 sites of a local @xmath104 pseudoparticle can have @xmath427 different positions inside the @xmath421-site domain of a local @xmath219 pseudoparticle belonging to a branch such that @xmath422 . we consider that in the initial rotated - electron site distribution configuration , the local @xmath104 and @xmath219 pseudoparticles are located side by side in the charge sequence . thus these two quantum objects occupy together a compact @xmath457-site domain . by simple counting arguments it is straightforward to realize that in order to reach each of these permitted positions each of the @xmath103 sites of the local @xmath104 pseudoparticle must perform a jump of @xmath103 sites in the charge sequence somewhere inside the @xmath421-site domain of the steady local @xmath219 pseudoparticle . this result is consistent with the expression of the number @xmath458 of the available positions for the moving local @xmath104 pseudoparticle involving sites of steady local @xmath219 pseudoparticles belonging to branches such that @xmath422 . this expression reveals that when located inside the site domain of such a pseudoparticle each of the @xmath103 sites of the local @xmath104 pseudoparticle can move into a possible @xmath459 sites out of the @xmath421 total number of sites of the steady local @xmath219 pseudoparticle . this agrees with the above result obtained from counting arguments alone . we note that the number of charge - sequence sites jumped by each local @xmath104 caterpillar leg equals precisely @xmath103 , the number of sites of such a @xmath104 pseudoparticle . thus these jumps ensure that the position in the effective electronic lattice of all the @xmath421 sites of the steady local @xmath219 pseudoparticle remains unchanged after the @xmath104 passes it , as requested by the property 3-iv . this result also applies to the general situation when a moving local @xmath104 pseudoparticle passes a domain containing several local steady @xmath219 pseudoparticles belonging to branches such that @xmath422 . also in the general situation that each of the legs of the moving caterpillar uses @xmath460 possible sites of each steady local @xmath219 pseudoparticle , out of a total of @xmath421 sites . the question which arises from the occurrence of the above jump mechanism is which are the @xmath103 sites of the steady local @xmath219 pseudoparticle that are jumped by each site of the moving local @xmath104 pseudoparticle . fortunately , the requirements mentioned in the property 3-iv provide the needed information for the precise definition of the @xmath427 permitted positions of the local @xmath104 pseudoparticle inside the steady local @xmath219 pseudoparticle . we find that in order to reach these permitted positions each site of the local @xmath104 pseudoparticle must perform two independent jumps of @xmath107 sites in the charge sequence . let us suppose that the local @xmath104 pseudoparticle is moving from the left to the right - hand side . these jumps bring the local @xmath104 pseudoparticle to and from respectively , a position where its charge - sequence location @xmath319 , as given by eq . ( [ bstringc ] ) , coincides with that of the steady local @xmath219 pseudoparticle . in such an event each of the @xmath103 legs of the caterpillar jumps a number @xmath107 of sites and reaches a position where its @xmath103 legs occupy @xmath103 sites of the charge sequence located around the center of the steady local @xmath219 pseudoparticle . in order to move forward to the right - hand side from that position , each of the @xmath103 legs of the caterpillar jump again a number of @xmath107 sites of the charge sequence . thus the possible positions of a local @xmath104 pseudoparticle when it passes from the left to the right - hand side of a steady local @xmath219 pseudoparticle are the following : first the local @xmath104 pseudoparticle performs a @xmath103-leg caterpillar step . the net result of such an elementary collective step is that the compact domain of @xmath103 sites of the local @xmath104 pseudoparticle interchanges position with the site of index @xmath461 of the local @xmath219 pseudoparticle . thus this first elementary collective step brings the local @xmath104 pseudoparticle inside the domain of @xmath421 sites of the steady local @xmath219 pseudoparticle . in the reached rotated - electron site distribution configuration all the @xmath103 sites of the local @xmath104 pseudoparticle are located between the sites of index @xmath461 and @xmath462 of the steady local @xmath219 pseudoparticle . this procedure is repeated @xmath463 times until the @xmath103 sites of the @xmath104 pseudoparticle are located between the sites of index @xmath464 and @xmath465 of the steady local @xmath219 pseudoparticle . in each of the corresponding @xmath103-leg caterpillar steps the local @xmath104 pseudoparticle moves forward by a single lattice site of the charge sequence . however , in the @xmath466 step each site of the @xmath104 pseudoparticle jumps @xmath107 sites of the charge sequence . this jump brings it to a rotated - electron site distribution configuration where the charge - sequence position @xmath319 of the local @xmath104 pseudoparticle given in eq . ( [ bstringc ] ) coincides with that of the steady local @xmath219 pseudoparticle . let @xmath307 where @xmath467 denote the indices of the @xmath103 sites of the @xmath104 pseudoparticle while it was located between the sites of index @xmath464 and @xmath465 of the local @xmath219 pseudoparticle . in the next collective movement the local @xmath104 pseudoparticle site of index @xmath450 jumps into the site which was associated with the local @xmath219 pseudoparticle site of index @xmath468 . out of the remaining @xmath452 sites of the local @xmath104 pseudoparticle , @xmath107 of these sites of indices @xmath307 such that @xmath469 jump into the local @xmath219 pseudoparticle sites of indices @xmath470 where again @xmath469 . the remaining @xmath471 sites of the local @xmath104 pseudoparticle of indices @xmath307 such that @xmath472 jump into the sites which were occupied by the local @xmath104 pseudoparticle sites of indices @xmath307 where @xmath473 . as a result of these events the local @xmath219 pseudoparticle site of index @xmath465 jumps @xmath471 charge - sequence sites from the right to the left - hand side and thus moves by a block of @xmath107 sites in such a direction . in contrast , all the remaining @xmath219 pseudoparticle sites of indices @xmath474 such that @xmath475 jump @xmath471 charge - sequence sites from the left to the right - hand side and move by a block of @xmath107 sites in such a direction . these collective jumps reach a configuration where the @xmath103 sites of the @xmath104 pseudoparticle are located between the sites of index @xmath465 and @xmath476 of the @xmath219 pseudoparticle . in order to reach the next permitted position each of the @xmath103 local @xmath104 pseudoparticle sites jump again a number @xmath107 of lattice sites of the charge sequence . this brings the local @xmath104 pseudoparticle to a distribution configuration where its @xmath103 sites are located between the sites of index @xmath476 and @xmath477 of the steady local @xmath219 pseudoparticle . these two jumps are followed by @xmath463 @xmath103-leg caterpillar steps where the @xmath104 pseudoparticle moves again by a single lattice site of the charge sequence . the net result of each of these elementary collective steps is that the compact domain of @xmath103 @xmath104 pseudoparticle sites interchanges position with one of the sites of the @xmath219 pseudoparticle . step number @xmath478 of these @xmath463 @xmath103-leg caterpillar steps leads to a distribution configuration where the @xmath103 sites of the @xmath104 pseudoparticle are located between the sites of index @xmath479 and @xmath480 of the steady local @xmath219 pseudoparticle . finally , the last step brings the local @xmath104 pseudoparticle to a distribution configuration where it is located outside and on the right - hand side of the domain of @xmath421 sites of the steady local @xmath219 pseudoparticle . the same intermediate positions are reached when the @xmath104 pseudoparticle passes the domain of @xmath421 sites of the @xmath219 pseudoparticle from the right to the left - hand side . we have just studied a situation where a local @xmath104 pseudoparticle passes a steady local @xmath219 pseudoparticle belonging to a branch such that @xmath422 . for simplicity we assumed that in the initial rotated - electron site distribution configuration the domain of @xmath103 charge - sequence sites located just on the right - hand side of the steady local @xmath219 pseudoparticle are occupied by yang holons . however , our results can be generalized to more complicated situations . an example is a situation where a local @xmath104 pseudoparticle passes a steady local @xmath219 pseudoparticle which has as first neighbor in the charge sequence a second steady local @xmath456 pseudoparticle . another example is when a local @xmath104 pseudoparticle passes a steady local @xmath219 pseudoparticle containing inside its @xmath421-site domain a third steady local @xmath456 pseudoparticle belonging to a branch such that @xmath481 . these more general situations can be described by the use of the properties 1-iv to 4-iv . one always finds that in order to move forward , the @xmath103 sites of the local @xmath104 pseudoparticle use @xmath460 sites out of the total number of @xmath421 sites of each steady local @xmath219 pseudoparticle belonging to branches such that @xmath422 . moreover , when passing the central region of the @xmath421-site domain of such a local @xmath219 pseudoparticle , the @xmath103 legs of the @xmath104 caterpillar always perform the two @xmath107-site jumps described above . in summary , the number of charge - sequence _ empty sites _ of a @xmath104 pseudoparticle equals the number of sites occupied by yang holons plus a number of sites given by @xmath482 for each @xmath219 pseudoparticle of the charge sequence belonging to branches such that @xmath422 . the number of permitted positions of the local @xmath104 pseudoparticle inside each of the site domains of such a @xmath219 pseudoparticle is @xmath427 . in these permitted positions the @xmath104 pseudoparticle is located between the sites @xmath483 and @xmath429 such that @xmath430 of the local @xmath219 pseudoparticle . in order to reach or to leave a position between the sites of index @xmath465 and @xmath476 of a steady local @xmath219 pseudoparticle belonging to a branch such that @xmath422 , each of the @xmath103 legs of the moving caterpillar always performs a jump of @xmath107 sites in the charge sequence . otherwise it moves throughout the charge sequence by @xmath103-leg caterpillar steps . in such steps the @xmath103-site block of the local @xmath104 pseudoparticle moves forward by a charge - sequence site . the net result of such an event is that the @xmath103-site domain of the moving quantum object interchanges position with either a yang holon or with a site of a @xmath219 pseudoparticle belonging to a branch such that @xmath422 . the present definition of the permitted positions of a local @xmath104 pseudoparticle when passing the remaining quantum objects which occupy a charge sequence provides a consistent , complete , and unique description for the rotated - electron distribution configurations of doubly occupied sites and empty sites which describe the charge sequence . the results obtained for the charge sequences and corresponding @xmath104 pseudoparticles are also valid for the spin sequences and their @xmath105 pseudoparticles . the events studied above are illustrated in figs . 5 and 6 for the case when a local @xmath339 pseudoparticle passes a steady local @xmath339 pseudoparticle and a steady local @xmath341 pseudoparticle respectively , from the left to the right - hand side . if instead we consider that the local @xmath339 pseudoparticle is steady , fig . 6 illustrates the possible positions of a local @xmath341 pseudoparticle when it passes from the right to the left - hand side a steady local @xmath339 pseudoparticle . in this case the position in the effective electronic lattice of the two sites of the @xmath339 pseudoparticle remains unchanged and the different relative positions of the two quantum objects are reached by movements of the @xmath341 pseudoparticle . importantly , we emphasize that in this case the @xmath341 pseudoparticle does not occupy the sites occupied by the @xmath339 pseudoparticle . again this is consistent with property 3-iv . the charge sequence ( and the spin sequence ) is obtained by omission of the rotated - electron singly occupied sites ( and rotated - electron doubly occupied / empty sites ) . such a procedure follows from the independent conservation of the charge and spin sequence . we note that each of the @xmath104 pseudoparticle ( and @xmath105 pseudoparticle ) branch occupancy configurations are also independently conserved . thus it follows our above analysis that the construction of the effective @xmath104 pseudoparticle lattice mentioned in property 8-iii also involves omission of the sites of the effective electronic lattice which are jumped by the movements of a @xmath104 pseudoparticle , since these sites are not _ seen _ by such a moving quantum object . the omitted sites correspond to the @xmath421-site domains of local @xmath219 pseudoparticles of branches such that @xmath484 and to @xmath103-site domains belonging to the @xmath421 sites of each @xmath219 pseudoparticle such that @xmath422 . moreover , for such an effective @xmath104 pseudoparticle lattice the @xmath103-site domain of the @xmath104 pseudoparticles is _ seen _ as a point - like occupied site . thus such an effective lattice has @xmath485 sites , a number @xmath486 of which are empty of @xmath104 pseudoparticles , as we confirm below . a similar analysis holds for the @xmath105 pseudoparticles . we have just found the rotated - electron distribution configurations of the @xmath103 sites which describe a local @xmath18 pseudoparticle . moreover , we also defined the @xmath322 charge - sequence _ empty sites _ of such a local @xmath18 pseudoparticle . although our results can be generalized to the case of the local spin sequences and corresponding local @xmath105 pseudoparticles , often our analysis focused explicitly on the case of the local charge sequences and associated local @xmath104 pseudoparticles . an important result which follows from our findings is that indeed the positions of the @xmath417 charge - sequence _ empty sites _ of each local @xmath104 pseudoparticle branch are fully determined by the location of the @xmath19 pseudoparticles and @xmath104 pseudoparticles alone . our results thus confirm that for fixed values of the numbers @xmath328 and @xmath433 the number of occupancy configurations of the local @xmath104 pseudoparticles is given by eq . ( [ nanns ] ) for @xmath114 . it follows that we can classify the local @xmath104 pseudoparticle occupancy configurations for @xmath274 branches with finite occupancy in a given state by providing the indices @xmath326 given in eqs . ( [ stringc ] ) . these indices correspond to the @xmath328 pseudoparticle locations provided in eq . ( [ localcnp ] ) . our results also confirm that the number of local pseudoparticle occupancy configurations of a cphs ensemble subspace is indeed given by eq . ( [ ncphs ] ) . finally , since our results can be generalized to spin sequences as well , they also validate the use of the representation associated with the local state defined in eqs . ( [ localst])-([args ] ) . such a local state represents each of the different local pseudoparticle occupancy configurations whose number for a specific cphs ensemble subspace is given in eq . ( [ ncphs ] ) . our results also confirm that the set of these local states whose number is given in eq . ( [ ncphs ] ) constitutes a complete set of states in the corresponding cphs ensemble subspace . thus in the case of cphs ensemble subspaces with no @xmath9 yang holons and no @xmath9 hl spinons the local states defined in eqs . ( [ localst])-([args ] ) constitute a complete set of states . our main goal here is the generalization of these local states to cphs ensemble subspaces with finite occupancies of both @xmath102 yang holons and @xmath102 hl spinons . we start by expressing the complete set of local states defined in eqs . ( [ localst])-([args ] ) in terms of the following more basic states , @xmath487 where the charge sequence refers to the following sets of rotated - electron distribution configurations of doubly - occupied and empty sites , @xmath488 and the same follows for the spin sequence with the index @xmath284 replaced by the index @xmath285 , rotated - electron doubly occupied sites @xmath198 replaced by spin - down rotated - electron singly occupied sites @xmath200 , and sites free of rotated electrons @xmath199 replaced by spin - up rotated - electron singly occupied sites @xmath201 . the states defined in eqs . ( [ initial])-([argcn ] ) provide the location of the @xmath103 sites associated with each of the @xmath18 pseudoparticles whose branches have finite occupancy in the corresponding charge and spin sequence . however , we note that these states include only rotated - electron site distribution configurations of the simple type given in eqs . ( [ lcnp ] ) and ( [ lsnp ] ) . the general charge and spin sequences can be obtained from these states by application of suitable operators , as we discussed above . these are thus the most basic and simple states . our first task is expressing the states defined in eqs . ( [ localst])-([args ] ) in terms of these simple states . each local @xmath104 and @xmath105 pseudoparticle is described by the superposition of @xmath337 rotated - electron site distribution configurations given in eqs . ( [ cnp ] ) and ( [ snp ] ) respectively . the type of configuration superposition given in these equations is common to the description of all local @xmath18 pseudoparticles of a charge or spin sequence . thus the proper description of the the charge and spin sequence of each of the states defined in eqs . ( [ localst])-([args ] ) involves the superposition of @xmath489 such a rotated - electron site distribution configurations . it follows that these normalized local states can be expressed as follows , @xmath490/2}\,\vert(j_{1},\,j_2, ... ,j_{n_c});\ , \{(h_{1},\,h_2, ... ,h_{n_{c,\,\nu}})\};\,\{(l_{1},\,l_2, ... ,l_{n_{s,\,\nu}})\}\rangle \nonumber \\ & = & \bigl[\prod_{\nu'=1}^{\infty}\prod_{j'=1}^{n_{c,\nu'}}\prod_{x=1}^{2\nu'}e^{i\pi h_{j',\,x}{\hat{d}}_{j',\,x}}\bigr]\,\bigl[\prod_{\alpha = c , s}\prod_{\nu''=1}^{\infty}\prod_{j''=1}^{n_{\alpha,\nu '' } } \prod_{g''=1}^{\nu''}(1-{\hat{\cal{t}}}_{\alpha,\,\nu'',\,j'',\,g''})\bigr]\nonumber \\ & \times & \vert(j_{1},\,j_2, ... ,j_{n_c});\ , \{(\bullet_{h_{j,\,1}}, ... ,\bullet_{h_{j,\,\nu } } , \,\circ_{h_{j,\,1+\nu } } , ... , \circ_{h_{j,\,2\nu}})\};\,\{(\downarrow_{l_{j,\,1 } } , ... , \downarrow_{l_{j,\,\nu}},\,\uparrow_{l_{j,\,1+\nu } } , ... , \uparrow_{l_{j,\,2\nu}})\}\rangle \ , , \label{prima}\end{aligned}\ ] ] where the operator @xmath491 measures the number of rotated - electron doubly occupied sites in the charge - sequence site of index @xmath474 ( whose value is given by @xmath246 or @xmath247 ) and the transformation laws for application of the operator @xmath492 were given above and are illustrated in eqs . ( [ tc ] ) and ( [ ts ] ) . the local states given in both eqs . ( [ localst])-([args ] ) and in eq . ( [ prima ] ) provide a complete basis of states for cphs ensemble subspaces spanned by lowest - weight states of both the @xmath12 @xmath8-spin and spin algebras . let us now consider the general case of states associated with cphs ensemble subspaces spanned by states with both finite occupancies of @xmath102 yang holons and @xmath102 hl spinons . the description of the local charge and spin sequences with finite occupancies of @xmath102 yang holons and @xmath102 hl spinons respectively , is quite similar to the ones we studied above . for simplicity let us consider first the case of the charge sequences . according to property 1-iii the @xmath9 yang holons and @xmath89 yang holons correspond to single rotated - electron doubly occupied lattice sites and single rotated - electron empty lattice sites of such a charge sequence respectively . in the case of a lowest weight state ( and highest weight state ) of the @xmath8-spin algebra all yang holons have the same @xmath8-spin projection @xmath493 ( and @xmath494 ) and correspond to rotated - electron empty sites ( and rotated - electron doubly occupied sites ) . the application of the off - diagonal generators of the @xmath12 @xmath8-spin algebra given in eq . ( [ sc ] ) leads to flips of the yang holons @xmath8-spins @xcite . following property 5-iii , in order to achieve the yang holon occupancy configurations required by the @xmath12 @xmath8-spin algebra , we find that each local charge sequence is described by a superposition @xmath495 of yang holon occupancy configurations . here @xmath496 stands for the total number of yang holons in the charge sequence . all the @xmath497 states associated with these configurations have the same set of @xmath281 lattice sites occupied by yang holons and the same numbers of @xmath9 yang holons and @xmath89 yang holons . however , these states differ in the distributions of the @xmath9 yang holons and @xmath89 yang holons over the @xmath281 lattice sites . the same analysis holds for spin sequences with finite occupancies of @xmath102 hl spinons . thus there are @xmath498 states with precisely the same occupancy configurations of local @xmath19 pseudoparticles and local @xmath18 pseudoparticles . each of these occupancy configurations is described by a state of form given in eq . ( [ prima ] ) . on the other hand , we denote each of the @xmath498 general states by , @xmath499 where @xmath500 and @xmath501 are the corresponding numbers of @xmath9 yang holons and @xmath9 hl spinons respectively . we note that eq . ( [ lcslws ] ) remains valid for the general local states ( [ nlwss ] ) . thus the values of the numbers @xmath281 and @xmath282 are determined by the values of the numbers of @xmath19 pseudoparticles and of @xmath18 pseudoparticles . the values of the numbers @xmath500 and @xmath501 were added to the local states ( [ nlwss ] ) because they are needed for the specification of the corresponding cphs ensemble subspace . in contrast , the values of the numbers @xmath502 and @xmath503 are dependent and given by @xmath504 and @xmath505 and are not explicitly provided in the expression of the local states ( [ nlwss ] ) . each of these normalized states is given by , @xmath506\,\bigl[\prod_{\alpha ' = c , s}\prod_{\nu'=1}^{\infty } 2^{\nu'\,n_{\alpha',\,\nu'}}\bigr]}\nonumber \\ & \times & \vert(l_{c,-1/2},\,l_{s,-1/2});\,(j_{1},\,j_2, ... ,j_{n_c});\ , \{(h_{1},\,h_2, ... ,h_{n_{c,\,\nu}})\};\,\{(l_{1},\,l_2, ... ,l_{n_{s,\,\nu}})\}\rangle \nonumber \\ & = & \bigl[\prod_{\alpha = c , s}\bigl({\hat{s}}^{\alpha}_{+}\bigr)^{l_{\alpha,\,-1/2}}\bigr]\ , \bigl[\prod_{\nu'=1}^{\infty}\prod_{j'=1}^{n_{c,\nu'}}\prod_{x=1}^{2\nu'}e^{i\pi h_{j',\,x}{\hat{d}}_{j',\,x}}\bigr]\,\bigl[\prod_{\alpha ' = c , s}\prod_{\nu''=1}^{\infty}\prod_{j''=1}^{n_{\alpha',\nu '' } } \prod_{g''=1}^{\nu''}(1-{\hat{\cal{t}}}_{\alpha',\,\nu'',\,j'',\,g''})\bigr]\nonumber \\ & \times & \vert(j_{1},\,j_2, ... ,j_{n_c});\ , \{(\bullet_{h_{j,\,1}}, ... ,\bullet_{h_{j,\,\nu } } , \,\circ_{h_{j,\,1+\nu } } , ... , \circ_{h_{j,\,2\nu}})\};\,\{(\downarrow_{l_{j,\,1 } } , ... , \downarrow_{l_{j,\,\nu}},\,\uparrow_{l_{j,\,1+\nu } } , ... , \uparrow_{l_{j,\,2\nu}})\}\rangle \ , , \label{completes}\end{aligned}\ ] ] where the operators @xmath507 with @xmath111 are the off - diagonal generators defined in eqs . ( [ sc ] ) and ( [ ss ] ) . importantly , application of these operators onto the local states given in eq . ( [ prima ] ) leaves the rotated - electron site distribution configurations of the @xmath18 pseudoparticles invariant and generates @xmath8-spin and spin flips in the @xmath89 yang holons and @xmath89 hl spinons respectively , as requested by property 5-iii . the set of all local states of general form given in eq . ( [ completes ] ) provides a complete basis of states in any cphs ensemble subspace of the hilbert space . however , these states do not ensure the periodic boundary conditions of the original electronic problem . such conditions are ensured by the construction of the energy eigenstates as fourier - transform superpositions of local states of general form given in eq . ( [ completes ] ) , as discussed in the ensuing section . in this section we finally reach a precise definition for the concept of effective @xmath18 pseudoparticle lattices . the precise definition of such a concept is a necessary condition for the construction of the energy eigenstates as fourier - transform superpositions of the local states introduced in the previous section . the concept of effective @xmath18 pseudoparticle lattice is related to the separation of the translational and internal degrees of freedom of these quantum objects , as we explain below . as mentioned in property 8-iii , in addition to the effective @xmath19 pseudoparticle lattice , there is an effective @xmath18 pseudoparticle lattice for each @xmath111 and @xmath112 pseudoparticle branch . the spatial coordinate associated with these lattices is the conjugate of the pseudoparticle band - momentum @xmath123 , which is a good quantum number of the many - electron problem . thus the fourier transforms which relate the local pseudoparticles introduced in this paper to the band - momentum pseudoparticles obtained from analysis of the bethe - ansatz solution in ref . @xcite involve the spatial coordinate of the effective pseudoparticle lattices . besides the introduction of the effective pseudoparticle lattices , in this section we also provide explicit expressions for the energy eigenstates in terms of slater determinants which involve fourier - transform superpositions of the local states introduced in the previous section . this clarifies the relation of these energy eigenstates to the rotated - electron site distribution configurations . indeed , analysis of the new found expressions of the energy eigenstates provides useful information on the processes involved in the diagonalization of the non - perturbative quantum problem after the electron - rotated - electron hilbert - space unitary rotation is performed . such a non - perturbative diagonalization leads to nothing but the energy eigenstates also reached by the bethe - ansatz solution . since the present problem is treated in a similar manner for the @xmath104 and @xmath105 pseudoparticles , for simplicity let us again concentrate namely on the case of the @xmath104 pseudoparticles . the separation of the @xmath18 pseudoparticle translational and internal degrees of freedom can be defined in terms of the following general properties which are obtained from the above properties 1-iii to 8-iii and 1-iv to 3-iv , and other features of the 1d hubbard model : 1-v one can separate the @xmath103-site internal structure of a local @xmath104 pseudoparticle from the problem of the description of the movements of these quantum objects in the effective electronic lattice . such separation leads to the concept of effective pseudoparticle lattice , whose spatial coordinates are associated with the pseudoparticle translational degrees of freedom only . according to the properties 1-iv to 4-iv , in addition to the @xmath103-leg caterpillar steps associated with the motion of a @xmath104 pseudoparticle , such a quantum object jumps each of the @xmath103-site domains of the remaining local pseudoparticles of the same branch . thus in the description of the motion of a local @xmath104 pseudoparticle , the @xmath103-site domains of these objects play the role of the occupied _ sites _ of an effective @xmath104 pseudoparticle lattice . from the point of view of the motion of these objects in such an effective @xmath104 pseudoparticle lattice these @xmath103-site domains are point - like sites without internal structure . roughly speaking , these point - like pseudoparticle occupied sites correspond to the _ center of mass _ of the @xmath103-site block associated with each local @xmath104 pseudoparticle . this description corresponds to a separation of the pseudoparticle translational and internal degrees of freedom . the latter degrees of freedom are described by the @xmath103-site rotated - electron distribution configurations studied in the previous section . such internal structure plays a key role both in the fulfilment of the @xmath104 pseudoparticle transformations under application of the off - diagonal generators of the @xmath12 @xmath8-spin algebra . in addition , such an internal structure is a necessary condition for the fulfilment of the periodic boundary conditions of the original electronic problem . on the other hand , the @xmath104 pseudoparticle translational degrees of freedom are closely related to the spatial coordinates of the effective pseudoparticle lattice . these spatial coordinates play the role of conjugate variable relative to the band momentum @xmath123 . such a band momentum obeys eq . ( [ pbcanp ] ) to ensure periodic boundary conditions for the original electronic problem . an important point is that the pseudoparticle spatial coordinate occupancy configuration of each @xmath104 pseudoparticle branch with finite occupancy in a given energy eigenstate is independently conserved . as the charge ( and spin ) sequence corresponds to the rotated - electron doubly occupied and empty sites ( and rotated - electron singly occupied sites ) only , one can also introduce an independent effective @xmath104 pseudoparticle lattice for each of these occupied pseudoparticle branches , whose coordinates correspond to some of the sites of the local charge sequence only . 2-v within the above separation of the pseudoparticle translational and internal degrees of freedom , the moving @xmath104 pseudoparticle _ sees _ the @xmath103-site domains of each of the local pseudoparticles belonging to the same branch as point - like occupied pseudoparticle sites . on the other hand , according to the property 3-iv a moving @xmath104 pseudoparticle jumps the @xmath421-site domains representative of each local @xmath219 pseudoparticles belonging to branches such that @xmath423 . this property can be understood as follows : since in its movements the @xmath104 pseudoparticle _ sees _ the @xmath103-site domains of the remaining @xmath104 pseudoparticles of the same branch as point - like unreachable occupied pseudoparticle sites , such a @xmath103-site domain width is the smallest width _ seen _ by the @xmath104 pseudoparticle . thus such a width plays the role of a site domain _ width uncertainty_. as a result , pseudoparticle site domains of width smaller than the @xmath103-site domain are not _ seen _ by the @xmath104 pseudoparticle . this is consistent with the exact property that the moving @xmath104 pseudoparticle does not _ see _ ( _ i.e. _ jumps ) smaller @xmath421-site @xmath219 pseudoparticle domains corresponding to branches such that @xmath484 . on the other hand , again according to property 3-iv , the @xmath103-leg caterpillar step movements of the @xmath104 pseudoparticle uses @xmath482 sites only , out of the @xmath421-site domain of each local @xmath219 pseudoparticle belonging to branches such that @xmath422 . thus , again consistently with the above site domain _ width uncertainty _ , in their motion throughout the effective electronic lattice , a local @xmath104 pseudoparticle does not _ see _ a number @xmath103 of sites out of the @xmath421 sites of such a @xmath422 local @xmath219 pseudoparticle . 3-v let us introduce the concepts of empty sites and lattice constant of the effective @xmath104 pseudoparticle lattice which follow from the above analysis . for the moving @xmath104 pseudoparticle all the @xmath417 charge - sequence _ empty sites _ corresponding to a number @xmath281 of @xmath102 yang holons and a number @xmath508 of sites belonging to the @xmath421-site domains of local @xmath219 pseudoparticles of the charge sequence such that @xmath422 are seen as equivalent and indiscernible point - like empty sites of a _ effective @xmath104 pseudoparticle lattice_. moreover , according to the property 1-v such a quantum object also _ sees _ the @xmath328 @xmath103-site domains of pseudoparticles belonging to the @xmath104 pseudoparticle branch as equivalent and indiscernible point - like occupied sites . thus while running through all its possible positions in the effective electronic lattice the moving @xmath104 pseudoparticle _ sees _ all the @xmath509 sites of the _ effective @xmath104 pseudoparticle lattice _ as equivalent and indiscernible point - like sites . therefore , these sites are equally spaced for the effective @xmath104 pseudoparticle lattice . the value of the corresponding lattice constant is uniquely defined as follows : in its movements a @xmath104 pseudoparticle _ does not see _ ( i ) the @xmath510 sites occupied by @xmath219 pseudoparticles belonging to branches such that @xmath423 ; ( ii ) a number @xmath511 of sites belonging to the @xmath421-site domains of @xmath219 pseudoparticles belonging to branches such that @xmath422 ; and ( iii ) the @xmath85 sites singly occupied by rotated electrons . thus the the @xmath104 pseudoparticle _ jumps _ all the above sites of the effective electronic lattice and _ sees _ the @xmath103-site pseudoparticle domains of the @xmath104 pseudoparticle branch as point - like occupied sites . on the other hand , in order to pass all the @xmath509 sites of the effective @xmath104 pseudoparticle lattice and return to its original position , the @xmath104 pseudoparticle should run through a distance which equals the length @xmath255 of the real - space and effective electronic lattice . therefore , a necessary condition to ensure the periodic boundary conditions of the original electronic problem is that the length of the effective @xmath104 pseudoparticle lattice must equal the length @xmath255 of the effective electronic lattice . this determines uniquely the value of the lattice constant of the effective @xmath104 pseudoparticle lattice which reads @xmath512 . that the length of the effective @xmath104 pseudoparticle lattice is @xmath255 is consistent with the spacing @xmath513 given in eq . ( [ qj1j ] ) for the corresponding pseudoparticle discrete band - momentum values provided by the exact bethe - ansatz solution @xcite . such a discrete band momentum @xmath123 is the conjugate of the coordinate @xmath514 of the effective @xmath104 pseudoparticle lattice where @xmath299 . 4-v a similar analysis is valid for the local @xmath105 pseudoparticles , provided that the above mentioned @xmath85 sites singly occupied by rotated electrons are replaced by the @xmath292 $ ] sites doubly occupied by rotated electrons and free of rotated electrons . the lattice constant of the effective @xmath105 pseudoparticle lattice is given by @xmath515 . on the other hand , the effective @xmath19 pseudoparticle lattice has @xmath5 lattice sites and its lattice constant @xmath3 is the same as for the effective electronic lattice . the positions of the @xmath85 @xmath19 pseudoparticles and @xmath138 $ ] @xmath19 pseudoparticle holes in this effective lattice equal the corresponding positions of the rotated - electron singly occupied sites and rotated - electron doubly occupied / empty sites , as already discussed in previous sections . 5-v the set of pseudoparticle occupancy configurations of the effective @xmath19 pseudoparticle , @xmath104 pseudoparticle , and @xmath105 pseudoparticle lattices belonging to branches @xmath274 with finite occupancy in a given local state of form ( [ completes ] ) together with the numbers @xmath117 and @xmath118 of such a state contain the same information as the corresponding description of the same local state in terms of the rotated - electron site distribution configurations . since the value of the number @xmath128 given in eqs . ( [ n*sum ] ) and ( [ nhag ] ) is distinct for different cphs ensemble subspaces , it follows from the expression of the effective @xmath18 pseudoparticle lattice constants , @xmath516 that the value of such constants changes accordingly . let us introduce the following notation for the spatial coordinates of the effective @xmath19 pseudoparticle lattice , @xmath517 where the index @xmath36 was called @xmath288 and @xmath289 in eq . ( [ chcl ] ) for the @xmath19 pseudoparticle occupied sites ( @xmath518 ) and empty sites ( @xmath519 $ ] ) respectively . we recall that the number of these occupied and empty sites of the effective @xmath19 pseudoparticle lattice equals the number of rotated - electron singly occupied sites and rotated - electron doubly occupied / empty sites respectively . moreover , let us introduce the following notation for the spatial coordinates of the effective @xmath520 pseudoparticle and @xmath105 pseudoparticle lattices , @xmath521 and @xmath522 respectively . the band - momentum limiting values given in eq . ( [ qag ] ) for the @xmath18 pseudoparticle bands can be expressed in terms of the corresponding lattice constants @xmath523 as follows , @xmath524 \approx { \pi\over a_{\alpha , \,\nu } } \ , . \label{qagaan}\ ] ] also the @xmath19 pseudoparticle limiting band - momentum value ( [ qc ] ) can be written as , @xmath525 \approx { \pi\over a } \ , . \label{qceac}\ ] ] thus the domain of available pseudoparticle band - momentum values corresponds to an effective _ first - brillouin zone _ associated with an underlying effective pseudoparticle lattice . the fact that the spatial coordinate introduced in eqs . ( [ xc ] ) , ( [ xe ] ) , and ( [ xg ] ) is the conjugate of the pseudoparticle band momentum given in eq . ( [ qj ] ) is used in refs . @xcite in fourier analysis involving pseudoparticle creation and annihilation operators . the ground - state plays an important role in the study of finite - energy few - electron spectral functions @xcite . thus let us use the ground - state eqs . ( [ n*csnu ] ) and ( [ qcangs ] ) in order to find the corresponding values for the effective @xmath18 pseudoparticle lattice constants in the case of electronic densities @xmath165 and spin densities @xmath166 such that @xmath526 and @xmath527 respectively . we find that in the case of a ground state corresponding to an electronic density and a spin density whose values are within these domains these constants read , @xmath528 where @xmath529 $ ] is the doping concentration away from half filling . moreover , the ground - state number @xmath181 given in eq . ( [ n*csnu ] ) and the ground - state limiting band - momentum values ( [ qcangs ] ) can be written in terms of the effective pseudoparticle lattice constants as follows , @xmath530 and @xmath531 respectively . we note that when the effective pseudoparticle lattice constants given in eq . ( [ acangs ] ) diverge , as is the case for @xmath532 as @xmath529\rightarrow 0 $ ] and for @xmath533 as @xmath534\rightarrow 0 $ ] when @xmath535 , the corresponding number @xmath181 is zero . this just means that in these limits these bands and corresponding effective lattices do not exist for states belonging to the ground - state cphs ensemble subspace . finally , let us relate the new introduced effective pseudoparticle lattices to previous results on the 1d hubbard model in the limit @xmath16 . indeed the effective @xmath19 and @xmath277 pseudoparticle lattices introduced here are related to known properties of the model in that limit . for instance , it is well known that in such a limit and at zero spin density the charge and spin degrees of freedom of the 1d hubbard model can be described by two independent systems of @xmath536 spin - less fermions and @xmath537 spin - down spins @xcite . the spin - less fermions can be associated with an effective lattice with @xmath30 sites , whereas the @xmath537 spin - down spins correspond to a squeezed effective lattice with @xmath538 sites . the spin - less fermion and spin occupancy configurations of these effective lattices describe electron occupancy configurations of the associated real - space lattice . in the limit @xmath16 such spin - less fermion and spin effective lattices are directly related to the effective @xmath19 pseudoparticle and @xmath277 pseudoparticle lattices respectively , introduced above . on the other hand , within our pseudoparticle description of the quantum problem these effective pseudoparticle lattices are well defined for all finite values of @xmath14 . in that case the pseudoparticle occupancy configurations of these effective lattices describe rotated - electron site distribution configurations . in the limit @xmath16 the energy eigenstates with finite occupancies in the effective @xmath104 pseudoparticle lattices have finite values of electron double occupation . thus according to the energy spectrum given in eq . ( [ ehuinf ] ) these states have in such a limit an infinite excitation energy relative to the ground state and do not contribute to the finite - energy physics . this explains why these states are unimportant for the @xmath16 physics and are in general not considered in that limit @xcite . at zero spin density the effective @xmath105 pseudoparticle lattices such that @xmath279 do not exist for the ground state because @xmath539 for such a state . however , at zero spin density excitations with finite occupancy of @xmath105 pseudoparticles belonging to branches such that @xmath279 have a gapless energy spectrum . in the limit @xmath16 the spin excitations involving both occupancy configurations of @xmath277 pseudoparticles and @xmath105 pseudoparticles belonging to branches such that @xmath279 are often described by the isotropic heisenberg chain which describes these excitations @xcite . our last goal is the construction of the energy eigenstates in terms of the local charge , spin , and @xmath19 pseudoparticle sequences studied in previous sections . the energy eigenstates are described in terms of band - momentum pseudoparticle , yang holon , and hl spinon occupancy configurations in ref . @xcite by direct use of the bethe - ansatz solution and @xmath25 symmetry of the model . in the previous section we expressed the local states given in both eqs . ( [ localst])-([args ] ) and eq . ( [ prima ] ) in terms of rotated - electron site distribution configurations and showed that they provide a complete basis of states . in order to relate these local states to the energy eigenstates we should first express the former states in terms of local @xmath19 and @xmath18 pseudoparticle occupancy configurations in the corresponding effective pseudoparticle lattices . we then fourier transform the obtained states into band - momentum space @xmath123 with respect to the spatial coordinates of the effective @xmath19 and @xmath18 pseudoparticle lattices . as in the case of eq . ( [ chcl ] ) for the case of the effective @xmath19 pseudoparticle lattice , we denote by @xmath288 and @xmath289 where @xmath540 and @xmath541 respectively , the occupied and empty sites respectively , of the effective @xmath18 pseudoparticle lattices . the spatial coordinate of these occupied and empty pseudoparticle sites is given by , @xmath542 where @xmath288 is the index which defines the position of the occupied pseudoparticle sites and @xmath289 the location of the empty pseudoparticle sites . we note that the empty sites of the effective pseudoparticle lattices correspond to the sites left over by the occupied sites and vice versa . thus we can uniquely specify a given effective pseudoparticle lattice site distribution configuration by providing the location of the occupied sites ( or empty sites ) only . here we choose the representation in terms of the locations in the effective @xmath19 and @xmath18 pseudoparticle lattices of the occupied sites @xmath543 of eqs . ( [ chcl ] ) and ( [ aeloe ] ) . in such a representation the states given in eq . ( [ completes ] ) are denoted as , @xmath544 here @xmath545 is the set of spatial coordinates associated with the set of indices @xmath546 given in eq . ( [ lcp ] ) . these spatial coordinates correspond to the set of sites occupied by @xmath19 pseudoparticles in the corresponding effective lattice . moreover , @xmath547 gives the local pseudoparticle coordinates of each of the effective @xmath18 pseudoparticle lattices with finite occupancy in the associated effective electronic lattice . we emphasize that from analysis of the specific rotated - electron site distribution configurations relative to a given state of form ( [ prima ] ) one can construct by use of the above properties 1-v to 3-v the @xmath19 pseudoparticle , @xmath104 pseudoparticle , and @xmath105 pseudoparticle occupancy configurations in the effective @xmath19 , @xmath104 , and @xmath105 pseudoparticle lattices respectively , of the corresponding state of form ( [ ellss ] ) . this relation was already mentioned in the above property 5-v and the inverse statement is obviously also true . ( [ prima ] ) and ( [ ellss ] ) refer to two different representations of the same local states . these local states constitute a complete set of states . however , the form of these local states does not ensure the periodic boundary conditions of the original electronic problem . the representation of the local states in terms of the effective pseudoparticle lattice occupancy configurations given in eq . ( [ ellss ] ) is the most suitable starting point for construction of the energy and momentum eigenstates . such a construction is fulfilled by fourier transforming the local states given in eq . ( [ ellss ] ) into band - momentum space with respect to the spatial coordinates of the effective pseudoparticle lattices given in eqs . ( [ xc ] ) , ( [ xe ] ) , and ( [ xg ] ) . provided that the discrete band - momentum values obey the boundary conditions associated with eqs . ( [ pbccp ] ) and ( [ pbcanp ] ) , the form of the obtained states ensures the periodic boundary conditions of the original electronic problem . such procedures lead to the following expression of the energy eigenstates in terms of the local states of eq . ( [ ellss ] ) , @xmath548^{n_{\alpha,\,\nu'}/2}\bigr)\nonumber \\ & \times & \vert ( l_{c,-1/2},\,l_{s,-1/2});\,(q_{j_1},\,q_{j_2}, ... ,q_{j_{n_c}});\ , \{(q_{j_1},\,q_{j_2}, ... ,q_{j_{n_{c,\,\nu}}})\};\,\{(q_{j_1},\,q_{j_2}, ... ,q_{j_{n_{s,\,\nu}}})\}\rangle \nonumber \\ & = & \sum_{j_{1}<j_2< ... <j_{n_c}}\sum_{{\cal{p}}}\,(-1)^{\cal{p } } \,e^{(ia\sum_{l=1}^{n_c}j_{{\cal{p}}(l)}\,q_{j_{l}})}\nonumber \\ & \times & \bigl[\prod_{\nu'=1}^{\infty}\,\bigl(\sum_{j_{1}<j_2< ... <j_{n_{c,\,\nu'}}}\sum_{{\cal{p}}}\,(-1)^{\cal{p } } \,e^{(ia_{c,\,\nu}\sum_{l=1}^{n_{c,\,\nu'}}j_{{\cal{p}}(l)}\,[{\pi\over a } - q_{j_{l}}])}\bigr)\bigr]\nonumber \\ & \times & \bigl[\prod_{\nu''=1}^{\infty}\,\bigl(\sum_{j_{1}<j_2< ... <j_{n_{s,\,\nu''}}}\sum_{{\cal{p}}}\,(-1)^{\cal{p } } \,e^{(ia_{s,\,\nu}\sum_{j'=1}^{n_{s,\,\nu''}}j_{{\cal{p}}(l)}\,q_{j_{l}})}\bigl)\bigr]\nonumber \\ & \times & \vert(l_{c,-1/2},\,l_{s,-1/2});\,(x_{j_{1}},\,x_{j_2}, ... ,x_{j_{n_c}});\,\{(x_{j_{1}},\,x_{j_2 } , ... , x_{j_{n_{c,\,\nu}}})\ } ; \,\{(x_{j_{1}},\,x_{j_2}, ... ,x_{j_{n_{s,\,\nu}}})\}\rangle \ , . \label{wffin}\end{aligned}\ ] ] the permutations @xmath549 on the right - hand side of eq . ( [ wffin ] ) generate a slater determinant of the band momenta and the spatial coordinates of the @xmath85 @xmath19 pseudoparticles and @xmath158 @xmath18 pseudoparticles belonging to branches @xmath550 and @xmath112 with finite occupancy in the corresponding local charge and spin sequences . on the left - hand side of eq . ( [ wffin ] ) the set of pseudoparticle occupied band - momentum values specify the energy eigenstate . there @xmath551 are the set of @xmath85 occupied band momentum values out of the @xmath5 available discrete @xmath123 values of the @xmath19 pseudoparticle band such that @xmath552 . moreover , @xmath553 are the set of @xmath158 occupied band momentum values out of the @xmath181 available discrete @xmath123 values of the @xmath18 pseudoparticle bands of branches @xmath550 and @xmath274 such that @xmath554 . such band - momentum pseudoparticle occupancy configurations are discussed in ref . @xcite . since equations ( [ prima ] ) and ( [ ellss ] ) refer to two different representations of the same states , the energy eigenstates ( [ wffin ] ) are fourier - transform superpositions involving permutations of the local charge - sequence and spin - sequence states ( [ ellss ] ) . the latter local states correspond to rotated - electron site distribution configurations . thus combination of the general expressions ( [ completes ] ) and ( [ wffin ] ) provides important information about the relationship between the band - momentum pseudoparticles and the rotated - electron occupancy configurations . we emphasize that the permutations on the right - hand side of eq . ( [ wffin ] ) are associated with different locations for the local @xmath18 pseudoparticles but that the internal structure of these quantum objects remains invariant under these permutations . both the fourier transforms and permutations of the general expression ( [ wffin ] ) do not touch the internal structure of the local @xmath18 pseudoparticles and refer only to their translational degrees of freedom . this results from the point - like character of the occupied sites of the effective pseudoparticle lattices . thus the local and band - momentum @xmath18 pseudoparticles have the same internal structure . while the complete set of symmetrized states used in ref . @xcite are energy eigenstates of the 1d hubbard model in the limit @xmath16 only , the states given in eqs . ( [ wffin ] ) and ( [ eanqj ] ) are energy eigenstates of that model for all values of @xmath14 . the local charge , spin , and @xmath19 pseudoparticle sequences defined in expressions ( [ prima ] ) and ( [ completes ] ) and used in the construction of the latter states , describe rotated - electron site distribution configurations which in the limit @xmath16 refer to similar electron configurations . thus in contrast to the complete set of symmetrized energy eigenstates used in ref . @xcite , the complete set of energy eigenstates of form ( [ wffin ] ) are in the limit @xmath16 the band - momentum states associated with the unitary operator @xmath90 . in the case of the energy eigenstates given in eq . ( [ wffin ] ) , a first condition for the fulfilment of the periodic boundary conditions of the original electronic problem is that in all the local charge and spin sequences described by the states ( [ completes ] ) , the local @xmath104 pseudoparticles and @xmath105 pseudoparticles are described by the same rotated - electron @xmath103-site distribution configurations given in eqs . ( [ cnp ] ) and ( [ snp ] ) respectively . these are obtained from the proper symmetrization of the corresponding single pseudoparticle @xmath103-site sequences . a second condition to ensure these periodic boundary conditions is that energy eigenstates given in eq . ( [ wffin ] ) are fourier transform superpositions of these charge and spin sequences of the form given on the right - hand side of that equation . provided that the discrete values of the band momentum @xmath123 of these slater determinant superpositions obey eqs . ( [ pbccp ] ) and ( [ pbcanp ] ) , the periodic boundary conditions of the original electronic problem are ensured . note that the general expression provided in eq . ( [ wffin ] ) includes suitable permutations of the local sequences . in the same equation the rotated - electron site distribution configurations of the local charge and spin sequences associated with the states ( [ prima ] ) or ( [ completes ] ) are reexpressed in terms of pseudoparticle occupancy configurations of the corresponding effective pseudoparticle lattices . let us consider the specific case of a ground state with electronic densities and spin densities belonging to the domains @xmath555 and @xmath168 respectively . such a state is a superposition of local normalized states ( [ completes ] ) which have no finite occupancies of @xmath9 yang holons , @xmath9 hl spinons , @xmath104 pseudoparticles , and @xmath105 pseudoparticles belonging to branches such that @xmath279 @xcite . thus in the case of such a ground state these local states are of the following simplified form , @xmath556\,\vert(0,\,0);\,(j_{1},\,j_2, ... ,j_{n_c } ) ; \,\{(\downarrow_{l_{j',\,1}},\,\uparrow_{l_{j',\,2}})\}\rangle \ , , \label{localgs}\end{aligned}\ ] ] where @xmath557 and the operators @xmath558 are the same as in the general eq . ( [ completes ] ) . the states defined in eqs . ( [ localgs ] ) and ( [ arggsl ] ) provide the position of the two rotated - electron singly occupied sites associated with the internal structure of each local @xmath277 pseudoparticle . these local states have an alternative representation in terms of the @xmath19 and @xmath277 pseudoparticle occupancy configurations of the effective @xmath19 and @xmath277 pseudoparticle lattices respectively . within such an effective pseudoparticle lattice representation the local states ( [ localgs ] ) are denoted by @xmath559 . here @xmath560 are the spatial coordinates which define the positions of the local @xmath277 pseudoparticles in the effective @xmath277 pseudoparticle lattice . these spatial coordinates correspond to the indices @xmath561 on the left - hand side of eq . ( [ localgs ] ) . the sites of such an effective @xmath277 pseudoparticle lattice are equally spaced , the corresponding ground - state lattice constant @xmath562 being given in eq . ( [ acangs ] ) . the ground state expression is a particular case of the fourier - transform superposition of local states given in eq . ( [ wffin ] ) . in the case of the ground - state such an expression simplifies to , @xmath563 where @xmath564 , @xmath565 , and @xmath566 . the concepts of _ local pseudoparticle _ and of _ effective pseudoparticle lattice _ introduced in this paper are applied in the studies of refs . @xcite to the evaluation of finite - energy few - electron spectral function expressions . in this paper we identified the conjugate variable of the pseudoparticle band momentum obtained naturally from the bethe - ansatz solution in ref . @xcite . such a conjugate variable is nothing but the spatial coordinate of the effective pseudoparticle lattice . we found that there is one of such lattices for each pseudoparticle branch . moreover , we studied the pseudoparticle site distribution configurations which describe the energy eigenstates . we have also related these configurations to the rotated - electron site distribution configurations which describe the same states . since electrons and rotated electrons are related by a mere unitary transformation , our results provide useful information about the relation between electronic excitations and energy eigenstates . the results obtained in this paper are used in refs . @xcite in the evaluation of the overlap between few - electron excitations and energy eigenstates . as discussed in ref . @xcite , the pseudoparticles studied here have residual interactions . in that reference it is found that a canonical unitary transformation maps the pseudoparticles onto non - interacting _ pseudofermions_. ( this transformation is other than the electron - rotated - electron canonical unitary transformation associated with the operator ( [ sv ] ) . ) interestingly , the latter quantum objects are more suitable for the evaluation of matrix elements between the ground state and excited states . the pseudoparticle - pseudofermion unitary transformation maps the @xmath19 pseudoparticle ( and @xmath18 pseudoparticle ) discrete band momentum value @xmath123 onto the @xmath19 pseudofermion ( and @xmath18 pseudofermion ) discrete _ shifted momentum _ value @xmath567 ( and @xmath568 ) . here @xmath569 and @xmath570 are momentum functionals which depend on the excited state occupancy configurations and involve the two - pseudofermion phase shifts . although the momenta @xmath571 and @xmath572 are of order @xmath185 and play a crucial role in the spectral properties of the quantum problem , the above transformation is such that the differences @xmath573/l$ ] and @xmath574/l$ ] are of order @xmath575 ^ 2 $ ] and thus vanish within our large-@xmath255 description . therefore , to first order in @xmath185 the shifted - momentum values are such that @xmath576 , like the corresponding band - momentum values . it follows that there is no level crossing between the sequence of @xmath577 discrete values and the corresponding sequence of @xmath578 discrete values such that @xmath228 and @xmath579 for the @xmath19 and @xmath18 branches respectively . moreover , the presence of the extra momentum term @xmath571 or @xmath572 in the shifted - momentum expression does not affect the underlying effective @xmath20 pseudoparticle lattice . indeed , that momentum just imposes a twisted boundary condition such that each @xmath19 pseudofermion ( and @xmath20 pseudofermion ) hopping from site @xmath5 to site @xmath247 ( and site @xmath128 to site @xmath247 ) of such an effective lattice will acquire a phase @xmath580 ( and @xmath581 ) . thus the effective pseudoparticle lattice introduced here is invariant under the pseudoparticle - pseudofermion transformation studied in ref . @xcite . in the case of the hilbert subspace where the few - electron excitations are contained there is an one - to - one correspondence between the pseudofermion shifted - momentum occupancy configurations ( and pseudofermion site distribution configurations ) and the corresponding pseudoparticle band - momentum occupancy configurations ( and pseudoparticle site distribution configurations ) that describe the energy eigenstates . thus the results found in the present paper are useful for the description of these states in terms of pseudofermion occupancy configurations as well . since the pseudofermions are non - interacting , it is found in ref . @xcite that the wave function of the excited states of the ground - state normal - ordered 1d hubbard model factorizes for all values of @xmath14 . importantly , such a factorization is used in ref . @xcite , where the few - electron spectral functions are expressed as a convolution of single pseudofermion spectral functions . this factorization involves the pseudofermion branches with finite occupancy in the excited states . thus due to such a factorization the problem of the evaluation of finite - energy few - electron spectral functions is reduced to the problem of the evaluation of pseudofermion spectral functions . the latter problem involves the derivation of matrix elements between the ground state and excited states for pseudofermion operators . the effective pseudoparticle lattices introduced in this paper play an important role in the evaluation of the pseudofermion matrix elements . it turns out that the derivation of the above matrix elements involves the anticommutator @xmath582 where the operators @xmath583 and @xmath584 create and annihilate respectively , a @xmath232 pseudofermion of sifted momentum @xmath585 , and similar operators can be introduced for the @xmath19 pseudofermions . we note that because of the functional character of the momenta @xmath569 and @xmath570 these anticommutation relations have not the usual fermionic form . this justifies the designation _ pseudofermion_. furthermore , without the concept of local pseudoparticle and effective pseudoparticle lattice introduced in this paper the evaluation of these anticommutators would be a complex open problem . however , our results imply the existence of _ local pseudofermions _ which correspond to the local pseudoparticles introduced in this paper . fortunately both the invariance of the pseudoparticle effective lattice under the pseudoparticle - pseudofermion transformation and the associated concept of local pseudofermion allows the expression of the above anticommutators in terms of the local pseudofermion operator anticommutators @xmath586 associated with spatial coordinates @xmath254 and @xmath587 of the effective @xmath20 and @xmath588 lattices respectively . such an expression is very useful for the evaluation of the pseudofermion matrix elements between the ground state and excited states @xcite . as in the case of the pseudoparticles , also the shifted - momentum and local pseudofermion descriptions are related by a simple fourier transform @xcite . thus one can express the above shifted - momentum pseudofermion operator anticommutators in terms of the local pseudofermion operator anticommutators as follows , @xmath589 as a result of the introduction of the concept of effective pseudoparticle ( and pseudofermion ) lattice the derivation of the local operator anticommutators on the right - hand side of this equation is trivial and expressions for the band - momentum anticommutators operators ( [ pfacrjj ] ) can be easily obtained by performing the @xmath36 and @xmath590 summations @xcite . corresponding expressions are found for the @xmath19 pseudofermions such that , @xmath591 and the anticommutators involving @xmath18 pseudofermions and @xmath19 pseudofermions vanish . ( also the anticommutators given in eq . ( [ pfacrjj ] ) vanish for @xmath592 @xcite . ) we note that the pseudofermion description introduced in ref . @xcite is a generalization for all values of @xmath14 of the well known @xmath16 quantum - object description used in refs . @xcite . for instance , in the limit @xmath16 the @xmath19 pseudofermions become the spin - less fermions of refs . @xcite and the spinons become the spins of these references . the @xmath277 pseudofermions describe a spin pair of opposite projections . also the remaining @xmath105 pseudofermions describe @xmath103 spins , @xmath107 of each direction . in the limit @xmath16 the general @xmath19 pseufofermion anticommutation relations ( [ pfacrjjc ] ) become nothing but the spin - less fermion anticommutation relations provided in the first ( unnumbered ) equation of sec . iv of ref . @xcite . as discussed in ref . @xcite and confirmed in ref . @xcite , except for the @xmath14 dependence of the two - pseudofermion phase shifts involved in the general expressions of the momentum functionals @xmath569 and @xmath570 , the evaluation of the pseudofermion matrix elements for finite values of @xmath14 is formally similar to the derivation of the spin - less fermion matrix elements presented in ref . @xcite for @xmath16 . this similarity results from the facts that both the energy - eigenstate rotated - electron occupancy configurations and corresponding pseudoparticle , holon , and spinon occupancy configurations introduced in this paper are independent of the value of @xmath14 and that in the limit @xmath16 considered in ref . @xcite electrons and rotated electrons are the same quantum objects . this example reveals how the local pseudoparticle ( and pseudofermion ) representation introduced in this paper is used for practical calculations at finite @xmath14 . the only formal difference between the finite @xmath14 studies of refs . @xcite and those of ref . @xcite for @xmath16 is that in the former references the evaluation of the spectral function for each pseudofermion branch proceeds as in latter reference for the spin - less fermions . in contrast , the studies of ref . @xcite use the well known @xmath16 connection between the spin degrees of freedom of the 1d hubbard model and the spin @xmath593 isotropic heisenberg chain @xcite to evaluate the spin part of the one - electron spectral function . such direct connection does not exist for finite values of @xmath14 . however , the method used in refs . @xcite for finite values of @xmath14 is also valid for @xmath16 and leads to the same final results as the analysis of ref . @xcite . in this paper we studied the relation between the description of the energy eigenstates of the 1d hubbard model in terms of rotated - electron occupancy configurations and of yang holon , hl spinons , and pseudoparticle band momentum occupancy configurations . such a relation involves the concepts of _ local pseudoparticle _ and _ effective pseudoparticle lattices_. we introduced the pseudoparticle spatial coordinates @xmath254 associated with and conjugate of the pseudoparticle band - momentum @xmath123 . the band - momentum pseudoparticle description is provided by the bethe - ansatz takahashi s thermodynamic equations @xcite . the pseudoparticle spatial coordinates introduced in this paper correspond to effective pseudoparticle lattices whose length and lattice constant are independent of the value of @xmath14 . the introduction of the concept of pseudoparticle effective lattice involved the study rotated - electron distribution configurations of doubly occupied and empty sites ( and of spin - down and spin - up singly occupied sites ) which describe the internal structure of the local @xmath104 pseudoparticles ( and local @xmath105 pseudoparticles ) . in what the translational degrees of freedom are concerned , the local @xmath18 pseudoparticle is a point - like quantum object , its internal structure being the same as that of the corresponding band - momentum pseudoparticle . we also found that the spatial coordinates of the occupied and empty sites of the effective @xmath19 pseudoparticle lattice are the same as the coordinates of the sites singly occupied by rotated electrons and rotated - electron doubly occupied / empty sites respectively . the energy eigenstates can be described in terms of pseudoparticle site distribution configurations in the corresponding effective pseudoparticle lattice . our results reveal that there is an one - to - one correspondence between the local pseudoparticle site distribution configurations in the effective pseudoparticle lattices which describe a given energy eigenstate and the rotated - electron site distribution configurations which describe the same state . as mentioned in sec . ii , the electron site distribution configurations which describe the energy eigenstates are very complex and dependent on the value of @xmath14 . this is confirmed by the @xmath14 dependence of the double occupation expectation value studied in ref . @xcite . however , the electron - rotated - electron canonical unitary transformation is such that it maps these complex and @xmath14 dependent electron site distribution configurations onto the corresponding @xmath14 independent rotated - electron site distribution configurations studied in this paper . this is an important property of the non - perturbative diagonalization of the 1d hubbard model . according to the results obtained in this paper , this property implies that the local pseudoparticle site distribution configurations of the effective pseudoparticle lattices which describe the energy eigenstates are also independent of the value of @xmath14 . moreover , since the spatial coordinate of these effective lattices is the conjugate of the pseudoparticle band - momentum obtained from analysis of the bethe - ansatz solution @xcite , it is again such a property which justifies why the pseudoparticle band - momentum occupancies which describe the energy eigenstates are independent of the values of @xmath14 . the invariance associated with the commutation of the momentum operator with the electron - rotated - electron unitary operator is related to the @xmath14 independence of the band - momentum pseudoparticle , rotated - electron , and local pseudoparticle occupancy configurations which describe the energy eigenstates . in this paper we profitted from such a @xmath14 independence of the pseudoparticle and rotated - electron occupancy configurations and extracted some of our results from analysis of the @xmath16 problem . in that limit the quantum problem simplifies because the @xmath8-spin and spin occupancy configurations are degenerated and electrons and rotated electrons are the same quantum object and thus electron double occupation is a good quantum number . we also provided preliminary information about the application of the concepts of local pseudoparticle and effective pseudoparticle lattice introduced in this paper to the evaluation of the spectral properties of few - electron operators . the study of the finite - energy few - electron spectral functions is required for the further understanding of the unusual finite - energy / frequency spectral properties observed in quasi-1d materials @xcite , which are far from being well understood . indeed there are clear indications that electronic correlations effects might play an important role in the finite - energy physics of these low - dimensional materials @xcite . for low values of the excitation energy the microscopic electronic properties of these materials are usually described by systems of coupled chains . on the other hand , for finite values of the excitation energy larger than the transfer integrals for electron hopping between the chains , 1d lattice models like the 1d hubbard model are expected to provide a good description of the physics of these materials . the studies of the present paper are a first necessary step for the evaluation of few - electron spectral function expressions for finite values of excitation energy , as confirmed in refs . @xcite . as mentioned in sec . i , application of a preliminary version of the theory introduced here and in these references to the study of the one - electron removal spectral properties of quasi-1d materials leads to a quantitative and successful description of the exotic independent charge and spin spectral lines observed in photoemission experiments for all values of excitation energy @xcite . we thank nuno peres for useful discussions related to the occupancy configurations represented in the figures of this paper and for valuable support in the production of these figures . we also thank jim w. allen , antnio castro neto , ralph claessen , ricardo dias , francisco ( paco ) guinea , katrina hibberd , peter horsch , lus miguel martelo , karlo penc , and pedro sacramento for stimulating discussions . in this appendix we review some aspects of the large @xmath14 physics that are useful for the contextualization of the problems studied in this paper . we start by considering the simple 1d hubbard hamiltonian given in eq . ( [ hh ] ) which differs from the hamiltonians given in eqs . ( [ h ] ) and ( [ hso4 ] ) in the definition of the zero - energy level . the energy spectrum of the stationary states of the model ( [ hh ] ) can written as a functional of the pseudoparticle band - momentum distribution functions and @xmath9 yang holon numbers such that @xcite , @xmath594 this energy functional depends on these distribution functions through the momentum - rapidity functional @xmath595 and rapidity functional @xmath596 @xcite . the momentum spectrum @xmath597 of the energy eigenstates has a simpler expression , being the linear superposition of the pseudoparticle band - momentum distribution functions and @xmath9 yang holon number . it reads , @xmath598 + { \pi\over a}\,m_{c,\,-1/2 } \ , . \label{pp}\end{aligned}\ ] ] as in the case of the momentum values of the quasi - particles of a fermi liquid , the momentum @xmath597 given in eq . ( [ pp ] ) is additive relative to the band - momentum values of the pseudoparticles . the momentum rapidity functional @xmath595 and the rapidity functional @xmath599 on the right - hand side of eq . ( [ eh ] ) can be defined in terms of the takahasi s thermodynamic equations @xcite rewritten in functional form @xcite . these quantities are functionals of the set of band - momentum distribution functions @xmath600 and @xmath601 , where @xmath602 and @xmath274 , whose possible different occupancy configurations classify the energy eigenstates . in the limit @xmath16 these functionals have the same value for all energy eigenstates and read @xmath603 and @xmath604 . thus in such a limit the energy and momentum expressions ( [ eh ] ) and ( [ pp ] ) respectively , simplify to expression ( [ ehuinf ] ) and @xmath598 + { \pi\over a}\,d \ , , \label{ppuinf}\end{aligned}\ ] ] respectively , where the electron double occupation @xmath196 is a good quantum number such that @xmath605 in that limit . for finite values of @xmath14 the latter relation is replaced by @xmath606 , where @xmath7 is the rotated - electron double occupation . according to eq . ( [ mas ] ) the number @xmath607 of @xmath9 holons can be written as @xmath608 in the case of zero spin density , electronic densities @xmath167 , and on - site repulsion @xmath609 the ground state energy @xmath610 of the 1d hubbard model is given by @xcite , @xmath611 and the ground state electron double occupation @xmath612 can be written as @xcite , @xmath613 here the term @xmath614 is the kinetic energy associated with the hopping processes which do not change electron double occupation . on the other hand , the energy term @xmath615 includes both kinetic and potential energy contributions . it arises from hopping processes which change electron double occupation and lead to the ground - state electron double occupation value given in eq . ( [ d0 ] ) . thus the physics associated with this second energy term corresponds to excitation processes of higher order in @xmath17 relatively to the @xmath239 limit where electron double occupation is a good quantum number and the ground - state electron double occupation of the 1d hubbard model is exactly zero . if we include energy contributions of the order @xmath616 the ground state contains a small but finite electron double occupation expectation value given in eq . ( [ d0 ] ) and the corresponding quantum problem is not equivalent to the physical situation of interest for the rotated - electron studies of this paper . indeed , electrons equal rotated electrons when the limit @xmath239 is reached and double occupation is a good quantum number . only in that limit do the electron occupancy configurations which describe the band - momentum energy eigenstates equal the corresponding rotated - electron configurations which are valid for all values of @xmath14 . note that at half filling the electronic density reads @xmath617 , the energy term of order @xmath618 on the right - hand side of eq . ( [ e0 ] ) vanishes , and the ground - state energy and electron double occupation expressions given in eqs . ( [ e0 ] ) and ( [ d0 ] ) simplify to @xmath619 and @xmath620 respectively . it is well known that this energy can be associated with an isotropic heisenberg model @xcite . the corresponding ground state leads to energy contributions of the order @xmath616 and thus contains a small but finite electron double occupation expectation value , @xmath621 . it follows that the usual description of the large-@xmath14 half - filling hubbard model in terms of an isotropic heisenberg model is not equivalent to our limit . in that limit only hopping processes which do not change the value of electron double occupation should be considered . to leading order in the parameter @xmath17 , the energy spectrum of the 1d hubbard model in the limit @xmath239 is of the form given in eq . ( [ ehuinf ] ) . the permitted hopping processes lead to contributions in the eigenstate energies of order @xmath622 and @xmath618 . these contributions are associated with the energy terms @xmath623 $ ] and @xmath624 respectively , on the right - hand side of eq . ( [ ehuinf ] ) . in the particular case of the ground state there are no contributions of order @xmath622 because the electron double occupation eigenvalue @xmath612 is zero . in the limit @xmath239 only hopping processes such that the electron singly occupied sites can move relatively to the electron doubly - occupied and empty site distribution configurations without changing these configurations are permitted . in such a limit there is a huge degeneracy of @xmath8-spin and spin occupancy configurations . thus there are several choices for complete sets of energy eigenstates with the same energy and momentum spectra given in eqs . ( [ ehuinf ] ) and ( [ ppuinf ] ) respectively . this is because in this limit there are also many choices for complete sets of compatible observables . the 1d hubbard model in the limit of @xmath16 has been studied in the literature by many authors @xcite . in the case of evaluation of quantities describing the physics of the model in the limit of @xmath16 , the alternative use of different complete sets of states leads to the same final expressions for correlation functions and other quantities of physical interest . in this paper we are interested in one of these choices of energy eigenstates only . it corresponds to the complete set of @xmath253 energy eigenstates generated from the corresponding energy eigenstates of the finite-@xmath14 1d hubbard model by turning off adiabatically the parameter @xmath17 . only that set of states correspond to the states obtained from the finite @xmath17 states by the electron - rotated - electron unitary transformation . these states are common eigenstates of both the 1d hubbard model as @xmath16 and of the set of number operators @xmath625 , @xmath626 , and @xmath218 with @xmath111 and @xmath121 in the same limit . together with the hamiltonian these operators constitute a complete set of compatible and commuting hermitian operators @xcite . these operators also commute with the momentum operator @xmath627 . for these energy eigenstates the band momentum @xmath123 of the @xmath19 pseudoparticles and @xmath104 and @xmath105 pseudoparticles such that @xmath274 is a good quantum number . this justifies the designation of band - momentum energy eigenstates . however , in the limit @xmath239 there are other choices for complete sets of energy and momentum eigenstates which are due to the periodic boundary conditions of the original electronic problem . in such a limit the 1d hubbard model can be mapped onto a problem for which the number of doubly occupied sites is conserved @xcite . since in the limit @xmath239 the ground state has zero double occupation , one usually introduces the concepts of lower and upper hubbard bands which are associated with the hilbert subspaces spanned by states with zero and finite values of the double occupation respectively @xcite . in the particular case of excitations involving creation of an electron the upper hubbard band corresponds to the hilbert subspace spanned by states with a single doubly occupied site . within the @xmath239 scheme used in refs . @xcite one spectrally decomposes the elementary electronic operators into those which solely act in the upper or lower hubbard bands , and eliminate perturbatively those parts which couple the two bands . this is achieved by application of a suitable transformation that eliminates these parts to a given order in @xmath17 . such a transformation is nothing but the electron - rotated - electron unitary transformation in the limit of small values of @xmath17 . this procedure leads to an effective hamiltonian which is equivalent to the 1d hubbard model to lowest order in @xmath17 . in reference @xcite that effective hamiltonian was called _ harris - lange model_. in reference @xcite the concepts of lower and upper hubbard bands were generalized to all values of @xmath17 and associated with the rotated - electron double occupation . in this appendix we illustrate for the specific case of a local @xmath375 pseudoparticle how the rotated - electron site distribution configurations which describe a local @xmath18 pseudoparticle always involve a number @xmath107 of site pairs such that @xmath368 and @xmath369 where @xmath370 . in the present case we have that @xmath114 and thus the rotated - electron site distribution configurations correspond to doubly occupied and empty sites . moreover , since @xmath628 a local @xmath375 pseudoparticle involves four site pairs such that @xmath368 and @xmath629 where @xmath630 . once there are two possible rotated - electron site occupancies for each pair and the number of pairs of each local @xmath375 pseudoparticle is four , the total number of different internal rotated - electron site distribution configurations is @xmath631 . according to the general expression ( [ cnp ] ) these @xmath632 different internal rotated - electron site distribution configurations of doubly occupied and empty sites are such that , @xmath633 \,\bigl[\prod_{g=1}^{4}(1-{\hat{\cal{t}}}_{c,\,4,\,j,\,g})\bigr ] \nonumber \\ & \times & ( \bullet_{h_{j,\,1}},\,\bullet_{h_{j,\,2}},\ , \bullet_{h_{j,\,3}},\,\bullet_{h_{j,\,4}},\ , \circ_{h_{j,\,5}},\,\circ_{h_{j,\,6}},\ , \circ_{h_{j,\,7}},\,\circ_{h_{j,\,8 } } ) = \nonumber \\ & + & e^{i\pi[h_{j,\,1}+h_{j,\,2}+h_{j,\,3}+h_{j,\,4}]}\ , ( \bullet_{h_{j,\,1}},\,\bullet_{h_{j,\,2}},\ , \bullet_{h_{j,\,3}},\,\bullet_{h_{j,\,4}},\ , \circ_{h_{j,\,5}},\,\circ_{h_{j,\,6}},\ , \circ_{h_{j,\,7}},\,\circ_{h_{j,\,8}})\nonumber \\ & - & e^{i\pi[h_{j,\,2}+h_{j,\,3}+h_{j,\,4}+h_{j,\,5}]}\,(\circ_{h_{j,\,1}},\,\bullet_{h_{j,\,2}},\ , \bullet_{h_{j,\,3}},\,\bullet_{h_{j,\,4}},\ , \bullet_{h_{j,\,5}},\,\circ_{h_{j,\,6}},\ , \circ_{h_{j,\,7}},\,\circ_{h_{j,\,8}})\nonumber \\ & + & e^{i\pi[h_{j,\,1}+h_{j,\,3}+h_{j,\,6}+h_{j,\,8}]}\,(\bullet_{h_{j,\,1}},\,\circ_{h_{j,\,2}},\ , \bullet_{h_{j,\,3}},\,\circ_{h_{j,\,4}},\ , \circ_{h_{j,\,5}},\,\bullet_{h_{j,\,6}},\ , \circ_{h_{j,\,7}},\,\bullet_{h_{j,\,8}})\nonumber \\ & - & e^{i\pi[h_{j,\,1}+h_{j,\,3}+h_{j,\,4}+h_{j,\,6}]}\,(\bullet_{h_{j,\,1}},\,\circ_{h_{j,\,2}},\ , \bullet_{h_{j,\,3}},\,\bullet_{h_{j,\,4}},\ , \circ_{h_{j,\,5}},\,\bullet_{h_{j,\,6}},\ , \circ_{h_{j,\,7}},\,\circ_{h_{j,\,8}})\nonumber \\ & + & e^{i\pi[h_{j,\,1}+h_{j,\,2}+h_{j,\,7}+h_{j,\,8}]}\,(\bullet_{h_{j,\,1}},\,\bullet_{h_{j,\,2}},\ , \circ_{h_{j,\,3}},\,\circ_{h_{j,\,4}},\ , \circ_{h_{j,\,5}},\,\circ_{h_{j,\,6}},\ , \bullet_{h_{j,\,7}},\,\bullet_{h_{j,\,8 } } ) \nonumber \\ & - & e^{i\pi[h_{j,\,1}+h_{j,\,2}+h_{j,\,4}+h_{j,\,7}]}\,(\bullet_{h_{j,\,1}},\,\bullet_{h_{j,\,2}},\ , \circ_{h_{j,\,3}},\,\bullet_{h_{j,\,4}},\ , \circ_{h_{j,\,5}},\,\circ_{h_{j,\,6}},\ , \bullet_{h_{j,\,7}},\,\circ_{h_{j,\,8 } } ) \nonumber \\ & + & e^{i\pi[h_{j,\,1}+h_{j,\,4}+h_{j,\,6}+h_{j,\,7}]}\,(\bullet_{h_{j,\,1}},\,\circ_{h_{j,\,2}},\ , \circ_{h_{j,\,3}},\,\bullet_{h_{j,\,4}},\ , \circ_{h_{j,\,5}},\,\bullet_{h_{j,\,6}},\ , \bullet_{h_{j,\,7}},\,\circ_{h_{j,\,8 } } ) \nonumber \\ & - & e^{i\pi[h_{j,\,1}+h_{j,\,2}+h_{j,\,3}+h_{j,\,8}]}\,(\bullet_{h_{j,\,1}},\,\bullet_{h_{j,\,2}},\ , \bullet_{h_{j,\,3}},\,\circ_{h_{j,\,4}},\ , \circ_{h_{j,\,5}},\,\circ_{h_{j,\,6}},\ , \circ_{h_{j,\,7}},\,\bullet_{h_{j,\,8}})\nonumber \\ & + & e^{i\pi[h_{j,\,2}+h_{j,\,3}+h_{j,\,5}+h_{j,\,8}]}\,(\circ_{h_{j,\,1}},\,\bullet_{h_{j,\,2}},\ , \bullet_{h_{j,\,3}},\,\circ_{h_{j,\,4}},\ , \bullet_{h_{j,\,5}},\,\circ_{h_{j,\,6}},\ , \circ_{h_{j,\,7}},\,\bullet_{h_{j,\,8}})\nonumber \\ & - & e^{i\pi[h_{j,\,1}+h_{j,\,6}+h_{j,\,7}+h_{j,\,8}]}\,(\bullet_{h_{j,\,1}},\,\circ_{h_{j,\,2}},\ , \circ_{h_{j,\,3}},\,\circ_{h_{j,\,4}},\ , \circ_{h_{j,\,5}},\,\bullet_{h_{j,\,6}},\ , \bullet_{h_{j,\,7}},\,\bullet_{h_{j,\,8}})\nonumber \\ & + & e^{i\pi[h_{j,\,3}+h_{j,\,4}+h_{j,\,5}+h_{j,\,6}]}\,(\circ_{h_{j,\,1}},\,\circ_{h_{j,\,2}},\ , \bullet_{h_{j,\,3}},\,\bullet_{h_{j,\,4}},\ , \bullet_{h_{j,\,5}},\,\bullet_{h_{j,\,6}},\ , \circ_{h_{j,\,7}},\,\circ_{h_{j,\,8}})\nonumber \\ & - & e^{i\pi[h_{j,\,2}+h_{j,\,5}+h_{j,\,7}+h_{j,\,8}]}\,(\circ_{h_{j,\,1}},\,\bullet_{h_{j,\,2}},\ , \circ_{h_{j,\,3}},\,\circ_{h_{j,\,4}},\ , \bullet_{h_{j,\,5}},\,\circ_{h_{j,\,6}},\ , \bullet_{h_{j,\,7}},\,\bullet_{h_{j,\,8}})\nonumber \\ & + & e^{i\pi[h_{j,\,2}+h_{j,\,4}+h_{j,\,5}+h_{j,\,7}]}\,(\circ_{h_{j,\,1}},\,\bullet_{h_{j,\,2}},\ , \circ_{h_{j,\,3}},\,\bullet_{h_{j,\,4}},\ , \bullet_{h_{j,\,5}},\,\circ_{h_{j,\,6}},\ , \bullet_{h_{j,\,7}},\,\circ_{h_{j,\,8}})\nonumber \\ & - & e^{i\pi[h_{j,\,3}+h_{j,\,5}+h_{j,\,6}+h_{j,\,8}]}\,(\circ_{h_{j,\,1}},\,\circ_{h_{j,\,2}},\ , \bullet_{h_{j,\,3}},\,\circ_{h_{j,\,4}},\ , \bullet_{h_{j,\,5}},\,\bullet_{h_{j,\,6}},\ , \circ_{h_{j,\,7}},\,\bullet_{h_{j,\,8}})\nonumber \\ & + & e^{i\pi[h_{j,\,5}+h_{j,\,6}+h_{j,\,7}+h_{j,\,8}]}\,(\circ_{h_{j,\,1}},\,\circ_{h_{j,\,2}},\ , \circ_{h_{j,\,3}},\,\circ_{h_{j,\,4}},\ , \bullet_{h_{j,\,5}},\,\bullet_{h_{j,\,6}},\ , \bullet_{h_{j,\,7}},\,\bullet_{h_{j,\,8}})\nonumber \\ & - & e^{i\pi[h_{j,\,4}+h_{j,\,5}+h_{j,\,6}+h_{j,\,7}]}\,(\circ_{h_{j,\,1}},\,\circ_{h_{j,\,2}},\ , \circ_{h_{j,\,3}},\,\bullet_{h_{j,\,4}},\ , \bullet_{h_{j,\,5}},\,\bullet_{h_{j,\,6}},\ , \bullet_{h_{j,\,7}},\,\circ_{h_{j,\,8 } } ) \ , . \label{c4p}\end{aligned}\ ] ] on the left - hand side of this equation the site-@xmath307 rotated - electron double occupation operator @xmath351 has eigenvalue @xmath246 and @xmath247 when that site is doubly occupied by rotated electrons and free of rotated electrons respectively , and the operator @xmath634 acts on the pair of sites of indices @xmath296 and @xmath635 only . this operator is designed to act onto the rotated - electron eight - site distribution configuration of the particular form illustrated on the left - hand side of eq . ( [ c4p ] ) . this specific configuration is such that the first four sites are doubly occupied by rotated electrons and the following four sites are free of rotated electrons . from the application of this operator onto such a rotated - electron eight - site distribution configuration a new distribution configuration is generated where the site of index @xmath296 is free of rotated electrons , the site of index @xmath635 is doubly occupied by rotated electrons , and the occupancy of the other six sites remains unchanged . this operation is repeated according to the products on the left - hand side of eq . ( [ c4p ] ) and leads to the @xmath632 internal rotated - electron site distribution configurations on the right - hand side of the same equation . these configurations are generated by considering that in each of the @xmath630 pairs of sites of indices @xmath359 and @xmath635 , the site of index @xmath359 is doubly occupied by rotated electrons and the site of index @xmath635 is free of rotated electrons and vice versa . the phase factors appearing on both sides of the above equation result from the momentum @xmath244 associated with each rotated - electron doubly occupied site . this is the momentum of the @xmath9 holon corresponding to such a site . finally note that a similar description is obtained for the @xmath378 pseudoparticle provided that one replaces these phase factors by one and the rotated - electron doubly occupied and empty sites by spin - down and spin - up rotated - electron singly occupied sites respectively . f. woynarovich , j. phys . c : solid state phys . * 15 * , 97 ( 1982 ) . masao ogata and hiroyuki shiba , phys . b * 41 * , 2326 ( 1990 ) . karlo penc , frdric mila , and hiroyuki shiba , phys . lett . * 75 * , 894 ( 1995 ) . karlo penc , karen hallberg , frdric mila , and hiroyuki shiba , phys . * 77 * , 1390 ( 1996 ) . karlo penc , karen hallberg , frdric mila , and hiroyuki shiba , phys . b * 55 * , 15 475 ( 1997 ) . elliott h. lieb and f. y. wu , phys * 20 * , 1445 ( 1968 ) . m. takahashi , prog . phys . * 47 * , 69 ( 1972 ) . b. ramos and m. j. martins , j. phys . a * 30 * , l195 ( 1997 ) ; m. j. martins and p.b . ramos , nucl . b * 522 * , 413 ( 1998 ) . m. rasetti , _ the hubbard model , recent results _ , word scientific , singapore ( 1991 ) ; a. montorsi , _ the hubbard model _ , word scientific , singapore ( 1992 ) . , edited by dionys baeriswyl , david k. campbell , jos m. p. carmelo , francisco guinea , and enrique louis ( nato asi series b , vol . 343 , plenum press , new york , 1995 ) . carmelo and j. m. romn , submitted for publication in phys . b ( 2002 ) [ cond - mat/0302044 ] . j. m. p. carmelo and p. d. sacramento , submitted for publication in phys . rev . b ( 2002 ) [ cond - mat/0302042 ] . j. m. p. carmelo and k. penc , submitted for publication in phys . b ( 2002 ) [ cond - mat/0303279 ] . n. e. hussey , m. n. mcbrien , l. balicas , j. s. brooks , s. horii , and h. ikuta , phys . lett . * 89 * , 086601 ( 2002 ) . a. menzel , r. beer , and e. bertel , phys . lett . * 89 * , 076803 ( 2002 ) . shin - ichi fujimori , akihiro ino , testuo okane , atsushi fujimori , kozo okada , toshio manabe , masahiro yamashita , hideo kishida , and hiroshi okamoto , phys . lett . * 88 * , 247601 ( 2002 ) . m. z. hasan , p. a. montano , e. d. isaacs , z .- x . shen , h. eisaki , s. k. sinha , z. islam , n. motoyama , and s. uchida , phys . lett . * 88 * , 177403 - 1 ( 2002 ) . r. claessen , m. sing , u. schwingenschlgl , p. blaha , m. dressel , and c. s. jacobsen , phys . lett . * 88 * , 096402 ( 2002 ) . gweon , j. d. denlinger , c. g. olson , h. hochst , j. marcus , and c. schlenker c , physica b * 312 - 313 * , 584 ( 2002 ) ; g .- h . gweon , j. d. denlinger , j. w. allen , r. claessen , c. g. olson , h. hochst , j. marcus , c. schlenker , and l. f. schneemeyer , j. electron spectrosc . relat . phenom . * 117 - 118 * , 481 ( 2001 ) . p. monceau , f. ya . nad , and s. brazovskii , phys . * 86 * , 4080 ( 2001 ) . k. takenaka , k. nakada , a. osuka , s. horii , h. ikuta , i. hirabayashi , s. sugai , and u. mizutani , phys . * 85 * , 5428 ( 2000 ) . t. mizokawa , c. kim , z .- x . shen , a. ino , t. yoshida , a. fujimori , m. goto , h. eisaki , s. uchida , m. tagami , k. yoshida , a. i. rykov , y. siohara , k. tomimoto , s. tajima , yuh yamada , s. horii , n. yamada , yasuji yamada , and i. hirabayashi , phys . lett . * 85 * , 4779 ( 2000 ) . j. moser , j. r. cooper , d. jrome , b. alavi , s. e. brown , and k. bechgaard , phys . * 84 * , 2674 ( 2000 ) . g. mihly , i. kzsmrki , f. zmborszky , and l. forr , phys . * 84 * , 2670 ( 2000 ) . v. vescoli , f. zwick , j. voit , h. berger , m. zacchigna , l. degiorgi , m. grioni , and g. grner , phys . lett . * 84 * , 1272 ( 2000 ) . j. d. denlinger , g.h . gweon , j. w. allen , c. g. olson , j. marcus , c. schlenker , and l .- s . hsu , phys . 82 * , 2540 ( 1999 ) . h. fujisawa , t. yokoya , t. takahashi , s. miyasaka , m. kibune , and h. takagi , phys . b * 59 * , 7358 ( 1999 ) . k. kobayashi , t. mizokawa , a. fujimori , m. isobe , y. ueda , t. tohyama , and s. maekawa , phys . * 82 * , 803 ( 1999 ) ; k. kobayashi , t. mizokawa , and a. fujimori , _ ibid . _ * 80 * , 3121 ( 1998 ) . claude bourbonnais and denis jrome , science * 281 * , 1155 ( 1998 ) . v. vescoli , l. degiorgi , w. henderson , g. grner , k. p. starkey , and l. k. montgomery , science * 281 * , 1181 ( 1998 ) . f. zwick , d. jrome , g. margaritondo , m. onellion , j. voit , and m. grioni , phys . * 81 * , 2974 ( 1998 ) ; f. zwick , s. brown , g. margaritondo , c. merlic , m. onellion , j. voit , and m. grioni , _ ibid . _ * 79 * , 3982 ( 1997 ) . r. neudert , m. knupfer , m. s. golden , j. fink , w. stephan , k. penc , n. motoyama , h. eisaki , and s. uchida , phys . lett . * 81 * , 657 ( 1998 ) . t. mori , t. kawamoto , j. yamaura , t. enoki , y. misaki , t. yamabe , h. mori , and s. tanaka , phys . lett . * 79 * , 1702 ( 1997 ) . c. kim , a. y. matsuura , z .- x . shen , n. motoyama , h. eisaki , s. uchida , t. tohyama , and s. maekawa , phys . lett . * 77 * , 4054 ( 1996 ) ; n. motoyama , h. eisaki , and s. uchida , phys . rev . lett . * 76 * , 3212 ( 1996 ) . m. granath , v. oganesyan , d. orgad , and s. a. kivelson , phys . b * 65 * , 184501 ( 2002 ) . d. orgad , s. a. kivelson , e. w. carlson , v. j. emery , x. j. zhou , and z. x. shen , phys . lett . * 86 * , 4362 ( 2001 ) . e. w. carlson , d. orgad , s. a. kivelson , and v. j. emery , phys . b * 62 * , 3422 ( 2000 ) . j. zaanen , o. y. osman , h. v. kruis , z. nussinov , and j. tworzydlo , philos . b * 81 * , 1485 ( 2001 ) . a. l. chernyshev , s. r. white , and a. h. castro neto , phys . b * 65 * , 214527 ( 2002 ) . a. brooks harris and robert v. lange , phys . * 157 * , 295 ( 1967 ) . o. j. heilmann and e. h. lieb , ann . n. y. acad . sci . * 172 * , 583 ( 1971 ) ; e. h. lieb , phys . * 62 * , 1201 ( 1989 ) . c. n. yang , phys . 63 * , 2144 ( 1989 ) . y. mizuno , k. tsutsui , t. tohyama , and s. maekawa , phys . b * 62 * , r4769 ( 2000 ) ; h. kishida , m. ono , k. miura , h. okamoto , m. izumi , t. manako , m. kawasaki , y. taguchi , y. tokura , t. tohyama , k. tsutsui , and s. maekawa , phys . rev . lett . * 87 * , 177401 ( 2001 ) . g. beni , t. holstein , and p. pincus , phys . b * 8 * , 312 ( 1973 ) . douglas j. klein , phys . rev . b * 8 * , 3452 ( 1973 ) . jos carmelo and dionys baeriswyl , phys . 7541 ( 1988 ) . a. h. macdonald , s. m. girvin , and d. yoshioka , phys . b * 37 * , 9753 ( 1988 ) . s. sorella and a. parola , j. phys . condens . matter * 4 * , 3589 ( 1992 ) . r. g. dias and j. m. lopes dos santos , j. de physique i * 2 * , 1889 ( 1992 ) . henk eskes and andrzej m. ole , phys . lett . * 73 * , 1279 ( 1994 ) ; henk eskes , andrzej m. ole , marcel b. j. meinders , and walter stephan , phys . b * 50 * , 17 980 ( 1994 ) . f. gebhard , k. bott , m. scheidler , p. thomas , and s. w. koch , phylo . b * 75 * , 13 ( 1997 ) . n. m. r. peres , r. g. dias , p. d. sacramento , and j. m. p. carmelo , phys . b * 61 * , 5169 ( 2000 ) . j. m. p. carmelo , to be submitted for publication in phys . b ( 2003 ) . j. m. p. carmelo and k. penc , to be submitted for publication in phys . b ( 2003 ) . m. sing , u. schwingenschlgl , r. claessen , p. blaha , j. m. p. carmelo , l. m. martelo , p. d. sacramento , m. dressel , and c. s. jacobsen , submitted for publication in phys . b ( 2003 ) [ cond - mat/0304283 ] . fabian h. l. essler , vladimir e. korepin , and kareljan schoutens , phys . lett . * 67 * , 3848 ( 1991 ) ; nucl . b * 372 * , 559 ( 1992 ) . v. emery , phys . b * 14 * , 2989 ( 1975 ) .
in this paper we describe the pseudoparticles , holons , and spinons whose occupancy configurations describe the energy eigenstates of the one - dimensional ( 1d ) hubbard model in terms of rotated electrons . rotated electrons are related to electrons by a mere unitary transformation such that rotated electron double occupation is a good quantum number and the _ effective electronic lattice _ occupied by rotated electrons is identical to the _ real - space lattice _ occupied by electrons . moreover , we find that the band - momentum pseudoparticle description associated with the bethe - ansatz takahasi s thermodynamic equations is related by fourier transform to a _ local pseudoparticle _ representation in terms of occupancy configurations of spatial coordinates . such spatial coordinates correspond to an _ effective pseudoparticle lattice_. there is an effective pseudoparticle lattice for each pseudoparticle branch with finite occupancy in a given state . this description introduces the local pseudoparticles whose spatial coordinate is the conjugate of the band momentum . we describe the energy eigenstates in terms of local pseudoparticle site distribution configurations in such lattices . moreover , we relate both the local pseudoparticle internal structure and the latter configurations to the rotated - electron site distribution configurations which describe the energy eigenstates . the electron - rotated - electron unitary transformation is such that the latter configurations are independent of the value of the on - site coulombian repulsion . our findings provide useful information about the relation of the exotic pseudoparticles , holons , and spinons that diagonalize the non - perturbative many - electron problem to the original electrons . we provide an example showing how the derived local pseudoparticle representation can be used in the evaluation of finite - energy few - electron spectral function expressions .
it is known that the persistent current ( i.e. a direct current in the thermodynamic equilibrium state ) can flow along a mesoscopic loop because of the quantization of the momentum circulation @xmath0 the persistent current @xmath1 in normal metal mesoscopic loops was predicted more than 30 years ago [ 1 ] and was observed not so long ago [ 2 ] . it is known also that a potential difference @xmath2 should be observed on a segment @xmath3 of an inhomogeneous conventional loop at a current density @xmath4 along the loop if the average resistivity along the segment @xmath5 differs from the one along the loop @xmath6 . therefore a possibility of a persistent voltage can be assumed at a segment of an inhomogeneous mesoscopic loop at @xmath7 . the latter , i.e. @xmath7 , can be only if the mean free path of electrons is not smaller than the length @xmath8 of loop circumference and the temperature is lower than the energy difference between adjacent permitted states @xmath9 . this difference is not large for electrons . for example , at @xmath10 , the @xmath11 value corresponds @xmath12 . therefore it is enough difficult to observe the persistent current in normal metal loop [ 2 ] . it is more easier to observe the persistent current in superconducting loop since the mean free path of superconducting pairs is infinite and the energy difference between adjacent permitted states is much higher than in normal metal loop . first experimental evidence of the persistent current at non - zero resistance - the little - parks experiment [ 3 ] was made as long ago as 1962 . in the present work the dc voltage proportional to the @xmath13 is observed on segments of an asymmetric superconducting loop in accordance with the analogy with a conventional inhomogeneous loop . an electron micrograph one of the aluminum loop samples . @xmath14 and @xmath15 are the current and potential contacts of the symmetric loop . @xmath16 and @xmath17 are the current and potential contacts of the asymmetric loop . @xmath18 are the additional potential contacts of the asymmetric loop . ] the dependencies of the dc voltage @xmath19 on the magnetic flux @xmath20 of some round symmetric and asymmetric al loops ( see fig.1 ) with a diameter @xmath21 = 1 , 2 and 4 @xmath22 and a linewidth w = 0.2 and 0.4 @xmath22 at the dc measuring current @xmath23 and different temperature close to @xmath24 were measured . here @xmath25 is the magnetic induction produced by a superconducting coil ; @xmath26 is the area of the loop . the al microstructures fig.1 are prepared using an electron lithograph developed on the basis of a jeol-840a electron scanning microscop . the sheet resistance @xmath27 at 4.2 k , the resistance ratio @xmath28 and the midpoint of the superconducting resistive transition @xmath29 . the voltage oscillations corresponded to the conventional little - parks oscillations were observed on the contacts @xmath15 of the symmetric loop [ 4 ] . these resistance oscillations @xmath30 observed in the temperature region of the superconducting transition ( i.e. at @xmath31 ) are explained by the @xmath24 oscillations @xmath32 because of the oscillation of the persistent current @xmath33 [ 5 ] . here @xmath34 is the flux quantum for superconducting pair , @xmath35 . the voltage @xmath36 at the measuring current @xmath37 whereas the voltage oscillations @xmath38 are observed on the contacts @xmath17 and @xmath18 of the asymmetric loop at @xmath39 fig.2 . the apparent ( with amplitude @xmath40 ) oscillations without an external dc current were observed in a narrow temperature region @xmath41 corresponds to the bottom of the resistive transition . its amplitude increases ( up to @xmath42 ) with temperature lowering down to the lowest temperature @xmath43 we could reach . oscillation of the voltage measured on the @xmath17 contacts ( upper curve ) and on the @xmath18 contacts ( lower curve ) of the asymmetric loop with 2r = 4 @xmath22 and w = 0.4 @xmath22 . @xmath44 . @xmath45 corresponded to the bottom of the resistive transition . ] the observation of the voltage oscillations in the asymmetric loop and its absence in the symmetric loop conform to the analogy with a conventional loop . although the oscillations @xmath33 [ 5 ] take place in the both cases in the symmetric loop @xmath46 and therefore @xmath47 at @xmath44 . the difference of @xmath48 from @xmath49 in the asymmetric loop is caused by the additional potential contacts @xmath18 . there is an important difference from the conventional loop where the potential difference appears in accordance with the ohm s law @xmath50 . the voltage on fig.2 is observed without the faraday s voltage @xmath51 and consequently the electric field @xmath52 and the persistent current @xmath13 should have opposite directions in a segment because @xmath53 , i.e. according to the result presented on fig.2 a segment of the asymmetric loop is a dc power source at @xmath54 and @xmath55 . the power @xmath56 can be obtained on a load with the resistance @xmath57 . because the resistance of the segment @xmath58 then @xmath59 at @xmath60 and @xmath61 . it should be noted that already the classical little - parks experiment is evidence of the dc power source . according to this experiment the persistent current @xmath63 is observed at non - zero resistance @xmath64 along the loop and consequently an energy dissipation with the power @xmath65 takes place . the persistent current is maintained in spite of this dissipation because of reiterated changes of the momentum circulation of superconducting pairs at switching of the loop between superconducting states with different connectivity [ 6 ] . because of the quantization the momentum circulation of pair @xmath66 changes from @xmath67 to @xmath68 at each closing of superconducting state . the reiterated changes @xmath69 with an average frequency @xmath70 is equivalent of the action of an average force @xmath71 which maintains the circulating current instead of the faraday s voltage . @xmath72 is the thermodynamic average of the quantum number @xmath73 . @xmath74 when @xmath75 is not close to @xmath76 and @xmath77 at @xmath78 , where the integer number @xmath73 in @xmath79 corresponds to a minimum possible value @xmath80 . the quantum force @xmath81 is not localized in principle in any loop segment [ 6 ] , i.e. @xmath82 . therefore a potential difference @xmath83 should be observed on a segment @xmath3 remaining in superconducting state when other segment is switched in normal state with the frequency @xmath70 . this relation explains the observed voltage oscillation fig.2 . loop segments are switched in normal state at @xmath84 by thermal fluctuations and an external electric noise . consequently the energy of thermal fluctuations or an external electric noise is transformed in the power @xmath62 observed in our work . the observation of the voltage oscillations fig.2 is evidence of a possibility of an analogous observation on segments of inhomogeneous normal metal mesoscopic loop . there is an important question : thermal fluctuations or an external noise induce the dc voltage . the temperature dependence of the amplitude @xmath85 observed in our work is evidence of the latter . but it is enough obvious that thermal fluctuations can also induce the dc voltage on segments of inhomogeneous mesoscopic loops [ 6 ] .
a dc voltage changed periodically with magnetic field is observed on segments of asymmetric aluminum loop without any external dc current at temperatures corresponded to superconducting transition . according to this experimental result a segment of the loop is a dc power source . a possibility of a persistent voltage on segments of an inhomogeneous normal metal mesoscopic loop follows from this result .
there are reasons to believe that at the planck scale , the scale at which gravity becomes a quantum phenomenon , the very structure of space - time may change . that this may happen is suggested even by general relativity . a quantum mechanical particle of momentum @xmath2 in the presence of a classical gravitational field ( the latter described by einstein s equations ) causes the metric @xmath3 to fluctuate . this induces an additional uncertainty in position , given by @xmath4 , where @xmath5 is the planck length . thus the usual quantum mechanical uncertainty relation gets modified to @xmath6 at high energies , the second term can become significant and lead to important deviations from the usual quantum mechanics . for example , in the usual quantum mechanics , @xmath7 is large at low values of momenta but can become small at high momenta which can provide higher resolution . with the modified uncertainty relation even at high momenta @xmath7 is limited in resolution because of strong curvature effects . in other words , independent of momentum , @xmath7 is always larger than a minimal length scale @xmath5 . the appearance of the minimal length in the classical theory of gravity should tell us that it is no surprise to expect that such a conclusion becomes even more inevitable in a quantum theory of gravity . indeed , a whole range of quantum gravity models predict the existence of a minimal length @xcite . however , because the planck length is so minuscule it is of consequence only to the physics of the early universe , if at all @xcite . there is a twist in the plot , however , and that is due to the development of brane - world physics . this new paradigm in physics has far - reaching implications but of our immediate interest is the fact that there are models of brane - world which allow a realisation of quantum gravity effects at very low energy scales as low as a tev . in this paper , we will deal with the simplest of such realizations which is a model due to arkani - hamed , dimopoulos and dvali . in the model of arkani - hamed , dimopoulos and dvali , the so - called add model @xcite , extra space dimensions are compactified to have large volume moduli and thereby it becomes possible to lower the scale of quantum gravity from the planck scale to the tev scale . in the add model , the gauge interactions are confined to a 3-brane while gravity propagates in all the available dimensions . the higher dimensional planck scale @xmath8 in the add model is related to the usual planck scale by @xmath9 , where @xmath10 is the radius of compactification and @xmath11 is the number of extra space dimensions . for appropriate choices of @xmath10 and @xmath11 , the fundamental ( higher dimensional ) planck or string scale can be brought down to @xmath12 1 tev . this scale can be probed at present and future colliders . various signals of graviton production and virtual graviton exchange have been studied @xcite and reviewed @xcite in recent years . in the context of add model with @xmath13 tev , the minimal length hypothesis is phenomenologically interesting if we take it to be around an inverse tev , viz . @xmath14 . for a review , see ref . @xcite . different applications of the minimal length scenario ( mls ) have been discussed in @xcite . in an earlier analysis @xcite , we had studied how the mls hypothesis influences the dilepton production process in hadron colliders . in the add model , for such processes involving virtual gravitons , one has to sum over an infinite tower of graviton kaluza - klein ( kk ) states at the amplitude level . the result is divergent and to cure the divergence an _ ad - hoc _ cutoff of the order of @xmath15 is used . however , in the mls scenario , the minimal length acts as an ultraviolet regulator and allows one to sum over the entire kk graviton tower by smoothly cutting off the contribution of higher energy kk states rendering the amplitude finite . another important modification comes from the rescaling of momentum measure leading to an alteration in the phase space integration . these , as we observed in @xcite , lead to a significant deviation of the bound on @xmath8 from the one obtained in the conventional add picture without the mls hypothesis . in the present analysis , we consider processes involving emission of real gravitons in hadron colliders , namely , the upgraded tevatron and the future large hadron collider ( lhc ) . the partonic sub - processes are @xmath16 , where @xmath17 are the appropriate quarks / gluons . here @xmath18 means a real graviton with kk index @xmath19 , where @xmath11 is the number of extra dimensions . for a given process , one must sum over such external modes at the cross section level . this kk summation gets modified in the mls scenario . the phase space integration also gets modified in the mls scenario . again , as we shall see towards the end , the bounds that one obtains on @xmath8 in the conventional add model from the jet plus missing energy mode in the hadron colliders are considerably weakened if we implement the mls hypothesis . specifically , by mls scenario we mean the add model with a minimal length @xmath5 of the order of tev@xmath20 incorporated in it . since the uncertainty in position measurement now can not be smaller than @xmath5 , now the standard commutation relation between position and momentum has to be modified @xcite . in the mls scheme , the compton wavelength ( @xmath21 ) of a particle can not be arbitrarily small . we suppose that even though the wave vector @xmath22 is bounded from above , the momentum @xmath2 can be as large as possible . in the same way , the frequency @xmath23 is restricted from above while the energy @xmath24 can go up arbitrarily . this means that the standard relations @xmath25 and @xmath26 need to be modified . this can be realised by introducing the unruh ansatz @xcite @xmath27 , \nonumber \\ l_{p } \omega(e)&=&\tanh^{1/\gamma } \left[\left(\frac{e}{m_{s}}\right)^{\gamma}\right],\end{aligned}\ ] ] where @xmath28 is a positive constant . ( [ dispersion ] ) captures the essence of a minimal length scale : as @xmath2 ( or @xmath24 ) becomes very large , @xmath22 ( or @xmath23 ) approaches the upper bound @xmath29 and hence one can not probe arbitrarily small length and time scales . the generalized position - momentum and energy - time uncertainty principle can now be written in a lorentz covariant form as @xmath30=i \,\,\frac{\partial p_{\mu}}{\partial k_{\nu}}.\ ] ] as explained in @xcite , the effect of eq . ( [ dispersion ] ) can conveniently be accounted for by a simple redefinition of the momentum measure , which in one dimension is given by @xmath31 the lorentz invariant 4-momentum measure is modified as @xmath32 where the jacobian matrix @xmath33 can be kept diagonal . in this paper we study how processes involving real graviton emissions in hadron colliders are going to be influenced by the mls hypothesis . the emitted gravitons will disappear carrying missing energy , so the events that we look for are jets plus missing energy . the partonic sub - processes that contribute to this channel are @xmath34 , @xmath35 , @xmath36 , and @xmath37 . the differential cross section for a particular sub - process ( labelled @xmath38 ) for a given kk graviton mode has the form @xmath39 where @xmath40 is the mass of @xmath18 , and @xmath41 are functions of the mandelstam variables @xmath42 whose explicit expressions are given in ref . these cross sections have to be convoluted with the density of kk states given by @xcite @xmath43 noting that one can express @xmath44 , the differential cross section summed over the kinematically allowed kk graviton tower is given by @xmath45 where the lower limit of integration is zero and the upper limit corresponds to the maximum allowed graviton energy as @xmath46 . an important thing that has to be noticed here is that while the usual 4-dimensional graviton emission cross section goes like @xmath47 , the effect of kk summation jacks up the contribution which goes effectively as @xmath48 . this is the add effect . now , how does the mls hypothesis modify the kk summation ? now for analytic simplicity we take @xmath49 and set @xmath50 . the departure from these assumptions will be discussed later . the unruh relations ( [ dispersion ] ) will modulate the integral in ( [ dscon ] ) by a factor @xmath51 . the mls - influenced differential cross section is given by @xmath52 for large argument , the sech - square function falls exponentially , thus the contributions from high mass kk states will be smoothly cut off . another important modification due to the mls is the change in the cross section due to the rescaling of the momentum measure given in eq . ( [ scale ] ) . it is straightforward to check , as derived in @xcite , that the phase space integration in the total cross section picks up the following modification factor : @xmath53 where @xmath54 runs over the four initial and final states in a @xmath55 process , and @xmath56 and @xmath57 are the total initial and final four momenta in a @xmath58 process . we work out this modification factor for the process we are studying viz . note , while @xmath17 are massless , @xmath18 s are massive . using eqs . ( [ dispersion ] ) and ( [ scale ] ) , setting @xmath60 and @xmath61 , we can easily show that @xmath62 above , @xmath63 and @xmath64 are the momentum fractions of the hadrons carried by the initial state partons , @xmath65 and @xmath66 are the rapidities of the final state parton and the graviton , @xmath67 is the transverse momentum of the final state parton , and @xmath68 is the transverse mass of the graviton @xmath18 . it is instructive to check that in the decoupling limit @xmath69 , the phase space correction factor goes to unity . the sech - square modulation in eq . ( [ dsmls ] ) together with the phase space modifications in eqs . ( [ mod ] ) and ( [ ebyomega ] ) constitute the complete correction . if we keep @xmath28 and @xmath70 as free parameters the analytic expressions of the modifications in eqs . ( [ dsmls ] ) and ( [ ebyomega ] ) will look more complicated , and we do not display them . instead , we have numerically demonstrated the sensitivity of cross section due to variations of @xmath28 and @xmath71 in the next section . we have numerically computed the cross section for the production of the graviton kk tower in association with a jet at the run ii of tevatron ( @xmath72 1.96 tev ) and at the lhc ( @xmath72 14 tev ) in both the add model and in the mls scenario . there is an irreducible sm background to the signal coming from @xmath73jet production followed by the invisible decay of the @xmath74 into neutrinos . we have computed the sm background and the signal in the add and the mls cases using the mrst 2001 lo parton distributions @xcite ( the mls modification of the parton densities is numerically insignificant ) . since the signal peaks at larger missing-@xmath75 as compared to the background the purity of the sample can be improved by employing a strong cut on the missing-@xmath75 . we require the missing-@xmath75 to be greater than 80 gev for the tevatron run ii analysis . the cut on the jet pseudorapidity is @xmath76 2.4 . our results are shown for the case @xmath77 . we display the results of this computation in figure 1 by plotting the cross section in the add and the mls scenarios as a function of @xmath8 . experimentally , the @xmath73jet cross section is not known very precisely . to have a rough estimate , we multiplied the sm cross section by the appropriate luminosity to obtain the number of events , and twice the square - root of those events ( back to pb unit ) correspond to the 2-@xmath78 spread around the sm value . thus , no detector effect has been taken into consideration . in other words , assuming purely statistical errors on the sm cross section , we have plotted the 2-@xmath78 allowed bands in the figure , which may be considered as a rough estimate of the allowed zone . as expected , the mls cross section has a steeper fall - off with @xmath8 . consequently , the lower bound on @xmath8 in the mls scenario is diluted by about 60 gev compared to what one obtains in the conventional add model at the @xmath79 confidence level . for the lhc analysis , we have used the much stronger cut of 200 gev on the missing @xmath75 and a jet pseudorapidity cut of @xmath80 3.5 . the cross section for the signal in the two cases and the sm - allowed band are shown in figure 2 . the effects of the suppression of cross section due to the minimal length is stronger at the lhc and hence the dilution of the bound ( by a few hundred gev ) is greater than at the tevatron . figures 1 and 2 have been drawn setting @xmath60 and @xmath61 . in figure 3 , we have displayed the variation of the total cross section , for tevatron run ii , with @xmath28 for different choices of @xmath81 . the value of @xmath82 has been fixed at 1.5 tev for this figure . in figure 4 , we have presented the analogous variation for the lhc energy for @xmath83 tev . we recall that the dependence of the cross section on @xmath28 and @xmath71 arise from the use of eq . ( [ dispersion ] ) while calculating the mls modulation factors . the existence of a minimal length is a generic prediction of quantum gravity theories . the model of large extra dimensions , the add model , is an example where quantum gravity effects become manifest at energy scales as low as a tev making it accessible to collider experiments . we incorporate the idea of a minimal length in the add model , which we now call the mls scenario , and study what impact the existence of a minimal length has on the collider effects of the conventional add model . in particular , we have considered graviton production in association with jets at hadron colliders to address this particular issue . by analysing the jet@xmath84missing energy channel at the upgraded tevatron and the large hadron collider , we demonstrate that the add model bounds get diluted when the minimum length hypothesis is invoked . finally , we make a few comments on the possible implications of minimal length on the astrophysical constraints on the add model . first we note that gravitons produced in supernovae core by nucleon nucleon bremsstrahlung process can carry away missing energy . but neutrino fluxes from sn1987a detected by the imb and kamiokande experiments have been observed to carry away most of the energy ( @xmath85 ergs ) released during core collapse . this has been used to obtain tight bounds on @xmath8 of the order of 50 tev for @xmath86 and about 4 tev for @xmath87 @xcite in the conventional add model . since the production process of the gravitons is the same as the one studied in the present paper , it is important to ask as to what extent the mls scenario dilutes the supernova bounds . note , the high nucleon density at the supernova core and the large multiplicity of kk graviton modes for low values of @xmath11 is responsible for the enhancement of the effects , which in turn places strong constraints on @xmath8 . the mls scenario would indeed suppress the graviton production cross section and tend to dilute the bounds . but the numerical impact of this dilution would not be large compared to the other uncertainties , e.g. strong core temperature dependence , inherited from pure supernova dynamics . a string theory calculation of this cross section has been done in @xcite , where the amplitude is multiplied by a form factor . it is worth comparing the string theory result with the unruh ansatz that we have used . in the string theory case the form factor goes like @xmath88 in the asymptotic limit . in our case , the modulation depends on the choice of @xmath28 . for @xmath60 , the high mass kk states are cut off by @xmath89 since the argument of sech - square in eq . ( [ dsmls ] ) has a linear dependence on graviton mass . for @xmath90 , the dependence on the graviton mass is quadratic , and we obtain a string - like modulation . we note at this point that the string calculation as done in @xcite is based on a toy scenario with an embedding of qed in string theory , and the calculation of jet plus missing energy relies on a rough extrapolation from this toy scenario . because of this model dependence , a more detailed comparison of the results in @xcite and our analysis is not possible . gb and ks acknowledge support by the indo - french centre for the promotion of advanced research , new delhi ( ifcpar project no : 2904 - 2 ) , and thank cern ph / th division for hospitality during the work . gb also thanks lpt - orsay and spht - saclay for hospitality during the final stage of the work . gb s research has also been supported , in part , by the dst , india , project no : sp / s2/k-10/2001 . n. arkani - hamed , s. dimopoulos and g. dvali , phys . b * 429 * ( 1998 ) 263 [ arxiv : hep - ph/9803315 ] ; i. antoniadis , n. arkani - hamed , s. dimopoulos and g. dvali , phys . b * 436 * ( 1998 ) 257 [ arxiv : hep - ph/9804398 ] ; n. arkani - hamed , s. dimopoulos and g. dvali , phys . d * 59 * ( 1999 ) 086004 [ arxiv : hep - ph/9807344 ] ; n. arkani - hamed , s. dimopoulos and j. march - russell , phys . d * 63 * ( 2001 ) 064020 [ arxiv : hep - th/9809124 ] . u. harbach , s. hossenfelder , m. bleicher and h. stocker , phys . b * 584 * ( 2004 ) 109 [ arxiv : hep - ph/0308138 ] , m. cavaglia and s. das , arxiv : hep - th/0404050 , m. cavaglia , s. das and r. maartens , class . * 20 * ( 2003 ) l205 [ arxiv : hep - ph/0305223 ] ; s. hossenfelder , phys . d * 70 * ( 2004 ) 105003 [ arxiv : hep - ph/0405127 ] ; phys . b * 598 * ( 2004 ) 92 [ arxiv : hep - th/0404232 ] ; arxiv : hep - ph/0409350 ; mod . a * 19 * ( 2004 ) 2727 [ arxiv : hep - ph/0410122 ] ; arxiv : hep - ph/0412265 ; arxiv : hep - th/0510245 ; c. barcelo , s. liberati and m. visser , arxiv : gr - qc/0505065 .
theories of quantum gravity suggest the existence of a minimal length scale . we study the consequences of a particular implementation of the idea of a minimal length scale in the model of large extra dimensions , the add model . to do this we have looked at real graviton production in association with a jet at hadron colliders . in the minimal length scenario , the bounds on the effective string scale are significantly less stringent than those derived in the conventional add model , both at the upgraded tevatron and at the large hadron collider . ` pacs nos : 11.25.wx , 13.85.qk ` + ` key words : extra dimension , minimal length , hadron collider ` ` hep - ph/0512050 ` + sinp / tnp/05 - 28 + tifr / th/05 - 46 + * studying the effects of minimal length in large extra dimensional models in the jet @xmath0 missing energy channels at hadron colliders * + , kumar rao @xmath1 , k. sridhar @xmath1 + \1 . saha institute of nuclear physics , 1/af bidhan nagar , kolkata 700064 , india + 2 . department of theoretical physics , tata institute of fundamental research , + homi bhabha road , mumbai 400005 , india
0.5 cm in treating nuclear many - body problems , one can often reduce the full " many - body problem to a much smaller and more manageable problem , referred to as a model - space problem . in so doing , an important step is to determine the model - space effective interaction @xmath0 . in this paper , we would like to present an introductory and pedagogical review of this topic , especially the derivation of @xmath0 from realistic meson - exchange nucleon - nucleon interactions . many - body problems are difficult as we are all aware of . they are particularly so , when a , the number of particles in the system , is large . in fact a many - body problem with a=3 is already a very hard problem . in the real world , a is usually much larger . for example , a=18 when the nucleus @xmath6 is treated as an 18-nucleon problem . it becomes an a=54 problem if it is treated as composed of 54 quarks . so the complexity of the problem depends largely on how we look at the problem , or , more precisely , it depends on what is our model space " within which we are making our observations . let us consider the nucleus @xmath6 as an example ; it is a system of 18 nucleons . the nuclear shell model has been the most successful model for nuclear structure . this model is in fact a model - space approach . the wave functions @xmath7 of this nucleus consists of , in principle , shell - model wave functions @xmath8 , @xmath9 , .... and @xmath10 ; @xmath11 and @xmath12 denoting particles and holes with respect to a closed @xmath13 core . the dimension of this full shell - model space is practically infinite " . for actual calculations , some truncation or renormalization of the full - space problem is indispensible . to describe the low - energy properties of this nucleus , it may not be necessary to consider all these wave - function components ; it may be sufficient to include just some low - energy parts of them . this is actually the approach commonly used in the shell - model description of @xmath6 , where this nucleus is treated as composed of two valence neutrons in the 0d1s shell outside a closed @xmath13 core . denoting this restricted model space as p , the hamiltonian for this nucleus is @xmath14 where @xmath0 is the effective interaction and @xmath15 denotes the single - particle ( sp ) hamiltonian . .binding energies of ca isotopes in mev given by talmi s model.@xcite [ cols="<,<,<",options="header " , ] as shown in the table , the @xmath17 coefficients for the @xmath18 channel are all very small . thus the counter term for the nn tensor force is nearly vanishing and consequently @xmath19 . it appears that the tensor force is exempted from renormalization . it should be of interest to further study this possible exemption property of the tensor force . 0.3 cm * iii.c . hermitian low - momentum interactions * 0.3 cm it should be pointed out that the low - momentum nn interaction given by the above folded - diagram expansion is not hermitian . this is not a desirable feature , and one would like to have an interaction which is hermitian . methods for deriving hermitian effective interactions have been developed by okubo @xcite and andreozzi @xcite . a general framework for constructing hermitian effective interactions was recently studied by holt , kuo and brown @xcite . in this section , we shall discuss their method for derivation of hermitian effective interactions . as we shall see shortly , one can in fact construct a family of hermitian low - momentum nn interactions that are phase - shift equivalent . let us denote the folded - diagram low - momentum nn interaction as @xmath20 . ( recall that we have used the lee - suzuki ( ls ) method for its derivation . ) @xmath20 preserves the half - on - shell @xmath3-matrix @xmath21 for @xmath22 . if we relax this half - on - shell constraint , we can obtain low - momentum nn interactions which are hermitian . ( note the hermitian potentials discussed below all preserve the fully on - shell @xmath3-matrix and are phase - shift equivalent @xcite . ) the model - space secular equation for @xmath20 is @xmath23 where @xmath24 is a subset of the eigenvalues @xmath25 of the full - space schroedinger equation @xmath26 . @xmath27 is the @xmath28-space projection of the full - space wave function @xmath29 , namely @xmath30 . the above effective interaction may be rewritten in terms of a wave operator @xmath31 , namely @xmath32 where @xmath31 possesses the usual properties : @xmath33 here @xmath34 is the complement of @xmath28 , @xmath35 . while the full - space eigenvectors @xmath7 are orthogonal to each other , the model - space eigenvectors @xmath27 are clearly not so and consequently the effective interaction @xmath20 is not hermitian . we now make a @xmath36 transformation such that @xmath37 where @xmath38 is the dimension of the model space . this transformation reorients the vectors @xmath39 such that they become orthonormal to each other . we assume that @xmath27 s ( @xmath40 ) are linearly independent so that @xmath41 exists , otherwise the above transformation is not possible . since @xmath42 and @xmath36 exist entirely within the model space , we can write @xmath43 and @xmath44 . using eq . ( 52 ) , we transform eq . ( 50 ) into @xmath45 which implies @xmath46 since @xmath47 is real ( it is an eigenvalue of @xmath48 which is hermitian ) and the vectors @xmath42 are orthonormal to each other , @xmath49 is clearly hermitian . the non - hermitian secular equation , eq . ( 12 ) , is now transformed into a hermitian model - space eigenvalue problem @xmath50 with the hermitian effective interaction @xmath51 or equivalently @xmath52 to calculate @xmath53 , we must first have the @xmath36 transformation . since there are certainly many ways to construct @xmath36 , this generates a family of hermitian effective interactions , all originating from @xmath20 . for example , we can construct @xmath36 using the familiar schmidt orthogonalization procedure , namely : @xmath54 with the matrix elements @xmath55 determined from eq . we denote the hermitian effective interaction using this @xmath36 transformation as @xmath56 . clearly there is more than one such schmidt procedure . for instance , we can use @xmath57 as the starting point , which gives @xmath58 , @xmath59 , and so forth . this freedom in choosing the orthogonalization procedure actually gives us many ways to generate a hermitian interaction , and this is our family of hermitian interactions produced from @xmath20 . we now show how some well - known hermitization transformations relate to ( and in fact , are special cases of ) ours . we first look at the okubo transformation @xcite . from the properties of the wave operator @xmath31 @xmath60 it follows that an analytic choice for the @xmath36 transformation is @xmath61 this leads to the hermitian effective interaction & & v_okb-1= p ( 1+^+ ) ^1/2p(h_0+v_ls)p ( 1+^+ ) ^-1/2p + & & 0.8 cm -ph_0p . it is easily seen that the above is equal to the okubo hermitian effective interaction & & v_okb = p ( 1+^+ ) ^-1/2(1+^+)(h_0+v ) + & & ( 1 + ) ( 1+^+ ) ^-1/2p - ph_0p , giving us an alternative expression , eq . ( 62 ) , for the okubo interaction . there is another interesting choice for the transformation @xmath36 . as pointed out by andreozzi @xcite , the positive definite operator @xmath62 can be decomposed into two cholesky matrices , namely @xmath63 where @xmath64 is a lower triangle cholesky matrix , @xmath65 being its transpose . since @xmath64 is real and it is within the @xmath28-space , we have @xmath66 and the corresponding hermitian effective interaction from eq . ( 56 ) is @xmath67 this is the hermitian effective interaction of andreozzi @xcite . the hermitian effective interaction of suzuki and okamoto @xcite is of the form @xmath68 with @xmath69 and @xmath70 . it has been shown that this interaction is the same as the okubo interaction @xcite . in terms of the @xmath36 transformation , it is readily seen that the operator @xmath71 is equal to @xmath72 with @xmath36 given by eq . thus , these three well - known hermitian effective interactions indeed belong to our family . the family of hermitian potentials are all phase - shift equivalent to @xmath20 @xcite . using a solvable matrix model , one finds that the above hermitian effective interactions @xmath73 and @xmath74 can be quite different @xcite from each other and from @xmath20 , especially when @xmath20 is largely non - hermitian . for the @xmath1 case , it is fortunate that the @xmath2 coresponding to @xmath20 is only slightly non - hermitian . as a result , the hermitian low - momentum nn inteactions corresponding to @xmath73 and @xmath74 are all quite similar to each other and to @xmath20 . thus in applications it is the @xmath2 corresponding to @xmath20 which is commonly used @xcite . as indicated by figs . 6 - 8 , the @xmath2 interactions derived from free - space @xmath1 have worked well in shell - model calculations involving mainly valence nucleons . but for infinite symmetric nuclear matter , such free - space two - nucleon interactions alone are unable to reproduce simultaneously the empirical saturation energy and density ( @xmath75mev and @xmath76@xmath77 ) @xcite . to illustrate , we display in fig . 15 results from a recent nuclear matter calculation @xcite using a non - pertubative ring - diagram resummation method that will be outlined later . there the lowest curve ( c ) is obtained with the free - space bonns potential @xcite . as seen this curve descends rapidly with density , showing no sign of reaching a local minimum at the saturation energy and density . in the present section we outline how this over - binding problem may be overcome by using density - dependent effective interactions generated by the brown - rho scaling mechanism @xcite or the inclusion of three - nucleon potential @xmath5 @xcite . ( the symbol @xmath1 will be used to denote the two - nucleon potential . ) realistic nucleon - nucleon potentials are mediated by the exchange of mesons such as the @xmath78 , @xmath79 , @xmath31 and @xmath80 mesons . in constructing these potentials the meson - nucleon coupling constants ( and in some cases their masses ) are adjusted to fit the ` free - space ' nn scattering data . mesons in a nuclear medium , however , can have properties ( masses and couplings ) that are different from those in free space , as the former are ` dressed ' or ` renormalized ' by their interactions with the medium . thus , the nn potential in a medium of density @xmath81 , usually denoted by @xmath82 , should be different from that in free space . the well - known brown - rho ( br ) scaling mechanism @xcite provides a schematic framework for deriving such density - dependent interactions . a main result is the universal scaling relation for nucleon and meson masses : @xmath83 where @xmath84 and @xmath85 denote respectively the in - medium and in - vacuum masses , and the parameter @xmath17 has the value @xmath86 . this scaling naturally renders @xmath1 a density - dependent interaction @xmath82 . as we shall describe later , it is remarkable that the above simple scaling law , derived in the context of chiral symmetry restoration in dense matter , would have important consequences for traditional nuclear structure physics . we have carried out several studies on the effects of br scaling on finite nuclei , nuclear matter and neutron stars @xcite . let us just briefly describe a few of them . for convenience in implementing br scaling , we have employed the bonna and/or bonns one - boson - exchange potentials @xcite whose parameters for @xmath79 , @xmath31 and @xmath80 mesons are scaled with the density ( in our calculations the meson masses and cut - off parameters are equally scaled ) . note that @xmath78 is protected by chiral symmetry and is not scaled . that we scale @xmath79 but not @xmath78 has an important consequence for the tensor force , which plays an important role in the famous gamow - teller ( gt ) matrix element for the @xmath87 @xmath88 @xmath89-decay @xcite . the tensor force from @xmath78- and @xmath79-meson exchange are of opposite signs . a lowering of only @xmath90 , but not @xmath91 , can significantly suppress the net tensor force strength and thus largely diminish the gt matrix element . in addition , the scaling of the @xmath31 meson introduces additional short - distance repulsion into the nucleon - nucleon interaction , which was found to also contribute to the suppression @xcite . with br - scaling we were able to satisfactorily account for the anomalously - long @xmath92-yr lifetime of this decay @xcite . it may be noted that the br scaling as given in eq . ( 65 ) is a linear scaling , and it may be suitable for low densities near @xmath93 only ( @xmath94 of eq . ( 67 ) may become negative for large @xmath81 ) . for high densities , we need a different scaling such as the new - br scaling @xcite . before describing this scaling , let us first describe briefly a @xmath2 ring - diagram formalism @xcite on which our nuclear - matter calculations with the new - br scaling are based . in this ring - diagram formalism @xcite the nuclear - matter ground - state energy is given by the all - order sum of the particle - particle hole - hole ( @xmath95 ) ring diagrams as shown in fig . the ground - state energy of asymmetric nuclear matter is expressed as @xmath96 where @xmath97 denotes the energy for the non - interacting system and @xmath98 is the energy shift due to the nn interaction . we include in general three types of ring diagrams , the proton - proton , neutron - neutron and proton - neutron ones . the proton and neutron fermi momenta are , respectively , @xmath99 and @xmath100 , where @xmath101 and @xmath102 denote respectively the proton and neutron density . the isospin asymmetry parameter is @xmath103 . with such ring diagrams summed to all orders , we have [ eng ] e(n,)&=&_0 ^ 1 d_m _ ijkl < y_m(ij , ) y_m^*(kl , ) + & & ij|v_low - k|kl , where @xmath104 is a strength parameter , integrated from 0 to 1 . the transition amplitudes @xmath105 are obtained from a @xmath95 rpa equation @xcite . these @xmath105 amplitudes represent & & y_m(,)=_0|a_^ a_^|_m , + & & ( , ) = ( h , h ) or ( p , p ) , where @xmath106 and @xmath29 are respectively the ground state and @xmath85th excited state of nuclear matter with two additional particles or holes . @xmath107 is a measure of the particle - particle excitations in the ground state . setting @xmath107=0 means the ground state is taken to be the closed fermi sea , and the above ring - diagram method reduces to the usual hf method where only the first - order ring diagram ( diagram ( a ) of fig . 3 ) is included for the energy shift . in this case , the above energy shift becomes @xmath108 where @xmath109=(1,0 ) if @xmath110 for protons and @xmath109=(1,0 ) if @xmath111 for neutrons . the above @xmath2 ring - diagram framework has been applied to symmetric and asymmetric nuclear matter @xcite and to the nuclear symmetry energy @xcite . this framework has also been tested by applying it to dilute cold neutron matter in the limit that the @xmath112 scattering length of the underlying interaction approaches infinity @xcite . this limit which is a conformal fixed point is usually referred to as the unitary limit . for many - body systems at this limit , the ratio @xmath113 is expected to be a universal constant of value @xmath114 . ( @xmath115 and @xmath116 are , respectively , the interacting and non - interacting ground - state energies of the many - body system . ) the above ring - diagram method has been used to calculate neutron matter using several very different unitarity potentials ( a unitarity cdbonn potential obtained by tuning its meson parameters , and several square - well unitarity potentials ) @xcite . the @xmath117 ratios given by our calculations for all these different unitarity potentials are all close to 0.44 , in good agreement with the quantum - monte - carlo results ( see @xcite and references quoted therein ) . in fact our ring - diagram results for @xmath117 are significantly better than those given by hf and bhf ( brueckner hf ) @xcite . the above unitary calculations have provided satisfactory results , supporting the reliability of our @xmath2 ring - diagram framework for calculating the nuclear matter eos . we now describe the new - br scaling @xcite on which the results shown in fig . 15 are based . the idea behind this scaling is that when a large number of skyrmions as baryons are put on an fcc ( face - centered - cubic ) crystal to simulate dense matter , the skyrmion matter undergoes a transition to a matter consisting of half - skyrmions @xcite in bcc ( body - centered - cubic ) configuration at a density that we shall denote as @xmath118 . this density is difficult to pin down precisely but it is more or less independent of the mass of the dilaton scalar , the only low - energy degree of freedom that is not well - known in free space . the density at which this occurs has been estimated to lie typically between 1.3 and 2 times normal nuclear matter density @xmath93 @xcite . in our model , nuclear matter is separated into two regions i and ii respectively for densities @xmath119 and @xmath120 . as inferred by our model , they have different scaling functions @xmath121 it has been found that both the br and new - br scalings are important for nuclear - matter saturation @xcite . the eos shown by ( a ) and ( b ) of fig . 15 are obtained with the new - br scaling with @xmath122 and 2.0@xmath93 respectively . both give an energy per nucleon @xmath123 mev , saturation density @xmath124 and compression modulus @xmath125=208 mev , all in satisfactory agreement with the empirical values @xcite . we believe that this scaling works well for low densities of @xmath126 . as described in @xcite , we employ in our new - br scaling calculations the bonns potential @xcite with scaling parameters @xmath127=0.13 , @xmath128=0.121 , @xmath129=0.139 , @xmath130=0.13 and @xmath131=0 for region i , where @xmath132 is the associated scaling of the meson - nucleon coupling constant . for region ii the scaling parameters are @xmath127=0.13 , @xmath128=0.121 , @xmath129=0.139 , @xmath133=0.13 , @xmath134=0 and @xmath135=0.77 for ( a ) and 0.78 for ( b ) . note that this scaling has some special features : in region i the coupling constants @xmath136 and @xmath137 are not scaled , while in region ii only the coupling constant @xmath136 is scaled . also in region ii the nucleon mass is a density - independent constant as given earlier . note that our choices for the @xmath17 parameters are consistent with the ericson scaling which is based on a scaling relation for the quark condensate @xmath138 @xcite . according to this scaling , at low densities one should have @xmath139 with @xmath140 . > from heavy - ion collision experiments , danielewicz _ et al . _ @xcite have obtained constraints for the pressure - density eos @xmath141 of nuclear matter up to densities @xmath142 . to test this new - br scaling in the high - density region , the pressure eos @xmath141 up to the above densities has been calculated . @xcite the results are presented in fig . 16 where the calculated @xmath141 for neutron matter ( upper panel ) and symmetric nuclear matter ( lower panel ) are compared with the danielewicz constraints @xcite . for neutron matter , there are two constraints , one for stiff eos ( upper box ) and the other for soft eos ( lower box ) . the eoss calculated with parameters a and b are denoted by solid- and open - square respectively . as seen , the results are in satisfactory agreements with the empirical constraints of @xcite . also based on such experiments , there has been much progress in determining the nuclear symmetry energy @xmath143 up to densities as high as @xmath144 @xcite . thus an application of the new - br scaling to the calculation of @xmath143 would provide an important test for this scaling in the region with @xmath120 . as displayed in fig . 17 , the symmetry energies calculated from the new - br eos @xcite are in good agreement with the experimental constraints of li et al . @xcite and tsang et al . as seen , there are two constraints from li , one for stiff eos and the other for soft eos . the calculated results are close to the stiff constraint at high denities , but slightly lower than both at low densities . the eos at high densities ( @xmath145 ) is important for neutron - star properties . thus an application of the new - br scaling to neutron star structure would provide a further test . as shown in fig . 18 , the maximum mass of pure - neutron stars calculated from the new - br eos is about @xmath146 , slightly larger than the maximum mass @xmath147 observed in nature @xcite . the central density of the neutron star is @xmath144 . at densities as high as this central density , how to scale the hadrons with the medium is an interesting question and should be further studied . we note that the above scaling is only ` inferred ' by the skyrmion - half - skyrmion model @xcite . although the initial results obtained with this model are promising , further studies of this model should be useful and of interest . for densities near @xmath93 , the br scaling of eq . ( 67 ) and new - br scaling of eq . ( 70 ) are practically the same , both rendering the nn interaction into a density - dependent one . this density dependence has played an important role in describing nuclear properties such as nuclear - matter saturation . @xcite it is of interest that this importance was already recognized in the well - known skyrme effective interaction @xcite of the form v_skyrme = v_sky(r_1 -r_2 ) + t_3 ( r_1 - r_2 ) ( r_2 -r_3 ) , where @xmath148 is a zero - range two - body interaction . the second term is a zero - range three - body interaction , and by integrating out one participating nucleon over the fermi sea it becomes a density - dependent two - body force @xmath149 , namely @xmath150 where @xmath151 and both @xmath152 and @xmath153 are strength parameters . in calculations with skyrme interactions , the inclusion of @xmath149 is essential for nuclear matter saturation . as observed in ref . @xcite the combined potential given by the sum of the unscaled-@xmath1 and @xmath149 , the skyrme density - dependent interaction shown above , can give equally satisfactory nuclear matter saturation properties as the br - scaled @xmath1 . ( we use @xmath1 to denote the two - nucleon interaction . ) similar equivalence is also noted for the new - br scaling , as illustrated in fig . there a qualitative agreement is seen up to densities @xmath154 between the eos for symmetric nuclear matter calculated with a new - br - scaled @xmath1 and that with a combined interaction ( @xmath155 ) . ( the parameters of @xmath156 and @xmath153=0 are used for the @xmath149 calculation of fig . 19 . ) we now come to the three - nucleon force @xmath5 . as we have just discussed , nuclear matter calculations with either br or new - br scaled @xmath1 can satisfactorily describe nuclear matter saturation properties but not so with the unscaled @xmath1 . it is well known that there are many - nucleon forces @xmath5 and beyond . especially for nucleons in dense medium , the effects of such many - body forces are likely to be important . can the calculations using @xmath1 ( unscaled ) plus @xmath5 also give satisfactory nuclear saturation properties ? in the following we shall study this question . the lowest - order ( nnlo ) chiral three - nucleon interaction @xmath5 will be considered . this interaction is of the form @xmath157 where @xmath158 @xmath159 @xmath160 where @xmath161 , @xmath162 mev , @xmath163 mev , @xmath164 mev , @xmath165 is the difference between the final and initial momentum of nucleon @xmath166 and f_ijk^ & = & ^(-4c_1m_^2 + 2c_3 _ i _ j ) + & & + c_4 ^ _ k^_k ( _ i _ j ) . [ 3n4 ] the parameters @xmath167 are constrained by @xmath168 scattering and can then be fitted within these uncertainties to peripheral nucleon - nucleon scattering phase shifts @xcite . the parameters @xmath169 and @xmath170 must be determined by three - nucleon observables . navratil et al . @xcite fit to the binding energies of @xmath171 and @xmath172 , which determined a curve of consistent @xmath169 and @xmath170 values . to uniquely determine @xmath169 and @xmath170 , one can fit also the @xmath89-decay lifetime of @xmath171 @xcite . recently coraggio et al . @xcite have carried out extensive nuclear matter calculations using a chiral @xmath1 @xcite and the above @xmath5 with its @xmath169 and @xmath170 parameters determined also from the binding energies of three - nucleon systems and the lifetime of @xmath171 . the values of the ( @xmath169 , @xmath170 ) parameters so determined are @xmath173 , @xmath174 and @xmath175 respectively for variations in the resolution scale defined by the momentum - space cutoff @xmath176 mev @xcite , where @xmath177 enters the regulating function f(p,p)=exp[-(p/)^2n-(p/)^2n ] [ regfunc ] that multiplies the nucleon - nucleon potential . in eq . ( [ regfunc ] ) @xmath178 and @xmath11 denote the relative momenta of the incoming and outgoing nucleons , and the regulator exponent @xmath81 for the above three values of @xmath177 are @xmath179 respectively . as shown above , the @xmath169 and @xmath170 parameters for @xmath180 mev do not differ largely from each other . we shall denote the two- and three - nucleon potentials corresponding to a chosen @xmath177 as @xmath181 and @xmath182 respectively . in refs . @xcite it was shown that low - momentum chiral potentials at the resolution scales @xmath183 and 450 mev have perturbative properties similar to the renormalization - group evolved potential @xmath2 . in the present work , we have employed only one of them , the @xmath177=414 set . we plan to carry out calculations using the other two @xmath177 choices in a future work . by integrating out one participating nucleon over the fermi sea , holt , kaiser and weise @xcite have reduced @xmath5 to a density - dependent two - body force @xmath184 . comparing with @xmath5 , @xmath184 is much more convenient for nuclear many - body calculations . briefly speaking , they are related by @xmath185 it is seen that @xmath184 is a density ( @xmath186 ) dependent two - body interaction . the basis states @xmath187 , @xmath188 , @xmath189 in eq . ( 78 ) are all anti - symmetrized and normalized . note as well that there is a @xmath81-body counting factor @xmath190 to be included in calculations : we have @xmath190 = ( 1 , 1/2 , and 1/3 ) respectively for ( 2- , 1- , 0- ) body vertices . thus the vertex in ( a ) of fig . 3 is ( @xmath191 ) , and each vertex in ( b ) and ( c ) is ( @xmath192 ) . in the 1-body diagram @xmath193 of fig . 4 the vertex is @xmath194 . as an initial application , we have performed ring - diagram calculations of nuclear matter eos and nuclear symmetry energy using the combined potential @xmath195 plus @xmath184 derived from @xmath196 . the ring - diagram method used here is the same as used earlier for the new - br calculations . the resulting eos for symmetric nuclear matter is presented in fig . 19 ( lower panel ) , denoted by ` @xmath197 ' . this eos has ground - state energy per nucleon @xmath198 mev , compression modulus @xmath199 mev and saturation density @xmath200 , all in satisfactory agreement with the empirical values . our eos calculated with @xmath196 has worked well in the low - density region near @xmath93 . in this region , this eos and the new - br one shown in the upper panel of the figure agree well with each other , indicating that the new - br - scaled @xmath1 and the combined potential @xmath201 can both describe nuclear matter at low density near @xmath93 satisfactorily . but for densities beyond @xmath202 , the eos given by these two potentials have significant differences as shown by fig . 19 . in the lower panel of fig . 19 , we also present the eos given by the combined potential ( @xmath203 ) where @xmath149 is the same skyrme - type density - dependent interaction used in the upper panel of the figure . the same ring - diagram method is used for both eos . as seen , this eos and the one from ( @xmath204 ) are in good agreement for densities lower than @xmath205 , indicating that for such low densities the effect of @xmath5 can be well reproduced by @xmath149 . but at higher densities , the @xmath149 eos is significantly lower . to have better agreement between the two , we may need to use a @xmath149 with different @xmath152 and @xmath153 parameters . this possibility is being studied by us and some recent work in this direction can be found in refs . @xcite . medium modifications to the nucleon - nucleon interaction resulting from brown - rho scaling or three - nucleon forces have also succeeded in explaining the very long lifetime ( @xmath92yrs ) of the @xmath87 @xmath88 @xmath89-decay . the expected lifetime is only several hours , and in order to achieve the empirical lifetime a precise cancellation in the gamow - teller matrix element to the order of one part in a thousand is required . calculations using @xmath1 ( unscaled ) find only a small suppression @xcite of the transition , and it is only with the inclusion of medium modifications that it has been possible to explain the long lifetime @xcite . based on the above @xmath201 eos we have also calculated the nuclear symmetry energy @xmath143 using a method as described in ref . lattimer and lim @xcite have studied the constraints on @xmath143 and @xmath64 ( defined as @xmath206 ) . at density @xmath207 their constraints are @xmath208 and @xmath209 . at @xmath207 , our results are @xmath210 mev and @xmath211 mev , both in good agreement with the lattimer - lim constraints . at densities higher than @xmath202 , our calculated @xmath143 exhibits , however , a supersoft behavior ( namely it decreases with density after reaching a maximum near @xmath212 ) . as discussed in refs . @xcite , predictions of @xmath143 at high densities are rather diverse , depending on the nucleon - nucleon interactions employed . further studies of @xmath143 at densities beyond @xmath202 are certainly needed and of interest . as illustrated by the pioneering work of talmi @xcite , nuclear shell - model calculations in small model spaces can describe empirical nuclear properties highly successfully . an important ingredient in such calculations is the use of empirically determined effective interactions . it is of interest to see if such interactions can be derived microscopically from an underlying nn interaction and how . this question has been widely studied in past years . in this paper several developments related to this question have been discussed . first we reviewed the @xmath213-box folded - diagram theory , with which the full - space nuclear hamiltonian can be reduced to a model - space effective one of the form @xmath214 . here @xmath215 is the sp hamiltonian extracted from experimental sp energies , and the effective interaction is given by a folded - diagram expansion in terms of the irreducible vertex function @xmath213-box . methods for summing up this series for both degenerate and non - degenerate model spaces are discussed . to calculate @xmath0 , a first step is to calculate the @xmath213-box , which is composed of irreducible valence - linked diagrams , from the input nn interactions . we have reviewed the low - momentum interaction @xmath2 which has been commonly used in the @xmath0 calculations . this interaction has been obtained from realistic @xmath1 by integrating out their high - momentum components beyond a decimation scale @xmath177 using a renormalization - group procedure . the resulting @xmath2 interaction is a smooth potential ( without strong repulsive core ) which is convenient for calculations . furthermore such interactions derived from different @xmath1 models are nearly the same , for @xmath177 less than @xmath216 , leading to a nearly unique low - momentum interaction . as indicated by the few sample calculations presented earlier , shell - model calculations using the effective interaction derived from @xmath2 have indeed worked rather successfully in describing low - energy experimental data . turning to nuclear systems at densities near and above @xmath93 , a long standing problem has been that calculations with the two - nucleon potential @xmath1 alone are unable to give satisfactory saturation properties . we have demonstrated that this shortcoming can be amended by including either the new - br scaling or @xmath196 . as illustrated in fig . 19 , the eos from new - br ( @xmath217 ) and ( @xmath197 ) are close to each other for densities below @xmath205 . both can reproduce the empirical nuclear saturation properties well . in our eos calculations we have employed a ring - diagram method where the particle - particle hole - hole ring diagrams are summed to all orders . the symmetry energy at densities near @xmath93 can also be well reproduced by the above two approaches . there may be some underlying equivalence between new - br scaling and @xmath5 at low densities , and its further study will be interesting as well as useful . a similar equivalence is observed concerning the famous gamow - teller ( gt ) matrix element for the @xmath87 @xmath88 @xmath89-decay , which has an anomalously long half - life of @xmath218yrs . the use of the br - scaled @xmath1 has been crucial in highlighting the important role played by medium effects not normally included in many - body perturbation theory calculations with free - space two - body interactions . these medium effects can be recast in the language of three - body forces , which have been equally successful in describing the transition . but at densities higher than @xmath202 , the nuclear - matter eos given by these two approaches are significantly different , as indicated by fig . 19 . at such high densities , the symmetry energies given by them are also very different . studies of these differences and their impact on neutron star properties , such as the mass - radius relationship and maximum mass , will be of interest . we have compared the new - br and @xmath5 eoss with the eos given by ( unscaled-@xmath1 + @xmath149 ) . it is observed that all three are in good qualitative agreement in the density region below @xmath205 . that the density - dependent effects on symmetric nuclear matter near @xmath93 from the new - br scaling and from @xmath5 may be well reproduced by an empirical density - dependent force of the skyrme type is an interesting result , indicating that the empirical skyrme force may have a microscopic connection with the new - br scaling and/or @xmath5 . 0.3 cm * acknowledgement * we would like to acknowledge the inspiration aage bohr and ben mottelson have been to our field over the decades , as well as to us personally . two of us ( ttsk and eo ) cherish vivid memories from our visits to copenhagen during the golden years of nuclear physics . we are grateful to mannque rho , ruprecht machleidt and luigi coraggio for many helpful discussions . this work was supported in part by the department of energy under grant no . de - fg02 - 88er40388 and de - fg02 - 97er-41014 . 99 a. de - shalit and i. talmi , nuclear shell theory ( academic press , new york and london , 1963 ) . g. e. brown and t. t. s. kuo , nucl . a92 * , 481 ( 1967 ) . t. t. s. kuo and g. e. brown , nucl . phys . * 85 * , 40 ( 1966 ) . m. hjorth - jensen , t. t. s. kuo and e. osnes , rev . phys . * 261 * , 205 ( 1995 ) . l. coraggio , a. covello , a. gargano , n. itaco and t. t. s. kuo , prog . * 62 * , 135 ( 2009 ) . m. lacombe et al . , phys . c * 21 * , 861 ( 1980 ) . r. machleidt , adv . * 19 * , 189 ( 1989 ) . r. machleidt , phys . c * 63 * , 024001 ( 2001 ) . r. b. wiringa , v. g. j. stoks and r. schiavilla , phys . c * 51 * , 38 ( 1995 ) . v. g. j. stoks and r. a. m. klomp , c. p. f. terheggen and j. j. de swart , phys . c * 49 * , 2950 ( 1994 ) . d. r. entem , r. machleidt and h. witala , phys . c * 65 * , 064005 ( 2002 ) . t. t. s. kuo and e. osnes , _ lecture notes in physics _ * vol.364 * ( springer - verlag 1990 ) p.1 . t. t. s. kuo , s. y. lee and k. f. ratcliff , nucl . phys . * a176 * , 65 ( 1971 ) . j. goldstone , proc . ( london ) * a293 * , 267 ( 1957 ) . l. w. siu , j. w. holt , t. t. s. kuo and g. e. brown , phys . c * 79 * , 054004 ( 2009 ) . h. dong , t. t. s. kuo and r. machleidt , phys . c * 80 * , 065803 ( 2009 ) . h. feshbach , ann . * 8 * , 49 ( 1958 ) . m. morita , prog . * 29 * , 351 ( 1963 ) . b. h. brandow , rev . * 39 * , 771 ( 1967 ) . m. b. johnson and m. baranger , ann . ( n.y . ) * 62 * , 172 ( 1971 ) . i. lindgren , j. phys . * b7 * , 2441 ( 1974 ) . j. w. holt , g. e. brown , t. t. s. kuo , j. d. holt and r. machleidt , phys . * 100 * , 062501 ( 2008 ) . j. w. holt , n. kaiser and w. weise , phys . c * 79 * , 054331 ( 2009 ) . j. shurpin , t. t. s. kuo and d. strottman , nucl . phys . * a408 * , 310 ( 1983 ) . e. m. krenciglowa and t. t. s. kuo , nucl . * a235 * , 171 ( 1974 ) . s. y. lee and k. suzuki , phys . lett . * b91 * , 173 ( 1980 ) . k. suzuki and s. y. lee , prog . . phys . * 64 * , 2091 ( 1980 ) . k. suzuki , r. okamoto , p. j. ellis and t. t. s. kuo , nucl . phys . * a567 * 576 ( 1994 ) . t. t. s. kuo , f. krmpotic , k. suzuki and r. okamoto , nucl . a582 * , 126 ( 1995 ) . n. tsunoda , k. takayanagi , m. hjorth - jensen and t. otsuka , phys . c * 89 * , 024313 ( 2014 ) . h. dong , t. t. s. kuo and j. w. holt , nucl . a930 * , 1 ( 2014 ) . r. okamoto , k. suzuki , h. kumagai and s. fujii , j. phys . . ser . * 267 * , 012017 ( 2010 ) . k. suzuki , r. okamoto , h. kumagai and s. fujii , phys . c * 83 * , 024304 ( 2011 ) . s. bogner , t. t. s. kuo and l. coraggio , nucl . a684 * , 432c ( 2001 ) . s. bogner , t. t. s. kuo , l. coraggio , a. covello and n. itaco , phys . c * 65 * , 051301(r ) ( 2002 ) . s. k. bogner , a. schwenk , r. j. furnstahl and a. nogga , nucl . phys . * a763 * , 59 ( 2005 ) . h. dong , t. t. s. kuo , h. k. lee , r. machleidt , m. rho , phys . c * 87 * , 054332 ( 2013 ) . r. machleidt , `` computational nuclear physics 2nuclear reactions '' ( langanke , maruhn , koonin , eds . , springer ny 1993 ) , chap.1 , p.1 . g. e. brown and m. rho , phys . * 66 * , 2720 ( 1991 ) . t. hatsuda and s. h. lee , phys . c * 46 * , r34 ( 1992 ) . g. e. brown and m. rho , phys . * 269 * , 333 ( 1996 ) . g. e. brown and m. rho , phys . * 363 * , 85 ( 2002 ) . g. e. brown and m. rho , phys . rep . * 396 * , 1 ( 2004 ) . k. hebeler , s. k. bogner , r. j. furnstahl , a. nogga and a. schwenk , phys . c * 83 * , 031301(r ) ( 2011 ) . j. w. holt , n. kaiser and w. weise , nucl . phys . * a876 * , 61 ( 2012 ) . a. gardestig and d.r . phillips , phys . lett . * 96 * , 232301 ( 2006 ) . d. gazit , phys . lett . * 666 * , 472 ( 2008 ) . l. e. marcucci , a. kievsky , a. rosati , s. schiavilla and m. viviani , phys . 108 * , 052502 ( 2012 ) . l. coraggio , j. w. holt , n. itaco , r. machleidt and f. sammarruca , phys . rev . c * 87 * , 014322 ( 2013 ) . l. coraggio , j. w. holt , n. itaco , r. machleidt , l. e. marcucci and f. sammarruca , phys . c * 89 * , 044321 ( 2014 ) . j. w. holt , g. e. brown , j. d. holt and t. t. s. kuo , nucl . * a785 * , 322 ( 2007 ) . h. dong , t.t.s . kuo and r. machleidt , c * 83 * , 054002 ( 2011 ) . j. w. holt , n. kaiser and w. weise , phys . c * 81 * , 024002 ( 2010 ) . h. dong , t. t. s. kuo , h. k. lee , r. machleidt and m. rho , phys . c * 87 * , 054332 ( 2013 ) . l. w. siu , t. t. s. kuo and r. machleidt , phys . c * 77 * , 034001 ( 2008 ) . h. dong , l. w. siu , t. t. s. kuo and r. machleidt , phys . c * 81 * , 034003 ( 2010 ) . a. s. goldhaber and n. s. manton , phys . * b198 * , 231 ( 1987 ) . b. y. park , d. p. min , m. rho and v. vento , nucl . a * 707 * , 381 ( 2002 ) ; h. j. lee _ et al . _ , nucl . a * 723 * , 427 ( 2003 ) ; m. rho , arxiv:0711.3895 [ nucl - th ] . m. ericson , phys . b301 * , 11 ( 1993 ) . p. danielewicz , r. lacey , and w. a. lynch , science * 298 * , 1592 ( 2002 ) . b. a. li and l. w. chen , phys . c * 72 * , 064611 ( 2005 ) . b. a. li , l. w. chen and c. m. ko , phys . rep . * 464 * , 113 ( 2008 ) . m. b. tsang , y. zhang , p. danielewicz , m. famiano , z. li , w. g. lynch and a. w. steiner , phys . lett . * 102 * , 122701 ( 2009 ) . p. b. demorest et al . , nature * 467 * , 1081 ( 2010 ) . j. antoniadis et al . , science * 340 * , 6131 ( 2013 ) t. t. s. kuo , j. phys . . ser . * 580 * , 012027 ( 2015 ) . p. ring and p. schuck , _ the nuclear many - body problem _ ( springer - verlag , new york , 1980 ) . d. r. entem and r. machleidt , phys . c * 68 * , 041001 ( 2003 ) . p. navratil , v. g. gueorguiev , j. p. vary , w. e. ormand and a. nogga , phys . 99 * , 042501 ( 2007 ) . j. m. lattimer and y. lim , astrophys . j. * 771 * , 51 ( 2013 ) . b. a. brown and a. schwenk , phys c * 89 * 011307 ( 2014 ) . e. rrapaj , a. roggero and j. w. holt , arxiv:1510.00444 . s. aroua et al . , nucl . phys . * a720 * 71 ( 2003 ) . p. maris , j. p. vary , p. navratil , w. e. ormand , h. nam and d. j. dean , phys . lett . * 106 * 202502 ( 2011 ) . b. a. brown , phys . rev . lett . * 85 * , 5296 ( 2000 ) . z. xiao et al . , phys . lett . * 102 * , 062502 ( 2009 ) , b. a. li , l. w. chen , d. h. wen , z. xiao , c. xu , g. c. yong and m. zhang , nucl . phys . * a834 * , 509c ( 2010 ) .
model - space effective interactions @xmath0 derived from free - space nucleon - nucleon interactions @xmath1 are reviewed . we employ a double decimation approach : first we extract a low - momentum interaction @xmath2 from @xmath1 using a @xmath3-matrix equivalence decimation method . then @xmath0 is obtained from @xmath2 by way of a folded - diagram effective interaction method . for decimation momentum @xmath4 , the @xmath2 interactions derived from different realistic @xmath1 models are nearly model independent , and so are the resulting shell - model effective interactions . for nucleons in a low - density nuclear medium like valence nucleons near the nuclear surface , such effective interactions derived from free - space @xmath1 are satisfactory in reproducing experimental nuclear properties . but it is not so for nucleons in a nuclear medium with density near or beyond nuclear matter saturation density . in this case it may be necessary to include the effects from brown - rho ( br ) scaling of hadrons and/or three - nucleon forces @xmath5 , effectively changing the free - space @xmath1 into a density - dependent one . the density - dependent effects from br scaling and @xmath5 are compared with those from empirical skyrme effective interactions .
we consider the basic multiple - access communication problem in a distributed setting . in gallager s survey paper @xcite , it is pointed out that the multiple - access problem has been studied from a number of different perspectives , each having its own advantages and shortcomings . in the data networking community , a well - known distributed multiple access scheme is aloha @xcite . in aloha , it is assumed that a collision " happens whenever more than one user transmit simultaneously . those packets involved in a collision are discarded and retransmitted according to some retransmission probability . the collision channel model , however , does not accurately describe the underlying physical multiple - access channel . it is well known that there exist coding techniques which can decode multiple users messages when simultaneous transmissions occur . indeed , more sophisticated models such as signal capture @xcite , spread aloha @xcite , and multi - packet reception @xcite have been developed to enhance aloha . even these improved schemes , however , are not optimal from the viewpoint of information theory . in the information theory literature , the capacity regions of various multiple - access channel ( mac ) models have been characterized ( see @xcite ) . rate splitting multiple - access techniques ( or generalized time - sharing ) are presented in @xcite to achieve every point in the gaussian or the discrete memoryless mac capacity region using only single - user codes . these schemes , however , require a pre - defined decoding order , which makes distributed implementation difficult . finally , in the spread spectrum community , cdma techniques are adopted . here , users are decoded regarding all other users signals as interference . this , however , is not optimal from the information theoretic viewpoint . to address some of the shortcomings mentioned above , medard _ et al . _ @xcite use information - theoretic techniques to analyze different notions of capacity for time - slotted aloha systems . a coding / decoding scheme which combines rate splitting and superposition coding is constructed . this scheme allows some bits to be reliably received even when collision occurs , and more bits to be reliably received in the absence of collisions . shamai @xcite proposes a similar scheme to apply a broadcast strategy to multiple - access channel under static fading where the fading coefficients are not available to the transmitters or the receiver . to implement the scheme in @xcite , however , a pre - defined decoding order is required , as in @xcite . in @xcite , cheng proposes a distributed scheme called `` stripping cdma '' for the @xmath1 out of @xmath2 gaussian mac . here , no pre - defined decoding order is required . it is shown in @xcite that stripping cdma is asymptotically optimal , although the optimal operating parameters are not specified . in this paper , we investigate distributed multiple - access schemes based on the idea of rate splitting for both the @xmath0-user additive white gaussian noise mac and the @xmath0-user discrete memoryless mac . we characterize the optimal operating parameters as well as the asymptotic optimality of these schemes from the viewpoint of information theory . assume that every user has an infinite backlog of bits to send , and that every user knows the total number of users @xmath0 . we propose a distributed scheme , called _ distributed rate splitting _ ( drs ) , to achieve the optimal communication rates allowed by information theory . in this scheme , each real user creates a number of virtual users via a power / rate splitting mechanism in the @xmath0-user gaussian channel or via a random switching mechanism in the @xmath0-user discrete memoryless channel . at the receiver , all virtual users are successively decoded . a possible advantage of the drs scheme is that it can be implemented with lower complexity when compared with multiple - access schemes such as joint coding and less coordination when compared with time - sharing and rate splitting . in sections [ sec : gau ] and [ sec : dmc ] , we focus first on symmetric situations where the channel capacity regions are symmetric and every real user creates the same number of virtual users . in this case , the drs scheme entails the following . each user @xmath3 creates @xmath1 virtual users indexed by @xmath4 , @xmath5 . the virtual user class @xmath6 consists of users @xmath7 ( i.e. we have altogether @xmath1 virtual user classes and there are @xmath0 virtual users in each class ) . in the @xmath0-user gaussian mac , virtual users are created via a power / rate splitting mechanism . the signal transmitted by a real user is the superposition of all its virtual users signals . the receiver receives the sum of the virtual users signals plus noise . all virtual users are then successively decoded in increasing order of their class . that is , all virtual users in class @xmath6 , @xmath8 , are decoded before any virtual user in @xmath9 , where @xmath10 , is decoded . in contrast to @xcite , the optimal operating parameters , such as power and rate , are explicitly specified for any finite @xmath1 . in the @xmath0-user discrete memoryless mac , virtual users with the same input distribution as the real users are created , and the transmitted signal of a real user is determined by a random switch . the receiver successively decodes all virtual users in increasing order of their class given the side information of already decoded virtual users . the optimal switch is found for any finite @xmath1 for the @xmath11-user case . finally , it is shown that for both channel models , the rate tuple achieved by the drs scheme converges to the maximum equal rate point allowed by the information - theoretic bound as the number of virtual users per real user tends to infinity . next , in section [ sec : variation ] , we consider more general situations where the capacity regions can be asymmetric and real users may generate different numbers of virtual users . for the case of asymmetric capacity regions , new operating parameters are specified for any finite number of virtual users per real user . we show that the drs scheme still can achieve a point on the dominant face as the number of virtual users per real user tends to infinity . for the case of unequal number of virtual users per real user , we present a variation of drs which supports differential user rate requirements in a distributed manner . in this new scheme , each user @xmath3 , independently from other users , generates @xmath12 virtual users according to its own rate requirement . all virtual users are then decoded reliably at the receiver . furthermore , as each real user generates more virtual users , the rate tuple achieved under this variation of drs converges the maximum equal rate point on the dominant face . we first examine a gaussian mac with a symmetric capacity region . later in section [ sec : variation ] , we consider the asymmetric case . consider an @xmath0-user gaussian mac where each transmitter has transmission power @xmath13 and the receiver has noise variance @xmath14 . the capacity region @xmath15 is the set of @xmath16 satisfying to denote @xmath17 throughout . ] @xmath18 where @xmath19 is the cardinality of the set @xmath20 . the _ dominant face _ @xmath21 is the subset of rate tuples which gives equality in for @xmath22 . for this symmetric setting , it is easy to see that the maximum common rate that every user can achieve is achievable if for any @xmath23 , there exists an @xmath24 multiple - access code with overall error probability @xmath25 , where @xmath26 is the block length . ] is @xmath27 . it is well - known that rate tuples on the dominant face other than the vertices can not be achieved via standard successive decoding @xcite . note that the optimal rate tuple @xmath28 , called the _ maximum equal rate point _ , is such a point . for the two - user gaussian mac , the maximum equal rate point is shown in fig . [ fig : twogaussian ] . currently , three methods are known to achieve general points on the dominant face : joint encoding / decoding , time - sharing , and rate - splitting . joint encoding / decoding is not practical because of its high complexity @xcite . in time - sharing , all @xmath0 users need to coordinate their transmissions . therefore , some communication overhead is required . the rate - splitting method in @xcite achieves every point in @xmath29 via a generalized successive decoding scheme . for the two - user case , user @xmath30 creates two virtual users , say @xmath31 and @xmath32 , by splitting its power @xmath13 into @xmath33 and @xmath34 and setting @xmath35 , @xmath36 . user @xmath11 does not split its power and sets its rate to @xmath37 . the decoding order is @xmath38 . in order to achieve the maximum equal rate point , we solve @xmath39 , yielding @xmath40 . thus , both time - sharing and rate splitting require some coordination among users . in this paper , we focus on _ distributed _ multiple - access communication schemes . in particular , we introduce the _ distributed rate splitting _ ( drs ) scheme . the drs scheme offers the possibility of multiple - access communication with lower complexity when compared with joint coding , and communication with less coordination when compared with the time - sharing or rate splitting method . moreover , we show that drs can achieve the maximum equal rate point of the mac capacity region asymptotically . we now formally present the drs scheme . in this scheme , each user creates @xmath1 virtual users by splitting its power @xmath13 into @xmath41 , where @xmath42 is the power allocated to the @xmath43th virtual user and @xmath44 . each user then assigns transmission rate @xmath45 to virtual user @xmath43 . note that the proposed drs scheme is symmetric , i.e. all @xmath0 users split their powers and set their rates in the same way . the signal transmitted by a user is the superposition of its virtual users signals . as defined in section [ sec : intro ] , virtual user class @xmath6 consists of all virtual users indexed by @xmath43 . the receiver receives the sum of all virtual users signals plus noise . all virtual users are then successively decoded in increasing order of their class . to illustrate the drs scheme , consider the case @xmath46 . each real user splits its power @xmath13 into @xmath33 and @xmath34 . notice there are two major differences between our scheme and the traditional rate splitting scheme in @xcite . first , in our scheme , all real users split in the same way , whereas there is at least one user who does not split in the traditional rate splitting scheme . second , virtual users in the same class , ( i.e. with the same index @xmath43 ) , are allocated the same rate in our scheme , whereas all virtual users have different rates according to the pre - defined decoding order in the traditional rate splitting scheme . is not optimal . later in this section , we demonstrate the asymptotic optimality of drs by taking @xmath1 to infinity . ] these differences are illustrated in fig . [ fig : difference ] . since we assume the receiver uses successive decoding method , some virtual user must be decoded first . without loss of generality we assume one of the @xmath33 virtual users is decoded first . for the case @xmath46 , we show that there is a unique way for a real user to split its power in order to maximize its total throughput . for @xmath47 and for a fixed @xmath33 , each real user s throughput is maximized by setting @xmath48 and @xmath49 . _ proof : _ the @xmath33 virtual user who is decoded first must have @xmath50 ( i.e. the virtual user regards all other virtual users as interference ) in order to be decoded successfully . due to symmetry , all other @xmath33 virtual users must have the same @xmath51 . then the problem of maximizing each real user s throughput reduces to @xmath52 , subject to ( i ) @xmath53 , ( ii ) one of the @xmath33 virtual users is decoded first and ( iii ) @xmath54 note that @xmath55 is maximized when the interference plus noise faced by all the @xmath56 virtual users is minimized , and the only way to minimize the interference plus noise faced by all the @xmath56 virtual users is to decode all the @xmath33 virtual users before decoding any @xmath56 virtual user . therefore , the minimum interference plus noise faced by any @xmath56 virtual user is @xmath57 . hence , the maximum rate associated with a @xmath56 virtual user is @xmath58 . @xmath59 + using the drs scheme with @xmath47 , each user can strictly increase its throughput relative to the case where users do not split their powers and decode against each other as noise . this is easily verified by observing that for any @xmath60 , @xmath61 now consider the case where each user creates more than two virtual users @xmath62 . here , we show that each user s throughput increases further . [ le : morevuser ] given a drs scheme with @xmath1 virtual users per real user , where @xmath63 are the virtual users powers , it is possible to strictly increase the throughput via an @xmath64 virtual user system with powers @xmath65 , where @xmath66 . _ proof : _ suppose that every user splits its power into @xmath1 virtual users : @xmath67 subject to @xmath44 , where @xmath1 is an arbitrary integer and @xmath42 is the power of @xmath43th virtual user . since virtual user @xmath1 is decoded last , following the reasoning in the proof of lemma @xmath30 , we have @xmath68 . we now split the virtual user with power @xmath69 into two new virtual users with powers @xmath70 and @xmath71 , where @xmath72 . we set @xmath73 now each real user has @xmath74 virtual users . notice that we do not change the power and decoding order of any of the other virtual users ( i.e. virtual users @xmath75 ) . from a real user s view point , the virtual user with @xmath76 is decoded second to last _ among all virtual users generated by this real user _ and the virtual user with @xmath77 is decoded last . thus , all virtual users can be decoded and from , @xmath78 . therefore , every real user with @xmath1 virtual users can strictly increase its throughput by splitting its power among @xmath74 virtual users . @xmath59 + before we examine the asymptotic behavior of drs , we solve the problem of how to split a user s power _ optimally _ among a fixed number of virtual users . the main difficulty here is that the objective function is not concave . in order to find the optimal splitting method , we prove the following lemma . [ lemma : maximization ] consider the following optimization problem : @xmath79 subject to @xmath42+@xmath80 = @xmath81 and @xmath82 , where @xmath83 , @xmath0 and @xmath81 are positive constants and @xmath84 . the unique solution to is also the unique solution to @xmath85 , where @xmath86 . _ proof : _ substitute @xmath87 into the objective function , we have @xmath88setting @xmath89 subject to @xmath90 , the unique solution is @xmath91 . thus , @xmath92 is the unique stationary point of @xmath93 . we can also verify that @xmath94 and @xmath95 . so @xmath96 is the unique solution to our maximization problem . we can directly solve @xmath97 subject to @xmath98 and @xmath82 . the unique solution is also @xmath96 . @xmath59 + we now present the optimal splitting method . theorem [ thm : nece ] states a necessary condition for the optimal splitting method , and theorem [ thm : unique ] implies there is a unique optimal splitting method . in corollary [ cor : optimal ] , we formally present the optimal splitting method and the required power levels . [ thm : nece ] let each real user split its power into @xmath1 virtual users . let @xmath42 be the power allocated to the @xmath43th virtual user and @xmath99 @xmath100 . if @xmath101 maximizes @xmath102 and satisfies @xmath103 , @xmath104 for @xmath105 , then @xmath106 , for all @xmath43 . that is , the optimal power split must lead to equal transmission rates for all virtual users . _ proof : _ we use a perturbation argument . suppose @xmath107 maximizes @xmath108 and satisfies @xmath109 , @xmath110 and the resulting @xmath111 is not the same for all @xmath43 . then we can find a pair of virtual users @xmath112 , where virtual user @xmath43 and @xmath113 are decoded at the @xmath43th and @xmath114th places respectively , and @xmath115 . without loss of generality , let us consider the case where @xmath116 @xmath117 . by the definition of @xmath45 , we have @xmath118 we can verify that if we change @xmath119 to @xmath120 and @xmath121 to @xmath122 , where @xmath123 is a small positive number , then the first term of decreases and the second term of increases . let @xmath124 be the solution to @xmath125 ( the existence of @xmath126 can be demonstrated ) . let @xmath127 and @xmath128 . notice that @xmath129 . since the maximization considered in lemma [ lemma : maximization ] has a unique solution , @xmath130 . this contradicts our assumption that @xmath107 maximizes + @xmath108 . therefore , the theorem follows . @xmath59 + by theorem [ thm : nece ] , if @xmath131 maximizes @xmath132 and satisfies @xmath133 , @xmath134 for @xmath105 , then we must have @xmath135 where @xmath136 . therefore , if we show that @xmath137 for all @xmath138 , has a unique solution , then there is _ at most _ one feasible solution to the maximization problem . [ thm : unique ] the set of equations : @xmath139 , subject to @xmath140 and @xmath141 @xmath142 , has a unique solution . _ proof : _ see appendix i. [ cor : optimal ] if a real user splits its power @xmath13 into @xmath1 virtual users , then the unique way to maximize this user s throughput is to set @xmath143 $ ] for @xmath144 . _ proof : _ since the constraint region @xmath145 is a simplex and @xmath108 is continuous , there exists at least one solution . we denote one solution by @xmath146 . by the necessary condition stated in theorem [ thm : nece ] , @xmath146 must satisfy @xmath147 . moreover , by the uniqueness property stated in theorem [ thm : unique ] , @xmath146 is the unique solution to @xmath148 subject to @xmath149 next , we plug @xmath42 into the expression of @xmath150 . let @xmath151 , we are able to verify that @xmath152 , which is independent of @xmath43 . hence , the corollary follows . @xmath59 + we now examine the asymptotic behavior of the drs scheme . we first demonstrate the interesting fact that the rate tuple converges to the maximum equal rate point for a general power split as long as all virtual users powers go to @xmath153 as @xmath154 . this implies a convergence result for the optimal power split . we then analyze the rate of convergence under the optimal power split . [ thm : genesplit ] given any power split @xmath155 , a sufficient condition for @xmath156 is @xmath157 as @xmath158 . _ @xmath159where @xmath160 , @xmath161 . the equality in is justified as follows . note that , @xmath162 . hence , for sufficiently small positive @xmath33 , @xmath163 . now , we examine the error term @xmath164 \right| \nonumber \\ & \leq & \sum_{k=1}^l \left| \log \left ( 1 + \frac{p_k}{mp - m\sum_{j < k}p_j - p_k + n } \right ) \right . \nonumber \\ & & \hspace{1 cm } \left . - \left(\frac{p_k}{mp - m\sum_{j < k}p_j - p_k + n } \right ) \right| \nonumber \\ & \leq & \sum_{k=1}^l \left(\frac{p_k}{mp - m\sum_{j < k}p_j - p_k + n } \right)^2 \label{eq : small } \\ & \leq & \sum_{k=1}^l \left(\frac{1}{n}\right)^2 p_k^2 \label{inq : powerconv}\\ & \leq & \left(\frac{1}{n}\right)^2 \max_{k } p_k \sum_{k=1}^l p_k \\ & = & p \left(\frac{1}{n}\right)^2 \max_{k } p_k\end{aligned}\ ] ] where inequality in holds because @xmath165 when @xmath166 , and the inequality in follows from the fact that @xmath167 . since @xmath168 as @xmath154 , the error term goes to zero in the limit . this justifies the equality in . using the capacity bound , we also have @xmath169 . therefore , @xmath170 @xmath59 + note that our optimal power split satisfies the sufficient condition in theorem [ thm : genesplit ] . therefore , its convergence is assured . [ cor : conv ] if each real user adopts the optimal splitting method specified in corollary [ cor : optimal ] , then @xmath171 next , we examine the rate of convergence to the maximum equal rate point under the optimal power split . define the error term @xmath172 \equiv r^ * - \sum_{k=1}^l r_k$ ] , we analyze how fast this error term tends to @xmath153 as @xmath154 . we prove the following : = \omega\left(g\left[n\right]\right)$ ] if there exist positive constants @xmath173 and @xmath174 such that @xmath175 \geq c_1 g\left[n\right ] $ ] for all @xmath176 , and @xmath177 = o(g[n])$ ] if there are positive constants @xmath178 and @xmath174 , such that @xmath179 \leq c_2 g[n]$ ] for all @xmath180 . finally , @xmath177 = \theta \left(g[n]\right)$ ] if @xmath177 = \omega \left(g[n]\right)$ ] and @xmath177 = o \left(g[n]\right)$ ] . ] [ thm : speed ] @xmath181 = \theta \left(\frac{1}{l}\right)$ ] . _ proof : _ let @xmath182 . @xmath183 \nonumber \\ & = & \lim_{l \rightarrow \infty } l \left ( \frac{1}{2m}\log \left(a\right ) - \frac{l}{2 } \log \left ( \frac{m}{m-1 + a^{-\frac{1}{l}}}\right)\right ) \nonumber \\ & = & \lim_{y \rightarrow 0 } \frac{\frac{y\log\left(a\right)}{2m}-\frac{1}{2 } \log \left(\frac{m}{m-1+a^{-y}}\right)}{y^2}\\ & = & \lim_{y \rightarrow 0 } \frac{\frac{\log\left(a\right)}{2m}-\frac{1}{2}\frac{a^{-y}\log\left(a\right)}{m-1+a^{-y}}}{2y } \label{eq : l1}\\ & = & \frac{\left(m-1\right)\left(\log\left(a\right)\right)^2}{4m^2}.\label{eq : l2}\end{aligned}\ ] ] note that equalities in and can be verified by lhospital s rule . consequently , given any @xmath23 , there exists a positive integer @xmath174 such that for all @xmath184 , we have @xmath185 - \frac{\left(m-1\right)\left(\log\left(a\right)\right)^2}{4m^2}\right| < \varepsilon$ ] . this implies @xmath186 < \frac{\left(m-1\right)\left(\log\left(a\right)\right)^2}{4m^2 } + \varepsilon$ ] . therefore , we can choose a small enough @xmath123 such that @xmath187 . let @xmath188 , we have @xmath189 \leq c_2 $ ] . this implies that there exists a positive integer @xmath174 and for all @xmath184 , we have @xmath190 \leq \frac{c_2}{l}$ ] . @xmath59 + finally , we note that all virtual users in one virtual user class can be decoded _ in parallel_. thus , the decoding delay of drs is proportional to the number of virtual users @xmath1 and independent of the number of real users . since @xmath1 is controlled by the designer , drs offers a tradeoff between the throughput of a real user and the decoding delay . in fig . [ fig : simulation ] , we present some numerical simulations illustrating the tradeoff between the number of virtual users and the throughput for both the high and low snr regimes . a system with @xmath191 real users is used in the simulations . an @xmath0-user discrete memoryless mac is defined in terms of @xmath0 discrete input alphabets @xmath192 , @xmath193 , an output alphabet @xmath194 and a stochastic matrix @xmath195 with entries @xmath196 . for any product input distribution @xmath197 , let the achievable region @xmath198 $ ] be the set of @xmath199 satisfying @xmath200 where @xmath201 and @xmath202 . the capacity region of the asynchronous mac is @xcite @xcite @xmath203.\ ] ] we fix the input product distribution @xmath204 and focus on achieving the desired operating point in @xmath205 $ ] . in this section , we consider only discrete memoryless channels satisfying the following symmetry condition : @xmath206 @xmath207 such that @xmath208 . later , in section [ sec : variation ] , we consider the more general asymmetric case . we further assume that for @xmath209 , if @xmath210 , then @xmath211 . under our symmetric setting , the maximum common rate that every user can achieve is @xmath212 . in the gaussian mac , virtual users are created via a power / rate splitting mechanism . for the discrete memoryless mac , we adopt the random switching mechanism of @xcite where virtual users with the same input distribution as the real users are created and the transmitted signal of a real user is determined by a random switch . we first consider the two - user discrete memoryless mac @xmath213 , and illustrate the random switching mechanism . the optimal random switches and the asymptotic behavior of drs under the optimal switching are presented . we then examine the @xmath0-user case @xmath214 , and present a sufficient condition for the random switching mechanism to converge to the information theoretic upper bound . finally , we investigate the rate of convergence for a simple suboptimal random switch . consider a two - user mac , @xmath215 . for a fixed input product distribution @xmath216 , the achievable region is given by : @xmath217 under our symmetry assumption ( cf ) , we have @xmath218 , @xmath219 , @xmath220 , and the optimal rate tuple is @xmath221 let us consider the random switching mechanism for this channel . we first consider the case where each real user generates two virtual users . later , we consider the case where the number of virtual users per real user goes to infinity . we split by means of two switches , as shown in fig . [ fig : switch ] . each switch has two inputs , @xmath222 and @xmath223 and one output @xmath224 . switch @xmath3 is controlled by a random variable @xmath225 with @xmath226 . the output is given by @xmath227 if @xmath228 , and @xmath229 if @xmath230 . the switching random variables @xmath231 are independent of the channel inputs . we also assume that @xmath231 are available at the receiver . in practice , one would generate @xmath232 and @xmath233 at the transmitters and at the receiver , e.g. by means of a pseudorandom sequence generator . assign to the channel inputs @xmath234 , @xmath235 , @xmath236 and @xmath237 the probability mass function @xmath238 . notice that @xmath239 and @xmath240 are independent and each has the same probability mass function as the random variable @xmath241 for @xmath242 . in successive decoding for the discrete memoryless mac , the signals of decoded virtual users are used as side information to aid the decoding process of subsequent virtual users . the first constituent decoder observes the output @xmath243 and tries to decode @xmath244 and @xmath236 . the second constituent decoder is informed of the decision about @xmath245 made by the previous constituent decoder and tries to decode @xmath235 and @xmath237 . without loss of generality , let us focus on real user 1 . @xmath246where the second equality follows from the independence between @xmath234 and @xmath232 , and the last equality follows from the fact that when @xmath247 , @xmath234 is independently of the output @xmath248 and @xmath233 . similarly , we have @xmath249.\end{aligned}\ ] ] it can be verified that in the two - user discrete memoryless mac , both real users throughput can be strictly increased by splitting their inputs via a random switch , relative to the case where they do not split , ( i.e. @xmath250 for @xmath251 ) . next , we show that by generating more virtual users , the throughput of each real user increases further . for @xmath252 , consider a distributed rate splitting scheme with @xmath1 virtual users per real user . the random switch for user @xmath3 is controlled by @xmath253 , where @xmath254 for @xmath144 . it is possible to strictly increase the throughput via an @xmath64 virtual user system by splitting the @xmath1th virtual user into two virtual users . _ proof : _ without loss of generality , we consider user @xmath30 . for the @xmath43th virtual user , we have @xmath255 \nonumber \\ & = & \lambda_k \left [ \left(\sum_{j < k } \lambda_j \right)i \left(x_1;y , x_2 \right ) \right . \nonumber \\ & & \hspace{2 cm } + \left . \left ( 1 - \sum_{j < k } \lambda_j \right ) i\left(x_1;y \right ) \right ] , \label{eq : lelast}\end{aligned}\]]where equality in is due to the independence between @xmath256 and @xmath232 , and equality in follows from the fact that when @xmath257 , @xmath256 is independent of the output @xmath248 and all the other random variables . finally , equality in holds because when @xmath258 , one of the random variables @xmath259 is the switch output , and when @xmath260 , none of them is the switch output . therefore , @xmath261.\ ] ] now let us split @xmath262 into @xmath263 and @xmath264 by using a switch controlled by a binary random variable @xmath265 with @xmath266 . we have @xmath267 where @xmath268 . hence , @xmath269 \\ & = & \alpha \overline{\alpha } \lambda_l^2 \left[i \left(x_1;y , x_2 \right ) - i \left(x_1;y \right ) \right ] \\ & \geq & 0\end{aligned}\ ] ] with strict inequality if @xmath270 . @xmath59 + before we examine the asymptotic behavior of the drs scheme for @xmath252 , we first solve the problem of how to find the optimal switches for a fixed number of virtual users per real user . for @xmath271 , if a real user has @xmath1 virtual users , then the optimal random variable to control the switch for user @xmath3 is @xmath272 with @xmath273 for @xmath144 and @xmath242 . _ proof : _ we use a perturbation argument . suppose the random variables @xmath274 with @xmath275 maximize @xmath276 , where @xmath277 and @xmath278 . moreover , suppose there exists @xmath279 such that @xmath280 . let @xmath279 be the first element which is not equal to @xmath281 . we consider the pair @xmath282 . we have @xmath283 @xmath284.\end{gathered}\ ] ] therefore , @xmath285i(x_1;y , x_2 ) \\ & & + \left[\left(\lambda_k + \lambda_{k+1}\right)\frac{l - k+1}{l } - \lambda_k \lambda_{k+1 } \right ] i(x_1;y).\end{aligned}\ ] ] first , we consider the case where @xmath286 . we let @xmath287 and @xmath288 for @xmath289 . we have @xmath290i(x_1;y , x_2 ) \\ + \left [ \left(\lambda_k + \lambda_{k+1 } \right)\frac{l - k+1}{l } - \left(\lambda_k-\varepsilon \right ) \left(\lambda_{k+1}+\varepsilon \right ) \right ] i(x_1;y ) . \end{gathered}\ ] ] thus , @xmath291 + @xmath292 - @xmath293 + @xmath294 = @xmath295 + @xmath296 @xmath123 -@xmath297 @xmath298-@xmath299 notice that the second term of the r.h.s . expression is positive and there exists @xmath123 such that @xmath300 + @xmath301 @xmath123-@xmath302 . this , however , contradicts our assumption that the random variables @xmath274 with @xmath303 , maximize @xmath276 . we obtain similar contradictions for the case where @xmath304 . @xmath59 + we now examine the asymptotic behavior of the drs scheme in the two - user discrete memoryless mac . [ thm : twoconv ] for @xmath252 , if both real users adopt @xmath305 for @xmath8 and @xmath306 , then @xmath307 for @xmath308 . moreover , if we define the error term @xmath172 \equiv r^ * - \sum_{k=1}^l r_{x_{ik}}$ ] , then @xmath172 = \theta ( \frac{1}{l } ) $ ] for @xmath306 . _ proof : _ without loss of generality , we consider real user @xmath30 . since @xmath309 we have @xmath310 \\ & = & \lim_{l \rightarrow \infty}\left[\frac{l+1}{2l}i(x_1;y ) + \frac{l-1}{2l}i(x_1;yx_2)\right ] \\ & = & \frac{1}{2 } \left[i(x_1;y ) + i(x_1;y , x_2 ) \right ] \\ & = & \frac{1}{2}i(x_1,x_2;y),\end{aligned}\]]where the last equality follows from the fact that @xmath219 in the symmetric setting . next , we examine the rate of convergence . we have the error term @xmath172 = \frac{1}{2}i(x_1;y ) + \frac{1}{2}i(x_1;y , x_2 ) - ( \frac{l+1}{2l}i(x_1;y ) + \frac{l-1}{2l}i(x_1;y , x_2))$ ] . thus , @xmath311 = \frac{1}{2l}\left ( i(x_1;y , x_2 ) - i(x_1;y ) \right),\ ] ] which implies @xmath312 = \frac{1}{2}\left ( i(x_1;y , x_2 ) - i(x_1;y ) \right ) \triangleq \frac{c}{2}.\ ] ] for any @xmath313 , there exists an @xmath174 such that for all @xmath184 , @xmath314 - \frac{c}{2}|<\varepsilon$ ] . hence , @xmath315 < \frac{c}{2}+\varepsilon $ ] . we can choose @xmath123 small enough such that @xmath316 . this implies @xmath172 = \theta ( \frac{1}{l } ) $ ] . @xmath59 + there are @xmath0 real users and each real user creates @xmath1 virtual users . we have @xmath0 switches @xmath317 to do the splitting with probabilities @xmath254 for @xmath318 and @xmath319 . we assume the receiver also knows @xmath317 . this may require common randomness to exist between all transmitters and the receiver . due to symmetry , we focus on one user , say user @xmath30 . for @xmath144 , we set @xmath320 , \label{eq : mlast}\end{aligned}\]]where @xmath321 and @xmath322 . equality in is due to the independence between @xmath256 and @xmath232 . equality in holds because when @xmath257 , @xmath256 is independent of the output @xmath248 and all the other random variables . the first term in follows from @xmath323 = @xmath324 . the second summation term in follows from the fact that the probability of @xmath3 switching random variables among @xmath325 having values less than @xmath43 is @xmath326 . it can be verified that real user @xmath30 with @xmath1 virtual users can strictly increase its total throughput via an @xmath74 virtual user system . in order to maximize the total throughput of real user 1 for fixed @xmath1 , we need to find the optimal @xmath327 to maximize @xmath328 . this is a non - convex optimization problem and appears to be difficult . we are able to verify that for the general @xmath0-user case ( unlike the two - user case ) , random switches with a uniform distribution are in general suboptimal . nevertheless , it is possible to generalize the asymptotic result of theorem [ thm : twoconv ] . we first demonstrate the fact that the convergence result holds for a general switch controlled by @xmath253 , where @xmath329 for @xmath8 , as long as @xmath330 as @xmath154 . we then analyze the rate of convergence for a particular suboptimal switch , the uniform switch . [ thm : mconv ] for a general random switch controlled by @xmath253 , where @xmath331 for @xmath144 , a sufficient condition for @xmath332 is @xmath330 as @xmath154 for @xmath318 . _ proof : _ without loss of generality , let us examine real user @xmath30 . @xmath333 \\ & = & \lim_{l \rightarrow \infty}\sum_{i=0}^{m-1}c_i i_i,\end{aligned}\]]where @xmath334 for @xmath335 . it is sufficient to prove @xmath336 for all @xmath3 . for @xmath335 , @xmath337where @xmath160 , @xmath338 and @xmath339 is the beta function . the term @xmath340 can be shown to converge to @xmath341 as @xmath154 by a similar argument . @xmath59 + next , we analyze a particular suboptimal switch , the uniform switch . since the uniform switch satisfies the sufficient condition in theorem [ thm : mconv ] , the convergence result holds . the next lemma presents its rate of convergence . consider an @xmath0-user discrete memoryless mac . let each real user have @xmath1 virtual users and each switch be controlled by an i.i.d . random variable @xmath342 with @xmath343 , @xmath8 . define the error term @xmath172 \equiv \frac{1}{m}i(x_1,x_2, ... ,x_m;y ) - \sum_{k=1}^l r_{x_{ik } } $ ] . then @xmath172 = o ( \frac{1}{l } ) $ ] for all @xmath3 . _ proof : _ in the uniform switch setting , @xmath343 for @xmath318 and @xmath144 . without loss of generality , we examine the total throughput of real user @xmath30 . @xmath344,\end{gathered}\]]where @xmath345 for @xmath346 . we denote @xmath347 , where @xmath348 is defined in the proof of theorem [ thm : mconv ] for @xmath349 , and @xmath350for @xmath335 . therefore , the error term can be calculated as follows , @xmath351 & = & \left| \frac{1}{m } i(x_1,x_2, ... \sum_{i=0}^{m-1}c_i i_i \right|\\ & = & \left| \sum_{i=0}^{m-1 } \left(\frac{1}{m } - c_i \right ) i_i \right| \\ & \leq & \sum_{i=0}^{m-1 } \left| \frac{1}{m } - c_i \right| i_i \\ & \leq & m \left[\max_i \left(\left| \frac{1}{m } - c_i \right| i_i \right ) \right].\end{aligned}\ ] ] note that @xmath352 is maximized at @xmath353 . for @xmath335 , it can be verified that @xmath354 @xmath355 multiplying both sides of the above two inequalities by @xmath356 , we have @xmath357 therefore , @xmath358 \frac{1}{l } \\ & \equiv & \alpha \frac{1}{l},\end{aligned}\ ] ] so @xmath172 \leq m \alpha \frac{1}{l}$ ] , where @xmath359 . for the term @xmath340 , a similar argument can be used to show that @xmath360 . therefore , @xmath172 = o \left ( \frac{1}{l } \right)$ ] . in section [ sec : gau ] and section [ sec : dmc ] , we imposed two symmetry constraints . the first is that the capacity region for the gaussian mac and the achievable rate region for the discrete memoryless mac are symmetric . the second is that users generate the same number of virtual users . in this section , we describe two variations of drs . the first variation is presented in section [ sec : unequalpower ] , where we relax the symmetric region constraint . in this case , we show that as the number of virtual user per real user tends to infinity , the rate tuple achieved under drs approaches a point on the dominant face . the second variation is presented in section [ sec : differentvuser ] , where each real user may generate a different number of virtual users . the main advantage of this variation is that it can accommodate different user rate requirements in a distributed fashion . in this section , we consider the case where real users in a gaussian mac may have different transmission powers ( i.e. the capacity region may not be symmetric ) . we assume that user @xmath3 has transmission power @xmath361 and the power vector @xmath362 is known to all users . we also assume that all real users split their powers into @xmath1 virtual users according to the common power splitting rule defined by the vector @xmath363 , where @xmath364 @xmath365 and @xmath366 . the power vector for the virtual users generated by user @xmath3 is @xmath367 for @xmath368 . for any real user with @xmath1 virtual users , the unique way to maximize this user s throughput is to set + + @xmath369 $ ] + + for @xmath144 . moreover , if all real users adopt this power allocation rule , then @xmath370for @xmath318 . _ proof : _ by replacing @xmath42 by @xmath371 for @xmath144 , we can use arguments similar to those in section [ sec : gau ] to prove the following : 1 . given a drs scheme with @xmath1 virtual users per real user , it is possible to strictly increase the throughput via an @xmath64 virtual user system . 2 . under the optimal power split , all virtual users generated by real user @xmath3 must have the same rate for @xmath372 . ( virtual users generated by different real users may have different rates . ) 3 . for any real user with @xmath1 virtual users , the unique way to maximize this user s throughput is to set + + @xmath373 $ ] + + for @xmath144 . if all real users adopt this power allocation rule , then @xmath374for @xmath318 @xmath59 + we illustrate this achievable point on the dominant face for a two - user gaussian mac in fig . [ fig : asy_reg ] . in section [ sec : dmc ] , we considered the symmetric setting ( cf ): @xmath375 @xmath207 such that @xmath208 . in this section , we relax this constraint and consider an asymmetric achievable region . we require only that for @xmath376 , if @xmath377 , then @xmath378 . the @xmath0 switches @xmath379 have probabilities @xmath380 for @xmath318 and @xmath381 . consider a general random switch controlled by @xmath253 , where @xmath380 , @xmath381 . define @xmath382 . if @xmath383 as @xmath154 , then @xmath384 for @xmath318 . _ proof : _ we can replace @xmath385 by @xmath386 and use arguments similar to those for theorem [ thm : mconv ] to prove the above lemma . @xmath59 + in this section , we retain the assumption that every user has the same transmission power @xmath13 , but we do not require all real users to create the same number of virtual users . that is , user @xmath3 and user @xmath387 create @xmath12 and @xmath388 virtual users independently , where @xmath12 may not be equal to @xmath388 . the signal transmitted by a real user is the superposition of all its virtual users signals . we also assume in this section that user @xmath3 transmits the number @xmath12 in a header message to the receiver . the receiver receives the sum of @xmath389 signals plus noise . we now describe a protocol which allows each user to split its power and set its rates independently , and allows the receiver to decode all virtual users one by one via a generalized successive decoding mechanism . recall that for the gaussian mac , successive decoding works as follows . users are decoded one after another regarding all other users that have not been decoded as interference , and the signals of decoded users are subtracted from the overall received signal . _ protocol @xmath30 _ : for user @xmath3 , the power split and rate allocation rule are defined as follows : for @xmath390 , @xmath391,\\ r_{ik } & = \frac{1}{2 } \log \left(1 + \frac{p_{ik}}{m\left(p - \sum_{j < k } p_{ij } \right ) - p_{ik } + n } \right ) . \ ] ] note that the power split and rate allocation rule in _ protocol @xmath30 _ are the same as that discussed in section [ sec : gau ] . the generalized successive decoding algorithm is given by the following pseudo - program . note that after a virtual user is decoded , its signal is subtracted from the overall received signal . * decode * virtual users @xmath392 in any order or in parallel . * set * @xmath393 * while * ( some virtual users are not decoded ) , find the minimal element in @xmath394 , say the @xmath3th entry ; decode the subsequent virtual user of user @xmath3 ; update the @xmath3th entry of @xmath394 : @xmath395 , where @xmath396 is the power of the virtual user being decoded in the previous step ; * end * [ le : adrs ] if all @xmath0 users adopt the power split and the rate allocation rule described in _ protocol @xmath30 _ , then for any @xmath397 , @xmath398 , the decoder can decode all virtual users one by one following the decoding algorithm . _ proof : _ by the rate allocation rule , @xmath399 for @xmath400 . thus , each of them can tolerate the maximum amount of interference plus noise , @xmath401 . it is then easy to see that virtual users @xmath402 can be decoded reliably in any order or in parallel . now , we set @xmath403 . in the first run of the * while * loop , if @xmath404 is the minimum entry in @xmath405 , the receiver decodes virtual user @xmath406 . by the rate allocation rule , @xmath407 this implies that the maximum amount of interference plus noise that virtual user @xmath406 can tolerate is @xmath408 . however , the real amount of interference plus noise it faces is @xmath409 , which is smaller than or equal to what it can tolerate because @xmath410 . therefore , virtual user @xmath406 can be decoded reliably at the receiver . suppose the decoding process succeeds in the @xmath411th run of the * while * loop . now , @xmath412 . let us consider the @xmath413th run of the * while * loop . suppose the @xmath3th entry , @xmath414 , is the minimum entry in @xmath394 . the receiver decodes the subsequent virtual user of user @xmath3 , denoted by @xmath415 . by the rate allocation rule , @xmath416 this implies the maximum amount of interference plus noise that virtual user @xmath415 can tolerate is @xmath417 = @xmath418 . however , the real amount of interference plus noise it faces is @xmath419 , which is smaller than or equal to what virtual user @xmath415 can tolerate because @xmath420 . therefore , virtual user @xmath415 can be decoded reliably . hence , the lemma follows by induction . @xmath59 + to illustrate the decoding algorithm , let us carefully examine a three - user example shown in fig . [ fig : examgau ] . the shaded regions correspond to the virtual users @xmath421 , which are decoded in any order . suppose we decode `` @xmath422 '' first . by the rate allocation rule , @xmath423 , which means the maximum amount interference plus noise that virtual user `` @xmath422 '' can tolerate is @xmath424 . this is exactly the amount of interference plus noise it faces . therefore , `` @xmath422 '' can be decoded reliably and we can subtract the signal of virtual user `` @xmath422 '' from the overall received signal . similarly , @xmath425 can be decoded reliably and subtracted from the overall received signal . now @xmath426 . the subsequent decoding order is illustrated by the numbers in fig . [ fig : examgau ] . in the first run of the * while * loop , since @xmath427 is the minimum in @xmath394 , the receiver decodes virtual user `` @xmath428 '' . by the rate allocation rule , @xmath429 , which implies the maximum amount interference plus noise it can tolerate is @xmath430 . however , the real interference plus noise it faces is @xmath431 , which is smaller than what it can tolerate since @xmath432 . so virtual user `` @xmath428 '' can be decoded reliably and subtracted from the received signal . by searching for the minimum entry in @xmath394 in each run , we always decode a virtual user that can tolerate more interference than what it really faces . this assures the validity of our decoding algorithm . all of the other virtual users can be decoded in a similar fashion . by lemma [ le : morevuser ] and lemma [ le : adrs ] , user @xmath3 can choose any @xmath397 , independently from other users , and have all virtual users decoded reliably at the receiver . therefore , user @xmath3 can choose @xmath12 according to its own service requirement . for example , if user @xmath3 wants to send low rate voice communication packets , it can set @xmath433 , which corresponds to the basic cdma scheme . if user @xmath3 wants to send high rate stream video , it can set @xmath12 equal to a large value in order to get higher throughput at the expense of higher coding complexity . thus , this variation of drs provides an explicit way for end users to trade off throughput and coding complexity , making differential rate requirements achievable in a distributed manner . finally , corollary [ cor : conv ] and theorem [ thm : speed ] demonstrate the asymptotic optimality of this scheme and its rate of convergence . in this section , we describe a variation of the drs scheme for the discrete memoryless mac which supports differential rate requirements to end users in a distributed manner . in this scheme , we adopt the uniform switch , but we do not require every user to have the same number of virtual users . we split by means of @xmath0 independent switches . without loss of generality , let us consider user @xmath3 . if user @xmath3 has @xmath12 virtual users , then switch @xmath3 has @xmath12 inputs , @xmath434 for @xmath390 , and one output @xmath224 . switch @xmath3 is controlled by a uniform random variable @xmath435 with @xmath436 for @xmath437 . the output is given by : @xmath438 , if @xmath439 . we now describe the protocol for the discrete memoryless mac which allows user @xmath3 to choose @xmath12 independently . we show the asymptotic optimality of this variation of drs under the protocol . _ protocol 2 _ : for user @xmath3 with @xmath12 virtual users , the switch @xmath3 is controlled by a uniform random variable @xmath253 where @xmath436 for @xmath437 . the rate allocation rule is defined as follows : @xmath440.\end{gathered}\ ] ] the decoding algorithm is given by the following pseudo - program . note that after a virtual user is decoded , its signal is used as side information to aid the decoding process of subsequent virtual users . * decode * virtual users @xmath392 in any order or in parallel . * set * @xmath441 * while * ( some virtual users are not decoded ) , find the minimal element in @xmath442 , say the @xmath3th entry ; decode the subsequent virtual user of user @xmath3 ; update the @xmath3th entry of @xmath442 : @xmath443 ; * end * if all @xmath0 users adopt the rate allocation rule in _ protocol 2 _ , then for any @xmath397 @xmath444 , the decoder can decode all virtual users one by one following the decoding algorithm . _ proof _ : virtual users @xmath392 can be decoded in any order if and only if @xmath445 for all @xmath3 . this is true because under our rate allocation rule , @xmath446 . we set @xmath447 in the first run of the * while * loop , if @xmath448 is the minimum entry of @xmath442 , the receiver decodes virtual user @xmath406 . by the rate allocation rule @xmath449 \label{eq : i2}\end{aligned}\ ] ] virtual user @xmath406 can be decoded reliably if @xmath450 this mutual information can be simplified in the same way as described in equations - . recall the definition @xmath382 . @xmath451 we can verify that the @xmath452 term in is less than or equal to the corresponding @xmath452 term in equation . this follows from @xmath453since @xmath454 for all @xmath455 . similarly , it can be verified that the second term in is less than or equal to the corresponding term in equation . therefore , virtual user @xmath406 can be decoded reliably . suppose the decoding process succeeds in the @xmath411th run of the * while * loop . now , @xmath456 . let us consider the @xmath413th run of the * while * loop . suppose the @xmath3th entry , @xmath457 , is the minimum entry in @xmath442 . the receiver decodes the subsequent virtual user of user @xmath3 , denoted by @xmath458 . by the rate allocation rule @xmath459.\end{aligned}\ ] ] again , we can simplify the mutual information @xmath460 and show that it is great than or equal to @xmath461 , which implies that virtual user @xmath458 can be decoded reliably . hence , the lemma follows by induction . @xmath59 + let us illustrate the decoding algorithm by the following example . we consider a two - user discrete memoryless mac where user @xmath30 creates @xmath11 virtual users and user @xmath11 creates @xmath462 virtual users . random switch @xmath30 is controlled by @xmath232 where @xmath463 , and random switch @xmath11 is controlled by @xmath233 where @xmath464 for @xmath465 . by the rate allocation rule , the virtual users rates can be simplified as follows @xmath466\end{aligned}\ ] ] @xmath467\\ r_{x_{23 } } & = & \frac{1}{3}\left[\frac{1}{3 } i(x_1;y ) + \frac{2}{3}i(x_1;y , x_2 ) \right].\end{aligned}\ ] ] we first decode @xmath468 in any order . suppose we decode `` @xmath469 '' first . virtual user `` @xmath469 '' can be decoded reliably if @xmath470 . the condition holds because @xmath471 virtual user `` @xmath422 '' can be decoded similarly . now , we set @xmath472 . in the first run of the * while * loop , the receiver decodes `` @xmath473 '' since @xmath474 . let us calculate the mutual information between @xmath237 and @xmath248 given @xmath475 and previously decoded @xmath476 . @xmath477 the third equality is due to our symmetric assumption . therefore , virtual user `` @xmath473 '' can be decoded . by searching for the minimum entry in @xmath442 in each run , we always decode a virtual user whose rate is smaller than or equal to the corresponding mutual information . this guarantees the correctness of our decoding algorithm . virtual users @xmath478 can be decoded reliably at the receiver in a similar fashion . the asymptotic optimality of this scheme in the discrete memoryless mac can be demonstrated by theorem [ thm : mconv ] . in this paper , we take an information - theoretic approach to the problem of distributed multiple - access communication . we present a distributed rate splitting scheme whereby each real user creates a number of virtual users and all virtual users are successively decoded at the receiver . one possible advantage of distributed rate splitting is that it can be implemented with lower complexity when compared with joint coding schemes , and less coordination among users when compared with either time - sharing or rate splitting . for the symmetric @xmath0-user gaussian mac , each real user creates the same number of virtual users via a power / rate splitting mechanism . the transmitted signal of a real user is the superposition of all its virtual users signals . for the symmetric @xmath0-user discrete memoryless mac , each real user creates the same number of virtual users via a random switching mechanism , and the transmitted signal of a real user is determined by a random switch . all virtual users are successively decoded at the receiver . it is shown that drs can achieve the maximum equal rate point for both channel models as the number of virtual users per real user tends to infinity . finally , we present two variations of the drs scheme . for the case of asymmetric capacity regions , we show that a point on the dominant face can be achieved asymptotically . for the case of an unequal number of virtual users , we show that different user rates requirements can be accommodated independently in a distributed manner . _ proof : _ we use induction on the number of virtual users . for @xmath46 , the original problem reduces to : @xmath479 subject to @xmath480 and @xmath481 . the unique solution is @xmath482 for @xmath483 , suppose @xmath484 uniquely solves @xmath485 for any @xmath486 , subject to @xmath487 and @xmath488 . let us consider the @xmath387 virtual users case . given any tuple @xmath489 such that @xmath490 , we can fix @xmath80 , so @xmath491 . we now solve @xmath492 subject to @xmath493@xmath42 = @xmath494 , @xmath42 @xmath495 for @xmath43=@xmath496 for fixed @xmath80 , by the induction hypothesis , we have a unique solution @xmath497@xmath498 , ... , @xmath499@xmath500 which solves equation ( [ eq : hminusone ] ) . let @xmath501 @xmath502 . we are able to verify that @xmath503 is a strictly decreasing function of @xmath80 , and @xmath504 = @xmath505 is a strictly increasing function of @xmath80 . moreover , the function values at boundary points satisfy @xmath506 and @xmath507 . so there exists a unique @xmath508 such that @xmath509 . hence , we conclude @xmath510 is the unique solution to the @xmath387 virtual users case . the theorem follows by induction . edmund yeh gratefully acknowledges useful discussions with prof . shlomo shamai on this topic . m. medard , j. huang , a. goldsmith , s. meyn , and t. coleman , `` capacity of time - slotted aloha packetized multiple - access systems over the awgn channel , '' _ ieee trans . wireless commun . . 3 , pp . 486499 , mar . 2004 . r. cheng , `` stripping cdma an asymptotically optimal coding scheme for l - out - of - k white gaussian channels , '' in _ proc . ieee global telecommunications conference _ , sorrento , italy , nov . 1996 , pp . 142146 . jian cao jian cao received the b.s . degree in electrical and electronics engineering with first class honor from the hong kong university of science and technology , hong kong , in 2002 , the m.s . degree in electrical engineering from yale university , new haven , in 2003 , and is currently working toward the ph.d . degree in electrical engineering at yale . his research interests include information theory and queueing theory . edmund yeh edmund yeh received his b.s . in electrical engineering with distinction from stanford university in 1994 , m.phil in engineering from the university of cambridge in 1995 , and ph.d . in electrical engineering and computer science from mit in 2001 . since july 2001 , he has been on the faculty at yale university , new haven , connecticut , where he is currently an associate professor of electrical engineering and computer science . yeh is a recipient of the army research office ( aro ) young investigator program ( yip ) award ( 2003 ) , the winston churchill scholarship ( 1994 ) , the national science foundation and office of naval research fellowships ( 1994 ) for graduate study , the frederick e. terman award from stanford university ( 1994 ) and the barry m. goldwater scholarship from the united states congress ( 1993 ) . yeh is a member of phi beta kappa , tau beta pi , and ieee . he has been visiting faculty at mit , princeton university , university of california at berkeley , and swiss federal institute of technology , lausanne .
we consider the multiple - access communication problem in a distributed setting for both the additive white gaussian noise channel and the discrete memoryless channel . we propose a scheme called distributed rate splitting to achieve the optimal rates allowed by information theory in a distributed manner . in this scheme , each real user creates a number of virtual users via a power / rate splitting mechanism in the @xmath0-user gaussian channel or via a random switching mechanism in the @xmath0-user discrete memoryless channel . at the receiver , all virtual users are successively decoded . compared with other multiple - access techniques , distributed rate splitting can be implemented with lower complexity and less coordination . furthermore , in a symmetric setting , we show that the rate tuple achieved by this scheme converges to the maximum equal rate point allowed by the information - theoretic bound as the number of virtual users per real user tends to infinity . when the capacity regions are asymmetric , we show that a point on the dominant face can be achieved asymptotically . finally , when there is an unequal number of virtual users per real user , we show that differential user rate requirements can be accommodated in a distributed fashion . multiple access , rate splitting , successive decoding , stripping , interference cancellation , aloha .
the cosmological requirement that half the baryons in the universe are in a warm / hot intergalactic medium ( whim ) has motivated uv and x ray absorption line studies to detect this otherwise invisible gas , using absorption against a bright agn to probe the line of sight material ( bregman 2007 ) . however , most agn ( except blazars ) have intrinsic columns of warm absorbing gas in their nuclei , and the fact that this is generally connected to a nuclear outflow ( blustin et al 2005 ) complicates separating this intrinsic absorption from extrinsic line of sight material using velocity information . nonetheless , various studies have looked for absorption lines ( predominantly ovii / viii ) as signatures of the whim , though all current detections are more likely to be associated with material in our galactic halo or local group ( bregman & lloyd jones 2007 ) . these have velocities within a few hundred km / s of the local standard of rest , and indicate columns of ovii / viii @xmath6 @xmath5 , equivalent to hydrogen columns of @xmath7 @xmath5 . much more controversial was the potential association of substantially higher columns of gas to such local material . mckernan et al ( 2004 ; 2005 ) noticed that several agn with strong he or h like fe absorption features had these lines at energies which were approximately consistent with the local standard of rest . for any single object the approximate match between the putative blueshifted outflow velocity and the galaxy redshift could be coincidental , but mckernan et al ( 2004 ; 2005 ) pointed out three agn showing this trend , with redshifts spanning @xmath8 ( mcg63015 , pg1211 + 143 and pds 456 ) . however , the derived columns of @xmath9 @xmath5 are high even for an intrinsic agn outflow ( pounds et al 2003 , hereafter p03 ; reeves et al 2003 ) . if these are instead of local origin then it requires a tremendously significant change to our understanding of the galactic halo environment ( mckernan et al 2004 ; 2005 ) . here we use new suzaku data to show that the highly ionised fe absorption in pg 1211 + 143 is variable on a timescale of years . this conclusively demonstrates that it is intrinsic to the agn , and not associated with our galaxy or local group . we collate recent results to show that this is also the case for most other agn with strong fe k absorption lines , showing that powerful outflows are associated with luminous accretion flows . chandra and xmm newton gratings gave the first high resolution x ray spectra of the warm absorbers in agn ( e.g. blustin et al 2005 , mckernan , yaqoob & reynolds 2007 ) . however , only the chandra hetgs can cover the fe k region , and this has very small effective area . thus most of the information on these absorption lines are from the large effective area , moderate resolution ccd s on xmm - newton . these showed that some powerful agn have strong absorption lines at 7 kev ( observed frame ) associated with large columns of highly ionised iron ( pg 1211 + 143 : p03 ; pounds & page 2006 , hereafter pp06 ; pds456 : reeves et al 2003 ) . the results on pg 1211 + 143 , at a redshift of @xmath0 , were initially controversial . p03 claimed that the dip in the spectrum at @xmath10 kev was an absorption line , which they initially identified h like fe k@xmath11 ( 6.97 kev , lab frame ) . given the source redshift this requires an outflow velocity of @xmath12 km / s i.e. 0.08c assuming an origin in the agn . other , weaker but still significant absorption lines in the ccd spectra at 2.68 and 1.47 kev could also be interpreted as arising from material with the same outflow velocity if these are from h like s and mg k@xmath11 , respectively . ( p03 , pp06 ) . the higher resolution rgs instrument also shows lines from lower z elements at low energies . p03 identified these with transitions which required similarly high velocities as the high ionisation material , while kaspi & behar ( 2006 ) used different identifications to obtain a lower velocity outflow solution for these lines . nonetheless , the origin of the rgs features is not the major issue . only the ccd s on xmm - newton can show the existence of material which is so highly ionised that heavy elements ( s , mg , fe ) are predominantly he or h like . surprisingly , kaspi & behar ( 2006 ) identified the s and mg k - shell lines with weaker he - like k@xmath13 lines ( with little velocity offset ) , instead of the blue - shifted h - like k@xmath11 transitions , which begs the question as to why the corresponding stronger k@xmath11 lines are not seen . these authors also model the fe k absorption in the 2001 xmm - newton data with an edge at 7.27 kev ( rest frame ) of optical depth @xmath14 , corresponding to a large column ( @xmath15@xmath5 ) of very low ionisation state matter ( fe vii / viii ) ; this should have substantial low energy absorption which is clearly not present in the data . as we show below , accounting for any low ionisation absorption ( via partial covering ) in a self consistent photoionisation model does not remove the requirement for the blue - shifted highly ionised iron k absorber . the only self consistent option for the features around 7 kev in the xmm - newton data is k@xmath11 absorption lines from h or he like iron . this requires a large blueshift ( 0.08c or 0.14c respectively ) if the absorption is intrinsic to the agn , with the implication being that the kinetic energy associated with this material can dominate the agn energetics ( pounds & reeves 2007 ) . pg 1211 + 143 was observed by suzaku ( mitsuda et al . 2007 ) between 24 - 27 november 2005 . in this paper we discuss data taken with the 4 xis ( x - ray imaging spectrometer ; koyama et al . 2007 ) ccds . events files from official version 2.0 of the suzaku pipeline processing were used . the pg 1211 + 143 suzaku event files were screened using identical criteria to those described in reeves et al . ( 2007 ) for the suzaku observation of mcg-5 - 23 - 16 . subsequently source spectra from the xis ccds were extracted from circular regions of 2 radius centred on the source , in the off - axis hxd nominal pointing position . background spectra were extracted from 4 circles offset from the source region , avoiding the calibration sources on the corners of the ccd chips . xis response files ( rmfs ) and ancillary response files ( arfs ) were generated using the xisrmfgen and xissimarfgen scripts including correction for the hydrocarbon contamination on the optical blocking filter ( ishisaki et al . a net xis source exposure of 97.5 ks was obtained for each of the 4 xis chips . the 3 front illuminated xis chips ( xis 0,2,3 ) were used in this paper , as they have the greatest sensitivity at iron k. these chips were found to produce consistent spectra within the statistical errors , so the spectra and responses were combined to maximise signal to noise . the net source count rate for the 3 xis combined was @xmath16countss@xmath17 , with background only 3% of the source rate . the xis source spectrum was binned to a minimum of 100 counts per bin to enable the use of @xmath18 minimisation . errors are quoted to 90% confidence for 1 parameter ( i.e. @xmath19 ) . xmm - newton observed pg 1211 + 143 twice , on 2001 june 15 for 53ks and 2004 june 21 for 57ks but this second exposure was affected by high particle background , resulting in only 25ks of usable data . the epic - pn spectra used in this paper are identical to those described in pp06 and pounds & reeves ( 2007 ) , which were processed using the xmm - newton sas v6.5 software . the epic - mos spectra are consistent with the epic - pn , with the same 7 kev absorption feature present in the iron k band , as discussed by pp06 . the 20012005 spectra from 0.4 - 10 kev are plotted in figure 1 , relative to a @xmath20 power - law continuum . the ccd s on both xmm - newton and suzaku have similar responses so the data are directly comparable . a galactic column of @xmath21@xmath5 is included in all the spectral fits . from the spectra , it is clear that there is little change in the overall spectral shape or normalisation . all 3 observations show a strong soft x - ray excess below 1 kev , though it is marginally more prominent during the 2001 xmm - newton and 2005 suzaku observations . however the deep iron k absorption feature observed at 7 kev in 2001 ( pounds et al . 2003 ) which is readily apparent even on this broad band spectral plot , is not present at the same strength in either the 2004 or 2005 data . this feature has varied significantly , ruling out a local origin for the absorber . to quantify this , we use only the 2001 xmm - newton and 2005 suzaku observations as these have the best signal to noise . we fit the datasets simultaneously over the 210kev bandpass to avoid the complexity associated with the soft x ray excess . a power law with fixed galactic absorption gives a very similar index between the two datasets as the continua are very similar ( see fig 1 ) , but the overall fit is poor ( @xmath22 and @xmath23 for the xmm - newton and suzaku data , respectively ) . 2 shows the ratio of the data to the model for this fit , where as well as the obvious absorption feature in the xmm - newton data , it is clear that both spectra have iron emission at @xmath24 kev . adding a gaussian line ( with intrinsic width constrained to be equal across the two datasets ) gives a line energy and equivalent width consistent between the two datasets , at @xmath25 kev and @xmath26 ev respectively . the intrinsic line width is resolved with a fwhm velocity of @xmath27kms@xmath17 i.e. typical of radii @xmath28@xmath29 from the black hole . [ cols="^,^,^,^,^,^,^,^,^,^ " , ] however , there is still significant curvature in the spectrum , as shown by the poor combined @xmath30 . this can be phenomenologically modelled by an additional gaussian line , again with intrinsic width tied between the two datasets . this gives an acceptable fit , with @xmath31 , for a very broad line , with an intrinsic width of @xmath32 kev . the line energy is consistent between the two datasets at @xmath33 kev but its equivalent width is much larger in the xmm - newton data than in the suzaku data , at @xmath34 ev compared to @xmath35 ev ( see table 1 ) , showing that there is more curvature in the xmm - newton data . since it is also the xmm - newton data which show the line absorption feature , it is possible that this curvature is associated with the absorber , rather than with a broad , redshifted iron line from reflection from the very innermost ( i.e. @xmath36 ) accretion disc . we phenomenologically model this by partial covering of the source by partially ionised material ( e.g. miller et al . , 2007 ; turner et al . , 2007 ) , using a grid of xstar ( v2.1l ) photoionisation models based upon the publicly available tabulated grid 25. this uses the fe k treatment of kallman et al . ( 2004 ) , a turbulent velocity of @xmath37kms@xmath17 and solar abundances ( grevesse & sauval 1998 ) , while the illuminating continuum from 11000 rydbergs has @xmath38 rather than @xmath20 , and spans a wider ( but coarser ) range of column ( 0.05 to 500@xmath39 @xmath5 , 10 points ) and @xmath40 ( -3 to 6 , 12 points ) . this partially ionised absorption is applied to a fraction @xmath41 of the assumed continuum , while the remaining ( @xmath42 ) is not affected by this material . we fix the redshift of the partial coverer to that of the agn , and tie the ionisation parameter of this material ( but not its covering fraction ) between the two datasets . this model gives an equally good fit as the broad gaussian model , with @xmath43 for @xmath44 . the column of the ionised material is consistent between the two datasets at @xmath45 @xmath5 , but the covering fraction is much larger in the xmm - newton spectrum ( 80 per cent compared to 40 per cent ) , showing that the spectral curvature is more evident in these data . we use these two different models ( hereafter termed gau and pcfxi , respectively ) to assess the robustness of the change in the narrow absorption features to different continuum placement . note the partial covering model is insensitive to the turbulence assumed , as its effect from 210kev is to reproduce the continuum curvature in the spectrum , from bound free absorption such as the iron k - shell edge . fig 3 shows the @xmath46 residuals to the power - law , gau and pcfxi models , respectively . the absorption line is still present in the xmm - newton data , even against the pcfxi model which contains the iron k - shell edge and its associated absorption line structure ( figure 13 , kallman et al . we fit an additional inverted gaussian of fixed width ( @xmath47kev ) to the xmm - newton data and find that the fit is significantly improved , to @xmath48 and @xmath49 for the gau and pcfxi continuum models , respectively . the line energy and equivalent width in both is @xmath50kev ( @xmath51kev observed ) and @xmath52 ev , respectively . we allow a line with the same energy and width in the suzaku data , and find a 90% confidence upper limit to its equivalent width of @xmath53 ev . thus the absorption line is significantly weaker in the 2005 suzaku data , so the absorption has changed on a timescale of @xmath1 years . either the column has physically moved , or its ionisation state has changed ( or both ) . the former puts a limit on the size scale of 4 light years , unfeasibly small for any diffuse galactic halo or intergalactic emission ( which should be on kpc scales ) , while the latter sets a limit on the density through the recombination timescale , @xmath54 , where @xmath55 is the recombination rate , which is dependent on the temperature , t. for a whim origin , the absorption line must be associated with fe xxvi in order that its energy matches that observed . such material should also be collisionally ionised , and requires temperatures @xmath56 k in order to be the dominant ion . this gives a recombination rate of @xmath57 @xmath58 s@xmath17 ( e.g. arnaud & rothenflug 1985 ) , so for this to be less than 4 years requires densities @xmath59 @xmath3 , which is 6 orders of magnitude above the typical density even for a local halo component ( e.g. bregman & lloyd davies 2007 ) . we model the fe k absorption line using the warmabs xstar photo - ionisation code incorporated into xspec , adopting solar abundances . we assume a turbulent velocity of 5000 km s@xmath17 , below the upper limit of @xmath60 km s@xmath17 set by the observed intrinsic width of the line of @xmath61 kev . this gives a column of @xmath62@xmath5 , with an ionisation parameter of @xmath63 at an observed redshift of @xmath64 i.e. implying an agn outflow velocity of @xmath65c . while the derived column is dependent on the turbulent velocity , the curve of growth is in the linear regime here , so increasing the turbulent velocity does not increase the equivalent width of the line , leading to a robust lower limit of @xmath66@xmath5 . the lower limit on the density , combined with the column density , can give an upper limit on the size scale @xmath67 of the absorber as @xmath68 . a huge column of @xmath69 @xmath5 is required in order to produce an iron absorption line with equivalent width of 100 ev if the gas velocity is simply set by its temperature of @xmath70 k ( kotani et al . 2000 ; 2006 ) as expected from any local hot gas . such a large column is inconsistent with the spectrum , as this produces a strong dip at 810 kev ( rest frame ) due to the fe k - shell edge structure ( kallman et al 2004 ) , which is not present in the data . indeed from the lack of an iron k edge in pg1211 + 143 , this sets an upper limit to the column of @xmath71@xmath5 and requires that there must be substantial turbulent velocity ( @xmath72kms@xmath17 derived from the curve of growth ) in excess of the thermal motions to model the line ew . this itself is inconsistent with a whim origin . with this , the upper limit on the absorber size scale is @xmath73 cm . this is unfeasibly low to be associated with even the smallest local diffuse hot gas , the kpc scale galactic halo . another constraint on the properties of the material comes from the emission measure @xmath74 where @xmath75 is the volume , assuming that the gas is distributed evenly in a spherical shell at distance @xmath76 . the total bremsstrahlung luminosity is @xmath77 , so the surface brightness in solid angle @xmath78 is @xmath79 . for a solid angle of 1 arcmin@xmath80 sr this gives @xmath81 ergs s@xmath17 @xmath5 arcmin@xmath82 , i.e. brighter than the crab in every square arcmin in x - rays . indeed the observed upper limit to the diffuse sky surface brightness in the 3/4 kev band is @xmath83 counts s@xmath17 arcmin@xmath82 ( bregman & lloyd davies 2007 ) , corresponding to a flux of @xmath84 ergs s@xmath17 @xmath5 arcmin@xmath82 , 10 orders of magnitude below the predicted value . thus the variability and column appear to rule out any sort of non - agn origin for the absorbing material . the best fit identification by pp06 of the highly ionised absorption as being dominated by he like fe also rules out a local origin as the line is then not close to the rest wavelength of the transition , showing a blueshift with respect to the local frame of @xmath85 . this is because the mg , si and s k - shell absorption seen in the xmm - newton ccd spectrum can be fit with the same layer of absorption if the ionisation state of iron is lower than h - like . this mismatch between the outflow velocity of the absorber and recessional velocity of the agn is common . there are now many other agn with blueshifted fe k@xmath11 indicating an outflow , and we collate these from the literature in fig . 4 . aside from the 3 low velocity examples in figure 4 , none of the agn outflow velocities coincide with the recession velocity . in ic 4329a and mcg-5 - 23 - 16 the absorption lines are observed at 7.7 kev ( markowitz et al . 2006 ; braito et al . 2007 ) inconsistent with the rest energy of any fe k@xmath11 transition line . a further example is from the new 2007 suzaku data from pds456 ( reeves et al . 2007 , in prep ) , which shows iron k absorption lines observed at 7.6 and 8.2 kev , inconsistent with local absorption from iron k@xmath11 . indeed the iron k absorption features seen in high redshift bal qso s are likewise very far from the iron k@xmath11 rest energies for local material ( chartas et al 2002 , 2003 ) . we show that the observed variability in the absorption line in pg 1211 + 143 between the xmm - newton and suzaku data taken 4 years apart conclusively rules out a diffuse gas origin such as the local galactic halo or whim . the gas must be associated with the agn , as is further evidenced by its large column density in the xmm - newton 2001 data which is far too high for local or intergalactic gas . the conclusive identification with the agn , and the implausibility of any alternative line transition other than iron k@xmath11 means that the large outflow velocity of @xmath86c is inescapable . such material is predicted from the winds which are produced from luminous accretion discs in agn ( proga & kallman 2004 ) . the coincidence of the outflow velocity with the source redshift in these data is not significant . other powerful agn which show iron k absorption features clearly show a range of ( intrinsic and observed ) velocities , removing the apparent trend for the line energy to appear at the rest energy for these transitions . thus it is clear that this material is a mildly relativistic outflow from the agn . its high velocity means that its kinetic energy can be comparable to the bolometric radiated luminosity of the agn ( king & pounds 2003 ; pounds et al 2003 ; pounds & reeves 2007 ) , yet its high ionisation means that it is effectively invisible in all other wavebands . while such a major component of agn energetics could go unnoticed in individual objects , there is increasing evidence that strong agn feedback controls galaxy formation and evolution . mildly relativistic winds provide more efficient heating than the jet due to their impact on a larger area , so these highly ionised disc winds may be the key to understanding the growth of structure in the universe .
we present a comparison between the 2001 xmm - newton and 2005 suzaku observations of the quasar , pg1211 + 143 at @xmath0 . variability is observed in the 7 kev iron k - shell absorption line ( at 7.6 kev in the quasar frame ) , which is significantly weaker in 2005 than during the 2001 xmm - newton observation . from a recombination timescale of @xmath1 years , this implies an absorber density @xmath2@xmath3 , while the absorber column is @xmath4@xmath5 . thus the sizescale of the absorber is too compact ( pc scale ) and the surface brightness of the dense gas too high ( by 9 - 10 orders of magnitude ) to arise from local hot gas , such as the local bubble , group or warm / hot intergalactic medium ( whim ) , as suggested by mckernan et al . ( 2004 , 2005 ) . instead the iron k - shell absorption must be associated with an agn outflow with mildly relativistic velocities . finally we show that the the association of the absorption in pg1211 + 143 with local hot gas is simply a coincidence , the comparison between the recession and iron k absorber outflow velocities in other agn does not reveal a one to one kinematic correlation . [ firstpage ] accretion , accretion discs atomic processes x - rays : galaxies
accretion onto a supermassive black hole is generally regarded to be powering active galactic nuclei ( agns ) . for a steady keplerian accretion disk around a black hole with mass @xmath11 , the released energy via viscosity dissipation is given by @xmath12 where @xmath13 , the eddington luminosity @xmath14 , @xmath15 , @xmath16 , @xmath17 and @xmath18 ( shakura & sunyaev 1973 ) . it is important to note that @xmath19 is independent of the specific mechanism of viscosity stress . in the regime of the standard accretion disk around a massive black hole , the disks can not produce x - ray emission . models of the accretion disks with hot corona have been suggested in order to explain the observed x - ray emission ( liang & price 1977 ) . in a popular model , a fraction @xmath8 of @xmath19 is assumed to be released in the hot corona and the left ( @xmath20 ) in cold disk , but the transportation of angular momentum is mainly taking place in cold disk ( haardt & maraschi 1991 ; svensson & zdziarski 1994 , kawaguchi et al . 2001 ) . however , the factor @xmath8 remains open since we poorly understand how the process happens in the disk . the important roles of the magnetic field turbulence have been realized in transportation of angular momentum ( balbus & hawley 1991 , tout & pringle 1992 , turner et al . 2002 , 2003 ) and formation of the hot corona ( galeev et al . 1979 , stella & rosner 1984 , merloni & fabian 2002 , liu et al . 2002 , turner 2004 , kuncic & bicknell 2004 ) . the maxwell stress is capable of transporting the angular momentum in the disk ( balbus 2003 ) . numerical simulations show the actual physical processes inside the disks are very complicated ( balbus & hawley 1991 , stone et al . 1996 ; blaes 2002 ; blaes & socrates 2003 ; machida & matsumoto 2003 ; igumenshchev et al . 2003 ; sano et al . 2003 , kato et al . we noted that these simulations are based on the assumption of radiative inefficiency , and can not be directly applied to explain observations . the seminal paper by galeev et al . ( 1979 ) showed that the gravitational energy will be mainly released in the hot corona since the strong buoyancy and magnetic field reconnection inevitably lead to the formation of the hot corona above cold disk ( stella & rosner 1984 ) . the formation of the hot corona is most likely naturally related with the transportation of angular momentum , the factor @xmath8 can be thus obtained if the magnetic stress and energy transportation are assumed ( merloni & fabian 2002 ) . this may lend us a possible opportunity to test the working magnetic stress from hard x - ray observations . there is growing evidence in the galactic black hole candidates for that the hot corona becomes weak when @xmath21 increases ( see a review of mcclintock & remillard 2004 ) . however it is elusive how the hot corona changes with @xmath21 in agn disks . two problems remain open : 1 ) what is the proper magnetic field stress ? 2 ) how to connect the two processes of angular momentum transportation and the energy release in the hot corona ? in this _ letter _ , we look into the hard x - ray data of radio - quiet agns to see @xmath21-dependence of hard x - rays and consider the roles of magnetic turbulence in the disks . we found a strong correlation between the factor @xmath8 and the eddington ratio , implying that the energy release is driven by the eddington ratio and the importances of magnetic turbulence . table 1 gives the data of the present sample , which is composed of 56 objects consisting of 29 narrow line seyfert 1 galaxies ( nls1s ) and 27 broad seyfert 1 galaxies ( bls1s ) . col ( 1 ) is the name of the objects , col ( 2 ) the 2 - 10kev luminosity from _ asca _ observations in units of erg s@xmath22 , col ( 3 ) the photon index between 2 - 10 kev , col ( 4 ) the bolometric luminosity in units of erg s@xmath22 , col ( 5 ) the black hole mass in the units of solar mass . col ( 6 ) gives some notes on the objects , indicating the references of x - ray luminosity and photon index as well as the references or methods of the estimations of the black hole mass , bolometric luminosity . the black hole masses are estimated by table 1 the sample 0.2 cm [ cols="<,<,^,^,^ , < " , ] fig 1 shows the plot of photon index , @xmath4 , and fraction of x - ray emission , @xmath1 , versus the eddington ratio , @xmath0 . except for phl 909 , the x - ray photon indices almost keep a constant for bls1s with a mean value of @xmath23 . there is a trend for nls1s : the higher is accretion rate , the steeper is the x - ray spectrum as @xmath24 . we find the pearson s correlation coefficient of @xmath25 is 0.37 for nls1s . the difference of @xmath26 relations in bls1s and nls1s show an interesting evidence for the bi - modal accretion in seyfert 1 galaxies . however , it is interesting to note that there is a strong correlation of @xmath4 and the eddington ratio @xmath2 for the entire sample , @xmath27 with the pearson s coefficient @xmath28 and probability @xmath29 . we use the least square method of multivariate regression and find a very strong correlation @xmath30 with coefficient @xmath31 and @xmath32 of the @xmath33-test ( much larger than @xmath34 at @xmath35 level of confidence ) . this correlation can be rewritten as @xmath36 from which an interesting conclusion follows : the hot corona is mainly controlled by the eddington ratio . shows this correlation . the correlation shows that the corona becomes dramatically weak with increases of the eddington ratio . while the eddington ratio is low , a larger fraction of the energy will be released in the hot corona . a naive expectation is that the magnetic stress responsible for the transportation of angular momentum increases with the accretion rates . otherwise the accretion would be halted because of the inefficient transportation . however , this does not agree with the observed correlation , which indicates that the dissipation in the cold disk ( not in the hot corona ) becomes more efficient with increases of the accretion rates . successful models of the hot corona should address this issue explicitly . the correlation of @xmath37 shows that the hard x - ray spectrum becomes harder with increases of @xmath2 , confirming the results of lu & yu ( 1999 ) and gierlinski & done ( 2004 ) . this is usually explained by that the seed photons efficiently cool the hot corona with increases of @xmath21 . however , this explanation could be replaced by that the hot corona really becomes weak since @xmath3 . indeed this agrees with that the broad line seyfert 1 galaxies have strong reprocessing of hard x - ray , such as in ngc 5548 ( clavel et al . 1992 ; chiang et al . 2000 ) , evidenced by that optical and uv variations follow the hard x - ray . in nsl1s the reprocessing is rather weak , such as akn 564 ( shemmer et al . 2001 ) , ngc 4051 ( done et al . 1990 ) and iras 13224 - 3809 ( young et al . assuming that the energy transportation from the cold disk to the hot corona is by buoyancy driving the magnetic tubes across extent of the disk , merloni & fabian ( 2002 ) proposed the fraction of the energy transported by magnetic buoyancy to be @xmath38 , where @xmath39 is the transporting velocity , @xmath40 the magnetic pressure and the dissipated energy @xmath41 via viscosity stress @xmath42 . we have @xmath43 where @xmath44 , @xmath45 alfvn velocity and the constant @xmath46 . here @xmath47 and @xmath48 . 0.5 cm [ fig2 ] according to eq.(4 ) , table 2 gives the properties of the factor @xmath8 for six distinct types of the magnetic stresses based on different assumptions . comments are given in table 2 based on the correlation ( eq . 3 ) . cases 1 ) , 3 ) , 4 ) and 6 ) predict properties inconsistent with the correlations in figure 1b , thus they could be ruled out . only cases 2 ) and 5 ) are possible since they provide characters of the hot corona consistent with the correlation . specifying the magnetic stress , we can solve the structure of disk and the factor @xmath8 at each radius formulated in svensson & zdziarski ( 1994 ) as described by merloni & fabian ( 2002 ) and merloni ( 2003 ) . we define the radial - averaged fraction @xmath49 , @xmath50 so that we can compare with the observation , if the magnetic stress is assumed . in case 2 ) the magnetic pressure @xmath51 is originally suggested by stella & rosner ( 1984 ) , who argued the buoyancy controls the growth of the magnetic field . the magnetic stress is simply scaled by the gas pressure . such a stress can stabilize the thermal instability of the cold disk ( svensson & zdziarski 1994 ) . on the other hand , ichimaru ( 1975 ) argued that the growth of the magnetic field may be controlled by the reconnection and gave a magnetic pressure tensor @xmath52 ( case 5 ) . this stress has been extensively studied by many authors ( burm 1985 , szuszkiewicz 1990 ) . recently merloni & fabian ( 2002 ) and merloni ( 2003 ) derived this simple formulation based on analytical work ( blaes & socrates 2001 ) and 3-d mhd numerical simulation ( turner et al . 2002 ) of magneto - rotational - instability in radiative accretion disks . figure 2 . shows the results of @xmath49 for the magnetic stresses @xmath51 and @xmath53 . we find that @xmath49 is not sensitive to the black hole mass . more interesting thing in figure 2 is that the slope of @xmath54 ( for @xmath55 ) is insensitive to both the black hole mass and the viscosity @xmath56 , but only sensitive to the magnetic stress . these properties of the theoretical relation of @xmath54 allow us to test the working magnetic stress in disk . we have @xmath57 for the merloni & fabian s model whereas @xmath58 for the stella & rosner s . mf02 predicts a flatter slope than the observation . sr84 model has a steeper slope , but it matches the observational correlation @xmath59 within the uncertainties in the present sample . a more fiducial test on the magnetic stress requires the accurate estimations of the black hole masses , bolometric luminosity and the total hard x - ray luminosity from observations . the near future observations of _ integral _ can provide entire hard x - ray spectra of a large sample so as to test the disk - corona models . the future 3-d mhd numerical simulations including the hot corona should be tested by the plot of @xmath60 . with the estimations of the black hole masses in radio - quiet agns , we find the hard x - ray spectrum index and @xmath61 strongly correlate with the eddington ratio . these correlations directly indicate that the fraction @xmath8 is controlled by the eddington ratio . such a correlation can be explained by the magnetic turbulence that plays two roles in transporting angular momentum and forming the hot corona . the correlation @xmath62 favors the magnetic stress @xmath9 in the present sample . balbus , s.a . & hawley , j.f . , 1991 , apj , 376 , 214 balbus , s.a . , 2003 , ara&a , 41 , 555 blaes , o.m . , 2002 , astro - ph/0211368 blaes , o.m . & socrates , a. , 2001 , apj , 553 , 987 blaes , o.m . & socrates , a. , 2003 , apj , 596 , 509 boroson , t. & green , r.f . , 1992 , apjs , 80 , 109 burm , h. , 1985 , a&a , 143 , 389 chiang , j. , et al . 2000 , apj , 528 , 292 clavel , j. et al . , 1992 , apj , 393 , 113 done , c. , madejski , g.m . & ycki , p.t . , 1990 , mnras , 243 , 713 elizalde , f. & steiner , j.e . , 1994 , mnras , 268 , l47 galeev , a.a . , rosner , r. & vaiana , g.s . , 1979 , apj , 229 , 318 gierlinski , m. & done , c. , 2004 , mnras , in press , astro - ph/0312271 grupe , d. , wills , b. , leighly , k.m . , meusinger , h. , 2004 , aj , 127 , 156 haardt , f. & maraschi , l. , 1991 , apj , 380 , l51 ichimaru , s. , 1975 , apj , 202 , 528 igumenshchev , i.v . , narayan , r. & abramowicz , m.a . , 2003 , apj , 592 , 1042 kaspi , s. , et a. , 2000 , apj , 533 , 631 kato , y. , mineshige , s. & shibata , k. , 2004 , apj , in press ( astro - ph/0307306 ) kawaguchi , t. , shimura , t. & mineshige , s. , 2001 , apj , 546 , 966 kuncic , z.a . & bicknell , g.v . , 2004 , apj , astro - ph/0402421 laor , a. , & netzer , h. , 1989 , mnras , 238 , 897 leighly , k.m . , 1999 , apjs , 125 , 317 liang , e.p.t & price , r.h . , 1977 , apj , 218 , 247 liu , b.f . , mineshige , s. & shibata , r. , 2002 , apj , 572 , l173 lu , y. & yu , q. , 1999 , apj , 526 , l5 mcclintock , j.e . & remillard , r.a . , 2003 , astro - ph/0306213 mclure , r.j . & dunlop , j.s . , 2001 , mnras , 327 , 1999 merloni , a. , 2003 , mnras , 341 , 1051 merloni , a. & fabian , a.c . , 2002 , mnras , 332 , 165 mineshige , s. et al . 2000 , pasj , 52 , 499 nandra , k. , et al . , 2000 , apj , 544 , 734 rndahl , e. , rnnback , j. & van groningen , r.e . , 2003 , a&a , 404 , 883 reeves , j.n . & turner , m.j.l . , 2000 , mnras , 316 , 234 sano , t. , inutsuka , s. , turner , n.j . & stone , j.m . , 2004 , apj , astro - ph/0312480 shakura , i. & sunyaev , r. , 1973 , a&a , 24 , 337 shemmer , o , et al . , 2001 , apj , 561 , 162 stella , l. & rosner , r. , 1984 , apj , 277 , 312 stone , j. , et al . , 1996 , apj , 463 , 656 svensson , r. & zdziarski , a. , 1994 , apj , 436 , 599 szuszkiewicz , e. , 1990 , mnras , 244 , 377 taam , r. & lin , d.n.c . , 1984 , apj , 287 , 761 tremaine , s. , et al . , 2002 , apj , 574 , 740 tout , c.a . & pringle , j.e . , 1992 , mnras , 259 , 604 turner , n.j . , 2004 , apj , astro - ph/0402539 turner , n.j , stone , j.m . & sano , t. , 2002 , apj , 566 , 148 turner , n.j . , stone , j.m . , krolik , j.h . & sano , t. , 2003 , apj , 593 , 992 turner , t.j . , goerge , i. , nandra , k. & turcan , d. , 1999 , apj , 524 , 667 turner , t.j . , et al . , 2001 , apj , 561 , 131 vaughan , s. , reeves , j. , warwick , r. & edelson , r. , 1999 , mnras , 309 , 113 vron - cetty , m .- p . & vron , p. , 2003 , a&a , 412 , 399 vestergaard , m. , 2002 , apj , 571 , 733 wandel , a. , & liang , e.p.t . , 1991 , apj , 380 , 84 woo , j .- h . & urry , m.c . , 2002 , apj , 579 , 530 young , a.j . , et al . , 1999 , mnras , 304 , l46
we compile a sample consisting of 56 radio - quiet active galactic nuclei so as to investigate statistical properties of hot corona of accretion disks from _ asca _ observations . the black - hole masses in the sample are estimated via several popular methods and the bolometric luminosities from the multi - wavelength continuum . this allows us to estimate the eddington ratio ( @xmath0 ) so that the undergoing physical processes can be tested via hard x - ray data . we find a strong correlation between @xmath1 and @xmath2 as @xmath3 with a multivariate regression . this indicates that the release of gravitational energy in the hot corona is controlled by the eddington ratio . on the other hand , the correlation between the hard x - ray spectral index ( @xmath4 ) and @xmath2 depends critically on the types of objects : @xmath4 is nearly constant ( @xmath5 ) in broad - line seyfert 1 s ( bls1s ) , whereas @xmath6 in narrow - line seyfert 1 s ( nls1s ) , although not very significant . we can set constraints on the forms of magnetic stress tensor on the condition that @xmath7 is proportional to the fraction @xmath8 of gravitational energy dissipated in the hot corona and that @xmath8 is proportional to magnetic energy density in the disk . we find that the shear stress tensor @xmath9 is favored by the correlation in the present sample , where @xmath10 is the gas pressure .
non - equilibrium growth and aging phenomena are of great interest due to their wide applications across various scientific fields of study , including many body statistical systems , condensed matter systems , biological systems and so on @xcite-@xcite . they are complex physical systems , and details of microscopic dynamics are widely unknown . thus it is best to describe these systems with a small number of variables , their underlying symmetries and corresponding universality classes , which have been focus of nonequilibrium critical phenomena . one particular interesting class is described by the kardar - parisi - zhang ( kpz ) equation @xcite@xcite . recently , this class is realized in a clean experimental setup @xcite@xcite , and their exponents for one spatial dimension @xmath11 is confirmed : the roughness @xmath12 , the growth @xmath13 and the dynamical @xmath14 exponents . along with the experimental developments , there have been also theoretical developments in the context of aging . the kpz class reveals also simple aging in the two - time response functions @xcite@xcite . in these works , it is shown that the autoresponse function of the class is well described by the logarithmic ( log ) and logarithmic squared ( log@xmath15 ) extensions of the scaling function with local scale invariance for @xmath11 . in a recent paper @xcite based on @xcite@xcite , we have considered log extensions of the two - time correlation and response functions of the scalar operators with the conformal schrdinger and aging symmetries for the spatial dimension @xmath16 and the dynamical exponent @xmath17 , in the context of gauge / gravity duality @xcite@xcite . the power - law and log parts are determined by the scaling dimensions of the dual field theory operators , the eigenvalue of the internal coordinate and the aging parameter , which are explained below . interestingly , our two - time response functions show several qualitatively different behaviors : growth , aging ( power - law decaying ) or both behaviors for the entire range of our scaling time , depending on the parameters in our theory @xcite . we further have made connections to the phenomenological field theory model @xcite in detail . the two - time response functions and their log corrections of our holographic model @xcite are completely fixed by a few parameters and are valid for @xmath18 , while the field theory model @xcite has log@xmath15 extensions and is valid for @xmath19 . closing the gap between these two models from the holographic side is one of the main motivation of this work . in this paper , we generalize our analysis @xcite in two different directions : ( 1 ) by applying to general dynamical exponent @xmath1 ( not conformal ) as well as to any number of spatial dimensions @xmath0 and ( 2 ) by including log@xmath15 corrections in two - time response functions . in [ sec : ntlft ] , we first analytically compute the correlation and response functions for general @xmath1 and @xmath0 along with their log and log@xmath15 extensions . then we add the aging generalizations of our non - conformal results in [ sec : aging ] . we try to make contact with kpz class in [ sec : kpz ] , and conclude in [ sec : conclusion ] . logarithmic conformal field theory(lcft ) is a conformal field theory(cft ) which contains correlation functions with logarithmic divergences . they typically appear when two primary operators with the same conformal dimensions are indecomposable and form a jordan cell . the natural candidates for the bulk fields , in the holographic dual descriptions of lcft , of the pair of two primary operators forming jordan cell are given by a pair of fields with the same spin and a special coupling . after integrating out one of them , it becomes the model with higher derivative terms . the ads dual construction of the lcft was first considered in @xcite@xcite@xcite using a higher derivative scalar field on ads background . recently , higher derivative gravity models on ads geometry with dual lcft have got much attention , starting in three dimensional gravity models @xcite-@xcite . in four and higher dimensional ads geometry , the gravity models with curvature - squared terms typically contain massless and ghostlike massive spin two fields . when the couplings of the curvature - squared terms are tuned , the massless and massive modes become degenerate and turn into the massless and logarithmic modes @xcite . this , so called , critical gravity has the boundary dual lcft which contains stress - energy tensor operator and its logarithmic counterpart . more recently , studies on generalizations of lcft in the context of ads / cft correspondence @xcite@xcite , in particular the correlation functions of a pair of scalar operators , have been made in two different directions . one is on the non - relativistic lcft . in @xcite the dual lcft to the scalar field in the lifshitz background has been investigated . the study on the dual lcft to the scalar field in the schrdinger and aging background was made in @xcite . the other is on the lcft with @xmath20 divergences . correlation functions with log@xmath15 corrections have been investigated in several works . in the context of the gravity modes of tricritical point , they are interpreted as rank-3 logarithmic conformal field theories ( lcft ) with log@xmath15 boundary conditions @xcite . explicit action for the rank-3 lcft is considered in @xcite . see also a recent review on these developments in @xcite . in this section we would like to accomplish two different things motivated by these developments along with those explained in the introduction . first , we compute correlation and response functions for ads in light - cone ( alcf ) and schrdinger type gravity theories , which are dual to some non - relativistic field theories with galilean invariance for general dynamical exponent @xmath1 and @xmath0 . second , we generalize our correlation and response functions with log@xmath15 as well as log contributions for @xmath1 and @xmath0 . the log correction has been investigated in @xcite in the context of schrdinger geometry and aging geometry for @xmath16 and @xmath17 as mentioned in the introduction . we first consider the alcf @xcite@xcite . due to several technical differences , we present our computations in some detail , building up correlation functions [ sec : alcfcorr ] , their log extensions [ sec : alcfcorrlog ] and log@xmath15 extension [ sec : alcfcorrlog2 ] . for schrdinger backgrounds @xcite@xcite , we comment crucial differences in [ sec : schrlog ] . and then we present the correlation functions for @xmath17 in [ sec : ztwo ] and for @xmath21 in [ sec : zthreehalf ] . we summarize our results in [ sec : ntlftsummary ] . let us turn to the ads in light - cone ( alcf ) with galilean symmetry studied in @xcite@xcite@xcite . the case for general @xmath22 is also considered in @xcite for zero temperature and in @xcite for finite temperature . the metric is given by @xmath23 which is invariant under the space - time translations @xmath24 , galilean boost @xmath25 , @xmath26 scale transformation @xmath27 , @xmath28 and translation along the @xmath29 coordinate , which represents the dual particle number or rest mass . , @xmath30 it turns out that this does not provide a closed algebra with other symmetry generators for @xmath31 . ] the geometry satisfies vacuum einstein equations with a negative cosmological constant . the finite temperature generalizations of the alcf for @xmath17 and @xmath16 have been considered in @xcite@xcite@xcite . we compute correlation functions of the geometry ( [ alcfmetric ] ) by coupling a probe scalar . @xmath32 where @xmath33 is a coupling constant , and @xmath34 . we use @xmath35 for our boundary cutoff . the field equation of @xmath36 for general @xmath1 and @xmath0 is @xmath37 note that we treat @xmath38 coordinate special and replace all the @xmath39 as @xmath40 . this is in accord with the fact that the coordinate @xmath38 plays a distinguished role in the geometric realizations of schrdinger and galilean symmetries @xcite@xcite@xcite . the general solution is given by @xmath41 where @xmath42 , @xmath43 represent bessel functions . we choose @xmath44 over @xmath45 due to its well defined properties deep in the bulk . we follow @xcite@xcite to compute correlation functions by introducing a cutoff @xmath35 near the boundary and normalizing @xmath46 , which fixes @xmath47 . we compute an on - shell action to find @xmath48 & = \int d^{d+1 } x \frac{l^{d+3}}{u^{d+3 } } ~\phi^ * ( u , t , \vec x ) ~\frac{u^2}{l^2}\partial_u \phi ( u , t , \vec x ) \big |_{u_b } \;. \end{aligned}\ ] ] using @xmath49 , @xmath50 , @xmath51 and @xmath52 the onshell action can be rewritten as @xmath53 \int [ x , \vec k ' , \vec k ] \quad \phi_0^ * ( k ' ) \mathcal f_1 ( u , k',k ) \phi_0 ( k ) \big |_{u_b } \ ; , \label{wholeeq}\end{aligned}\ ] ] where @xmath54 \int [ x , \vec k ' , \vec k ] = \int d t ~\frac{d \omega'}{2 \pi } \frac{d \omega}{2 \pi } e^{-i ( \omega ' -\omega ) t } \int d^d x \int \frac{d^d k'}{(2\pi)^d } \frac{d^d k}{(2\pi)^d } e^{i(\vec k ' - \vec k ) \cdot \vec{x } } $ ] . @xmath55 is given by @xmath56 this function appears again when we construct the log and log@xmath15 extensions below . for general @xmath0 we find @xmath57 one can evaluate the @xmath58-dependent part of @xmath55 at the boundary , @xmath59 , by expanding it for small @xmath35 . we obtain the following non - trivial contribution @xmath60 note that the function @xmath55 is only a function of @xmath61 when it is evaluated at the boundary . by inverse fourier transform of ( [ f1function ] ) for an imaginary parameter @xmath62 , we get the following coordinate space correlation functions for a dual field theory operator @xmath36 @xmath63 these are our correlation functions evaluated for general dynamical exponent @xmath1 and for the number of spatial dimensions @xmath0 , which are direct generalizations of a previous result for @xmath16 and @xmath17 @xcite . note that the result is valid for the field theories with galilean boost without conformal symmetry . in particular , the parameter @xmath2 carries scaling dimensions @xmath64 = z-2 $ ] , and the exponent is actually dimensionless . this is consistent with the scaling properties written in ( [ alcfscaling ] ) . the result ( [ f1correlator ] ) is independent of @xmath1 , while depending on the number of dimensions @xmath0 . this is expected because the alcf metric ( [ alcfmetric ] ) is independent of @xmath1 , which is a special feature for the alcf . this is not true for schrdinger background as we see below . motivated by the recent interests on lcft from the holographic point of view @xcite@xcite , we consider two scalar fields @xmath36 and @xmath65 in the background ( [ alcfmetric ] ) @xmath66 where @xmath67 represents a cutoff near the boundary . we take @xmath68 and @xmath69 . the field equations for @xmath36 and @xmath65 of the action ( [ schractionlcft ] ) become @xmath70 where @xmath71 \ ; , \end{aligned}\ ] ] which is a differential operator for alcf . following @xcite@xcite , we construct bulk to boundary green s functions @xmath72 @xmath73\ ; , \nonumber \\ & \tilde \phi ( u , x ) = \int \frac{d^{d+1 } k'}{(2\pi ) ^{d+1 } } e^{-i k ' \cdot x } \left [ g_{21 } ( u , k ' ) j(k')+ g_{22 } ( u , k ' ) \tilde j(k')\right]\;. \end{aligned}\ ] ] we have @xmath74 , which follows from the structure of the equations of motion given in ( [ diffeqschr ] ) and the action ( [ schractionlcft ] ) . the green s functions satisfy @xmath75 where @xmath76 and @xmath77 . solutions of @xmath78 and @xmath79 are given by @xmath80 where @xmath81 . the normalization constants @xmath82 can be determined by requiring that @xmath83 @xcite@xcite . there exists another green s function @xmath84 due to a coupling between @xmath36 and @xmath65 in the action ( [ schractionlcft ] ) , which satisfies @xmath85 to evaluate @xmath84 , we use the same methods used in @xcite@xcite . using @xmath86 = 2 \nu \ ; , \end{aligned}\ ] ] and the fact that @xmath87 , we get @xmath88 thus we have an explicit form . @xmath89 after plugging the bulk equation of motion into the action ( [ schractionlcft ] ) , the boundary action becomes of the form @xmath90 \int [ x , \vec k ' , \vec k ] ~~\tilde j ( k ' ) \big [ \mathcal f_1 ( u_b , k ' , k ) j ( k ) + \mathcal f_2 ( u_b , k ' , k ) \tilde j ( k ) \big ] \;. \end{aligned}\ ] ] for alcf , the system has the time translation invariance , thus the time integral is trivially evaluated to give a delta function . the @xmath91 s are given by @xmath92 we note that @xmath55 leads the same result as ( [ f1function ] ) and ( [ f1correlator ] ) . let us evaluate @xmath93 , which is @xmath94 derivative of @xmath55 given in ( [ g12expressionlcft ] ) . the result is @xmath95 \right ) \;. \end{aligned}\ ] ] these are our correlation and response functions with log extensions for general @xmath1 and @xmath0 . this is a direct generalization of the previous result for @xmath16 and @xmath17 @xcite . note that the result is valid for the field theories with galilean boost without conformal symmetry similar to the result ( [ f1correlator ] ) . again it is independent of @xmath1 . motivated by the recent interests on tricritical log gravity @xcite , we consider three scalar fields @xmath96 , @xmath97 and @xmath98 in the background ( [ alcfmetric ] ) with the following action @xmath99 + \frac{\phi_3 \phi_2}{2l^2 } \right ] \ ; , \end{aligned}\ ] ] where we take @xmath100 for @xmath101 . this action is previously considered in @xcite in a different context . the field equations for @xmath36 s of the action ( [ schractionlcft ] ) become @xmath102 where @xmath103 is given in ( [ diffopalcf ] ) . we construct the bulk to boundary green s function @xmath72 in terms of @xmath104 as @xmath105 we choose @xmath106 , which is in accord with the structure of the equations of motion given in ( [ diffeqschrtlcft ] ) . the green s functions satisfy @xmath107 where @xmath108 and @xmath77 . the green s functions @xmath109 are @xmath110 where @xmath111 with the same normalization constant given in ( [ onshact1 ] ) . there exist other green s functions @xmath112 for the action ( [ schractiontlcft ] ) , which satisfies @xmath113 in particular , we have @xmath114 to evaluate them , we generalize the methods used in @xcite@xcite to the tricritical case . using again @xmath115 = 2 \nu $ ] , and the fact that @xmath116 , we get @xmath117 thus @xmath118 note that the last expression has second order derivative of @xmath94 , which leads log@xmath15 contributions . after plugging the bulk equation of motion into the action ( [ schractiontlcft ] ) , the boundary action becomes of the form @xmath119 \int[\vec x , \vec k ' , \vec k ] ~~ j_i ( k'_\mu ) \mathcal f_{ij } ( u_b , k'_\mu , k_\mu ) j_j ( k_\mu ) \;. \end{aligned}\ ] ] the system has time translation invariance , thus the time integral is trivially evaluated to give delta function . the @xmath91 s are given by @xmath120 note that the first terms in @xmath121 and @xmath122 are @xmath123 when evaluated at @xmath59 . we also notice that @xmath124 and @xmath125 are identical to ( [ f1function ] ) and ( [ f2function ] ) , and thus the corresponding correlation functions ( [ f1correlator ] ) and ( [ f2correlator ] ) , respectively . now we are ready to evaluate @xmath126 . using ( [ f3function ] ) and ( [ f1correlator ] ) , we get @xmath127 - \nu^2 \ln [ \frac{\mathcal m u_b^2}{2(t_2 - t_1)}]^2 \right ] \ ; , \nonumber \end{aligned}\ ] ] where @xmath128 . this is our main result in this section , response functions with log and log@xmath15 extensions , which is valid for general @xmath1 and @xmath0 . note also that this result is valid for the systems without non - relativistic conformal invariance . we notice that various coefficients in the square bracket are completely determined once @xmath94 is fixed . we first establish the schrdinger type solutions with galilean symmetry with @xmath129 following @xcite , see also @xcite@xcite . finite temperature generalizations for general @xmath1 is considered in @xcite , while those for @xmath17 are considered in @xcite@xcite@xcite@xcite@xcite . the metric at zero temperature is given by @xmath130 which is invariant under the space - time translations @xmath24 , galilean boost @xmath25 , scale transformation @xmath131 and translation along the @xmath29 coordinate . their explicit forms are given in ( [ galileantr1 ] ) and ( [ scaletr1 ] ) . there exists additional special conformal transformation for @xmath17 , which has been focus of previous investigations . there have been more general class of gravity backgrounds with so - called hyperscaling violation . these backgrounds are described by @xmath132 $ ] considered in @xcite , where @xmath133 is a hyperscaling violation exponent . @xmath133 is first introduced in @xcite based on @xcite . this hyperscaling violation might be also interesting in the general context of aging and growth phenomena . the associated matter fields are a gauge field , a scalar and the non - trivial coupling between them . the geometry ( [ schrmetric ] ) is not a solution of vacuum einstein equations . thus we require to support it with some matter fields . one particular example is the ground state of an abelian higgs model in its broken phase @xcite @xmath134 where @xmath135 , @xmath136 , and @xmath0 is the number of spatial dimensions . it is not hard to find a different matter system that supports the metric @xcite . @xmath137 where @xmath138 . we are interested in constructing correlation and response functions using three different actions , ( [ scalaractionalcf ] ) , ( [ schractionlcft ] ) and ( [ schractiontlcft ] ) , as in the previous section [ sec : alcf ] . here we briefly show that the procedure is the same as before . thus we can compute the logarithmic ( squared ) extensions by taking a simple @xmath94 derivatives of the correlation function obtained from the action ( [ scalaractionalcf ] ) . we start by considering correlation functions of the geometry ( [ schrmetric ] ) by coupling a probe scalar with the same action as ( [ scalaractionalcf ] ) . the field equation for @xmath36 becomes @xmath140 note @xmath141 , which is one of the main differences between the schrdinger background ( [ schrmetric ] ) and alcf ( [ alcfmetric ] ) . again , we treat @xmath38 coordinate special and replace all the @xmath39 as @xmath40 . with this differential equation , one can compute the correlation function @xmath55 . for general @xmath1 , analytic solutions are not available . the resulting correlation function for @xmath17 is already computed in @xcite@xcite@xcite , while that of @xmath21 is computed below in [ sec : zthreehalf ] . previously , several special cases also have been computed in @xcite . to compute the corresponding log and log@xmath15 extensions , we consider a schrdinger differential operator @xmath142 \ ; , \end{aligned}\ ] ] where @xmath143 and @xmath144 for @xmath129 . for the special case @xmath17 , we have @xmath145 from ( [ diffopschr ] ) . with the differential operator @xmath146 , we can still use the relation @xmath147 = 2 \nu \ ; , \end{aligned}\ ] ] to compute the correlation ( response ) functions with the logarithmic extensions . for that purpose , we use the equations ( [ bulktoboundarygreensfunctionlcft ] ) - ( [ f1function2 ] ) with appropriate @xmath148 and @xmath149 . we also get the response functions with the log@xmath15 extension using ( [ bulktoboundarygreensfunctiontlcft ] ) - ( [ lastfstlcft ] ) with appropriate @xmath150 and @xmath151 . the upshot is that the logarithmic extensions can be computed by taking one or two derivatives of the correlation functions available . let us compute these correlation and response functions for @xmath17 and @xmath21 in turn . we comment for the conformal case @xmath17 here . the differential equation ( [ blukscalareqschr ] ) simplifies to @xmath152 this is similar to that of alcf given in ( [ blukscalareqalcf ] ) , the only difference is the presence of the parameter @xmath153 , which modifies @xmath94 as @xmath154 . this observation leads us that we can compute correlation and response functions with logarithmic extensions as in [ sec : alcf ] . these are ( [ f1correlator ] ) , ( [ f2correlator ] ) and ( [ f3correlator ] ) with modified @xmath94 . the case @xmath19 is our main interest for the application to kpz universality class . for the time being , we work on general spatial dimensions @xmath0 . the corresponding solution is @xmath155 \ ; , \end{aligned}\ ] ] where @xmath156 , @xmath157 , @xmath144 and @xmath143 . @xmath158 and @xmath159 represent the confluent hypergeometric function and the generalized laguerre polynomial . we choose @xmath158 for our regular solution . the momentum space correlation function can be evaluated as the ratio between the normalizable and non - normalizable contributions at the boundary expansion of the solution ( [ schr32sol ] ) , which is given by @xcite @xmath160 } { \gamma [ 2 \nu ] } \frac{\gamma \left[\frac{1 + 2\nu}{2}+\frac{m^2}{2 q}\right ] } { \gamma \left[\frac{1 - 2\nu}{2}+\frac{m^2}{2 q}\right ] } \ ; , \end{aligned}\ ] ] where we only keep momentum dependent parts . one can restore the @xmath35 dependence using scaling arguments . for the general case , fourier transforming back analytically to the coordinate space is difficult . thus we would like to consider some special cases . * @xmath161 : the momentum space correlator has the same form as aging in alcf @xmath162 } { \gamma [ 2 \nu ] } \frac{\gamma \left[\frac{1 + 2\nu}{2}\right]}{\gamma \left[\frac{1 - 2\nu}{2}\right ] } \;. \end{aligned}\ ] ] for imaginary parameter @xmath62 , we get @xmath163 this result is for @xmath164 . the dependence on time and space is identical to the result of the aging in alcf . + we are interested in log and log@xmath15 extensions . for this purpose , we consider @xmath165 where @xmath166 collectively denotes the other @xmath94 dependent parts . we also use the same actions ( [ schractionlcft ] ) and ( [ schractiontlcft ] ) to get the correlation functions with the logarithmic extension @xmath167 \right ) \langle \phi ( x_{2 } ) \phi ( x_{1 } ) \rangle _ { \mathcal f_1}^{m_0 } \ ; , \end{aligned}\ ] ] and the log@xmath15 extension @xmath168 + a_2 \ln [ \frac{\mathcal m u_b^2}{(t_2 - t_1)}]^2 \right ] \langle \phi _ 2 \phi_1 \rangle _ { \mathcal f_1}^{m _ 0 } \ ; , \end{aligned}\ ] ] where @xmath169+\nu h''[\nu ] } { h[\nu ] } $ ] , @xmath170 + 2 \nu h'[\nu ] } { h[\nu ] } $ ] and @xmath171 . * @xmath172 : we use the asymptotic expansion form from 5.11 of @xcite @xmath173 } { \gamma [ z+b ] } \sim z^{a - b } \sum_{k=0}^{\infty } \frac{g_k ( a , b)}{z^k } \ ; , \quad g_{k}(a , b)=\binom{a - b}{k } b^{{(a - b+1)}}_{{k } } ( a ) \;,\end{aligned}\ ] ] where @xmath174 are binomial coefficients and @xmath175 s are generalized bernoulli polynomials . for our case , @xmath176 . + the momentum space correlation function is @xmath177 } { \gamma [ 2 \nu ] } \left ( \frac{m^2}{2 q } \right)^{2\nu } \sum_{n=0}^{\infty } g_{2n } \left(\frac{q}{m^2 } \right)^{2n } \;,\end{aligned}\ ] ] the first term is independent of momenta , which we ignore . for the rest of the terms , the inverse fourier transform of @xmath178 gives us @xmath179 , which vanishes for integer @xmath180 . thus the coordinate correlation function identically vanishes except the case @xmath181 . thus we get , using @xmath182 @xmath183 } \frac{\mathcal m^{\frac{d}{2}-2\nu}}{\pi 2^{1 - 2\nu } } \cdot \frac{\theta ( x_2^+ ) \theta(t_2 - t_1 ) } { ( t_2 - t_1)^{\frac{d+2}{2}+2\nu } } \cdot \exp \left ( - { \frac{\mathcal m ( \vec{x}_2 - \vec{x}_1)^2 } { 2 ( t_2 - t_1)}}\right ) \;. \end{aligned}\ ] ] + we are interested in the response functions with log and log@xmath15 extensions , we consider @xmath184 where @xmath185 collectively denotes the other @xmath94 dependent parts . we also use the same actions ( [ schractionlcft ] ) and ( [ schractiontlcft ] ) to get the correlation functions with the logarithmic extension @xmath186 \right ) \langle \phi^ * \phi \rangle _ { \mathcal f_1}^{m_\infty } \ ; , \end{aligned}\ ] ] and the log@xmath15 extension @xmath187 + \tilde a_2 \ln [ \frac{\mathcal m u_b^2}{(t_2 - t_1)}]^2 \right ] \langle \phi^ * \phi \rangle _ { \mathcal f_1}^{m_\infty } \ ; , \end{aligned}\ ] ] where @xmath188+\nu \tilde h''[\nu ] } { \tilde h[\nu ] } $ ] , @xmath189 + 4 \nu \tilde h'[\nu ] } { \tilde h[\nu ] } $ ] and @xmath190 . these two extreme cases , @xmath191 and @xmath192 , signal that the parameter @xmath2 can bring some quantitatively different behaviors of the correlation and response functions because of the different power in time dependent denominators @xmath193 and @xmath194 in ( [ schrcorrf1 ] ) and ( [ schrcorrlargemf1 ] ) , respectively . in this section we summarize [ sec : ntlft ] by considering the two - time correlation and response functions with logarithmic extensions . from the various results of alcf and schrdinger backgrounds , equations ( [ f3correlator ] ) , ( [ schrcorrf3 ] ) and ( [ schrcorrlargemf3 ] ) , we observe that the correlation functions with(out ) log extensions show qualitatively similar properties . some typical two - time correlation and response functions can be obtained by putting @xmath195 in equation ( [ f3correlator ] ) . @xmath196 + a_2 \ln [ \frac{\mathcal m_b}{t_2 - t_1}]^2 \right ] \ ; , \end{aligned}\ ] ] where @xmath197 , and the coefficients @xmath198 we note that these coefficients , @xmath199 and @xmath200 , are determined once @xmath94 is fixed . @xmath201 is invariant under the time translation transformation , and so @xmath202 . the so - called `` waiting time '' @xmath203 does not have a physical meaning . thus @xmath204 is completely fixed as a function of @xmath205 , once @xmath0 , @xmath94 and @xmath206 are given . physically , this time translation invariant two - time response functions describe either constant growth or constant aging ( decaying ) phenomena . further physical significances are considered in detail in [ sec : kpz ] . equipped with the generalization of our correlation and response functions for general @xmath1 and @xmath0 , non - relativistic and ( non-)conformal geometries , we would like to add yet another ingredient to them : aging , one of the simplest time - dependent physical phenomena . typically aging is realized when the system is rapidly brought out of equilibrium . for this simple time - dependent phenomena , time translational invariance is broken . there are two important time scales : ( 1 ) _ waiting time _ which marks the time scale when the system is perturbed after it is put out of equilibrium and ( 2 ) _ response time _ which marks when the perturbation is measured . typical properties of aging are described by the two - time response functions in terms of these waiting time and response time , and are power law decay , broken time translation invariance and dynamical scaling between the time and spatial coordinates . these are shown in holographic model in @xcite as well as various field theoretical models , see _ e.g. _ @xcite@xcite@xcite . in the context of anti - de sitter space / conformal field theory correspondence ( ads / cft ) @xcite@xcite and its extension to schrdinger geometries @xcite@xcite@xcite@xcite , the geometric realizations of aging have been put forward in @xcite@xcite by generalizing the background with explicit time dependent terms . these terms are generated by a singular time dependent coordinate transformation , which itself has significant physical meaning in the context of holography @xcite . furthermore , there exists a time boundary at @xmath207 and physical boundary conditions are explicitly imposed : ( 1 ) by complexifying time in @xcite or ( 2 ) by introducing some decay modes of the bulk scalar field along the ` internal ' spectator direction @xmath38 , which is not explicitly visible from the dual field theory in @xcite . we prefer the option ( 2 ) in this paper as in @xcite , where the resulting two - time correlation functions show a dissipative behavior and exhibit the three characteristic features of the aging system mentioned above . thus the time translation symmetry is broken globally , and the aging symmetry is realized as conformal schrdinger symmetry modulo time translation symmetry @xcite@xcite . their finite temperature properties with asymptotic aging invariance are also investigated in @xcite . see also a recent review @xcite . in this section we would like to generalize this aging construction to the case with general dynamical exponent @xmath1 and for general dimensions @xmath0 . the generalization of the singular coordinate transformation and the corresponding aging geometries are constructed in [ sec : agingtransform ] . in [ sec : agingalcf ] , we construct the two point correlation and response functions for alcf in the context of @xcite@xcite , while similarly in [ sec : agingback ] for schrdinger background in the context of @xcite@xcite . their log and log@xmath15 extensions are explained in [ sec : aginglogextension ] . physical properties of aging is explored in holography by using a _ singular _ coordinate transformation @xmath208 which is first introduced in @xcite , specifically for @xmath17 case . it is important to impose physical boundary conditions on the time boundaries in addition to the spatial boundaries . the simplest possibility in this context has been explored in @xcite we would like to extend this singular transformation for general @xmath1 in a direct manner . @xmath209 note that for general @xmath1 , the coordinate @xmath38 has non - trivial dimensions , @xmath210 = 2-z $ ] , under the scaling transformation . one immediate consequence is a nontrivial scaling dimension of our parameter @xmath64 = z-2 $ ] . this is already observed in the exponent of the correlation and response functions in previous section . now for the aging extension , we observe that the parameter @xmath3 also has a definite scaling dimension @xmath211 = 2-z$ ] . these two parameters conspire to provide us a rather simple and elegant generalization to the aging correlation and response functions for general @xmath1 . the background metric extended to the aging is correspondingly modified to @xmath212 where @xmath213 corresponds to aging in alcf . there exists a slight change in metric compared to ( [ alcfmetric ] ) or ( [ schrmetric ] ) : the coefficient of the term @xmath214 has a factor of @xmath1 instead of @xmath215 . one can check that the matter contents without the singular transformation would solve the corresponding einstein equation . these cases can be considered as _ locally galilean_. to compute the correlation functions of the probe scalar fields in the background geometry ( [ agingmetric ] ) with general @xmath1 and @xmath0 , we consider the action given in ( [ scalaractionalcf ] ) . the field equation for @xmath36 becomes @xmath216 \phi \\ & = \frac{\partial^2 \phi}{\partial u^2 } + \frac{z i m\alpha -d-1}{u } \frac{\partial \phi}{\partial u } -\left [ \frac { 4 m^2 l^2 + 2(d+2 ) z i m \alpha + z^2 m^2 \alpha^2 } { 4 u^2 } + \frac{\gamma m^2 } { u^{2z-2 } } + \vec k^2 \right ] \phi \;. \nonumber \end{aligned}\ ] ] note that here we treat @xmath38 coordinate special and replace all @xmath39 as @xmath40 , because this coordinate plays a distinguished role in galilean and corresponding aging holography @xcite@xcite . to find the solution of the equation ( [ blukscalareqagingalcf ] ) , we use the fourier decomposition as @xmath217 where @xmath218 is the momentum vector for the corresponding coordinates @xmath219 . @xmath220 is introduced for the calculation of the correlation functions and is determined by the boundary condition with the normalization @xmath221 . and @xmath222 is the kernel of integral transformation that convert @xmath223 to @xmath224 , which is necessary for our time dependent setup @xcite . with this fourier mode , the differential equation ( [ blukscalareqagingalcf ] ) decomposes into time dependent part and radial coordinate dependent one . the time dependent equation and solution read @xmath225 the radial dependent equation is given by @xmath226 where @xmath227 . from this point we can not carry on the analysis for both the aging in alcf , @xmath228 , and aging background @xmath141 simultaneously . thus , we first consider the correlation functions of the scalar operator in aging alcf . for @xmath228 , an analytic solution of the equation ( [ feq ] ) is available as @xmath229 where @xmath230 and @xmath231 are bessel functions with @xmath232 and @xmath233 . note the overall @xmath3 dependent factor , which is a non - trivial feature of our model . we also consider the boundary condition along time direction near the boundary . the solution behaves as @xmath234 along with the time dependent factor @xmath235 in ( [ agingtimesol ] ) , whose inverse fourier transform is given by @xmath236 this wave function converges for @xmath237 if @xmath238 , and for @xmath239 due to the exponential factor if @xmath240 . in particular , this condition allows the parameter @xmath241 to be negative @xmath242 especially for the case @xmath243 . similar result for @xmath17 and @xmath16 is already considered in @xcite . note that we only consider the imaginary @xmath244 . we follow @xcite to compute the correlation functions by introducing a cutoff @xmath35 near the boundary and normalizing @xmath46 , which fixes @xmath245 . the on - shell action is given by @xmath246 & = \int d^{d+1 } x \frac{l^{d+3}}{u^{d+3 } } ~\phi^ * ( u , t , \vec{x } ) ~\left(\frac{u^2}{l^2}\partial_u + i m \frac{z \alpha u}{2 l^2 } \right ) \phi ( u , t , \vec{x } ) \big |_{u_b } \;. \end{aligned}\ ] ] this can be recast using @xmath247 as @xmath248 where @xmath249 represents the existence of a physical boundary in the time direction , @xmath250 , and @xmath91 is @xmath251 note that the spatial integration along @xmath252 can be done trivially to give a delta function @xmath253 . one can bring @xmath254 factors in @xmath255 and @xmath256 together to cancel each other . this removes the second part in @xmath257 . from this point it is straight forward to check that @xmath91 is given by ( [ f1function ] ) at the boundary . further details can be found in @xcite . for imaginary parameter @xmath62 , we get the same correlation function as @xmath258 this is one of our main results . the aging correlation functions for general dynamical exponent @xmath1 and @xmath0 have a direct relation with those of schrdinger as @xmath259 the overall time dependent factor @xmath260 comes from the inverse fourier transform of the time part , which has been evaluated in great detail @xcite . thus the result is independent of @xmath1 , while depending on the number of dimensions @xmath0 . the corresponding extensions with log and log@xmath15 are considered below in [ sec : aginglogextension ] . for aging backgrounds , we have nontrivial @xmath1 dependence , and we need to treat them separately . fortunately , an analytic solution is available for @xmath21 @xmath261 + c_2 l_a^{2\nu } ( 2 q u ) \right ) \ ; , \end{aligned}\ ] ] where @xmath262 , @xmath263 and @xmath143 . @xmath158 and @xmath159 represent the confluent hypergeometric function and the generalized laguerre polynomial . we choose @xmath158 for our regular solution . the momentum space correlation function turns out to be the same as ( [ schrmomentumcorr ] ) as explained there . for the rest , we follow similarly [ sec : agingalcf ] to get the aging correlation and response functions . finally , we arrive general conclusion @xmath264 the overall time dependent factor @xmath265 comes from the inverse fourier transform of the time part , which has been evaluated in great detail @xcite . as we mentioned in [ sec : ntlftsummary ] , the aging in alcf and aging background have similar properties as far as the correlation and response functions are concerned . thus we present the logarithmic extension of the aging correlation functions using ( [ ficorrelatoraginglc ] ) . in the previous sections , [ sec : agingalcf ] and [ sec : agingback ] , we establish the fact that the correlation functions have the overall time dependent factor @xmath265 from the time dependent part of the momentum correlation function . in section [ sec : ntlft ] , on the other hand , we developed the algorithm to generate the logarithmic extensions using @xmath94 derivatives from the fact @xmath266 = 2 \nu$ ] . these two generalizations are independent of each other . thus we safely generate the logarithmic extensions of the aging correlation functions by differentiating the aging correlation functions in terms of @xmath94 . @xmath267\ ; , \label{generalformulalogaging2 } \\ & \langle \phi ( x_{2 } ) \phi ( x_{1 } ) \rangle_{aging}^{\mathcal f_3 } = \left ( \frac{t_2}{t_1 } \right)^{-\frac{\alpha \mathcal m}{2 } } \frac{1}{4\nu } \frac{\partial}{\partial \nu } \left [ \frac{1}{2\nu } \frac{\partial}{\partial \nu } \left [ \langle \phi ( x_{2 } ) \phi ( x_{1 } ) \rangle_{schr}^{\mathcal f_1 } \right ] \right ] \;. \label{generalformulalogaging3 } \end{aligned}\ ] ] the results @xmath268 are given in equations ( [ f1correlator ] ) , ( [ schrcorrf1 ] ) and ( [ schrcorrlargemf1 ] ) . their specific forms are @xmath269 \right ) \langle \phi ( x_{2 } ) \phi ( x_{1 } ) \rangle_{aging}^{\mathcal f_1 } \ ; , \\ \langle \phi_3 ( x_{2 } ) \phi_3 ( x_{1 } ) \rangle^{\mathcal f_3}_{aging } \label{f3correlatoraging } & = - \frac{1}{8 \nu^4 } \left [ a_0 + a_1 \ln [ \mathcal m_{\delta t } ] + a_2 \ln [ \mathcal m_{\delta t}]^2 \right ] \langle \phi ( x_{2 } ) \phi ( x_{1 } ) \rangle_{aging}^{\mathcal f_1 } \ ; , \end{aligned}\ ] ] where @xmath270 these are main results of our aging response functions . physical significances related to them are discussed in the following section . in this section we would like to seek a connection to kpz universality class , its growth , aging or both phenomena at the same time . our investigation is concentrated on the generalizations of two - time response functions for general dynamical exponent @xmath1 and for general spatial dimensions @xmath0 , along with their generalizations with the log and log@xmath15 contributions . previously , we observed that our two - time response functions reveal several qualitatively different behaviors , such as growth , aging or both in our holographic setup @xcite . in a particular case , @xmath17 and @xmath16 , both growing and aging behaviors have been observed for a parameter range @xmath271 @xcite . this was motivated by a recent progress on field theory side @xcite@xcite along with some clear experimental realization of the kpz class in one spatial dimension @xcite@xcite . here we obtain additional properties of the two - time response functions as well as to extend our results for general @xmath1 and @xmath0 . before presenting the details , we comment their general behaviors . * due to the simple broken time translation invariance of our system , signified by the parameter @xmath241 , our two - time response function reveals a power - law scaling behavior at early time region , which is distinct from another power - law scaling at late time region . the turning point between the two time regions , @xmath272 , is marked by the waiting time @xmath203 . if @xmath273 , there exists either only growth or aging behavior . * the initial power scaling behaviors , growth or aging , are crucially related to the parameter @xmath274 , especially the combination @xmath275 , which is the scaling dimensions of the dual field theory operators we consider . the late time scaling behavior is further modified by @xmath276 , which is aging parameter , in addition to the scaling dimensions . * the power - law part of the two - time response functions show the growth and aging behaviors , while the log and log@xmath15 corrections provides further modifications that would match detailed data by tuning available parameters . we consider a typical correlation and response functions for general @xmath1 and @xmath0 , extending previous results for @xmath17 and @xmath16 @xcite @xmath277 with a waiting time , @xmath203 , a scaling time , @xmath278 and two other free parameters @xmath279 and @xmath241 , which satisfies the condition ( [ conditiononalpham ] ) coming from the time boundary @xmath280 . the response function ( [ twotimeresponsetlcfty ] ) is the general form for the aging alcf for all the cases considered in [ sec : agingalcf ] . this is also valid for the alcf in [ sec : alcf ] without the condition ( [ conditiononalpham ] ) if we set @xmath281 . . the parameters @xmath0 and @xmath282 do not change the qualitative behaviors , while the parameters @xmath94 , actually @xmath274 , and @xmath241 are important for the early time and late time power law scaling . , title="fig:",scaledwidth=47.0% ] . the parameters @xmath0 and @xmath282 do not change the qualitative behaviors , while the parameters @xmath94 , actually @xmath274 , and @xmath241 are important for the early time and late time power law scaling . , title="fig:",scaledwidth=47.0% ] let us comment for the case with positive @xmath94 . this case has only aging properties if the parameters @xmath94 and @xmath241 are not too large . for @xmath283 and @xmath284 , which is allowed by the time boundary condition ( [ conditiononalpham ] ) , the bending point , around @xmath285 in the figure [ fig : positivenus ] , sits deep down and the second leg of the plot becomes horizontal . as we increase either @xmath94 or @xmath241 , the bending point goes up . this is depicted in the figure [ fig : positivenus ] . for @xmath286 or a particular value of @xmath274 , we can get a straight line , which is identical to the time independent case . if one is interested in growth phenomena , it is more interesting to consider @xmath287 . due to the form of the response function ( [ twotimeresponsetlcfty ] ) , the part @xmath288 determines the properties at early time @xmath289 . for @xmath290 , actually @xmath291 determines the slope at early time . for @xmath292 , the slope of the first leg is negative , while that is positive for @xmath293 . this can be verified directly in the figure [ fig : negativenus ] . . each panel has a fixed value of @xmath274 , which determines the slope of the first leg at early time , while the sign of @xmath241 determines the relative slope of the second leg at late time , compared to the first leg . left : the case saturated with the bf bound @xmath294 and @xmath295 . three plots are for @xmath296 , which are blue straight , black dotted and red dashed lines , respectively . similarly for the middle with @xmath297 and right plots with @xmath298 with the same values of @xmath296 . , title="fig:",scaledwidth=49.0% ] . each panel has a fixed value of @xmath274 , which determines the slope of the first leg at early time , while the sign of @xmath241 determines the relative slope of the second leg at late time , compared to the first leg . left : the case saturated with the bf bound @xmath294 and @xmath295 . three plots are for @xmath296 , which are blue straight , black dotted and red dashed lines , respectively . similarly for the middle with @xmath297 and right plots with @xmath298 with the same values of @xmath296 . , title="fig:",scaledwidth=49.0% ] + . each panel has a fixed value of @xmath274 , which determines the slope of the first leg at early time , while the sign of @xmath241 determines the relative slope of the second leg at late time , compared to the first leg . left : the case saturated with the bf bound @xmath294 and @xmath295 . three plots are for @xmath296 , which are blue straight , black dotted and red dashed lines , respectively . similarly for the middle with @xmath297 and right plots with @xmath298 with the same values of @xmath296 . , title="fig:",scaledwidth=49.0% ] on the other hand , the late time behavior is determined by a factor @xmath299 . if @xmath300 , the slope does not change . the relative slope of the second leg is determined by the sign of @xmath241 . this is verified in the figure [ fig : negativenus ] . the real slope of the second leg is governed by the sign of @xmath301 . in particular , early time growth and late time aging happens for @xmath302 the second condition is similar to our time boundary condition ( [ conditiononalpham ] ) , but not identical . for growth phenomena , the roughness of interfaces is quantified by their inter - facial width , @xmath303 ^ 2 \rangle_l } $ ] , defined as the standard deviation of the interface height @xmath304 over a length scale @xmath305 at time @xmath306 @xcite . an equivalent way to describe the roughness is the height - difference correlation function @xmath307 ^ 2 \rangle $ ] . @xmath308 and @xmath309 denote the average over a segment of length @xmath310 and all over the interface and ensembles , respectively . both @xmath311 and @xmath312 are common quantities for characterizing the roughness , for which the so - called family - vicsek scaling @xcite is expected to hold . the dynamical scaling property is @xmath313 with two characteristic exponents : the roughness exponent @xmath314 and the growth exponent @xmath315 . the dynamical exponent is given by @xmath316 , and the cross over length scale is @xmath317 . for an infinite system , the correlation function behaves as @xmath318 at some late time region @xmath319 . from the two - time correlation function in equation ( [ twotimeresponsetlcfty ] ) , we can get the growth exponent @xmath320 where @xmath321 . note that the parameters satisfy the condition ( [ conditiononalpham ] ) from the time boundary conditions . we notice that our system size is infinite , and thus it is not simple matter to obtain the corresponding roughness exponent . the dynamical exponent is not fixed in alcf , even though the differential equation has @xmath1 dependence , which can be checked in ( [ blukscalareqagingalcf ] ) . for kpz universality class , there is a nontrivial scaling relation between the roughness exponent @xmath322 and the dynamical exponent @xmath1 , chapter 6 in @xcite @xmath323 while this relation is remained to be checked in our holographic model , we assume it is valid to make contact with some field theoretical models . using the relation @xmath324 , we get @xmath325 let us examine these critical exponents against the known case for @xmath11 @xmath326 these can be reproduced with the condition @xmath327 which can be matched for @xmath328 and @xmath329 for negative @xmath94 . we choose @xmath329 for the simple growth behavior . there are two independent critical exponents . one particular interesting exponent is the so called growth exponent @xmath330 , where @xmath321 . we consider a dual field theory operator with @xmath331 , @xmath332 . then by expanding for small @xmath333 , we get @xmath334 using again @xmath335 and the relation @xmath324 , we get @xmath336 if we further restrict our attention to the case @xmath337 for considering only the growth phenomena , we have the following dependence on the number of spatial dimensions @xmath338 where @xmath339 . for @xmath340 , these exponents match ( [ kpzexponents ] ) for @xmath11 . the corresponding roughness exponent is depicted as blue line in the left panel of the figure [ fig : roughnessexponents ] , which is referred as `` alcf '' . to compare with other growth models @xcite , we also depicted the roughness exponents of the kim - kosterlitz model @xmath341 @xcite as well as wolf - kertsz model @xmath342 @xcite . of alcf for @xmath343 and @xmath344 that matches kpz exponents for @xmath11 given in ( [ kpzexponents ] ) . kk represents @xmath341 from kim - kosterlitz @xcite , while wk @xmath342 from wolf - kertsz @xcite . right panel : plot for @xmath322 of alcf@xmath345 for @xmath343 and @xmath346 that matches ( [ kkexponents ] ) only when @xmath347 . , title="fig:",scaledwidth=48.0% ] of alcf for @xmath343 and @xmath344 that matches kpz exponents for @xmath11 given in ( [ kpzexponents ] ) . kk represents @xmath341 from kim - kosterlitz @xcite , while wk @xmath342 from wolf - kertsz @xcite . right panel : plot for @xmath322 of alcf@xmath345 for @xmath343 and @xmath346 that matches ( [ kkexponents ] ) only when @xmath347 . , title="fig:",scaledwidth=48.0% ] for @xmath348 , we get @xmath349 our results ( [ kkexponents ] ) are only valid for @xmath347 , which is referred as `` alcf@xmath345 '' in the right panel of the figure [ fig : roughnessexponents ] . these exponents have been conjectured for growth in a restricted solid - on - solid model by kim - kosterlitz @xcite@xcite . we have shown in previous sections that the response functions reveal growth and aging behaviors without log or log@xmath15 corrections . the log and log@xmath15 corrections have been considered to match further details at early time region @xcite@xcite . in this section , we would like to investigate some more details related to those corrections based on previous results @xcite . our two - time response functions with log correction are given by @xmath350 \bigg\ } \ ; , \end{aligned}\ ] ] where @xmath203 , @xmath351 , @xmath94 and @xmath241 are free parameters , while the coefficients are given by @xmath352 } \;.\end{aligned}\ ] ] we note that the coefficients are completely fixed by two fixed parameters @xmath353 note that similar result for @xmath17 and @xmath16 has been available in @xcite . the detailed comparisons between ( [ twotimeresponsetlcftylog ] ) and the phenomenological field theory model @xcite@xcite were investigated . we noted that the terms proportional to @xmath354 and @xmath355 are not considered in @xcite@xcite , which do not modify qualitative features of the response functions . for this case , the analysis done in @xcite is still valid . we obtain the log@xmath15 extension of the response function using holographic approach @xmath356 \ln [ 1 - \frac{1}{y } ] + \tilde r_2 \ln [ 1 - \frac{1}{y}]^2 \bigg\ } \ ; , \end{aligned}\ ] ] where @xmath357}{\tilde a_0 + \tilde a_1 \ln [ \mathcal m_b ] + \tilde a_2 \ln [ \mathcal m_b]^2 } \ ; , \\ & \tilde r_2 = \frac{\tilde a_2}{\tilde a_0 + \tilde a_1 \ln [ \mathcal m_b ] + \tilde a_2 \ln [ \mathcal m_b]^2 } \ ; , \\ & \tilde a_0 = 1 + \nu \psi - \nu^2 ( \psi^2- \psi ' ) \ ; , \\ & \tilde a_1 =( 2\nu^2 \psi - \nu ) \ ; , \qquad \tilde a_2 = - \nu^2 \;.\end{aligned}\ ] ] these coefficients are also completely fixed by two fixed parameters given in ( [ mbpsiparameters ] ) . compared to @xmath358 , a new parameter @xmath359 determines the behaviors of response functions related to the log and log@xmath15 contributions . as we explicitly check in the figure [ fig : lognegativenus ] , the qualitative behavior of the response functions does not change with the log and log@xmath15 contributions for reasonably small @xmath359 . the growth exponent and aging properties are determined by the two parameters @xmath274 and @xmath241 , which define our theory . with blue straight , @xmath360 with red dashed and @xmath361 with black dot - dashed lines for @xmath243 , @xmath11 , @xmath362 and @xmath363 . for the response functions with log corrections , we need one more input @xmath359 , which we took @xmath364 for these plots . the smaller the value of @xmath359 , the smaller the differences between @xmath365 and its log extensions . , title="fig:",scaledwidth=47.0% ] with blue straight , @xmath360 with red dashed and @xmath361 with black dot - dashed lines for @xmath243 , @xmath11 , @xmath362 and @xmath363 . for the response functions with log corrections , we need one more input @xmath359 , which we took @xmath364 for these plots . the smaller the value of @xmath359 , the smaller the differences between @xmath365 and its log extensions . , title="fig:",scaledwidth=47.0% ] with black dot - dashed and @xmath366 with brown straight lines for @xmath243 , @xmath11 , @xmath362 and @xmath363 . left : @xmath364 , right : @xmath367 . we check that there is no qualitative changes due to the unwanted terms explained in ( [ twotimeresponsetlcftylog2wanted ] ) . , title="fig:",scaledwidth=47.0% ] with black dot - dashed and @xmath366 with brown straight lines for @xmath243 , @xmath11 , @xmath362 and @xmath363 . left : @xmath364 , right : @xmath367 . we check that there is no qualitative changes due to the unwanted terms explained in ( [ twotimeresponsetlcftylog2wanted ] ) . , title="fig:",scaledwidth=47.0% ] we would like to compare our results ( [ twotimeresponsetlcftylog2 ] ) with the following equation , which is equation ( 10 ) of @xcite ( or equation ( 4.3 ) of @xcite ) , obtained from the phenomenological field theory model . @xmath368-\frac{1}{2 } f_{0}\tilde \xi'^2\ln^2 [ 1-\frac{1}{y}]-g_{21,0}\xi'\ln[y-1]+\frac{1}{2}f_{0}\xi'^2\ln^2 [ y-1]\right ) \ ; , \nonumber \end{aligned}\ ] ] where the parameters @xmath369 come from the logarithmic extension in the field theory side . from the phenomenological input @xcite `` the parenthesis becomes essentially constant for sufficiently large @xmath370 , '' the condition @xmath371 is imposed to remove the last two terms . the other parameters in ( [ ftresponsefunction ] ) are determined to match available data . we can identify the exponents @xmath372 by comparing the equation to our result @xcite @xmath373 by including the log@xmath15 corrections , we provide the necessary terms @xmath374 ^ 2 $ ] as well as @xmath375 $ ] . the relative coefficients between them are fixed by ( [ log2coefficients ] ) . on the other hand , there exist also several unwanted terms inside the parenthesis of ( [ twotimeresponsetlcftylog2 ] ) . first , we note new terms @xmath376 and @xmath377 $ ] due to log@xmath15 corrections , in addition to @xmath378 coming from the log correction . all these terms might spoil the desired properties of the phenomenological response function ( [ ftresponsefunction ] ) . to examine the effects coming from these unwanted terms , we plot the response function with only wanted terms as @xmath379 + \tilde r_2 \ln [ 1 - \frac{1}{y}]^2 \bigg\ } \;. \end{aligned}\ ] ] we explicitly check in the figure [ fig : logunwantedterms ] that the full response functions ( [ twotimeresponsetlcftylog2 ] ) have a qualitatively similar behavior compared to those ( [ twotimeresponsetlcftylog2wanted ] ) with only wanted terms in the field theory approach . while the dynamical exponent @xmath1 for the alcf is not fixed in obtaining the correlation and response functions , those of the schrdinger backgrounds crucially depend on @xmath1 . in fact , obtaining analytic solutions of the differential equation ( [ blukscalareqschr ] ) is a highly non - trivial task . fortunately , we are able to get response functions for @xmath164 with some approximations as @xmath380 where @xmath203 , @xmath351 , @xmath94 and @xmath241 are free parameters . this is valid for the aging background ( [ agingschrcorrf1 ] ) as well as the schrdinger background , @xmath281 , given by ( [ schrcorrf1 ] ) with @xmath381 for @xmath191 and by ( [ schrcorrlargemf1 ] ) with @xmath382 for @xmath383 . from the two - time response functions in equation ( [ twotimeresponsetlcfty ] ) , we can get the growth exponent @xmath384 where @xmath321 . now the dynamical exponent is fixed as @xmath164 for the aging background . using the relation @xmath324 , we get @xmath385 the critical exponents for the kpz universality class given in ( [ kpzexponents ] ) can be reproduced with the condition @xmath386 which can be matched for @xmath387 and @xmath388 for negative @xmath94 . the value of @xmath274 for @xmath382 becomes negative , yet is allowed as we can see from the expression of @xmath94 . we have extended our geometric realizations of aging symmetry in several different ways based on previous works for @xmath17 and @xmath16 @xcite@xcite . first , we generalize our correlation and response functions to the non - conformal setup with general dynamical exponent @xmath1 and for arbitrary spatial dimensions @xmath0 . they have galilean symmetries with time translation symmetry , which are summarized in the equations ( [ f1correlator ] ) , ( [ schrcorrf1 ] ) and ( [ schrcorrlargemf1 ] ) . for convenience , we reproduce equation ( [ f1correlator ] ) here @xmath389 which is valid for general @xmath1 and @xmath0 . second , these are extended with log and log@xmath15 corrections with appropriate bulk actions . practically , these corrections can be computed using simple properties of the differential operators ( [ diffopalcf ] ) , ( [ diffopschr ] ) and their commutation relations ( [ commutationrelation ] ) . the results are listed in ( [ f2correlator ] ) , ( [ f3correlator ] ) for alcf , ( [ schrcorrf2 ] ) , ( [ schrcorrf3 ] ) for schrdinger backgrounds . all these response functions have time translation invariance . third , on top of these extensions , we also compute response functions with aging symmetry , by breaking the global time translation invariance using a singular coordinate transformation ( [ coordinatechangez ] ) . we check the general relation between the aging response functions and those of schrdinger backgrounds holds ( [ generalformulalogaging1 ] ) @xmath390 this generalization is independent of the logarithmic extensions . with these results , we investigate our two - time response functions for general @xmath1 and especially for arbitrary number of spatial dimensions @xmath0 with log and log@xmath15 extensions ( [ twotimeresponsetlcftylog2 ] ) @xmath391 where @xmath392 represent various contributions from the log and log@xmath15 extensions . from the systematic analysis , we have found that our two - time response functions reveal a power - law scaling behavior at early time region , which is distinct from another power - law scaling at late time region . this can be explicitly checked in the figure [ fig : negativenus ] . the early time power scaling behaviors are governed by the scaling dimensions of the dual field theory operators . in particular , their growth and aging is determined by the sign of the parameter @xmath274 . the late time behaviors are modified by @xmath276 , the aging parameter . if @xmath286 , the initial behaviors persist without change , which is expected due to its time translation invariance . the turning point between these two time regions is marked by the waiting time @xmath203 . the log and log@xmath15 corrections provide further modifications that would match detailed data by turning available parameters . let us conclude with some observations and future directions toward holographic realizations of kpz universality class . our generalizations of the holographic response functions to general @xmath1 and @xmath0 open up some possibilities to have contact with the higher dimensional growth and aging phenomena . we have done the first attempt to do so in [ sec : alcfcriticalexponent ] . we make some contacts with kim - kosterlitz model @xcite at higher spatial dimensions with an assumptions ( [ azrelationkpz ] ) for some particular dual scalar operators . although it is not perfect , we consider this as a promising sign for the future developments along the line . we mention two pressing questions we would like to answer in a near future . our holographic model is an infinite system , and thus obtaining the roughness exponent `` @xmath322 '' is rather challenging . progresses on this point will provide a big step toward realizing holographic kpz class . assumption ( [ azrelationkpz ] ) is well understood in the field theoretical models @xcite . there galilean invariance was a crucial ingredient , which is also important in our holographic model . verifying this relation would be an important future challenge we thank to e. kiritsis , y. s. myung , v. niarchos for discussions and valuable comments on the higher dimensional kpz class and the logarithmic extensions of cft . sh is supported in part by the national research foundation of korea(nrf ) grant funded by the korea government(mest ) with the grant number 2012046278 . sh and jj are supported by the national research foundation of korea ( nrf ) grant funded by the korea government(mest ) through the center for quantum spacetime(cquest ) of sogang university with grant number 2005 - 0049409 . jj is supported in part by the national research foundation of korea(nrf ) grant funded by the korea government(mest ) with the grant number 2010 - 0008359 . bsk is grateful to the members of the crete center for theoretical physics , especially e. kiritsis , for his warm hospitality during his visit . bsk is supported in part by the israel science foundation ( grant number 1468/06 ) . k. takeuchi , m. sano , _ `` universal fluctuations of growing interfaces : evidence in turbulent liquid crystals , '' _ phys . lett . * 104 * , 230601 ( 2010 ) . . k. takeuchi , m. sano , t. sasamoto and h. spohn , _ `` growing interfaces uncover universal fluctuations behind scale invariance , '' _ sci . * 1 * , 34 . m. henkel , j. d. noh and m. pleimling , _ `` phenomenology of ageing in the kardar - parisi - zhang equation , '' _ phys . rev . * e 85 * , 030102(r ) ( 2012 ) . m. henkel , _ `` on logarithmic extensions of local scale - invariance , '' _ nucl . b * 869 * , 282 ( 2013 ) . s. hyun , j. jeong and b. s. kim , _ `` aging logarithmic conformal field theory : a holographic view , '' _ jhep * 1301 * , 141 ( 2013 ) . j. i. jottar , r. g. leigh , d. minic and l. a. pando zayas , _ `` aging and holography , '' _ jhep * 1011 * , 034 ( 2010 ) . s. hyun , j. jeong and b. s. kim , _ `` finite temperature aging holography , '' _ jhep * 1203 * , 010 ( 2012 ) . j. m. maldacena , _ `` the large n limit of superconformal field theories and supergravity , '' _ adv . * 2 * , 231 ( 1998 ) [ int . j. theor . phys . * 38 * , 1113 ( 1999 ) ] . o. aharony , s. s. gubser , j. m. maldacena , h. ooguri and y. oz , _ `` large n field theories , string theory and gravity , '' _ phys . rept . * 323 * , 183 ( 2000 ) . m. flohr , _ `` bits and pieces in logarithmic conformal field theory , '' _ int . j. mod . a * 18 * , 4497 ( 2003 ) . m. r. gaberdiel , _ `` an algebraic approach to logarithmic conformal field theory , '' _ int . j. mod . a * 18 * , 4593 ( 2003 ) . a. m. ghezelbash , m. khorrami and a. aghamohammadi , _ `` logarithmic conformal field theories and ads correspondence , '' _ int . j. mod . a * 14 * , 2581 ( 1999 ) . i. i. kogan , _ `` singletons and logarithmic cft in ads / cft correspondence , '' _ phys . b * 458 * , 66 ( 1999 ) . y. s. myung and h. w. lee , _ `` gauge bosons and the ads(3 ) / lcft(2 ) correspondence , '' _ jhep * 9910 * , 009 ( 1999 ) . d. grumiller and n. johansson , _ `` instability in cosmological topologically massive gravity at the chiral point , '' _ jhep * 0807 * , 134 ( 2008 ) ; s. ertl , d. grumiller and n. johansson , _ `` erratum to ` instability in cosmological topologically massive gravity at the chiral point ' , arxiv:0805.2610 , '' _ . k. skenderis , m. taylor and b. c. van rees , _ `` topologically massive gravity and the ads / cft correspondence , '' _ jhep * 0909 * , 045 ( 2009 ) . d. grumiller and i. sachs , _ `` ads ( 3 ) / lcft ( 2 ) @xmath393 correlators in cosmological topologically massive gravity , '' _ jhep * 1003 * , 012 ( 2010 ) . d. grumiller and o. hohm , _ `` ads(3)/lcft(2 ) : correlators in new massive gravity , '' _ phys . lett . b * 686 * , 264 ( 2010 ) . m. alishahiha and a. naseh , _ `` holographic renormalization of new massive gravity , '' _ phys . d * 82 * , 104043 ( 2010 ) . j. maldacena , _ `` einstein gravity from conformal gravity , '' _ . s. hyun , w. jang , j. jeong and s. -h . yi , _ `` noncritical einstein - weyl gravity and the ads / cft correspondence , '' _ jhep * 1201 * , 054 ( 2012 ) . s. hyun , w. jang , j. jeong and s. -h . `` on classical equivalence between noncritical and einstein gravity : the ads / cft perspectives , '' _ jhep * 1204 * , 030 ( 2012 ) . h. lu and c. n. pope , _ `` critical gravity in four dimensions , '' _ phys . lett . * 106 * , 181302 ( 2011 ) . e. a. bergshoeff , s. de haan , w. merbis and j. rosseel , _ `` a non - relativistic logarithmic conformal field theory from a holographic point of view , '' _ jhep * 1109 * , 038 ( 2011 ) . e. a. bergshoeff , s. de haan , w. merbis , j. rosseel and t. zojer , _ `` on three - dimensional tricritical gravity , '' _ . t. moon and y. s. myung , _ `` rank-3 finite temperature logarithmic conformal field theory , '' _ phys . d * 86 * , 084058 ( 2012 ) . d. grumiller , w. riedler , j. rosseel and t. zojer , _ `` holographic applications of logarithmic conformal field theories , '' _ . w. d. goldberger , _ `` ads / cft duality for non - relativistic field theory , '' _ jhep * 0903 * , 069 ( 2009 ) . j. l. f. barbon and c. a. fuertes , _ `` on the spectrum of nonrelativistic ads / cft , '' _ jhep * 0809 * , 030 ( 2008 ) . d. t. son , _ `` toward an ads / cold atoms correspondence : a geometric realization of the schroedinger symmetry , '' _ phys . d * 78 * , 046003 ( 2008 ) . k. balasubramanian and j. mcgreevy , _ `` gravity duals for non - relativistic cfts , '' _ phys . * 101 * , 061601 ( 2008 ) . b. s. kim , _ `` schrdinger holography with and without hyperscaling violation , '' _ jhep * 1206 * , 116 ( 2012 ) . b. s. kim , _ `` hyperscaling violation : a unified frame for effective holographic theories , '' _ jhep * 1211 * , 061 ( 2012 ) . j. maldacena , d. martelli and y. tachikawa , _ `` comments on string theory backgrounds with non - relativistic conformal symmetry , '' _ jhep * 0810 * , 072 ( 2008 ) . b. s. kim and d. yamada , _ `` properties of schroedinger black holes from ads space , '' _ jhep * 1107 * , 120 ( 2011 ) . b. s. kim , e. kiritsis and c. panagopoulos , _ `` holographic quantum criticality and strange metal transport , '' _ new j. phys . * 14 * , 043045 ( 2012 ) . d. t. son and a. o. starinets , _ `` minkowski - space correlators in ads / cft correspondence : recipe and applications , '' _ jhep * 0209 * , 042 ( 2002 ) . c. p. herzog , m. rangamani and s. f. ross , _ `` heating up galilean holography , '' _ jhep * 0811 * , 080 ( 2008 ) . a. adams , k. balasubramanian and j. mcgreevy , _ `` hot spacetimes for cold atoms , '' _ jhep * 0811 * , 059 ( 2008 ) . d. yamada , _ `` thermodynamics of black holes in schroedinger space , '' _ class . * 26 * , 075006 ( 2009 ) . l. mazzucato , y. oz , s. theisen and , _ `` non - relativistic branes , '' _ jhep * 0904 * , 073 ( 2009 ) . m. ammon , c. hoyos , a. obannon and j. m. s. wu , _ `` holographic flavor transport in schrodinger spacetime , '' _ jhep * 1006 * , 012 ( 2010 ) . b. gouteraux and e. kiritsis , _ `` generalized holographic quantum criticality at finite density , '' _ jhep * 1112 * , 036 ( 2011 ) . c. charmousis , b. gouteraux , b. s. kim , e. kiritsis and r. meyer , _ `` effective holographic theories for low - temperature condensed matter systems , '' _ jhep * 1011 * , 151 ( 2010 ) [ . f. w. j. olver , d. w. lozier , r. f. boisvert , and c. w. clark , editors , _ `` nist handbook of mathematical functions . '' _ cambridge university press , new york , ny , 2010 . http://dlmf.nist.gov/5.11[gamma functions ] . m. henkel and m. pleimling , _ `` local scale - invariance in disordered systems , '' _ . n. gray , d. minic and m. pleimling , _ `` on non - equilibrium physics and string theory , '' _ . f. family and t. vicsek , _ `` scaling of the active zone in the eden process on percolation networks and the ballistic deposition model , '' _ j. phys . * a18 * , l75 ( 1985 ) .
we analytically compute correlation and response functions of scalar operators for the systems with galilean and corresponding aging symmetries for general spatial dimensions @xmath0 and dynamical exponent @xmath1 , along with their logarithmic and logarithmic squared extensions , using the gauge / gravity duality . these non - conformal extensions of the aging geometry are marked by two dimensionful parameters , eigenvalue @xmath2 of an internal coordinate and aging parameter @xmath3 . we further perform systematic investigations on two - time response functions for general @xmath0 and @xmath1 , and identify the _ growth exponent _ as a function of the scaling dimensions @xmath4 of the dual field theory operators and aging parameter @xmath3 in our theory . the initial growth exponent is only controlled by @xmath4 , while its late time behavior by @xmath3 as well as @xmath4 . these behaviors are separated by a time scale order of the waiting time . we attempt to make contact our results with some field theoretical growth models , such as kim - kosterlitz model at higher number of spatial dimensions @xmath0 . cquest-2013 - 0593 + taup-2963/13 seungjoon hyun@xmath5 , jaehoon jeong@xmath6 and bom soo kim@xmath7 @xmath8_department of physics , college of science , yonsei university , seoul 120 - 749 , korea _ + @xmath9_center for quantum spacetime , sogang university , seoul 121 - 742 , korea _ + @xmath10_raymond and beverly sackler school of physics and astronomy , _ + _ tel aviv university , 69978 , tel aviv , israel _ + [email protected] , [email protected] , [email protected]
the spectral problem for possible sums of two random hermitian matrices with given spectra had been a long - standing problem , formulated as horn s conjecture , until it was completely solved only recently by klyachko , knutson and tao ( @xcite , @xcite ) . in @xcite , klyachko then observed a close connection between @xmath0-biinvariant random walks on @xmath1 and random walks on the additive group @xmath2 of all hermitian @xmath3-matrices with trace 0 whose transition probabilities are conjugation - invariant under @xmath0 . he used this connection to reduce the description of the possible singular spectra of products of random matrices from @xmath1 with given singular spectra to the spectral problem for sums of hermitian matrices . basically , klyachko s observation is a connection between the convolution algebras of the gelfand pairs @xmath5 and @xmath6 . it is closely related to a similar correspondence between the convolutions of conjugation - invariant measures on a compact lie group ( here @xmath0 ) on one hand and @xmath7-invariant measures on its lie algebra ( here @xmath8 ) on the other , in terms of the so - called wrapping map ( see @xcite and remark [ wrapping ] ) . klyachko noticed in @xcite , but did not not explain that his connection can be well expressed in terms of hypergroups ; his proof goes via random walks in the group @xmath0 and relies on various identities between the spherical functions of @xmath5 , the characters of @xmath0 and the euclidean group @xmath2 , as well as poisson s summation formula . the main purpose of the present paper is to clarify and simplify klyachko s approach @xcite by using so - called deformations of hypergroup convolutions by positive semicharacters as introduced in @xcite , @xcite . our description in particular implies that the banach algebra @xmath9 of @xmath0-biinvariant signed measures on @xmath1 with total variation norm is isometrically isomorphic to a certain banach subalgebra of @xmath10 , whereby probability measures are being prserved . we finally show how this leads to new proofs for ( known ) limit theorems for @xmath0-biinvariant random walks on @xmath1 . the explicit construction of this isomorphism runs as follows : 1 . @xmath0 acts on @xmath2 by conjugation as a group of orthogonal transformations . the space @xmath11 of all orbits can be identified with @xmath12 where @xmath13 represents the ordered eigenvalues of a matrix in @xmath2 . the banach algebra @xmath14 of all @xmath0-invariant bounded measures on @xmath2 can be identified with the banach algebra @xmath15 of the associated orbit hypergroup @xmath16 . the polar decomposition of @xmath17 shows that the double coset space @xmath18 may also be identified with @xmath19 where now for @xmath20 corresponds to the singular spectrum @xmath21 of some @xmath22 . this leads to a canonical banach algebra isomorphism between @xmath9 and the banach algebra @xmath23 of the double coset hypergroup @xmath24 . the characters of the hypergroups @xmath25 and @xmath26 are spherical functions of the corresponding symmetric spaces , and it is well - known that as functions on @xmath27 they only differ by a known factor @xmath28 ( see @xcite ) . we prove that such a connection between the characters of two hypergroup structures on @xmath19 implies that their convolutions are related by a so - called hypergroup deformation . in particular , the supports of the convolution products of two point measures are the same in both hypergroups . this immediately implies the equivalence of the two spectral problems for hermitian and unitary matrices , the main result of klyachko @xcite . the deformation isomorphism , however , is not isometric and not probability preserving . to achieve this , a final correction is needed : 4 . instead of looking at the banach algebra @xmath14 of all @xmath0-invariant measures on @xmath2 , we take a suitable exponential function @xmath290,\infty[$ ] and observe that the norm - closure of @xmath30 is a banach algebra which turns out to be isometrically isomorphic to @xmath23 . putting all steps together one obtains the claimed probability preserving isometric isomorphism . for @xmath31 this construction reflects the known close connection between the so - called naimark hypergroup @xmath32 and the bessel - kingman hypergroup @xmath33 ; see @xcite , @xcite . all results will be derived in the general setting of a complex connected semisimple lie group @xmath34 with finite center and maximal compact subgroup @xmath35 ; in the special situation above , @xmath36 . for further results concerning the associated hypergroup convolutions considered in this paper we refer to @xcite and @xcite . the paper is organized as follows : in section 2 we collect some relevant facts on commutative hypergroups and their deformations . in section 3 we then use this deformation to show how the banach algebras of all @xmath35-biinvariant bounded complex measures on @xmath34 appear as subalgebras of the banach algebra of bounded complex measures on some euclidean space . the final section is devoted to probabilistic applications . in this section we give a quick introduction to hypergroups . we in particular prove a general result on deformations of hypergroup convolutions in proposition [ deform ] which is crucial for step ( 3 ) of our construction . moreover , step ( 4 ) will be explained . we first fix some notations . for a locally compact hausdorff space @xmath37 , @xmath38 denotes the space of all positive radon measures on @xmath37 , and @xmath39 the banach space of all bounded regular complex borel measures with the total variation norm . moreover , @xmath40 is the set of all probability measures , @xmath41 the set of all measures with compact support , and @xmath42 the point measure in @xmath43 . the spaces @xmath44 of continuous functions are defined as usual . for details on the following we refer to @xcite and @xcite . a hypergroup @xmath45 consists of a locally compact hausdorff space @xmath35 and a convolution @xmath46 on @xmath39 such that @xmath47 becomes a banach algebra , where @xmath46 is weakly continuous and probability preserving and preserves compact supports of measures . moreover , there exists an identity @xmath48 with @xmath49 for @xmath43 , as well as a continuous involution @xmath50 on @xmath37 such that for @xmath51 , @xmath52 is equivalent to @xmath53 , and @xmath54 . here for @xmath55 , the measure @xmath56 is given by @xmath57 for borel sets @xmath58 . a hypergroup @xmath45 is called commutative if and only if so is the convolution @xmath46 . hence , for a commutative hypergroup @xmath45 the triple @xmath59 is a commutative banach-@xmath46-algebra with identity @xmath60 . [ ex_1 ] 1 . if @xmath34 is a locally compact group , then @xmath61 is a hypergroup with the usual group convolution @xmath46 . 2 . let @xmath35 be a compact subgroup of a locally compact group @xmath34 . then @xmath62 is a banach-@xmath46-subalgebra of @xmath63 with identity @xmath64 where @xmath65 is the normalized haar measure of @xmath35 embedded into @xmath34 . moreover , the double coset space @xmath66 is a locally compact hausdorff space , and the canonical projection @xmath67 induces a probability preserving , isometric isomorphism @xmath68 of banach spaces by taking images of measures . the transport of the convolution on @xmath69 to @xmath70 via @xmath71 leads to a hypergroup structure @xmath72 with identity @xmath73 and involution @xmath74 , and @xmath71 becomes a probability preserving , isometric isomorphism of banach-@xmath46-algebras . 3 . let @xmath75 be a finite - dimensional euclidean vector space and @xmath76 a compact subgroup of the orthogonal group of @xmath77 acting continuously on @xmath78 . for @xmath79 , denote the image measure of @xmath80 under @xmath81 by @xmath82 then the space of @xmath35-invariant measures @xmath83 is a banach-@xmath46-subalgebra of @xmath84 ( with the group convolution ) , with identity @xmath85 . moreover , the space @xmath86 of all @xmath35-orbits in @xmath78 is locally compact , and the canonical projection @xmath87 induces a probability preserving , isometric isomorphism @xmath88 of banach spaces and an associated so - called orbit hypergroup @xmath89 such that @xmath71 becomes a probability preserving , isometric isomorphism of banach-@xmath46-algebras . the involution on @xmath89 is given by @xmath90 . we next collect some data of a commutative hypergroup @xmath45 . by a result of r. spector , there exists a ( up to normalization ) unique haar measure @xmath91 which is characterized by @xmath92 for all @xmath93 and @xmath94 where we use the notation @xmath95 similar to the dual of a locally compact abelian group , one defines the spaces 1 . @xmath96 ; 2 . @xmath97 @xmath98 the elements of @xmath99 and @xmath100 are called semicharacters and characters , respectively . all spaces above are locally compact hausdorff spaces w.r.t . the topology of compact - uniform convergence . the fourier transform on @xmath101 is defined by @xmath102 similar , the fourier - stieltjes transform of @xmath55 is defined by @xmath103 @xmath104 . both transforms are injective , c.f . @xcite . in the following , we consider different hypergroup convolutions on @xmath37 , and we write @xmath105 etc . in order to specify the relevant convolution . [ ex_2 ] 1 . assume that in the situation of [ ex_1](ii ) , @xmath106 is commutative . then a @xmath35-biinvariant function @xmath107 with @xmath108 is by definition a spherical function of @xmath109 if @xmath110 multiplicative functions @xmath111 are in one - to - one correspondence with spherical functions on @xmath34 via @xmath112 for the projection @xmath67 . in this way , the fourier(-stieltjes ) transform on @xmath106 corresponds to the spherical fourier(-stieltjes ) transform . 2 . in the situation of example [ ex_1](iii ) , the functions @xmath113 are continuous multiplicative functions of the orbit hypergroup @xmath114 for @xmath115 , the complexification of @xmath78 , where @xmath116 holds if and only if @xmath117 . it is also well - known ( see @xcite ) that @xmath118 . in @xcite , positive semicharacters were used to construct deformed hypergroup convolutions . more precisely , the following was proven there : let @xmath119 be a positive semicharacter on the commutative hypergroup @xmath45 , i.e. , @xmath120 for @xmath43 . then the convolution @xmath121 extends uniquely to a bilinear , associative , probability preserving , and weakly continuous convolution @xmath122 on @xmath39 . moreover , @xmath123 becomes a commutative hypergroup with the identity and involution of @xmath45 . for @xmath124 , one has @xmath125 note that by eq.([deformconvo ] ) , the mapping @xmath126 establishes a canonical algebra isomorphism between @xmath127 and @xmath128 which usually when @xmath129 is unbounded can not be extended to @xmath39 . the hypergroup @xmath123 is called the deformation of @xmath45 w.r.t . @xmath129 . clearly , many data of @xmath123 can be expressed in terms of @xmath129 and corresponding data of @xmath45 . in the above setting , we have=-1pt 1 . @xmath130 is a haar measure of @xmath123 . the mapping @xmath131 is a homeomorphism ( w.r.t . the compact - uniform topology ) between @xmath132 and @xmath133 , and also between @xmath134 and @xmath135 . for ( i ) and the first part of ( ii ) see @xcite ; the second part of ( ii ) is analogous . we next turn to the following converse statement ; it will be crucial for this paper : [ deform ] let @xmath45 and @xmath123 be commutative hypergroups on @xmath37 . assume there is a positive semicharacter @xmath129 of @xmath45 such that the spaces of multiplicative continuous functions for @xmath45 and @xmath123 are related via @xmath136 then @xmath123 is the deformation of @xmath45 w.r.t . @xmath129 . let @xmath137 denote the deformation of @xmath45 via @xmath129 . . then by our assumption and the proposition above , @xmath139 is multiplicative w.r.t . @xmath140 as well , and the fourier - stieltjes transforms of @xmath141 and @xmath142 w.r.t . @xmath123 satisfy @xmath143 by the injectivity of the fourier - stieltjes - transform on @xmath39 , we obtain @xmath144 . thus the convolutions of @xmath123 and @xmath137 coincide , and so do the involutions , because they are uniquely determined by the convolutions . it is well - known ( see @xcite,@xcite,@xcite,@xcite ) that the double coset hypergroup @xmath32 may be realized as hypergroup @xmath145 with multiplicative functions @xmath146 via the correspondence of @xmath147 with @xmath148 , we have @xmath149 . on the other hand , the orbit hypergroup @xmath150 may be realized as the bessel - kingman hypergroup @xmath151 with multiplicative functions @xmath152 see @xcite , @xcite . clearly , @xmath153 is a positive semicharacter on @xmath45 abd @xmath154 for @xmath155 . proposition [ deform ] implies the known fact that @xmath123 is a deformation of @xmath45 , c.f . @xcite . we next consider further examples which explain step ( 4 ) in the introduction ; they are similar to a construction in @xcite . let @xmath156 be a euclidean vector space of finite dimension @xmath157 , @xmath76 a compact subgroup of the orthogonal group of @xmath78 , and @xmath89 the associated orbit hypergroup . fix @xmath158 with @xmath159 , and consider the exponential @xmath160 on @xmath78 . let further @xmath161 the multiplicativity of @xmath162 yields that @xmath163 , where @xmath46 denotes the group convolution on @xmath78 . hence @xmath164 is a subalgebra of the banach-@xmath46-algebra @xmath84 , and its norm - closure @xmath165 is a banach subalgebra of @xmath84 . notice that for @xmath166 this algebra is not closed under the involution on @xmath84 ; for instance , the @xmath157-dimensional normal distribution @xmath167 with density @xmath168 is contained in @xmath169 while this is not the case for @xmath170 . nevertheless , we prove that @xmath169 is isometrically isomorphic as a banach algebra to the banach algebra of measures of a suitable deformation of the orbit hypergroup @xmath89 . more precisely : [ euclidean - deform ] let @xmath158 with @xmath159 and define @xmath171 then the following hold : = -1pt 1 . @xmath129 is a positive semicharacter on @xmath172 2 . if @xmath173 is the deformation of @xmath89 w.r.t . @xmath129 , then the canonical projection @xmath174 induces ( by taking image measures ) a probability preserving isometric isomorphism of banach algebras from @xmath169 onto @xmath175 . 1 . @xmath129 is obviously continuous and positive . for @xmath176 , we further have @xmath177 moreover , by the condition on @xmath178 above and the properties of the haar measure of @xmath35 , @xmath179 therefore , @xmath129 is a positive semicharacter on @xmath89 . 2 . consider the diagram + m_c^0,k(v ) & ^ e _ & m_c^,k(v ) + & & + ( m_c(v^k ) , * ) & ^ _ 0 & ( m_c(v^k ) , ) + where @xmath180 always stands for a space of measures with compact support , and the vertical mappings are obtained by taking image measures w.r.t . the mapping @xmath71 restricted to @xmath181 and @xmath164 is probability preserving and isometric . moreover , for @xmath182 and @xmath183 we obtain by the @xmath35-invariance of @xmath80 and the definition of @xmath129 , @xmath184 + which proves that the diagram commutes . therefore , as both horizontal mappings and the left vertical mapping are algebra isomorphisms , the right vertical mapping is also a probability preserving isometric isomorphism of banach algebras from @xmath185 onto @xmath186 . a continuity and density argument shows that the banach algebras @xmath169 and @xmath175 are isomorphic via the probability preserving mapping @xmath71 . we here identify the banach algebra of all bounded borel measures on a connected semisimple noncompact lie group with finite center , which are biinvariant under a maximal compact subgroup , as a banach algebra of measures in some euclidean setting . for the general background we refer to @xcite . let @xmath34 be a complex , noncompact connected semisimple lie group with finite center and @xmath35 a maximal compact subgroup of @xmath34 . consider the corresponding cartan decomposition @xmath187 of the lie algebra of @xmath34 , and choose a maximal abelian subalgebra @xmath188 . @xmath35 acts on @xmath189 via the adjoint representation as a group of orthogonal transformations with respect to the killing form as scalar product . let further @xmath190 be the weyl group of @xmath35 , which acts on @xmath191 as a finite reflection group , with root system @xmath192 . here and lateron , @xmath191 is always identified with its dual @xmath193 via the killing form , which we denote by @xmath194 . we fix some weyl chamber @xmath195 in @xmath191 and denote the associated system of positive roots by @xmath196 . the closed chamber @xmath197 is a fundamental domain for the action of @xmath190 on @xmath191 . later on we shall need the half sum of roots , @xmath198 we now identify @xmath19 with the orbit hypergroup @xmath199 where each @xmath35-orbit in @xmath200 corresponds to its unique representative in @xmath201 . then in view of example [ ex_2](ii ) above and prop . iv.4.8 of @xcite , the multiplicative continuous functions of @xmath26 , considered as @xmath35-invariant functions on @xmath189 , are given by @xmath202 where @xmath148 runs through the complexification @xmath203 of @xmath191 . moreover , @xmath204 iff @xmath148 and @xmath80 are in the same @xmath190-orbit . this is a special case of harish - chandra s integral formula for the spherical functions of a cartan motion group . according to theorem ii.5.35 and cor . ii.5.36 of @xcite , they can also be written as @xmath205 with the fundamental alternating polynomial @xmath206 on the other hand , @xmath19 can be identified with @xmath106 where @xmath207 corresponds to the double coset @xmath208 . according to this identification and the explicit formula for the spherical functions in theorem iv.5.7 of @xcite , the multiplicative continuous functions on the commutative double coset hypergroup @xmath209 are ( as functions on @xmath191 ) given by @xmath210 with @xmath211 . thus in particular , @xmath212 notice that @xmath213 is a positive semicharacter of @xmath26 . by weyl s formula ( @xcite , prop . i.5.15 . ) , @xmath214 the square of @xmath213 ( called @xmath215 in @xcite ) determines the ratio of the volume elements in @xmath189 and @xmath216 . proposition [ deform ] and eq . ( [ zshg ] ) show that @xmath209 is the deformation of the orbit hypergroup @xmath217 via @xmath213 . moreover , we have @xmath218 with the half sum @xmath219 and @xmath220 . as the condition @xmath159 is satisfied , proposition [ euclidean - deform ] further implies that the banach algebra of measures of @xmath209 can be identified with @xmath221 , which is the closure of @xmath222 the claimed isomorphism between @xmath221 and @xmath223 is now given by taking image measures w.r.t . the canonical projection @xmath224 . [ wrapping ] the algebra isomorphism @xmath225 with the semicharacter @xmath226 is closely related to the so - called wrapping map for the compact lie group @xmath35 , see @xcite . in fact , as @xmath34 is complex , we have @xmath227 in the cartan decomposition @xmath228 . thus @xmath229 ( on which @xmath35 acts by conjugation ) is the dual symmetric space of @xmath216 . the wrapping map @xmath230 is defined by @xmath231 where @xmath232 and @xmath233 is the @xmath35-invariant extension of @xmath234 notice that @xmath235 . as shown in @xcite , @xmath236 is an algebra homomorphism from @xmath35-invariant measures in @xmath237 to conjugation - invariant measures in @xmath238 . the proof thereof is based on weyl s integration formula and kirillov s character formula for compact groups . in contrast to our situation , @xmath236 does not preserve positivity and can therefore not be associated with a hypergroup deformation . the group @xmath239 is a maximal compact subgroup of the connected semisimple lie group @xmath240 . in the cartan decomposition @xmath241 we obtain @xmath189 as the additive group @xmath2 of all hermitian @xmath3-matrices with trace @xmath4 , on which @xmath0 acts by conjugation . moreover , @xmath191 consists of all real diagonal matrices with trace 0 and will be identified with @xmath242 on which the weyl group acts as the symmetric group @xmath243 as usual . we thus may take @xmath244 this set parametrizes the possible spectra of matrices from @xmath2 . a system of positive roots corresponding to @xmath19 is @xmath245 where @xmath246 denotes the standard basis of @xmath247 . the root system is of type @xmath248 . in order to describe the probability preserving isometric isomorphism for this example explicitly , we realize the canonical projections @xmath249 and @xmath250 as follows : for @xmath251 , define @xmath252 as the tuple of eigenvalues of @xmath253 , ordered by size . for @xmath254 define @xmath255 as the element @xmath207 such that the singular spectrum @xmath256 of @xmath257 is @xmath258 . we have @xmath259 which implies that the banach algebra @xmath221 above may be described as the closure of @xmath260 with @xmath261\bigr).\ ] ] as discussed above , the mappings @xmath262 induce ( by taking images of measures ) probability preserving isometric isomorphisms from the banach algebras @xmath221 and @xmath263 onto @xmath264 for the double coset convolution @xmath122 on @xmath265 . let us finally come back to the two spectral problems studied by klyachko @xcite . the hypergroup deformation between the two convolutions @xmath46 and @xmath122 on @xmath19 gives the natural explanation for the close connection of these problems : first , for fixed @xmath266 the probability measure @xmath267 is the distribution of possible spectra of sums @xmath268 where the @xmath269 run through all matrices from @xmath270 with @xmath271 ( @xmath272 ) . on the other hand , @xmath273 describes the distribution of possible singular spectra @xmath274 of products @xmath275 where the @xmath276 run through all matrices from @xmath277 with given singular spectra @xmath278 . as @xmath279 we obtain : ( c.f . @xcite , theorem b ) for elements @xmath280 the following are equivalent:=-2pt 1 . there exist matrices @xmath281 with given spectra @xmath282 such that @xmath283 . 2 . there exist matrices @xmath284 with given singular spectra @xmath285 such that @xmath286 . we now briefly discuss the further classical series of complex , noncompact connected simple lie groups with finite center ( c.f . appendix c in @xcite . ) 1 . * the @xmath287-case . * for @xmath288 consider @xmath289 with the maximal compact subgroup @xmath290 . in this case @xmath191 may be identified with @xmath291 with standard basis @xmath292 , and we may choose @xmath293 and @xmath294 . the weyl group @xmath190 is isomorphic to the semidirect product @xmath295 , and @xmath296 2 . * the @xmath297-case . * for @xmath298 consider @xmath299 with the maximal compact subgroup @xmath300 . in this case , again @xmath301 with @xmath19 and @xmath190 as in the @xmath287-case . a positive root system is @xmath302 and @xmath303 . comparing this with the @xmath287-case , we see from eq . ( [ euclidean ] ) that the spherical functions @xmath304 of the orbit hypergroups @xmath305 are the same in both cases . the preceding results on hypergroup deformations therefore imply that the double coset hypergroups @xmath306 and @xmath307 are deformations of each other w.r.t . certain positive semicharacters . * the @xmath308-case . * for @xmath309 consider @xmath310 with the maximal compact subgroup @xmath311 . in this case @xmath301 with @xmath312 and @xmath313 . in this final section we use our previous results to translate limit theorems for random walks on @xmath189 into corresponding results for @xmath35-biinvariant random walks on @xmath34 . we first sketch the method . [ method ] let @xmath315 be a sequence on independent @xmath34-valued random variables with distributions @xmath316 ( @xmath317 ) . then @xmath318 with @xmath319 forms a @xmath35-biinvariant random walk on @xmath34 . using the canonical projection @xmath320 and the associated isomorphism @xmath321 of banach algebras , we see that @xmath322 is a markov process on @xmath19 with initial distribution @xmath85 and transition probabilities @xmath323 for @xmath324 , @xmath207 , and borel sets @xmath325 . this means that @xmath326 is a random walk on the hypergroup @xmath25 with transition probabilities @xmath327 . clearly , @xmath328 has distribution @xmath329 on the other hand , when considering the canonical projection @xmath330 , we find unique probability measures @xmath331 with @xmath332 for @xmath333 . if @xmath334 is a sequence on independent @xmath189-valued random variables with distributions @xmath335 , we get a random walk @xmath336 on @xmath189 whose projection @xmath337 is a random walk on @xmath25 with the same transition probabilities as @xmath338 , i.e. , @xmath337 and @xmath338 have the same finite dimensional distributions and therefore admit the same limit theorems . hence , limit theorems for the random walk @xmath339 on @xmath189 may be translated into limit theorems for @xmath338 . here is an example . assume that the @xmath340 are independent of @xmath157 , and that for the associated @xmath341 on @xmath189 the first moment vector @xmath342 exists . then by kolmogorov s strong law , @xmath343 almost surely as @xmath344 . in other words , @xmath345 a.s .. as @xmath330 is homogeneous of degree 1 and contractive ( see below ) , we obtain @xmath346 and thus @xmath347 a.s .. therefore , @xmath348 a.s . for @xmath344 . this strong law for @xmath338 has the computational drawback that @xmath349 is described in terms of @xmath350 . to overcome this , we introduce a suitable modified moment function @xmath351 such that for all probability measures @xmath350 having first moments , @xmath352 we shall show that @xmath353 is determined uniquely by ( [ momentenforderung ] ) and give an explicit formula . this allows to compute the limiting constant @xmath349 directly on @xmath19 . we next define @xmath353 . motivated by ( [ momentenforderung ] ) for @xmath359 with @xmath360 for @xmath207 , we put @xmath361 notice here that @xmath362 . the function @xmath353 is obviously @xmath35-invariant , continuous , and satisfies @xmath363 for @xmath364 with respect to the norm induced by the killing form . we first check that @xmath368 for @xmath366 . the definition of @xmath353 and the harish - chandra formula ( [ har ] ) show that for @xmath369 , @xmath370 where @xmath371 is the derivative in direction @xmath372 w.r.t . the open weyl chamber @xmath373 corresponding to @xmath19 is an orthogonal transversal manifold for the adjoint action of @xmath35 on @xmath189 ( @xcite , ch.ii , 3.4.(vi ) ) . therefore the orthogonal complement @xmath374 of @xmath191 in @xmath189 coincides with the tangent space of the orbit @xmath375 in @xmath178 . as @xmath376 is constant on @xmath35-orbits , @xmath377 for @xmath378 . hence @xmath379 . in order to check @xmath380 , we recall from ch . 3 of @xcite that @xmath381 on the other hand , an elementary rearrangement inequality ( theorem 368 of @xcite ) shows that for @xmath382 and @xmath383 , @xmath384 as @xmath385 , we obtain for @xmath383 that @xmath386 eq . ( [ charc ] ) now shows that @xmath380 . this completes the proof of part ( i ) . part ( ii ) follows from eqs . ( [ ableitungm1 ] ) and ( [ euclidean ] ) for the spherical functions @xmath304 on @xmath191 . [ moments ] let @xmath350 be a probability measure and @xmath387 . we say that @xmath341 admits @xmath388-th moments if @xmath389 , or equivalently , if @xmath390 for all @xmath391 . this condition can be translated into a corresponding condition for @xmath392 . in fact , as @xmath35 acts on @xmath189 as a group of orthogonal transformations , @xmath393.\ ] ] therefore , @xmath350 admits @xmath388-th moments if and only if @xmath394 admits @xmath388-th moments . proposition [ m1inc ] and the estimate @xmath395 for @xmath364 show that that this condition for @xmath396 implies that the modified moment vector @xmath397 exists . moreover , as @xmath398 we obtain from propos . [ m1inc ] that @xmath399 and hence , as claimed , @xmath400 [ strong - law ] let @xmath4010,2[$ ] and @xmath402 such that @xmath403 admits @xmath388-th moments . let @xmath315 be a sequence of i.i.d . @xmath34-valued @xmath80-distributed random variables . then @xmath404 for @xmath344 , with @xmath405 in case @xmath406 , while @xmath407 is arbitrary for @xmath4010,1[$ ] . let @xmath406 . by section [ moments ] , the @xmath388-th moment of the associated @xmath408 with @xmath409 exists . let @xmath334 be i.i.d . @xmath189-valued random variables with distribution @xmath341 , and @xmath410 the associated random walk on @xmath189 . the classical marcinkiewicz - zygmund law ( theorem 5.2.2 of @xcite ) yields that for all @xmath391 , @xmath411 as @xmath412 but this means that @xmath413 a.s . and hence , by [ m1inc](i ) and ( [ momentenforderung ] ) , @xmath414 as @xmath415 and @xmath416 have the same finite - dimensional distributions , the claim follows . the case @xmath4010,1[$ ] is similar . [ strong - law - spezial ] let @xmath422 and @xmath423 such that its projection @xmath403 admits @xmath388-th moments . then , for each sequence @xmath315 of i.i.d . @xmath1-valued and @xmath80-distributed random variables , @xmath424 the strong laws [ strong - law ] and [ strong - law - spezial ] are in principle well - known especially for @xmath425 ; see for instance @xcite and references cited therein . nevertheless our approach may be of some interest , because it uses the close connection between ( biinvariant ) random walks on @xmath34 and those on @xmath189 in a simple explicit way . t. koornwinder , _ jacobi functions and analysis of noncompact semisimple lie groups_. in : r. a. askey et al . ( eds . ) , special functions : group theoretical aspects and applications . dordrecht - boston - lancaster : d. reidel publishing company 1984 , pp . 1 - 85 .
we establish a deformation isomorphism between the algebras of @xmath0-biinvariant compactly supported measures on @xmath1 and @xmath0-conjugation invariant measures on the euclidean space @xmath2 of all hermitian @xmath3-matrices with trace @xmath4 . this isomorphism concisely explains a close connection between the spectral problem for sums of hermititan matrices on one hand and the singular spectral problem for products of matrices from @xmath1 on the other , which has recently been observed by klyachko @xcite . from this deformation we further obtain an explicit , probability preserving and isometric isomorphism between the banach algebra of bounded @xmath0-biinvariant measures on @xmath1 and a certain ( non - invariant ) subalgebra of the bounded signed measures on @xmath2 . we demonstrate how this probability preserving isomorphism leads to limit theorems for the singular spectrum of @xmath0-biinvariant random walks on @xmath1 in a simple way . our construction relies on deformations of hypergroup convolutions and will be carried out in the general setting of complex semisimple lie groups .
grb afterglows are typically observed to decay as power laws in time ( as reviewed by , e.g. , * ? ? ? the leading afterglow model is the synchrotron fireball @xcite . it describes the afterglow as synchrotron emission from shock - accelerated electrons with a broken power law spectral energy distribution , with several characteristic break frequencies . when the typical synchrotron frequency ( @xmath13 ) or the cooling frequency ( @xmath7 ) passes through the optical bands , the model predicts a break in the lightcurve . the fireball model s spectral breaks have different power law indices depending upon the circumburst medium ( as discussed by * ? ? ? the different resulting lightcurves allow important physical distinctions to be inferred from early afterglow observations . one example is the anticipated difference between a constant circumburst density ( called `` ism '' ) or @xmath14 density gradient ( called `` wind '' as it resembles the environment produced by a steady mass - loss wind outflow ) . the grb follow - up community has worked to produce ever - earlier observations in order to detect lightcurve breaks . the passage of the cooling frequency , @xmath7 , has been inferred from changes in the broadband spectral index . the first such example was grb 970508 @xcite , where the optical - to - x - ray slope steepened as if @xmath7 were passing between them . direct evidence of a cooling break passage in lightcurves is rarer . grb 030329 showed a shallow change of slope at @xmath15 d postulated to be the cooling break @xcite , although the standard fireball picture is complicated in that case , possibly with `` layered '' ejecta giving 2 jets and 2 jet breaks @xcite . more recently , @xcite found a shallow break in the grb 040924 afterglow better explained by @xmath7 than a jet break . here we report observations of the grb 050502a afterglow in unfiltered optical and @xmath16 bands . the optical data span nearly 3 logarithmic decades in time and begin at @xmath17 min after the start of the grb . we discuss the lightcurves in the context of the fireball model . this paper s optical and near - ir lightcurves are the result of three observation teams with different instruments at separate sites : rotse - iii , pairitel , and mdm . the rotse - iii array is a worldwide network of 0.45 m robotic , automated telescopes , built for fast ( @xmath18 s ) responses to grb triggers from satellites such as hete-2 and _ swift_. they have a wide ( @xmath19 ) field of view imaged onto a marconi @xmath20 back - illuminated thinned ccd , and operate without filters , with a bandpass from approximately 400 to 900 nm . rotse - iiib is located at mcdonald observatory in texas . the rotse - iii systems are described in detail in @xcite . the 1.3 m diameter pairitel ( peters automated infrared imaging telescope ) is a fully automated incarnation of the system used for the two micron all - sky survey ( 2mass ) . it is located on the ridge at mt . hopkins , arizona . the camera consists of three 256@xmath21256 nicmos3 arrays that image simultaneously the same @xmath22 portion of the sky at @xmath23 , @xmath24 , and @xmath25 bands ( central wavelengths 1.2 , 1.6 , and 2.2 @xmath26 m ) . the mdm observatory is located at kitt peak , arizona . it includes the 1.3 m mcgraw - hill telescope , which covers an @xmath27 field of view . this instrument operates with a standard set of filters . the camera is a site @xmath28 thinned , backside illuminated ccd , with a pixel scale of @xmath29 . on 2005 may 2 , integral detected grb 050502a ( integral trigger 2484 ) at 02:13:57 ut . the position was distributed as a gamma - ray burst coordinates network ( gcn ) notice at 02:14:36 ut , with a @xmath30 radius error box , 39 s after the start of the burst @xcite . the burst had a duration of 21 s , with a fluence of @xmath31 in the 20 - 200 kev band @xcite . rotse - iiib responded automatically to the gcn notice in 5.0 s with the first exposure starting at 02:14:41.0 ut , 44 s after the burst and only 23 s after the cessation of @xmath32-ray activity . the automated scheduler began a program of ten 5-s exposures , ten 20-s exposures , and 412 60-s exposures . the first exposure was taken during evening twilight hours , and rotse - iiib was able to follow the burst position until the morning twilight . the early images were affected by scattered clouds and a bright sky background . near real - time analysis of the rotse - iii images detected a @xmath33 magnitude fading source at @xmath34 , @xmath35 ( j2000.0 ) that was not visible on the digitized sky survey red plates , which we reported via the gcn circular e - mail exploder within 40 minutes of the burst @xcite . the pairitel instrument on mt . hopkins received the grb 050502a trigger before dusk and generated a rapid response too . when an interrupt is generated by a new burst at nighttime , a new set of observations is queued , overriding all other targets . typical time from gcn alert to slew is 10 seconds and typical slew times are 1 - 2 minutes . since this burst occurred before nighttime science operations , it overrode the scheduled observations related to telescope pointing . this behavior yielded a bad pointing model in the first set of observations and has subsequently been prohibited . the first set of usable imaging began at 02:52:39.5 ut , or approximately 39 minutes after the trigger . several imaging epochs were conducted over the following 5 hours . pairitel images were acquired with double - correlated sampling with effective exposure times of 7.848 seconds and dithered over several positions during a single epoch of observation . typical total integration times are between 2 to 30 minutes with approximately 1030 different dither positions . the mdm observatory began @xmath1-band observations 1.8 hours after the burst , following the initial rotse gcn report . 21 exposures were taken of the grb field spanning a total of 6.4 hours . the three diverse datasets required different photometric reductions . all the data described below are given in table [ tab : photometry ] . the rotse - iiib images were dark - subtracted and flat - fielded with its standard pipeline . the flat - field was generated from 30 twilight images . sextractor @xcite was applied for the initial object detection . the images were then processed with a customized version of the daophot psf fitting package @xcite that has been ported to the idl astronomy user s library @xcite . the psf is calculated for each image from a set of @xmath36 well - measured stars within @xmath37 of the target location . the psf is fit simultaneously to groups of stars using the nstar procedure @xcite . relative photometry was then performed using 13 neighboring stars . rotse - iii magnitudes are calibrated as @xmath1-equivalent to accomodate rotse s peak optical bandpass sensitivity at red wavelengths . the @xmath1-equivalent magnitude zero - point is calculated from the median offset to the usno 1 m @xmath1-band standard stars @xcite in the magnitude range of @xmath38 with the color of typical stars in the field such that @xmath39 . as we have no data on afterglow color information at the early time , no additional color corrections have been applied to the unfiltered data , and the magnitudes quoted are then treated as @xmath1-band and referred to as `` @xmath40 '' . the pairitel data were processed through its custom reductions pipeline and then combined into mosaics as a function of epoch and filter . the mosaicking used a cross - correlation between reduced images to find the pixel offsets for the dithers . final images were constructed using a static bad pixel mask and drizzling @xcite . bias and sky frames are obtained for each frame by median combining several images before and after that frame . flat fields are applied from archival sky flats , and are known to be highly stable over long periods of time . aperture photometry was performed in a 3.5 pixel radius aperture ( the average seeing fwhm was 2.4 pixels ) . the pairitel photometric zeropoints for each of @xmath23 , @xmath24 , and @xmath25 were determined from 2mass catalog stars , of which there are @xmath41 in the field . the mdm data were processed using standard iraf / daophot procedures . aperture photometry was performed in a @xmath42 radius aperture ( average seeing was @xmath6 @xmath42 ) centered on the ot and nearby field stars . instrumental magnitudes were then transformed to the @xmath1 system using the latest calibration provided by henden et al . ( 2005 ) , using differential photometry for each image with respect to two henden stars , the only ones available in the mdm field ( ra , dec = 13:30:04.3 , + 42:41:06.0 and 13:29:59.8,+42:43:00.2 , both j2000 ) . in order to analyze the results , we convert all magnitudes from table [ tab : photometry ] to spectral flux densities . for the @xmath1 ( or rotse @xmath43 ) magnitudes , we use the effective frequencies and zeropoint fluxes of @xcite . the @xmath44 magnitudes are converted using @xcite . there is scant galactic extinction at the high galactic latitude of the transient . however , we do correct the fluxes for 0.028 , 0.01 , 0.006 , and 0.004 mags of extinction in the @xmath1 ( & @xmath43 ) , @xmath23 , @xmath24 , and @xmath25 bands respectively , as found from the extinction estimates of @xcite . @xcite established the source redshift , @xmath45 . at such high @xmath46 , absorption from the lyman-@xmath47 forest becomes important in the rotse - iii bandpass . the bandpass covers the wavelength range of approximately @xmath48 to @xmath49 , but the ly@xmath47 absorption from this @xmath46 significantly depresses wavelengths corresponding to @xmath48 and @xmath50 . ordinarily the blue spectrum of an ot would cause an overestimation of the @xmath43-band flux from rotse instrumental magnitudes , but in this case extinction across a significant fraction of the bandpass makes the conversion using the zeropoint from unextincted field stars _ underestimate _ the @xmath43-band flux . as a result , we expect a color term between the rotse @xmath43 and the mdm @xmath1 magnitudes . the precise color term is a complicated function of the ly@xmath47 absorption and the ot spectrum . we fit a constant color term parameter when examining @xmath1 and @xmath43 data together , as a constant multiplicative offset of the rotse points relative to the mdm data . this constancy presumes that any ot color changes within the rotse bandpass have a small effect upon the color term parameter by comparison to the data uncertainties . the lightcurves are plotted in figure [ fig : fig1 ] . the @xmath1-band lightcurve appears to be a broken power law . the need for a break is made obvious by the precise mdm observations at later times . the first ( rotse ) point in the lightcurve at @xmath51 s seems underluminous relative to subsequent decay . the ir data are less well - sampled , but can be described as a set of power laws with decay rates similar to those seen in the @xmath1 band . to quantify these trends , we fit the data using power law and broken power law forms . the smooth broken power law form is @xmath52 where @xmath53 is the break time , @xmath54 is the flux at the break time , @xmath55 and @xmath56 are the two power law indices , and @xmath57 is a sharpness parameter . the data ( fig . [ fig : fig1 ] ) clearly require a very sharp break transition . we use the form above for convenience , but do not fit the sharpness , setting @xmath57 at a very high value ( @xmath58 ) . as noted above , we also fit a constant multiplicative factor for the rotse @xmath43 points , as a color term . the rotse points after the break observed in the data have significant uncertainties . we expect any color change that may be associated with the break will not change the color term significantly in comparison to the rotse error bars . the `` separate '' fit uses the form discussed above for the @xmath1-band data and individual power laws for each ir band ( there is no substantial evidence for breaks in the ir data ) . this allows each band to have different temporal decays and flux normalizations . the `` multiband '' fit connects the wavelength bands via a spectral index and coupled temporal decays , as expected in the fireball model . this fit uses the above functional form for @xmath1-band data , and simultaneously fits single power laws for the ir of the form @xmath59 where @xmath60 designates the frequency of an ir band . we attempt two `` multiband '' forms , @xmath61 ( the initial @xmath1-band decay index ) and @xmath62 ( the final @xmath1-band decay index ) . first , we examined the @xmath1-band lightcurve ( rotse & mdm ) . a single power law does not produce an acceptable fit . we fit all @xmath1-band data ( including `` @xmath43 '' ) to the `` broken power law + color term '' function described above , and repeated the exercise excluding the first point . both produced formally acceptable fits , but excluding the first point improved the fit notably from @xmath63 for 39 degrees of freedom ( dof ) to @xmath64 for 38 dof . the first observation , at 44 s post - grb , appears to plateau , not joining the overall decay seen later . the flux density is approximately 4@xmath65 below the back - extrapolation of the fit excluding the first point , marginal evidence for a deviation at early times from the later decay . while a single point can not definitively establish such a deviation , the first point does not fit the overall later decay and thus we exclude it from the lightcurve fits discussed below and recorded in table [ tab : fitpars ] . the potential constraints placed by this first point are discussed later , in [ sec : fireball ] . overluminosity , or flux above the expectations from the subsequent decay , would be expected if a reverse shock s emission were observable as it passes through the ejecta at early times comparable to our first observation ( e.g. , * ? ? ? * ; * ? ? ? this is certainly not seen in our lightcurve . the maximum flux we observe in the first observation ( 99% confidence upper limit ) is only 90% of the back - extrapolated flux from the fit excluding the first data point . a flux level significantly above this extrapolation would have been readily measurable . the @xmath1 ( & @xmath43 ) band , excluding the first point , were fit to the broken power law as described above , and each of the @xmath23 , @xmath24 , and @xmath25 bands were independently fit to a power law . nondetections are included as flux densities of @xmath66 to force the fit not to overestimate undetected fluxes . however , the elimination of nondetections from the fits does not change the parameters significantly . the resulting fit parameters are given in table [ tab : fitpars ] . overall , the total @xmath67 for 61 dof . due to the ly@xmath47 absorption discussed above , the @xmath1-band fit requires a significant color term , 0.25 magnitudes , for the rotse @xmath43 data relative to the filtered mdm @xmath1 measurements . with this systematic correction , the @xmath1-band lightcurve has a sharp break at a time @xmath68 s post - burst . the @xmath1-band lightcurve break is from a decay index @xmath47 of @xmath69 to @xmath70 , a @xmath71 . in the fireball model , the expected steepening from a `` jet break '' ( observing a lack of flux due to the edge of conical ejecta ) is at least @xmath72 @xcite . the only shallow steepening expected in the fireball model is from the passage of the cooling frequency . the model predicts @xmath73 , consistent with the observed break at the @xmath74 level . the nir data are consistent with power law fits . the fitted @xmath23 , @xmath24 , and @xmath25 decay rates lie between the initial and final @xmath1-band decay rates . they are internally consistent at the 1@xmath65 level and thus show no evidence for nir color changes . the nir decay indices are not as tightly constrained as those describing the @xmath1-band , with individual uncertainties @xmath75 . as the @xmath1-band break is shallow , each nir band s decay is consistent with either the initial or the final @xmath1-band decay index ( at the 1.6@xmath65 level or better ) . consistency with the _ initial _ @xmath1-band decay would be expected for an ism fireball in which @xmath7 drops as @xmath76 , while consistency with the _ final _ @xmath1-band decay would be expected for a wind fireball where @xmath7 rises as @xmath77 ( as reviewed by * ? ? ? * ) . the @xmath25 band is slightly more consistent with the ism picture than the wind , the @xmath24 and @xmath23 with the wind than the ism . as evident from the poorly - determined nir decay rates , the nir data do not have the sensitivity to determine the presence or absence of a lightcurve break as shallow as that observed in the @xmath1-band . moreover , the nir time coverage is from 2320 s to 20700 s and the observed @xmath1-band break is at @xmath68 s. even if the nir data were more sensitive , given the timing of the @xmath1-band lightcurve break , in the ism ( wind ) model prediction , @xmath7 would not be observed falling ( rising ) to ( from ) the @xmath23 band during the nir coverage . the ism model would predict that @xmath7 would fall from @xmath1 to @xmath23 only at @xmath78 s. the wind model would predict that @xmath7 would rise from @xmath23 to @xmath1 , with the break in @xmath23 at @xmath79 s. given no obvious color changes , we also fit the multiband function as described in [ earlyunderlum ] , [ separatefit ] . we gave initial parameters to search for two cases , with the nir decay either the power law behavior of the initial @xmath1-band lightcurve decay , or of the final @xmath1 decay . the two results are shown in table [ tab : fitpars ] . the @xmath1 decay indices , the @xmath1-band break time , and the @xmath1 flux at the break time do not change significantly . the break amplitude @xmath80 is consistent with the cooling break passage at the 1.5@xmath65 level in both cases . both multiband fits are formally quite good , with @xmath81/dof@xmath821 . the data are insufficiently constraining to distinguish between @xmath83 ( ism expectation ) , and @xmath84 ( wind expectation ) . we show both of these fits in figure [ fig : fig1 ] , where the shallower @xmath11 favored by the @xmath84 model and the steeper ir decay make the fitted model underestimate all the @xmath25 data slightly . as the fit is nevertheless good , this only mildly favors @xmath85 and the ism interpretation . the optical - nir spectrum of the data is well fit by a power law @xmath86 . this is observed in the nir data alone ( @xmath44 at one common epoch ) , as well as @xmath87 data ( interpolating the @xmath1-band point ) . there is no significant change across observation epochs as there is no evidence for color changes from the lightcurves . the temporal decays of the ir bands are not strongly constrained ( @xmath88 , table [ tab : fitpars ] ) , are all mutually consistent , and are consistent ( at @xmath89 ) with either the initial or the final @xmath1-band decay . moreover , the complete dataset is well fit by the `` multiband '' function that includes the spectral index @xmath11 to connect the bands . the @xmath44 spectrum at the initial epoch yields @xmath90 , while @xmath87 gives @xmath91 . these are not as well - constrained as the @xmath11 fit using the `` multiband '' function . when @xmath83 , @xmath92 is fit , and when @xmath84 , @xmath93 . these results are all mutually consistent . given their variation , we take our measure of the optical - nir spectral index to be @xmath94 . the observed @xmath1-band lightcurve break will produce color changes relative to the single power laws of the nir bands , but they are not detectable as we do not have good ir detections in the period from the break to the end of our nir observations . the power law break in the @xmath1-band lightcurve is consistent with @xmath95 , the value expected for the passage of a cooling break @xmath7 in a fireball model with either an ism or wind circumburst density . the fireball model can also steepen a decay via a jet break , or via the passage of the typical frequency @xmath13 in the fast - cooling ( @xmath96 ) case . the latter , for an observing frequency @xmath97 , is the transition from @xmath98 to @xmath99 . as previously noted , a jet break should be significantly steeper than the @xmath100 observed @xcite . the @xmath13 passage in a fast - cooling case requires the initial decay to be @xmath101 before the break ( * ? ? ? * viib ) which is not compatible with any of the lightcurves . from the temporal indices , we conclude that we are observing the passage of @xmath7 through @xmath1 at @xmath102 s. in the ism case , this is a passage from @xmath103 to @xmath104 . in the wind case , it is from @xmath105 to @xmath103 . both require that @xmath13 is below the optical by the first fit data point at @xmath106 s. the fireball model predicts relations between observed temporal and spectral indices and the value @xmath107 ( @xmath108 ) of the power law index for the spectral energy distribution of synchrotron - emitting electrons . the relations are functions of synchrotron spectral break ordering , the spectral segment , and whether the fireball is the ism or wind case , compiled by @xcite . with an initial and a final decay index , there are two relations to determine @xmath107 for each assumed case ( ism or wind ) . using the @xmath1-band decay indices from the `` separate '' fit gives values of @xmath107 , @xmath109 ( from the initial @xmath110 ) and @xmath111 ( final @xmath112 ) , for the constant - density ism case , and @xmath113 ( @xmath110 ) and @xmath114 ( @xmath112 ) , assuming the wind case . given the relations between @xmath107 and @xmath47 , as well as those between @xmath107 and @xmath11 , @xmath47 and @xmath11 are related ( again compiled in * ? ? ? * ) . with the @xmath1 lightcurve break modelled by the passage of @xmath7 in either an ism or a wind - like medium , there are four possible @xmath47@xmath11 `` closure '' relations of @xmath115 . we summarize them in table [ tab : alphabeta ] , along with the values for the electron energy spectral index @xmath116 in these cases , and @xmath117 . as before , the ism model better represents the data . the ism closure @xmath118 results are more consistent with 0 , but the wind cases do not deviate by more than @xmath119 . similarly , the values of @xmath107 inferred from the ism relations @xmath116 better agree with its @xmath117 than in the wind case , with a similar level of consistency as the closure relations . the observed spectral index of @xmath120 is consistent with either the ism or wind picture , as seen in the closure relations , but is slightly steeper than the wind model expects ( table [ tab : alphabeta ] ) , by @xmath121 . at the source redshift of @xmath45 @xcite , the @xmath25-band in the observer frame is roughly @xmath48-band in the local frame , and the @xmath1-band is near - uv at @xmath122 hz . over this frequency range in the source frame , a dust extinction law such as that of the large magellanic cloud ( lmc ) will be linear , leading to a simple power law steepening of the spectrum @xcite . dust extinction such as that observed in the milky way would have a significant bump in the middle of that frequency range , so we did not consider it . we use the lmc prescription of @xcite and determine that a small local extinction @xmath123 would give @xmath124 . the amount of extinction would be sufficient to provide an intrinsic spectrum compatible with the wind model . the extinction law prescribed for the small magellanic cloud bar is steeper than the lmc case , and so if applicable should require less dust . there can not be a very large amount of extinction at the burst redshift . to get the observed spectrum with as little extinction as @xmath125 requires a flat or rising intrinsic spectrum . such an intrinsic spectrum is inconsistent with the fireball model . @xcite present the _ swift_/xrt s x - ray upper limits , from observations later than those presented here . we compare the x - ray limits to the optical by extrapolating the @xmath1-band flux density to their epoch , using our fitted models . we convert their flux limit to a spectral flux density limit using the spectral information given , both for their earliest epoch at @xmath126 ksec , and for the overall observation . we find that the early optical spectral index of @xmath127 is marginally consistent with the x - ray limit , overestimating the x - ray 90% upper limit by approximately a factor of 3 . ( given that the uncertainty is @xmath11 is 0.1 , note that @xmath128 would only overestimate the x - ray limit by a factor of 1.5 , and @xmath129 would match it . ) the expected initial intrinsic @xmath127 ( ism model ; table [ tab : alphabeta ] ) would be marginally consistent , but the expected @xmath130 ( wind ; table [ tab : alphabeta ] ) would be inconsistent , overestimating the x - ray limit by at least a factor of 14 . this x - ray optical spectrum is consistent with the passage of @xmath7 . with an ism model , @xmath7 has passed below the optical , steepening the spectrum between @xmath1 and the x - ray at later times . with a wind profile , @xmath7 increases with time , placing it between @xmath1 and the x - ray at the time of the x - ray limit . however , @xmath7 would not get far above the optical and the spectrum steepens between @xmath7 and the x - ray . in both cases , the steepening @xmath131 above @xmath7 results in a model - predicted x - ray flux well below the _ swift_/xrt limit . if the sub - luminosity of the earliest @xmath1-band observation ( 44 s ) were known to be due to the passage of the peak @xmath13 , there would be three numerical constraints for the fireball model s spectral parameters : the peak time in @xmath1-band , the peak flux density value at that time , and the cooling break time at the @xmath1-band frequency . the early lightcurve behavior is neither strong evidence for a rollover from an earlier flatter lightcurve evolution , nor for a particular physical mechanism for a rollover . therefore the first two constraints are inequalities ; the peak @xmath13 could have have passed the optical band earlier . the observed decay from @xmath132 to @xmath133 s requires that the passage @xmath134 is occurring or has occurred at @xmath135 s. if @xmath13 passed the @xmath1-band before @xmath132 s , the synchrotron spectral peak would be at a lower frequency and a higher flux density than the @xmath1-band s . thus the peak flux density @xmath136 at @xmath132 s must be at least as bright as the initial @xmath1-band point . the peak passage @xmath13 is quite early , having occurred at a time no later than 2.1 times the grb duration measured by @xcite . the cooling break passage time of @xmath137 s is significant , placing @xmath7 in the optical at a fairly early time . this is quite early in comparison to several other cases where broadband data has indicated @xmath7 is above the optical at a time @xmath6 days post - burst , e.g. , 970508 @xcite , 000301c @xcite , or 011211 @xcite . the implication in the ism case with @xmath138 would be a low value of @xmath7 during the first observation , with @xmath139hz at @xmath132 s. we consider the observational constraints in light of both the ism and the wind models for the data . the synchrotron spectrum of a spherical fireball expanding into the ism at a known redshift depends upon five parameters : energy @xmath140 , density @xmath141 , the power law slope of the electron energy distribution @xmath107 , the energy fraction partitioned to the electrons , @xmath142 , and the energy partition to the magnetic fields , @xmath143 ( see * ? ? ? * and references therein ) . with three constraints and the value of @xmath107 , we can put three of @xmath140 , @xmath141 , @xmath143 , and @xmath142 in terms of the fourth , thus placing a limit upon a combination of two parameters . we use the equations of @xcite for the ism slow - cooling case for @xmath13 , @xmath7 , and @xmath144 . adopting @xmath145 , we find the numerical relations using the redshift @xmath45 @xcite and a cosmology with @xmath146 , @xmath147 , @xmath148 km s@xmath149mpc@xmath149 to determine the appropriate luminosity distance . we can then isolate @xmath150 where @xmath142 and @xmath143 are scaled to 10% and 1% respectively as these are reasonable expectations for the parameters from broadband fits in many afterglow cases ( see , e.g. , * ? ? ? * ; * ? ? ? within the simple fireball model that well - describes this afterglow case in the optical - ir , the implication is that the microphysical energy partitions must be quite small . for example , if @xmath151 , then @xmath152 . other parameters can be put in terms of these , thus density @xmath153 . keeping @xmath154 @xmath155 requires @xmath143 to be not much less than 1% . we note that the microphysical parameter combination in eq . 3 is proportional to the peak flux density at 44 s as @xmath156 , and to the frequency of the peak @xmath13 at 44 s as @xmath157 . thus if the peak was significantly below the @xmath1 band at that time , this combination of microphysical parameters would be notably smaller . for a wind model , the physical parameter set is the same except that the density is parameterized by @xmath158 , where @xmath159 g cm@xmath149 . we do a similar analysis using @xcite s equations . in this case , we find @xmath161 once again the microphysical parameters must be small ( and increasingly small as the initial value of @xmath13 is placed further below the @xmath1-band frequency ) . there is an additional constraint : @xmath162 this inequality indicates that , if @xmath163 at the initial observation , @xmath164 would be larger than 0.0004 . however , from the previous relation the low @xmath13 would require smaller microphysical parameters , @xmath142 or @xmath143 or both . if @xmath142 is decreased in the case @xmath163 , then @xmath158 would remain low . a value @xmath165 is a very small wind outflow parameter ; @xmath166 is observed in wolf - rayet stars ( as reviewed by * ? ? ? the physical parameter constraints are somewhat easier to fulfill with the ism fireball model than with the wind model . we observe a shallow break of @xmath100 in the @xmath1 lightcurve of the grb 050502a afterglow at approximately 6000 s post - burst . we note that @xmath23 , @xmath24 , and @xmath25 observations during this period , not as well - sampled , do not constrain the presence or absence of such a lightcurve break in the nir . with the shallow @xmath80 , we conclude that the optical break represents the passage of the synchrotron cooling frequency @xmath7 through @xmath1 . the observed spectral index is @xmath9 . this is consistent with an ism model for the broken @xmath1-band lightcurve , and slightly steeper than expected in a wind model . the temporal and spectral index closure relations slightly favor an ism over a wind interpretation . a small amount of dust at the host redshift would steepen an intrinsically flatter spectrum sufficiently to accomodate the wind interpretation . there is no evidence for an overluminosity at our earliest observation , 44 s after the initial gamma - rays . the first lightcurve point appears suppressed , but from a single point we can not conclude that this is the case . at a minimum , the early decay requires @xmath13 at or below @xmath167 at that time . using that constraint and the observation of @xmath7 s passage indicates that in both the ism and the wind explanations small microphysical energy partitions are required . the wind interpretation expects an exceptionally low wind outflow parameter @xmath158 , which may again somewhat favor an ism interpretation . @xcite find multicolor evidence for a lightcurve bump at @xmath168 s. a single rotse data point during this period has a slight rise above the overall decay , but not at a significant level given its signal to noise . @xmath44 sampling is insufficient to address whether the bump occurs at these wavelengths . without the well - sampled mdm @xmath1-band observations presented here from 5800 - 28000 s , @xcite did not detect the passage of @xmath7 through @xmath1 . @xcite find the data favors a density variation over a refreshed shock as the source of the bump . in that case , the significance of the bump is greater than expected for @xmath169 ( see * ? ? ? * ) . for the wind case the cooling break passage observed in the dataset presented here would be from @xmath169 to @xmath170 at @xmath826000 s , so this again favors an ism over a wind interpretation . we note that the observed steeper , steady decay following the lightcurve break is evident in the data over at least 0.7 logarithmic decades in time . despite the evidence for a density variation in the data of @xcite , such an impulsive event can not explain the sustained change in lightcurve evolution seen here . there are several lines of evidence , including the spectral index , the closure relations of spectral and temporal indices , and the optical bump seen by @xcite , that all somewhat favor an ism model over a wind one for this afterglow . with _ swift _ and integral providing rapid triggers , and rapid response instruments such as rotse and pairitel providing followup , the grb community is accumulating bursts with prompt , detailed observations . such data sets should further probe the physics of grb afterglows . rotse - iii has been supported by nasa grant nng-04wc41 g , nsf grant ast-0407061 , the australian research council , the university of new south wales , the university of texas , and the university of michigan . work performed at lanl is supported through internal ldrd funding . special thanks to the observatory staff at mcdonald observatory , especially david doss . the peters automated infrared imaging telescope ( pairitel ) is operated by the smithsonian astrophysical observatory ( sao ) and was made possible by a grant from the harvard university milton fund , the camera loan from the university of virginia , and the continued support of the sao and uc berkeley . partial support is also supplied by a nasa _ swift _ cycle 1 guest investigator grant . 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 . lcrrc rotse - iiib & none & 44.0 & 49.0 & @xmath171 + rotse - iiib & none & 59.0 & 78.9 & @xmath172 + rotse - iiib & none & 88.5 & 108.0 & @xmath173 + rotse - iiib & none & 117.2 & 136.8 & @xmath174 + rotse - iiib & none & 146.7 & 180.5 & @xmath175 + rotse - iiib & none & 190.0 & 239.6 & @xmath176 + rotse - iiib & none & 248.7 & 327.7 & @xmath177 + rotse - iiib & none & 337.5 & 416.7 & @xmath178 + rotse - iiib & none & 426.5 & 546.1 & @xmath179 + rotse - iiib & none & 555.3 & 754.4 & @xmath180 + rotse - iiib & none & 763.6 & 962.7 & @xmath181 + rotse - iiib & none & 971.9 & 1309.9 & @xmath182 + rotse - iiib & none & 1319.7 & 1658.2 & @xmath183 + rotse - iiib & none & 1667.4 & 2214.4 & @xmath184 + rotse - iiib & none & 2224.1 & 2908.7 & @xmath185 + rotse - iiib & none & 2918.0 & 3881.3 & @xmath186 + rotse - iiib & none & 3890.5 & 5075.0 & @xmath187 + rotse - iiib & none & 5084.1 & 6741.0 & @xmath188 + rotse - iiib & none & 6750.1 & 8823.3 & @xmath189 + rotse - iiib & none & 8832.5 & 11670.7 & @xmath190 + rotse - iiib & none & 11680.6 & 15415.2 & @xmath191 + rotse - iiib & none & 15425.1 & 20274.6 & @xmath192 + rotse - iiib & none & 20284.1 & 28421.6 & @xmath193 + mdm & @xmath1 & 5816.0 & 6416.0 & @xmath194 + mdm & @xmath1 & 6458.0 & 7058.0 & @xmath195 + mdm & @xmath1 & 7105.0 & 7705.0 & @xmath196 + mdm & @xmath1 & 7750.0 & 8350.0 & @xmath197 + mdm & @xmath1 & 8386.0 & 8986.0 & @xmath198 + mdm & @xmath1 & 9025.0 & 9625.0 & @xmath199 + mdm & @xmath1 & 9928.0 & 10828.0 & @xmath200 + mdm & @xmath1 & 10871.0 & 11771.0 & @xmath201 + mdm & @xmath1 & 11812.0 & 12712.0 & @xmath202 + mdm & @xmath1 & 12750.0 & 13650.0 & @xmath203 + mdm & @xmath1 & 13687.0 & 14587.0 & @xmath204 + mdm & @xmath1 & 14621.0 & 15461.0 & @xmath205 + mdm & @xmath1 & 15645.0 & 16545.0 & @xmath206 + mdm & @xmath1 & 16587.0 & 17487.0 & @xmath207 + mdm & @xmath1 & 17537.0 & 18737.0 & @xmath208 + mdm & @xmath1 & 18797.0 & 19997.0 & @xmath209 + mdm & @xmath1 & 20115.0 & 21915.0 & @xmath210 + mdm & @xmath1 & 22041.0 & 23841.0 & @xmath211 + mdm & @xmath1 & 24057.0 & 25857.0 & @xmath212 + mdm & @xmath1 & 26358.0 & 28158.0 & @xmath213 + mdm & @xmath1 & 28319.0 & 30119.0 & @xmath214 + pairitel & @xmath23 & 2322.5 & 3112.4 & @xmath215 + pairitel & @xmath23 & 4942.0 & 6335.4 & @xmath216 + pairitel & @xmath23 & 6365.4 & 7583.0 & @xmath217 + pairitel & @xmath23 & 8158.9 & 9903.5 & @xmath218 + pairitel & @xmath23 & 9970.9 & 11722.3 & @xmath219 + pairitel & @xmath23 & 11790.4 & 13538.4 & @xmath220 + pairitel & @xmath23 & 13568.0 & 15316.8 & @xmath221 + pairitel & @xmath23 & 15382.5 & 17134.5 & @xmath222 + pairitel & @xmath23 & 17201.3 & 17979.7 & @xmath223 + pairitel & @xmath23 & 18951.5 & 20742.1 & @xmath222 + pairitel & @xmath24 & 2322.5 & 3112.4 & @xmath224 + pairitel & @xmath24 & 4942.0 & 6335.4 & @xmath225 + pairitel & @xmath24 & 6365.4 & 7583.0 & @xmath226 + pairitel & @xmath24 & 8150.3 & 9903.5 & @xmath227 + pairitel & @xmath24 & 9970.9 & 11722.3 & @xmath228 + pairitel & @xmath24 & 11790.4 & 13538.4 & @xmath229 + pairitel & @xmath24 & 13568.0 & 15316.8 & @xmath230 + pairitel & @xmath24 & 15382.5 & 17134.5 & @xmath231 + pairitel & @xmath24 & 17201.3 & 17979.7 & @xmath232 + pairitel & @xmath25 & 2322.5 & 3112.4 & @xmath233 + pairitel & @xmath25 & 4942.0 & 6335.4 & @xmath234 + pairitel & @xmath25 & 6365.4 & 7583.0 & @xmath235 + pairitel & @xmath25 & 8150.3 & 9903.5 & @xmath236 + pairitel & @xmath25 & 9970.9 & 11722.3 & @xmath237 + pairitel & @xmath25 & 11790.4 & 13538.4 & @xmath238 + pairitel & @xmath25 & 13568.0 & 15316.8 & @xmath239 + pairitel & @xmath25 & 15382.5 & 17134.5 & @xmath240 + pairitel & @xmath25 & 17201.3 & 17979.7 & @xmath241 + pairitel & @xmath25 & 18951.5 & 20742.1 & @xmath240 + lccc + @xmath110 & early @xmath1 power law index & @xmath242 & - + @xmath112 & late @xmath1 power law index & @xmath243 & - + @xmath244 & @xmath1 break time & @xmath245 & sec + @xmath246 & @xmath247 at @xmath53 & @xmath248 & @xmath26jy + @xmath249 & rotse color term & @xmath250 & mag + @xmath251 & @xmath81 of @xmath1 fit / dof & 44.9/38 & - + @xmath252 & @xmath23 power law index & @xmath253 & - + @xmath254(@xmath255 ) & @xmath256 of 1st ir epoch & @xmath257 & @xmath26jy + @xmath258 & @xmath81 of @xmath23 fit / dof & 8.0/8 & - + @xmath259 & @xmath24 power law index & @xmath260 & - + @xmath261(@xmath255 ) & @xmath262 of 1st ir epoch & @xmath263 & @xmath26jy + @xmath264 & @xmath81 of @xmath24 fit / dof & 3.6/7 & - + @xmath265 & @xmath25 power law index & @xmath266 & - + @xmath267(@xmath255 ) & @xmath268 of 1st ir epoch & @xmath269 & @xmath26jy + @xmath270 & @xmath81 of @xmath25 fit / dof & 3.4/8 & - + + @xmath110 & early power law index & @xmath271 & - + @xmath112 & late power law index ( @xmath1 only ) & @xmath272 & - + @xmath244 & @xmath1 break time & @xmath273 & sec + @xmath246 & @xmath247 at @xmath53 & @xmath274 & @xmath26jy + @xmath11 & spectral index @xmath275 & @xmath276 & - + @xmath249 & rotse color term & @xmath277 & mag + @xmath81 & @xmath81 of multiband fit / dof & 65.7/66 & - + + @xmath110 & early power law index ( @xmath1 only ) & @xmath278 & - + @xmath112 & late power law index & @xmath279 & - + @xmath244 & @xmath1 break time & @xmath280 & sec + @xmath246 & @xmath247 at @xmath53 & @xmath281 & @xmath26jy + @xmath11 & spectral index @xmath275 & @xmath282 & - + @xmath249 & rotse color term & @xmath283 & mag + @xmath81 & @xmath81 of multiband fit / dof & 68.2/66 & - + cccccc ism , initially @xmath103 & @xmath284 & @xmath285 & @xmath109 & @xmath286 & @xmath287 + ism , initially @xmath103 & @xmath288 & @xmath289 & @xmath111 & & @xmath290 + wind , initially @xmath105 & @xmath291 & @xmath292 & @xmath113 & & @xmath293 + wind , initially @xmath105 & @xmath294 & @xmath295 & @xmath114 & @xmath286 & @xmath296 +
we present lightcurves of the afterglow of grb 050502a , including very early data at @xmath0 s. the lightcurve is composed of unfiltered rotse - iiib optical observations from 44 s to 6 h post - burst , @xmath1-band mdm observations from 1.6 to 8.4 h post - burst , and pairitel @xmath2 observations from 0.6 to 2.6 h post - burst . the optical lightcurve is fit by a broken power law , where @xmath3 steepens from @xmath4 to @xmath5 at @xmath65700 s. this steepening is consistent with the evolution expected for the passage of the cooling frequency @xmath7 through the optical band . even in our earliest observation at 44 s post - burst , there is no evidence that the optical flux is brighter than a backward extrapolation of the later power law would suggest . the observed decay indices and spectral index are consistent with either an ism or a wind fireball model , but slightly favor the ism interpretation . the expected spectral index in the ism interpretation is consistent within @xmath8 with the observed spectral index @xmath9 ; the wind interpretation would imply a slightly ( @xmath10 ) shallower spectral index than observed . a small amount of dust extinction at the source redshift could steepen an intrinsic spectrum sufficiently to account for the observed value of @xmath11 . in this picture , the early optical decay , with the peak at or below @xmath12 hz at 44 s , requires very small electron and magnetic energy partitions from the fireball .
the dynamic of a small quantum system interacting weakly with uncontrollable degrees of freedom is well understood when a markovian approximation applies . in this situation , after tracing - out the environment , the system density matrix evolution can be well approximated by a lindblad equation @xcite . besides that the applicability of the markovian approximation range over many areas of physics @xcite there exist several real systems whose dynamics present strong departures from it . the main signature of this departure is the presence of strong non - exponential decay behaviors , such as power law and stretched exponential . some examples are nanocrystal quantum dots under laser radiation schlegel , brokmann , grigolini , superconducting qubits makhlinreport , makhlin , falci , spin environments @xcite , dephasing in atomic and molecular physics @xcite , electron transfer and exciton dynamics in proteins @xcite , and molecular systems maintained in a glassy environment @xcite , to name but a few . these and another specific experimental situations rise up the necessity of finding formalism and effective evolutions able to describe the corresponding non - markovian dynamics . when the environment is modeled as an infinite set of normal modes , departure from a markov approximation can be related to the corresponding spectral density function . this situation was extensively studied for the spin - boson and boson - boson models , where exacts solutions are available leggett , grifoni , lang , hanggi , milena . exact solutions can also be formulated for more general systems . nevertheless , due to the huge analytical and numerical efforts needed for getting the non - markovian system dynamics , alternative numerical methods based in a special decomposition of the spectral density function were formulated @xcite . anomalous system dynamics also arise from a random matrix modeling of the system - environment interaction @xcite . this approach naturally applies when describing environments characterized by a complex dynamics as for example chaotic ones . outside from a microscopic point of view , there exist an increasing interest in describing non - markovian effects in open quantum systems by introducing memory contributions in lindblad evolutions wilkie , budini , cresser , lidar . this procedure provide easy manageable dynamics . while most of these models are phenomenological , in this paper we will relate the presence of strong memory effects in the standard lindblad theory @xcite with the microscopic interaction of a system with a complex structured environment . we will base our considerations in a splitting of the full hilbert space of the bath in a direct sum of sub - reservoirs , constructed in such a way that each one guarantees the conditions for the applicability of a markov approximation . our motivation for formulating this splitting comes from systems embedded in glassy environments , where the underlying disorder produce localized bath states , inducing a natural shell structure of modes , each set having a different coupling strength with the system @xcite . thus , we associate a different markovian sub - bath to each set of states . as we will demonstrate , the splitting assumption allows us to generalize the usual born - markov approximation , which in a natural way leads to the formulation of lindblad equations characterized by a random dissipative rate . as is well known from a classical context , master equations with random rates @xcite are useful for describing strong non - markovian effects @xcite . here we will demonstrate that the same scheme can also be applied in a quantum context under the previous conditions , i.e. , a complex environment under the splitting condition . as an example , we will study the anomalous dissipative dynamics of a quantum tunneling system described in a two level approximation @xcite . strong non - exponential behaviors , such as stretched exponential and power law , arise from the interplay between the unitary hopping dynamics and the memory effects induced by the environment . the conditions under which our modeling can be mapped to the spin - boson model and stochastic dynamics are established . in this context , the differences between our framework and other approaches @xcite introduced to deal with non - markovian environments are established . in general , the evolution of a system interacting with a complex environment can not be described in a markovian approximation . while a general treatment is clearly not possible , when the splitting condition applies , for weak system - bath interactions , the reduced dynamics can be described through a generalization of the born - markov approximation . we start by assuming a full microscopic hamiltonian description@xmath0where @xmath1 correspond to the hamiltonian of a system s , and @xmath2 correspond to the hamiltonian of the bath b. the term @xmath3 describes their mutual interaction , with the operators @xmath4 and @xmath5 acting on the system and bath hilbert spaces respectively . in an interaction representation with respect to @xmath6 , the total density matrix @xmath7 evolves as@xmath8,\]]where @xmath9 is the interaction hamiltonian in the heisenberg representation . integrating formally this equation , and substituting the solution in it , the evolution of the reduced system density matrix @xmath10 can be written as@xmath11\ } , \label{exact}\]]where , as usual , a first order contribution was discarded after assuming @xmath12=0 $ ] . from this evolution , the well known born - markov approximation can be deduced blum , cohen , carmichael , weiss . the born approximation consists into assume , at all times , an uncorrelated structure for the total density matrix@xmath13where @xmath14 define a stationary state of the bath . this assumption is consistent up to second order in the interaction hamiltonian . when the decay of the bath correlation defines the small time - scale of the problem , after introducing eq . ( [ born ] ) in eq . ( [ exact ] ) , the markovian approximation leads to a closed local in time density matrix evolution . we remark that the born - markov approximation does not rely in any specific model of bath dynamics @xcite , such as an infinite set of harmonic oscillators . in fact , its master equation is independent of model assumptions used in its derivation @xcite . now we consider a complex environment for which the previous approximations are not valid . as is usual when dealing with complex environments kuznezov , dcohen , gaspard , kottos , instead of defining the bath hamiltonian @xmath15 as an infinite set of normal modes , here we specify it through its eigenstates basis @xmath16 , which in a weak interaction limit , results unmodified by the interaction with the system . as a central hypothesis , we will assume that , while the full action of the environment can not be described in a markov approximation , it is possible to split the full hilbert space of the bath as a direct sum of subspaces , in such a way that each one defines a sub - reservoir able to induce by itself a markovian system evolution @xcite . these hypotheses are the main assumptions that allow us to formulate our results . in conformity with the splitting condition , we write the interaction hamiltonian as a direct sum of sub - hamiltonians@xmath17where @xmath18 here , each operator @xmath19 defines the interaction between the system and each sub - reservoir @xmath20 in order to describe the joint action of all sub - reservoirs over the system , instead of the uncorrelated form eq . ( [ born ] ) , we introduce the _ _ generalized born - approximation__@xmath21where @xmath22=1,$ ] and we have defined@xmath23@xmath24 is the base of eigenvectors that span the subspace corresponding to each sub - reservoir . therefore , each contribution in eq . ( [ rhotot ] ) consists in the external product between a system state @xmath25 and the projection of the stationary bath state @xmath14 over each subspace @xmath26 . in physical terms , each state @xmath27 takes in account the dissipative effects induced by each sub - reservoir . after introducing eq . ( [ rhotot ] ) in eq . ( [ exact ] ) , we get the approximated evolution@xmath28\ } , \notag \end{aligned}\]]where @xmath29 means a trace operation with the states @xmath24 corresponding to each subspace . furthermore , we have introduced the sub - bath density states @xmath30 where@xmath31the normalization condition @xmath32=1 $ ] implies the relation @xmath33 . thus , the set @xmath34 can be seen as a set of probabilities defined by the weight of each subspace in the total stationary bath state . from eq . ( [ rhotot ] ) we can write@xmath35\approx \sum_{r}p_{r}\rho _ { r}(t ) . \label{sum}\]]then , the evolution eq . ( [ borngeneral ] ) is in fact a linear combination of the evolutions corresponding to the set @xmath36 , each one participating with weight @xmath37 . after introducing the _ markovian approximation _ @xcite to the evolution of each state @xmath25 , in a schrdinger representation , we get@xmath38-\left ( \frac{1}{\hbar } \right ) ^{2}\int_{0}^{\infty } dt^{\prime } \label{markov } \\ & & \mathrm{tr}_{b_{r}}\{[h_{i_{r}},[h_{i_{r}}(-t^{\prime } ) , \rho _ { r}(t)\otimes \rho _ { b_{r } } ] ] . \notag \end{aligned}\]]this evolution correspond to the usual born - markov approximation when considering a bath consisting only of the subset of states @xmath39 and characterized by the stationary state @xmath40 . the system density matrix is defined by the linear combination eq . ( [ sum ] ) . the evolution eq . ( [ markov ] ) , disregarding transients of the order of the sub - bath hamiltonian correlation time , can be always well approximated by a lindblad equation @xcite@xmath41+\gamma _ { r}\mathcal{l}% [ \rho _ { r}(t ) ] , \label{lindblad}\]]where @xmath42=(-i/\hbar ) [ h_{s},\bullet ] $ ] is the system liouville superoperator and the lindblad superoperator is defined by@xmath43=\sum_{\alpha } \frac{1}{2}([v_{\alpha } , \bullet v_{\alpha } ^{\dagger } ] + [ v_{\alpha } \bullet , v_{\alpha } ^{\dagger } ] ) .\]]as the underlying microscopic interaction between the system and the environment is the same in each subspace , the set of operators @xmath44 does not depend on index @xmath26 . nevertheless , each subspace has associated a different characteristic dissipative rate @xmath45 . as this rate arises from the interaction of the system with the manifold of states @xmath24 , consistently with the fermi golden rule @xcite , it is proportional to the characteristic interaction strength of each subspace , denoted as @xmath46 multiplied by the corresponding density of states @xmath47 evaluated in a characteristic frequency @xmath48 of the system , i.e. , @xmath49 . correspondingly , the system state follows from the average@xmath50random rate equations were extensively used to model classical anomalous diffusion processes in disordered media klafter , bernasconi , odagaki , miguel . here , we have derived a similar structure for a different physical situation , i.e. , quantum systems embedded in a complex structured environment . as in a classical context , while the density matrixes @xmath25 follows a markovian evolution , the average state @xmath51 evolve with a non - markovian evolution . this evolution can be easily obtained in a laplace domain , where the average eq . ( [ average ] ) takes the form@xmath52with @xmath53 being the laplace variable , and we have used the solutions @xmath54\rho _ { s}(0).$ ] consistently with the uncorrelated initial condition @xmath55 , there not exist any statistical correlation between @xmath56 and the random rate set . thus , the average evolution can be obtained without appealing to a projector technique klafter , bernasconi , odagaki , miguel . in fact , after introducing in eq . ( rholaplace ) the identity in the form @xmath57\rangle ^{-1}$ ] , it is immediate to get the deterministic , closed , non - markovian evolution equation@xmath58+\int_{0}^{t}d\tau \,% \mathbb{l}(t-\tau ) [ \rho _ { s}(\tau ) ] , \label{nomarkov}\]]where the superoperator @xmath59 is defined in the laplace domain by the equation@xmath60=\langle g_{r}(u)\rangle \mathbb{l}(u)[\bullet ] . \label{memory}\]]depending on the set @xmath61 , eq . ( [ nomarkov ] ) may lead to the presence of strong non - markovian decay behaviors in the system dynamics . this characteristic originates from the entanglement of the system which each sub - reservoir , situation explicitly introduced by eq . ( rhotot ) . an example of complex structured environment where the generalized born - markov applies straightforwardly is a bath hamiltonian whose eigenvectors can be labelled with two indexes @xmath62 . the index @xmath63 is continuous , and for each @xmath26 the corresponding sub - manifold of states is able to induce a different system markovian - decay . the difference between the markovian dynamics may originates in the coupling strength of each sub - manifold with the system . on the other hand , it may originates due to strong variations of the bath density of states with index @xmath26 . the system dynamics follows as a superposition of markovian dynamics whose weights are taken in account through the generalized born approximation eq . ( [ rhotot ] ) . if the decay induced by each sub - manifold is the same , the generalized born approximation reduces to the usual one , and then a markovian evolution is obtained . further examples can be established in the context of random band - matrix models @xcite , where the markovian sub - baths , for example , may be associated to subspaces with a different characteristic bandwidth . classical master equations with random rates are characterized by equations similar to those obtained previously . nevertheless , as in general the underlying numbers of states is infinite , some kind of approximation is necessary in order to obtain the operator @xmath59 , as for example an effective medium approximation @xcite . here , we introduce a similar approximation in order get a general characterization of the dynamics . thus , in eq . ( [ memory ] ) we discard the dependence introduced by the lindblad superoperator @xmath64 in the propagator @xmath65 , i.e. , @xmath66 . then , we get the approximated solution @xmath67 from where it follows the evolution@xmath68+\int_{0}^{t}d\tau k(t-\tau ) e^{(t-\tau ) \mathcal{l}_{h}}\mathcal{l}% [ \rho _ { s}(\tau ) ] . \label{aproximada}\]]in this approximation all information about the random rate is introduced through the kernel function@xmath69as in a classical context , this kernel can be associated with a waiting time distribution @xmath70 and a survival probability @xmath71 defined by@xmath72 in classical master equations , these objects define a continuous time random walk @xcite . in the quantum case , a similar stochastic dynamics can be constructed @xcite . it consists in the application at random times of the superoperator @xmath73 , implying the transformation @xmath74 $ ] , while during the intervals between these disruptive actions the system evolves with its unitary dynamics , @xmath75 $ ] . the intervals between the successive applications of @xmath76 follows from the waiting time distribution @xmath70 . the function @xmath77 defines the corresponding survival probability , @xmath78 . thus , the average over different realizations of the random times can be written as@xmath79 . \label{estocastica } \end{aligned}\]]from here , in a laplace domain , it is straightforward to recuperate the evolution eq . ( [ aproximada ] ) . when @xmath80i , with @xmath81 a completely positive superoperator @xcite , a similar stochastic dynamics can be formulated after introducing a limit procedure @xcite . we remark that the stochastic interpretation [ eq . ( [ estocastica ] ) ] was constructed after associating to the kernel @xmath82 a waiting time distribution and a survival probability , eq . ( [ survival ] ) . this association does not rely in the generalized born - markov approximation , neither it was deduced from a conditional continuous time measurement theory @xcite . therefore , it is not clear if one can associate to the stochastic dynamics a random signal of a measurement apparatus . if this is the case , contradictions between environmental decoherence and wave - function collapse may arise @xcite . as an example of our formalism , in this section we will characterize the dissipative dynamics of a quantum tunneling system described in a two level approximation @xcite and driven by a complex environment . then , the system hamiltonian can be written as @xmath83 the first term , proportional to the @xmath84-pauli matrix @xmath85 define the energy of the effective levels , and the second one , proportional to the @xmath86-pauli matrix @xmath87 , introduce the reversible hopping between the two effective states . the complex environment will be represented by the lindblad superoperator@xmath43=\frac{1}{2}([\sigma _ { z}\bullet , \sigma _ { z}]+[\sigma _ { z},\bullet \sigma _ { z } ] ) , \label{lindblad}\]]and an arbitrary set @xmath88 of random rates and weights . for fixed rate , this superoperator induces a dynamics equivalent to a thermal environment in a high temperature limit @xcite . the evolution of the system density matrix is defined by eqs . ( [ nomarkov ] ) and ( [ memory ] ) . here , we write the evolution in terms of the components of the bloch vector , which are defined by the mean value of the pauli matrixes , @xmath89 , with @xmath90 @xmath91 and @xmath84 . we get [ bloch ] @xmath92thus , the system evolution is completely characterized by three memory kernels @xmath93 , @xmath94 , and @xmath95 . in appendix a , we give the exact expressions of these kernels for arbitrary random rates , joint with the kernels that arise from the effective approximation eq . ( [ aproximada ] ) . from eq . ( [ bloch ] ) it is straightforward to write the system density matrix evolution [ eq . ( [ nomarkov ] ) ] as a sum of lindblad superoperators , each one characterized by a different kernel . when the hopping frequency is zero , @xmath96 , the dynamics reduce to a dispersive one . thus , the coherences decay continuously while the population of each effective level remains constant . in this limit , from appendix a , for arbitrary set @xmath61 we get the exact kernels [ kerneldispersivo ] @xmath97 , \\ \gamma _ { y}(t ) & = & k(t)\cos [ \omega _ { a}t ] , \\ \upsilon ( t ) & = & k(t)\sin [ \omega _ { a}t ] . \end{aligned}\]]where @xmath82 is defined in the laplace domain by eq . ( [ k(u ) ] ) . we note that these kernels also arise from the effective approximation eq . ( aproximada ) , indicating that for @xmath96 , both evolutions coincide . from eqs . ( [ bloch ] ) , the exact solution of the bloch vector is given by [ disperso ] @xmath98s_{x}(0)-\sin [ \omega _ { a}t]s_{y}(0)\},\ \ \ \ \ \ \ \ \\ s_{y}(t ) & = & p_{0}(t)\{\sin [ \omega _ { a}t]s_{x}(0)+\cos [ \omega _ { a}t]s_{y}(0)\},\ \ \ \ \ \ \ \ \\ s_{z}(t ) & = & s_{z}(0 ) , \end{aligned}\]]where @xmath71 is the survival probability defined by its laplace transform eq . ( [ survival ] ) , which in the time domain reads @xmath99.$ ] consistently , we note that the exact solutions eqs . ( [ disperso ] ) correspond to an average over markovian solutions , each one characterized by a rate @xmath45 and participating with weight @xmath37 . depending on the distribution of the dissipation rate , arbitrary forms of the decay can be obtained from this average over exponential functions . hence the non - markovian behavior can be observed in the relaxation of the density matrix to the stationary state . the form of the set @xmath61 depends on the specific structure of the complex environment . here , we will determine this set in a phenomenological way as a function of the system decay behavior . we will be interested in obtaining anomalous decay dynamics such as _ power law_. a possible set consistent with this decay is @xmath100,\;\;\;\;\;\;p_{r}=(1-e^{-a})\exp [ -ar ] , \label{expo}\ ] ] where @xmath101,$ ] @xmath102 scale the random rates , and the constants @xmath103 and @xmath104 measure the exponential decay of the random rates and their corresponding weights . with these definitions , it is simple to demonstrate that after a transient of order @xmath102 , the waiting time distribution and its associated survival probability , eq . ( [ survival ] ) , present a power law decay behavior @xcite , @xmath105 , and @xmath106 where @xmath107 . clearly , this behavior is reflected in the system dynamics . , @xmath108 , @xmath109 , @xmath110 , @xmath111 , and the markovian limit @xmath112 . in all cases we take @xmath113.,height=272 ] when @xmath114 , the kernel @xmath82 corresponding to the set eq . ( expo ) can be well approximated by the expression@xmath115with the definitions@xmath116 the scaling of these parameters can be motivated by considering a two dimensional set of random rates @xcite . from eq . ( [ k(u ) ] ) and ( [ survival ] ) , the waiting time distribution and its associated survival probability can be obtained as @xmath117from here , is it is simple to proof that @xmath118 is a completely monotone function @xcite , which implies that @xmath71 decays in a monotonous way or equivalently , @xmath119 . in figure 1 we plot the survival probability @xmath71 by assuming the kernel eq . ( [ kernel ] ) for different values of @xmath120 . we note that in a short time regime , the decay is an exponential one , while in an asymptotic regime a power law behavior is present@xmath121,\ \ \ \ \ \ \ \ \ \ p_{0}(t)\simeq \frac{% \beta ^{1-\alpha } } { \gamma \gamma ( 1-\alpha ) } \frac{1}{t^{\alpha } } , \]]where @xmath122 is the gamma function . these asymptotic behaviors follows immediately from eq . ( [ sur ] ) . when the dispersion of the random rate @xmath123 is zero @xmath124 , consistently the dynamics reduce to a markovian one , @xmath125 , which implies the pure exponential decay @xmath126 $ ] and @xmath127 $ ] . in the next subsection we will characterize the tunneling dynamics by assuming a complex environment characterized by the random rate set eq . ( expo ) or equivalently by the kernel eq . ( [ kernel ] ) . here we will analyze the tunneling dynamics for a symmetric case @xmath128 , which arise when the two effective levels have the same energy . from appendix a , the exact kernels read [ kernelhoping ] @xmath129as before , the kernel @xmath130 is defined by eq . ( [ k(u ) ] ) . the exact solution of the bloch vector can be obtained in a laplace domain . we get [ hoping ] @xmath131s_{z}(0)+\delta s_{y}(0)\},\ \ \ \ \ \ \end{aligned}\]]where we have defined @xmath132 which can also be expressed as @xmath133 . in this case it is not possible to find in the time domain a general exact solution for arbitrary memory kernels . a simple analytical solution is only available in a markovian case [ @xmath125 ] [ markovian ] @xmath135 \notag \\ & & -\lambda ^{-1}[(\gamma /2)s_{y}(0)+\delta s_{z}(0)]\sinh [ \lambda t],\ \ \ \ \ \\ s_{z}(t ) & = & e^{-\gamma t/2}\{s_{z}(0)\cosh [ \lambda t ] \notag \\ & & + \lambda ^{-1}[(\gamma /2)s_{z}(0)+\delta s_{y}(0)]\sinh [ \lambda t],\ \ \ \ \ \end{aligned}\]]where @xmath136 , and @xmath137 defines the unique dissipative rate . notice that in the limit of null dissipation , a periodic hopping between the effective levels is obtained . considering as initial condition the upper eigenstate of @xmath138 . the envelopes are given by @xmath139 . from top to bottom , the parameters are @xmath140 , @xmath141 , and @xmath142 . in all cases we take @xmath143 @xmath144 , and @xmath145,height=257 ] for arbitrary random rates @xmath61 , the dynamics can be characterized in different regimes . first , in the case @xmath146 , from eqs . ( [ hoping ] ) it is possible to get the approximated solutions [ aproxhoping ] @xmath147s_{y}(0)-\sin [ \delta t]s_{z}(0)\},\ \ \ \ \ \ \ \\ s_{z}(t ) & \simeq & p_{0}(t/2)\{\sin [ \delta t]s_{y}(0)+\cos [ \delta t]s_{z}(0)\}.\ \ \ \ \ \ \ \end{aligned}\]]thus , the dynamics consist in a periodic tunneling between the two effective states , and whose decay can be written in terms of the survival probability . as in the previous case , this solution correspond to an average over the corresponding markovian solutions , i.e. , eq . ( [ markovian ] ) written in the limit of small decay rate when compared to the tunneling frequency @xmath148 . in figure 2 we plot the average of the @xmath84-pauli matrix which follows from eq . ( [ hoping ] ) with the kernel eq . ( [ kernel ] ) . as initial condition we take the upper eigenstate of @xmath85 . we verified that the exact solutions are well described by the approximation eq . ( [ aproxhoping ] ) for parameters values satisfying @xmath149 . as the envelope decay is given by @xmath150 , by increasing the average rate @xmath151 , the dynamics decay in a faster way . this dependence is broken when the average rate is much greater than the hopping frequency . in the limit @xmath152 , the dissipative dynamics dominates over the tunneling one . in figure 3 we plot @xmath153 [ eq . ( [ hoping ] ) ] for different values of the characteristic parameters of the kernel eq . ( [ kernel ] ) . we note that by increasing the average rate @xmath137 , a slower decay is obtained . thus , the dynamics develops a zeno - like effect @xcite . from the exact solution eq . ( hoping ) , the characteristic decay of the bloch vector can be approximated by the expressions @xmath154 where we have introduced @xmath155 and the function @xmath156^{-1}\simeq u^{-1}p_{0}(\delta ^{2}/u)$ ] . for the kernel defined by eq . ( [ kernel ] ) , the characteristic decay @xmath157 results @xmath158 as can be seen in figure 3 ( dotted line ) , besides the oscillatory behavior , after the transient @xmath159 , this function provides an excellent fitting of the decay dynamics . considering as initial condition the upper eigenstate of @xmath138 . the fitting decay curves ( dotted lines ) are given by eq . ( [ zenodecay ] ) . from top to bottom , the parameters are @xmath160 , @xmath161 , @xmath162 , @xmath163 , @xmath164 , and @xmath165 . in all cases we take @xmath143 @xmath166 and @xmath145 in the inset we show the same graphic in a log - log scale.,height=279 ] the function @xmath167 is characterized by a reach variety of behaviors . first , we note that in the markovian limit , @xmath168 , we get an exponential decay with rate @xmath169 , which clearly diminish by increasing @xmath137 . in the non - markovian case , in a short time regime , we can approximate@xmath170\},\]]while in an asymptotic long time limit we get@xmath171thus , the dispersion of the random rate ( measured by @xmath172 ) induce , at short times , an extra stretched exponential decay , while in the asymptotic regime it scales a power law behavior [ @xmath173 . the characteristic rates of both regimes arise from a competence between the unitary and dissipative dynamics . we notice that by increasing the dispersion rate @xmath172 , the characteristic rate of the stretched exponential decay is increased , while the rate for the power law regime is decreased . the dependence in the hopping frequency @xmath174 is the inverse one . the zeno - like effect can be _ qualitatively _ understood in terms of the stochastic evolution corresponding to the effective approximation eq . ( [ aproximada ] ) . this stochastic process develops in the system hilbert space and consists in the application at random times of the superoperator @xmath73 , which in view of eq . ( [ lindblad ] ) reads @xmath175=\sigma _ { z}\bullet \sigma _ { z}$ ] , while in the intermediates times the system evolves with its unitary evolution @xmath176.$ ] the superoperator @xmath81 implies the disruptive transformations @xmath177 , @xmath178 , @xmath179 , while the unitary dynamics is equivalent to a rotation around the @xmath180-direction . in the limit of vanishing hopping frequency @xmath174 , the continuous applications of the superoperator @xmath76 kill the @xmath180-@xmath181 components and frozen the dynamics in the initial condition @xmath182 . thus , a pure zeno effect is recuperated . for @xmath183 the decay dynamics is determined from the competence between the transformations induced by @xmath81 and @xmath184 , defining the zeno - like regime . this interpretation is exact in a markovian limit and always valid for the effective master equation eq . ( aproximada ) . in obtaining the previous results we have assumed an infinite set of random rates , eq . ( [ expo ] ) , whose effects can be approximated by the kernel eq . ( [ kernel ] ) . while this election guarantees the presence of an asymptotic power law decay , strong non - exponential behaviors can be obtained in an _ intermediate regime _ by considering only a finite set , @xmath185 , of random rates @xmath61 . on the other hand , for a finite set , the asymptotic system dynamics is always markovian and characterized by the inverse rate @xmath186 . this result follows from @xmath187 in figure 4 we show the decay dynamics induced by an environment characterized by a finite set of random rates @xmath45 ( @xmath188 ) with equal weights , @xmath189 . each curve follows from a superposition of markovian solutions , eq . ( [ markovian ] ) with @xmath190 . the set of rates @xmath191 of each plot differ in a multiplicative factor , in such a way that the relation between the average rate @xmath192 and the corresponding fluctuation rate @xmath193/\langle \gamma _ { r}\rangle $ ] remains constant in all curves . for the case @xmath194 the random rates are @xmath195 , @xmath142 , @xmath196 , @xmath197 , @xmath198 , @xmath199 and @xmath200 . considering as initial condition the upper eigenstate of @xmath201 and a finite set of random rates with equal weights . from top to bottom , the parameters are @xmath202 , @xmath163 , @xmath203 , @xmath204 , @xmath205 , and @xmath206 . in all cases we take @xmath207 , @xmath208 and @xmath145 for @xmath209 the dotted lines correspond to the fitting @xmath210 $ ] , while for @xmath211 they corresponds to @xmath212^{-\protect\delta } $ ] ( see text).,height=264 ] in order to enlighten the intermediate non - exponential regime , we have plotted the shifted average @xmath213 , with @xmath214 . in the deep zeno - like regime [ @xmath215 , @xmath216 can be well approximated by an stretched exponential behavior @xmath217 $ ] , with @xmath218 and @xmath219 . for @xmath220 , a power law fitting is more adequate @xmath221^{-\delta } , $ ] with @xmath222 and @xmath223 . we note that a similar non - exponential fitting was found in ref . @xcite by considering the action of a finite bath , which can be associated with a glassy environment . on the other hand , the oscillatory effects in the decay of @xmath216 arise from the markovian solutions , eq . ( markovian ) , corresponding to the rates satisfying @xmath224 . in fact , for the markovian solution , this condition delimits the change between a monotonous and an oscillatory decay behavior . consistently , we notice that by increasing the average rate , the amplitude of the oscillations are smaller . a similar effect can be seen in figure 3 . our formalism relies on the applicability of the generalized born - markov approximation . here we explore the possibility of mapping its dynamics with other models that also induce anomalous decay behaviors . _ spin - boson model _ : the spin - boson model is defined by the total hamiltonian @xmath225where the bath hamiltonian @xmath226 $ ] corresponds to a set of harmonic oscillators , and @xmath227 . the bath is characterized by the spectral density function@xmath228and assumed to be in equilibrium at temperature @xmath229 . as is well known , the reduced system dynamics can be obtained in an exact way leggett , grifoni , lang , hanggi , milena . it reads [ spin - boson ] @xmath230\ \ \ \ \ \ \ \\ & & + y_{b}^{(s)}(t)s_{x}(0)+y_{b}^{(a)}(t)s_{y}(0 ) , \notag \\ s_{y}(t ) & = & \frac{1}{\delta } \frac{ds_{z}(t)}{dt } , \\ \frac{ds_{z}(t)}{dt } & = & \int_{0}^{t}d\tau \lbrack k_{a}^{(a)}(t-\tau ) -k_{a}^{(s)}(t-\tau ) s_{z}(\tau ) ] \ \ \ \ \ \ \ \notag \\ & & + k_{b}^{(a)}(t)s_{x}(0)+k_{b}^{(s)}(t)s_{y}(0)\ } , \end{aligned}\]]where the corresponding kernels can be written as functions of @xmath231 . on the other hand , it is possible to write the exact averaged evolution eq . ( [ bloch ] ) in the form eqs . ( [ spin - boson ] ) . in appendix b we present the kernels corresponding to each dynamics . from these expressions , it is simple to proof that to first order in @xmath174 , after disregarding a phase factor , both set of kernels can be mapped under the condition @xmath232=\exp [ -q^{\prime } ( t ) ] , \label{map}\ ] ] where@xmath233,\]]define the real part of the thermal bath correlation . we remark that the mapping eq . ( [ map ] ) is only valid in a high temperature limit , condition consistent with the lindblad structure eq . ( [ lindblad ] ) . in this context , from eq . ( [ map ] ) , it is possible to enlighten the difference between the present approach and that developed in refs . tannor , ulrich . in our approach , which relies in splitting the hilbert space of the bath as a _ direct sum _ of subspaces , @xmath234 $ ] is written as a sum of exponential functions , each one associated to each markovian sub - reservoir . instead , in refs . @xcite , @xmath235 is expressed as a sum of exponential functions . this representation relies in an artificial discomposing of the spectral density function @xmath231 as a sum of individuals terms . thus , the hilbert space of the bath is effectively split in an _ external product _ of subspaces , each one associated to a non - markovian sub - reservoir . as in our approach , the system density matrix can be written in terms of a set of auxiliary sub - density - matrixes . nevertheless , their evolution involves coupling among them all . _ stochastic hamiltonian _ : decoherence in small quantum systems is also modeled by introducing stochastic elements in the system evolution . this situation arises naturally in many physical systems makhlinreport , makhlin , falci , adrian . consistently with the spin - boson model we consider the stochastic hamiltonian@xmath236\sigma _ { z}+\delta \sigma _ { x}\},\]]where @xmath237 is a classical non - white noise term . by assuming @xmath238 , where @xmath239 means an average over realizations of the noise , in the limit of vanishing @xmath174 it is simple to solve the stochastic dynamics and obtain the average of the pauli matrixes . the final evolution is the same as in eq . ( [ disperso ] ) after replacing @xmath71 with the average dephasing factor @xmath240\rangle \rangle _ { \xi } $ ] . thus , the generalized born - markov approximation can be mapped to the stochastic hamiltonian evolution under the condition @xmath241 , which explicitly reads @xmath232=\big\langle\big\langle\exp [ i\int_{0}^{t}d\tau \xi ( \tau ) ] \big\rangle\big\rangle_{\xi } .\]]this condition can be consistently satisfied if the dephasing factor @xmath242 decays in a monotonous way . we have presented a theoretical approach intended to describe the dynamic of small quantum systems interacting with a complex structured environment . our formalism is based in an extension of the well known born - markov approximation , which relies in the possibility of splitting the environment as a direct sum of sub - reservoirs , each one being able to induce by itself a markovian system dynamics . then , we have demonstrated that the full action of the complex environment can be described through a random lindblad master equation . the set of random rates follows from a fermi golden rule . thus , they are proportional to the characteristic coupling strength of each subspace multiplied by the corresponding sub - density of states evaluated in a characteristic frequency of the system . the associated probabilities are defined by the weight of each subspace in the stationary state of the bath . from a phenomenological point of view , the set of random rates and weights can be determined in a consistent way in function of the system decay . in fact , the system dynamics is characterized by a non - markovian master equation that in function of the random rate set can develop strong non - exponential decays . as an example we worked out the dissipative dynamic of a quantum tunneling system in a two level approximation . we have introduced a set of random rates that lead to the presence of asymptotic power law decay . in the limit of small hopping frequency , when compared with the average rate , we have showed that a zeno - like phenomenon arises , which is characterized by a stretched exponential and a power law decay . these behaviors follow from the interplay between the unitary dynamics and the entanglement - memory - effects induced by the reservoir . for the tunneling dynamics , we have also demonstrated that non - exponential decays arise even by considering a small set of random rates . furthermore , we have established the conditions under which the random lindblad evolution can be mapped to a spin - boson model and a stochastic hamiltonian evolution . finally , we want to emphasize that the present results define a new framework for describing anomalous quantum system dynamics , which consists in taking the characteristic rate of a lindblad equation as a random distributed variable . we remark that this approach was not derived from an ensemble of identical systems whose local interactions with the environment can be approximated by different markovian evolutions . in fact , the underlying microscopic physics can be related to a _ single _ quantum system coupled to an environment with a complex structured spectral density function and whose dynamical influence over the system can be approximated by a direct sum of markovian sub - reservoirs . thus , our approach may be relevant for the description of anomalous decay processes in individual mesoscopic systems embedded in a condensed phase environment schlegel , brokmann , grigolini . a natural example for which the generalized born - markov approximation may applies are glassy reservoirs , where the underlying configurational disorder produce a hierarchical distribution of coupling strength between the single system and the corresponding localized eigenstates of the reservoir @xcite . here we present the exact expressions for the kernels @xmath243 , @xmath244 , and @xmath245 that define the evolution of the pauli operators average , eqs . ( [ bloch ] ) . for arbitrary rates @xmath246 , the kernels read [ exactkernels ] @xmath247(u+b)+u\omega _ { a}^{2}\},\ \ \ \ \ \\ \gamma _ { y}(u ) & = & d\{[u(u+b)+\delta ^{2}](u+c)+u\omega _ { a}^{2}\},\ \ \ \ \ \\ \upsilon ( u ) & = & d(b - c)u\omega _ { a},\ \ \ \ \ \end{aligned}\]]where @xmath248 denotes the function @xmath249+\delta ^{2}\}[u+b(u)]+u\omega _ { a}^{2}}.\ ] ] the extra functions @xmath250 and @xmath251 are defined by@xmath252where we have introduced@xmath253(u+\gamma _ { r})+u\omega _ { a}^{2}}% .\]]using that the laplace transform of @xmath254 is given by @xmath255 , in the case @xmath96 it is possible to recuperate the expressions of section iii - a , eqs . ( [ kerneldispersivo ] ) . on the other hand , taking @xmath256 it is straightforward to get the results of section iii - c , eqs . ( [ kernelhoping ] ) . in an effective approximation , ( [ aproximada ] ) , the corresponding kernels read @xmath257\ } , \\ \gamma _ { y}(t ) & = & k(t)\cos [ \varphi t ] , \\ \upsilon ( t ) & = & k(t)\frac{\omega _ { a}}{\varphi } \sin [ \varphi t ] , \\ \phi _ { x}(t ) & = & k(t)\frac{\omega _ { a}\delta } { \varphi ^{2}}\{1-\cos [ \varphi t]\ } , \\ \phi _ { y}(t ) & = & k(t)\frac{\delta } { \varphi } \sin [ \varphi t ] , \end{aligned}\ ] ] where @xmath258 the extra kernels @xmath259 and @xmath260 couples the derivative of @xmath153 to the averages @xmath261 and @xmath262 respectively , i.e. , @xmath263 for the exact evolution , these kernels vanish . the kernels of the spin - boson model eqs . ( [ spin - boson ] ) , in lowest order in @xmath174 read @xcite [ kernelspinboson ] @xmath264 , \\ k_{a}^{(a)}(t ) & \simeq & \delta ^{2}y_{b}^{(a)}(t)\sin [ q^{\prime \prime } ( t ) ] , \\ k_{a}^{(s)}(t ) & \simeq & \delta ^{2}y_{b}^{(s)}(t)\cos [ q^{\prime \prime } ( t ) ] , \\ y_{a}^{(a)}(t ) & \simeq & -\delta y_{b}^{(a)}(t)\cos [ q^{\prime \prime } ( t ) ] , \\ k_{b}^{(s)}(t ) & \simeq & \delta y_{b}^{(s)}(t ) , \\ k_{b}^{(a)}(t ) & \simeq & -\delta y_{b}^{(a)}(t ) , \\ y_{b}^{(s)}(t ) & \simeq & \cos [ \omega _ { a}t]e^{-q^{\prime } ( t ) } , \\ y_{b}^{(a)}(t ) & \simeq & -\sin [ \omega _ { a}t]e^{-q^{\prime } ( t ) } , \end{aligned}\]]where @xmath265 and @xmath266 are defined by [ kerneldisorder ] @xmath268 @xmath269 , \\ y_{a}^{(a)}(u ) & = & t(u)\delta \lbrack \omega _ { a}-\upsilon ( u ) ] , \\ y_{b}^{(s)}(u ) & = & t(u)\delta \lbrack u+\gamma _ { x}(u ) ] , \\ k_{b}^{(a)}(u ) & = & t(u)\delta \lbrack \omega _ { a}-\upsilon ( u ) ] , \\ y_{b}^{(s)}(u ) & = & t(u)[u+\gamma _ { y}(u ) ] , \\ y_{b}^{(a)}(u ) & = & -t(u)[\omega _ { a}-\upsilon ( u ) ] , \end{aligned}\]]where we have introduced the structure of these kernels is the same as those of the spin - boson model in the limit of vanishing @xmath174 , which implies that @xmath174 only appears through the unitary evolution . in fact , in this limit we can approximate eqs . ( [ exactkernels ] ) by @xmath271/2 $ ] and @xmath272/2i$ ] . after introducing these expressions in eqs . ( [ kerneldisorder ] ) , it is simple to get @xmath273p_{0}(t)$ ] , and @xmath274p_{0}(t)$ ] . then , disregarding in eqs . ( [ kernelspinboson ] ) the phase contribution proportional to @xmath275 , which is valid in a high temperature limit wilkie , a mapping with eqs . ( [ kerneldisorder ] ) can be done after imposing the equality @xmath276 . the stochastic dynamics can be formally extended to the case @xmath80i @xcite after introducing the superoperator @xmath277=\{$]i@xmath278i@xmath279\}\rho $ ] , where @xmath280 must be intended as a control parameter . then , the evolution eq . ( [ aproximada ] ) is recuperated in the limit in which simultaneously @xmath281 and the number of events by unit of time go to infinite , the last limit being controlled by the sprinkling distribution @xmath282 @xcite . this object is defined as the probability density for an event at time @xmath283 , disregarding the possibility of extra events in @xmath284 . therefore , it is given by @xmath285 where @xmath286 is the step function [ here defined as @xmath287 for @xmath288 , and @xmath289 for @xmath290 , which implies @xmath291.$ ] in the laplace domain it reads @xmath292 $ ] , which allows to write @xmath293 . when the random rate assumes only two different values @xmath294 , with probabilities @xmath295 , the exact kernel reads @xmath296 $ ] where @xmath297 $ ] . the rate @xmath172 is defined in terms of the fluctuations @xmath193/\langle \gamma _ { r}\rangle .$ ] the extra rates are @xmath298 , and @xmath299
in this paper we demonstrate that lindblad equations characterized by a random rate variable arise after tracing out a complex structured reservoir . our results follows from a generalization of the born - markov approximation , which relies in the possibility of splitting the complex environment in a direct sum of sub - reservoirs , each one being able to induce by itself a markovian system evolution . strong non - markovian effects , which microscopically originate from the entanglement with the different sub - reservoirs , characterize the average system decay dynamics . as an example , we study the anomalous irreversible behavior of a quantum tunneling system described in an effective two level approximation . stretched exponential and power law decay behaviors arise from the interplay between the dissipative and unitary hopping dynamics .
broad absorption line quasi stellar objects ( bal qsos ) form a rare class comprising @xmath2% of the qso population at moderate to high redshift ( foltz et al . the standard view , based largely on estimates of upper limits to resonant scattering by the bal clouds , is that such clouds have a small covering factor as seen from the qso nucleus , implying that essentially all radio - quiet qsos would be classified as bal qsos if observed from the proper angle . bal qsos appear to be found more frequently among those with enhanced far - ir emission , strong emission , and/or weak or absent [ ] narrow emission lines ( low et al . 1989 ; boroson & meyers 1992 ; turnshek et al . while the strong emission might be due to an orientation effect , it seems unlikely that the other properties mentioned would have a strong dependence on viewing angle . one interpretation of the bal phenomenon running counter to the standard unified view is that these qsos are very young and are ejecting their gaseous envelopes at very high velocities following the initial turn - on of the agn ( hazard et al . in particular , the high incidence of low - ionization bal qsos in iras - selected , luminous qsos , along with the reddening from dust in these objects , suggests that this subclass , at least , may comprise transition objects between ultraluminous infrared galaxies and the classical qso population ( lpari 1994 , voit et al 1993 , egami et al . 1996 ) . in principle , characterization of the host galaxies of sufficiently large samples of normal radio - quiet qsos and bal qsos might tell us whether the bal phenomenon is mostly due to orientation or mostly due to intrinsic properties . here we report the chance discovery of a bal qso at moderately high redshift , for which hst imaging already exists . we found the qso in the field of the quasar 3c48 as part of a keck spectroscopic program studying the host galaxies of low redshift qsos ( canalizo & stockton 1997 , 1998 ) . the only previous mention of this object in the astronomical literature is its identification as a `` star '' by kirhakos et al . it is because of its proximity to 3c48 that it is present on hst wfpc archival images , and , as we shall describe , it does show some extended emission . spectroscopic observations of the host galaxy of 3c48 were carried out on 1996 october 13 with the low - resolution imaging spectrometer ( lris ) on the keck ii telescope . we used a 600 groove mm@xmath0 grating blazed at 5000 to obtain a dispersion of 1.24pixel@xmath0 and a useful wavelength range of 46007200 . one of our slit positions happened to include a stellar object to the south of 3c48 . the slit ( 1 wide , projecting to @xmath35 pixels on the tektronix 2048@xmath42048 ccd ) was centered @xmath315 e of 3c48 , with a pa of 14 , as indicated on fig . the spectrum of what we will refer to as the bal qso 0134 + 3253 fell 41 below that of the host galaxy of 3c48 . the total exposure time was 60 minutes . the spectra were reduced with iraf , using standard reduction procedures . after dark , bias and flat field correction , the images were background subtracted to remove sky lines . wavelength calibrations were made using the least - mean - squares fit to identified lines in a comparison spectrum . the spectra were flux calibrated using spectrophotometric standards from massey _ et al . _ the distortions in the spatial coordinate were removed with the _ apextract _ routines . wfpc2 images of the 3c48 field were obtained from the hst data archive . the images used in this analysis included two 1400 s exposures in the f555w filter and two 1700 s exposures in the f814w filter . in each case , 3c48 was centered on the pc1 detector and the bal qso fell near pixel coordinates ( 308 , 244 ) in the wfc2 detector . most cosmic rays were removed by subtracting one of the frames ( `` image 2 '' ) from the other frame obtained with the same filter ( `` image 1 '' ) , then thresholding the difference at a 3@xmath5 level , setting points above this threshold to the median of the difference image . pixels near the position of the peak of the qso were excluded from this process . the corrected difference image was then added back to image 2 , giving a corrected version of image 1 . the few cosmic rays within the relevant region that escaped this process were removed manually with the iraf task _ imedit_. the procedure was repeated , interchanging the images , to correct image 2 , and the two corrected images for each filter were averaged . in order to study extended luminous material around the qso , we subtracted a model point - spread function ( psf ) . this psf was based on a star of similar brightness to the qso , located 15 away . while the wfpc2 shows significant field variations in the psf , at the detection level of the qso features of the psf beyond a radius of 04 are insignificant compared with the noise . we found we could achieve a good match simply with a slight adjustment of the ellipticity of the psf model . finally , the field including 3c48 and the bal qso was imaged with a notched @xmath6 filter on 1997 october 2 with the uh quirc camera ( hodapp et al . 1996 ) on the university of hawaii 2.2 m telescope . the total exposure time was 50 minutes , and we used standard ir reduction procedures . the field of the qso is shown in fig . 1 . the redshift of the qso , determined from the and ] broad emission lines , is @xmath7 = 2.169 . the coordinates , measured from the hst archival images , are @xmath8 and @xmath9 . from these same images , we obtain @xmath10 and @xmath11 . kirhakos et al . ( 1994 ) give @xmath12 . figure 2 shows the observed frame spectrum of the bal qso . the only trough evident in this region is , which shows at least three components and outflow velocities extending to @xmath1312000 km s@xmath0 . the trough is superposed on the emission line , and the red edge is around + 1,000 km s@xmath0 , as can be seen clearly in fig . 2 . unfortunately , @xmath14 did not fall within our observed range , so we can not determine whether this is a low - ionization bal qso . weymann et al . ( 1991 ) found that the continua of low - ionization bal qsos are substantially redder than those of high - ionization bal qsos and non - bal qsos , presumably due to higher amounts of dust ( sprayberry & foltz 1992 ) . this bal qso appears reddened with respect to the faint object spectrograph composite qso spectrum ( zheng et al . 1997 ) , but this reddening could be due to the intervening metal - line absorption system described in 3.2 . there are hints of and emission superposed on ] + , and a possible weak trough bluewards of this feature , both of which are common in low - ionization bal qsos . figure 2 shows several unresolved absorption features , which we identify as arising from an intervening low - ionization metal - line system at @xmath15 . table 1 gives the line identifications and measured equivalent widths . the absorption features are unusually strong , with al iii and fe ii rest - frame equivalent widths being of strength similar to or greater than that seen in the relatively rare damped ly@xmath1 systems ( eg . lu et al . 1996 ) . we have performed a single - component curve - of - growth analysis on the five lines ( @xmath16 km s@xmath0 ) , yielding an estimated column density of log @xmath17()@xmath18 @xmath19 . assuming a solar metallicity ( log @xmath20 , anders & grevesse 1989 ) , this implies the absorber has an column density of @xmath17()@xmath21 @xmath19 . this is likely to be a damped ly@xmath1 absorber since the metallicity of such systems is typically @xmath22% solar ( pettini et al . furthermore , the fact that fe depletes readily onto dust in h i gas and our selection of a conservatively high doppler value in the curve - of - growth analysis both add to the likelihood that this column density estimate is low by a factor of at least 10 . low - ionization qso metal - line absorption systems have been identified with the gaseous halos of luminous galaxies ( bergeron & boiss 1991 ; steidel 1993 ) with large absorption cross sections , @xmath23 kpc ( @xmath24 ) . the damped absorbers , however , are believed to arise from the inner disk regions ( few 10s of kpc ) of the same absorbers based on their column density ( wolfe et al . 1986 ) , mass density ( wolfe 1987 ) , metallicity , and asymmetric absorption profile kinematics ( prochaska & wolfe 1997 ) . attempts to image high - redshift damped ly@xmath1 absorbers have met with limited success , presumably due in part to the close proximity of the intervening system to the qso image . high - spatial - resolution _ hst _ images have recently revealed candidate galaxies within a few arcsec of the qso for several moderate - redshift ( @xmath25 ) damped ly@xmath1 absorbers ( le brun et al . 1997 ) , but the absorbers appear to be morphologically diverse . the interpretation of extended structure associated with this qso is therefore likely to be confused if indeed the intervening @xmath15 system arises from a damped ly@xmath1 absorber lying very near the qso line of sight . figure [ fuzz ] shows the extent and distribution of the faint , extended luminous material around the qso , from the hst archival images . in both the f555w and f814w images , the overall impression is of a roughly elliptical morphology , with a major axis at position angle @xmath26 , and a center offset slightly to the sw from the qso . there is evidence for a discrete object , or at least a luminosity peak , @xmath27 ne of the qso , present in both images but most obvious in the f814w image . the southern edge of the ellipse appears to form an arc - like feature ( fig . [ fuzz]@xmath28 ) , but this impression is dependent on only a few crucial pixels and could easily result from noise fluctuations . by interpolating over the central 04 region of our psf - subtracted images ( fig . [ fuzz]@xmath28 and @xmath29 ) , we can obtain a conservative estimate of the magnitudes of the extended component alone . we find @xmath30 and @xmath31 in 26-diameter apertures , where the errors include only measurement uncertainties . if the luminosity were due to stars at the redshift of the possible damped absorber ( @xmath32 ) , the slope of the spectral - energy distribution at rest - frame 20003000 , compared with bruzual & charlot ( 1997 ) models , indicates that it would likely be dominated by a stellar population with an age of @xmath33 years for solar metallicity , or roughly twice this age for 20% of solar metallicity . on the other hand , if we are observing stars in the qso host galaxy , at @xmath34 , the corresponding ages of the dominant population at rest - frame 18002500 would be @xmath35 and @xmath36 years , respectively . we do not detect any extension in our deep , ground - based @xmath37 image . this image has a 1-@xmath5 detection threshold of @xmath38 erg s@xmath0 @xmath19 @xmath0 arcsec@xmath0 . for a 2-@xmath5 detection , our surface - brightness limit ( in @xmath39 ) is essentially the same as that for the f814w filter , although we lose some additional detection efficiency because we are looking in the wings of the qso profile . in any case , our failure to detect the extension in the ir is consistent with the stellar populations inferred above and would be inconsistent with the presence of any additional red component that would dominate either of these populations by a factor of @xmath40 or more in the observed ir ( rest - frame 60007000 ) . thus , we can not distinguish between an intervening or host - galaxy origin for the extended material in terms of our current knowledge of its sed . a galaxy at @xmath32 dominated by a @xmath33-year - old population would not be sufficiently unusual to constitute an objection to an intervening origin . we have been forced by our low s / n to deal with average colors , but inspection of fig . [ fuzz ] indicates that the peak or discrete object just ne of the qso may be somewhat redder than the rest of the extended material . it is tempting to suggest that this object may be responsible for the intervening system , and that the more diffuse extended material may be the qso host galaxy ; but this suggestion is little more than speculation at this stage . this research was partially supported by nsf under grant ast95 - 29078 . k.c.r . acknowledges support provided by nasa through the hubble fellowship grant # hf-01076.01 - 94a awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc . , for nasa under contract nas 5 - 26555 . lrrr 4947.9 & al iii 1854.72 & 1.13 & 0.424969.9 & al iii 1862.79 & 0.76 & 0.286252.9 & fe ii 2344.21 & 2.48 & 0.936333.7 & fe ii 2374.46 & 1.43 & 0.546355.2 & fe ii 2382.76 & 2.99 & 1.126899.2 & fe ii 2586.65 & 3.29 & 1.236935.3 & fe ii 2600.17 & 2.99 & 1.12[table : eqws ]
we report the serendipitous discovery of a bal qso at z = 2.169 , located 41 southwest of 3c48 . we present keck lris spectroscopy covering rest frame 1500 to 2300 . the civ bal has three components and it extends to outflow velocities of at least 12,000 km s@xmath0 . the bal qso has an intervening low - ionization metal - line absorption system at z = 1.667 which is likely to be a damped ly@xmath1 absorber . hst images show extended luminous material around the qso , which could be either the host galaxy or the intervening system .
during the last two decades there has been much interest in using the electron spin in electronic devices . this research field , often referred to as spintronics , has already made great impact on metal - based information storage systems . there are hopes that a similar success can also be achieved in semiconductor based systems @xcite . manipulating the spins of the electrons via external magnetic fields over nanometer length scales is not considered feasible . another , more attractive , method is to use electric fields to manipulate electron spins via spin - orbit interaction . the spin - orbit interaction arises from the fact that an electron moving in an external electrical field experiences an effective magnetic field in its own reference frame , that in turn couples to its spin via the zeeman effect@xcite . in condensed matter systems , the spin - orbit interaction is found in crystals with asymmetry in the underlying structure @xcite . in bulk this is seen in crystals without an inversion center ( e.g zincblende structures ) and is termed the dresselhaus spin - orbit interaction@xcite . on the other hand the structural asymmetry of the confining potential in heterostructures gives rise to the so called rashba term@xcite . the rashba interaction has practical advantages in that it depends on the electronic environment of the heterostructure which can be modified in sample fabrication and in - situ by gate voltages@xcite . this results in the possibility of varying the spin - orbit interaction on the nanometer scale . interestingly , even structurally symmetric heterostructures can present spin - orbit interaction provided that coupling between subbands of distinct parities is allowed@xcite . the spin - orbit strength can be measured in a variety of different experimental setups@xcite : in a magnetoresistance measurement via shubnikov - de haas oscillations@xcite , weak ( anti- ) localization @xcite , or electron spin resonances in semiconducting nanostructures@xcite or quantum dots@xcite , or optically via spin relaxation @xcite , spin precession @xcite , spin - flip raman scattering @xcite , or radiation - induced magnetoresistance oscillations @xcite . in this paper we propose a method for extracting the strength of the rashba spin - orbit interaction via a charge conductance measurement . this method does not require the use of external magnetic fields or radiation sources . we consider a quantum wire modulated by an external periodic potential , fig . [ fig : schematicpic ] . essentially , the method relies on the fact that the rashba - induced shifts of the band gap positions in energy dramatically alter the charge conductance of the superlattice . via direct diagonalization , we determine the band structure of an infinite parabolically confined quantum wire . the energy bands clearly show band gaps that are renormalized by the rashba interaction . interestingly , the band gaps shift in energy as the strength of the rashba interaction is varied . the location of the band gaps are at the crossing of energy bands from adjacent brillouin zones . these energy bands can be calculated via an analytical approximation scheme@xcite . using non - equilibrium green s functions ( negf ) and the landauer formalism we calculate the charge conductance through a finite region containing both a periodic potential and the rashba interaction . in the conductance we find several dips appearing at different location in energy . moreover , these conductance dips coincide with renormalized band gaps of the superlattice . as with the band gaps the positions of the conductance dips are at the crossing points of the energy bands from next neighbor brillouin zones . for a wide range of rashba spin - orbit interaction strengths some of the conductance dips shift linearly in energy as function of the strength of the rashba coupling . for this range we derive a relation , eq . ( [ eq : linearealpha ] ) , describing the location of linearly shifting dips in energy as a function of rashba interaction strength . the rashba coupling can therefore be extracted by fitting the shift of conductance dips via eq . ( [ eq : linearealpha ] ) . figure 1 shows a schematic of the proposed experimental setup . this paper is organized as follows . in sec . [ sec : infwire ] we calculate the energy bands of the infinite wire superlattice via direct diagonalization . by using an analytical approximation scheme@xcite we then show how the resulting band gaps depend on the rashba spin - orbit interaction . via the negf method and the landauer formalism we introduce in sec . [ sec : finwire ] a numerical scheme to calculate the charge conductance through a finite region . this region is connected to electron reservoirs to the left and right , and contains both a periodic potential and a rashba spin - orbit coupling . in sec . [ sec : result ] we then show that the band picture of the infinite wire superlattice , developed in sec . [ sec : infwire ] , is applicable to the finite length periodic potential . we close sec . [ sec : result ] with a discussion about possible experimental procedures . lastly , in sec . [ sec : pervar ] we show that our results are robust against fluctuations in the strength and width of the periodic potential . we investigate an infinite quasi-1d parabolic wire with a uniform rashba spin - orbit coupling in the presence of a longitudinal modulation described by the potential @xmath0 where @xmath1 is the period of the superlattice . the hamiltonian that describes this system is @xmath2 were @xmath3 is the effective mass , @xmath4 and @xmath5 are the momentum operators in the longitudinal and transverse direction of the wire , @xmath6 is the rashba spin - orbit strength , and @xmath7 is the confinement frequency of the parabolic potential . to find the eigenvalues of the hamiltonian in eq . ( [ eq : originalh ] ) it is convenient to introduce the standard ladder operator @xmath8 of the parabolic confinement and rotate the spin operators so that the @xmath4 part in the rashba interaction term couples to the @xmath9 operator@xcite @xmath10 + v{_{\mbox{\scriptsize p}}}(x)\nonumber\\ & = \widetilde{h}_0+\widetilde{h}_1 , \label{eq : rotatedh}\end{aligned}\ ] ] here @xmath11 is the spin ladder operator , @xmath12 is the rescaled rashba strength and @xmath13 are the eigenvalues of the operator @xmath4 . in eq . ( [ eq : rotatedh ] ) we scale all lengths in oscillator length @xmath14 and all energies in @xmath15 . we separate the hamiltonian @xmath16 into a diagonal part , @xmath17 and a non - diagonal part @xmath18 + \sum_{n\neq0 } c_n e^{in\frac{2\pi}{\lambda}}.\ ] ] the eigenstates of the @xmath19 hamiltonian are represented by the kets @xmath20 . here @xmath21 is the quantum number of the harmonic transverse energy bands , i.e. the eigenvalue of the @xmath22 operator , and @xmath23 is the eigenvalue of the @xmath9 operator with @xmath24 and @xmath25 denoting the spin up and spin down states , respectively . the corresponding eigenenergies are @xmath26 a plot of the @xmath27 energy bands vs @xmath28 in half of the brillouin zone , i.e. , @xmath29 , can be seen in fig . [ fig : h0enbands ] ( a ) . for @xmath30 nm , @xmath31 nm@xmath32 , @xmath33 mevnm ( @xmath34 ) , @xmath35 , @xmath36 , and @xmath37 . ( b ) crossing of the energy bands corresponding to the @xmath38 , @xmath39 , and @xmath40 states . ( c ) crossing of the energy bands corresponding to the @xmath41 , @xmath42 , @xmath43 , and @xmath44 states . ( d ) same as ( c ) but for higher bands , i.e. @xmath45.,scaledwidth=49.0% ] the hamiltonian @xmath46 introduces couplings _ i ) _ between the @xmath20 and the @xmath47 states , due to the rashba interaction , and _ ii ) _ between the @xmath20 and the @xmath48 states , due to the periodic potential . note that the coupling is strongest where the energy bands corresponding to these states cross each other . for some particular rashba strength @xmath49 the energy bands associated with the states @xmath38 and @xmath39 of @xmath19 cross at @xmath50 , where @xmath51 . if we choose the period @xmath1 of the superlattice potential , eq . ( [ eq : periodiceq ] ) , as @xmath52 the @xmath40 energy band will also cross at @xmath53 . this crossing point occurs in energy at @xmath54 see fig . [ fig : h0enbands ] ( b ) . in the following we refer to this particular choice of parameter @xmath49 as the reference spin - orbit coupling strength . similar crossing also occur at @xmath53 with the energy bands associated with the states @xmath55 , @xmath56 , and @xmath57 for energies @xmath58 with the energy band corresponding to the state @xmath59 crossing close by . the crossings for @xmath60 and @xmath61 can be seen , respectively , in fig . [ fig : h0enbands ] ( c ) and ( d ) . we calculate the eigenenergies of the full hamiltonian , @xmath16 , via direct diagonalization . the resulting energy bands are plotted in fig . [ fig : hfullenbands ] ( a ) . the energy bands that cross at @xmath53 and the one crossing close by , see sec . [ sec : h0 ] , corresponds to the states coupled by @xmath46 . a blow up of the resulting energy gaps can be seen in fig . [ fig : hfullenbands ] ( b ) . at @xmath62 the coupling results in a double energy gap , see fig . [ fig : hfullenbands ] ( c ) , and a triple energy gap at the higher energy crossings , see fig . [ fig : hfullenbands ] ( d ) and ( e ) . in fig . [ fig : hfullenbands ] the spin - orbit strength is at the reference value @xmath49 . note that for a non - zero rashba coupling , as in fig . [ fig : hfullenbands ] , the energy gaps have shifted from the bragg plane at @xmath63 . this is because the spin - orbit interaction shifts the wave - number @xmath28 of the electrons in the longitudinal direction and thus renormalizes the locations of interferences . for the same parameters as used in fig . [ fig : h0enbands ] . the energy bands are calculated via direct diagonalization of @xmath16 . ( b ) a blow up focusing on the band gaps of the first three crossings at @xmath53 . ( c ) the band gaps formed at the energy band crossing shown in fig . [ fig : h0enbands ] ( b ) . in ( d ) and ( e ) , the energy band crossings shown in fig . [ fig : h0enbands ] ( c ) and ( d ) , respectively.,scaledwidth=47.0% ] when the strength of the rashba interaction is changed the crossing points @xmath64 of the energy - bands shift and with them the energy gaps . these crossing points can be worked out analytically by using the eigenenergies of the effective rashba hamiltonian of the parabolic wire@xcite , which are @xmath65 and @xmath66 for @xmath67 . for @xmath68 @xmath69 . in eq . ( [ eq : effenup ] ) and ( [ eq : effendown ] ) @xmath70^{1/2},\end{aligned}\ ] ] and we have added the energy constant @xmath71 resulting from the periodic potential and replaced @xmath28 with @xmath72 . note that the energy bands described by eq . ( [ eq : effenup ] ) and eq . ( [ eq : effendown ] ) are derived for wires without a longitudinal periodic potential and therefore do not contain energy gaps that result from the periodic potential , i.e. they are the solution of the @xmath16 , see eq . ( [ eq : rotatedh ] ) , with @xmath73 and @xmath74 for @xmath75 . having determined the crossing points , we insert them into either of the crossing energy bands , eq . ( [ eq : effenup ] ) or eq . ( [ eq : effendown ] ) , to obtain the location of the crossing points in energy as a function of @xmath6 . in the appendix we present the equation for the crossing point between the energy bands @xmath76 and @xmath77 corresponding to states @xmath56 and @xmath57 , see eq . ( [ eq : kcross ] ) . to determine the location in energy of this crossing point as a function of @xmath6 we insert @xmath64 from eq . ( [ eq : kcross ] ) into either @xmath76 or @xmath77 . . we also mark the values @xmath78 mevnm and @xmath79 mevnm onto the @xmath80-axis . these values are used for the conductance results in fig . [ fig : gapvsg].,scaledwidth=45.0% ] now , as the crossing points shift in energy with @xmath6 they can be thought of as trajectories . the trajectories of the crossing points , calculated from eq . ( [ eq : effenup ] ) and ( [ eq : effendown ] ) , can be seen in fig . [ fig : trajectories ] . for higher energies there are further trajectories . we index the trajectories via @xmath21 and the letters `` c '' , `` l '' , and `` r '' . the trajectories that we label with the letter c only shift a little to the left for @xmath81 ( see lowest horizontal line in fig . [ fig : trajectories ] ) . relative to the c labeled trajectories and again for @xmath81 , we see trajectories that make a large shift to the left and right . those trajectories we , respectively , label with the letters `` l '' and `` r '' . for spin - orbit strengths close to the reference strength , @xmath49 , the crossing points follow nonlinear trajectories . as the spin - orbit coupling becomes larger or lower than the reference strength the trajectories quickly become more linear . the choice of the reference rashba spin - orbit strength @xmath49 determines the period of the periodic potential , see eq . ( [ eq : period ] ) . we will show in sec . [ sec : finwire ] that the band gaps appear as dips in the charge conductance through a finite periodic potential . by fitting the measured energy shift of the conductance dips to the trajectories in fig . [ fig : trajectories ] it is possible to extract the value of @xmath6 . the linear parts of the trajectories are best suited for fitting . it would therefore be convenient that the range of @xmath6 extracted via the fitting is contained within the linear region . this can be achieved by choosing a sufficiently low @xmath49-value ( and thus @xmath1 , eq . [ eq : period ] ) . we present below a linearized equation for the @xmath6-value of the crossing point between the energy bands corresponding to the states @xmath56 and @xmath57 as a function of fermi energy . by taylor expanding around some point @xmath82 in the linear region of the trajectories and making a linear approximation we obtain @xmath83 , \label{eq : linearealpha}\ ] ] here @xmath84 and @xmath85 are known functions of the energy band index @xmath21 , see eq . ( [ eq : f ] ) and eq . ( [ eq : g ] ) . the derivation of eq . ( [ eq : linearealpha ] ) is shown in the appendix . in fig . [ fig : trajectories ] we plot as dashed lines linearized trajectories described by eq . ( [ eq : linearealpha ] ) where we have chosen @xmath86 mevnm . in this section we calculate in the linear response the conductance through a finite periodic potential . this is done via the landauer formula together with the negf method . we consider a hardwalled wire of width @xmath87 with a transverse parabolic potential . we divide the wire into a finite central region of length @xmath88 and semi - infinite left and right parts , see fig . [ fig : numschema ] . the central region includes a rashba spin - orbit interaction described by a symmetrized hamiltonian which is turned on smoothly , at both the left and right ends . in the central region we also assume a longitudinal periodic potential that represents the potential due to the fingergates . we describe the central region by the hamiltonian @xmath89 here @xmath90 denotes an anticommutator . the left and right leads contain the same parabolic potential as in the center region but neither a rashba spin - orbit interaction nor a periodic potential in the longitudinal direction , i.e. , they are described by the hamiltonian @xmath91 the total system , @xmath92 , is discretized via the finite difference method on a grid of @xmath93 points with a mesh size @xmath94 . a schematic of the system can be seen in fig . [ fig : numschema ] . , described by the hamiltonian @xmath95 and two semi - infinite wires , described by the hamiltonians @xmath96 and @xmath97 . the total system is discretized on a grid with mesh size @xmath94 . at the left and right edges of the central region the rashba spin - orbit coupling is smoothly turned on.,scaledwidth=47.0% ] here we use the negf method@xcite to calculate the charge conductance . the method requires us to find the retarded green s function of the central region @xmath98 . to do this we have to isolate @xmath98 from the infinite matrix equation describing the retarded green s function of the total system @xmath99 by separating the total green s function into a left , right , and central part we can write eq . ( [ eq : infinitegreen ] ) as @xmath100 \left[\begin{array}{lll } \mathbf{g}{_{\mbox{\scriptsize l}}}{^{\mbox{\scriptsize r } } } & \mathbf{g}{_{\mbox{\scriptsize lc}}}{^{\mbox{\scriptsize r } } } & \mathbf{g}{_{\mbox{\scriptsize lr}}}{^{\mbox{\scriptsize r } } } \\ \mathbf{g}{_{\mbox{\scriptsize cl}}}{^{\mbox{\scriptsize r } } } & \mathbf{g}{_{\mbox{\scriptsize c}}}{^{\mbox{\scriptsize r } } } & \mathbf{g}{_{\mbox{\scriptsize cr}}}{^{\mbox{\scriptsize r } } } \\ \mathbf{g}{_{\mbox{\scriptsize rl}}}{^{\mbox{\scriptsize r } } } & \mathbf{g}{_{\mbox{\scriptsize rc}}}{^{\mbox{\scriptsize r } } } & \mathbf{g}{_{\mbox{\scriptsize r}}}{^{\mbox{\scriptsize r } } } \\ \end{array}\right]\nonumber\\ = & \left[\begin{array}{ccc } \mathbf{i } & 0 & 0 \\ 0 & \mathbf{i } & 0 \\ 0 & 0 & \mathbf{i } \\ \end{array}\right ] . \label{eq : hgmatrixeq}\end{aligned}\ ] ] the matrices @xmath101 , couple together the central region to the left ( @xmath102 l ) and right leads ( @xmath102 r ) . here @xmath103 is the tight - binding hopping parameter that results from the discretization . multiplying out eq . ( [ eq : hgmatrixeq ] ) gives nine matrix equations from which we can isolate a _ finite _ matrix equation for @xmath98 . this is done by treating the contributions from the infinite leads as self - energies@xcite . the matrix equation for @xmath98 is @xmath104 and the self - energy of lead @xmath102l , r is @xmath105 where @xmath106 is the green s function of lead @xmath107 and is determined analytically@xcite . here @xmath108 is the discretized hamiltonian of eq . ( [ eq : originalh ] ) over the central region . all the matrices are @xmath109 matrices . in order to save computational power only the necessary green s function matrix elements are calculated with the recursive green s function method@xcite . here we are interested in low biases and hence focus on the linear response regime . from the green s function the charge conductance at the fermi energy @xmath110 can be calculated , via the fisher - lee relation@xcite , as @xmath111 , \label{eq : conductance}\ ] ] where @xmath112 and @xmath113 . in what follows we use eq . ( [ eq : conductance ] ) to calculate the charge conductance through our system . in our simulations we consider a ga@xmath114in@xmath115as alloy with an effective mass @xmath116 , with @xmath117 being the bare electron mass . the width of the wire is set as @xmath118 nm and its length as @xmath119 m . the oscillator length is @xmath30 nm which corresponds to @xmath120 mev . for the discretization we use @xmath121 points with @xmath122 nm being the distance between nearest neighbors . this corresponds to a tight - binding hopping parameter @xmath123 mev@xmath124 . we choose the reference rashba strength as @xmath78 mev nm ( i.e. @xmath125 ) which results in an energy band crossing point at @xmath126 and a period @xmath127 nm . this corresponds to @xmath12880 fingergates . the strength of the periodic potential is set as @xmath129 mev . in fig . [ fig : gapvsg ] we compare the calculated conductance , fig . [ fig : gapvsg ] ( a ) , through the finite wire to the energy bands of the infinite wire , fig . [ fig : gapvsg ] ( b ) and ( c ) , as a function of fermi energy . results for two values of @xmath6 are presented , @xmath78 mevnm and @xmath79 mevnm . for comparison we also plot in fig . [ fig : gapvsg ] ( a ) the conductance through a non - periodic potential @xmath130 for the same values @xmath6 mentioned above , i.e. the corresponding gapless systems@xcite . each energy gap is associated to its corresponding conductance dip via a labeled arrow . the labeling on the arrows is the same labeling as those of the trajectories in fig . [ fig : trajectories ] . we see that there is a good correspondence between the dips in conductance and the gaps in the energy band for both @xmath49 and @xmath131 . this indicates that the behavior of the _ finite _ periodic potential can be properly described with the band model from the infinite periodic potential . mevnm , and @xmath78 mevnm . note that the conductance curve for @xmath131 has been shifted by 2@xmath132 for clarity . the results for a wire with a non - periodic potential @xmath130 is plotted with black solid lines . in ( b ) and ( c ) the energy bands of the infinite wire superlattice are plotted for the same @xmath6 values as in ( a ) , respectively . note that ( c ) is the same figure as fig . [ fig : hfullenbands ] ( b ) , but rotated clockwise by @xmath133 . the correspondence between the nine pairs of dips and band gaps are indicated by the labeled arrows . , scaledwidth=45.0% ] in fig . [ fig : trajectories ] we can see how the band gaps shift in energy as a function of @xmath6 . we want see if the energy shift of conductance dips correlates with the shift of the band gaps . this is most easily seen by considering the differential conductance @xmath134 in fig . [ fig : standardsurface ] we plot @xmath135 as a function of @xmath136 and @xmath6 . there we see the conductance steps of the parabolic wire as relatively straight vertical trajectories appearing at the tic marks on the horizontal axis . the rest of the trajectories in fig . [ fig : standardsurface ] correspond to dips in the charge conductance . by comparing the trajectories , that the dips in charge conductance form , to the ones of the band gaps in fig . [ fig : trajectories ] we see an excellent match . this firmly confirms that the conductance dips of the finite wire can be described via the band model for the superlattice . we also plot in fig . [ fig : standardsurface ] as light - gray lines the linearized trajectories of crossing points between the energy bands corresponding to the @xmath56 and @xmath57 states , see eq . ( [ eq : linearealpha ] ) . the trajectories are linearized around @xmath6=20 mevnm and are identical to the straight dashed lines in fig . [ fig : trajectories ] . through the finite superlattice as a function of the fermi energy @xmath110 and spin - orbit coupling , @xmath6 . the linearized trajectories of the crosspoints between the energy bands corresponding to the @xmath56 and @xmath59 states are shown with gray solid lines . the trajectories are linearized around @xmath86 mevnm . the rashba spin - orbit values , @xmath78 mevnm and @xmath79 mevnm , corresponding to the curves in fig . [ fig : gapvsg ] are marked on the right border . note that @xmath137 and that the differential conductance @xmath138 is in arbitrary units.,scaledwidth=47.0% ] from fig . [ fig : standardsurface ] we can extract the rashba spin - orbit coupling by fitting the linear energy shift of the conductance dips , via eq . ( [ eq : linearealpha ] ) . the energy shift , i.e. the change in fermi energy , is controlled by the chemical potential of the leads . the backgate that controls the spin - orbit strength introduces an extra shift in the fermi energy ( due to the electrostatic coupling of the backgate to the 2deg ) . this shift can be compensated by changing the chemical potential of the leads by an equal amount . similar methods have been used to probe the energy spectrum of quantum dots using transport methods @xcite . another way to compensate for this shift is to use a combination of back and front gates . this has been experimentally demonstrated to control the rashba spin - orbit interaction strength without introducing charging in the 2deg@xcite . the periodic potential plays a crucial role in in the formation and behavior of the energy gaps . we therefore devote this section to study the effect that variations of parameters in the periodic potential has on the conductance . in what follows we consider variations on the length and strength of the periodic potential ( sec . [ sec : varlenstrength ] ) as well as random fluctuations on the period and strength of each finger gate potential ( sec . [ sec : imprecision ] ) . through the finite periodic potential as a function of fermi energy @xmath110 for ( a ) different values of @xmath139 and ( b ) different number of fingergates . all the conductances were calculated with @xmath140 mevnm . note that the conductances have been separated by @xmath141.,scaledwidth=47.0% ] here we examine the effects of changing the strength of the periodic potential , @xmath139 , and the length @xmath88 . in fig . [ fig : diff ] ( a ) the length of the wire is varied such that it contains 10 , 40 , 80 , or 160 fingergates with a fixed period @xmath142 nm . there we see the dips easily for @xmath143 40 fingergates and they become clearer for more fingergates . there is not much change between the conductance results for 80 and 160 fingergates , i.e. @xmath144 80 fingergates is an adequate number of fingergates . [ fig : diff ] ( b ) shows results where the strength of the periodic potential has been varied between values of 0.2 , 0.5 , 1.0 , and 1.5 @xmath15 . we see that the width of the conductance dips grows with larger @xmath139 . but as @xmath139 increases the total conductance strength @xmath145 deteriorates due to interferences . this makes the detection of the conductance dips difficult . through the finite periodic potential as a function of fermi energy @xmath110 and spin - orbit coupling , @xmath6 . four cases are considered : a system with ( a ) 320 , ( b ) 80 , ( c ) 40 , and ( d ) 10 fingergates . in all the plots we introduce a 5% gaussian error in the length and height of the potential in each period . , scaledwidth=49.0% ] the periodic potential used in the previous section can be considered as being ideal . a more realistic case would be if there were some fluctuations in the potential . to verify the robustness of the results in the previous section we rerun the simulations with a non - ideal periodic potential . this is done by introducing a @xmath146 gaussian error in both the height and length of each potential hill created by the fingergates . figure [ fig : gausserror ] ( b ) shows that extra noise is added to our conductance result ; still the trajectories of the conductance dips are fairly visible . to counter against the noise introduced by fluctuations we could add extra periods to the periodic potential . this helps averaging out the noise , as can be seen in fig . [ fig : gausserror ] ( a ) where we have quadrupled the length of the wire and the number of fingergates . this ( self)averaging is however slow and requires a large number of fingergates . another solution , as we are mainly interested in the slope of the crosspoint trajectories , is actually to reduce the number of fingergates to an optimal number . as can be seen in fig . [ fig : gausserror ] ( c ) we obtain many extra trajectories resulting from the imprecision of the period . but the slope of the trajectories of the conductance dips can be easily fitted for @xmath147 . as expected if we go down to as few as 10 fingergates the effects of the periodic potential is nearly washed out . we have studied a parabolic quantum wire with rashba spin - orbit interaction and a longitudinal periodic potential . we find that the energy gaps resulting from the periodic potential split up and shift from the bragg plane due to the rashba spin - orbit interaction . above a certain spin - orbit strength some of the new energy gaps shift linearly in energy as a function of the rashba spin - orbit strength . we propose that this effect can be used to measure the change in the rashba spin - orbit strength , e. g. resulting from a voltage gate@xcite . the energy gaps result in dips in the charge conductance . the energy shift of these dips can be fitted , by an analytical equation that we derive , eq . ( [ eq : linearealpha ] ) , to extract the strength of the rashba spin - orbit interaction . the advantage of the this method is that it only requires conductance measurement and not any external magnetic or radiation source . in this appendix we calculate the crossing point @xmath64 of the @xmath55 and @xmath59 energy bands . we need to solve the equation @xmath148 for @xmath28 . here we used eq . ( [ eq : effenup ] ) and ( [ eq : effendown ] ) and have defined @xmath149 equation [ eq : encrosseq ] is a nonlinear equation which is hard to solve directly . an easier way is to linearize the @xmath150 and @xmath151 functions around the point @xmath152 . this approximation introduces little errors and allows us to write the crossing point as @xmath153 where @xmath154 and @xmath155 the trajectories of the crossing points across the energy - rashba spin - orbit coupling surface are then @xmath156 or @xmath157 . the crossing points and their trajectories for the other energy bands , namely the ones corresponding to the states @xmath56 and @xmath57 , the @xmath55 and @xmath57 , and the @xmath56 and @xmath59 , can be worked out in the same way . the @xmath28 value of the crossing point between the @xmath76 and @xmath77 bands remains almost constant as a function of @xmath6 . this is because these bands move at similar rate in opposite directions in @xmath28 space as @xmath6 is changed . we take an advantage of this when finding a linear equation around some point @xmath82 in the straight segments of the trajectories . we know that @xmath53 is a crossing point for @xmath158 so we can , instead of using eq . ( [ eq : kcross ] ) , make the approximation that @xmath159 for all @xmath6 . we then linearize @xmath160 by taylor expanding around @xmath82 . solving for @xmath161 in @xmath160 and scaling back to @xmath6 we obtain @xmath162 , \label{eq : applinearealpha}\ ] ] where @xmath163 and @xmath164 in the last steps of the eq . ( [ eq : f ] ) and eq . ( [ eq : g ] ) we have plugged in the values @xmath165 and @xmath166 , which correspond to @xmath78 mevnm and @xmath86 mevnm . note that all lengths are scaled in the oscillator length @xmath14 and all energies in @xmath15 .
in this work we study the effects of a longitudinal periodic potential on a parabolic quantum wire defined in a two - dimensional electron gas with rashba spin - orbit interaction . for an infinite wire superlattice we find , by direct diagonalization , that the energy gaps are shifted away from the usual bragg planes due to the rashba spin - orbit interaction . interestingly , our results show that the location of the band gaps in energy can be controlled via the strength of the rashba spin - orbit interaction . we have also calculated the charge conductance through a periodic potential of a finite length via the non - equilibrium green s function method combined with the landauer formalism . we find dips in the conductance that correspond well to the energy gaps of the infinite wire superlattice . from the infinite wire energy dispersion , we derive an equation relating the location of the conductance dips as a function of the ( gate controllable ) fermi energy to the rashba spin - orbit coupling strength . we propose that the strength of the rashba spin - orbit interaction can be extracted via a charge conductance measurement .
be / x - ray binaries are the most common type of accreting x - ray pulsar systems . they consist of a pulsar and a be ( or oe ) star , a main sequence star of spectral type b ( or o ) that shows balmer emission lines ( see e.g. , slettebak 1988 and coe 2000 for reviews . ) the line emission is believed to be associated with an equatorial outflow of material expelled from the rapidly rotating be star which probably forms a quasi - keplerian disk near the be star @xcite . x - ray outbursts are produced when the pulsar interacts with this disk . be / x - ray binaries typically show two types of outburst behavior : ( a ) giant outbursts ( or type ii ) , characterized by high luminosities ( @xmath4 ergs s@xmath5 ) and high spin - up rates ( i.e. , a significant increase in pulse frequency ) and ( b ) normal outbursts ( or type i ) , characterized by lower luminosities ( @xmath6 ergs s@xmath5 ) , low spin - up rates ( if any ) , and recurrence at the orbital period @xcite . as a population be / x - ray binaries show a correlation between their spin and orbital periods @xcite . on 1998 september 8 a 15.8 s pulsar , which was designated gro j1944 + 26 , was discovered with the burst and transient source experiment ( * ? ? ? * batse ) on the _ compton gamma ray observatory ( cgro)_. gro j1944 + 26 was localized to a @xmath7 error box @xcite . at the same time , the all - sky monitor ( * ? ? ? * asm ) on the _ rossi x - ray timing explorer ( rxte ) _ discovered a new source which they localized to a @xmath8 error box and designated xte j1946 + 274 @xcite . subsequent observations with the proportional counter array ( * ? ? ? * pca ) also on the _ rxte _ on 1998 september 16 revealed 15.8 s pulsations , confirming that batse and _ rxte _ were seeing the same object @xcite . scanning observations with the pca further improved the position to a 2.4 error circle @xcite . observations with _ bepposax _ further improved the x - ray position to r.a . = @xmath9 , decl . = + 2721.5 ( equinox 2000.0 ; error radius 1 at 95% confidence ) @xcite . xte j1946 + 274 lies in the error box of the ariel 5 transient 3a 1942 + 274 ; however , there is a non - negligible probability that this is a chance association @xcite . a likely optical counterpart has been identified as an optically faint @xmath10 mag , bright infrared ( @xmath11 ) be star at r.a . = @xmath12 , decl . = 272200 , which lies in the refined bepposax error circle@xcite . spectroscopic and photometric data indicate a b0 - 1iv - ve star at 8 - 10 kpc . the counterpart s mass is not well constrained , but is expected to be in the @xmath13 range . two previously proposed candidates @xcite lie @xmath14 outside the _ bepposax _ error circle . using data from the 1998 outburst taken with the pca and the high energy x - ray timing experiment ( * ? ? ? * hexte ) on _ rxte _ , @xcite revealed a cyclotron resonance scattering feature or cyclotron line near 35 kev . this feature implies a magnetic field strength of @xmath15 g , where @xmath16 is the gravitational redshift of the emission region . using about 1 year of _ rxte _ asm 2 - 10 kev flux measurements , beginning with the 1998 outburst , @xcite observed 5 outbursts in which they found evidence of an @xmath17 day modulation in the outburst flux . they found that the first outburst in september 1998 was markedly different in rise and decay timescales than the 4 following outbursts and based on this suggested that the first outburst might be classified as a giant outburst . xte j1946 + 274 was observed with the indian x - ray astronomy experiment ( * ? ? ? * ixae ) 1999 september 18 - 30 ( mjd 51408 - 51421 ) and 2000 june 28-july 7 ( mjd 51723 - 51733 ) . they detected 15.8 s pulsations in the 2 - 6 and 6 - 18 kev bands , with similar double peaked profiles in both observations . the pulse period history was consistent with a constant intrinsic spin - up plus an eccentric orbit . the reported pulse periods were @xmath18 s and on mjd 51445.0 and @xmath19 s on mjd 51727.5 . xte j1946 + 274 was active from 1998 september through 2001 july , when it dropped below the detection threshold of the _ rxte _ pca . as of 2002 july , no additional outbursts have been seen with the _ rxte _ asm . table [ tab : obs ] lists the dates and instruments observing xte j1946 + 274 s location . in this paper we will present results from observations of xte j1946 + 274 with batse and _ rxte _ , including histories of pulse frequency and total flux for xte j1946 + 274 . from the pulse frequency history , we derive an orbital solution . applying this orbital solution , we will investigate the observed correlation between spin - up torques and observed flux and discuss implications for the accretion mechanism . we will describe the unusual orbital phasing of the outbursts . using data from the _ rxte _ pca and hexte , we will investigate spectral variations in the initial bright outburst and the last 2 outbursts . we compare our x - ray results to optical h@xmath3 observations . to determine pulse frequencies for xte j1946 + 274 , we performed a grid search over a range of candidate frequencies using data from the large area detectors on batse@xcite . this search technique is described in detail in @xcite and @xcite . we will briefly summarize the technique here . variations of this technique , including searches in pulse period , have been widely used ( * ? ? ? * ; * ? ? ? * for example ) . the pulsar search technique consists of 3 steps ( 1 ) data selection and combination , ( 2 ) 20 - 50 kev pulse profile estimates , and ( 3 ) a grid search in frequency . first , the batse discla channel 1 ( 20 - 50 kev , 1 s time resolution ) data were selected for which the source was visible , the high voltage was on , the spacecraft was outside the south atlantic anomaly , and no electron precipitation events or other anomalies were flagged by the batse mission operations team . the count rates were combined over the 4 lads viewing xte j1946 + 274 , using weights optimized for an exponential energy spectrum , @xmath20 with temperature @xmath21 kev , and grouped into @xmath22 s segments . a segment length of 300 s was used because it was short enough that the background was well - fitted by a quadratic model . the segment boundaries were chosen to avoid earth occultation steps from bright sources@xcite . the second step in this process was estimation of an initial 20 - 50 kev pulse profile for each segment of data . in each segment , the combined rates were fitted with a model consisting of a sixth - order fourier expansion in pulse phase ( representing the 20 - 50 kev pulse profile ) @xmath23 where @xmath24 is the estimated complex fourier coefficient for harmonic where @xmath25 and @xmath26 is the pulse frequency . ] @xmath27 in segment @xmath28 and a spline function with quadratics in time ( representing the background plus mean source count rate ) . a sixth - order fourier expansion was chosen based on the number of harmonics that are significant in a 1-day observation , the typical integration time required to detect xte j1946 + 274 . our initial pulse phase model was of the form @xmath29 , where @xmath30 mhz before mjd 51270 and @xmath31 mhz after mjd 51270 were barycentric frequencies and @xmath32 was the time corrected to the solar system barycenter using the jpl de-200 ephemeris @xcite . the initial pulse frequency was changed after mjd 51270 because continued spin - up of the pulsar moved the observed frequency too far away from the folding frequency . the value and slope of the spline function were required to be continuous across adjacent 300 s segment boundaries , but not across data gaps . ideally we would like to be able to fit a pulse profile to the entire 4-day interval of data for each frequency grid point ; however this is computationally very very expensive . instead we have formulated an equivalent method in which we first fit pulse profiles to short segments of data using a single fixed frequency . then the profiles are combined and shifted by a frequency offset corresponding to each grid point . fitting pulse profiles to short data segments has other advantages in addition to improving computational efficiency . ( 1 ) if the folding frequency chosen was incorrect , the profile will be smeared out . however , the degree of smearing in pulse phase @xmath33 is approximately given by @xmath34 , where @xmath35 is an offset in frequency , @xmath36 is the segment length , and @xmath37 is the number of fourier coefficients . for @xmath38 s , @xmath39 mhz , and @xmath40 , @xmath41 , which corresponds to a drop in the amplitude of the sixth harmonic by 15% . the amplitude of the first harmonic only drops by about 0.4% . using a short segment reduces smearing of the pulse profile . ( 2 ) typically 470 pulse profiles from the 300 s segments are accumulated within each 4-day interval . since we have a large number of profiles we can estimate sample variances for each fourier coefficient . sample variances give us a measure of the noise present in the data and account for aperiodic noise if it is present . the final step in our advanced pulsar search was a grid search in frequency using the set of typically several hundred estimated 20 - 50 kev pulse profiles from 4-day intervals of data . new xte j1946 + 274 pulse frequencies were determined from an initial grid search over 1200 evenly spaced trial barycentric frequencies in the range 62.97 - 63.32 mhz for mjd 48361 - 51270 and 63.10 - 63.45 mhz for mjd 51270 - 51690 . the search range was changed because continued spin - up of the pulsar moved the pulse frequencies outside the original search range . for each frequency offset @xmath35 , a mean pulse profile is computed for each 4-day interval . the mean pulse profile is of the form @xmath42 where @xmath43 , @xmath44 , and @xmath45 is the midpoint barycentric time of segment @xmath28 . the mean coefficients @xmath46 are given by @xmath47 with weights @xmath48 . the typical statistic used for a search in pulse frequency or pulse period is the @xmath49 statistic @xcite given by @xmath50 where @xmath51 is the formal ( i.e. , poisson statistical ) error on the fourier coefficient @xmath46 , respectively . if aperiodic noise is present , either due to xte j1946 + 274 or due to other sources in the large batse field of view , e.g. , cygnus x-1 , the poisson statistical error is an underestimate because the actual underlying noise level is higher than poisson noise this can create a problem with the @xmath49 statistic , causing it to depend on the noise level . to account for aperiodic noise , we used all of the typically @xmath52 pulse profiles from 300 s segments within each 4-day interval to estimate sample variances for the fourier coefficients for each harmonic in the mean profile . each harmonic was treated separately . we then modified the @xmath49 statistic by replacing the poisson variances with our computed sample variances to create a new statistic which we will call @xmath53 after @xcite where it is described in detail . due to the large field - of - view of batse , other pulsars were often also present when we were measuring xte j1946 + 274 . if we limited our statistic to use the first 3 harmonics where xte j1946 + 274 was the brightest , we reduced the chances of contamination of the search results from other pulsars that happen to have harmonics near the higher harmonics of xte j1946 + 274 . the best - fit frequency for each 4-day interval was then determined using the @xmath54 statistic . using a similar method we also generated pulse frequency measurements for _ rxte _ pca observations during the initial outburst in 1998 september - october and during the last 2 outbursts , 2001 march - july . barycentered standard 1 ( 125 ms , no energy resolution ) data were fitted with a model consisting of a constant background plus a 6th - order fourier expansion in pulse phase model , creating an estimated 2 - 60 kev pulse profile for each pca observation . the pulsed phase model consisted of a constant barycentric frequency , @xmath55 mhz for the 1998 september - october observations and @xmath56 mhz for the 2001 observations , estimated from projections of batse measurements . for each set of observations , we searched over a grid of 151 evenly spaced frequencies spanning the range @xmath57 mhz , the same size interval as used for batse . figures [ fig : freqs]a & c show the barycentered pulse frequency history and the 2 - 10 kev total flux history , respectively , for xte j1946 + 274 . in figure [ fig : freqs]a , the trend in the pulse frequencies repeatedly changes from spin - up to spin down and back to spin - up at regular intervals ( every 170 days ) . this repeated pattern is most likely due to the pulsar s orbit . within each 170 day cycle in the pulse frequencies , there are two outbursts . examination of the observed pulse frequency history showed us two things : ( 1 ) a strong orbital signature and ( 2 ) strong intrinsic torque variations . to attempt to extract orbital parameters , we first fitted the observed pulse frequencies @xmath58 with a model consisting of a global orbit and a global polynomial frequency model , i.e. , @xmath59 where @xmath60 is a model of the emitted frequency . the velocity relative to the observer @xmath61 is given by @xmath62 where @xmath63 is the projected semi - major axis of the pulsar s orbit ; @xmath64 is the orbital period ; @xmath65 is the periapse angle ; and @xmath66 is the orbital eccentricity ; @xmath67 is the eccentric anomaly , given by @xmath68 and @xmath69 is the epoch of periastron passage . we fitted models containing polynomials of orders 1 - 10 . all of these models were very poor fits . table [ tab : orb ] lists the orbital parameters obtained using a first and 10th order polynomial for comparison . errors on these parameters have been inflated by a factor of @xmath70 , but are still expected to be considerably underestimated because the @xmath71 values are so large . figure [ fig : poly ] shows the emitted frequency model @xmath72 , @xmath73 the constant term from the emitted frequency model times the velocity relative to the observer to illustrate the amplitude of orbital effects , the full frequency model including the emitted frequency and orbit , and the frequency residuals for the 10th order polynomial model . clearly a simple model such as a global polynomial did not suffice to describe the intrinsic torques . instead we decided to model the intrinsic torques in a piecewise fashion . each outburst was split into 2 - 6 intervals , depending on the length of the outburst , where each interval contained 3 - 4 frequency measurements and spanned 16 - 24 days or 16 - 32 days for the batse and the _ rxte _ pca data , respectively . in each interval , we fitted the observed pulse frequencies with a global orbit plus an independent linear frequency model , equation [ eqn : fobs ] with @xmath74 as a model of the emitted frequency for each time interval @xmath75 . table [ tab : orb ] lists our best fit orbit . figure [ fig : resid ] shows @xmath72 , @xmath76 to illustrate the amplitude of orbital variations ( note @xmath77 was chosen because it lies in near the center of the frequency range . ) , the full frequency model including the emitted frequency and orbit , and the frequency fit residuals . this fit is not formally acceptable either ; however , we believe this is mostly because this model still does not completely describe the intrinsic torques . errors on the individual parameters in table [ tab : orb ] have been inflated by @xmath78 . comparing the three columns of table [ tab : orb ] shows that the orbital parameters appear to be similar , whether we use a simple polynomial model or a more complicated piecewise model for @xmath72 . figure [ fig : freqs]b shows the orbit - corrected spin frequencies . the regular pattern seen in figure [ fig : freqs]a has been removed , while intrinsic torque effects remain in the data . a close look at figure [ fig : freqs ] indicates that the outbursts are not fixed in orbital phase and that we are seeing approximately 2 outbursts per orbit . to study the orbital phase of the outbursts , we fitted the 2 - 10 kev _ rxte _ asm flux in figure [ fig : freqs]c with a quartic polynomial , and used this polynomial to normalize the outbursts . we then computed the orbital position using @xmath79 where @xmath80 is the semi - major axis , estimated from @xmath63 assuming @xmath81 which gives a companion mass of @xmath82 . figure [ fig : orbphase ] shows the normalized intensity versus orbital position , with darker colors indicating higher intensities . the outer loop of the spiral has @xmath83 , the estimated radius assuming @xmath84 . for each successive orbit , we reduced @xmath80 by 10% to offset the orbits . this plot spirals in solely to allow comparison of outburst phases . we are not claiming that the neutron star is spiraling in toward its companion . the outburst peaks clearly are not fixed in orbital phase , although there is typically one peak on each side of the orbit . just before periastron and just before apastron , the flux is typically relatively low . the cause of these low flux periods is unclear . although the fluxes show considerable modulation , they remain detectable with _ rxte _ throughout the period 1998 september - 2001 july . to look for variations in energy spectra during the outbursts observed with the _ rxte _ pca , we first studied hardness ratios . using ftools 5.1 @xcite and standard2 ( 16-s , 129 energy channel ) data , we created light curves in 4 energy bands corresponding to 2 - 5 , 5 - 10 , 10 - 15 , and 15 - 20 kev for both the 1998 and 2001 outbursts . these light curves were background subtracted and their times were corrected to the solar system barycenter using ftools . next we formed 3 hardness ratios from adjacent energy bands . figure [ fig : hi ] shows each of these ratios vs. total pca source count rate in the 2 - 30 kev band . figure [ fig : cc ] shows color - color diagrams . in both figures , the grey - scale denotes intensity , with darker points denoting higher intensities . ( note : this is the measured 2 - 30 kev intensity , not the normalized 2 - 10 kev intensity used in section [ sec : orbph ] and figure [ fig : orbphase ] . ) for all of the hardness ratios , we see a correlation between hardness and intensity in the 2001 outburst that is not present in the 1998 outburst . in the color - color diagrams , the hardness ratios are correlated for both sets of outbursts and have similar slopes . however , during the 2001 outbursts , points move from left to right along the correlation as intensity increases ; while in the 1998 outburst , the points move around in hardness as intensity increases . to better quantify the observed spectral changes , we generated response matrices , count spectra , and background spectra using ftools for pca standard2 data and hexte event mode data for each observation . we then fitted each of the observations in xspec 11.1 @xcite with two models : ( 1 ) an absorbed power law with a high - energy cutoff and an iron line , xspec model : phabs ( powerlaw + gaussian ) highecut@xmath85 and ( 2 ) an absorbed power law with an iron line and a cyclotron absorption feature , xspec model : phabs ( powerlaw + gaussian ) cyclabs@xmath85 . for both models we used only data in the 2.7 - 30 kev energy range for pca data and in the 15 - 50 kev range for hexte . the power law normalizations were allowed to be independent for the pca and each hexte cluster . figure [ fig : highecut ] shows @xmath71 , spectral index , cutoff energy , folding energy , and flux in the fe line vs. 2 - 60 kev flux . figure [ fig : cyclabs ] shows @xmath71 , spectral index , the cyclotron line energy , and the flux in the fe line vs. 2 - 60 kev flux . other spectral parameters did not show significant variations with intensity . a comparison of the two models showed that model 2 generally provided a better fit to the data ( see figures [ fig : highecut ] & [ fig : cyclabs ] , top panel ) . in both models , the power law photon index remains relatively independent of flux for fluxes above about @xmath86 ergs @xmath87 s@xmath5 , however , below that flux level , the power law index softens as the intensity decreases . the integrated flux in the iron line is correlated with the 2 - 60 kev flux in both models , indicating that the iron line is indeed a feature of xte j1946 + 274 s spectrum and not a background feature . in model 2 a cyclotron absorption feature , when it is constrained , tends to be near the value observed by @xcite . to estimate an upper limit to the quiescent flux from xte j1946 + 274 , we fitted pca data from each of 6 observations after pulsations were undetectable in 2001 august - september with an absorbed power law . these observations were consistent with a power law index of @xmath88 , a hydrogen column of about @xmath89 @xmath87 , and a 2 - 30 kev flux of @xmath90 ergs @xmath87 s@xmath5 . since _ rxte _ is not an imaging instrument , we can not distinguish between low - level emission from xte j1946 + 274 and background sources such as the galactic ridge . however , this gives us an upper limit on the flux from xte j1946 + 274 . to estimate a bolometric correction to the _ rxte _ asm measurements , we plotted the 2 - 60 kev fluxes from model 2 in section [ sec : spec ] vs. the 5-day average asm count rate surrounding each pca observation . we then fitted a line to these data , obtaining a bolometric correction of 1 asm ct s@xmath91 ergs @xmath87 s@xmath5(2 - 60 kev ) . the error on this correction has been estimated including systematic effects . figure [ fig : bc ] shows the best fit line to these data . near the peak flux level , the asm fluxes show evidence for a turnover relative to the pca fluxes . we have attempted to account for this turn - over by including a systematic error of 10% on our bolometric correction . using this bolometric correction , we then plotted the average bolometrically corrected asm flux vs. spin - up rate for each interval in our frequency model . figure [ fig : fdotvsflux ] clearly shows a correlation between spin - up rate and flux . such a correlation suggests a disk is present because direct wind accretion is believed to be less efficient at transferring angular momentum @xcite . if enough angular momentum is present , a disk will form . simple accretion theory assumes that the material from the companion star is flowing onto a rotating neutron star with a strong magnetic field . this field determines the motion of material in a region of space called the magnetosphere . the size of this region is denoted by the magnetospheric radius @xmath92 , given by @xcite @xmath93 where @xmath94 is the gravitational constant ; @xmath95 is the mass of the neutron star ; and @xmath96 is the mass accretion rate , which is assumed to be related to the observed bolometric flux @xmath97 by @xmath98 , where @xmath99 is the radius of the neutron star and @xmath100 is the distance to the pulsar . @xmath28 is a constant factor of order 1 . equation [ eqn : rm ] with @xmath101 gives the alfvn radius @xmath102 for spherical accretion and with @xmath103 gives the magnetospheric radius derived by @xcite . the torque applied by accretion of matter onto a neutron star , assuming torques due to matter leaving the system are negligible , is given by @xcite @xmath104 where @xmath105 is the moment of inertia of the neutron star and @xmath106 is the specific angular momentum of the material . if @xmath105 is assumed constant , then @xmath106 is given by @xmath107 where @xmath108 is the spin - up rate . to estimate @xmath106 for xte j1946 + 274 , we assumed typical pulsar parameters , @xmath109 , @xmath110 km , @xmath111 g @xmath112 , typical values of @xmath113 hz s@xmath5 and @xmath114 ergs @xmath87 s@xmath5(see figure [ fig : fdotvsflux ] ) , and a distance @xmath115 kpc @xcite . this yielded @xmath116 @xmath117 s@xmath5 . an accretion disk will form if the specific angular momentum of the material accreted from the be star s disk is comparable to the keplerian specific angular momentum at the magnetospheric radius , i.e. , @xmath118 for xte j1946 + 274 , using the magnetic field measurement @xmath119 g @xcite which implies @xmath120 g @xmath121 , then @xmath122 @xmath117 s@xmath5 for @xmath123 and @xmath124 @xmath117 s@xmath5 for @xmath125 . for xte j1946 + 274 , @xmath126 for @xmath123 and @xmath127 for k = 0.47 , hence an accretion disk is most likely present . in contrast , for the wind - fed system vela x-1 where a disk is not expected to be present , @xmath128 hz s@xmath5 @xcite , @xmath129 ergs s@xmath5 , and @xmath130 g @xmath131 @xcite leading to @xmath132 @xmath112 s@xmath5 and @xmath133 @xmath112 s@xmath5 , i.e. , @xmath134 . further , three - dimensional simulations of wind accretion show that the average specific angular momentum accreted via wind accretion is always smaller than the keplerian value @xcite . since an accretion disk appears to be present , we can use simple accretion theory to derive a distance to xte j1946 + 274 . substituting equations [ eqn : ellrm ] into equation [ eqn : torque ] and solving for @xmath108 gives @xmath135 assuming @xmath123 and the parameters defined earlier , @xmath136 where @xmath137 is the spin - up rate in units of @xmath138 hz s@xmath5 and @xmath139 is the bolometric flux in units of @xmath86 ergs @xmath87 s@xmath5 . we then fitted this model to the data in figure [ fig : fdotvsflux ] . to account for the flux and @xmath108 errors , we computed @xmath71 as @xmath140 where @xmath141 is a free parameter fitted to each flux and @xmath142 is equation [ eqn : nudot ] evaluated at @xmath141 . the solid line shows our best fit , with @xmath143 , which gives us a distance of @xmath144 kpc . this distance is consistent with optical observations of the counterpart . the errors on the distance are computed from the fit itself and take into account only the statistical errors on the flux and spin - up rate and the estimated systematic error on the bolometric correction . however , there is considerable uncertainty in our assumed neutron star parameters . the estimated distance depends on the following combination of neutron star parameters : @xmath145 . using the neutron star equations of state given by @xcite and assuming @xmath146 , we estimate the error on the distance due to uncertainty in the neutron star parameters is about 30% , i.e. , @xmath147 kpc . using ftools and _ rxte _ pca event mode data , we created light curves and corresponding background light curves in 5 energy bands : 2 - 5 , 5 - 10 , 10 - 15 , 15 - 20 , and 20 - 30 kev for each of 29 observations , 12 from the 1998 outburst and 17 from the 2001 outbursts . each light curve was barycentered , background subtracted , and folded at the appropriate frequency from our frequency search . each pulse profile contained 50 phase bins . first we examined profiles from the entire 2 - 30 kev band to search for intensity dependent variations . we aligned all of the pulse profiles by finding the phase of the minimum of the profile using a quadratic interpolation and placing that minimum at phase 0.0 . we found definite variations with intensity in both the 1998 and 2001 outbursts and also evidence that the pulse profile was different at similar intensities during the rise and fall of the initial bright outburst . profiles with similar shapes and similar intensities were averaged to better illustrate the observed shape changes . figure [ fig : intprof]a shows the average 2 - 30 kev pulse profile at 6 different intensities . figure [ fig : intprof]b shows the peak - to - peak pulse fraction versus mean flux for each profile . in the following descriptions and in figures [ fig : intprof]a and b , the pulse profiles are numbered from one to six . ( 1 ) at the lowest intensities ( 6 - 17 counts s@xmath5 pcu@xmath5 , 2 - 30 kev ) , during the 2001 outbursts , the profile s main feature was a deep notch , which we used to align the profiles . the profile consisted of an asymmetric structured main peak that peaked near phase 0.2 . ( 2 ) as the intensity increased ( 23 - 65 counts s@xmath5 pcu@xmath5 ) , a second notch becomes prominent near phase 0.45 , making the profile consist of two main peaks . these peaks were approximately equal in intensity at first , but the second peak brightens as the overall intensity increases . a small peak near phase 0.87 also appears . no _ rxte _ pca observations were taken at intensities between 65 and 178 counts s@xmath5 pcu@xmath5 . ( 3 ) the next profile shape occurs at intensities of 178 - 266 counts s@xmath5 pcu@xmath5 , near the end of the 1998 outburst . this profile is markedly different from that near the peak of the 2001 outburst ( profile 2 ) . both notches are broader , the peak near phase 0.65 is brighter than the peak near phase 0.3 and the small peak at phase 0.87 has disappeared . the main notch appears to be less deep than at lower intensities . ( 4 ) profiles at intensities of 285 - 308 and 330 counts s@xmath5 pcu@xmath5 , all from the rise of the 1998 outburst , show a broader first peak and a less intense shoulder following the second peak than the lower intensity profiles from the fall of the outburst ( profile 3 ) . ( 5 ) at intensities of 322 - 328 counts s@xmath5 pcu@xmath5 , during the decline of the 1998 outburst , the first peak is narrower than at similar intensities during the outburst rise ( profile 4 ) , but broader than at lower intensities during the decline ( profile 3 ) . ( 6 ) at the peak of the outburst , the width of the first peak is at an intermediate width between the profiles from the rise and fall of the outburst . during the 2001 outbursts , the profile shape appeared to depend primarily on intensity , while during the 1998 bright outburst , the profile shape depended on both intensity and whether the profile was from the rise or fall of the outburst . next we looked for variations in the pulse profile versus energy . figure [ fig : profvsenergy ] shows pulse profiles in 5 energy bands from the peak of the 1998 outburst ( left panel ) and near one of the peaks in the pair of outbursts in 2001 ( right panel ) . in the 1998 profiles , the profile consists of two main peaks at lower energies ( @xmath148 kev ) . the first peak is dominant at lower energies and the second peak becomes more dominant as energy increases . also , as energy increases , the notch between the two main peaks fills in . as the 2 - 30 kev intensity decreases , the profile shape " appears to move down in energy . for example , in the right hand panel of figure [ fig : profvsenergy ] , the 2 - 5 kev profile consists of two nearly equal peaks , reminiscent of the 10 - 15 kev profile at the peak of the 1998 outburst . a regular monitoring program was established to study the strength and structure of the h@xmath3 optical emission line during the x - ray observations in 2001 . this line arises from the circumstellar disk surrounding the be star and the line properties are strongly related to the extent and dynamic structure of the disk . data were primarily collected from the 2.5 m isaac newton telescope ( int ) in la palma , spain . the int was equipped with the 235-mm camera and eev#10 ccd . the use of the r1200r grating results in a nominal dispersion of @xmath149 / pixel . intermediate resolution spectroscopy was also performed on 17th july 2001 and 22nd october 2001 using the 1.93-m telescope at the observatoire de haute provence , france . that telescope was equipped with the long - slit spectrograph _ carelec _ and the @xmath150 eev ccd . we used the 1200 ln / mm grating in first order , resulting in a nominal dispersion of @xmath151 / pixel . see table [ tab : ha ] for details . the h@xmath3 profile obtained from each observation is presented in figure [ fig : ha ] . the lines have been shifted vertically by an arbitrary amount in order to present each profile clearly . the data of the 17 july and 7 october 2001 are of a lower standard due to poor weather conditions . nonetheless it is possible to accurately determine the equivalent width of each of the emission lines ; values are presented in table [ tab : ha ] . from these numbers one can see that there is little evidence for any changes in the line flux , assuming a constant continuum . however , from figure [ fig : ha ] , one can see obvious shifts in the line position after 2001 july . if one takes the first spectrum ( 3 may 2001 ) as a base line and subtracts it from each of the other spectra after they have been normalized to the same peak value , then significant changes are apparent . the result of this process is shown in figure [ fig : diff ] . it is apparent that little happens to the profiles ( and , presumably , the circumstellar disk ) until july / august 2001 . starting with the spectrum of 17 july 2001 one can see significant perturbations occurring in the line profile indicative of density changes in the disk structure . between august and september the direction of the perturbation changes , suggesting that the density enhancement / rarefication had changed sides ( or rotated around ) in the disk . the size of the perturbation also seems to have been increasing from july to october , though the separation of the red and blue peaks remains approximately constant at @xmath152340 km / s . xte j1946 + 274 is a very unusual be / x - ray binary system and its behavior does not fit well into the standard normal / giant outburst behavior . instead it showed an extended period of activity from 1998 september - 2001 july when the x - ray flux showed considerable modulation , with two peaks or outbursts per orbit , but the x - ray flux never dropped below the detection threshold of _ this extended period of activity more closely resembles a series of normal outbursts than a single giant outburst ; however , unlike typical normal outbursts , these outbursts shift rapidly in orbital phase , there are two outbursts per orbit , and the x - ray flux does not drop dramatically between outbursts . the initial outburst shows considerable spin - up @xmath153 hz s@xmath5 , like a giant outburst , but this spin - up is not substantially larger than that seen in later outbursts . in addition , the peak flux ( 2 - 60 kev ) of the initial outburst is about @xmath154 ergs @xmath87 s@xmath5 , while the peak flux ( 2 - 60 kev ) of the last two outbursts observed with the pca is about @xmath155 ergs @xmath87 s@xmath5 , only a factor of @xmath156 fainter than the brightest outburst . these fluxes correspond to luminosities of @xmath157 ergs s@xmath5and @xmath158 ergs s@xmath5 , respectively , assuming our best - fit distance of 9.5 kpc . in giant outbursts of be / x - ray binaries , accretion disks are expected to be present and indeed , evidence for an accretion disk , based on correlations between the observed flux and spin - up rate , has been found for several sources during giant outbursts @xcite . independent evidence for an accretion disk based on the detection of quasi - periodic oscillations during a giant outburst has been found for exo 2030 + 375 @xcite and a0535 + 262 @xcite . until recently , normal outbursts were believed to be due to direct wind accretion from the be disk , so significant spin - up was not expected because direct wind accretion is not believed to be very efficient at transferring angular momentum @xcite . if enough angular momentum is present in the accreted material , an accretion disk will form . however , evidence for spin - up during normal outbursts has been observed in gs 0834430 @xcite , 2s 1417624 @xcite , 2s 1845024 @xcite , and in exo 2030 + 375 @xcite . we see a correlation between spin - up and flux for xte j1946 + 274 as shown in figure [ fig : fdotvsflux ] . the spin - up rate and its correlation with bolometric flux during xte j1946 + 274 s outbursts suggest an accretion disk may be present . in addition , our calculations show that the specific angular momentum of the accreted material is comparable to the keplerian specific angular momentum , hence a disk is expected to form in xte j1946 + 274 . further , our fit to the spin - up vs. bolometric flux correlation , which assumed a disk was present , yielded a distance of @xmath2 kpc which is consistent with the distance of 8 - 10 kpc derived from optical observations @xcite . hence an accretion disk is likely to be present in xte j1946 + 274 . between 2001 july 31 ( mjd 52121 ) and 2001 august 9 ( mjd 52130 ) , xte j1946 + 274 dropped below the pca s detection threshold . at approximately the same time , between 2001 june 29 ( mjd 52089 ) and 2001 july 17 ( mjd 52107 ) , the h@xmath3 profile began to change rapidly . the profile was stable for 3 observations on 2001 may 3 , may 10 , and june 29 ( mjd 52032 , 52039 , & 52089 ) that corresponded to x - ray observations during the decline of the second to last outburst , during the low state between the last two outbursts , and near the peak of the last outburst . the coincidence of the change in h@xmath3 with the x - ray turn - off suggests that changes in the be disk caused the x - ray outbursts to cease . to determine whether or not the observed x - ray turn - off was due to centrifugal inhibition of accretion @xcite , we estimate the flux at the onset of this effect by equating the magnetospheric radius to the corotation radius . the magnetospheric radius is given by equation [ eqn : rm ] and the corotation radius is given by @xmath159 where @xmath26 is the spin frequency of the pulsar . setting @xmath160 gives the threshold flux for the onset of centrifugal inhibition of accretion , i.e. , @xmath161 where @xmath162 , @xmath163 , @xmath164 , and @xmath165 are the pulsar s magnetic moment in units of @xmath166 g @xmath131 , mass in units of 1.4 @xmath167 , radius in units of @xmath168 cm , and spin period in units of 15.8 seconds , respectively . using @xmath169 kpc and @xmath170 g @xmath121 yields @xmath171 ergs @xmath87 s@xmath5@xmath172 . for @xmath173 , @xmath174 ergs @xmath87 s@xmath5 . our measured upper limit fluxes are in the range @xmath175 ergs @xmath87 s@xmath5 , consistent with xte j1946 + 274 entering the centrifugal inhibition of accretion regime when the x - rays became undetectable with the pca . we propose that xte j1946 + 274 is a system in which the be star s equatorial plane and the orbital plane are not aligned . this is suggested by : ( 1 ) the optical observations show a relatively narrow fwhm ( 8.6 on 2001 may 10 , which corresponds to @xmath176 200 km s@xmath5 ) for the single - peaked h@xmath3 line , indicating that the be star is viewed from a relatively low inclination angle , i.e. , nearly pole - on . ( 2 ) the orbital signature is quite obvious in the pulse frequencies , indicating that the orbital inclination angle is likely not low : the derived mass function @xmath177 from the piecewise frequency model indicates an inclination angle @xmath178 for the expected mass range of @xmath13 derived from optical observations . figure [ fig : mc ] shows the range of allowed inclinations using the 1-@xmath179 errors on the mass function . we see two outbursts per orbit in xte j1946 + 274 ; however , these outbursts are not fixed in orbital phase . this combination of behaviors is unique to xte j1946 + 274 and is likely not due to a single mechanism . the optical data and mass function suggest that the orbital plane and the be disk are not aligned in this system . such a misaligned system would be expected to produce two outbursts per orbit , each approximately corresponding to the neutron star s passage through the be disk ; hence these outbursts would be expected to be fixed in orbital phase . for xte j1946 + 274 , we are faced with the difficult task of explaining not only why we see the expected two outbursts per orbit , but also why these outbursts are not fixed in orbital phase . we propose that given the relatively high luminosities and spin - up rates in all of the outbursts that perhaps xte j1946 + 274 s unusual outburst behavior can be explained as a giant outburst . in normal outbursts , the be disk is expected to be truncated at a resonance radius by tidal forces from the neutron star s orbit @xcite ; however , in giant outbursts , where much more material is believed to be present in the disk , it is believed that the disk is no longer truncated . hence more material likely means a larger disk around the be star . density perturbations propagating in the be disk could change the contact points of the disk and the orbit . these perturbations would have to be moving quite rapidly in the disk . this idea of a giant outburst combined with density perturbations seems most likely to work in a system where the inclination angle between the be disk and the orbital plane is fairly small . it is not clear that this angle is small in xte j1946 + 274 . we do not claim that the idea of a giant outburst in an inclined system fully explains the behavior observed from xte j1946 + 274 . instead we put the idea forth as a suggestion to those doing simulations of these systems . the phasing of the outbursts of xte j1946 + 274 ( see figure [ fig : orbphase ] ) is difficult to understand . however , it is not the first be / x - ray binary to show shifts in outburst phase . exo 2030 + 375 underwent a fairly sudden ( within 4 orbits ) shift in the peak phase of the outbursts from 6 days after periastron ( phase @xmath180 ) to 2.5 days before periastron ( phase -0.05 ) , followed by a gradual shift in outburst phase to 2.5 days after periastron ( phase 0.05 ) . this shift was believed to be associated with a density perturbation observed in the be disk via h@xmath3 observations @xcite . gs 0834430 also underwent sudden shifts in outburst phase , undergoing 9 outbursts approximately centered on periastron , followed by dramatic shifts to outbursts centered on phases 0.37 and 0.75 @xcite . a companion was not known at the time for gs 0834430 , so we can not confirm that density perturbations in the be disk were responsible for these phase shifts . in xte j1946 + 274 , we have a similar problem , with the companion not observed for much of the time the x - ray outbursts were occurring . in addition , the problem is compounded by the fact that we are likely viewing the be star nearly pole - on ; hence the range of projected rotational velocities is very small , making it difficult to detect perturbed velocities which indicate density perturbations . only large density perturbations , such as the one that coincided with the x - ray turn - off , are likely to be detected . we are grateful to luisa morales for providing one of the optical spectra . this research has made use of data obtained from the high energy astrophysics science archive research center ( heasarc ) , provided by nasa s goddard space flight center ( gsfc ) . asm quick - look results were provided by the asm/ _ rxte _ teams at mit and at the gsfc sof and gof . the int is operated on the island of la palma by the isaac newton group in the spanish observatorio del roque de los muchachos of the instituto de astrofsica de canarias . based in part on observations made at observatoire de haute provence ( cnrs ) , france . angelini , l. , stella , l. , & parmar , a.n . 1989 , , 346 , 906 arnaud , k.a . 1996 , astronomical data analysis software and systems v , eds . g. jacoby & j. barnes , asp conf . series , 101 , 17 benensohn , j.s . , lamb , d.q . & taam , r.e . 1997 , apj , 478 , 723 bildsten , l. et al . 1997 , , 113 , 367 blackburn , j.k . 1995 , in asp conf . 77 , astronomical data analysis and software systems iv , ed . r.a . shaw , h.e . payne , & j.j.e . haynes ( san francisco : asp ) , 367 buccheri , r. et al . 1983 , , 128 , 245 campana , s. , israel , g. , & stella , l. 1998 , iau circ . 7039 campana , s. , israel , g. , & stella , l. 1999 , , 352 , l91 coe , m.j . 2000 , in the be phenomenon in early - type stars , iau colloquium 175 , asp conf . 214 , eds . smith & h.f . henrichs , ( san francisco : asp),656 corbet , r.h.d . 1986 , , 220 , 1047 finger , m. h. et al . 1999 , , 517 , 449 finger , m.h . , wilson , r.b . & chakrabarty , d. 1996 , a&as , 120 , 209 finger , m.h . , wilson , r.b . , harmon , b.a . 1996 , apj , 459 , 288 fishman , g. j. et al . 1989 , in proc . gro science workshop , ed . w.n . johnson ( greenbelt : nasa / gsfc ) , 2 ghavamian , p. & garcia , m. 1998 , iau circ . 7022 ghosh , p. & lamb , f.k . 1979 , apj , 234 , 296 hanuschik , r.w . 1996 , , 308 , 170 heindl , w.a . et al . 2002 , , 563 , l35 inam , s.c . & baykal , a. 2000 , , 353 , 617 israel , g.l . , polcaro , v.f . , & covino , s. 1998 , iau circ . 7021 jahoda , k. et al . 1996 , euv , x - ray , and gamma - ray instrumentation for astronomy vii , spie proc . 2808 , ed . sigmund & m.gumm ( bellingham : spie ) , 59 lamb , f.k . , pethick , c.j . , & pines , d. 1973 , , 184 , 271 levine , a.m. , bradt , h. , cui , w. , jernigan , j.g . , morgan , e.h . , remillard , r. , shirey , r. e. , & smith , d.a . 1996 , , 469 , l33 makishima , k. , mihara , t. , nagase , f. , tanaka , y. 1999 , , 525 , 978 okazaki , a.t . & negueruela , i. 2001 , , 377 , 161 parmar , a.n . , white , n.e . , stella , l. , & ferri , p. 1989 , , 338 , 359 paul , b. et al . 2001 , , 370 , 529 pringle , j.e . , & rees , m.j . 1972 , , 21 , 1 quirrenbach , a. et al . 1997 , , 479 , 477 reynolds , a.p . 1996 , , 312 , 872 rothschild , r.e . 1998 , , 496 , 538 ruffert , m. 1997 , , 317 , 793 ruffert , m. 1999 , , 346 , 861 slettebak , a. 1988 , , 100 , 770 smith , d.a . & takeshima , t. 1998 , iau circ . 7014 standish , e. m. , newhall , x. x. , williams , j. g. , & yeomans , d. k. 1992 , orbital ephemerides of the sun , moon , & planets , in explanatory supplement to the astronomical almanac , ed . k. seidelmann ( mill valley : university science books ) , 279 stella , l. , white , n.e . , & rosner , r. 1986 , , 308 , 669 stollberg , m.t . et al . 1999 , , 512 , 313 takeshima , t. & chakrabarty , d. 1998 , iau circ . 7016 verrecchia , f. et al . 2002 , , in press waters , l.b.f.m . & van kerkwijk , m.h . 1989 , , 223 , 196 wilson , c.a . 1997 , apj , 479 , 388 wilson , c.a . , finger , m.h . , wilson , r.b . , scott , d.m . 1998 , iau circ . 7014 wilson , c.a . , finger , m.h . , coe , m.j . , laycock , s. , fabregat , j. 2002 , apj , 570 , 287 wilson - hodge , c.a . 1999 , phd dissertation , university of alabama in huntsville wiringa , r.b . , fiks , v. & fabriocini , a. 1988 , phys . c. , 38 , 1010 lllll 1991 apr 15 - 2000 may 27 & 48361 - 51691 & batse & & + 1996 feb 23 - 2002 apr 11 & 50136 - 52376 & _ rxte _ asm & & + 1998 sep 16 - 1998 oct 14 & 51072 - 51100 & _ rxte _ pca , hexte & 12 & 85.87 + 2001 mar 9 - 2001 sep 25 & 51977 - 52177 & _ rxte _ pca , hexte & 24 & 129.49 + [ tab : obs ] llll @xmath64 ( days ) & @xmath181 & @xmath182 & @xmath183 + @xmath69 & jd@xmath184 & jd@xmath185 & jd@xmath186 + @xmath63 ( lt - sec ) & @xmath187 & @xmath188 & @xmath189 + @xmath66 & @xmath190 & @xmath191 & @xmath192 + @xmath65 ( degrees ) & @xmath193 & @xmath194 & @xmath195 + @xmath196 @xmath197 & @xmath198 & @xmath199 & @xmath200 + @xmath201d.o.f & @xmath202 & @xmath203 & @xmath204 + [ tab : orb ] cccccc 3 may 2001&int&ids+500 camera&r1200r & tek5 & [email protected] + 10 may 2001&int&ids+235 camera&r1200r & eev10&[email protected] + 29 jun 2001&int&ids+235 camera&r1200r & eev10&[email protected] + 17 jul 2001&ohp&carelec & 1200l / mm&eev & [email protected] + 12 aug 2001&int&ids+235 camera&r1200r & eev10&[email protected] + 26 sep 2001&int&ids+235 camera&r1200r & eev10&[email protected] + 7 oct 2001&int&ids+235 camera&r1200r & eev10&[email protected] + 22 oct 2001&ohp&carelec & 1200l / mm&eev & [email protected] + [ tab : ha ]
xte j1946 + 274 = gro j1944 + 26 is a 15.8 s be / x - ray pulsar discovered simultaneously in 1998 september with the burst and transient source experiment ( batse ) on the _ compton gamma ray observatory ( cgro ) _ and the all - sky monitor ( asm ) on the _ rossi x - ray timing explorer ( rxte)_. here we present new results from batse and _ rxte _ including a pulse timing analysis , spectral analysis , and evidence for an accretion disk . our pulse timing analysis yielded an orbital period of 169.2 days , a moderate eccentricity of @xmath0 , and implied a mass function of 9.7 m@xmath1 . we observed evidence for an accretion disk , a correlation between measured spin - up rate and flux , which was fitted to obtain a distance estimate of @xmath2 kpc . xte j1946 + 274 remained active from 1998 september - 2001 july , undergoing 13 outbursts that were not locked in orbital phase . comparing _ rxte _ pca observations from the initial bright outburst in 1998 and the last pair of outbursts in 2001 , we found energy and intensity dependent pulse profile variations in both outbursts and hardening spectra with increasing intensity during the fainter 2001 outbursts . in 2001 july , optical h@xmath3 observations indicate a density perturbation appeared in the be disk as the x - ray outbursts ceased . we propose that the equatorial plane of the be star is inclined with respect to the orbital plane in this system and that this inclination may be a factor in the unusual outburst behavior of the system .
in 1996 , l. kauffman introduced the notion of a virtual knot , which is motivated by study of knots in a thickened surface and abstract gauss codes , cf . . according to m. goussarov , m. polyak and o. viro , two classical knot diagrams represent the same knot type if and only if they represent the same virtual knot type . thus , the notion of a virtual knot is a generalization of a classical knot in 3-space . some properties and applications of virtual knots are found in , etc . using the basic moves appearing in the definition of a virtual knot , we obtain the notion of a virtual braid ( cf . ) . it is closely related with the welded braid group @xmath0 and the braid - permutation group @xmath1 . r. fenn , r. rimnyi and c. rourke defined the groups @xmath0 and @xmath1 and proved that they are isomorphic . there is a canonical epimorphism from the virtual braid group @xmath2 to the welded braid group @xmath0 , and the group @xmath2 contains the braid group @xmath3 and the symmetric group @xmath4 as subsets in a natural way . braid theory plays an important role in classical knot theory . the two theories are related by alexander s and markov s theorems which state that every knot ( or link ) type is represented by the closure of a braid and such a braid presentation is unique up to conjugations and stabilizations ( cf . ) . there is a one - to - one correspondence between links and braids modulo these operations . analogously , virtual braid theory is expected to be so in virtual knot theory . it is quite easy to prove the alexander theorem for virtual knots ; that is , every virtual link type is represented by the closure of a virtual braid . in fact , this is obvious from the relationship between virtual links and gauss diagrams , or gauss codes given in . in , kauffman asked whether there is a generalization of the markov theorem for virtual knots . our main result is the following , which is an answer to his question and ensures a relationship between virtual braid theory and virtual knot theory . * theorem 3.2 * _ two virtual braids ( or virtual braid diagrams ) have equivalent closures as virtual links if and only if they are related by a finite sequence of the following moves @xmath5 @xmath6 ( or @xmath7 @xmath6 ) : _ * a braid move ( which is a move corresponding to a defining relation of the virtual braid group ) , * a conjugation ( in the virtual braid group ) , * a right stabilization of positive , negative or virtual type , and its inverse operation , * a right / left virtual exchange move . @xmath8- , @xmath9- and @xmath10-moves correspond to classical markov moves . the last move ( @xmath11-move ) is an analogue of an exchange move ( cf . ) . in the category of classical braids , an exchange move is a consequence of markov moves . however its analogy does not hold in the category of virtual braids ( proposition 3.3 ) . thus @xmath11-moves are essential . by theorem 3.2 , a left stabilization of any type ( defined in 3 ) is realized by @xmath8- , @xmath9- , @xmath10- and @xmath11-moves . if the left stabilization is of virtual type , then it can be realized without @xmath11-moves ( proposition 3.4 ) . this is analogous to a fact that a left stabilization of positive / negative type for classical braids is realized by markov moves . it is rather surprising that a left stabilization of positive / negative type for virtual braids is not realized by @xmath8- , @xmath9- and @xmath10-moves ( proposition 3.5 ) . most of known virtual knot invariants , as groups , quandles , alexander polynomials , and f - polynomials ( jones polynomials , normalized bracket polynomials ) , are considered and calculated easily via virtual braids ( see for these invariants ) : group presentations and quandle presentations are obtained from virtual braids by a method which is completely analogous to that in classical knot theory ( although upper presentation and lower presentation yield different groups and quandles in general , ) . the burau representation is easily defined for virtual braids and it brings alexander polynomials . the notion of temperley - lieb algebra is generalized to virtual ones and its basis is much simpler than the classical one , . using the virtual temperley - lieb algebra , one can obtain the f - polynomials via braids . very recently , d. silver and s. williams found a new group invariant for a virtual link ( with @xmath12 components ) and a @xmath13-variable polynomial invariant derived from it . j. sawollek also found a similar invariant . their invariants are so powerful as to distinguish the trivial knot and kauffman s example of a virtual knot which can not be distinguished by all of the above invariants . ( j. s. carter proved this fact independently by an argument in ) . d. silver informed the author that their invariant in was motivated by the burau representation and it is natural to consider via virtual braids . related topics to these invariants from a point of view of virtual braids will be discussed elsewhere . welded braid theory due to fenn , rimnyi and rourke yields the notion of welded knots and links , which are also known as virtual knots and links in the weak sense ( ) . we prove alexander s and markov s theorems for welded knots and links . * theorem 7.2 * _ two welded braids ( or welded braid diagrams ) have equivalent closures as welded links if and only if they are related by a finite sequence of the following moves @xmath14 @xmath15 ( or @xmath16 @xmath15 ) : _ * a welded braid move ( which is a move corresponding to a defining relation of the welded braid group ) , * a conjugation in the welded braid group , * a right stabilization of positive , negative or virtual type , and its inverse operation . the author wishes to thank j. s. carter , n. kamada , l. h. kauffman , d. s. silver , x .- s . lin and o. dasbach for many stimulating conversations . he also thanks to department of mathematics and statistics , university of south alabama for hospitality . let @xmath17 be a positive integer and @xmath18 a set of @xmath17 interior points of the interval @xmath19 $ ] . we denote by @xmath20 the 2-disk @xmath19\times [ 0,1]$ ] and by @xmath21 $ ] the second factor projection . * definition 2.1 * a _ virtual braid diagram of degree @xmath17 _ is an immersed 1-manifold @xmath22 in @xmath20 such that * @xmath23 , * for each @xmath24 , @xmath25 $ ] is a homeomorphism , * the singularity ( the multiple point set ) @xmath26 consists of transverse double points , * @xmath27 $ ] is injective , * each point of @xmath26 is assigned information of _ positive , negative _ or _ virtual crossing _ as in figure 1 ( where the labels @xmath28 are used later ) . the arcs @xmath29 are assumed to be oriented from the top ( @xmath19\times\{1\}$ ] ) to the bottom ( @xmath19\times\{0\}$ ] ) of @xmath20 and two virtual braid diagrams are identified if they are transformed into each other continuously keeping the above conditions . the set of virtual braid diagrams of degree @xmath17 ( with the concatenation product ) forms a monoid which is generated by @xmath30 ( @xmath31 ) as in figure 2 . the trivial element is @xmath32 \subset e$ ] . * definition 2.2 * the _ virtual braid group @xmath2 of degree @xmath17 _ is the group obtained from the monoid of virtual braid diagrams of degree @xmath17 by introducing the following relations : @xmath33 a _ virtual braid of degree @xmath17 _ is an element of @xmath2 . we denote it by the same symbol @xmath34 as its representative ( a virtual braid diagram ) @xmath34 unless it makes confusion . the welded braid group @xmath0 ( defined in ) is obtained from @xmath2 by introducing additional relations @xmath35 ( @xmath36 . there is a canonical epimorphism @xmath37 . in particular , we see that the subgroup of @xmath2 generated by @xmath38 ( @xmath39 ) is isomorphic to the braid group @xmath3 and the subgroup generated by @xmath40 ( @xmath39 ) is isomorphic to the symmetric group @xmath4 , cf . . a _ virtual link diagram _ is a closed oriented 1-manifold @xmath41 immersed in @xmath42 such that the singularity set @xmath43 consists of transverse double points each of which is assigned information of positive , negative or virtual crossing as in figure 1 . positive and negative crossings are also called _ real crossings_. virtual link diagrams are considered up to isotopy of @xmath42 . _ virtual reidemeister moves _ are the local moves illustrated in figure 3 . notice that the moves indicated by ( b ) are obtained from the moves indicated by ( a ) by use of rii - moves or vii - moves . two virtual link diagrams are _ equivalent _ ( as virtual links ) if they are related by a finite sequence of virtual reidemeister moves . a _ virtual link _ ( or a _ virtual link type _ ) is the equivalence class of a virtual link diagram , . for a virtual braid diagram @xmath34 , the _ closure _ of @xmath34 is a virtual link diagram constructed as in figure 4 . if @xmath34 and @xmath44 are equivalent as virtual braids , then their closures are equivalent as virtual links . thus in virtual knot theory the closure makes sense for a virtual braid . * proposition 3.1 * _ every virtual link type is represented by the closure of a virtual braid diagram . _ for virtual braids @xmath45 , we say that @xmath46 is obtained from @xmath47 by a _ conjugation_. for a virtual braid ( diagram ) @xmath34 of degree @xmath17 , we denote by @xmath48 the virtual braid ( diagram ) of degree @xmath49 obtained from @xmath34 by adding @xmath50 trivial arcs to the left of @xmath34 and @xmath51 trivial arcs to the right . this defines a monomorphism @xmath52 . a _ right stabilization _ of _ positive _ , _ negative _ or _ virtual type _ is a replacement of @xmath53 by @xmath54 , @xmath55 or @xmath56 , respectively . see figure 5 . similarly , a _ left stabilization _ is a replacement of @xmath53 by @xmath57 , @xmath58 or @xmath59 . a _ right virtual exchange move _ is a replacement @xmath60 and a _ left virtual exchange move _ is a replacement @xmath61 where @xmath62 , see figure 6 . * theorem 3.2 * _ two virtual braids ( or virtual braid diagrams ) have equivalent closures as virtual links if and only if they are related by a finite sequence of the following moves @xmath5 @xmath6 ( or @xmath7 @xmath6 ) : _ * a braid move ( which is a move corresponding to a defining relation of the virtual braid group ) , * a conjugation ( in the virtual braid group ) , * a right stabilization of positive , negative or virtual type , and its inverse operation , * a right / left virtual exchange move . a replacement @xmath63 where @xmath64 , is called an _ exchange move _ , cf . . a @xmath11-move is an analogue of this move . in classical braid theory , an exchange move is a consequence of braid moves , conjugations and right stabilizations . however , in virtual braid theory , a virtual exchange move is independent of these moves . * proposition 3.3 * _ a @xmath11-move is not a consequence of @xmath8- , @xmath9- and @xmath10-moves ; namely , there is a pair of virtual braids which are related by a @xmath11-move and never related by a sequence of @xmath8- , @xmath9- and @xmath10-moves . _ by theorem 3.2 , a left stabilization of any type for virtual braids is a consequence of @xmath8- , @xmath9- , @xmath10- and @xmath11-moves . if it is of virtual type , then we do not need @xmath11-moves . * proposition 3.4 * _ a left stabilization of virtual type is a consequence of a @xmath10-move and some @xmath8- and @xmath9-moves . _ this is analogous to a fact that a left stabilization ( of positive / negative type ) for classical braids is a consequence of a right stabilization and some braid moves and conjugations . if the left stabilization for virtual braids is of positive / negative type , then we need @xmath11-moves in general . * proposition 3.5 * _ a left stabilization of positive / negative type for virtual braids is not a consequence of @xmath8- , @xmath9- and @xmath10-moves ; namely , there is a pair of virtual braids which are related by a left stabilization of positive / negative type and never related by a sequence of @xmath8- , @xmath9- and @xmath10-moves . _ for a virtual link diagram @xmath41 , we denote by @xmath65 the set of real crossings and by @xmath66 the map assigning the real crossings their signs . for a real crossing @xmath67 , let @xmath68 be a regular neighborhood of @xmath69 as in figure 1 . we denote by @xmath70 the four points of @xmath71 ordered as in the figure . put @xmath72 and @xmath73 , where @xmath74 means the closure . the restriction of @xmath41 to @xmath75 , denoted by @xmath76 , is the union of some oriented arcs and loops immersed in @xmath75 such that the singularities are virtual crossings of @xmath41 and the boundaries of the arcs are the set @xmath77 . define a subset @xmath78 such that @xmath79 if and only if @xmath76 has an arc starting from @xmath80 and terminating at @xmath34 . we denote by @xmath81 the number of components of @xmath41 . for example , for a virtual link diagram illustrated in figure 7 , @xmath82 the _ gauss data _ of @xmath41 is the quadruple @xmath83 . we say that two virtual link diagrams @xmath41 and @xmath84 have the _ same gauss data _ if @xmath85 and if there is a bijection @xmath86 such that @xmath87 preserves the signs of the crossing points and that @xmath79 implies @xmath88 , where @xmath89 is the bijection induced from @xmath86 . this condition is equivalent to that @xmath41 and @xmath84 have the same gauss diagram in the sense of or the same gauss code in the sense of . let @xmath41 be a virtual link diagram and @xmath90 as before . suppose that @xmath84 is a virtual link diagram with the same gauss data as @xmath41 . then we can deform @xmath84 by an isotopy of @xmath42 such that * @xmath41 and @xmath84 are identical in @xmath68 for every @xmath67 , * @xmath84 has no real crossings in @xmath75 , and * there is a one - to - one correspondence between the arcs / loops of @xmath76 and those of @xmath91 with respect to the end points of the arcs . in this situation , we say that @xmath84 is obtained from @xmath41 by _ replacing @xmath76_. * lemma 4.1 ( ) * _ if two virtual link diagrams @xmath41 and @xmath84 have the same gauss data , then @xmath41 is equivalent to @xmath84 . moreover , such an equivalence can be realized by @xmath92- , @xmath93- , @xmath94- and @xmath95-moves . _ _ without loss of generality we may assume that @xmath84 is obtained from @xmath41 by replacing @xmath76 . let @xmath96 be the arcs / loops of @xmath76 , and let @xmath97 be corresponding ones for @xmath91 . we may assume that @xmath98 intersects @xmath99 transversely . the arc or loop @xmath100 is homotopic to @xmath98 in @xmath42 ( relative to the boundary of @xmath100 if @xmath100 is an arc ) . taking the homotopy generically with respect to the arcs / loops @xmath99 and the 2-disks @xmath101 , we see that the arc / loop @xmath100 is transformed into @xmath98 by a finite sequence of moves as in figure 8 up to isotopy of @xmath42 . each move is a vi- , vii- , viii- , or mi - move . inductively , every @xmath102 is transformed into @xmath103 by such moves . let @xmath104 be the origin of @xmath42 . identify @xmath105 with @xmath106 by the polar coordinate and let @xmath107 be the projection , where @xmath108 is the half - line consisting of positive numbers . a _ braided virtual link diagram _ ( of degree @xmath17 ) is a virtual link diagram @xmath41 such that * it is contained in @xmath105 , * for an underlying immersion @xmath109 of @xmath41 , the composition @xmath110 is an orientation preserving covering map of degree @xmath17 ( where @xmath111 is the disjoint union of @xmath81 circles ) , and * @xmath112 is injective . a point @xmath113 of @xmath114 is called a _ regular value _ if @xmath115 . by cutting @xmath41 along the half - line @xmath116 for a regular value @xmath113 , we obtain a virtual braid diagram whose closure is @xmath41 . such a virtual braid is unique up to conjugation . * braiding process ( proof of proposition 3.1 ) . * let @xmath41 be a virtual link diagram and let @xmath101 be regular neighborhoods of the real crossings of @xmath41 . by an isotopy of @xmath42 , we may assume that all @xmath117 ( @xmath118 ) are in @xmath119 , @xmath120 for @xmath121 and the restriction of @xmath41 to @xmath117 satisfies the condition of a braided virtual link diagram . replace the remainder @xmath122 arbitrarily such that the result is a braided virtual link diagram . by lemma 4.1 , @xmath41 is equivalent to this diagram . the terminologies `` braid moves '' , `` right stabilizations '' and `` right / left virtual exchange moves '' are also used for braided virtual link diagrams . ( conjugations are just braid moves . ) these moves and their inverse moves are also called @xmath8- , @xmath10- and @xmath11-moves . for example , the moves illustrated in figure 9 are right stabilizations ( @xmath10-moves ) for braided virtual link diagrams . if two braided virtual link diagrams are related by a finite sequence of @xmath8- and @xmath10-moves , then we say that they are _ virtually markov equivalent in the strict sense_. if they are related by a finite sequence of @xmath8- , @xmath10- and @xmath11-moves , then we say that they are _ virtually markov equivalent_. * lemma 5.1 * _ let @xmath41 and @xmath84 be braided virtual link diagrams ( possibly of distinct degrees ) such that @xmath84 is obtained from @xmath41 by replacing @xmath122 . then @xmath41 and @xmath84 are virtually markov equivalent in the strict sense . _ _ proof . _ let @xmath101 be regular neighborhoods of the real crossings of @xmath41 ( and @xmath84 ) and @xmath123 . take a common regular value @xmath124 for @xmath41 and @xmath84 such that @xmath125 is not in @xmath126 . assume that there exists an arc / loop @xmath102 of @xmath76 and the corresponding one @xmath103 of @xmath91 such that @xmath127 . move a small part of @xmath102 or @xmath103 toward the origin by a series of @xmath8-moves corresponding to @xmath128 and apply some @xmath10-moves of virtual type so that @xmath129 . thus we may assume that @xmath129 for every arc / loop @xmath102 of @xmath76 . let @xmath130 and @xmath131 be underlying immersions @xmath132 of @xmath41 and @xmath84 such that they are identical near the preimages of the real crossings . let @xmath133 be intervals or circles in @xmath111 with @xmath134 for @xmath135 , and put @xmath136 . let @xmath137 be such immersions obtained from @xmath84 . note that @xmath138 and @xmath139 are orientation preserving immersions and @xmath140 . since @xmath102 and @xmath103 have the same degree with respect to @xmath125 , we have a homotopy @xmath141}$ ] between @xmath142 and @xmath143 relative to the boundary @xmath144 such that for each @xmath145 $ ] , @xmath146 is an immersion . taking such a homotopy generically with respect to the other arcs / loops of @xmath76 ( and @xmath91 ) and the 2-disks @xmath101 , we have a finite sequence of @xmath8-moves transforming @xmath102 to @xmath103 ( recall the proof of lemma 4.1 ) . applying this procedure inductively , we see that @xmath41 is transformed into @xmath84 by @xmath8-moves . applying the above argument , we obtain proposition 3.4 . _ proof of proposition 3.4 . _ in figure [ fig : vbraidd ] , we show a process that @xmath53 is transformed into @xmath147 ( the figure is for the case of @xmath148 ) . the step ( 2 ) @xmath149 ( 3 ) is a @xmath10-move , up to @xmath9-moves . the other steps are @xmath8-moves and @xmath9-moves . * lemma 5.2 * _ two braided virtual link diagrams with the same gauss data are virtually markov equivalent in the strict sense . _ _ proof . _ let @xmath41 and @xmath84 be braided virtual link diagrams with the same gauss data . let @xmath101 be regular neighborhoods ( as in figure [ fig : crossings ] ) of the real crossings @xmath150 of @xmath41 , and @xmath151 be regular neighborhoods of the corresponding real crossings @xmath152 of @xmath84 . ( case 1 ) suppose that @xmath153 and @xmath154 appear in @xmath114 in the same ( cyclic ) order . by an isotopy of @xmath42 , deform @xmath41 keeping the condition of a braided virtual link diagram such that @xmath155 and the restrictions of @xmath41 and @xmath84 to these disks are identical . by lemma 5.1 , @xmath41 and @xmath84 are virtually markov equivalent in the strict sense . ( case 2 ) suppose that @xmath153 and @xmath154 do not appear in @xmath114 in the same ( cyclic ) order . it is sufficient to consider a special case that @xmath153 and @xmath154 appear in @xmath114 in the same order except a pair , say @xmath156 and @xmath157 . applying @xmath8-moves , we may assume that @xmath41 is the closure of a virtual braid diagram which looks like the left one of figure [ fig : vbraida ] , where @xmath47 is a virtual braid diagram without real crossings and @xmath158 is a virtual braid diagram . the middle of the figure is obtained from the left by @xmath8- and @xmath10-moves . the right one is obtained from the middle by @xmath8-moves . by case 1 , the right one and @xmath84 are virtually markov equivalent in the strict sense . thus @xmath41 and @xmath84 are virtually markov equivalent in the strict sense . since the braiding process ( given in 4 ) does not change the gauss data of a virtual link diagram , we have the following . * corollary 5.3 * _ for a virtual link diagram @xmath41 , a braided virtual link diagram obtained by the braiding process is unique up to virtual markov equivalence in the strict sense . _ _ proof of theorem 3.2 . _ the if part is obvious . we prove the only if part . let @xmath41 and @xmath84 be braided virtual link diagrams which are equivalent as virtual links . there is a finite sequence of virtual link diagrams from @xmath41 to @xmath84 each step of which is one of the moves in figure [ fig : vrmoveoric ] . ( for ri - moves and vi - moves , there are other cases of orientations of the arcs . these cases are obtained from the moves in the figure by rii- and vii - moves . this is called _ the whitney trick_. for riii- , viii- and mi - moves , there are other cases of orientations on the arcs . these cases are also obtained from the moves in the figure by rii- and vii - moves . ) by use of vii - moves , an riic - move and an riid - move are obtained from an xa - move and an xb - move in figure [ fig : vrmoveorid ] , respectively . therefore , there is a finite sequence of virtual link diagrams @xmath159 such that each @xmath160 is obtained from @xmath161 by an ria- , rib- , vi- , riia- , riib- , xa- , xb- , viia- , viib- , viic- , riii- , viii- or mi - move . apply the braiding process to each @xmath160 and let @xmath162 be a braided virtual link diagram with the same gauss data as @xmath160 . note that @xmath163 is uniquely determined up to virtual markov equivalence in the strict sense ( lemma 5.2 ) . we assume that @xmath164 and @xmath165 . then it is sufficient to prove that for each @xmath166 @xmath167 , @xmath163 and @xmath168 are virtually markov equivalent . if @xmath160 is obtained from @xmath161 by a vi- , viia- , viib- , viic- , viii- or mi - move , then @xmath160 and @xmath161 have the same gauss data and so do @xmath163 and @xmath168 . by lemma 5.2 , @xmath163 and @xmath168 are virtually markov equivalent . suppose that @xmath160 is obtained from @xmath161 by an ria- , rib- , riia- , riib- , xa- , xb- , or riii - move . let @xmath169 be a 2-disk in @xmath42 where the move is applied , and let @xmath170 be the complement of @xmath169 in @xmath42 so that @xmath171 . if the move is not an xb - move , then we can deform @xmath160 and @xmath161 by an isotopy of @xmath42 such that @xmath172 and @xmath173 satisfy the condition of a braided virtual link diagram . apply the braiding process to the remainder @xmath171 , and we have braided virtual link diagrams , say @xmath174 and @xmath175 such that @xmath176 , @xmath177 , and @xmath178 . if the move is an ria- , rib- , or xa - move , then @xmath169 contains the origin @xmath104 of @xmath42 and @xmath174 and @xmath175 are related by a right stabilization of positive / negative type or a right virtual exchange move . if the move is an riia- , riib- , or riii - move , then @xmath169 is disjoint from @xmath104 and @xmath174 and @xmath175 are related by a @xmath8-move . since @xmath174 has the same gauss data as @xmath160 , it is virtually markov equivalent to @xmath163 by lemma 5.2 . similarly @xmath175 is virtually markov equivalent to @xmath168 . therefore @xmath163 and @xmath168 are virtually markov equivalent . if the move is an xb - move , then deform @xmath160 and @xmath161 by an isotopy of @xmath42 such that they are the closures of the ( virtual ) tangles depicted as ( a1 ) and ( b1 ) in figure [ fig : vbraidb ] , say @xmath179 and @xmath180 , where @xmath47 and @xmath158 are virtual braid diagrams . ( first deform @xmath172 and @xmath173 such that they are as in the thick boxes of ( a1 ) and ( b1 ) . then apply the braiding process to the remainder . ) let @xmath174 and @xmath175 be the closures of the virtual braid diagrams depicted as ( a2 ) and ( b2 ) in the figure . note that @xmath174 has the same gauss data as @xmath179 and hence as @xmath160 . thus @xmath174 is virtually markov equivalent to @xmath163 ( lemma 5.2 ) . similarly @xmath175 is virtually markov equivalent to @xmath168 . on the other hand , @xmath174 and @xmath175 are related by a left virtual exchange move . therefore @xmath163 and @xmath168 are virtually markov equivalent . we fix a positive integer @xmath181 and an integer @xmath182 . let @xmath41 be a virtual link diagram and let @xmath101 be regular neighborhoods of the real crossings of @xmath41 . let @xmath183 be the arcs / loops of @xmath122 as before . a _ state _ @xmath184 is assignment of elements of @xmath185 to the arcs / loops @xmath183 . a state @xmath184 is _ admissible _ if at each crossing @xmath69 of @xmath41 the labels around @xmath69 are one of figure [ fig : vweights ] . then we give @xmath69 an element of @xmath186 $ ] indicated in the figure , which is denoted by @xmath187 . for an admissible state @xmath184 of @xmath41 , let @xmath188 for an admissible state @xmath184 of @xmath41 , let @xmath189 be the virtual link diagram obtained from @xmath41 by changing the crossing points of @xmath41 as in figure [ fig : vsmooth ] . then each component , say @xmath190 , of @xmath189 inherits a unique element of @xmath191 from the state @xmath184 , which we denote by @xmath192 . let @xmath193 where @xmath190 runs over all components of @xmath189 . we denote by @xmath194 the number of positive crossings minus the number of negative crossings of @xmath41 . for a virtual link diagram @xmath41 , we define @xmath195 by @xmath196,\ ] ] where @xmath184 runs over all admissible states of @xmath41 . for a virtual braid diagram @xmath34 , we define @xmath197 by @xmath198 . * lemma 6.1 * _ if virtual braid diagrams @xmath34 and @xmath44 are related by @xmath8- , @xmath9- and @xmath10-moves , then @xmath199 . ( if braided virtual link diagrams @xmath41 and @xmath84 are virtually markov equivalent in the strict sense , then @xmath200 . ) _ _ proof . _ by a standard argument of state models ( cf . ) , it is directly checked that @xmath201 does not change under each move . details are left to the reader . _ the function @xmath201 is a modification of the state model of the braid invariant @xmath202 given by v. turaev . @xmath203 and @xmath204 correspond to @xmath205 and @xmath206 in . we changed @xmath206 into @xmath204 so that the function does not change under a vi - move in figure [ fig : vrmoveoric ] . this yields loss of invariability under riic- , riid - moves , that helps us to prove propositions 3.3 and 3.5 . _ proof of proposition 3.3 . _ let @xmath207 and @xmath208 . they are related by a right virtual exchange move . by a direct calculation , we have @xmath209 and @xmath210 . therefore , a right virtual exchange move is not a consequence of @xmath8- , @xmath9- and @xmath10-moves . let @xmath211 and @xmath212 . then @xmath213 and @xmath214 . therefore , a left virtual exchange move is not a consequence of @xmath8- , @xmath9- and @xmath10-moves . _ proof of proposition 3.5 . _ by a direct calculation , we have @xmath215 , @xmath216 , and @xmath217 . therefore , we have the result . in this section a virtual link diagram is referred to as a _ welded link diagram_. we call the local move illustrated in the left hand side of figure [ fig : wrmovea ] a _ @xmath218-move_. two welded link diagrams are _ equivalent _ _ as welded links _ if they are related by a finite sequence of virtual reidemeister moves and w - moves . the equivalence class is called a _ welded link _ or a _ welded link type_. it is easily verified that the oriented w - move illustrated in the right of figure [ fig : wrmovea ] is sufficient to realize all possible orientations for a w - move up to oriented moves in figure [ fig : vrmoveoric ] . we refer to a virtual braid diagram as a _ welded braid diagram_. recall that the welded braid group @xmath0 is the quotient of @xmath2 by adding the relations @xmath219 ( @xmath220 ) corresponding to w - moves . _ two welded braids ( or welded braid diagrams ) have equivalent closures as welded links if and only if they are related by a finite sequence of the following moves @xmath14 @xmath15 ( or @xmath16 @xmath15 ) : _ * a welded braid move ( which is a move corresponding to a defining relation of the welded braid group ) , * a conjugation in the welded braid group , * a right stabilization of positive , negative or virtual type , and its inverse operation . _ proof . _ if it is of virtual type , then it follows from proposition 3.4 . if it is of positive / negative type , then replace the virtual crossings of ( 2 ) in figure [ fig : vbraidd ] with real crossings so that the step ( 4 ) @xmath149 ( 5 ) is allowed in the welded braid group . _ proof . _ a right virtual exchange move is realized by @xmath221- , @xmath222- and @xmath223-moves as follows : @xmath224 where @xmath225 ( and we also denote by @xmath226 ( @xmath227 ) the natural images @xmath228 and @xmath229 ) . similarly , a left virtual exchange move is realized by wm0- , wm1-moves and left stabilizations . by lemma 7.3 , we have the result . we call a braided virtual link diagram a _ braided welded link diagram_. two braided welded link diagrams are _ welded markov equivalent _ if they are related by wm0- and wm2-moves . ( wm1-moves are regarded as wm0-moves . ) by lemma 7.4 , if two braided welded link diagrams are virtually markov equivalent , then they are welded markov equivalent . _ proof of theorem 7.2 . _ the if part is obvious . we prove the only if part . let @xmath41 and @xmath84 be braided welded link diagrams which are equivalent as welded links . there is a finite sequence of welded link diagrams @xmath159 such that each @xmath160 is obtained from @xmath161 by an ria- , rib- , vi- , riia- , riib- , xa- , xb- , viia- , viib- , viic- , riii- , viii- , mi- or w - move ( in figures [ fig : vrmoveoric ] , [ fig : vrmoveorid ] and [ fig : wrmovea ] ) . apply the braiding process to each @xmath160 and let @xmath162 be a braided welded link diagram with the same gauss data as @xmath160 . by lemma 5.2 ( and lemma 7.4 ) , @xmath163 is uniquely determined up to welded markov equivalence . we assume that @xmath164 and @xmath165 . it is sufficient to prove that for each @xmath166 @xmath167 , @xmath163 and @xmath168 are welded markov equivalent . in the proof of theorem 3.2 , we have already seen that @xmath163 and @xmath168 are welded markov equivalent , except the case that @xmath160 is obtained from @xmath161 by a w - move . suppose that @xmath160 is obtained from @xmath161 by a w - move . let @xmath169 be a 2-disk in @xmath42 where the w - move is applied , and let @xmath170 be the complement of @xmath169 so that @xmath171 . deform @xmath160 and @xmath161 by an isotopy of @xmath42 such that @xmath172 and @xmath173 satisfy the condition of a braided virtual ( welded ) link diagram . apply the braiding process to the remainder @xmath171 , and we have braided welded link diagrams , say @xmath174 and @xmath175 such that @xmath176 , @xmath177 , and @xmath178 . @xmath174 and @xmath175 are related by a wm0-move corresponding to @xmath230 . since @xmath174 has the same gauss data as @xmath160 , it is welded markov equivalent to @xmath163 . similarly @xmath175 is welded markov equivalent to @xmath168 . therefore @xmath163 and @xmath168 are welded markov equivalent . _ remark . _ ( 1 ) s. satoh showed that welded links are related with ribbon surfaces in 4-space whose components are tori . from the point of view of , welded braids are related with the motion group of a trivial link in 3-space ( cf . ) . \(2 ) when we use a move illustrated in figure [ fig : wrmoveb ] , called a _ @xmath231-move _ , instead of a w - move , we have another notion which is similar to a welded link . define a group @xmath232 by the quotient of @xmath2 by the relations @xmath233 ( @xmath220 ) , instead of @xmath234 . then we have results similar to those in this section . it should be noticed that we can not use both of w - moves and @xmath231-moves simultaneously . if we use both moves , every virtual ( or welded ) knot diagram changes into the unknot , .
virtual knots were first introduced by l. kauffman , which are a generalization of classical knots and links . they lead us to the notion of virtual braids , which are closely related with welded braids of r. fenn , r. rimnyi and c. rourke . it is proved that any virtual knot is uniquely described as the closure of a virtual braid modulo certain basic moves . this is analogous to the alexander and markov theorem for classical knots and braids . a similar result is proved for welded knots and braids _ mathematics subject classification _ : primary 57m25 . _ key words and phrases _ : virtual knot , braid , welded braid , alexander and markov theorem , exchange move .
tetraquarks were proposed decades ago . early in 1977 , jaffe make a calculation using the color - magnetic interaction in the bag model @xcite . he suggested that the light scalar mesons below 1 gev , @xmath0 , @xmath1 , @xmath2 and @xmath3 , be interpreted as a nonet of light tetraquarks . in recent years , the light scalar mesons are observed in decays of charmed mesons . the @xmath3 is observed as a peak in decay @xmath4 @xcite and @xmath2 in @xmath5 @xcite . from the process @xmath6 , bes collaboration determined the pole position of @xmath3 to be @xmath7 mev@xcite . also bes collaboration found a @xmath2 like structure in the decay @xmath8@xcite . the accumulation of experimental data allows us to study the structure of the light scalar nonet based on their decay properties @xcite . as a many - body system , a tetraquark state is quite different from a baryon or a conventional @xmath9 meson . the color structure is no longer trivial . it is quite sensitive to the hidden color structure of qcd interaction . a tetraquark state , if its existence is confirmed , may provide us important information about the qcd interaction that is absent from the ordinary baryons or the @xmath9 mesons . for instance , some authors had investigated the tetraquark system with the three - body @xmath10 and @xmath11 interaction , whose existence has no direct effect on the ordinary hadron states @xcite . the newly updated experimental data can shed more light on the relation of the possible tetraquark states and qcd interaction . in this article , we will study the possible tetraquark state within the framework of the non - relativistic potential model . we will calculate mass spectra and wave functions of the light tetraquark using the bhaduri potential@xcite . to fit the experimental masses , we will extend the model with the three - body @xmath10 and @xmath11 interaction . using the wave functions of tetraquarks , we will determine the coupling constants of tetraquarks to mesons under the `` fall apart '' mechanism . the article is organized as follows : in sec . [ sec:2 ] , we introduce the model hamiltonian and the multiple gaussian function method which is used to obtain the tetraquark wave functions . in sec . [ sec:3 ] , we present the `` fall apart '' decay calculation with tetraquark wave function . in sec . [ sec:4 ] , we present the numerical results . finally we will give a brief summary . in a non - relativistic quark model , usually the potentials are limited to the two - body interaction , which mainly consists of two parts : the @xmath12 color interaction including the confinement and the coulomb interaction of one - gluon exchange , and the @xmath13 color - magnetic interaction . the hamiltonian reads @xmath14\ ] ] where @xmath15 are quark masses , @xmath16 , and @xmath17 is the distance between quark @xmath18 and quark @xmath19 . among the various potential forms used in different quark models , the bhaduri potential @xcite is rather simple and gives a unified description of conventional hadron spectroscopy . also it is often used to discuss the tetraquark system @xcite . the potential reads @xmath20 the parameter values are @xmath21 in a tetraquark , some new interactions which have no direct effect on the ordinary hadrons may have significant contribution . for instance , one can introduce the following three - body @xmath10 and @xmath11 interactions@xcite @xmath22,\\ v_{q\bar{q}\bar{q}}(\bm{r}_i,\bm{r}_j,\bm{r}_k ) & = d^{abc } f_i^a f_j^{b * } f_k^{c * } u_0 \exp[-(r_{ij}^2 + r_{jk}^2 + r_{ki}^2)/r_0 ^ 2].\end{aligned}\ ] ] in this article , since we will only treat the ground states of tetraquark , the spatial dependence of the three - body interaction is less important . so we will only add the following simplified interaction into the model hamiltonian ( [ eq:1 ] ) @xmath23 this interaction is diagonal in the diquark anti - diquark color base of tetraquark @xmath24_{\bar{3 } } [ \bar{q}\bar{q}]_{3 } \mid v_{3b } \mid [ qq]_{\bar{3 } } [ \bar{q}\bar{q}]_{3 } \rangle = - \frac{20}{9}u_0,\\ \langle [ qq]_{6 } [ \bar{q}\bar{q}]_{\bar{6 } } \mid v_{3b } \mid [ qq]_{6 } [ \bar{q}\bar{q}]_{\bar{6 } } \rangle = + \frac{10}{9}u_0 . \end{aligned}\ ] ] an immediate consequence is that this three - body interaction has no direct contribution to any meson - meson coupling channel . to explain our calculation method , we first define some convenient coordinates for tetraquark system as illustrated in figure [ fig1 ] @xcite , [ coord - x ] @xmath25 [ coord - y ] @xmath26 [ coord - z ] @xmath27 @xmath28 the base wave function for the tetraquark will be the product of color , spin , flavor and spatial wave functions . the color and spin @xmath29 base function we use is of the following diquark anti - diquark coupling form : * @xmath30 @xmath31 * @xmath32 @xmath33 * @xmath34 @xmath35 here the color wave function of the two ( anti-)quarks is labeled by @xmath36 dimension and the spin wave function is by the total spin . the anti - symmetric diquarks @xmath37 $ ] , @xmath38 $ ] , @xmath39 $ ] form the @xmath40 representation of flavor @xmath41 . the @xmath40 diquarks and @xmath42 anti - diquarks further form a tetraquark nonet . they are assumed to be the light scalar mesons @xcite . so the flavor wave functions are : @xmath43[\bar{u}\bar{s}]-[ds][\bar{d}\bar{s}])\\ f_0(i=0 ) & = \frac{1}{\sqrt{2}}([us][\bar{u}\bar{s}]+[ds][\bar{d}\bar{s}])\\ \sigma_0(i=0 ) & = [ ud][\bar{u}\bar{d } ] \\ \kappa^+ & = [ ud][\bar{s}\bar{d } ] \end{aligned}\ ] ] as for the spatial wave functions , we will start from the multi - dimensional gaussian function @xmath44 where @xmath45 are the function parameters . the wave function of this form is well convergent and there exists many analytical expressions for different matrix elements . we will use it to construct the spatial wave function base @xcite . under the hypothesis of jaffe , the color - spin wave function of `` good '' diquark is the symmetric one @xmath46 . as the flavor wave function of the scalar nonet state is anti - symmetric , so the spatial wave function should be symmetric . that is the spatial wave function of tetraquark state should be invariant under @xmath47 and/or @xmath48 . if we use the gaussian function ( [ eq:12 ] ) as the base wave function , it is easy to see that @xcite @xmath49 we will use the following symmetric combination as the base function @xmath50.\end{aligned}\ ] ] if the non - diagonal parameters @xmath51 are small , we have @xmath52 \notag \\ & \times \left[1 + 2a_{12}^{s2}(\bm{x}_1\cdot\bm{x}_2)^2 + 2a_{13}^{s2}(\bm{x}_1\cdot\bm{x}_3)^2 + 2a_{23}^{s2}(\bm{x}_2\cdot\bm{x}_3)^2 \right].\end{aligned}\ ] ] this allow us to study the correlations in the quark alignment . we will choose @xmath53 independent symmetric gaussian functions ( [ eq:13 ] ) , @xmath54 , to span an @xmath53-dimensional nonorthogonal basis . the @xmath53 independent gaussian functions are obtained by the following process . first , we use one such symmetric gaussian function as the test wave function in variation to determine a base parameter set @xmath55 . the matrix @xmath56 will be specified by three principal values denoted @xmath57 , @xmath58 , @xmath59 and three euler angels @xmath60 which specified the orientation . then a complete parameter set @xmath61 is generated by first scaling to the principal values@xcite @xmath62 where @xmath63 , @xmath64 , and @xmath65 a scaling factor ; then we make an euler rotation @xmath60 . by diagonalizing the hamiltonian in the above nonorthogonal basis , we will obtain the mass and wave function of tetraquark states . the wave function can be expressed in the above base functions as @xmath66 where @xmath67 is the flavor wave function , @xmath68 is the superposition coefficient . similar to case in pseudo - scalar mesons , the @xmath69 members @xmath70 , @xmath71 in the scalar nonet will mix with each other . to consider the mixing , we further introduce a mixing angle @xmath72 @xcite @xmath73 then @xmath74 and @xmath3 are the physically observable states . in this article , we do not discuss the underlying mechanism of this mixing . so we will merely treat the mixing angle @xmath72 as one additional parameter . several authors have used the effective lagrangian with @xmath41 symmetry to discuss the decay of light scalar nonet @xcite . here we can calculate the coupling constants using the tetraquark wave functions . the general coupling lagrangian reads @xmath75 \notag \\ & + \sigma_0\left [ + g_{\sigma_0\pi\pi } \frac{\pi\cdot\pi}{2 } + g_{\sigma_0\bar{k}k } \bar{k}k + ... \right ] \notag \\ & + a \cdot \left [ g_{a_0 \bar{k}{k } } \bar{k}\tau k + g_{a \eta_s \pi } \eta_s \pi + c_{a\eta_q\pi } \eta_q \pi + ... \right ] \notag \\ & + g_{\kappa\bar{k}\pi } \left ( \bar{k}\tau \kappa \cdot \pi + h.c . \right ) + ... \end{aligned}\ ] ] at present , the quark interaction underlying those meson decaying couplings is still unclear to us . here we will assume that the decaying is the fusion process and can be depicted by the `` fall apart '' mechanism in figure [ fig2 ] . more specific , we assume that the coupling constant of a tetraquark @xmath76 to two mesons @xmath77 and @xmath78 is proportional to the wave function overlapping @xmath79 the meson wave functions will also be approximated by multiple gaussian wave functions determined by a similar variation process @xmath80 where @xmath67 is meson flavor wave function , and the spatial base function is @xmath81 a tetraquark system @xmath82 can fall apart into two different flavor combinations @xmath83 and @xmath84 , and the corresponding final meson - meson states are different @xmath85 the spatial wave functions are in the coordinates @xmath86 and @xmath87 defined in eqs . ( [ coord - y ] ) and ( [ coord - z ] ) respectively . `` fall apart '' mechanism for decays of @xmath88 tetraquark states ] in the decay of the light scalars to pseudo - scalar mesons , we need to consider the @xmath89-@xmath90 mixing @xmath91 where @xmath92 , @xmath93 and @xmath94 @xcite . we obtain the following expressions for the coupling constants ( a proportional constant is dropped ) [ eq:25 ] @xmath95 [ eq:26 ] @xmath96 [ eq:27 ] @xmath97 [ eq:28 ] @xmath98 besides the explicit flavor overlapping factors , @xmath99 is the overlapping of the color , spin and spatial wave function . after considering the @xmath3-@xmath70 mixing effect , the coupling constants @xmath100 for the decays of @xmath3 and @xmath70 are further modified to [ eq:29 ] @xmath101 bhaduri potential gives a unified description of the spectroscopy of ordinary mesons and baryons . the hamiltonian ( [ eq:1 ] ) itself is an eigenvalue problem of the differential equation which can be solved numerically . however , the multiple gaussian function method can still give an impressively good approximation of the ground state mesons and the gaussian wave function is rather simple to use . in table [ tab:1 ] , we show some results of the pseudo - scalar meson calculation . we can see that the multiple gaussian function method greatly improve the single gaussian function approximation . .pseudo - scalar meson calculation . in col . 1 , we direct solve the schrdinger equation . in col . 2 , we use the variation method with a single gaussian function . in col . 3 , we use the multiple gaussian function method with 7 gaussian functions . [ cols="<,^,^,^ " , ] in summary , we have performed a tetraquark calculation of light scalar mesons using the quark potential model . if we only consider the two - body quark interaction as in the conventional hadron calculation , the masses of the tetraquark states will be several hundred mev higher than experimental data . also the major component of the light tetraquark wave functions consists of the color sextet diquark and anti - diquark . after including a three - body interaction in the hamiltonian , we find the masses of the light tetraquark nonet agree with experimental data and their wave functions are composed of mainly the `` good '' diquark and anti - diquark . we have used the multiply gaussian function to approximate the tetraquark wave functions and noticed that there is a small mixing of d - waves in the wave functions . with wave functions we obtain , we also calculate the coupling constants for scalar meson decays according to the `` fall apart '' mechanism . by introducing the isoscalar mixing angle @xmath72 , we obtain a fit of the ratios of coupling constants for scalar meson decays similar to other analysis based on tetraquark picture .
we study the possible tetraquark interpretation of light scalar meson states @xmath0 , @xmath1 , @xmath2 , @xmath3 within the framework of the non - relativistic potential model . the wave functions of tetraquark states are obtained in a space spanned by multiple gaussian functions . we find that the mass spectra of the light scalar mesons can be well accommodated in the tetraquark picture if we introduce a three - body quark interaction in the quark model . using the obtained multiple gaussian wave functions , the decay constants of tetraquarks are also calculated within the `` fall apart '' mechanism .
the study of adjoint actions and variants thereof , and in particular the classification of orbits for such actions and the description of the orbit closures , are a common theme in lie representation theory . the archetypical example is the jordan - gerstenhaber theory for the conjugacy classes of complex @xmath2-matrices . + a more recent case is a. melnikov s study of the action of the borel subgroup @xmath0 acting on upper - triangular @xmath1-nilpotent matrices via conjugation @xcite . the orbits and their closures are described there combinatorially in terms of so - called link patterns , which we will recapitulate in section [ rom ] . + our aim in this paper is to generalize the work of a. melnikov by extending the variety of upper - triangular @xmath1-nilpotent matrices to all @xmath1-nilpotent matrices . the basic setup to reach this goal is a translation of the classification problem to a problem in representation theory of finite - dimensional algebras . more precisely , this translation yields a bijection between the orbits and the isomorphism classes of certain representations of a specific finite - dimensional algebra , see section [ trans ] . after a brief summary of methods from the representation theory of algebras ( see , for example , @xcite ) in section [ roa ] , we are able to calculate all indecomposable representations using auslander - reiten theory @xcite in section [ covsection ] and to classify the required representations . this gives a combinatorial classification in terms of oriented link patterns in section [ cob ] . + since several results on orbit closures for representations of finite - dimensional algebras are available through work of g. zwara @xcite , we can also characterize the orbit closures of @xmath1-nilpotent matrices in section [ close ] . + finally , we study the conjugation action of upper - triangular matrices on arbitrary nilpotent matrices . we provide a generic normal form for the orbits of this action in section [ gnf ] and construct a large class of semiinvariants in section [ si ] . + * acknowledgments : * the authors would like to thank k. bongartz and a. melnikov for valuable discussions concerning the methods and results of this work . in this section , we fix some notation and collect information about the aforementioned group action . in addition , we summarize material from the representation theory of finite - dimensional algebras . + let @xmath3 be the field of complex numbers . we denote by @xmath4 the borel subgroup of upper - triangular matrices , by @xmath5 the variety of nilpotent @xmath2-matrices , and by @xmath6 the closed subvariety of @xmath1-nilpotent such matrices . obviously , @xmath7 and @xmath0 act on @xmath8 and on @xmath6 via conjugation . + in case of the action of @xmath7 on @xmath8 , the classical jordan - gerstenhaber theory gives a complete classification of the orbits and their closures in terms of partitions ( or , equivalently , young diagrams ) . + our aim is to classify the orbits @xmath9 of @xmath1-nilpotent matrices @xmath10 under the action of @xmath0 . such a classification will be given in terms of _ oriented link patterns _ ; these are oriented graphs on the set of vertices @xmath11 such that every vertex is incident with at most one arrow . this is followed by a description of the orbit closures , by giving a necessary and sufficient condition to decide whether one orbit is contained in the closure of another , and by a method to construct all orbits contained in a given orbit closure . these descriptions are also given in terms of oriented link patterns . the group @xmath0 also acts on @xmath12 , the space of all upper - triangular matrices in @xmath8 , and on @xmath13 . the orbits and their closures for the latter action are described by a. melnikov in @xcite . since these results will be generalized in the following , we describe them in more detail . + let @xmath14 be the set of involutions in the symmetric group @xmath15 in @xmath16 letters . an element @xmath17 of @xmath18 is represented by a so - called link pattern , an unoriented graph with vertices @xmath11 and an edge between @xmath19 and @xmath20 if @xmath21 . for example , the involution @xmath22 corresponds to the link pattern ( 100,30 ) ( 10,10)(20,0)5 ( 10,-5)1 ( 30,-5)2 ( 50,-5)3 ( 70,-5)4 ( 90,-5)5 ( 10,10)(20,40)(30,10 ) ( 50,10)(70,40)(90,10 ) . for @xmath23 , define @xmath24 by @xmath25 and denote by @xmath26 the @xmath0-orbit of @xmath27 . [ mel1]@xcite every orbit of @xmath0 in @xmath28 is of the form @xmath29 for a unique @xmath23 . the next step is to look at the ( zariski-)closures of the orbits @xmath29 . + for @xmath30 , consider the canonical projection @xmath31 deleting the first @xmath32 and the last @xmath33 columns and rows of a matrix in @xmath28 . define the rank matrix @xmath34 of @xmath35 by @xmath36 the rank matrix @xmath34 is @xmath0-invariant , and we denote @xmath37 for @xmath23 . we define a partial ordering on the set of rank matrices by @xmath38 if @xmath39 for all @xmath19 and @xmath20 , inducing a partial ordering on @xmath14 by @xmath40 if @xmath38 . @xcite the orbit closure of @xmath41 is given by @xmath42 . moreover , the entry @xmath43 of the rank matrix equals the number of edges with end points @xmath44 and @xmath45 such that @xmath46 in the link pattern of @xmath17 . the theorem thus gives a combinatorial characterization of the @xmath0-orbits in @xmath28 and their orbit closures in terms of link patterns . as we make key use of results from the representation theory of finite - dimensional algebras for the study of the action of @xmath0 on @xmath6 , we now recall the basic setup of this theory and refer to @xcite and @xcite for a thorough treatment . let @xmath47 be a finite quiver , that is , a directed graph @xmath48 consisting of a finite set of vertices @xmath49 and a finite set of arrows @xmath50 , whose elements are written as @xmath51 ; the vertices @xmath52 and @xmath53 are called the source and the target of @xmath54 , respectively . a path in @xmath47 is a sequence of arrows @xmath55 such that @xmath56 for all @xmath57 ; we formally include a path @xmath58 of length zero for each @xmath59 starting and ending in @xmath19 . we have an obvious notion of concatenation @xmath60 of paths @xmath55 and @xmath61 such that @xmath62 . + the path algebra @xmath63 is defined as the @xmath64-vector space with basis consisting of all paths in @xmath47 , and with multiplication @xmath65 the radical @xmath66 is defined as the ( two - sided ) ideal generated by paths of positive length . an ideal @xmath67 of @xmath63 is called admissible if @xmath68 for some @xmath69 . + the key feature of such pairs @xmath70 consisting of a quiver @xmath47 and an admissible ideal @xmath71 is the following : every finite - dimensional @xmath64-algebra @xmath72 is morita - equivalent to an algebra of the form @xmath73 , in the sense that their categories of finite - dimensional @xmath64-representations are ( @xmath64-linearly ) equivalent . + a finite - dimensional @xmath64-representation @xmath74 of @xmath47 consists of a tuple of @xmath64-vector spaces @xmath75 for @xmath76 , and a tuple of @xmath64-linear maps @xmath77 indexed by the arrows @xmath78 in @xmath50 . a morphism of two such representations @xmath79 and @xmath80 consists of a tuple of @xmath64-linear maps @xmath81 such that @xmath82 for a representation @xmath74 and a path @xmath83 in @xmath47 as above , we denote @xmath84 . we call @xmath74 bound by @xmath67 if @xmath85 whenever @xmath86 . + the abelian @xmath64-linear category of all representations of @xmath47 bound by @xmath67 is denoted by @xmath87 ; it is equivalent to the category of finite - dimensional representations of the algebra @xmath73 . we have thus found a `` linear algebra model '' for the category of finite - dimensional representations of an arbitrary finite - dimensional @xmath64-algebra @xmath72 . + we define the dimension vector @xmath88 of @xmath74 by @xmath89 for @xmath76 . for a fixed dimension vector @xmath90 , we consider the affine space @xmath91 ; its points naturally correspond to representations @xmath74 of @xmath47 of dimension vector @xmath92 with @xmath93 for @xmath59 . via this correspondence , the set of such representations bound by @xmath67 corresponds to a closed subvariety @xmath94 . it is obvious that the algebraic group @xmath95 acts on @xmath96 and on @xmath97 via base change @xmath98 . by definition , the @xmath99-orbits @xmath100 of this action naturally correspond to the isomorphism classes of representations @xmath74 in @xmath87 of dimension vector @xmath92 . + by the krull - schmidt theorem , every representation in @xmath87 is isomorphic to a direct sum of indecomposables , unique up to isomorphisms and permutations . thus , knowing the isomorphism classes of indecomposable representations in @xmath87 and their dimension vectors , we can classify the orbits of @xmath101 in @xmath102 . + for certain classes of finite - dimensional algebras , a convenient tool for the classification of the indecomposable representations is the auslander - reiten quiver @xmath103 of @xmath73 . its vertices @xmath104 $ ] are given by the isomorphism classes of indecomposable representations of @xmath73 ; the arrows between two such vertices @xmath104 $ ] and @xmath105 $ ] are parametrized by a basis of the space of so - called irreducible maps @xmath106 . several standard techniques are available for the calculation of @xmath103 , see for example @xcite and @xcite . we will illustrate one of these techniques , namely the use of covering quivers , in subsection [ covsection ] in a situation relevant for our setup . our aim in this section is to translate the classification problem for the action of @xmath0 on @xmath6 into a representation - theoretic one . the following is a well - known fact on associated fibre bundles : [ basicthm ] let @xmath107 be an algebraic group , let @xmath108 and @xmath109 be @xmath110varieties , and let @xmath111 be a @xmath107-equivariant morphism . assume that @xmath109 is a single @xmath107-orbit , @xmath112 . define @xmath113 and @xmath114 . then @xmath108 is isomorphic to the associated fibre bundle @xmath115 , and the embedding @xmath116 induces a bijection between @xmath117-orbits in @xmath118 and @xmath119-orbits in @xmath108 preserving orbit closures . we consider the following quiver , denoted by @xmath47 from now on , \(m ) [ matrix of math nodes , row sep=0.05em , column sep=2em , text height=1.5ex , text depth=0.2ex ] : & & & & & & & + & 1 & 2 & 3 & & n-2 & n-1 & n + ; ( m-1 - 2 ) edge node[above=0.05 cm ] @xmath120 ( m-1 - 3 ) ( m-1 - 3 ) edge node[above=0.05 cm ] @xmath121(m-1 - 4 ) ( m-1 - 6 ) edge node[above=0.05 cm ] @xmath122(m-1 - 7 ) ( m-1 - 7 ) edge node[above=0.05 cm ] @xmath123 ( m-1 - 8 ) ( m-1 - 8 ) edge [ loop right ] node@xmath54 ( m-1 - 8 ) ; together with the ideal @xmath71 generated by the path @xmath124 . we consider the full subcategory @xmath125 of @xmath87 consisting of representations @xmath74 for which the linear maps @xmath126 are injective . corresponding to this subcategory , we have an open subset @xmath127 , which is stable under the @xmath101-action . we consider the dimension vector @xmath128 . [ thm ] there exists a closure - preserving bijection between the set of @xmath0-orbits in @xmath6 and the set of @xmath101-orbits in @xmath129 . consider the subquiver @xmath130 of @xmath47 with @xmath131 and @xmath132 . we have a natural @xmath101-equivariant projection @xmath133 . the variety @xmath134 consists of tuples of injective maps , thus the action of @xmath101 on @xmath135 is easily seen to be transitive . namely , @xmath134 is the orbit of the representation @xmath136 with @xmath137 being the canonical embedding from @xmath138 to @xmath139 . the stabilizer @xmath117 of @xmath140 is isomorphic to @xmath0 , and the fibre of @xmath141 over @xmath140 is isomorphic to @xmath6 . thus , @xmath142 is isomorphic to the associated fibre bundle @xmath143 , yielding the claimed bijection . by the results of the previous section , it suffices to classify the indecomposable representations in @xmath87 to obtain a classification of the orbits of @xmath0 in @xmath6 . we compute the auslander - reiten quiver @xmath144 of @xmath73 using covering theory , which is described in @xcite as mentioned before . we consider the ( infinite ) quiver @xmath145 given by \(m ) [ matrix of math nodes , row sep=0.85em , column sep=2em , text height=1.5ex , text depth=0.2ex ] & & & & & & & + & & & & & & & + : & & & & & & & + & & & & & & & + & & & & & & & + & 1 & 2 & 3 & & n-2 & n-1 & n + ; ( m-2 - 2 ) edge node[above=0.05 cm ] ( m-2 - 3 ) ( m-2 - 3 ) edge node[above=0.05 cm ] ( m-2 - 4 ) ( m-2 - 6 ) edge node[above=0.05 cm ] ( m-2 - 7 ) ( m-2 - 7 ) edge node[above=0.05 cm ] ( m-2 - 8 ) ( m-3 - 2 ) edge node[above=0.05 cm ] ( m-3 - 3 ) ( m-3 - 3 ) edge node[above=0.05 cm ] ( m-3 - 4 ) ( m-3 - 6 ) edge node[above=0.05 cm ] ( m-3 - 7 ) ( m-3 - 7 ) edge node[above=0.05 cm ] ( m-3 - 8 ) ( m-4 - 2 ) edge node[above=0.05 cm ] ( m-4 - 3 ) ( m-4 - 3 ) edge node[above=0.05 cm ] ( m-4 - 4 ) ( m-4 - 6 ) edge node[above=0.05 cm ] ( m-4 - 7 ) ( m-4 - 7 ) edge node[above=0.05 cm ] ( m-4 - 8 ) ( m-1 - 8 ) edge node[right=0.05 cm ] ( m-2 - 8 ) ( m-2 - 8 ) edge node[right=0.05 cm ] @xmath137 ( m-3 - 8 ) ( m-3 - 8 ) edge node[right=0.05 cm ] @xmath146 ( m-4 - 8 ) ( m-4 - 8 ) edge node[right=0.05 cm ] ( m-5 - 8 ) ; with the ideal @xmath147 generated by all paths @xmath148 , and the quiver @xmath149 given by \(m ) [ matrix of math nodes , row sep=0.85em , column sep=2em , text height=1.5ex , text depth=0.2ex ] : & & & & & & & + & & & & & & & + & 1 & 2 & 3 & & n-2 & n-1 & n + ; ( m-1 - 2 ) edge node[above=0.05 cm ] ( m-1 - 3 ) ( m-1 - 3 ) edge node[above=0.05 cm ] ( m-1 - 4 ) ( m-1 - 6 ) edge node[above=0.05 cm ] ( m-1 - 7 ) ( m-1 - 7 ) edge node[above=0.05 cm ] ( m-1 - 8 ) ( m-2 - 2 ) edge node[above=0.05 cm ] ( m-2 - 3 ) ( m-2 - 3 ) edge node[above=0.05 cm ] ( m-2 - 4 ) ( m-2 - 6 ) edge node[above=0.05 cm ] ( m-2 - 7 ) ( m-2 - 7 ) edge node[above=0.05 cm ] ( m-2 - 8 ) ( m-1 - 8 ) edge node[right=0.05 cm ] @xmath54 ( m-2 - 8 ) ; the quiver @xmath145 carries a natural action of the group @xmath150 by shifting the rows , such that @xmath151 . moreover , @xmath149 naturally embeds into @xmath145 , such that the composition of this inclusion with the projection @xmath152 is surjective . by results of covering theory @xcite , we have corresponding maps of the auslander - reiten quivers , namely an embedding @xmath153 and a quotient @xmath154 , such that the composition is surjective . since @xmath149 is nothing else than a dynkin quiver of type @xmath155 , it is routine to calculate its auslander - reiten quiver ( see @xcite ) , and we derive the auslander - reiten quiver @xmath156 just by making the identifications resulting from the action of @xmath150 , which can be read off from the dimension vectors of indecomposable representations . more examples and details concerning the calculation of auslander - reiten quivers using covering theory can also be found in @xcite . + we finally arrive at the picture ( the marked regions have to be identified ) given in figure [ ark ] . the auslander - reiten quiver of @xmath87,scaledwidth=100.0% ] we define the following representations @xmath157 for @xmath158 , @xmath159 for @xmath160 and @xmath161 for @xmath162 in @xmath87 ( graphically represented by dots for basis elements and arrows for a map sending one basis element to another one ) : + @xmath157 for @xmath163 : @xmath164 { 0 & { \xrightarrow}{0 } & \cdots & { \xrightarrow}{0 } & 0 & { \xrightarrow}{0 } & k & { \xrightarrow}{id } & \cdots & { \xrightarrow}{id } & k & { \xrightarrow}{e_1 } & k^{2 } & { \xrightarrow}{id } & \cdots & { \xrightarrow}{id } & k^{2 } \\ & & & & & & \bullet & { \rightarrow } & \cdots & { \rightarrow } & \bullet & { \rightarrow } & \bullet & { \rightarrow } & \cdots & { \rightarrow } & \bullet \\ & & & & & & j & & & & & & i & & & & n\\ & & & & & & & & & & & & \bullet & { \rightarrow } & \cdots & { \rightarrow } & \bullet \\ } ; \path[- > ] ( m-2 - 17 ) edge [ bend left=80 ] ( m-4 - 17 ) ( m-1 - 17 ) edge [ loop right ] node{$\alpha$ } ( m-1 - 17 ) ; \end{tikzpicture}\ ] ] @xmath157 for @xmath165 : @xmath164 { 0 & { \xrightarrow}{0 } & \cdots & { \xrightarrow}{0 } & 0 & { \xrightarrow}{0 } & k & { \xrightarrow}{id } & \cdots & { \xrightarrow}{id } & k & { \xrightarrow}{e_2 } & k^{2 } & { \xrightarrow}{id } & \cdots & { \xrightarrow}{id } & k^{2 } \\ & & & & & & \bullet & { \rightarrow } & \cdots & { \rightarrow } & \bullet & { \rightarrow } & \bullet & { \rightarrow } & \cdots & { \rightarrow } & \bullet \\ & & & & & & i & & & & & & j & & & & n\\ & & & & & & & & & & & & \bullet & { \rightarrow } & \cdots & { \rightarrow } & \bullet \\ } ; \path[- > ] ( m-4 - 17 ) edge [ bend right=80 ] ( m-2 - 17 ) ( m-1 - 17 ) edge [ loop right ] node{$\alpha$ } ( m-1 - 17 ) ; \end{tikzpicture}\ ] ] @xmath166 for @xmath160 : @xmath167 { 0 & { \xrightarrow}{0 } & \cdots & { \xrightarrow}{0 } & 0 & { \xrightarrow}{0 } & k & { \xrightarrow}{id } & \cdots & { \xrightarrow}{id } & k \\ & & & & & & i & & & & n\\ & & & & & & \bullet & { \rightarrow } & \cdots & { \rightarrow } & \bullet \\ } ; \path[- > ] ( m-1 - 11 ) edge [ loop right ] node{$0 $ } ( m-1 - 11 ) ; \end{tikzpicture}\ ] ] @xmath161 for @xmath168 : @xmath167 { 0 & { \xrightarrow}{0 } & \cdots & { \xrightarrow}{0 } & 0 & { \xrightarrow}{0 } & k & { \xrightarrow}{id } & \cdots & { \xrightarrow}{id } & k & { \xrightarrow}{0 } & 0 & { \xrightarrow}{0 } & \cdots & { \xrightarrow}{0 } & 0 \\ & & & & & & i & & & & j & & & & & & n\\ & & & & & & \bullet & { \rightarrow } & \cdots & { \rightarrow } & \bullet & & & & & & \\ } ; \path[- > ] ( m-1 - 17 ) edge [ loop right ] node{$0 $ } ( m-1 - 17 ) ; \end{tikzpicture}\ ] ] here we denote @xmath169 , @xmath170 and @xmath171 . [ uvw ] the representations @xmath157 , @xmath159 and @xmath161 form a system of representatives of the indecomposable objects in @xmath87 . the representations @xmath157 and @xmath159 form a system of representatives of the indecomposable objects in @xmath125 the endomorphism rings of these representations are easily computed to be @xmath172/(x^2)\mbox { for } i\leq j,\ ] ] @xmath173 thus they are indecomposable . their dimension vectors are @xmath174 respectively . these are precisely the dimension vectors appearing in @xmath103 , thus we have found all indecomposables . it is clear from the definition that the indecomposable representations belonging to @xmath125 are the @xmath157 and the @xmath159 . our next aim is to parametrize the isomorphism classes of representations in @xmath125 of dimension vector @xmath92 . as mentioned before , the krull - schmidt theorem states that every representation can be decomposed into a direct sum of indecomposables in an essentially unique way . [ classification ] the isomorphism classes @xmath74 in @xmath125 of dimension vector @xmath92 are in natural bijection to 1 . @xmath2-matrices @xmath175 with entries @xmath176 or @xmath177 , such that @xmath178 for all @xmath179 , 2 . oriented link patterns on @xmath11 , that is , oriented graphs on the set @xmath11 such that every vertex is incident with at most one arrow . moreover , if an isomorphism class @xmath74 corresponds to a matrix @xmath72 under this bijection , the orbit @xmath180 and the orbit @xmath181 correspond to each other via the bijection of lemma [ thm ] . let @xmath74 be a representation in @xmath125 of dimension vector @xmath92 , so @xmath182 for some multiplicities @xmath183 by theorem [ uvw ] . since @xmath184 , we simply need to calculate all tuples @xmath185 such that @xmath186 applying the automorphism @xmath187 of @xmath188 defined by @xmath189 this condition is equivalent to @xmath190 if we fix @xmath191 , this condition states that @xmath192 we can extract an oriented graph on the set of vertices @xmath11 from @xmath193 as follows : for all @xmath158 , we have an arrow from @xmath20 to @xmath19 if @xmath194 . the conditions on @xmath193 ensure that this graph is in fact an oriented link pattern . the matrix @xmath193 is obviously @xmath1-nilpotent . + the decomposition of @xmath74 into indecomposables can be visualized as follows . @xmath195 { & \bullet & \rightarrow & \bullet & \rightarrow & \bullet & \cdots & \bullet & \rightarrow & \bullet & \rightarrow & \bullet & ~~ & 1 \\ & & & \bullet & \rightarrow & \bullet & \cdots & \bullet & \rightarrow & \bullet & \rightarrow & \bullet & ~~ & 2\\ & & & & & \bullet & \cdots & \bullet & \rightarrow & \bullet & \rightarrow & \bullet & ~~ & 3\\ m : & & & & & & & & \vdots & & \vdots & & & \\ & & & & & & & \bullet & \rightarrow & \bullet & \rightarrow & \bullet & ~~ & n-2\\ & & & & & & & & & \bullet & \rightarrow & \bullet & ~~ & n-1\\ & & & & & & & & & & & \bullet & ~~ & n\\ } ; \path[- > ] ( m-1 - 12 ) edge [ bend left=80 ] ( m-3 - 12 ) ( m-7 - 12 ) edge [ bend right=50 ] ( m-2 - 12 ) ( m-6 - 12 ) edge [ bend right=120 ] ( m-5 - 12 ) ; \end{tikzpicture}\ ] ] the arrows in the rightmost column of the diagram allow us to read off the indecomposable direct summands of @xmath74 . namely , @xmath157 is a direct summand of @xmath74 if and only if there is an arrow @xmath196 . if there is no arrow at @xmath64 , the indecomposable @xmath197 is a direct summand of @xmath74 . + shortening the above picture to the rightmost column , @xmath74 corresponds to an oriented link pattern : @xmath198 for a given matrix @xmath10 , we would like to decide to which oriented link pattern it corresponds . define @xmath199 , the span of the first @xmath19 coordinate vectors in @xmath200 , and define a matrix @xmath201 by setting @xmath202 ( we formally define @xmath203 for @xmath204 or @xmath205 ) . the matrix @xmath206 is obviously an invariant for the @xmath0-action on @xmath6 . it is easy to extract an oriented link pattern from @xmath206 as follows : the matrix @xmath72 belongs to the orbit of a matrix @xmath193 as above if and only if @xmath207 or , conversely , @xmath208 for all @xmath158 . by @xmath0-invariance , we just have to compute @xmath206 for @xmath175 as in the previous theorem . we have @xmath209 if and only @xmath210 and there exists @xmath211 such that @xmath212 or , equivalently , such that there exists an arrow @xmath213 in the corresponding oriented link pattern . since both @xmath214 and @xmath215 are spanned by coordinate vectors @xmath216 , we thus have @xmath217 . the second formula follows . we can also rederive theorem [ mel1 ] of a. melnikov : every @xmath218orbit of an upper - triangular @xmath1-nilpotent matrix corresponds to the orbit of a representation in @xmath125 of dimension vector @xmath92 which does not contain @xmath157 for @xmath219 as a direct summand . in this case , the corresponding link pattern consists of arrows pointing in the same direction . we can thus delete the orientation and arrive at a link pattern as in @xcite . + * remark : * our method is easily generalized to obtain a classification of orbits for a more general group action : let @xmath220 be the parabolic subgroup consisting of block - upper triangular matrices with block - sizes @xmath221 . then @xmath222 acts on @xmath6 by conjugation , and the same reasoning as above yields a bijection between @xmath222-orbits in @xmath6 and isomorphism classes of representations in @xmath125 of dimension vector @xmath223 . using the analysis of this section , the @xmath222-orbits in @xmath6 correspond bijectively to matrices @xmath193 such that @xmath224 for all @xmath225 . consequently , they correspond bijectively to `` enhanced oriented link patterns of type @xmath221 '' , namely , to oriented graphs on the set @xmath226 such that the vertex @xmath19 is incident with at most @xmath227 arrows for all @xmath19 . after classifying the orbits via oriented link patterns , we describe the corresponding orbit closures . again , we will solve this problem using results about the geometry of representations of algebras . two theorems of g. zwara are the key to calculating these orbit closures , see @xcite and @xcite for more details . let @xmath74 and @xmath228 be two representations in @xmath87 of the same dimension vector @xmath92 . we say that @xmath74 degenerates to @xmath228 if @xmath229 in @xmath230 , which will be denoted by @xmath231 . since the correspondence of lemma [ thm ] preserves orbit closure relations , we know that @xmath231 if and only if the corresponding @xmath1-nilpotent matrices , denoted by @xmath175 and @xmath232 , respectively , fulfill @xmath233 in @xmath6 . [ zw](zwara ) suppose an algebra @xmath73 is representation - finite , that is , @xmath73 admits only finitely many isomorphism classes of indecomposable representations . let @xmath74 and @xmath228 be two finite - dimensional representations of @xmath73 of the same dimension vector . + then @xmath234 if and only if @xmath235 for every representation @xmath236 of @xmath73 . to simplify notation , we set @xmath237 : = \dim_k \operatorname{hom}(u , v ) $ ] for two representations @xmath236 and @xmath238 . since the dimension of a homomorphism space is additive with respect to direct sums , we only have to consider the inequality @xmath239\leq[u , m']$ ] for indecomposable representations @xmath236 to characterize a degeneration @xmath231 furthermore , since @xmath240=0 $ ] for all representations @xmath74 in @xmath125 by a direct calculation , we can restrict these indecomposables @xmath236 to those of type @xmath157 and @xmath159 of the previous section . + we can easily calculate the dimensions of homomorphism spaces between these indecomposable representations . [ hom ] for @xmath241 we have * @xmath242 = \delta_{i\leq k}$ ] , * @xmath243 = \delta_{i\leq k}$ ] , * @xmath244 = \delta_{i\leq l}$ ] , * @xmath245 = \delta_{i\leq l } + \delta_{j\leq l } \cdot \delta_{i\leq k}$ ] , where @xmath246 for a representation @xmath74 in @xmath125 of dimension vector @xmath92 ( or equivalently , for the corresponding @xmath1-nilpotent matrix @xmath72 ) , consider the corresponding oriented link pattern . define @xmath247 as the number of vertices to the left of @xmath64 which are not incident with an arrow , plus the number of arrows whose target vertex is to the left of @xmath64 . define @xmath248 as @xmath249 plus the number of arrows whose source vertex lies to the left of @xmath250 and whose target vertex lies to the left of @xmath64 . [ pq ] we have @xmath234 ( or equivalently , @xmath233 in the notation above ) if and only if @xmath251 and @xmath252 for all @xmath253 . given two representations @xmath74 and @xmath228 , we write @xmath254 for an indecomposable @xmath236 , the condition @xmath255 \leq [ u , m ' ] $ ] is then equivalent to @xmath256 + \sum_{i=1}^n n_i[u , { \mathcal{v}}_i ] \leq \sum_{i , j=1}^{n } m'_{i , j } [ u , { \mathcal{u}}_{i , j}]+ \sum_{i=1}^n n'_i[u , { \mathcal{v}}_i ] .\ ] ] using the dimensions of homomorphism spaces between indecomposable representations stated in the previous lemma , we calculate @xmath257 and @xmath258 and the condition @xmath239\leq [ u , m']$ ] is equivalent to the conditions @xmath251 and @xmath252 for all @xmath253 . the claimed interpretation of these values @xmath247 and @xmath259 in terms of oriented link patterns follows from theorem [ classification ] . as a next step , we develop a combinatorial method to produce all degenerations of a given representation @xmath74 in @xmath125 of dimension vector @xmath92 out of its oriented link pattern . it is sufficient to construct all minimal degenerations , that is , degenerations @xmath260 such that if @xmath261 , then @xmath262 or @xmath263 . minimal degenerations are denoted by @xmath264 . + in @xcite , g. zwara describes all minimal degenerations ; the result is stated here in a generality sufficient for our purposes . denote by @xmath265 the transitive closure of the relation on representations given by @xmath266 if there exists a short exact sequence @xmath267 such that @xmath268 . let @xmath74 and @xmath228 be representations in @xmath125 . + if @xmath269 , then one of the following holds : 1 . . there are representations @xmath271 , @xmath272 , @xmath273 in @xmath125 such that 1 . @xmath274 2 . @xmath275 3 . @xmath276 4 . @xmath277 is indecomposable . combining this theorem with the technique of ( * * theorem 4 ) , we obtain a characterization of minimal disjoint degenerations , that is , minimal degenerations @xmath264 such that @xmath74 and @xmath228 do not share a common direct summand : let @xmath264 be a minimal disjoint degeneration as before . then either @xmath228 is indecomposable or @xmath278 , where @xmath236 and @xmath238 are indecomposables and there exists an exact sequence @xmath279 or @xmath280 . thus we see that all minimal degenerations are of the form @xmath281 , where @xmath74 and @xmath228 are as in the corollary , thus @xmath228 involves at most two indecomposable direct summands . translating this to the language of oriented link patterns using theorem [ classification ] , we have `` localized '' the problem to the consideration of at most four vertices of an oriented link pattern . in this local case , we can apply theorem [ pq ] and easily work out all minimal degenerations . every minimal degeneration is of the form given in one of the following diagrams showing parts of the degeneration posets in terms of oriented link patterns . we assume that @xmath282 ( resp . @xmath283 , resp . @xmath284 ) are vertices of an oriented link pattern , and only indicate the changes to the arrows incident with one of these vertices ; all other arrows are left unchanged . @xmath285 @xmath286 * remark : * note that , although every minimal degeneration is of the form @xmath281 as above , the choice of @xmath271 is not arbitrary , that is , addition of an arbitrary @xmath271 might lead to a non - minimal degeneration . the precise conditions on @xmath271 neccessary for this degeneration to be minimal will be described in @xcite ; as a consequence , it will be shown in @xcite that all minimal degenerations are of codimension @xmath177 . + we have thus obtained a constructive way of describing an orbit closure @xmath287 of a @xmath1-nilpotent matrix @xmath72 in terms of its corresponding link pattern : by repeated application of the local changes to the arrows as in the theorem , we produce a list of all link patterns corresponding to matrices @xmath288 such that @xmath289 ( although this list will contain repetitions due to non - minimal degenerations ) . in this section , we consider the action of @xmath0 on @xmath8 by conjugation in general ( for the analogous problem of the action of @xmath0 on @xmath290 , see @xcite ) . + the starting point is the following observation ( see @xcite ) : + * example : * consider the action of @xmath291 on @xmath292 via conjugation . then the matrices @xmath293 $ ] for @xmath294 are pairwise non - conjugate . furthermore , on the open set @xmath295 of nilpotent matrices @xmath296 $ ] where @xmath297 or @xmath298 , the map @xmath299 given by @xmath300 is surjective and @xmath291-invariant . + we generalize some aspects of this example to arbitrary @xmath16 . it is appropriate to reformulate the problem as follows : we consider the action of @xmath301 on pairs @xmath302 consisting of a complete flag @xmath303 and a nilpotent operator @xmath304 of an @xmath16-dimensional @xmath64-vector space @xmath238 . then the orbits of this action are precisely the orbits of @xmath0 in @xmath8 since the variety of complete flags is isomorphic to the homogeneous space @xmath305 . 1 . @xmath306 for all @xmath307 , 2 . @xmath308 for all @xmath307 , or , equivalently , all induced maps @xmath309 are invertible , 3 . there exists a unique basis @xmath310 of @xmath238 such that 1 . @xmath311 for all @xmath312 , 2 . @xmath313 for all @xmath314 . obviously , the second property implies the first . we show that the third property implies the second one ; so assume there exists a basis @xmath310 with the properties ( a ) and ( b ) . by an easy induction , we have @xmath315 for all @xmath316 , and @xmath317 if @xmath318 . we thus have @xmath319 and the second property follows since @xmath320 . + conversely , assume that @xmath321 for all @xmath64 . in particular , we have @xmath322 , and thus @xmath323 and @xmath324 for all @xmath64 . we choose an arbitrary basis @xmath325 of @xmath238 which is adapted to @xmath326 , that is , such that @xmath327 for all @xmath64 . then , for all @xmath64 , the elements @xmath328 generate the @xmath64-dimensional space @xmath329 , thus they form a basis of this space . we can thus write the element @xmath330 uniquely as @xmath331 and we define @xmath332 for all @xmath64 . note that the elements @xmath333 do not depend on the choice of basis elements @xmath325 . we have @xmath334 : otherwise @xmath335 , and application of @xmath336 yields @xmath337 and thus @xmath338 for all @xmath19 by linear independence of the elements @xmath339 . then @xmath340 , a contradiction . since the elements @xmath341 form a basis and the @xmath342 are non - zero , the elements @xmath333 form a basis , too , which is again adapted to @xmath326 . + we have @xmath343 by definition . for @xmath318 , we thus have @xmath344 it follows that @xmath345 belong to @xmath346 , thus they form a basis of this space for dimension reasons . it also follows that @xmath347 form a basis of @xmath348 . + writing @xmath349 , we apply @xmath350 and calculate @xmath351 @xmath352 and thus @xmath353 and @xmath354 for all @xmath355 by linear independence of @xmath356 . we thus have @xmath357 for all @xmath316 , and , in particular , @xmath358 for all @xmath64 . the basis @xmath310 thus has the claimed properties . 1 . for @xmath307 , the first @xmath64 columns of @xmath365 are linearly independent , 2 . for @xmath307 , the minor @xmath366 is non - zero , 3 . @xmath72 is @xmath0-conjugate to a unique matrix @xmath117 such that @xmath367 for @xmath368 and @xmath369 for all @xmath370 . we apply the previous theorem to the vector space @xmath371 with coordinate basis @xmath372 , the standard flag defined by @xmath373 and the endomorphism @xmath336 given by multiplication by @xmath72 . the first property of the theorem immediately translates into linear independence of column vectors , whereas the second property translates to the non - vanishing of minors . the basis @xmath310 of the theorem yields an upper - triangular base change matrix , and representing @xmath72 with respect to this basis yields the desired @xmath0-conjugate @xmath117 . + we construct a class of determinantal @xmath0-semiinvariants on @xmath8 , that is , regular functions @xmath374 on @xmath8 such that @xmath375 for all @xmath376 and @xmath360 ; here @xmath377 is a character on @xmath0 called the weight of @xmath374 . for @xmath179 , we denote by @xmath378 the character defined by @xmath379 ; the @xmath380 form a basis for the group of characters of @xmath0 . + fix non - negative integers @xmath381 such that @xmath382 . moreover , fix polynomials @xmath383 $ ] for @xmath384 and @xmath385 , and denote the datum @xmath386 by @xmath387 . for all such @xmath19 and @xmath20 , consider the @xmath388-submatrices @xmath389 as defined in the previous section , and form the block matrix @xmath390 ; this is a @xmath391-matrix . for @xmath376 and @xmath395 , denote by @xmath396 ( resp . by @xmath397 ) the submatrix formed by the last @xmath362 rows and columns ( resp . by the first @xmath363 rows and columns ) of @xmath398 . with these definitions , it follows immediately that @xmath399 this yields the following equalities of block matrices @xmath400 @xmath401 and thus @xmath402 with the aid of these semiinvariants , we can see that the entries of the normal form @xmath117 associated to a matrix @xmath72 fulfilling the conditions of corollary [ hnf ] depend polynomially on @xmath72 , by describing them as the value of a special semiinvariant @xmath393 : for @xmath19 and @xmath20 such that @xmath403 and @xmath404 , consider the datum @xmath387 as above defined by @xmath405 , @xmath406 , @xmath407 , @xmath408 , @xmath409 , @xmath410 , @xmath411 , @xmath412 . then , for a matrix @xmath117 in the form of corollary [ hnf ] , we have @xmath413 . by a direct calculation , the matrix @xmath414 consists of the blocks @xmath415 @xmath416 thus , the matrix @xmath414 is lower triangular , all diagonal entries being @xmath177 except the @xmath417-entry , which equals @xmath418 . it seems likely that the semiinvariants @xmath393 generate the ring of all semiinvariants at least for a certain cone of weights . the generic normal form of corollary [ hnf ] allows to find identities between the @xmath393 by evaluation on matrices @xmath117 in normal form .
the orbits of the group @xmath0 of upper - triangular matrices acting on @xmath1-nilpotent complex matrices via conjugation are classified via oriented link patterns , generalizing a. melnikov s classification of the @xmath0-orbits on upper - triangular such matrices . the orbit closures as well as the `` building blocks '' of minimal degenerations of orbits are described . the classification uses the theory of representations of finite - dimensional algebras . furthermore , we initiate the study of the @xmath0-orbits on arbitrary nilpotent matrices .
ms & ph supervised the project , suggested the idea and its physical implementation . they provided the discussion of results and wrote up the paper . ar participated in discussions , suggested and proved the technical version of the theorem . ab developed numerical methods and analytical computations supporting the proposition in discussion with ms . ab prepared the figures . 99 einstein a. , podolsky b. & rosen n. can quantum mechanical description of physical reality be considered complete ? _ physical review _ * 47 * , 777780 ( 1935 ) . junge m. , palazuelos c. , prez - garca d. , villanueva i. & wolf m. m. , operator space theory : a natural framework for bell inequalities , _ phys . lett . _ * 104 * , 170405 ( 2010 ) . junge m. & palazuelos c. large violation of bell inequalities with low entanglement , _ comm . phys . _ * 306 * , 695746 ( 2011 ) . buhrman h. , regev o. , scarpa g. & wolf r. near - optimal and explicit bell inequality violations , _ theor . comput . _ * 8 * , 623645 ( 2012 ) . brit j. & vidick t. , explicit lower and upper bounds on the entangled value of multiplayer xor games , _ commun . phys . _ * 321 * , 181207 ( 2013 ) . pawlowski m. & brukner c. monogamy of bell s inequality violations in non - signaling theories , _ phys . lett . _ * 102 * , 030403 ( 2009 ) . weiss w. , benenti g. , casati g. , guarneri i. , calarco t. , paternostro m. & montangero s. violation of bell inequalities in larger hilbert spaces : robustness and challenges , _ new j. phys . _ * 18 * , 013021 ( 2016 ) . marciniak m. , rutkowski a. , yin z. , horodecki m. & horodecki r. , unbounded violation of quantum steering inequalities , _ phys . lett . _ * 115 * , 170401 ( 2015 ) . schrdinger e. discussion of probability relations between separated systems , _ math . soc . _ * 31 * , 555563 ( 1935 ) . schrdinger e. probability relations between separated systems , _ math . . soc . _ * 32 * , 446452 ( 1936 ) . gisin n. , stochastic quantum dynamics and relativity , _ helv . . acta _ * 62 * , 363371 ( 1989 ) . hughston l. p. , jozsa r. & wootters w. k. , a complete classification of quantum ensembles having a given density matrix , _ phys . lett . a _ * 183 * , 1418 ( 1993 ) . wiseman h. m. , jones s. j. & doherty a. c. , steering , entanglement , nonlocality , and the einstein podolsky rosen paradox , _ phys . lett . _ * 98 * , 140402 ( 2007 ) . oppenheim j. & wehner s. , the uncertainty principle determines the nonlocality of quantum mechanics , _ science _ * 330 * , 10721074 ( 2010 ) . ramanathan r. , goyeneche d. , mironowicz p. & horodecki p. , the uncertainty principle does not entirely determine the non - locality of quantum theory , arxiv.1506.05100 ( 2015 ) . wittmann b. , ramelow s. , steinlechner f. , langford n. k. , brunner n. , wiseman h. m. , ursin r. & zeilinger a. , loophole - free einstein podolsky rosen experiment via quantum steering , _ new j. phys . _ * 14 * , 053030 ( 2012 ) . fuwa m. , takeda sh . , zwierz m. , wiseman h. m. & furusawa a. , experimental proof of nonlocal wavefunction collapse for a single particle using homodyne measurements , _ nat . commun . _ * 6 * , 6665 ( 2015 ) . smith d. h. , gillett g. , de almeida m. p. , branciard c. , fedrizzi a. , weinhold t. j. , lita a. , calkins b. , gerrits t. , wiseman h. m. , nam s. w. & white a. g. , conclusive quantum steering with superconducting transition - edge sensors , _ nat . commun . _ * 3 * , 625 ( 2012 ) . sauders d. j. , jones s. j. , wiseman h. m. & pryde g. j , experimental epr - steering using bell - local states , _ nature phys . _ * 6 * , 845849 ( 2010 ) . cavalcanti d. , skrzypczyk p. , aguilar g. h. , nery r. v. , souto ribeiro p. h. & walborn s. p. , detection of entanglement in asymmetric quantum networks and multipartite quantum steering , _ nat . _ * 6 * , 7941 ( 2015 ) . tavakoli a. , hameedi a. , marques b. & bourennane m. quantum random access codes using single @xmath1-level systems , _ phys . lett . _ * 114 * , 170502 ( 2015 ) . maassen h. & uffink j. b. m. , generalized entropic uncertainty relations , _ phys . _ * 60 * , 1103 ( 1988 ) . deutsch d. , uncertainty in quantum measurements , _ phys . lett . _ * 50 * , 631 ( 1983 ) . goyeneche d. a new method to construct families of complex hadamard matrices in even dimensions , _ j. math . * 54 * , 032201 ( 2013 ) . bandyopadhyay s. , boykin p. o. , roychowdhury v. & vatan f. , a new proof of the existence of mutually unbiased bases , _ algorithmica _ , * 34 * , 512528 , ( 2002 ) . li c .- m . , chen l .- y . & yabushita a. , experimental verification of multidimensional quantum steering , arxiv:1602.07139 ( 2016 ) . rosoek k. , stobiska m. , wieniak m. & ukowski m. , two copies of the einstein rosen state of light lead to refutation of epr ideas , _ phys . lett . _ * 114 * , 100402 ( 2015 ) . stobiska m. , tppel f. , sekatski p. , buraczewski a. , ukowski m. , chekhova m. v. , leuchs g. & gisin n. filtering of the absolute value of photon - number difference for two - mode macroscopic quantum superpositions , _ phys . a _ * 86 * , 063823 ( 2012 ) . stobiska m. , tppel f. , sekatski p. & buraczewski a. towards loophole - free bell inequality test with preselected unsymmetrical singlet states of light , _ phys . rev . a _ * 89 * , 022119 ( 2014 ) . berry , d. , jeong , h. , stobiska , m. & ralph t. c. fair - sampling assumption is not necessary for testing local realism , _ phys . rev . a _ * 81 * , 012109 ( 2010 ) . goh k. t. , bancal j .- d . & scarani v. quantifying the amount of entanglement based only on the hilbert space dimension , arxiv:1509.08682 ( 2015 ) . iskhakov t. sh . , chekhova m. v. , rytikov g. o. & leuchs g. macroscopic pure state of light free of polarization noise , _ phys . lett . _ * 106 * , 113602 ( 2011 ) . iskhakov t. sh . , agafonov i. n. , chekhova v. & leuchs g. polarization - entangled light pulses of @xmath148 photons , _ phys . lett . _ * 109 * , 150502 ( 2012 ) . pan j .- w . , chen z .- b . , lu c .- y . , weinfurter h. , zeilinger a. & ukowski m. multiphoton entanglement and interferometry , _ rev . _ * 84 * , 777 ( 2012 ) . rdmark m. , ukowski m. & bourennane m. experimental test of fidelity limits in six - photon interferometry and of rotational invariance properties of the photonic six - qubit entanglement singlet state , _ phys . rev . * 103 * , 150501 ( 2009 ) . wieczorek w. , schmid ch . , kiesel n. , pohlner r. , ghne o. & weinfurter h. experimental observation of an entire family of four - photon entangled states , _ phys . lett . _ * 101 * , 010503 ( 2008 ) . lita a. e. , miller a. j. & nam s. w. counting near - infrared single - photons with 95% efficiency , _ opt . express _ * 16 * , 3032 ( 2008 ) . gerrits t. , calkins b. , tomlin n. , lita a. e. , migdall a. , mirin r. & nam s. w. extending single - photon optimized superconducting transition edge sensors beyond the single - photon counting regime , _ opt . express _ * 20 * , 23798 ( 2012 ) . mair a. , vaziri a. , weihs g. & zeilinger a. entanglement of the orbital angular momentum states of photons , _ nature _ * 412 * , 313316 ( 2001 ) . bobrov i. , kovlakov e. , markov a. , straupe s. & kulik s. tomography of spatial mode detectors , _ opt . express _ * 23 * , 649654 ( 2015 ) . krenn m. , handsteiner j. , fink m. , fickler r. & zeilinger a. twisted photon entanglement through turbulent air across vienna , _ proc . sci . _ * 112 * , 14197 ( 2015 ) . milne a. , jevtic s. , jennings d. , wiseman h. & rudolph t. quantum steering ellipsoids , extremal physical states and monogamy , _ new j. phys . _ * 16 * , 083017 ( 2014 ) . gheorghiu a. , wallden p. , kashefi e. rigidity of quantum steering and one - sided device - independent verifiable quantum computation , arxiv:1512.07401 ( 2015 ) . massar s. , pironio s. & pitala - garca d. hyperdense coding and superadditivity of classical capacities in hypersphere theories , _ new j. phys . _ * 17 * , 113002 ( 2015 ) . konno n. , limit theorem for continuous - time quantum walk on the line , _ phys . * 72 * , 026113 ( 2005 ) . watabe k. , kobayashi n. , katori m. & konno n. , limit distributions of two - dimensional quantum walks , _ phys . rev . a _ * 77 * , 062331 ( 2008 ) . pahor m. , largest coefficients in binomial expansions , _ australian senior mathematics journal _ * 20 * , 5359 ( 2006 ) . * supplementary information : + * quantum steering inequality with tolerance + for measurement - setting - errors : + experimentally feasible signature of unbounded violation _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the supplementary information comprises detailed proofs of the theorem and the proposition from the main text , as well as an elementary argument that any unbounded violation of bell inequalities requires divergence of at least one party s settings . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ given a quantum steering scenario involving @xmath33 settings , @xmath34 outcomes , and the set of orthonormal eingenbases @xmath149 defining the receiver s ( bob s ) measurements , the lhs steering functional is bounded from above @xmath150 where @xmath37 and and where @xmath151 for @xmath39 as in the deutsch maassen uffink uncertainty relations . this implies : @xmath152 from the above in particular the weaker variant follows , namely @xmath153 where @xmath154 . _ first , we will compute the quantum value @xmath155 for the steering functional @xmath70 defined as a set @xmath156 . in our case it is enough to show that @xmath157 , where @xmath15 is the number of bases @xmath158 second , we will describe a general method of computation of the classical bound @xmath159 . let @xmath160 . then , the following inequality holds @xmath161 where @xmath162 . for any @xmath163 , let @xmath164 . using the purification of @xmath165 and its schmidt decomposition , we can show that @xmath166 in further considerations we will omit the index @xmath163 . let us define the shift operator @xmath167 acts on the bases vectors in the following way : @xmath168 and observe that @xmath169 every element of @xmath170 is equal to @xmath171 . we decompose @xmath139 in the following way @xmath172 where @xmath173 and the set @xmath174 . next , we use the following fact @xmath175 hence , in order to estimate the norm of @xmath139 we have to estimate the norm of @xmath176 . this is just the maximal singular value of @xmath176 or , equivalently , the maximal eigenvalue of @xmath177 , squared . since this operator is block diagonal ( @xmath178 are permutation operators ) , we have to calculate the maximal singular value of @xmath179 , taking into account the proper index of @xmath52 . to this end , let us estimate this singular value of @xmath179 , which possess the following general form @xmath180 where @xmath181 and @xmath182 while @xmath183 are phases for given indices @xmath12 , @xmath99 , @xmath49 and @xmath50 . this results in @xmath184 here we use the fact that @xmath185 . this means that @xmath186 , where @xmath187 are eigenvalues of @xmath188 . let us denote the maximal eigenvalue as @xmath189 . from ( [ eq : supplmat_re ] ) we obtain the maximal singular value of @xmath179 , @xmath190 @xmath191 therefore , in order to estimate the norm of @xmath139 we must calculate the maximal absolute value of the overlap between vectors @xmath192 of the basis given by the number @xmath142 . then , the norm @xmath193 where @xmath194 ( @xmath46 is just mussen - uffink value for each @xmath52 ) . let us observe that for @xmath195 it is just identity transformation between these two bases ( it corresponds to the case @xmath196 ) hence @xmath197 and @xmath198 finally , the violation of the steering inequality @xmath199 which ends the proof of main theorem . in the main text we assume the worst case to calculate the norm of @xmath139 @xmath200 where @xmath201 for @xmath202 . thus , @xmath203 at the end of our consideration let us define the set of matrices @xmath204_{a , b=1}^{d}$ ] and consider the new matrix which has the form @xmath205 let us notice that matrix @xmath206 is just matrix @xmath139 without probabilities @xmath207 and @xmath208 , then we can formulate the main theorem in the stronger form ( in the sense of the shorter summation ) : observation : if for the matrix @xmath139 the matrix @xmath206 has a block toeplitz form namely @xmath209 for any @xmath210 then the inequality for @xmath211 takes the form @xmath212 where @xmath213 and @xmath214 , @xmath215 are ceiling and floor functions respectively . here we provide an explanation of this observation instead of the formal proof . this observation comes from symmetry of the block toeplitz hermitian matrix @xmath206 what implies @xmath216 for @xmath217 . let us explain in detail what we mean by symmetry in this case . the reader should notice that we have @xmath218 sets @xmath219 and the fundamental fact is that due to toeplitz form each of the sets is formed of the two matrices and each of them are repeated some number of times , we may , slightly abusing the notation , write @xmath220 for the toeplitz matrices . now , due to the hermicity of the global matrix @xmath206 which determines that @xmath221 and only the sets @xmath222 for @xmath223 ( if @xmath15 is odd otherwise we have to add one additional term in summation , namely @xmath224 ) should be taken into account , while calculating the quantities @xmath225 since any set for higher @xmath52 reproduces one of the previous ones modulo hermitian conjugate of its elements , which does not change the singular values of the matrices contained in the set . in this section we recall an elementary argument that any unbounded violation of bell inequalities requires divergence of at least one party s settings ( here this is bob , but considering alice makes no difference ) . quantum monogamy relation says that if we consider a no - signaling box @xmath226 among @xmath227 parties @xmath228 , where each bob has @xmath229 observables at his disposal , the following monogamy relation holds for bell functionals @xmath230 @xcite : @xmath231 with the original bipartite bell inequality for alice and bob defined as @xmath232 . let us assume that the bell functionals are non - negative . then , taking an arbitrary extension of an arbitrary no - signaling bipartite box @xmath233 , @xmath234 ( in a sense that averaging @xmath235 over the @xmath236 pairs of variables @xmath237 reproduces @xmath238 ) , gives @xmath239 due to non - negativity of the functional . note that minimizing over all the extensions within a given sub - theory of a no - signaling theory would give a bound within that `` subtheory '' ( including quantum mechanics ) . thus , for any no - signaling bipartite @xmath233 , one has the bell - violation rate bounded by @xmath240 . since any bipartite quantum statistics represents a no - signaling box , it must obey the same bound . _ given a set of @xmath15 bob s measurement bases @xmath241 with @xmath112 and @xmath113 , defined by some set of angles @xmath114 , the maximal overlap @xmath115 equals the maximal overlap between @xmath116 and @xmath117 with @xmath118 : @xmath242 where @xmath120@xmath122 goes to zero as fast as @xmath123{d}$ ] . _ let us consider a two - mode state expressed in @xmath243 basis , which is rotated by the angle @xmath147 @xmath244 then , the state @xmath245 in @xmath246 basis takes the following form in @xmath243 @xmath247 \dfrac{1}{\sqrt{(d - m)!\,m ! } } \sum_{p=0}^{d - m } \sum_{q=0}^m & \binom{d - m}{p}\binom{m}{q}(-1)^q ( \cos\theta)^{m+p - q } ( \sin\theta)^{d - m - p+q}\times{}\\ { } \times{}&\sqrt{(p+q)!(d - p - q ) ! } \ket{(p+q)_h , ( d - p - q)_v}. \end{aligned}\ ] ] overlap with @xmath248 equals to @xmath249}^{\min\bigl[(d - n),2(d - m)-(d - n)\bigr ] } \kern0.5em \binom{d - m}{\tfrac{d - n+k}{2}}\binom{m}{\tfrac{d - n - k}{2}}(-1)^{\tfrac{d - n - k}{2 } } ( \cos\theta)^{m+k } ( \sin\theta)^{d - m - k},\ ] ] where @xmath250 symbol denotes sum of @xmath251 for index @xmath1 increasing by 2 , i.e. @xmath252 [ cols="^,^,^,^ " , ] fig . [ fig : suppl_overlap ] depicts @xmath253 for @xmath254 and several values of @xmath147 . numerics indicates that @xmath253 reaches its all maximal values only for @xmath255 or @xmath256 and they are equal . therefore , we focus exclusively at @xmath257 and @xmath258 . for @xmath258 : @xmath259 and for @xmath257 : @xmath260 in order to find maximal values of @xmath261 and @xmath262 one should note that they are just binomial coefficients of the form @xmath263 thus , they reach maximal values for @xmath264 ( see e.g. @xcite ) . thus , for @xmath258 @xmath265 and for @xmath257 @xmath266 which gives also @xmath267 for given @xmath1 and @xmath147 . finally , let us notice that @xmath268 where @xmath269 . the second step is to find how fast @xmath122 goes to zero . for this task , the floor function might be omitted , thus @xmath270 for large @xmath1 . the binomial coefficient can be approximated with stirling s formula @xmath271 in the following way @xmath272 \binom{d}{d\sin^2\theta } = { } & \dfrac{d!}{(d\sin^2\theta)!\,(d - d\sin^2\theta ) ! } = \dfrac{d!}{(d\sin^2\theta)!\,(d\cos^2\theta ) ! } \\ { } \approx { } & \dfrac{(d / e)^d\sqrt{2\pi d}}{(d / e\sin^2\theta)^{d\sin^2\theta}\sqrt{2\pi d\sin^2\theta}(d / e\cos^2\theta)^{d\cos^2\theta}\sqrt{2\pi d\cos^2\theta } } \\ { } = { } & \dfrac{1}{\sqrt{2\pi d}\bigl[(\sin^2\theta)^{\sin^2\theta}(\cos^2\theta)^{\cos^2\theta}\bigr]^d\sin\theta\cos\theta}. \end{aligned}\ ] ] therefore , @xmath272 \sqrt{\binom{d}{q_{\theta , d } } } \approx { } & \dfrac{1}{\sqrt[4]{2\pi d}\bigl[(\sin\theta)^{\sin^2\theta}(\cos\theta)^{\cos^2\theta}\bigr]^{d}\sqrt{\sin\theta\cos\theta}}. \end{aligned}\ ] ] additionally , @xmath273^{d}.\ ] ] this leads to the following formula @xmath274^{d}}{\sqrt[4]{2\pi d}\bigl[(\sin\theta)^{\sin^2\theta}(\cos\theta)^{\cos^2\theta}\bigr]^{d}\sqrt{\sin\theta\cos\theta } } = \dfrac{1}{\sqrt[4]{2\pi d}\sqrt{\sin\theta\cos\theta } } = \dfrac{1}{\sqrt[4]{\tfrac{\pi}{2 } d}\sqrt{\sin2\theta } } \sim \dfrac{1}{\sqrt[4]{d}}.\ ] ] here we present our numerical method applied to find the experimental settings allowing to observe violation of the quantum steering inequality . the goal is to compute the optimal number of settings @xmath128 and the set of angles @xmath106 ( @xmath275 ) and @xmath276 which would maximize @xmath126 given by eq . ( [ supeq : v_q ] ) . maximal value of @xmath126 ensures that @xmath277 also reaches its maximum . the denominator of @xmath126 , @xmath138 , depends on the sum of @xmath279 which has to be minimized . therefore , our first step is based on analysis of @xmath122 ( eq . [ supeq : c ] ) , depicted in fig . [ supfig : c ] as a function of @xmath147 for @xmath278 . this function is periodical with period @xmath280 and symmetrical with respect to @xmath281 . it always reaches @xmath1 minima , one or two of them being global . in case of odd @xmath282 , the global minimum is achieved for @xmath281 . in contrast , even @xmath283 regularly lead to two equal global minima , located symmetrically to the left and right of @xmath281 . interestingly , all minima found for @xmath1 are always higher than minima for @xmath5 and lower than for @xmath284 . in case of odd @xmath1 we used @xmath285 and @xmath286 . we numerically optimized @xmath287 and found best @xmath288 , thus @xmath289 . this result was confirmed analytically for @xmath131 and @xmath92 . in case of this set is not the best possible , since it may not include the global minima , but allows us to assure that @xmath290 . the maximal number of settings @xmath15 is limited by the number of settings of the stepper motor driving the polarizer , usually around 400 per full rotation . in the second step our algorithm goes through all @xmath291 , computes @xmath126 for given @xmath1 and finds the maximal value . the result is a pair @xmath292 . the bounds for the lossy case are then computed as @xmath293 figs . 45 in the main text display @xmath126 and @xmath130 computed with the above method . it is worth noting that the erratic shape of @xmath129 is not accidental and results from eq . ( [ supeq : c ] ) . generally , the higher @xmath1 , the more settings have to be used in order to maximize @xmath126 . however , this relation is not monotonic and allows to find higher values of @xmath1 for which it is possible to use smaller @xmath128 . this feature allows to better suit the limitations of the experimental setup .
quantum steering is a relatively simple test for quantumness of correlations , proving that the values of quantum - mechanical measurement outcomes come into being only in the act of measurement . by exploiting quantum correlations alice can influence steer bob s physical system in a way inaccessible in classical world , leading to violation of some inequalities . demonstrating this and similar quantum effects for systems of increasing size , approaching even the classical limit , is a long - standing challenging problem . here we provide experimentally feasible signature of unbounded violation of a steering inequality . we derive its universal form where tolerance for measurement - setting - errors is explicitly build - in by means of the deutsch - maassen - uffink uncertainty relation . then , generalizing the mutual unbiasedness , we apply the inequality to the multi - singlet and multi - particle bipartite bell - state . however , the method is general and opens the possibility of employing multi - particle bipartite steering for randomness certification and development of quantum technologies , e.g. random access codes . in their famous paper , einstein , podolsky and rosen ( epr ) highlighted the phenomenon of entanglement @xcite : it is possible to see perfect correlations between measurement outcomes obtained by two observers , alice and bob , at distant locations , while for each observer his / her outcomes appear to be statistically random . these are the epr correlations . validation of entanglement requires designing a specific experimental scenario where measurements on a quantum state give outcomes which violate a classical inequality . the inequality can be constructed , for example , on the basis of probability distribution satisfying the kolmogorov axioms . its unbounded violation is equivalent to observation of the epr correlations which become more and more pronounced when size of a system increases , reaching even classical limit of macroscopic population . this is very challenging to accomplish in the paradigm of bell - nonlocality testing : if specific observables with @xmath0 settings and @xmath1 possible outcomes are used , bipartite quantum states with local hilbert space dimension @xmath1 can violate a bell inequality by a factor of @xmath2 @xcite , later improved to @xmath3 @xcite . thus , an unbounded violation of a bell inequality requires exponentially many observables ( or equivalently , settings ) . according to the monogamy relation @xcite , this scaling can be improved only up to the linear one ( see the supplementary information ) . however , the present results are still far from this limit and this makes them purely academic as far as the experimental perspective is concerned @xcite . in case of quantum steering , the task has been found to be less difficult : violation of a steering inequality by a factor of @xmath4 requires @xmath5 observables in the form of mutually unbiased bases ( mubs ) @xcite . however , this scenario necessitates the complementarity relation among the bases to be fulfilled exactly , which is experimentally impossible to attain . here we provide a general tight formulation of a quantum steering inequality in which , for the first time , tolerance to the errors of measurement settings is incorporated automatically . this allows for taking a significant step towards feasibility of experimental observation of a signature of unbounded violation of a steering inequality for a multi - particle bipartite bell - singlet state . in particular , the method enables constructing an inequality which is robust with respect to the degree of unbiasedness of bases and it employs generalized mubs , where unbiasedness may vary from one observable to another or even form one eigenvector element to another . the inequality is governed by the very parameters involved in the deutsch - maassen - uffink uncertainty relations . our formula can be adapted to various physical systems but , since mesoscopic quantum effects are very subtle , technological requirements for an experimental apparatus are high . here we focus on quantum - optical implementation based on polarization entangled squeezed vacuum states generated by parametric - down - conversion sources , polarization rotations and photon counting detection . quantum steering was originally introduced by schrdinger in 1935 @xcite and it was elaborated in the course of time @xcite . recently , it gained a lot of attention because it has been shown to be a weaker test of quantumness of correlations than bell inequalities thus , less experimentally demanding @xcite . it is different from it in that it quantifies discrepancy between the full quantum - mechanical treatment and the situation where only one of the observers is performing classical measurements ( e.g. alice ) . the latter amounts to the local hidden state model ( lhs ) , where the assumption of objective existence of measurement outcomes without taking any measurements bounds alice in the possibility of steering bob s state . however , intriguingly , steering is intimately linked to the bell nonlocality . in the case of simple chsh - bell inequality it is directly limited by the strength of some fine - grained uncertainty relations @xcite . in a general scenario this relation is more complex @xcite . in our result we see a connection between multi - particle bipartite steering and the deutsch maassen uffink entropic uncertainty relations . violation of steering inequalities has been confirmed in numerous experimental demonstrations involving a single photon @xcite , a two - photon singlet @xcite or a werner state @xcite . since quantum steering can be formulated as a quantum - information task where the classical measurements simulate an untrusted device , it has been extended to a multipartite scenario useful for semi - device - independent entanglement certification in quantum networks @xcite . our generalization of mubs will enable developments in the young field of quantum random access codes by allowing for more involved and optimal scenarios than those discussed until now @xcite . _ main result . _ let us consider a quantum steering scenario shown in fig . [ fig : scheme ] . alice and bob have locally access to subsystems of a bipartite system described by a quantum state @xmath6 . alice chooses one of her settings @xmath7 , measures a nondegenerate observable @xmath8 with eigenvectors @xmath9 and receives a result @xmath10 with probability @xmath11 . only after alice has collected the result @xmath12 , the following conditional state @xmath13 was `` created at a distance '' at bob s location . the set @xmath14 is steerable if for every @xmath12 bob can find a basis of @xmath15 states @xmath16 which is identical to @xmath14 . this is possible only if @xmath6 is maximally entangled . the degree of the match between @xmath16 and @xmath14 for a given @xmath6 is measured by a steering functional : @xmath17 where @xmath18 denotes the set of all bipartite quantum states . the maximal quantum value of ( [ sq ] ) equals the number of settings , @xmath19 , and reveals the perfect match . within the lhs model alice s measurements are considered to be untrusted , i.e. they are treated as a black box that receives inputs ( ) and reports outcomes ( ) . the state @xmath6 is now of a hybrid classical - quantum nature and it determines the form of the bob s conditional state as an average over ensemble of local hidden states @xmath20 , which existence is independent of taking any measurements by the observers @xcite : @xmath21 the steering functional equals @xmath22 where @xmath23 denotes the set of all bipartite states fitting to the lhs model . its maximal value depends on the choice of bob s bases @xmath16 and will be estimated below . quantum violation of our steering inequality is observed if @xmath24 we say that it is unbounded if the ratio @xmath25 between the quantum and classical value of the steering functional is an increasing function of some experimental parameters , for example of the amount of entanglement in @xmath6 or of a number and characteristics of the measured observables . an unbounded violation of ( [ vq ] ) has been shown for a sequence of maximally entangled states with an increasing local dimension @xmath1 ( e.g. maximally entangled states of larger and larger spin ) @xcite . however , it necessitates employing very complicated measurements at bob s side the mutually unbiased bases ( mubs ) with linearly diverging number of settings @xmath26 . in general , for a large dimension their existence is known only in same special cases and precise verification of their defining feature that @xmath27 for @xmath28 is experimentally very challenging . rather than trying to cure these problems our method shall simply overcome them . to this end , we will relax the mub condition by assuming near - uniform overlaps and will seek for the maximal overlap between the bases . in this way , our result is naturally linked to the deutsch maassen uffink uncertainty relations @xcite saying that the shannon entropies of measurement outcomes of two nondegenerate observables obtained in two bases @xmath16 and @xmath29 for a quantum state @xmath30 satisfy @xmath31 where the maximal overlap @xmath32 is a quantitative measure of complementarity of the bases . the inequality is nontrivial if the two observables do not commute . our main result is the following _ theorem . given a quantum steering scenario involving @xmath33 settings , @xmath34 outcomes , and a set of @xmath15 orthonormal eingenbases @xmath35 defining the receiver s ( bob s ) measurements , the lhs steering functional is bounded from above @xmath36 where @xmath37 and @xmath38 for @xmath39 is defined as in the deutsch maassen uffink uncertainty relations this implies : @xmath40 in particular , a weaker bound can be derived : @xmath41 with @xmath42 . _ let us observe that any state @xmath6 leading to @xmath43 , where @xmath44 is given by ( [ sq ] ) but without taking supremum over @xmath18 , violates the steering inequality ( [ steeringineq ] ) . the dependence solely on the maximal overlap @xmath45 frees us from the necessity of a precise control of all mub conditions separately , as it was done in the approaches taken before ours @xcite . while the physical interpretation of the parameter @xmath45 is very clear , that of @xmath46 is less comprehensible . @xmath45 simply quantifies the two most `` compatible '' eigenvectors of the measured observables and equals the maximal element of the @xmath47 @xmath48 matrix with columns and rows enumerated by the settings indexes @xmath49 and @xmath50 . alternatively one may say that @xmath45 picks the two particular bob s measurements settings for which the deutsch maassen uffink relation exhibits the weakest uncertainty . interestingly , our mathematical techniques allow us finding the classical bound if we analyze this matrix along its `` diagonals '' consisting of the elements with indexes @xmath51 , and @xmath46 denotes the maximal element in such `` @xmath52th diagonal '' . now we shall define the relaxed mub condition by introducing the maximal overlap equal @xmath53 with @xmath54 , where @xmath55 corresponds to the original mub measurements . any quantum bases satisfying this condition form a generalized mub . let us now examine violation of inequality ( @xmath56 ) for this @xmath57-generalized mub condition . unbounded violation is possible whenever the number of settings @xmath58 is an increasing function of @xmath1 but , now it can increase arbitrarily slowly . then the ratio @xmath59 and it diverges as @xmath60 if @xmath1 goes to infinity . for @xmath55 the behavior @xmath61 reproduces the result reported in @xcite but , without the need for the linear scaling @xmath62 . equivalently , considering the dependence on the number of settings by putting @xmath63 , we obtain at most @xmath64 in the limit of large @xmath15 . however , the best this steering inequality may offer is @xmath65 , which coincides with the maximal possible degree of bell inequality violation @xmath66 . indeed , it follows immediately from the fact that @xmath67 which is easy to derive by considering alice s and bob s measurement statistics on a pure product state ( independently on the number of settings and outcomes ) . explanation of the discrepancy between the bounds @xmath68 and @xmath15 calls for further research , but it might hold true that the present @xmath68 scaling is optimal and addresses the issue of optimality of the mub settings . we now would like to turn our abstract mathematical result into a form which could be tested in a laboratory . we first consider a source of independent pairs of photons entangled in their polarizations , i.e. many copies of singlet states @xmath69 . we assume a single pair fidelity @xmath70 , efficiency of each detector at alice side @xmath71 , which is required for conclusive version of a steering experiment ( see @xcite ) , and the relaxed mub condition @xmath72 . in this case the local dimension of the state equals @xmath73 ( allowing existence of @xmath5 original mubs with @xmath74 , see e.g. @xcite ) and we take the number of settings growing slower than the dimension , @xmath75 for @xmath76 ( @xmath15 is assumed to be integer ) . this leads to the ratio @xmath77 . exponential unbounded violation of order of @xmath78 is observed if @xmath79 ( similar analysis can be performed for @xmath80 leading however , to more complicated formulas ) . it is remarkable that for any fidelity and efficiency satisfying @xmath81 , there exist @xmath57 and @xmath30 such that violation grows exponentially with the number of the entangled pairs @xmath82 . ) as a function of @xmath82 copies of a singlet state and the parameter @xmath57 setting the relaxed mub condition . the best violation is obtained for @xmath83 , i.e. the original mubs . we took fidelity of a singlet state @xmath84 , detection efficiency @xmath85 and @xmath86 . the dark blue line corresponds to the plot depicted in fig . [ fig : violation_eta ] for @xmath85.,height=226 ] of steering inequality ( @xmath56 ) computed for the relaxed mub condition with @xmath87 , singlet state fidelity @xmath84 and @xmath86 as a function of detection efficiency @xmath71.,height=188 ] violation of the steering inequality ( @xmath56 ) as a function of @xmath82 copies of a singlet state and the parameter @xmath57 setting the relaxed mub condition is shown in fig . [ fig : violation_epsilon ] . we assumed exemplary non - perfect fidelity of the singlet @xmath84 , detection efficiency @xmath85 and @xmath86 . the best violation is obtained for the original mubs , @xmath55 . relaxing this condition results in monotonic decrease of violation for increasing @xmath57 but , its exponential character is preserved . fig . [ fig : violation_eta ] depicts dependence of violation on the efficiency @xmath71 for fidelity @xmath84 , @xmath87 and @xmath86 . the minimal required efficiency to observe the violation equals @xmath88 . both figures reveal an interesting property of quantum steering : exponential decay of the global fidelity @xmath89 and efficiency @xmath90 , observed for multi - pair sources of entangled qubits , is suppressed by exponential number of measurements . the latter is still possible to handle for small number of qubits ( as it is done in tomographic experiments ) . in fact , this proposal has recently been demonstrated for the case of @xmath91 copies , and @xmath92 settings corresponding to two mub elements @xcite . the main challenge in these types of experiments is extending the number of settings and generation of more mubs . this seems to require a significant nonlinearity , a resource which is rare . hence , below we shall consider a system where an unbounded violation of a steering inequality is possible using solely linear optics . we will now employ a quantum - optical scheme based on a parametric - down - conversion source but working in a regime where it generates polarization entangled squeezed vacuum states . their quantum correlations posses the same rotational invariance as a usual two - photon polarization singlet and can be seen as two copies of approximate original epr correlations . due to this property , these states have recently been successfully used to reveal a bell nonlocality which does not vanish in the limit of large population @xcite . using the same key feature and implementing the generalized mubs by merely polarization rotations , we will show that entangled squeezed vacuum states lead to unbounded violation of our steering inequality . entangled squeezed vacuum is a superposition of @xmath93-photon polarization bell - singlet states @xmath94 with a probability amplitude @xmath95 , @xmath96 , where @xmath97 ( @xmath98 ) is creation operator for a spatial mode @xmath12 ( @xmath99 ) and @xmath100 ( @xmath101 ) denotes horizontal ( vertical ) polarization . perfect correlations present in each multi - particle polarization singlet are manifested by equal photon numbers in orthogonal polarizations in the spatial modes : @xmath102 they are preserved with respect to the global rotations of polarization . each spatial mode in @xmath103 contains fixed number of particles equal @xmath1 . projections on fock states constitute a natural framework for bob s photon counting measurements . they reveal not only the correlations but lead to post - selective creation of @xmath103 from the squeezed vacuum state @xmath104 . thus , for simplicity , in our considerations we will now focus on a particular @xmath105 . in order to show unbounded violation ( [ quantum - violation ] ) for ( [ psi - n ] ) we will adopt a strategy similar to the one used in @xcite : we will examine the correlations after applying incremental rotations on the spatial modes . in fact , due to rotational symmetry of @xmath105 , it is enough to consider the relative rotation angle between the modes @xmath12 and @xmath99 , further denoted by @xmath106 . each mode represents a @xmath107-dimensional hilbert space spanned by one of the basis chosen from a collection of @xmath15 bases , enumerated by @xmath108 , generated by an appropriate polarization rotation : @xmath109 they correspond to bob s @xmath110 bases discussed in the theorem and represent generalized mubs ( see fig . [ fig : suppl_overlap ] in the supplementary information ) . we then have the following _ proposition . given a set of @xmath15 bob s measurement bases @xmath111 with @xmath112 and @xmath113 , defined by some set of angles @xmath114 , the maximal overlap @xmath115 equals the maximal overlap between @xmath116 and @xmath117 with @xmath118 : @xmath119 where @xmath120 and @xmath121 denotes the floor function . @xmath122 goes to zero as fast as @xmath123{d}$ ] . _ including experimental imperfections in their simplest form , we assume equal efficiency @xmath71 for each of the two detectors at alice s . for the multi - particle bell - singlet states ( [ psi - n ] ) this modifies the quantum value of the steering functional to @xmath124 and condition ( [ quantum - violation ] ) to @xmath125 . non - unit detection efficiency prevents from observation of the unbounded violation , which is expected . however , if the losses are moderate it will be possible to witness quantum steering for multi - particle quantum states . the price to pay with respect to our abstract result presented in the theorem is that in presence of losses the number of settings grows with the local dimension quicker . large number of measurements requires in turn very high accuracy in the polarization rotations . fig . [ fig : violation ] depicts the violation @xmath126 as a function of the local dimension @xmath1 , for the optimal angle @xmath127 and number of settings @xmath128 . the dependence @xmath129 shown in the upper inset of fig . [ fig : violation ] is rather complex ( see the supplementary information ) . similarly , fig . [ fig : violation_eta2 ] depicts @xmath130 as a function of the dimension @xmath1 and detection efficiency @xmath71 . as expected , the violation gets stronger for increasing population of the system , i.e. larger @xmath1-s , but also for higher detection efficiencies . however , we notice discrepancy between the value of critical efficiency for @xmath131 shown in this figure @xmath88 , and the one reported in literature @xmath132 @xcite . we think that it reflects the fact that our estimation of the classical bound is not tight . of steering inequality ( [ quantum - violation ] ) for multi - particle bell - singlet states @xmath133 ( [ psi - n ] ) and rotation angle @xmath127 . the upper inset shows the optimal number of settings @xmath128 for a given @xmath1 used in this computation . grey area indicates a range of values of @xmath126 for @xmath134 . ] of steering inequality ( [ quantum - violation ] ) computed for multi - particle bell - singlet states @xmath133 ( [ psi - n ] ) , detection efficiency @xmath71 and the best @xmath15 found for given @xmath1 ( see fig . [ fig : violation ] ) . white area indicates no violation . ] _ discussion . _ we would like to emphasize the following two aspects of the result we present . this is the first quantum steering inequality which is formulated in an error - tolerant way . its robustness with respect to non - perfect measurement settings stems from its link to the entropic deutsch maassen uffink uncertainty relation : the same parameter governs strength of both inequalities . violation of this inequality with multi - particle quantum correlations seems feasible . we have suggested two experimental proposals . the first one is a proof of concept , which may serve as an illustrative example of unbounded violation of steering inequality involving relaxed mub measurements . it can be performed with small number of two - photon singlet states and measurements . the second proposal seems particularly remarkable and stunning : this is a linear optics experiment which , also for the first time , can confirm that increasing arbitrarily slowly the local dimension of a maximally entangled state is sufficient to observe an unlimited discrepancy between the classical and quantum description of the experiment shown in fig . [ fig : scheme ] . our result highlights very clear distinction between steering and bell - nonlocality correlations . violation of bell inequalities with small number of settings vanishes for singlet states build from macroscopic qubits @xcite or is constant for multi - particle bell - singlet states @xcite . in the case of quantum steering , unbounded violation is possible for multi - particle generalization of the original epr correlations and almost constant number of settings in the form of mubs , or their feasible counterparts . the latter applied to multiple copies of a singlet state may enable violation of order of @xmath135 observed experimentally . on the other hand , multi - particle bipartite steering based on polarization entangled squeezed vacuum allows violation of order of @xmath136{d})$ ] at its best . its realization necessitates linear optics equipped with photon - number - resolving detection and the generalized mubs are implemented merely by polarization rotations , equally easy for every value of the local dimension . in our considerations we took into account experimental imperfections in the form of losses . further analysis of the possibility of loophole - free violation and the role of fair sampling assumption similar to @xcite remains to be addressed . our approach is very likely to be developed and exploited in several quantum information tasks which are of great technological interest . for example , recently the idea of partial characterization of the devices has been developed and has been expressed in terms of semi - device - independent scenarios @xcite . this approach lies between the two extremes of device - independent and tomographic strategies and quantum steering is its natural prototype . multi - particle bipartite steering can be used for randomness certification and may foster development of optimal strategies in quantum random access codes . the above potential applications result also from the fact that our result is getting feasible although , technological requirements for an experimental setup are high . polarization entangled squeezed vacuum states produced in parametric down conversion process , with mean photon number of the order of ten , are available in laboratories @xcite . the states ( [ psi - n ] ) are created postselectively : each time alice or bob takes a measurement , the number of photons in each mode is known . experiments could employ the techniques of @xcite and integrated optics setups equipped with superconducting transition - edge sensors ( tess ) @xcite , which posses photon - counting efficiencies near @xmath137 and have well - resolved photon - number peaks , up to around ten photons @xcite . for an alternative experimental implementation it is worth considering the angular momentum of light @xcite . our findings indicate new directions in the future research on quantum steering , by posing several open questions . the most intriguing is : how far does the general scaling in the number of settings @xmath15 go for the best steering monogamy relations ? in this context the first results on that type of monogamy @xcite need further development . to address these issues it might be useful to analyze the possible cases of steering inequalities with unequal number of settings at alice s and bob s . existence of the monogamy relations here is conceivable if the number of settings for the state estimation at bob s was of the order of the local dimension @xmath1 . then the asymmetric steering functional might satisfy a no - go limit in the form of an upper bound which value is fixed by diverging number of settings . verifying this question is an important problem since it would determine what is the relation between bell correlations and steering inequalities from the perspective of monogamy relation on the ground of no - signaling theories . another interesting puzzle is : can an asymmetric steering functional reveal quantumness in a more adequate way than the original one ( cf . @xcite ) ? if yes , its potential applications in information processing may be worth examining . our result highlights this point and we are planning to consider it in our future work . our findings open new conceptual and experimental perspective for steering inequalities and therefore , naturally stimulate questions in the context of quantum - information resources theory . for example , if and how the recent application of rigidity of quantum steering to device - independent one - sided computation @xcite can be adopted and made useful in our case ? further , it would be interesting to examine our result on the ground of general probabilistic theories referred to the receiver s ( bob ) system ( cf . @xcite and references therein ) . we hope that the results and discussion presented in this paper will stimulate the future research on quantum correlations in a broader context , including their applications and demonstrations . _ methods . _ the proof of our theorem involves a surprisingly powerful modification of the former analysis presented in @xcite : it builds - in tolerance for possible errors in measurement settings . this was not possible in the framework of @xcite . in @xcite the steering functional @xmath138 is bounded by the operator norm of a gramm matrix @xmath139 with matrix elements equal @xmath140 . its estimation is very troublesome because the matrix involves probabilities of an unknown value , resulting from the local hidden state model . hence , any upper bound on the matrix norm @xmath141 must optimize over all probability values . our technique was to exploit the tensor structure appearing due to double indices in a new , much efficient way . this was done by means of specific decomposition of the @xmath142 part of the structure , which employs mutually orthogonal shift @xmath143 matrices . this key trick allows to go much farther beyond the rough estimates of hidden probabilities parameters used so far . indeed , the latter allows to bound the maximal eigenvalue of the product @xmath144 thus , the singular value of @xmath139 as well , by factorizing out the probabilities due to their normalization rather then simply binding them all by unity . this allows to control the degree of mutual unbiasedness among all the bases individually i.e. from one pair to another in terms of deutsch maassen uffink parameters . derivation of the proposition exploited the properties of rotational symmetry of multi - particle bell - singlet states and the fact that the overlaps between multi - particle polarization bases ( [ polarizationbases ] ) generate a probability distribution @xmath145 which corresponds to a two - dimensional continuous - time quantum walk @xcite ( see the supplementary information ) . thus , @xmath146 does not follow the central limit theorem and its maximal values are taken at the boundaries of its domain . further , to obtain the formula ( [ c ] ) , some specific properties of binomial coefficients were used @xcite . the asymptotic value of the maximal overlap was computed using the stirling s formula . optimal number of settings @xmath128 and set of angles for given @xmath1 and @xmath147 were found by a numerical algorithm . technical details are presented in the supplementary information . finally , we would like to stress that the experimental implementation requires correct space - time arrangement @xcite and any faster - than - light signaling due to apparatus correlations must be excluded ( in order to avoid controversies similar to the ones which arose in some tests of bell inequalities ) . ms and ab were supported by the eu 7fp marie curie career integration grant no . 322150 `` qcat '' , ncn grant no . 2012/04/m / st2/00789 , mnisw co - financed international project no . 2586/7.pr/2012/2 and mnisw iuventus plus project no . ip 2014 044873 . ar thanks for support erc adg qolaps and ncn grant no . 2014/14/m / st2/00818 . ph acknowledges support of erc adg qolaps . the authors thank ryszard horodecki for a discussion .
unusual magnetic behavior of helimagnet mnsi unusual , which triggered the suspicion , that the magnetic states in such helimagnetic materials might be of topological origin opological , top2,bogdannature06,bogdanmnsi , is confirmed as skyrmions in quantum hall ( qh ) qhexperim and small angle neutron scattering ( sans ) experiments @xcite . these experiments , though made pioneering observations , could give only indirect evidence for the existence of skyrmion spin textures , due to their confinement in the momentum space . however , more recently , in a real space experiment with lorentz transmission electron microscopy ( tem ) , direct photographic evidence of skyrmion crystals was obtained in a thin film of helimagnet fe@xmath0co@xmath0si , on a plane perpendicular to the applied magnetic field @xcite . skyrmion crystals were found to be in a beautiful hexagonal form , with the localized individual skyrmions serving as its constituent molecules , extended over lattice spacing of 30 nm range . for theoretical description of these fascinating experimental observations in helimagnets , the accepted basic model is given by a ferromagnetic spin exchange hamiltonian in combination with a dzyaloshinskii - moriya ( dm ) interaction bogdanmnsi , lorentzexp , helimodel , helimodelprl10 . it is argued , that the occurrence of topological solitons in helimagnets is due to the competing forces between the ferromagnetic and the effective dm interactions , where the ferromagnetic spin exchange tends to align the neighboring spins parallel to each other , while the dm spin - orbital interaction with its broken inversion symmetry , orients them to be mutually perpendicular , resulting to a helical order of topological origin . since the experimentally observed skyrmions are found to have extended nature , slowly varying in comparison with the lattice structure of the original magnetic crystals , the long wavelength limit is justified and at low energy and weak dm coupling we can consider the continuum approximation . at this semiclassical limit an intriguing topological property sets in , with a spatially dependent magnetization as a unit vector , wrapping around a 2-sphere , while the 2d coordinate space , due to the fixed orientation of the spins at space - infinities , compactifies to another 2-sphere , inducing a sphere to sphere mapping . the degree of this mapping , counting the number of times the skyrmion magnetic field sweeps the target sphere , when the coordinate space is covered once , defines the integer valued topological charge @xmath1 @xcite . it is a remarkable fact , that at the continuum limit , the ferromagnetic hamiltonian itself , defined on a 2d plane , allows exact skyrmion solutions with arbitrary integer charge @xmath2 , which can be linked to the holomorphic functions @xcite . however , such solutions exhibiting in general a noncircular symmetry , posses a scale invariance property , which does not allow to fix the skyrmion size and makes the skyrmions in a ferromagnetic model unphysical metastable states . however , the addition of a dm type interaction with broken symmetry can change this picture significantly by providing the necessary scale through its coupling parameter and stabilizes the magnetic skyrmions , that have been observed in several recent experiments in different helimagnetic materials . however , at the theoretical level , the addition of the dm hamiltonian can not sustain the analytic solutions obtained in the original ferromagnetic model and at the same time introduces a high level of asymmetry to the resulting equations , due to explicitly broken inversion symmetry in the dm interaction . therefore , the appearance of skyrmion spin patterns with noncircular symmetry seems to be more generic in such systems , in particular for describing the skyrme crystals . however based on the experimental data , available so far , in observing the magnetic skyrmions in helimagnets , any definitive statement about the symmetry and isotropy of the individual skyrmions located in the skyrme lattice seems to be difficult to make . in qh experiments the indirect evidence of skyrmions was obtained in the form of an unusual magnetic field giving no details about the structure of the soliton . in sans experiment , though the situation was improved with the detection of the hexagonal form of skyrme crystals through a 6-fold symmetry , any symmetry statement about the individual skyrme molecules was beyond the scope of this experiment . finally in recent thin film lorentz tem experiment images of extended skyrme molecules inside hexagonal skyrme crystals were obtained in real space , revealing much detailed structures of these objects appearing in a helimagnetic material . however , even from this best available data , one can not confirm with certainty about the nature of symmetry and isotropy of the individual skyrmions , since the color - code of the images shows directional asymmetry and color inhomogeneity ( see fig . 1e - f of @xcite ) . therefore , for a clearer picture about the precise symmetry of the skyrmions and the exact directional orientations of the magnetic field lines one has to look for further finer and precision experiments . nevertheless , at the theoretical level , most of the models tend to support the symmetric form of the individual skyrmions and their isotropy inside a skyrme crystals , neglecting the inter - skyrmion interactions . for example , in a numerical simulation by minimizing the discretized model hamiltonian using the monte carlo method , circular symmetric skyrme molecules forming a skyrme crystal in symmetric form have been obtained @xcite . similarly , in the prevalent theoretical models bogdanmnsi , helimodelprl10 a circular symmetric ansatz is assumed for the solution of individual skyrmions . however , since the general nature of the governing euler equations are highly asymmetric , such a symmetric assumption put strong restrictions on the corresponding solutions , confining it to the unit charge sector and imposing a fixed initial phase shift , which limits the spin pattern only to a particular orientation ( @xmath3 type ) bogdanmnsi , helimodelprl10 . it is important to stress , that since the individual skyrmions as well as the skyrme crystals in general should be solutions of the energy minimizing equations , they should ideally be derived from the euler equations dictated by the governing helimagnetic model without restriction of symmetry , orientation and charge . however , no satisfactory theoretical proposal seems to have been offered to this challenging problem of describing the hexagonal skyrme crystal structure with noncircular symmetry , as a solution of the underlying equations with higher topological charges , which would induce nonlinear interactions between the constituent skyrme molecules . it is indeed a difficult problem to solve the governing equations , representing highly nonlinear coupled partial differential equations allowing no separation of variables . current studies on the helimagnetic model avoid this difficult problem by completely neglecting the inter - molecular interactions between the skyrmions , by assuming a circular cell approximation bogdanmnsi , helimodelprl10 , though such a linear superposition of individual solutions does not represent a solution of the governing nonlinear equations . moreover , such a configuration as a equidistant collection of individual skyrmions can not be reproduced as a limiting case of a general solution with a higher charge . other investigations on the skyrme lattice in the helimagnetic model focus usually on different other aspects like the effect of current and temperature gradients on the rotating of skyrme lattice @xcite , pinning of skyrmion through inhomogeneity of magnetic exchange @xcite , skyrmion spin lattice coupling with electric dipole @xcite etc . , leaving aside the basic problem of the formation of skyrme crystals as a topological solution . our aim here is to focus on the above posed problem of constructing a skyrme crystal as a solution of the helimagnetic equations . however , we bypass the direct problem , which is hard to solve and obtain the solutions with higher topological charges linked to the skyrme crystal , by looking first for the skyrmions appearing in the pure ferromagnetic model ( fm ) and subject them subsequently , to the dm interaction perturbatively , by assuming the coupling to be weak . this stabilizes the solitons through dm coupling and goes beyond the limitations of the earlier studies by allowing skyrmion solutions with higher charges and more general initial phase admitting different types of orientations for the spin texture . this would mean , however that one has to deal with noncircular symmetric fields without separation of variables , which unlike earlier proposals , is a more difficult task to handle . nevertheless , this approach solves to some extend the problem of formation of hexagonal skyrme crystals within the helimagnetic model with weak dm coupling , by finding a perturbed skyrmion solution with charge @xmath4 , and without the usual circular cell approximation . . the basic helimagnetic model , as motivated above , is given by the hamiltonian & & h= j h_fm+d h_dm , + & & h_fm= 1 2 d^2x ( ) ^2 , + & & h_dm = d^2x(*m * ) , where @xmath5 with @xmath6 is the magnetization vector and @xmath7 , @xmath8 are the competing ferromagnetic and dm interactions , respectively bogdanmnsi , lorentzexp , bakjensen . since the magnetization , taking values on a 2-sphere is restricted to a 2d plane , the system may be described by two spherical angles @xmath9 and @xmath10 defined through polar coordinates @xmath11 with the magnetization components given as @xmath12 the helimagnetic hamiltonian ( [ hheli ] ) may be expressed through these angle variables for convenience as & & h_fm=1 2 d^2x(()^2+^2 ( ) ^2 ) , + & & h_dm= d^2x ( ( -)(_-1 2 2 _ ) - ( - ) ( + 1 2 2 _ ) ) . the associated topological charge takes the explicit form q= d^2 x | [ ] | , note that the energy minimizing euler equations for the fields @xmath13 and @xmath14 , may be derived in the static case from the model hamiltonian ( [ cfm ] , [ cdm ] ) as & & ^2- 1 2 2 ( ) ^2 + 2 m _ ( , ) = 0 , = d j for angle @xmath15 and & & 2 ( ) + ^ 2 ( ^2 ) + 2 m_(,)=0 , and with respect to angle @xmath16 , where the additional dm terms are & & m _ ( , ) ^ 2 ( _ ( - ) + _ ( - ) ) and & & m _ ( , ) - ^ 2 ( _ ( -)+ _ ( - ) = 0 . note that these equations for finding the field configurations minimizing the hamiltonian , represent highly nonlinear inhomogeneous coupled partial differential equations in two variables and are difficult to solve in general . a simple possibility to overcome this difficulty , as adopted in most of the earlier studies @xcite , is to focus on an individual skyrmion with a circular symmetric ansatz = ( ) , = n + _ 0 , , with initial phase @xmath17 , as an arbitrary constant and @xmath18 as an arbitrary integer . specific values of @xmath19 represent precise crystallographic forms , describing the spin orientation in the related magnetic pattern bogdanmnsi . a closer look however reveals that for the circular symmetric ansatz ( ) to work , the equation ( ) for the angle @xmath20 must be zero , which due to @xmath21 reduces to the vanishing of @xmath22 ( [ eqdmgamma ] ) or in turn to the vanishing condition for the term @xmath23 clearly it becomes zero ( for @xmath24 ) , when the cosine function vanishes under the combined condition : @xmath25 and @xmath26 . therefore , for the validity of the symmetric ansatz ( ) for the skyrmions in helimagnets , one gets a solution restricted only to unit topological charge @xmath27 and with fixed value for the initial phase linked to @xmath3 type of magnetic pattern @xcite . therefore , though this is a convenient way to achieve separation of variables in a coupled nonlinear equation , it can not get generic skyrmion solutions with higher topological charges @xmath28 and misses the possibility of finding other crystallographic forms for other values of @xmath17 , involving noncircular symmetric solutions . more importantly , such a symmetric ansatz can not describe the skyrmion lattice structure having multiple centers and higher topological charge , including the skyrme crystals in hexagonal form , as a solution of the full set of governing equations ( [ betaeq],[gammaeq ] ) . nevertheless , a formal insertion of such circular symmetric skyrmions at each lattice site , sometimes with anisotropic terms like @xmath29 etc . added to the helimagnetic model ( [ hheli ] ) , was proposed theoretically for describing a skyrme crystal @xcite , which however neglects completely the interaction between the skyrme molecules , under circular cell approximation . going beyond the conventional circular cell approximation , we intend to look for more general noncircular symmetric solutions by considering interactions between the individual skyrme molecules , consistent with the coupled euler equations ( - ) , minimizing the helimagnetic model . however , since such asymmetric solutions are difficult to obtain in general , we adopt a bypass route by considering the deformation of skyrmion solutions of pure ferromagnetic model ( [ fm ] ) , by switching on the dm interaction ( [ dm ] ) perturbatively , by taking the dm coupling @xmath30 weaker in comparison with the ferromagnetic exchange coupling @xmath31 with @xmath32 as a small parameter . the assumption of a weak relative coupling @xmath33 however has gone already in the justification of the continuum approximation for the present model ( see e.g. lorentzexp ) . note , that this approach also brings in the necessary scaling parameter for stabilizing the skyrmions . in the first step , we focus on the nonlinear euler equations associated with the ferromagnetic model alone , without the dm interaction and look for the exact skyrmions with higher topological charge , which in general would have noncircular symmetry . in the next step , this exact solution is subjected to the dm interaction perturbatively , where though the perturbing fields do not allow separation of variables , the governing equations become linear , allowing numerical solutions . therefore , we start with suitable solutions @xmath34 for the ferromagnetic model ( [ fm ] ) ( or ( [ cfm ] ) ) , satisfying the related euler equations ( - ) , at @xmath35 . it is remarkable , that the topological skyrmions for this model can be given by the celebrated exact solutions found way back by belavin and polyakov ( bp ) bp . a deep theoretical concept and beautiful geometrical picture go into this construction , where an important lower bound for the energy ( [ cfm ] ) is revealed through the topological charge @xmath36 ( [ qba ] ) : @xmath37 . note that due to the conservation of charge the lower bound guarantees , that the finite energy topological solitons can not decay into the vacuum or any other topological states . interestingly , the lower bound is saturated under the so called _ self duality _ condition _ _ 0=2 _ 0 _ _ 0 , _ _ 0=2 _ 0 _ _ 0 and since this is the minimizing condition for the energy functional , the corresponding field configuration also becomes a solution of the associated euler equations ( ( - ) at @xmath35 ) , which are our current focus . note , interestingly that these euler equations are actually solved without solving them directly , but by solving the self - duality condition and therefore , for finding the skyrmion solutions in pure ferromagnetic model , one has to consider only the solution of ( ) , which remarkably can be solved exactly . deep reason behind this intriguing fact is that the relations ( ) for attaining the energy minimum represent the well known cauchy - riemann ( cr ) condition ( expressed in the polar coordinates ) , required for the analyticity of the complex function @xmath38 linked through the stereographic projection f(z)= e^i_0 . consequently , as shown in @xcite , any analytic function of the form f(z)= , n=_i n_i-_jn_j , would be an exact skyrmion solution of the ferromagnetic model with topological charge @xmath39 , allowing a scale invariance : @xmath40 . our strategy in constructing the skyrmions for the helimagnetic model , is to take the ferromagnetic soliton solutions ( ) as the unperturbed skyrmions @xmath41 through mapping ( [ stereo ] ) , giving the angle variables as inverse functions = 2 ^-1 ( |f | ) , = arg f. these are guaranteed to satisfy the self - duality ( cr condition ) ( ) and consequently , the euler - equations ( -[gammaeq ] ) for @xmath35 , as an exact solution . therefore , for describing an individual skyrmion we may consider the simplest case with @xmath42 , obtained from ( ) through a reduction @xmath43 giving an exact solution @xmath44 , where scale transformations @xmath45 are allowed . this isolated skyrmion of the ferromagnetic model , linked to the magnetization component m^3== = , having a perfect circular symmetry is shown in fig . 1a in the graphical form . for constructing skyrmion lattice in a ferromagnetic model , a naive way would be to insert the same circular symmetric bp skyrmion with @xmath42 , constructed above , at each skyrme lattice site on the 2d plane , separated uniformly from its neighbors at a distance of @xmath46 . this would give the magnetization in the form @xmath47 neglecting the nonlinear interactions between the skyrme molecules , induced by the governing euler equations . note , that a similar construction was adopted for the skyrme crystals , but for the helimagnetic models , in most of the current studies assuming a circular cell approximation . however , unfortunately such a formal construction of skyrme lattice using superposition of individual skyrme solitons is _ not _ a solution of the energy minimizing equations dictated by the model hamiltonian , both for the pure ferromagnetic as well as for the helimagnetic models , since due to nonlinearity of the governing equations the superposition principle of individual solutions does not hold . therefore any solution for the skyrme crystal with higher charges must satisfy the energy minimizing euler equations , where the nonlinear interactions between the individual skyrmions with unit charge , sitting at each lattice site , can not be neglected . for constructing such a skyrme crystal solution , we have to look therefore for skyrmions with a higher topological charge , which for the pure ferromagnetic model can be reduced again from the exact bp solution , which naturally can no longer exhibit circular symmetry . for the construction of a possible solution for the skyrmion crystal in hexagonal form , within the ferromagnetic model , we may consider an exact soliton with topological charge @xmath4 , obtained from the bp solution ( [ bpsol ] ) under certain reduction of parameters @xmath48 $ ] etc . , chosen suitably at seven centers of the equidistant lattice sites , to form a hexagonal lattice with equal sides @xmath49 in explicit form this solution may be given simply by f_7 ^ -1 ( z)= _ j=1 ^ 7 ( z - z_j)= - a_s^6 e^i + ^7 e^i 7 . this exact solution may be linked to the hexagonal skyrme crystal pattern for the angle variables @xmath50 through the magnetization component m^3=_0= , _ 0= -arg f_7(z ) , which is obviously noncircular symmetric and expressed through bp solution ( [ f7 ] ) , as graphically represented in fig . this topological soliton solution with higher charge describing the hexagonal skyrme crystal in a ferromagnetic model is guaranteed to satisfy the required energy minimizing equations , where parameter @xmath46 may serve as the skyrme lattice constant fixing the size of the crystal . however this parameter remains arbitrary due to the scale invariance @xmath51 of the solution , which also makes the skyrmion to vanish at @xmath52 inducing the well known unphysical metastable skyrmionic states in a pure ferromagnetic model . however , the additional dm interaction ( [ dm ] ) , which is included in the helimagnetic model could stabilize the skyrmion solutions by introducing the required scaling in the system through coupling parameter @xmath30 , as we will see below . note , that though in fig 1a an isolated skyrmion with unit charge exhibits perfect isotropy and circular symmetry , such individual symmetries no longer remain intact in the skyrme solution of crystalline form with higher charge @xmath4 , as shown in fig 1b , due to interactions between the skyrme molecules induced unavoidably by the underlying nonlinear equations . this characteristic behavior of the skyrme crystals with higher charges exhibiting anisotropy and noncircular symmetry , as obtained here in the ferromagnetic model , is expected to remain prevalent also in the helimagnetic model with the inclusion of an additional symmetry breaking dm term ( [ dm ] ) , at least in the excited states . this is the emphasis of our proposal in what follows . in pure ferromagnetic model ( [ fm ] ) for a ) individual skyrmion with @xmath53 and b ) hexagonal skyrmion crystal with @xmath4 . , title="fig:",width=226,height=4 ] in pure ferromagnetic model ( [ fm ] ) for a ) individual skyrmion with @xmath53 and b ) hexagonal skyrmion crystal with @xmath4 . , title="fig : " ] our next step is to consider the helimagnetic model ( [ hheli ] ) by switching on the dm interaction ( ) to the ferromagnetic model ( ) investigated above . for finding the spin texture in this helimagnetic model described by the skyrmion solution , one has to find the energy minimizing configuration given by the the solutions of the associated euler equations for the fields @xmath13 and @xmath14 derived from ( - ) . however , as discussed above , since these coupled highly nonlinear partial differential equations ( , ) for the two fields involving both the polar coordinates without separation of variables , are difficult to solve in general for @xmath54 , our strategy would be to treat the system perturbatively by considering @xmath55 the relative coupling between the dm and the ferromagnetic exchange interaction , as a small parameter . the perturbative solution therefore may be given by ( , ) = _ 0 ( , ) + _ 1 ( , ) , ( , ) = _ 0 ( , ) + _ 1 ( , ) , in the first order of approximation , where @xmath56 are the unperturbed solutions induced by the ferromagnetic hamiltonian ( [ cfm ] ) , while @xmath57 are the deformations suffered , when the dm interaction ( ) is switched on , perturbatively . the parameter @xmath58 also serves as the scaling parameter , breaking the scale invariance of the unperturbed skyrmions and thus providing the required stability to the soliton solutions . recall , that we have derived already the set of solutions @xmath59 in the analytic form for both the individual skyrmion and the skyrme crystal , as shown in fig 1 . therefore , it remains now to extract the deforming solutions @xmath60 . for this we insert the perturbative ansatz ( [ pbetgam ] ) in the full set of helimagnetic equations ( , ) and derive the corresponding linear set of equations in the first order of approximation @xmath61 : & & ^2_1 - 2 _ 0 ( _ 0 _ 1)-_1 2 _ 0 ( _ 0)^2 + 1 2 m_(_0,_0)=0 , + & & 2 _ 0 ( _ 0 _ 1+_1 _ 0 + _ 1 ^2 _ 0 ) + _ 1 2 _ 0 ( _ 0 _ 0)+ ^ 2 _ 0 ( ^2 _ 1)+ 1 2 m_(_0,_0)=0 , where @xmath62 are as in ( , ) . however , due to the anisotropic and symmetry breaking contribution of the dm interaction the overall symmetry of the unperturbed skyrmions gets broken generically , even for the isolated skyrmions and these coupled set of equations , unlike in the pure ferromagnetic model , can not be solved analytically . nevertheless , the linearity of these equations allows numerical solutions for the fields @xmath60 ( see supplementary material for mathematica 10 code ) . adding now these perturbing solutions for the dm interaction to the corresponding analytic solutions @xmath59 found above for the ferromagnetic model , which also stabilizes the solutions , we finally obtain the skyrmions for the helimagnetic model in the form ( ) as shown in fig . 2 , for both isolated and crystal solutions . in helimagnetic model ( [ fm]- ) for a ) individual skyrmion with @xmath53 and b ) hexagonal skyrmion crystal with @xmath4 . these skyrmion solutions are perturbations of ferromagnetic skyrmions obtained in fig . 1 for weak dm coupling , with small parameter @xmath63 . , title="fig:",width=226,height=4 ] in helimagnetic model ( [ fm]- ) for a ) individual skyrmion with @xmath53 and b ) hexagonal skyrmion crystal with @xmath4 . these skyrmion solutions are perturbations of ferromagnetic skyrmions obtained in fig . 1 for weak dm coupling , with small parameter @xmath63 . , title="fig : " ] it is evident by comparing the skyrmion solutions for the helimagnetic model presented in fig . 2 with those for the pure ferromagnetic model shown in fig . 1 , that the inclusion of the symmetry breaking dm interaction ( [ dm ] ) deforms the solutions and introduces further anisotropy and noncircular symmetry to the individual skyrme molecules as well as to the skyrme crystals appearing in the helimagnet model . this result generalizes therefore the known circular symmetric skyrmions and skyrmion crystals obtained in the helimagnetic model to the asymmetric case . appearance of skyrme crystals in the hexagonal form has been observed experimentally in the magnetization pattern in helimagnetic materials . however a full theoretical description for the formation of crystalline structures as a general topological skyrmion solution of the governing equations minimizing the model hamiltonian , which is difficult to achieve due to high asymmetry and nonlinearity inherent to the system is not yet available in the literature . assuming weaker dm interaction relative to the ferromagnetic exchange coupling , which is also a requirement for the conventional continuum approximation , we present here such solutions for skyrme crystals with higher topological charge and for individual skyrmions with unit charge in the helimagnetic model . our approach of perturbing and stabilizing the skyrmions of the ferromagnetic model with the symmetry breaking dm interaction , bypasses the major difficulty of nonlinearity and solves the energy minimizing equations without assuming any separation of variables or circular cell approximation , for the first time ( see the supplementary material ) . note that , though in recent experiments the evidence of skyrmion and skyrme crystals was found , the details about the circular symmetry or isotropy of skyrmions in their individual state or when they appear as interacting molecules in the skyrme crystal are not much clear ( see color - code of experimental images in @xcite ) therefore , the theoretical possibility of asymmetric skyrme crystals found here , can not be ruled out without verifying in precision experiments , where deformation of symmetries and isotropy in the magnetic pattern could be detected in finer details . to tune with the validity region of the present solutions obtained for low values of @xmath64 , care should be taken in choosing the helimagnetic materials , so that their spin - orbital dm interaction is weaker in comparison with the ferromagnetic spin exchange . though the helimagnetic crystals mnsi and fe@xmath65co@xmath66si share similar magnetic features , they differ significantly in their electronic structures , which is likely to make the later system more suitable for investigating the nature of skyrmion lattices over a large temperature range , with the possibility of revealing more refined structures . the recent lorentz tem experiment lorentzexp might well be revisited and rescrutinized for detecting the possible symmetry breaking in the observed magnetic pattern in skyrme crystals , in the light of the present result , at least in the excited states and in the higher temperature range . 99 c. pfleiderer , s. r. julian and g. g. lonzarich , nature * 414 * , 427 ( 2001 ) . ; n. dolron - leyraud _ et . nature * 425 * , 595 ( 2003 ) ; c. pfleiderer _ et . nature * 427 * , 227 ( 2004 ) b. binz and a. vishwanath , physica b * 403 * , 1336 ( 2008 ) mhlbauer , s. _ et . science * 323 * , 915 ( 2009 ) u. k. r@xmath13ler , a. n. bogdanov and c. pfleiderer , nature * 442 * , 797 ( 2006 ) . u. k. r@xmath67ler , a. a. leonov , a. n. bogdanov , _ chiral skyrmionic matter in noncentrosymmetric magnets _ , arxiv 1009.4849 ( 2010 ) , u. k. r@xmath67ler , a. a. leonov , a. n. bogdanov , _ skyrmion textures in chiral magnets _ , arxiv 0907.3651 ( 2009 ) karin everschor et al , phys . b * 86 * , 054432 ( 2012 ) ye - hua liu and you - quan li , j. phys . : cond . mat.*25 * 076005 ( 2013 ) s. seki , x. z. yu , s. ishiwata , and y. tokura , science * 336 * , 198 ( 2012 ) . p. bak and h. jensen , j. phys . * c 13 * , l881 ( 1980 )
intricate spin textures in helimagnets , identified as stable topological skyrmions , were observed experimentally , where skyrme lattice was supposed to exhibit symmetric structures in the ground state . we show the possibility of asymmetric skyrmions in a helimagnetic model , for individual skyrmion as well as for the hexagonal skyrme crystal with higher charge , as perturbative deformation and stabilization of exact ferromagnetic solitons . such nonsymmetric configurations for the skyrme lattice , predicted here theoretically , need to be verified in precision experiments . _ key words _ : solitons with topological charge , exact skyrmions in ferromagnet , asymmetric skyrmions and skyrme crystals in helimagnets , perturbative solutions +
photonic crystals ( pcs ) were first introduced theoretically by yablonovitch @xcite . , and experimentally by john @xcite . pcs , constructed with periodic structure of artificial dielectrics or metallic materials , have attracted many researchers in the past two decades for their unique electromagnetic properties and scientific and engineering applications @xcite . these crystals indicate a range of forbidden frequencies , called photonic band gap , as a result of bragg scattering of the electromagnetic waves passing through such a periodical structure @xcite . as the periodicity of the structure is broken by introducing a layer with different optical properties , a localized defect mode will appear inside the band gap . enormous potential applications of pcs with defect layers in different areas , such as light emitting diodes , filters and fabrication of lasers have made such structures are interesting research topic in this field . negative index materials ( nims ) with simultaneously negative permittivity and permeability , known as left - handed materials or metamaterials , were introduced by veselago @xcite nearly forty years ago . after the experimental realization of metamaterials by smith et al . @xcite , such materials have received attention for their very unusual electromagnetic properties @xcite . by possibility of manufacturing pcs with metamaterials , called metamaterial photonic crystals ( mpcs ) in recent years , a new research area was opened to researchers where numerous interesting results have been reported so far . among the papers published in mpcs properties , there are a number of reports in such structures with defect layers . the properties of the defect modes in different one - dimensional ( 1d ) conventional pc and 1d mpc have been reported by several authors @xcite . in this paper the characteristic matrix method is used to investigate the effects of the lossless negative and positive defect layers on the defect modes in 1d lossy mpc transmission spectra . the paper is organized as follows : the mpc structure and theoretical formulation ( characteristic matrix method ) are described in section 2 , the numerical results and discussions are presented in section 3 and the paper is concluded in section 4 . a schematic diagram of a one - dimensional mpc structure with defect layer is shown in figure 1 . the mpc structure is composed of alternative nim layers ( layers _ a _ ) and pim layers ( layers _ b _ ) , where nim is dispersive and dissipative . the thickness and the refractive index of the layers are @xmath0 and @xmath1 , @xmath2 and @xmath3 , respectively . layer _ c _ is a defect layer and is assumed to be a lossless and non dispersive material with thickness and refractive index of @xmath4 and @xmath5 , respectively . the calculations are performed using the characteristic matrix method @xcite , which is the most effective technique to analyze the transmission properties of pcs . the characteristic matrix of the structure is given by : @xmath6=(m_{a } m_{b})^{n/2 } m_{c } ( m_{b } m_{a})^{n/2}$ ] where @xmath7 , @xmath8 and @xmath9 are the characteristic matrices of layers _ a _ , _ b _ , and _ c_. the characteristic matrix @xmath10 for te waves at incidence angle @xmath11 in vacuum is given by @xmath12\ ] ] where @xmath13 , _ c _ is speed of light in vacuum , @xmath14 is the angle of refraction inside the layer _ i _ with refractive index @xmath15 and @xmath16 , where @xmath17 . the refractive index is given as @xmath18 @xcite , where the positive and negative signs are assigned for the pim and nim layers , respectively . the final characteristic matrix for an _ n _ period structure is given by : @xmath19^n= \prod^n_{i=1 } m_i \equiv \left ( \begin{array}{cc } m_{11 } & m_{12}\\ m_{21 } & m_{22 } \end{array}\right)\ ] ] where @xmath20 are the matrix elements of @xmath21^n$ ] . the transmission coefficient of the multilayer is calculated by : @xmath22 where @xmath23 and @xmath24 . the transmissivity of the multilayer is given by @xmath25 the transmissivity of the multilayer for tm waves can be obtained by changing @xmath26 , @xmath27 , and @xmath28 . the permittivity and permeability of layer _ a _ with negative refracting index in the microwave region are complex and are defined as @xcite , @xmath29 @xmath30 where _ f _ and @xmath31 are the frequency and damping frequency in ghz , respectively . the behaviours of the real parts of the permittivity and permeability of layer _ a _ , @xmath32 and @xmath33 , versus frequency have been discussed in our previous report @xcite . the transmission spectrum of the mpc structure was calculated based on the theoretical model described on the previous section . the calculations are carried out in the region where @xmath32 and @xmath33 are simultaneously negative , where a new gap , called the zero-@xmath34 gap @xcite , will appear @xcite . layer _ b _ is assumed to be the vacuum layer with @xmath35 . the thickness of layers _ a _ and _ b _ are chosen as @xmath36 mm and @xmath37 mm , respectively . the defect layer _ c _ is either nim or pim non dispersive and non dissipative . the total number of the lattice period is selected to be @xmath38 @xcite . the transmission spectrum for @xmath39 mm with positive refractive index of @xmath40 and @xmath41 , for two different loss factors ( @xmath42 ghz and @xmath43 ghz ) and for normal incidence are shown in figure 2 . as it is clearly seen , the rate of the transmittance and the height of the defect modes are affected by the loss factor . the frequency of the defect modes as a function of the refractive index of the defect pim and nim layers for normal incidence and for two different small lattice s loss factors are show in figure 3 . as it is seen , the frequency of the defect modes decreases as @xmath5 increases and the modes are nearly insensitive to the small loss factors . in addition , for the pim defect layer ( figure 3(a ) ) two defect modes appear for some specified refractive indices . the behaviour is quite different in the nim defect layer ( figure 3(b ) ) where no defect modes are observed in some specific regions . the effects of @xmath5 on the height of the defect modes for regions i , ii and iii are shown in figure 3(a ) for pim defect layer and for two different loss factors in figure 4 . it is seen that the height of the defect modes almost quadratically depends on the refractive index . the minimum height decreases for the regions with large @xmath5 . moreover , the height of the modes is very sensitive to the loss factor and decreases as @xmath31 increases . the transmission spectrum for @xmath44 and @xmath45 and @xmath46 mm and for two different loss factors ( @xmath42 ghz and @xmath47 ghz ) are presented in figure 5 . as it is seen , the frequency of the defect modes move toward the lower frequencies ( figure 5(a ) ) and the height of modes decrease as the thickness of the defect layer increases ( figure 5(b ) ) . the behaviors of the frequency and height of the defect modes versus @xmath4 for two different @xmath31 are shown in figures 6 and 7 , respectively . as it is seen , the frequency of the defect modes is nearly insensitive to the loss factors but the height is very sensitive to @xmath31 and decreases as the loss factors increases . in this section , the effects of the incidence angle on the defect transmission spectra are investigated . the transmission spectra of te and tm polarized waves for @xmath48 , @xmath49 , @xmath50 and @xmath51 incidence angles for @xmath52 mm , @xmath53 , and for two different loss factors ( @xmath42 ghz and @xmath47 ghz ) are shown in figures 8 and 9 , respectively . as it is observed from figure 8 , and reported in @xcite , the width , the depth and the central frequency of the band gap for te waves increase as the incidence angle increases . but , for tm waves , the gap disappears for the incidence angles greater than @xmath54 . as @xmath11 increases the width and the depth of gap decrease and the central frequency shifts toward the higher frequencies ( figure 9 ) . the behaviour of the defect mode s frequency for te and tm waves as a function of the incidence angle for two different loss factors are shown in figure 10 . here it is clearly seen that the frequency of the defect mode is nearly insensitive to the small loss factor and it disappears for tm waves for the incidence angle greater than @xmath54 . the change of the height of the defect mode for both polarizations as a function of the incidence angle for two different loss factors is shown in figure 11 . it is seen that the height of the defect modes decreases for te polarized waves while it increases for tm polarized waves as the incidence angle increases . it is also seen that the height of the modes are very sensitive to the loss factor for both polarizations . in the last part the effects of the number of the lattice period , _ n _ , and the loss factor on the transmission properties of the defect mode are studied . so far @xmath38 was assumed . the results are presented in figures 12 and 13 for normal incidence angle . as it is seen from figure 12(a ) and as reported in @xcite , the width of the band gap for @xmath55 is almost independent of _ n_. however the depth and the height of the defect mode reduce as _ n _ increases for @xmath42 ghz . it is interesting to note that the frequency of the defect mode remains unchanged but the width of the mode decreases as the number of the lattice period increases ( figure 12(b ) ) . the transmission spectra of the structure for four different loss factors and for @xmath38 are shown in figure 13 . as it is seen from figure 13(a ) the width and the depth of the band gap and the frequency of the defect mode are nearly insensitive to small loss factor but the transmission is affected by @xmath31 @xcite . moreover , the height of the defect mode decreases as the loss factor increases ( figure 13(b ) ) . in this paper the transmission properties of one - dimensional lossy photonic crystals with negative and positive refractive indices layers and a lossless defect layer at the center of the crystal were investigated by characteristic matrix method . it was shown that by adding a lossless defect layer to a 1d lossy mpc , a localized defect mode appears inside the band gap . the position of the defect mode depends on the physical properties of the defect layer such as the refractive index and thickness but it is independent of the loss factor . the results show that as the refractive index and thickness of the defect layer increases the frequency of the defect mode decreases . in addition , it was shown that the frequency and the height of the defect mode are sensitive to the incidence angle and polarization . for a pim defect layer with the refractive index between @xmath56 to @xmath57 , two defect modes appear in a specific range of the transmittance spectrum , while for an nim defect layer the spectrum is quite different where no defect modes is observed in a certain range of the refractive index ( @xmath58 to @xmath59 ) . the absence of the defect mode in some specified region is not clear and is the subject of our future studies . it was also seen that : _ i _ ) by increasing the incidence angle the frequency of the defect mode increases . _ ii _ ) by increasing the thickness of the defect layer the frequency of the defect mode moves toward lower frequencies . _ iii _ ) the height of the defect modes is very sensitive to the loss factors . _ iv _ ) the depth of the band gap increases and consequently the height and the width of the defect mode decrease , so the mode becomes narrower for @xmath55 . _ v _ ) the width and the depth of the band gap and the frequency of the defect modes are nearly insensitive to small loss factors but the height of the defect mode is very sensitive to small @xmath31 and decreases as the loss factor increases . such properties are quite useful in designing new types of optical devices in microwave engineering . 99 yablonovitch , e. , `` inhibited spontaneous emission in solid - state physics and electronics , '' _ phys . lett . _ , vol . 58 , 2059 - 2062 , 1987 . john , s. , `` strong localization of photons in certain disordered dielectric superlattices , '' _ phys . 58 , 2486 - 2489 , 1987 . noda , s. , a. chutinan , and m. imada , `` trapping and emission of photons by a single defect in a photonic bandgap structure , '' _ nature _ , vol . 407 , 608 - 610 , 2000 . massaoudi , s. , a. de lustrac , and i. huynen , `` properties of metallic photonic band gap material with defect at microwave frequencies : calculation and experimental verification , '' _ j. electromagn . waves appl . _ , vol . 14 , 1967 - 1980 , 2006 mandal , b. and a. r. chowdhury , `` spatial soliton scattering in a quasi phase matched quadratic media in presence of cubic nonlinearity , '' _ j. electromagn . waves appl . _ , vol . 1 , 123 - 135 , 2007 . zheng , q. r. , b. q. lin , and n. c. yuan , `` characteristics and applications of a novel compact spiral electromagnetic band - gap ( ebg ) structure , '' _ j. electromagn . waves appl . _ , vol . 2 , 199 - 213 , 2007 yablonovitch , e. , `` photonic band structure : the face - centered - cubic case employing nonspherical atoms , '' _ phys . _ , vol . 67 , 2295 - 2298 , 1991 . joannopoulos , j. d. , s. g. johnson , j. n. winn , and r. d. meada , _ photonic crystals : modeling the flow of light _ , princeton university press , princeton , 2008 . veselago , v. g. , `` the electrodynamics of substances with simultaneously negative values of @xmath60 and @xmath61 , '' _ sov . phys . 10 , 509 - 514 , 1968 . smith , d. r. , w. j. padilla , d. c. vier , s. c. nemat - naser , and s. schuitz , `` composite medium with simultaneously negative permeability and permittivity , '' _ phys . 84 , 4184 - 4187 , 2000 . engheta , n. and r. w. ziolkowski , _ metamaterials : physics and engineering explorations _ , john wiley @xmath62 sons , new jersey , 2006 . caloz , c. and t. itoh , _ electromagnetic metamaterials : transmission line theory and microwave applications _ , john wiley @xmath62 sons , new jersey , 2006 . cai , w. and v. shalaev , _ optical metamaterials : fundamentals and applications _ , springer , new york , 2010 . eleftheriades , g. v. and k. g. balmain , _ negative refraction metamaterials : fundamentals properties and applications _ , john wiley @xmath62 sons , new jersey , 2005 . novitsky , a. v. and l. m. barkovsky , `` guided modes in negative - refractive - index fibres , '' _ j. opt . a pure appl _ , vol . 7 , s51-s56 , 2005 cory , h. and t. blum , `` surface - wave propagation along a metamaterial cylindrical guide , '' _ microw . 44 , 31 - 35 , 2005 . kim , k. y. , j. h. lee , y. k. cho , and h. s. tae , `` electromagnetic wave propagation through doubly dispersive subwavelength metamaterial hole , '' _ opt . , vol . 13 , 3653 - 3665 , 2005 . lotfi , e. , k. jamshidi - ghaleh , f. moslem , and h. masalehdan , `` comparison of photonic crystal narrow filters with metamaterials and dielectric defects , '' _ eur . 60 , 369 - 372 , 2010 . zhu , q. and y. zhang , `` defect modes and wavelength tuning of one - dimensional photonic crystal with lithium niobate , '' _ optik _ , vol . 120 , 195 - 198 , 2009 . abdel - rahman , e. and a. shaarawi , `` defect mode in periodic and quasiperiodic one - dimensional photonic structures , '' _ j mater sci : mater electron _ 20 , s153-s158 , 2009 . ansari , n. , m. m. tehranchi , and m. ghanaatshoar , `` characterization of defect modes in one - dimensional photonic crystals : an analytic approach , '' _ physica b _ , vol . 404 , 1181 - 1186 , 2009 . li , x. , k. xie , and h. m. jiang , `` properties of defect modes in one - dimensional photonic crystals containing two nonlinear defects , '' _ opt . 282 , 4292 - 4295 , 2009 . wu , c. -j . and z. h. wang , `` properties of defect modes in one - dimensional photonic crystals , '' _ progress in electromagnetics research _ , vol . 103 , 169 - 184 , 2010 . ansari , n. and m. m. tehranchi , `` influence of filling fraction on the defect mode and gap closing of a one - dimensional photonic crystal : an analytical approach , '' _ physica b _ , vol . 405 , 2900 - 2906 , 2010 . gharaati , a. and h. azarshab , `` characterization of defect modes in one - dimensional ternary metallo - dielectric nanolayered photonic crystal , '' _ progress in electromagnetics research b _ , 37 , 125 - 141 , 2012 . zhang , w. , p. han , a. lan , y. li , and x. zhang , `` defect modes tuning of one - dimensional photonic crystals with lithium niobate and silver material defect , '' _ physica e _ , 44 , 813 - 815 , 2012 . aly , a. h. and h. a. elsayed , `` defect mode properties in a one - dimensional photonic crystal , '' _ physica b _ , 407 , 120 - 125 , 2012 . chen , y. h. , g. q. liang , j. w. dong , and h. z. wang , `` derivation and characterization of dispersion of defect modes in photonic band gap from stacks of positive and negative index materials , '' _ phys . 351 , 446 - 451 , 2006 . kang - song , t. , x. yuan - jiang , and w. shuang - chun , `` defect in photonic crystal with negative index material , '' _ optoelectron . _ , vol . 2 , 118 - 121 , 2006 jiang , h. , h. chen , h. li , and y. zhang , `` omnidirectional gap and defect mode of one - dimensional photonic crystals containing negative - index materials , '' _ appl . 83 , 5386 - 5388 , 2003 . wang , l. g. , h. chen , and s. y. zhu , `` omnidirectional gap and defect mode of one - dimensional photonic crystals with single - negative materials , '' _ phys . b _ , vol . 70 , 245102 - 245107 , 2004 . xu , q. , k. xie , h. yang , and j. tang , `` periodic defect modes of one - dimensional crystals containing single - negative materials , '' _ optik _ , vol . 121 , 1558 - 1562 , 2010 . xiang , y. , x. dai , s. wen , and d. fan , `` properties of omnidirectional gap and defect mode of one - dimensional photonic crystal containing indefinite metamaterials with a hyperbolic dispersion , '' _ _ , vol . 102 , 093107 - 093111 , 2007 . born , m. and e. wolf , _ principles of optics : electromagnetic theory of propagation , interference and diffraction of light _ , cambridge university press , uk , 2005 . zhang , l. , y. zhang , l. he , z. wang , h. li , and h. chen , `` zero-@xmath34 gaps of photonic crystals consisting of positive and negative index materials in microstrip transmission lines , '' _ j. phys . 40 , 2579 - 2587 , 2007 . li , j. , l. zhou , c. t. chan , and p. sheng , `` photonic band gap from a stack of positive and negative index materials , '' _ phys . 90 , 083901 - 083904 , 2003 . aghajamali , a. and m. barati , `` effects of normal and oblique incidence on zero-@xmath34 gap in periodic lossy multilayer containing double - negative materials , '' _ physica b _ , vol . 407 , 1287 - 1291 , 2012 . daninthe , h. , s. foteinopoulou , c. m. soukoulis , `` omni - reflectance and enhanced resonant tunneling from multilayers containing left - handed materials , '' _ photoni . _ , vol . 4 , 123 - 131 , 2006 awasthi , s. k. , a. mishra , u. malaviya , and s. p. ojha , `` wave propagation in a one - dimensional photonic crystal with metamaterial , '' _ solid state commun . 149 , 1379 - 1383 , 2009 .
transmission properties of one - dimensional lossy photonic crystals composed of negative and positive refractive index layers with one lossless defect layer at the center of the crystal are investigated by the characteristic matrix method . the results show that as the refractive index and thickness of the defect layer increase the frequency of the defect mode decreases . in addition , it is shown that the frequency of the defect mode is sensitive to the incidence angle , polarization and physical properties of the defect layer but it is insensitive to the small lattice loss factor . the height of the defect mode is very sensitive to the loss factor , incidence angle , polarization , refractive index and thickness of the defect layer . it was also shown that the height and the width of the defect mode are affected by the number of the lattice period and the loss factor . the results can lead to designing new types of narrow filter structures and other optical devices .
common envelope evolution , in which the components of a binary system are engulfed by a common gaseous envelope , is a brief but crucial phase in the formation of many compact binary systems . drag forces , due to velocity differences between the orbiting components of the binary and the surrounding gas , work to shrink the binary orbit , while the potential energy released acts to expel the common envelope . extensive numerical work ( e.g. bodenheimer & taam 1984 ; livio & soker 1988 ; taam , bodenheimer & rozyczka 1994 ; sandquist et al . 1998 ) has only partially succeeded in reducing the large uncertainties surrounding the efficiency of this process ( iben & livio 1993 ; livio 1994 ; rasio & livio 1996 ) . for neutron stars , an equally fundamental uncertainty concerns the _ accretion rate _ during the common envelope phase . at the densities of a typical giant envelope , the bondi - hoyle accretion rate @xmath1 for an inspiralling neutron star would be extremely large , typically of the order of @xmath2 , and often much larger still . this is many orders of magnitude in excess of the eddington rate of @xmath3 obtained by equating the outward radiation pressure with gravity leading to starkly different predicted outcomes . if @xmath4 is limited to the eddington rate ( or is even several orders of magnitude higher ) , then ejection of the stellar envelope will occur when the total accreted mass is @xmath5 , and the neutron star will survive . conversely , uninhibited accretion at close to @xmath1 will lead almost inevitably to collapse to a black hole . recent work has tended to suggest that @xmath1 is the appropriate rate for neutron star accretion in a stellar envelope ( see e.g. bethe & brown 1998 and references therein ) . at sufficiently high accretion rates , the eddington limit becomes irrelevant because photons are trapped and advected inward with the flow ( rees 1978 ; begelman 1979 ; blondin 1986 ) . a hot envelope develops around the neutron star , and eventually high enough densities and temperatures are reached that the accretion energy can be radiated away by neutrinos . hypercritical _ mode of accretion has been considered in the context of fallback onto neutron stars in supernovae ( e.g. colgate 1971 ; zeldovich , ivanova & nadezhin 1971 ; chevalier 1989 ; houck & chevalier 1991 ) , as a component of models for gamma - ray bursts ( popham , woosley & fryer 1999 ) , and applied to the problem of accretion in a stellar envelope ( chevalier 1993 , 1996 ; brown 1995 ; fryer , benz & herant 1996 ) . a striking consequence of ubiquitous hypercritical accretion would be that neutron star - neutron star binaries _ can not _ form from massive binaries via the usual common envelope evolution route , which would instead lead to black hole - neutron star binaries . the observed population of neutron star - neutron star binaries would have to form instead via a rarer channel in which the binary progenitors are of nearly equal mass ( brown 1995 ) , and would be an order of magnitude rarer than black hole - neutron star binaries ( bethe & brown 1998 ) . the robustness of this conclusion is of great interest , since these binaries are both potential gamma - ray burst progenitors , and the most promising targets for the early ligo detection of gravitational waves ( abramovici et al . 1992 ) . the velocity of the neutron star relative to the common envelope ensures that the accretion flow can not be spherically symmetric , at least at large radii of the order of the accretion radius . moreover , gradients in the envelope structure across the accretion radius introduce angular momentum into the flow , which modifies its properties even at small radii , via the formation of a rotationally supported disk . in this paper we investigate how these complexities affect hypercritical accretion via hydrodynamic simulations of the outer parts of the accretion flow . our approach thus complements the consideration of angular momentum presented by chevalier ( 1996 ) , though as discussed later we draw different conclusions as to its likely importance . the plan of this paper is as follows . in 2 we briefly review the properties of spherical hypercritical accretion that are relevant for the common envelope application , and in 3 we outline the numerical methods used . the principal computational limitation is the restriction to two dimensional simulations . results for flows with zero and non - zero angular momentum are presented in 4 and 5 , and the implications and remaining uncertainties discussed in 6 . the simplest approximation to the accretion flow in a common envelope is to describe it as bondi - hoyle - lyttleton accretion from a uniform medium onto a point mass . the properties of such flows have been extensively studied , both analytically ( hoyle & lyttleton 1939 ; bondi & hoyle 1944 ) , and numerically ( e.g. ruffert 1997 ; kley , shankar & burkert 1995 ; benensohn , lamb & taam 1997 , and references therein ) . for an accretor of mass @xmath6 , the critical radius is the accretion radius in the literature , but all lead to accretion rates that are essentially the same for our purposes ] , @xmath7 where @xmath8 and @xmath9 are the velocity and sound speed far upstream of the accreting star . material falling into a cylinder with this radius is accreted , so that , @xmath10 where @xmath11 is the density of the gas far upstream of the accretor , and the last factor is included as it provides better results for low mach number accretors . the timescale for setting up the flow is of the order of the sound crossing time at the accretion radius , @xmath12 . generally , since we are interested in inspiral through a pressure supported atmosphere , we expect mach numbers of order unity , though this will vary somewhat at different radii in the star . numerical values for @xmath13 and @xmath1 require specification of a stellar model . fryer , benz & herant ( 1996 ) tabulate values for several giant and main sequence stars ( the giants are of greater interest here , though somewhat analagous considerations apply to neutron star - main sequence collisions , for example in globular clusters ) . for example , for a @xmath14 giant @xmath15 , while @xmath1 varies from @xmath16 on upwards . for a more distended @xmath17 model lower accretion rates of @xmath18 are of course expected in the outer regions . these accretion rates can only be realized in the hypercritical regime , since the eddington limit on the luminosity , @xmath19 where @xmath20 is the proton mass and @xmath21 the thomson cross - section , corresponds to a few @xmath22 . for hypercritical accretion to occur the timescale for photons to diffuse out of the flow must exceed the timescale for them to be advected inwards . for a spherical flow whose opacity @xmath23 is dominated by electron scattering , the optical depth is @xmath24 and the diffusion time is roughly , @xmath25 for a flow in free - fall , where @xmath26 , equating @xmath27 to the free - fall time @xmath28 yields an estimate for the trapping radius ( blondin 1986 ) , @xmath29 in addition to the photons being trapped , we must also be able to eventually lose the accretion energy from the system . for a black hole , this is possible at any accretion rate since the energy can be invisibly advected across the horizon ( e.g. narayan , garcia & mcclintock 1997 ) . in a neutron star system , high accretion rates are required because the pressure at the base of the accretion flow must be able to become large enough to allow neutrino losses ( primarily from electron positron annihilation into neutrino pairs ) to balance the accretion energy , @xmath30 . in the spherically symmetric case , this requirement usually requires the formation of a hot envelope at small radii , in which @xmath31 is subsonic , bounded from free falling gas by a shock at radius @xmath32 . within this inner region , the density and pressure scale as a radiation dominated @xmath33 ( @xmath34 ) polytrope ( e.g. brown 1995 ; fryer , benz & herant 1996 ) , @xmath35 a detailed calculation ( houck & chevalier 1991 ; brown 1995 ) finds for @xmath32 , @xmath36 requiring that @xmath37 gives a limit on the accretion rate , of roughly @xmath38 . even allowing for a generous margin of error in this crude estimate , for example due to a non - spherical geometry , this is substantially less than the bondi - hoyle rate in an envelope , which should thus be safely in the hypercritical regime . of course we also require in the common envelope case that @xmath39 , but for spherical accretion and @xmath40 this is an easily satisfied constraint , @xmath41 . if angular momentum and other possible complications can be neglected , the scenario for the inspiral of a neutron star then has three stages . first , the neutron star accretes at less than the eddington rate as it encounters the tenuous outer atmosphere of the companion . negligible mass is accumulated in this phase . as the density grows , the implied bondi - hoyle rate first rises above @xmath42 , but is below @xmath43 . in this regime outward diffusion of photons is effective at limiting accretion , again to negligible levels . finally , the bondi - hoyle rate rises well above the critical value for hypercritical accretion , and gas starts accreting freely onto the neutron star . the outcome then depends critically on how much mass can be accumulated before enough energy has been lost to unbind the envelope . brown ( 1995 ) estimates this energy as , @xmath44 where @xmath45 is the accreted mass , and @xmath46 is the average mean velocity squared during the hypercritical accretion phase . for typical binding energies and velocities , @xmath47 , sometimes vastly so , and the mass accreted will probably exceed the maximum mass for a neutron star and force collapse to a black hole . we investigate how angular momentum could affect the above scenario of hypercritical accretion by studying the inviscid , purely hydrodynamic behavior of infalling gas at large radii , @xmath48 . the hypercritical regime corresponds to the complete dominance of advection of photons over diffusion , so that the effects of radiation transport are negligible . we also neglect magnetic fields , which is potentially less forgivable . to attain the required dynamic range in three dimensional calculations including magnetic fields unfortunately remain difficult . we have investigated two dimensional simulations both in spherical polar geometry ( with @xmath49 parallel to @xmath50 and axisymmetry in the @xmath51 direction ) , and in cylindrical geometry @xmath52 . the spherical polar simulations represent the most faithful two dimensional representation of the flow , but necessarily exclude the possibility of studying angular momentum or density gradients in the ambient medium . for such zero angular momentum flows , we found only that the analytic bondi - hoyle estimate of equation ( [ eq2 ] ) provides a good estimate of the accretion rate ( for low mach numbers the actual accretion rate is somewhat higher than the estimate , but the difference is not so great as to qualitatively affect the result ) . for these reasons ( and following benensohn , lamb & taam 1997 and most other two dimensional calculations ) we focus on the cylindrical polar calculations . the neutron star is represented as a point mass at @xmath53 , surrounded by a fixed computational mesh in @xmath54 . we use uniform zoning in the @xmath51 direction , and choose the radial grid such that @xmath55 with @xmath56 a constant . this amounts to choosing cells that have the same shape , in our case roughly square , at all radii . cylindrical polar simulations in which @xmath57 is the ignored co - ordinate allow for angular momentum , but correspond to the rather unphysical assumption that the central mass moves through a thin sheet of gas with zero gradients in the perpendicular direction . this may lead to qualitative errors , for example in the strength and prominence of ` flip - flop ' instabilities at high mach numbers ( compare e.g. livio 1992 ; benensohn , lamb & taam ( 1997 ) ; ruffert ( 1997 ) ) , though untangling the effects of geometry and resolution in the various calculations is difficult . we will not be interested here in transient disks or instabilities in the wake , where these issues are most worrisome . we assume that the gas is dominated by radiation pressure out to the outer boundary of the simulation at a few @xmath13 . the equation of state can then be modelled using a simple adiabatic relation , @xmath58 where @xmath59 is the energy per unit volume and we take @xmath33 , corresponding to a radiation dominated gas where @xmath60 . the calculations use the zeus-3d code developed by the laboratory for computational astrophysics ( clarke , norman & fiedler 1994 ) . zeus is an eulerian finite difference code which employs an artificial viscosity to handle shocks . the algorithms and design of the code are closely similar to those detailed by stone & norman ( 1992a , 1992b ) . for the outer boundary condition , we impose inflow of gas at velocity @xmath8 and sound speed @xmath9 at the outer boundary , @xmath61 , for @xmath62 . over the remainder of the outer boundary outflow boundary conditions are specified , implemented as simple continuation of fluid variables on the grid into the boundary zones . we take a mach number @xmath63 , appropriate to the generally mildly supersonic accretion flows in common envelopes . an inspiralling neutron star encounters radial gradients in @xmath11 , @xmath9 and @xmath8 across the accretion radius . in most cases the density gradient is the principal effect ( e.g. fryer , benz & herant 1996 ) , which we model as a simple exponential , @xmath64 where @xmath65 is the radial distance in the common envelope of the unperturbed medium from the neutron star orbit , and @xmath66 measures the strength of the gradient across the accretion radius . values of @xmath66 for various stars vary substantially , and have been tabulated by fryer , benz & herant ( 1996 ) . the inner boundary condition is reflecting ( @xmath67 at @xmath68 ) , allowing for the formation of a pressure supported inner envelope for zero angular momentum accretion . this implicitly assumes low accretion rates where the shock radius @xmath32 is large enough to exceed @xmath69 . for runs with angular momentum , the choice of inner boundary condition is less important as angular momentum provides substantial support against gravity at small radii . our boundary condition then amounts to assuming that radial flow through the disk is slow the plausibility of this can be verified post facto by studying the properties of the disks formed in the calculation . figure 1 shows results for a series of bondi - hoyle accretion simulations in which the density gradient in the ambient medium was varied from @xmath70 to @xmath71 . all the calculations used a grid with @xmath72 and @xmath73 . the inner boundary was at @xmath74 and the outer boundary at @xmath75 , giving a grid with @xmath76 . the calculations were run until @xmath77 , where the time unit @xmath78 is the sound crossing time of the accretion radius @xmath13 . we plot in fig . 1 only the inner region of the accretion flow . the @xmath79 run was also recomputed at modestly higher resolution ( @xmath80 ) until @xmath81 , in order to ascertain how quickly mass accumulated in the outer parts of the disk at a later time . no qualitative changes were observed to occur during this longer simulation . in the absence of density gradients in the ambient gas , the structure of the flow resembles closely that seen in simulations of bondi - hoyle accretion with the same parameters and an absorbing inner boundary condition . a pressure supported , roughly symmetric envelope has developed around the central object , and this displaces the bow shock upstream into the flow . at this resolution and mach number , the flow is found to be only rather weakly transient . increasing the density gradient first leads to an asymmetric displacement of the bow shock , followed for more extreme density gradients by the formation of a clear disk in the inner regions of the simulation . for the parameters adopted here , @xmath82 suffices to create a clear and persistent disk surrounding the central object . figure 2 plots the velocity and density fields corresponding to the images in fig . 1 . over this radial range there is a large variation in velocity , which we normalise to the local keplerian value @xmath83 . for the zero density gradient case , the outer flow is clearly not cylindrically symmetric , but within @xmath84 of the central object the envelope is both reasonably symmetric and characterised by only small rotational motions . conversely , a clear disk is produced in the calculations with stronger density gradients , @xmath82 , the flow here is disk - like out to at least @xmath85 . figure 3 shows how the volume averaged rotational velocity from the simulations , normalised to @xmath86 , varies as a function of radius . negligible rotation is seen in the @xmath70 run , with @xmath87 at all radii , while the runs with density gradients all show the signature of a disk in which both pressure gradients and rotational support are significant . for the two runs with the largest gradients , a sizeable disk extending out to almost @xmath13 , with typical angular velocity @xmath88 , is clearly produced . for hypercritical accretion , the disks formed around the neutron star will be advection dominated . the properties of advection dominated flows have been extensively studied , both at the high accretion rates of interest here ( e.g. begelman & meier 1982 ) and more recently at lower accretion rates ( e.g. narayan & yi 1994 , 1995 ) . these disks are hot , geometrically thick , and only weakly bound to the accreting object . the presence of a disk may affect the outcome of common envelope evolution in two ways , via a modification of the accretion rate onto the neuton star or via feedback of energy into the stellar envelope . the most direct potential influence of a disk is via the accretion rate . as infall continues , the disk will reach a quasi - steady state in which the rate of infall onto the outer disk balances the rate of disk accretion . this rate need not be the bondi - hoyle rate , and in principle could be lower ( for an illustration , in our inviscid simulations , where the viscosity is very low , we eventually reach a steady state where there is close to _ zero _ ongoing accumulation of mass in the disk ) . we estimate whether this is important below , and show that for the usually considered values of @xmath89 , the shakura - sunyaev ( 1973 ) viscosity parameter , the disk is probably able to transport mass inwards at the bondi - hoyle rate . the presence of a disk also makes the formation of outflows or jets probable ( e.g. livio 1999 ) . strong outflows could themselves affect the fate of the system by reducing the accretion rate onto the neutron star . much weaker jets , if they arose from deep in the neutron star potential well , would still provide an important additional energy input into the envelope , and shorten the common envelope phase . the accretion rate through the disk can be estimated from the measured disk mass and the inferred viscous timescale . for a viscosity parameterized via the shakura - sunyeav ( 1973 ) form , @xmath90 , and at radius @xmath91 the viscous timescale is , @xmath92 the mass in the disk will drain onto the central object on the viscous timescale at the outer radius , with an accretion rate @xmath93 . the value of @xmath89 appropriate to thick , radiation dominated disks is unknown . if mhd instabilites ( balbus & hawley 1991 ) are responsible for the viscosity , then simulations of thin gas pressure dominated disks find typically that @xmath94 ( stone et al . 1996 ; brandenburg et al there remain large uncertainties in the theoretical expectation for disks of the kind that we are interested in . however , taking @xmath95 , and evaluating the accretion rate from the disk mass and run of sound speed obtained at @xmath96 in the long duration @xmath79 run , we find that @xmath97 . this is a crude estimate , which in particular ignores completely the expansion of the disk expected from viscous evolution ( lynden - bell & pringle 1974 ) . as a result , the value @xmath1 should be regarded as an upper limit to the mean accretion rate through the disk . furthermore , at least at small radii the accretion is unlikely even to be steady ( chevalier 1996 ) , since the much slower fall - off of pressure with radius in an advection dominated disk as compared to a spherical envelope ( @xmath98 as compared to @xmath99 ) requires non - steady accretion to reach the extreme conditions required for neutrino emission . however , it suggests that there is no strong reason to believe that , in the absence of outflows , the formation of a disk creates an insurmountable bottleneck to rapid accretion . most accretion disk systems are observed to generate jets or less well collimated outflows . although a thick disk such as that formed in the simulations has a relatively small density contrast between the equatorial plane and the polar regions , advection dominated flows are also hot enough that the gas is only weakly bound to the central mass . detailed solutions show that the bernoulli constant , @xmath100 which measures the energy the gas would possess if adiabatically moved to infinity , is positive for an often wide range of angles close to the poles ( narayan & yi 1994 ; 1995 ) . although the outcome depends additionally on the outer boundary conditions and the detailed physics of outflow generation , this positivity of @xmath101 is likely to imply that outflows are a generic feature of advection dominated disks ( narayan & yi 1995 ; blandford & begelman 1999 ) . for our purposes , we can distinguish two extreme possibilities , in which outflows are either self - similar or generated exclusively from the inner disk at @xmath102 . _ self - similar _ outflows from advection dominated flows represent the model considered by blandford & begelman ( 1999 ) . in this case the fraction of accreting mass lost in the outflow is the same for each decade in disk radius , so that the remaining mass accretion rate through the disk decreases inwards as @xmath103 with @xmath104 . the appropriate value of @xmath105 , which within this model measures the efficiency with which mass is ejected , has not been determined for any realistic thick disk model , though it is has been suggested that @xmath105 is large for simulations of convection in thick disks ( stone , pringle & begelman 1999 ) . empirically @xmath105 must be @xmath106 if this model is to be successful in explaining the extremely low radiative efficiencies of disks around black holes in low luminosity galactic nuclei ( e.g. reynolds et al . 1996 ) . in the common envelope case , the disk extends over an extremely large range of radii , from the neutron star surface at @xmath107 out to of the order of the accretion radius at @xmath108 . as a result , the consequence of outflows that are inefficient at removing angular momentum would be to greatly reduce the accretion rate onto the neutron star . cygnus x-2 may be an example of a neutron star that has survived an episode of super - eddington mass transfer ( though here during thermal timescale mass transfer rather than common envelope evolution ) without accreting a significant fraction of the transferred mass ( king & ritter 1999 ; king & begelman 1999 ) . if this is indeed the case , it provides some support for the scenario of strong outflows under physical conditions analagous to those encountered during common envelope inspiral . alternatively , outflows may arise predominantly from the _ inner disk_. in this case , the energy feedback into the common envelope from outflows originating deep in the neutron star potential well could be highly significant . if , as observations of jet systems suggest ( livio 1999 ) , the jet is launched with a velocity roughly equal to the local keplerian velocity , @xmath109 , then the energy deposition after @xmath110 of mass has been ejected is just , @xmath111 where @xmath112 is an efficiency factor that will depend on the specific jet model . ejecting the common envelope requires an energy deposition of around @xmath113 ( brown 1995 ) , which could be achieved for a mass loss in a jet as low as , @xmath114 the energetic feedback from such an outflow could thus be important for hastening the ejection of the envelope , even if the mass loss itself was far too small to significantly impact the accretion rate . indeed for this type of jet a large energy deposition could only be avoided if either the accretion rate was @xmath115 , or if the efficiency of the jet production or coupling to the envelope was extremely low . a low efficiency of coupling to the envelope could arise if the jet was extremely well collimated and able to escape the star entirely , which remains a possibility . however , even for a collimation angle of @xmath116 , which is typical for many jet sources , the ejection of @xmath117 is sufficient to unbind the envelope . if jets are able to deposit energy into the envelope then an immediate consequence would be that the outcome of common envelope evolution should depend on the depth of the potential well at the surface of the inspiralling compact object more compact objects should lead to a higher value of the common envelope efficiency parameter @xmath118 and eject the envelope more easily . in this paper we have considered the probable fate of neutron stars during common envelope evolution . numerical simulations of the hydrodynamics of bondi - hoyle accretion at large radii from the neutron star show that modest density gradients , typical of those expected in giant envelopes , lead to persistent rotationally supported disks around the accreting object . the presence of angular momentum means that spherically symmetric treatments of hypercritical accretion are likely to be a poor approximation for studying the the common envelope phase . in particular , whether the neutron star accretes enough mass to force it to collapse to a black hole prior to loss of the envelope depends entirely on the subsequent evolution of a thick , advection dominated , accretion disk . little is known about the behavior of such disks , opening plenty of room for legitimate dispute as to the outcome . simple estimates suggest that the disks formed in the simulations are sufficiently hot and thick to support accretion rates which could be as large as the inferred bondi - hoyle rate in the common envelope phase , provided only that the disk viscosity is not surprisingly small . the presence of a disk does not create a serious bottleneck in the accretion flow at @xmath119 . however , more complex and uncertain physics is likely to come into play close to the neutron star . jets are observed almost universally from accreting systems possessing thick , hot accretion disks , of the kind envisaged here . in the case of ss433 , which arguably is the closest prototype to the physical condition of a neutron star in a common envelope , the mass outflow in the jets is extremely strong ( e.g. watson et al . 1986 ) . during common envelope evolution , a jet would allow for a strong feedback of energy from close to the neutron star surface into the stellar envelope , leading to a more rapid ejection of the envelope than would be possible from gas and gravitational drag alone . since the fate of the neutron star depends on a sensitive balance between the rate at which it accretes and the rate at which energy is deposited into the common envelope ( brown 1995 ) , this would improve the chances of neutron star survival both by lowering the accretion rate and by reducing the epoch of common envelope evolution . in general , jets must be avoided at all costs if the neutron star is to be able to accrete a large mass during inspiral . it is also possible that enough of the radiated neutrino energy could be absorbed at a larger radius to drive explosions , as discussed in the context of spherical accretion models by fryer , benz & herant ( 1996 ) . observationally , several binary pulsars are known whose properties would be consistent with the neutron star having survived a phase of common envelope evolution . camilo et al . ( 1996 ) identify four pulsars which have relatively large companion masses ( in excess of 0.45 @xmath120 ) , and which do not follow the eccentricity orbital period relation expected for lower mass binary pulsars ( phinney 1992 ) . these systems are likely to have undergone deep common envelope evolution ( van den heuvel 1994 ) , although probably the companions were of rather lower mass ( 16 @xmath120 ) than we have been discussing here . nonetheless , what is striking is that these neutron stars appear to have not only survived , but are now observed to be rotating with spin periods that are rapid , yet significantly slower than pulsars believed to have been spun - up via disk accretion . this is consistent with the scenario argued here , in which common envelope evolution involves accretion rates that are vastly super - eddington yet still insufficient to force collapse of the neutron star to a black hole . it is also consistent with the suggestion of brown , lee & bethe ( 1999 ) , who point out that the inferred masses of black holes in transient sources ( which appear to be a large fraction of the mass of the immediate progenitor ) imply that hypercritical accretion onto black holes in supernovae must be reasonably efficient . in supernovae , mass must be physically ejected from the system to avoid being accreted eventually , whereas in the common envelope case a modestly lowered accretion rate can suffice to enable the neutron star to survive until the envelope is lost .
neutron stars inspiralling into a stellar envelope can accrete at rates vastly exceeding the eddington limit if the flow develops pressures high enough to allow neutrinos to radiate the released gravitational energy . it has been suggested that this hypercritical mode of accretion leads inevitably to the formation of stellar mass black holes during common envelope evolution . we study the hydrodynamics of this flow at large radii ( @xmath0 ) , and show that for low mach number flows , in two dimensions , modest density gradients in the stellar envelope suffice to produce a hot , advection dominated accretion disk around the accreting object . the formation of outflows from such a disk is highly probable , and we discuss the impact of the resultant mass loss and feedback of energy into the envelope for the survival of the neutron star . unless outflows are weaker than those inferred for well observed accreting systems , we argue that in most cases insufficient accretion occurs to force collapse to a black hole before the envelope has been ejected . this conclusions is of interest for black hole formation in general , for some models of gamma ray bursts , and for predictions of the event rate in future ligo observations .
we thank fan wei and jeroen van den brink for helpful discussions . x liu was supported by the strategic priority research program of the chinese academy of sciences ( grant no : xdb07020200 ) . the work at brookhaven national laboratory was supported by the u.s . department of energy ( doe ) , division of materials science , under contract no . de - sc0012704 . j liu is supported by the director , office of science , office of basic energy sciences , materials sciences and engineering division , of the u.s . department of energy under contract no . de - ac02 - 05ch11231 through the quantum materials program . 20 ament l j p 2011 _ rev . phys . _ * 83 * 705 dean m p m 2015 _ journal of magnetism and magnetic materials _ * 376 * 3 kim y j 2012 _ synchrotron radiation news _ * 25 - 4 * 3 ament l j p 2009 _ phys . lett . _ * 103 * 117003 braicovich l 2010 _ phys . lett . _ * 104 * 077002 kim j 2012 _ phys . lett . _ * 108 * 177003 kim j 2012 _ phys . lett . _ * 109 * 157402 yin w g 2013 _ phys . lett . _ * 111 * 057202 kim j 2014 _ nature communications _ * 5 * 4453 lee w s 2013 _ phys . lett . _ * 110 * 265502 six project at the national syncrotron light source ii , http://www.bnl.gov/ps/nsls2/beamlines/files/pdf/sixposter.pdf . upbl07 project at the european synchrotron radiation facility . i21 webpage at the diamond light source , http://www.diamond.ac.uk / home / beamlines / i21.html/. webpage for the veritas beamline at max iv , https://www.maxlab.lu.se/veritas . huang d j and chang s l 2014 _ synchrotron radiation news _ * 27 - 1 * 10 chiuzbian s g 2014 _ review of scientific instruments _ * 85 * 043108 serrao c r 2013 _ phys . rev . b _ * 87 * 085121 sala m m 2014 _ phys . rev . b _ * 89 * 121101 kim b h and brink j v d _ arxiv e - prints _ 1404.2040 dean m p m 2012 _ nat . mater . _
resonant inelastic x - ray scattering ( rixs ) at the @xmath0-edge of transition metal elements is now commonly used to probe single magnon excitations . here we show that single magnon excitations can also be measured with rixs at the @xmath1-edge of the surrounding ligand atoms when the center heavy metal elements have strong spin - orbit coupling . this is demonstrated with oxygen @xmath1-edge rixs experiments on the perovskite sr@xmath2iro@xmath3 , where low energy peaks from single magnon excitations were observed . this new application of rixs has excellent potential to be applied to a wide range of magnetic systems based on heavy elements , for which the @xmath0-edge rixs energy resolutions in the hard x - ray region is usually poor . _ keywords : single magnon excitation , rixs , oxygen @xmath1-edge , iridates _ + recently , the resonant inelastic x - ray scattering technique has achieved great progress in condensed matter physics research @xcite . it has now been clearly demonstrated that rixs is not only suitable for studying charge and orbital excitations , but it is also a powerful tool to probe magnetic dynamics . with rixs at the @xmath0-edge of the magnetic elements , the magnetic dispersion curves were mapped out on various transition metal compounds such as cuprates @xcite and iridates @xcite . although the leading term of the photon - electron interaction does not directly couple to the spin , the strong spin - orbit coupling ( soc ) of the core level _ p _ states accessed at the @xmath0-edge makes direct spin - flip scattering possible in the rixs process @xcite . combined with the large energy transfer and appreciable momentum transfer capability of x - ray scattering , this makes rixs an attractive tool to measure magnetic dispersions , especially in situations where small sample volumes make inelastic neutron scattering difficult to apply . depending on the resonant energy of the targeted element , rixs is naturally divided into two classes : soft x - ray and hard x - ray @xcite . in a soft rixs setup , x - ray dispersing elements are diffraction gratings , which typically operate efficiently in an x - ray energy range up to about 2000 ev . the @xmath0-edges of the @xmath4 transition metal elements all fall below this energy scale . for high atomic number , @xmath5 , transition metal elements , the @xmath0-edge resonant energies are in the hard x - ray region and the rixs setup is totally different . instead of continuously tunable gratings , high quality single crystals are needed as the key optical elements . the rixs resolution crucially depends on the availability of a high quality single crystal with a bragg diffraction peak close to back - scattering at the energy of the @xmath0-edge of the targeted element . this requirement severely limits the application of rixs , and thus by far the majority of hard x - ray rixs studies of single spin - flip magnetic excitations have focused on @xmath6 iridates @xcite . to overcome this resolution limitation , we suggest that it is possible to probe single magnon excitations with rixs at the @xmath1-edge of the surrounding ligand atoms whose @xmath7 orbitals hybridize with the @xmath8 orbitals of a heavy element . in this rixs process , a @xmath9 core electron of the ligand atom is excited into the outer level @xmath7 states and the @xmath8 shell electronic configuration of the transition metal element is altered through the @xmath7-@xmath8 hybridization . the key ingredient to flip a spin , namely the strong spin - orbit coupling , is fulfilled by the @xmath8 electrons of the high @xmath5 index transition metal elements instead of the 1@xmath10 core level of the ligand atoms . the energy resolution at the @xmath1-edge of the common ligand atoms is often particularly high . for example , the best energy resolution currently achieved for rixs at the oxygen @xmath1-edge in the soft x - ray region is about 50 mev @xcite , which is much better than the majority of @xmath6 elements probed using @xmath0-edge rixs in the hard x - ray region . furthermore the energy resolution is set to improve significantly in coming years due to the development of new soft x - ray rixs beamlines @xcite . for demonstration , we carried out oxygen @xmath1-edge rixs experiments on the perovskite sr@xmath2iro@xmath3 . this material is antiferromagnetically ordered at low temperature ( @xmath11 k ) , and its spin - wave magnetic dispersion has been mapped out in detail using ir @xmath12-edge rixs @xcite . by measuring the rixs spectrum at the oxygen @xmath1-edge , peaks consistent with single magnon excitations were observed . our results demonstrate that the oxygen @xmath1-edge rixs is a powerful tool in probing the magnetic excitations in 5@xmath8 transition metal oxides . the oxygen @xmath1-edge rixs experiments were performed using the aerha instrument at the sextants beamline at the soleil synchrotron in france @xcite . the sr@xmath2iro@xmath3 samples were 200 nm thick epitaxial films deposited on srtio@xmath13 substrates using pulsed laser deposition @xcite . the sample was orientated such that the tetragonal crystal axes @xmath14 and @xmath15 were in the scattering plane . all data were collected at 25 k and the rixs spectra were taken with the detector at a fixed scattering angle of @xmath16 . for different incident x - ray polarization , the x - ray couples to ir @xmath6 orbitals differently , which gives rise to different pre - peak features at the oxygen @xmath1-edge absorption curve @xcite . to maximize the cross - section , we chose the excitation into the @xmath17 and @xmath18 ir orbitals via the in - plane oxygen atoms . the resonant energy of these excitations is determined by oxygen @xmath1-edge total electron yield spectra , shown in figure [ fig : tey ] . the data were collected with the incident x - ray polarization vector to be 15 degrees away from surface normal . thus it is mainly along the crystal @xmath19 direction . the pre - peak at 529.5 ev ( indicated by a dashed line ) is the excitation of in - plane oxygen @xmath9 core electron in to the ir @xmath17 and @xmath18 orbitals via @xmath7-@xmath8 hybridization . -edge taken with the incident x - ray polarization vector to be @xmath20 away from surface normal . the dashed line indicates the prepeak at 529.5 ev.,width=8 ] with the incident x - ray energy at 529.5 ev on the prepeak at the oxygen @xmath1-edge , the rixs spectra were taken with the detector at a fixed scattering angle of @xmath16 . by rotating the sample @xmath21 angle , the total momentum transfer , , varies in the @xmath22-@xmath0 plane as @xmath23 . depending on the relative angle between the incident -@xmath8 transition and higher energy features are dominated by fluorescence.,width=8 ] x - ray polarization and the scattered beam directions @xcite , the intensity of elastic scattering in soft x - ray rixs measurements can sometimes be much weaker than the inelastic rixs intensity @xcite . in cases like these it can be difficult to unambiguously determine the exact elastic energy in the rixs spectrum . therefore for each point measured in the experiments , we first used a circularly polarized incident beam to get appreciable elastic line signal to precisely calibrate the zero energy loss definition on the ccd detector . the result is shown in figure [ fig : circular ] . except the sharp elastic line at zero energy loss , there is a peak at 0.66 ev and multiple features above 2 ev . the peak at 0.66 ev corresponds to the @xmath8-@xmath8 transition whose energy is consistent with that observed with ir @xmath0-edge rixs @xcite . features at higher energy are mainly from the oxygen @xmath1-edge fluorescence . to better observe the low energy excitations , such as magnons , the overwhelming contribution from the elastic line needs to be minimized . this was done by switching to a @xmath24 polarized incident beam . with the @xmath24 incident polarization , the non - spin - flipped scattering process is strictly forbidden at the scattering angle of @xmath25 @xcite . since our detector was placed close this ( @xmath16 ) , the elastic line was largely suppressed . figure [ fig : qdep ] plots rixs spectra collected at this condition . indeed the elastic scattering is almost invisible , and low energy excitation peaks appear . incident polarization . the dashed line is with circular polarization , as shown in figure [ fig : circular].,width=8 ] since the magnetic interactions in sr@xmath2iro@xmath3 are almost completely two dimensional , we index the momentum transfer with only its @xmath22 component . for @xmath26 and @xmath27 , the low energy peaks are at 52 and 44 mev respectively , with an error bar of 10 mev ( estimated from half of the scanning step size ) . these energies agree very well with the magnon dispersion measured earlier @xcite . thus we conclude that the observed low energy peaks are from single magnon excitations . the observed low energy magnon excitation shows distinct dependence : its intensity quickly reduces when the in - plane momentum transfer moves towards @xmath28 . for the other two points closer to @xmath29 , the magnon excitations are barely observable this behavior is consistent with recent theoretical calculations by b. h. kim _ et al_. @xcite . they predict that oxygen @xmath1-edge rixs measurements of single magnons in sr@xmath2iro@xmath3 have a minimum in intensity at @xmath28 , and its size is governed by the buckling of the ir - o - ir bond angle . the agreement with this theoretical prediction further confirms that the observed low energy peaks are indeed from single magnon excitations . in conclusion , we show that single magnon excitations in antiferromagnetically ordered sr@xmath2iro@xmath3 can be probed with rixs at the @xmath1-edge of the ligand oxygen atoms . this technique helps to overcome the resolution limitation of @xmath0-edge rixs in the hard x - ray region for high @xmath5 index transition metal compounds . it has excellent potential to be applied to a wide range of magnetic systems based on heavy elements , and crucially , it is furthermore possible to measure very small sample volumes including thin films and heterostructures , something that until now as been largely limited to cuprate - based systems @xcite .
the recent discovery of a vast population of free - floating brown dwarfs in the vicinity of the sun @xcite has led to the definition of two new spectroscopic classes , l dwarfs and t dwarfs . these are the first additions to the standard stellar sequence in over 60 years , encompassing objects cooler than the m spectral class . l dwarfs @xcite are characterized by absorption bands of metal hydrides ( e.g. , feh , crh , and mgh ) , strong alkali lines ( li , na , k , rb , and cs ) , and h@xmath2o and co absorption in the near - infrared . below effective temperatures @xmath3 , co reduces to @xmath4 @xcite , a molecule which has distinct absorption bands at 1.6 and @xmath5 . objects exhibiting these @xmath4 features have grossly different spectroscopic and photometric properties as compared to l dwarfs , and are designated t dwarfs @xcite . the spectroscopic sequence m to l to t is the evolutionary cooling sequence for a typical brown dwarf . l dwarfs are further characterized by the presence of condensates near or just below their photospheres . these condensates include liquid iron ; solid vo ; and aluminum , calcium , magnesium , and titanium - bearing minerals such as enstatite ( @xmath6 ) , grossite ( @xmath7 ) , and perovskite ( @xmath8 ) @xcite . the opacity of the condensate grains and droplets gives l dwarfs fairly red near - infrared colors , with @xmath9 @xcite . conversely , condensates in mid- and late - type t dwarfs appear to lie deep below the photosphere . these objects have neutral near - infrared colors , @xmath10 , governed by strong absorption features of @xmath11 , @xmath4 , and collision - induced @xmath12 , particularly at k - band @xcite . three classes of models have been developed to account for condensates in brown dwarf atmospheres . the dusty model @xcite assumes that condensates remain dispersed in the gas and in chemical equilibrium throughout the atmosphere . the clear model @xcite removes all condensates from the photosphere as they form , presumably through gravitational settling . the cloudy model @xcite incorporates condensate cloud formation based on an assumed sedimentation efficiency , @xmath13 . the dusty and clear models are extreme cases of the cloudy model with very low and very high values of @xmath13 , respectively . in this paper , we present an exploratory model of partial cloud clearing in cool dwarf atmospheres derived from the cloudy and clear models . in @xmath142 we describe our model and compare its derived photometry to the near - infrared color - magnitude diagram of late - type dwarfs . this analysis indicates that a rapid clearing of clouds takes place across the l / t transition . in @xmath143 we present further evidence of cloud clearing in the persistance and strengthening of feh absorption at 9896 in the early- and mid - type t dwarfs . we discuss our results in @xmath144 . figure 1 shows the near - infrared color - magnitude diagram for a sample of m , l , and t dwarfs with known parallaxes . photometry were obtained from the two micron all - sky survey ( * ? ? ? * hereafter 2mass ) and parallaxes from the usno parallax program @xcite or from hipparcos @xcite in the case of companions to nearby stars . as described in @xcite , late - type m and l dwarfs are increasingly redder and fainter toward later spectral types , but the transition to the bluer t dwarfs is marked by a distinct brightening at j - band , with both the t2 dwarf sdss j1254 - 0122 @xcite and the t5 dwarf 2mass 0559 - 1404 @xcite being brighter at j - band than the latest - type l dwarfs . note that neither of these objects is expected to be young and hence less compact ( r @xmath15 1 r@xmath16 ; burrows et al . 1997 ) based on extremely weak or absent h@xmath17 emission ( burgasser et al . 2002a ; kirkpatrick et al . , in preparation ) . late - type t dwarfs are increasingly bluer in @xmath18 color and fainter at j - band . also shown in this figure are color - magnitude tracks for the three atmosphere models described above . the dusty model ( * ? ? ? * dashed line ) reproduces the colors of early - type l dwarfs , but continues to redden to colors far greater than those observed for the mid- and late - type l dwarfs . this is due to the continued influence of dust opacity , which mutes the otherwise strong absorption bands of h@xmath2o , ch@xmath19 , and h@xmath2 . the clear model ( solid line , left ) reproduces the colors of mid- and late - type t dwarfs , but not those of l dwarfs . finally , the cloudy model ( solid line , right ) successfully reproduces the near - infrared color limit for the latest - type l dwarfs , @xmath20-@xmath21 @xcite , and subsequently turns to bluer near - infrared colors . however , the predicted transition occurs over a greater decrease in brightness , @xmath22 mag from @xmath23 to @xmath24 , than is observed . increasing @xmath13 in this model ( i.e. , moving toward a more cleared photosphere ) allows the color transition to occur at brighter magnitudes , but fails to match the colors and magnitudes of the latest l dwarfs . therefore , the condensate cloud model , while adequately reproducing the brightness and color evolution of l dwarfs , can not reproduce the observed rapid transition to the t dwarfs . is it possible to resolve this apparent discrepancy between the cloudy model and the observations ? all three atmosphere models assume horizontal homogeneity and therefore uniform condensate distribution over the visible disk of the brown dwarf . however , clouds in the planets of the solar system , for instance jupiter s upper cloud decks of solid nh@xmath25 and nh@xmath19sh , display discrete structures of spots and bands caused by tropospheric weather patterns . breaks between the clouds on jupiter coincide with bright `` hot - spots '' at 5 @xmath26 @xcite , as light emerges from warmer regions well below the upper cloud decks . since horizontally non - homogeneous clouds are ubiquitous in planetary atmospheres , it is reasonable to expect that similar features could be present in the condensate cloud layers of brown dwarfs as well @xcite . connecting the cloudy and clear tracks in figure 1 are the predicted fluxes for an alternative model we have developed that allows for partial clearing of condensate clouds . these were constructed by first linearly interpolating between the fluxes of the clear model and a modified cloudy model in which the cloud opacity has been removed , weighing by the fraction of cloud coverage . this combination approximates the effect of partial cloud opacity on the underlying atmospheric pressure - temperature profile . we then linearly interpolated the resulting fluxes with those of the standard cloudy model , again weighing by the fractional cloud coverage , to derive the relative brightness contributions of cloudy and cloud - free regions . our simple approach assumes that many clouds are distributed randomly across the observed disk . when compared to the data , we see that the evolution from the red and dusty l dwarfs to the blue and clear t dwarfs can be traced ( heavy line ) down the cloudy track to an effective temperature @xmath27 , across the 1200 k partly cloudy model , and subsequently down the cloud - free track . sdss j1254 - 0122 is a true transition object , lying close to the 1200 k partly cloudy track at roughly 40% cloud coverage . 2mass j0559 - 1404 , however , appears to have a cloud - free photosphere . the t@xmath0 at which the color transition appears to take place is only slightly cooler than that estimated for one of the latest - type l dwarfs , gliese 584c ( @xmath28 ; kirkpatrick et al . 2000 ; dahn et al . the clearing of clouds at nearly constant t@xmath0 explains the apparent brightening of the early - type t dwarfs at 1 @xmath26 , as the emitting layers below the clouds in these objects would be warmer than the layers above the cloud decks in late - type l dwarfs . note that there is some intrinsic scatter among the empirical data , likely due to the influence of gravity , rotation , and metallicity on cloud coverageo and/or ch@xmath19 absorption features , resulting in later - type spectral morphologies and photometric colors for a given t@xmath0 . this phenomenon would not explain the rapid l / t transition , however , again due to the lack of strong h@xmath17 emission in either sdss j1254 - 0122 or 2mass j0559 - 1404 , as well as the very low photospheric magnetic reynolds numbers in these cool , neutral atmospheres @xcite . ] . nonetheless , the apparently sudden color transition from l to t clearly can not be explained by the progressive sinking of a uniform cloud deck with decreasing @xmath29 , but is possible through the disruption of condensate clouds . there is additional evidence that light emitted at @xmath30 by early and mid - type t dwarfs is seen through opening holes in condensate cloud decks . figure 2 plots the optical spectra of a series of late - type l and t dwarfs ( kirkpatrick et al . 1999 , 2000 ; burgasser et al . , in preparation ) , obtained using the low resolution imaging spectrograph @xcite on keck . spectral types listed are from @xcite and @xcite , and are consistent with a monotonic t@xmath0 sequence @xcite . here we focus on the region centered at the @xmath31 feh band . this feature weakens in the later - type l dwarfs , in concert with the observed disappearance of the @xmath32 feh band @xcite . this is consistent with the segregation of liquid iron into condensate clouds , which depletes the atmosphere of iron - bearing gases above the cloud layer . as the clouds settle below the photosphere with decreasing temperature , the amount of detectable feh drops as well however , the @xmath31 feh band unexpectedly reappears in sdss j1254 - 0122 and is quite pronounced in 2mass j0559 - 1404 . the retention of feh absorption from late - type l dwarfs to early- and mid - type t dwarfs is a general trend , as shown in figure 3 . the feh band clearly weakens from l4 to l8 , but persists and even strengthens slightly toward type t5.5 , subsequently weakening in the latest t dwarfs . the plateau in band strength across the l / t transition can not be explained with a cloud deck that gradually sinks below the photosphere with decreasing t@xmath0 . in this case , we would only see the atmosphere above the cloud deck in which feh is depleted . however , if holes are present in the clouds , deeper and warmer regions in which feh is not depleted become observable . these deep layers can be detected at @xmath30 , where @xmath11 , @xmath4 , and @xmath12 absorption is relatively weak . since the reappearance of feh coincides with the turnover in @xmath18 color , the hypothesis of cloud clearing simultaneously explains both phenomena . the disruption of condensate clouds in brown dwarf atmospheres likely begins when the cloud deck forms within the atmospheric convection zone and becomes subject to coherent vertical motions . at higher effective temperatures , the clouds lie primarily above the radiative - convective boundary and would likely remain spatially uniform , perhaps like the stratospheric hazes on titan and the giant planets . evidence for vertical transport has already been seen in cool brown dwarfs , with the detection of co absorption at @xmath33 in the t6.5 dwarf gliese 229b @xcite in excess of its equilibrium abundance @xcite . photometric studies of late - type l and t dwarfs also indicate far greater absorption in the @xmath34 band than predicted by chemical equilibrium models @xcite , likely caused by enhanced co abundances brought up by vertical flows . since the principal cloud decks lie within 0.5 pressure scale heights of one another , the convective motion need not be particularly vigorous to disrupt all layers . note that upwelling brings substantial co into the photosphere because of its stable bond and hence relatively slow conversion to @xmath4 @xcite . the fragility of the feh bond ( 1.63 ev versus 11.09 ev for co ) precludes any similar non - equilibrium enhancement of this species above the cloud layer . the picture of a brown dwarf s atmospheric evolution presented here has a number of observable consequences . first , the t@xmath0 scale between the late - type l and early - type t dwarfs should be very narrow ; this has already been surmised by a number of studies @xcite . a narrow temperature scale also implies fewer early - type t dwarfs than mid- or late - type t dwarfs , given the shorter time a brown dwarf spends in this phase of its thermal evolution . the statistics are currently insufficient to adequately test this prediction . we also expect substantial photometric and/or spectroscopic variability in late - type l and early - type t dwarfs , caused by the formation , evolution , and motion ( rotational or wind - driven ) of cloud holes in their upper atmospheres . such weather phenomena have been previously cited to explain non - periodic variability in late - type m and early - type l dwarfs @xcite . similar variability could be significant at @xmath30 in the l / t transition . filling in the color - magnitude sequence of late - type l and early - type t dwarfs will enable us to further characterize the behavior of condensate clouds in brown dwarf atmospheres . regardless , the observed spectral and photometric properties of brown dwarfs indicate that the planetary - like phenomena of cloud formation and disruption are important in the evolution of free - floating brown dwarfs . we thank a. burrows and i. mclean for valuable discussions , m. putman and k. zahnle for assistance in the preparation of the manuscript , and g. chabrier for an electronic version of his dusty models in the 2mass j and k@xmath35 bands . we also thank our anonymous referee for bringing up the issue of magnetic activity in the l / t transition . ab acknowledges support by nasa through hubble fellowship grant hst - hf-01137.01 awarded by the space telescope science institute , which is operated by the association of universities for research in astronomy , inc . , for nasa , under contract nas 5 - 26555 . mm acknowledges support from nsf grant ast 00 - 86288 and nasa grant nag5 - 9273 . work by kl and ds is supported by nsf grant ast 00 - 86487 and nasa grant nag5 - 4988 respectively . portions of the data presented herein were obtained at the w. m. keck observatory which is operated as a scientific partnership among the california institute of technology , the university of california , and the national aeronautics and space administration . the observatory was made possible by the generous financial support of the w. m. keck foundation . this publication makes use of data from the two micron all sky survey , which is a joint project of the unversity of massachusetts and the infrared processing and analysis center , funded by the national aeronautics and space administration and the national science foundation . alekseev , i. yu . , & gershberg , r. e. 1997 , in stellar magnetic fields , proceedings of the international conference , eds . yu . glagolevskij & i. romanyuk ( moscow : special astrophysical observatory press ) , p. 76 - 83
clouds of metal - bearing condensates play a critical role in shaping the emergent spectral energy distributions of the coolest classes of low - mass stars and brown dwarfs , l and t dwarfs . because condensate clouds in planetary atmospheres show distinct horizontal structure , we have explored a model for partly cloudy atmospheres in brown dwarfs . our model successfully reproduces the colors and magnitudes of both l and t dwarfs for the first time , including the unexpected brightning of the early- and mid - type t dwarfs at j - band , provided that clouds are rapidly removed from the photosphere at t@xmath0 @xmath1 1200 k. the clearing of cloud layers also explains the surprising persistence and strengthening of gaseous feh bands in early- and mid - type t dwarfs . the breakup of cloud layers is likely driven by convection in the troposphere , analogous to phenomena observed on jupiter . our results demonstrate that planetary - like atmospheric dynamics must be considered when examining the evolution of free - floating brown dwarfs . # 1/#2.1em .5ex-.1em /-.15em.25ex # 1to 0pt#1 # 1
about forty years ago , veselago firstly introduced the concept of left - handed material ( lhm ) in which both the permittivity @xmath0 and the permeability @xmath1 are negative @xcite . he predicted that lhm would have unique and potentially interesting properties , such as the negative refraction index , the reversed doppler shift and the backward cerenkov radiation . veselago pointed out that electromagnetic waves incident on a planar interface between a regular material and a lhm will undergo negative refraction . hence a lhm planar slab can act as a lens and focus waves from a point source . lhm did not receive much attention as it only existed in a conceptual form . after the first experimental observation of negative refraction using a metamaterial composed of split ring resonators ( srrs ) @xcite , the study of such materials has received increasing attention over the last few years . while negative refraction is most easily visualized in an isotropic metamaterial @xcite , negative refraction can also be realized in photonic crystals @xcite and anisotropic metamaterials @xcite have also been reported . recently , pendry extended veslago s analysis and further predicted that a lhm slab can amplify evanescent waves and thus behaves like a perfect lens @xcite . he proposed that the amplitudes of evanescent waves from a near - field object could be restored at its image . therefore , the spatial resolution of the superlens can overcome the diffraction limit of conventional imaging systems and reach the subwavelength scale . the great research interests were initiated by the revolutionary concept . more recently , the anisotropic metamaterials have been proved to be good candidates for slab lens application @xcite . although the focusing is imperfect , the substantial field intensity enhancement can readily be observed . in these cases , the anisotropic metamaterials under consideration are characterized by a hyperboloid dispersion relation , and the focusing is restricted to either e- or h - polarized radiation . the recent development in quasiisotropic metamaterial ( qimm ) offers us further opportunities to extend the previous work and further predict that both e- and h - polarized waves can be refocused . the main purpose of the present work is to construct a polarization insensitive lens by a qimm slab . for shallow incident angles the qimm slab will provide some degree of refocusing in the same manner as an isotropic lhm slab . we are particularly interested in exploiting polarization insensitive effect . first , starting from the representation of plane - wave angular spectrum , we derive the propagation of paraxial beams in the qimm slab . our formalism permits us to introduce ideas for beam focusing and phase compensation of paraxial beams by using the qimm slab . next we want to introduce the inverse gouy phase shift and negative rayleigh length when waves propagating in the qimm slab . as an example , we obtain the analytical description for a gaussian beam propagating through a qimm slab . we find that the phase difference caused by the gouy phase shift in vacuum can be compensated by that caused by the inverse gouy phase shift in the qimm slab . if certain matching conditions are satisfied , the intensity and phase distributions at object plane can be completely reconstructed at the image plane . finally , we will discuss what happen when the evanescent wave transmission through the qimm slab . following the standard procedure , we consider a monochromatic electromagnetic field @xmath7 $ ] and @xmath8 $ ] of angular frequency @xmath6 incident from vacuum into the anisotropic metamaterial . the field can be described by maxwell s equations @xcite @xmath9 the previous maxwell s equations can be combined in a straightforward way to obtain the well - known equation for the complex amplitude of the electric field , which reads @xmath10 where @xmath11 is the speed of light in vacuum . ( color online ) ( a ) the qimm with ellipsoid wave - vector surface ; ( b ) the qimm with double - sheeted wave - vector surface.,title="fig:",width=226 ] ( color online ) ( a ) the qimm with ellipsoid wave - vector surface ; ( b ) the qimm with double - sheeted wave - vector surface.,title="fig:",width=226 ] in the principal coordinate system , maxwell s equations yield a scalar wave equation . in free space , the accompanying dispersion relation has the familiar form @xmath12 where @xmath13 is the @xmath14 component of the incident wave vector . we note that the maxwell s equations are symmetrical in electric and magnetic fields . so as far as maxwell s equations is concerned , what we can do for electricity we can also do for magnetism . to achieve the polarization insensitive effect , we will focus our interesting on the anisotropic metamaterial , in which the permittivity and permeability tensor elements satisfy the condition : @xmath15 where @xmath16 is a constant . a careful calculation of the maxwell s equations gives the dispersion relation : @xmath17 where @xmath18 represents the @xmath14 component of transmitted wave - vector . the above equation can be represented by a three - dimensional surface in wave - vector space . this surface is known as the normal surface and consists of two shells @xcite . under the condition of eq . ( [ qc ] ) , we can find e- and h - polarized waves exhibit the same wave - vector surface . thus the anisotropic medium also be regard as qimm @xcite . clearly , we can find the dispersion surface has the following two types : ellipsoid and double - sheeted hyperboloid , as show in fig . [ fig1 ] . we are currently investigating the possibilities for the manufacture of the qimm . in fact , it is now conceivable that a metamaterial can be constructed whose permittivity and permeability values may be designed to vary independently and arbitrarily throughout a metamaterial , taking positive or negative values as desired . hence the permittivity and permeability tensor elements of qimm can be controlled by modulating the length scale of the srrs . the permittivity @xmath19 and permeability @xmath20 can be approximated by the lorentz model . we trust that the qimm can be constructed , because similar technology has been exploited in anisotropic metamaterial @xcite . in addition , photonic crystals might be a good candidate for constructing a qimm . the periodicity in photonic crystals is on the order of the wavelength , so that the distinction between refraction and diffraction is blurred . nevertheless , many novel dispersion relationships can be realized in photonic crystals , including ranges where the frequency disperses negatively with wave vector as required for a negative refraction @xcite . now we want to enquire : whether e- and h - polarized exhibit the same propagation characteristic . to answer the question we first discuss the transmission of wave vector . we choose the @xmath21 axis to be normal to the interface , the @xmath22 and @xmath23 axes locate at in the plane of the interface . the @xmath21-component of the transmitted wave vector can be found by the solution of eq . ( [ d2 ] ) , which yields @xmath24 @xmath25 for e- and h - polarized waves , respectively . here @xmath26 is the wave number in vacuum and @xmath27 . this choice of sign ensures that power propagates away from the boundary to the @xmath28 direction . the frequency contours of isotropic and quasiisotropic media : ( a ) the circle and ellipse represent the frequency contours of vacuum and quasiisotropic media , respectively . both the wave vector and the energy flow exhibit negative refraction . ( b ) the circle and hyperbola denote the frequency contours of vacuum and quasiisotropic media , respectively . the wave vector undergoes a positive refraction , while the energy flow undergoes a negative refraction.,title="fig:",width=302 ] the frequency contours of isotropic and quasiisotropic media : ( a ) the circle and ellipse represent the frequency contours of vacuum and quasiisotropic media , respectively . both the wave vector and the energy flow exhibit negative refraction . ( b ) the circle and hyperbola denote the frequency contours of vacuum and quasiisotropic media , respectively . the wave vector undergoes a positive refraction , while the energy flow undergoes a negative refraction.,title="fig:",width=302 ] by now , we know that e- and h- polarized waves propagate with same wave vector and poynting vector . it is a significantly different property from general anisotropic media . note that there is a bending angle between @xmath40 and @xmath41 , and therefore @xmath42 , @xmath43 and @xmath44 do not form a strictly right - handed or left - handed system in qimm . hence it is also different from isotropic media . for this reason , this kind of special anisotropic media is regarded as quasiisotropic . it should be mentioned that if @xmath45 in eq . ( [ qc ] ) , e- and h - polarized waves will exhibit the same single - sheeted dispersion relation . while the two polarized waves will undergo different amphoteric refraction , the special anisotropic media can not be regarded as quasiisotropic @xcite . now we are in the position to study the negative refraction in the qimm . unlike in isotropic media , the poynting vector in the qimm is neither parallel nor antiparallel to the wave vector , but rather makes either an acute or an obtuse angle with respect to the wave vector . in general , to distinguish the positive and negative refraction in qimm , we must calculate the direction of the poynting vector with respect to the wave vector . positive refraction means @xmath46 , and negative refraction means @xmath47 . from eqs . ( [ se ] ) and ( [ sh ] ) we get @xmath48 the negative refraction phenomenon is one of the most interesting properties of the qimm . we can see that the refracted waves will be determined by @xmath49 for e - polarized incident waves and @xmath50 for h - polarized incident waves . because of the importance of negative refraction in refocusing effect , we are interested in the two types of qimm , which can formed from appropriate combinations of material parameter tensor elements . + type i. in this case all of the @xmath3 and @xmath4 are negative . the frequency contour is an ellipse as shown in fig . [ fig2](a ) . here @xmath51 and @xmath52 , so the refraction angle of wave vector and poynting vector are always negative . + type ii . in this case @xmath53 , @xmath54 and @xmath55 . the frequency contour is a double - sheeted hyperbola as depicted in fig . [ fig2](b ) . here @xmath56 and @xmath47 . it yields that the refraction of poynting vector refraction is always negative even if the wave - vector refraction is positive . as noted above , the poynting vector will exhibit negative refraction in the two types of qimm . the negative refraction is the important effect responsible for the slab lens . hence , the two kind of qimm can be employed to construct a polarized insensitive lens . in this section , we consider the slab lens constructed by the qimm . as depicted in fig . [ fig3 ] , the qimm slab in region @xmath57 is surrounded by vacuum in region @xmath58 and region @xmath59 . a point source is place the object plane @xmath60 . a single ray will pass the interfaces @xmath61 and @xmath62 before it reaches the image plane @xmath63 . let us investigate what happens when the single ray pass through the qimm slab . because of the anisotropy the ray incident for different angle will exhibit different image plane . the ray tracing picture showing the focussing by qimm slab . the qimm slab is surrounded by vacuum in region 1 and region 3 . the solid line and dash - dotted lines are the theoretical objective and focusing planes , respectively.,width=377 ] first , we explore the aberration effect caused by the anisotropic effect . subsequent calculations of eq . ( [ as ] ) give the relationship between @xmath64 and @xmath65 by @xmath66 @xmath67 note that the expressions are sightly different from those in conventional uniaxial crystal @xcite . it is instructive to compare these results with snell s law , which for refraction from vacuum into an isotropic medium , gives @xmath68 , where @xmath69 represents the refractive index of the refracting medium . evidently , in the present case we can find the relationship between @xmath64 and @xmath70 is nonlinear , which is caused by the anisotropic effect @xmath71 or @xmath72 . the nonlinear relationship will result in a significant aberration effect in the image plane , hence the achievable resolution of image is limited . next , we want to discuss another aberration effect caused by frequency dispersion . for a certain incident angle , the ray with different frequency will exhibit different image plane as shown in fig . we assume that both @xmath73 and @xmath74 be approximated by the lorentz model . the other tensor elements components are approximated as constants . the resonate frequency , plasma frequency , and damping constant are identical in all respects to those utilized in ref . @xcite . ignoring the metallic structure , the other tensor elements assume the values of the background material which is dominantly air . obviously the frequency dispersion of @xmath73 and @xmath74 will place some practical limitations on the resolution of image . fortunately the limitations can be reduced , if the ray incident at a small angle . the significant effect has be illustrated in fig . it is interesting to noted that the image distance will slowly vary with frequency in a certain band . for a certain incident angle , the ray with different frequency @xmath75 will exhibit different image plane . note that for smaller incident angles , the qimm slab will provide some degree of frequency insensitive in a certain ghz band.,width=377 ] remarkable as the qimm slab lens is it suffers from some problem : how to cancel the aberration effect ? does the polarization insensitive lens have any use , if the image is not imperfect ? the main effect of anisotropy associated with qimm will limit the resolution of the image . although the image is imperfect , we trust that the reconstruct effect of intensity and phase in paraxial regime will lead to some potential applications . such as the qimm slab can be used to provide phase compensation and beam focusing in cavity resonator . furthermore , sightly anisotropy in qimm slab lens can improve the beam parameter . most importantly , the qimm can be used to design polarization - insensitive modulators and polarization - insensitive all - optical switching in fiber communication system . hence it is very desirable to investigate the polarization insensitive effect in paraxial regime . from a mathematical point of view , the approximate paraxial expression for the field can be obtained by the expansion of the square root of @xmath76 to the first order in @xmath77 @xcite , which yields @xmath78 @xmath79 where we have introduced the boundary condition @xmath80 . from eqs . ( [ pqze ] ) and ( [ pqzh ] ) we can easily find that for shallow incident angles the qimm slab will provide some degree of refocusing in the same manner as an isotropic lhm . hence the aberration effect can be cancelled in paraxial beam region . the interesting property allow us to introduce the idea to construct a qimm slab lens in paraxial beam region . equation ( [ mpe ] ) can be conveniently solved by employing the fourier transformations , so the complex amplitudes in qimm for e- and h - polarized beams can be conveniently expressed as @xmath81.\label{fielde}\ ] ] @xmath82.\label{fieldh}\ ] ] here @xmath83 , @xmath84 , and @xmath85 is the unit vector in the @xmath14-direction . substituting eqs . ( [ pqze ] ) and ( [ pqzh ] ) into eq . ( [ fielde ] ) and ( [ fieldh ] ) , respectively , we obtain @xmath86\tilde{\bf e}({\bf k}_\perp),\label{fielde}\end{aligned}\ ] ] @xmath87\tilde{\bf h}({\bf k}_\perp).\label{fieldh}\end{aligned}\ ] ] the fields @xmath88 and @xmath89 in eqs . ( [ fielde ] ) and ( [ fieldh ] ) are related to the boundary distributions of the fields by means of the relation @xmath90 , \label{ase}\ ] ] @xmath91 , \label{ash}\ ] ] for e- and h - polarized beams , respectively . evidently , eqs . ( [ ase ] ) and ( [ ash ] ) are standard two - dimensional fourier transform @xcite . in fact , after the field distribution in the plane @xmath60 is known , eqs . ( [ fielde ] ) and ( [ fieldh ] ) provide the expression of the e- and h - polarized field in the space @xmath92 , respectively . since our attention will be focused on beam propagating along the @xmath28 direction , we can write the paraxial fields as @xmath93 @xmath94 where the field @xmath95 is the slowly varying envelope amplitude which satisfies the parabolic equation : @xmath96{\bf a}_e({\bf r}_\perp , z)=0,\label{pee}\ ] ] @xmath97{\bf a}_h({\bf r}_\perp , z)=0.\label{peh}\ ] ] under the quasiisotropic condition of eq . ( [ qc ] ) , we can easily find that e- and h - polarized paraxial filed exhibit the same propagating characteristics in paraxial regime . the interesting properties allow us to introduce the idea to construct a polarization lens by qimm slab . for simplify , we introduce the effective refraction indexes : @xmath98 from eqs . ( [ pee ] ) and ( [ peh ] ) we can find that the field of paraxial beams in qimm can be written in the similar way to that in regular material , while the sign of the effective refraction index could be reverse . to simplify the proceeding analyses , we will focus our attention on the qimm with ellipsoid frequency contour . in the previous section we have understood both e- and h - polarized beams have the same propagation characteristic in qimm slab . hence we do not wish to get involved in the trouble to discuss the focusing effect of two polarized waves . instead , we will investigate the analytical description for e - polarized beam with a boundary gaussian distribution . this example allows us to describe the refocusing features of beam propagation in qimm slab . to be uniform throughout the following analysis , we introduce different coordinate transformations @xmath99 in the three regions , respectively . first we want to explore the field in region @xmath58 . without any loss of generality , we assume that the input waist locates at the object plane @xmath60 . the fundamental gaussian spectrum distribution can be written in the form @xmath100,\label{s1}\ ] ] where @xmath101 is the spot size . the rayleigh lengths give by @xmath102 . by substituting eq . ( [ s1 ] ) into eq . ( [ fielde ] ) , the field in the region @xmath58 can be written as @xmath103,\label{g1}\ ] ] @xmath104 here we have chosen different waists , @xmath105 and @xmath106 , in order to deal with a more general situation . because of the isotropy in vacuum , we can easily obtain @xmath107 and @xmath108 . the corresponding rayleigh lengths give by @xmath109 . we are now in a position to investigate the field in region @xmath57 . in fact , the field in the first boundary can be easily obtained from eq . ( [ g1 ] ) by choosing @xmath110 . substituting the field into eq . ( [ ase ] ) , the angular spectrum distribution can be obtained as @xmath111.\label{s2}\ ] ] for simplicity , we assume that the wave propagate through the boundary without reflection . substituting eq . ( [ s2 ] ) into eq . ( [ fielde ] ) , the field in the qimm slab can be written as @xmath112,\label{g2}\ ] ] @xmath113 here @xmath114 and @xmath115 . the interesting point we want to stress is that there are two different rayleigh lengths , @xmath116 and @xmath117 , that characterize the spreading of the beam in the direction of @xmath22 and @xmath23 axes , respectively . a further important point should be noted that we have introduce the negative rayleigh length . the inherent physics underlying the negative rayleigh length is the waves undergo a negative phase velocity in the qimm slab . as can be seen in the following , the negative rayleigh length will give rise to the corresponding reverse gouy phase shift . ( color online ) the numerically computed intensity distribution in object and image planes . ( a ) the intensity distribution for normal gaussian beam in object plane . the in tensity distribution in image plane for gaussian beam propagating through the qimm slab with different anisotropic parameters : ( b ) @xmath118 , @xmath119 . ( c ) @xmath120 , @xmath121 . ( d ) @xmath118 , @xmath121 . we can easily find the intensity distribution at the object plane can be completely reconstructed at the image plane.,title="fig:",width=226 ] ( color online ) the numerically computed intensity distribution in object and image planes . ( a ) the intensity distribution for normal gaussian beam in object plane . the in tensity distribution in image plane for gaussian beam propagating through the qimm slab with different anisotropic parameters : ( b ) @xmath118 , @xmath119 . ( c ) @xmath120 , @xmath121 . ( d ) @xmath118 , @xmath121 . we can easily find the intensity distribution at the object plane can be completely reconstructed at the image plane.,title="fig:",width=226 ] ( color online ) the numerically computed intensity distribution in object and image planes . ( a ) the intensity distribution for normal gaussian beam in object plane . the in tensity distribution in image plane for gaussian beam propagating through the qimm slab with different anisotropic parameters : ( b ) @xmath118 , @xmath119 . ( c ) @xmath120 , @xmath121 . ( d ) @xmath118 , @xmath121 . we can easily find the intensity distribution at the object plane can be completely reconstructed at the image plane.,title="fig:",width=226 ] ( color online ) the numerically computed intensity distribution in object and image planes . ( a ) the intensity distribution for normal gaussian beam in object plane . the in tensity distribution in image plane for gaussian beam propagating through the qimm slab with different anisotropic parameters : ( b ) @xmath118 , @xmath119 . ( c ) @xmath120 , @xmath121 . ( d ) @xmath118 , @xmath121 . we can easily find the intensity distribution at the object plane can be completely reconstructed at the image plane.,title="fig:",width=226 ] finally we want to explore the field in region @xmath59 . the field in the second boundary can be easily obtained from eq . ( [ g2 ] ) under choosing @xmath122 . substituting the field into eq . ( [ ase ] ) , the angular spectrum distribution can be written as @xmath123.\label{s3}\ ] ] substituting eq . ( [ s3 ] ) into eq . ( [ fielde ] ) , the field in the region 3 is given by @xmath124,\label{g3}\ ] ] @xmath125 here @xmath126 and @xmath127 . the corresponding rayleigh lengths give by @xmath128 , that denote the beam exhibit the same diffraction distance in the direction of @xmath22 and @xmath23 axes . the effect of the anisotropic diffraction is that these two beam widths keep their difference even if the rayleigh lengths , @xmath129 and @xmath130 , are equal , implying that generally the gaussian beam is astigmatic . up to now , the fields are determined explicitly in the three regions . comparison of eq . ( [ g2 ] ) , eq . ( [ g3 ] ) with eq . ( [ g1 ] ) shows that the field distributions in region @xmath57 and region @xmath59 may no longer remain gaussian . we take the image position @xmath131 to be the place of the second focusing waist . for the purpose of illustration , the intensity distribution in object plane is plotted in fig . [ fig5](a ) . in general , the shape of intensity distribution is distorted in image plane as shown in fig . [ fig5](b ) and fig . [ fig5](c ) . careful evaluation of eq . ( [ g3 ] ) reveals that the secret underlying the intensity distortion is the anisotropic diffraction . now , the most obvious question is whether the intensity distribution at the object plane can be completely reconstructed at the image plane . in the next step , we want to explore the matching condition of focusing . we can easily obtain the place of the focusing waist by choosing @xmath132 . let us assume the incident beam waist locates at plane @xmath60 . to eliminate the astigmatic effect , the beam waists should locate at the same place , namely @xmath133 . using these criterions , the matching condition for focusing can be written as @xmath134 under the focusing matching condition , the intensity distribution at the object plane can be completely reconstructed at the image plane as shown in fig . [ fig5](d ) . a further point should be noted is that the thickness of the qimm slab should satisfy the relation @xmath135 , otherwise there is neither an internal nor an external focus . in this section , we attempt to investigate the matching condition for phase compensation . in isotropic lhm , plane waves can propagate with negative phase velocity directed opposite to the direction of poynting vector . hence the phase difference can be compensated by the lhm slab @xcite . however , the negative tensor parameters associated with qimm provides a wealth of opportunities for observing and exploiting negative phase - velocity behavior . first let us investigate the phase distribution in region 1 . a more rigorous calculation of eq . ( [ g1 ] ) gives @xmath136 @xmath137 here @xmath138 and @xmath139 are the radius of curvature . because of the isotropy in vacuum , we can easily find @xmath140 . the gouy phase shift in vacuum is given by @xmath141 . next , we attempt to explore the phase distribution in region 2 . matching the boundary condition , the phase term in eq . ( [ g2 ] ) can be written as @xmath142 @xmath143 the gouy phase shift in qimm is given by eq . ( [ gouy2 ] ) . we should mention that there are two different radius of curvature , @xmath144 and @xmath145 , that characterize the beam undergo different diffraction effects in the direction of @xmath22 and @xmath23 axes , respectively . the phase difference caused by the gouy phase shift in vacuum can be compensated by that caused by the inverse gouy phase shift in the qimm slab . the phase fronts of gaussian beam ( solid lines ) differ from those of a perfect spherical wave ( dashed lines).,width=377 ] now , we are in the position to explore the phase distribution in region 3 . analogously , we make some serious calculation of eq . ( [ g3 ] ) , then obtain the phase distribution @xmath146 @xmath147 the radius of curvatures are given by eq . ( [ r3 ] ) and the corresponding gouy phase shift is given by eq . ( [ gouy3 ] ) . the anisotropic effect result in the two radius of curvatures keep their difference even if the rayleigh lengths are equal . it is known that an electromagnetic beam propagating through a focus experiences an additional @xmath148 phase shift with respect to a plane wave . this phase anomaly was discovered by gouy in 1890 and has since been referred to as the gouy phase shift @xcite . it should be mentioned that there exists an effect of accumulated gouy phase shift when a beam passing through an optical system with positive index @xcite . while in the qimm slab system we expect that the phase difference caused by gouy phase shift can be compensated by that caused by the inverse gouy shift in the qimm slab . ( color online ) the numerically computed phase distribution in object and image planes . ( a ) the phase distribution in object plane . the phase distribution in image plane after the gaussian beam propagating through the qimm slab with different anisotropic parameters : ( b ) @xmath118 , @xmath119 . ( c ) @xmath120 , @xmath121 . ( d ) @xmath118 , @xmath121 . the phase distribution can be completely reconstructed at the image plane . the parameters are the same as in fig . [ fig5].,title="fig:",width=226 ] ( color online ) the numerically computed phase distribution in object and image planes . ( a ) the phase distribution in object plane . the phase distribution in image plane after the gaussian beam propagating through the qimm slab with different anisotropic parameters : ( b ) @xmath118 , @xmath119 . ( c ) @xmath120 , @xmath121 . ( d ) @xmath118 , @xmath121 . the phase distribution can be completely reconstructed at the image plane . the parameters are the same as in fig . [ fig5].,title="fig:",width=226 ] ( color online ) the numerically computed phase distribution in object and image planes . ( a ) the phase distribution in object plane . the phase distribution in image plane after the gaussian beam propagating through the qimm slab with different anisotropic parameters : ( b ) @xmath118 , @xmath119 . ( c ) @xmath120 , @xmath121 . ( d ) @xmath118 , @xmath121 . the phase distribution can be completely reconstructed at the image plane . the parameters are the same as in fig . [ fig5].,title="fig:",width=226 ] ( color online ) the numerically computed phase distribution in object and image planes . ( a ) the phase distribution in object plane . the phase distribution in image plane after the gaussian beam propagating through the qimm slab with different anisotropic parameters : ( b ) @xmath118 , @xmath119 . ( c ) @xmath120 , @xmath121 . ( d ) @xmath118 , @xmath121 . the phase distribution can be completely reconstructed at the image plane . the parameters are the same as in fig . [ fig5].,title="fig:",width=226 ] we might suspect whether the phase difference caused by the gouy phase shift in vacuum can be compensated by that caused by the inverse gouy phase shift in qimm slab . to obtain the better physical picture , the schematic distribution of phase fronts are plotted in fig . [ fig6 ] . the phase fronts of a focused gaussian beam are plotted with solid lines , and the phase fronts of a perfect spherical wave are depicted with the dashed lines . the phase difference on the optical axis is caused by the gouy phase shift . the inherent secret underlying the reverse gouy phase shift in the qimm slab is the waves undergo a negative phase velocity . let us investigate what happens if we consider the phase difference caused by the gouy shift . under the focusing matching conditions , the phase difference caused by the gouy phase shift in the three regions are @xmath149 the first and third equations dictate the phase difference caused by the gouy shift in regions @xmath58 and @xmath59 , respectively . the second equation denotes the phase difference caused by the inverse gouy phase shift in the qimm slab . subsequent calculations of eq . ( [ gd ] ) show @xmath150 this implies that the phase difference caused by the gouy phase shift in vacuum can be compensated by the counterpart caused by the inverse gouy phase shift in qimm slab . therefore the condition for phase compensation can be simply written as @xmath151 the first term in eq . ( [ pha ] ) is the phase deference caused by the plane wave in vacuum , and the other term is the phase deference caused by the plane wave in the qimm slab . for the purpose of illustration , the phase distribution in object plane is plotted in fig . [ fig7](a ) . generally , the phase distributions in image plane is distorted as shown in fig . [ fig7](b ) and fig . [ fig7](c ) . as mentioned above , the phase distortion is caused by the effect of anisotropic diffraction . to cancel the phase distortion , the beam waists should locate at the same place , namely @xmath133 . under the phase matching condition , the phase distribution at the object plane can also be completely reconstructed at the image plane as depicted in fig . [ fig7](d ) . now an interesting question naturally arises : whether the matching conditions of focusing and the phase compensation can be satisfied simultaneously . clearly , if we seek a solution satisfying eqs . ( [ foc ] ) and ( [ pha ] ) , the only possibility is @xmath152 under the matching conditions , the intensity and phase distributions at the object plane can be completely reconstructed at the image plane . it should be mentioned that , for the qimm slab with double - sheeted hyperboloid wave - vector surface , both e- and h - polarized beams can also exhibit the same intensity and phase reconstructed effect . because of the positive phase velocity embedded in this type of qimm , the paraxial beam will experience the positive rayleigh distance and gouy phase shift . therefore the accumulated phase delay effect give rise to a large phase deference between the object and image planes . in this section , we discuss under what conditions anomalous transmission will occur when an evanescent wave transmitted through the qimm slab . it is well known that when an evanescent wave is transmitted through a slab of regular media with simultaneously positive permittivity and permeability , the amplitude of the transmitted wave will decay exponentially as the thickness of the slab increases . while an evanescent wave transmitted through an isotropic lhm slab , the amplitude of the transmitted wave would be amplified exponentially . this anomalous transmission of evanescent waves is a very peculiar property of lhm and it may lead to subwavelength image @xcite . now we will explore what happen when the evanescent wave transmission through the qimm slab . for e - polarized incident waves , the incident and reflected fields in region 1 can be written as @xmath153+r_e e_0 { \bf e}_y \exp[i(k_x x - k_z z)],\label{e1}\ ] ] where @xmath154 is the reflection coefficient . some of the incident wave is transmitted into the qimm slab , and conversely a wave inside the qimm slab incident on its interfaces with the surrounding vacuum also experiences transmission and reflection , so the electric field of the wave inside the slab is given by @xmath155+t e_0{\bf e}_y \exp[i ( q_x^{e } x- q_z^{e}z ) ] .\label{e2}\ ] ] here @xmath156 and @xmath157 are coefficients which need to be determined by boundary conditions @xcite . matching the boundary conditions for each wave - vector component at the plane @xmath122 gives the propagation field in the form @xmath158,\label{e3}\ ] ] where @xmath37 is the overall transmission coefficient . the @xmath21-component of the wave vectors of evanescent waves can be found by the solution of eq . ( [ d2 ] ) , which yields @xmath159 for e- and h - polarized waves , respectively . an evanescent wave ( @xmath160 ) interacting with a qimm slab ( a ) @xmath161 ( b ) @xmath162 . the qimm with different values of absorption @xmath163 , ranging from 0.001 to 0.1 . the qimm slab is supposed to be well satisfied the conditions : @xmath164 and @xmath165 . the vertical gray lines denote the two surfaces of the qimm slab.,title="fig:",width=302 ] an evanescent wave ( @xmath160 ) interacting with a qimm slab ( a ) @xmath161 ( b ) @xmath162 . the qimm with different values of absorption @xmath163 , ranging from 0.001 to 0.1 . the qimm slab is supposed to be well satisfied the conditions : @xmath164 and @xmath165 . the vertical gray lines denote the two surfaces of the qimm slab.,title="fig:",width=302 ] by matching the electric and magnetic fields at the two interfaces between the qimm slab and the surrounding vacuum , the coefficients in eqs . ( [ e1 ] ) , ( [ e2 ] ) , and ( [ e3 ] ) can be determined . we can get that the overall transmission through both surfaces of the qimm slab is given by @xmath166 from eq . ( [ te ] ) we can see that in general cases , when an evanescent wave is transmitted through a qimm slab , its amplitude will decay exponentially as the thickness of the slab increases . but if the following conditions are satisfied : @xmath167 @xmath168 the amplitude of the transmitted evanescent wave will be amplified exponentially by the transmission process through the qimm slab . from eq ( [ te ] ) , we can see that if the conditions ( [ eeci ] ) and ( [ eecii ] ) are satisfied , the overall transmission coefficient @xmath37 will be equal to @xmath169 , hence the amplitude of the transmitted evanescent wave will be amplified exponentially as the thickness of the qimm slab increases . note that the conditions mentioned above are sightly different from those as hu and chui have obtained @xcite . as for h - polarized evanescent waves , the overall transmission through both interfaces of the qimm slab can be obtained by similar procedures , and we can get the overall transmission coefficient : @xmath170 from eq . ( [ th ] ) , we can see that in general cases , the amplitude of the transmitted h - polarized evanescent waves will also decay exponentially as the thickness of the qimm slab increases . but if the following conditions are satisfied , the overall transmission coefficient @xmath38 will be equal to @xmath171 , and the amplitude of the transmitted h - polarized evanescent wave will be amplified exponentially by the transmission process through the qimm slab . so for h - polarized evanescent waves , @xmath172 @xmath173 if conditions ( [ ehci ] ) and ( [ ehcii ] ) are satisfied , the qimm slab will enhance exponentially the transmitted waves . comparing ( [ eeci ] ) and ( [ eecii ] ) with ( [ ehci ] ) and ( [ ehcii ] ) , we find that in the presence of qimm , the conditions for e - polarized means automatically the conditions for h - polarized are satisfied . hence the qimm also exhibit the significant insensitive effect for evanescent waves . we stress that , for the qimm slab with double - sheeted hyperboloid dispersion relation , the evanescent waves can not be amplified . in the above analysis the permittivity and permeability tensor elements are assumed to be lossless . however the effect of absorption , necessarily present in such materials , may drastically suppress any evanescent amplifying wave into a decaying one @xcite . comparing fig . [ fig8](a ) with fig . [ fig8](b ) suggests that large @xmath174 is not favored for the amplification of the evanescent wave inside the qimm slab . it is also clearly seen from figs . [ fig8](a ) and [ fig8](b ) that a low reflection occurs at the interfaces , since the qimm slab is supposed to be well satisfied the conditions ( [ eeci ] ) and ( [ eecii ] ) . so far we have shown the simulation results of the evanescent wave interacting with qimm slabs with different absorptions and thicknesses . we find that the suppression of evanescent - wave amplification can be effectively relaxed by reducing the thickness of the qimm slab . the numerical examples provide direct evidence that an evanescent wave could be amplified in a qimm slab with finite absorption . unfortunately , the loss of a realistic metamaterial could not be reduced to a very small level @xcite . therefore , the thickness of the qimm slab should be much smaller than the wavelength in order to realize the subwavelength imaging for such a system . as a result of the amplification of evanescent waves inside the lossy qimm slab , superlensing effect with subwavelength image resolution could be achieved practically . while the main effect of anisotropy associated with qimm will limit the resolution of the image . although the image is imperfect , we trust that the reconstruct effect of intensity and phase in paraxial regime will lead to some potential applications . several recent developments make the polarization insensitive lens a practical possibility . some time ago it was shown that a double - periodic array of pairs of parallel gold nanorods will exhibit negative permittivity and permeability in the optical range @xcite . another extremely promising material has been previously explored in certain designs of photonic crystals , which can be effectively modelled with anisotropic permittivity and permeability tensors @xcite . experimentally , the goal to realize the insensitive lens , chiefly lies in reducing the loss of qimm . practical polarization insensitive lens will require the frequency independent , a great challenge to the designers is to realize the negative material parameters in a wide band . the recent developments lead us to be optimistic that the polarization insensitive lens can be designed in future . in conclusion , we have proposed how to employ the qimm slab to create a polarization insensitive lens , in which both e- and h - polarized waves exhibit the same refocusing effect . for shallow incident angles the qimm slab will provide some degree of refocusing in the same manner as an isotropic negative index material . we have investigated the focusing and phase compensation of paraxial beams by the qimm slab . we have introduced the concepts of inverse gouy phase shift and negative rayleigh length of paraxial beams in qimm . we have shown that the phase difference caused by the gouy phase shift in vacuum can be compensated by that caused by the inverse gouy phase shift in the qimm slab . if certain matching conditions are satisfied , the intensity and phase distributions at object plane can be completely reconstructed at the image plane . the qimm slab exhibits the significant insensitive effect for both transmitted and evanescent waves . our simulation results show that the superlensing effect with subwavelength image resolution could be achieved in the form of a qimm slab . we wish the essential physics described in this paper will provide reference in the road to construct the polarization insensitive lens . we trust that the significant insensitive properties will lead to further novel effects and applications . h. luo are sincerely grateful to professor thomas dumelow for many fruitful discussions . we also wish to thank the anonymous referees for their valuable comments and suggestions . this work was partially supported by projects of the national natural science foundation of china ( nos . 10125521 and 10535010 ) , and the 973 national major state basic research and development of china ( no . g2000077400 ) .
we propose to employ the quasiisotropic metamaterial ( qimm ) slab to construct a polarization insensitive lens , in which both e- and h - polarized waves exhibit the same refocusing effect . for shallow incident angles , the qimm slab will provide some degree of refocusing in the same manner as an isotropic negative index material . the refocusing effect allows us to introduce the ideas of paraxial beam focusing and phase compensation by the qimm slab . on the basis of angular spectrum representation , a formalism describing paraxial beams propagating through a qimm slab is presented . because of the negative phase velocity in the qimm slab , the inverse gouy phase shift and the negative rayleigh length of paraxial gaussian beam are proposed . we find that the phase difference caused by the gouy phase shift in vacuum can be compensated by that caused by the inverse gouy phase shift in the qimm slab . if certain matching conditions are satisfied , the intensity and phase distributions at object plane can be completely reconstructed at image plane . our simulation results show that the superlensing effect with subwavelength image resolution could be achieved in the form of a qimm slab .
in this section , we focus on the tunneling amplitudes at finite @xmath52 . we rewrite eq . ( [ eq : energy ] ) of the main text as p_x(x)^4+p_x(x)^3+p_x(x)^2+p_x(x)+m_0+(1-v(x)^2)=0 , [ eq : sol ] where @xmath114 and @xmath115 ; @xmath116 and @xmath117 , while @xmath118 with @xmath119 ; @xmath120 , @xmath121 , and @xmath122 , where @xmath123 . note that we have written eq . ( 11 ) in dimensionless units . in the barrier regime @xmath124 , the effective wkb action @xmath125 leads to four solutions : s_i = s_i+i s_i , where @xmath126 , and all @xmath40 and @xmath41 are real numbers . of these four solutions , two propagate to the right and two to the left . focusing on the right - moving solutions , we obtain wkb amplitudes as discussed in the main text . for @xmath55 , the decay lengths and the phase factors have identical magnitude for all @xmath127 . thus the tunneling probability oscillates , irrespective of the value of @xmath128 . this is illustrated in fig . ( 5)a for different values of @xmath128 , for @xmath55 . ( 5)b , shows transmission as a function of @xmath128 for fixed @xmath77 . note that for fixed @xmath77 , the transmission also decays as @xmath128 increases . thus large @xmath128 solutions contribute insignificantly to the integrated transmission and hence to the net current . on the other hand , for @xmath129 and @xmath130 , both the decay lengths and the phase factors differ in magnitude . consequently , the oscillatory features of the transmission slowly die off with the increase of both @xmath7 and @xmath128 . this is illustrated in fig . ( 6 ) . for small @xmath128 , we see oscillatory transmission as long as @xmath54 . in contrast , for @xmath131 , we see more conventional decaying transmission with no oscillations . for large @xmath128 , the tunneling amplitudes are almost decaying for both limiting cases of @xmath7 . thus for large @xmath7 and large @xmath128 we obtain conventional transmission , and hence no n - shaped feature in the integrated current . in this section , we provide a detailed discussion of the numerical solutions of tunneling amplitudes presented in the main text . writing eq . ( 1 ) of the main text in rotated coordinate and mapping it into a landau - zener problem , we obtain first - order differential equation in momentum space as if_p_x= , [ eq : diff ] note that we take @xmath132 . following the definition of zener tunneling amplitudes , it is convenient to work in the eigenstate basis of @xmath64 at @xmath133 , which are given by @xmath68 , where @xmath69 is matrix transpose . substituting @xmath134 ( with @xmath135 ) into eq . ( 13 ) and setting @xmath26 for simplicity , we obtain ( in dimensionless units ) i _ p_xa=(4/)(p_x - p_y)^2a - i ( 4/ ) ( p_x+p_y)a+b + i _ p_xb=-(4/)(p_x - p_y)^2b+i ( 4/ ) ( p_x+p_y)a+a.we now solve these coupled differential equations within a suitably chosen range of @xmath136 . we take @xmath137 and @xmath138 as boundary conditions for this problem . with this , the transmission is given by the coefficient @xmath139 evaluated at @xmath140 . for our numerical calculation , we take @xmath141 , with @xmath142 . the inset of fig . ( 2 ) in the main text shows numerical results for @xmath143 . in fig . ( 7 ) , we present additional numerical results for finite @xmath128 and @xmath7 , and compare with the wkb results . although for small @xmath128 and small @xmath7 , we obtain reasonable agreement between wkb and numerical results , we observe deviations for large @xmath128 and/or large @xmath7 . we believe these are owing to instability of the numerical integration of the more complex situation where @xmath63 is the solution to a general quartic equation ( rather than one that has only even powers , as in blg ) . for different values of @xmath128 , where @xmath120 . b ) same as ( a ) as a function of @xmath128 for different values of @xmath77 . we take @xmath55 for both ( a ) and ( b ) . we take parameters appropriate to ( tio@xmath144/(vo@xmath145 . ]
we study transport across p - n junctions of gapped two - dimensional semi - dirac materials : nodal semimetals whose energy bands disperse quadratically and linearly along distinct crystal axes . the resulting electronic properties relevant to materials such as tio@xmath0/vo@xmath0 multilayers and @xmath1-(bedt - ttf)@xmath0i@xmath2 salts continuously interpolate between those of mono- and bi - layer graphene as a function of propagation angle . we demonstrate that tunneling across the junction depends on the orientation of the tunnel barrier relative to the crystalline axes , leading to strongly non - monotonic current - voltage characteristics , including negative differential conductance in some regimes . in multi - valley systems these features provide a natural route to engineering valley - selective transport . _ introduction. _ quantum tunneling is at the heart of modern semiconductor technology : the simplest incarnation of the venerable p - n junction relies on controlling electron tunneling primarily by tuning device parameters such as the width of the depletion region and its characteristic energy barrier . richer behavior more sensitive to microscopic material properties is possible in the presence of multiple bands , where applied fields can induce interband landau - zener transitions @xcite , reflected in highly nonlinear current - voltage ( @xmath3 ) characteristics . quantum interference , on the other hand , has historically been of less relevance to semiconductor devices , and was largely ignored in this context until the recent revolution in two - dimensional materials : for instance , interference phenomena stemming from the dirac nature of electronic quasiparticles in graphene have been successfully probed in transport experiments . recent work @xcite involving one of the present authors examined the interplay of interference and tunneling in bilayer graphene ( blg ) , whose pristine electronic dispersion consists of a pair of bands touching quadratically at isolated nodes in momentum space . as a consequence of this unusual electronic structure , when electrons in gapped blg tunnel across a barrier between p- and n - type regions , they may be transmitted with multiple possible phase shifts . this leads to an oscillatory dependence of the tunneling amplitude on the applied bias , and consequently a non - monotonic current - voltage response , with negative differential conductance in specific bias regimes . while such ` common - path ' tunneling interference occurs in any semiconductor @xcite , it is subleading to conventional tunneling except near quadratic band crossings hence its starring role in blg , versus its relative unimportance in monolayer graphene@xcite and conventional semiconductors . this dichotomy is particularly relevant to a new class of 2d semimetals ( sometimes termed zero - gap semiconductors ) with ` semi - dirac ' dispersion , characterized by electronic bands meeting in a discrete set of nodes about which the bands split linearly or quadratically along independent high - symmetry directions @xcite . such accidental degeneracies can be stabilized by symmetry , and semi - dirac dispersion has been argued to emerge in tio@xmath0/vo@xmath0 heterostructures @xcite and @xmath1-(bedt - ttf)@xmath0i@xmath2 organic salts under pressure @xcite . there are also proposals to engineer similar behavior in optical lattices or photonic metamaterials @xcite . owing to the anisotropic dispersion , semi - dirac systems have quasiparticle properties that continuously interpolate between those of mono- and bi - layer graphene as a function of orientation . it is therefore natural to ask how this can be leveraged when building quantum devices from them a question that has received surprisingly little attention to date . we study the behavior of a p - n junction in a gapped semi - dirac material , focusing on the interband landau - zener problem in the presence of an applied bias . as we demonstrate below , the anisotropy of a single semi - dirac node is imprinted on the tunneling behavior , via its dependence on the orientation of the barrier potential . when electrons propagating along the quadratic axis are incident on a barrier , they experience common - path interference and hence the tunneling amplitude oscillates with bias . the interference is absent when the barrier is oriented so that incident electrons propagate along the linear axis , with a critical angle separating the two regimes . as a consequence , the differential conductance of a p - n junction is orientation - dependent . appropriately designed p - n junctions may therefore facilitate valley - selective transport in materials with multiple valleys with distinctly oriented semi - dirac dispersions , potentially paving the way to several new applications . ) is defined in the @xmath4 plane with quadratic ( linear ) dispersion along the major ( minor ) axis of the ellipse . the barrier @xmath5 is defined in the @xmath6 plane oriented at an angle @xmath7 to @xmath8 . ( b ) two perpendicularly oriented nodes at the edges of the first bz ( shaded region ) of a semi - dirac material . filled ellipses denote the nodes in the extended - zone scheme . ] _ model and wkb solutions. _ we begin with the low - energy hamiltonian for a single gapped semi - dirac node , h= p_^2_1+v p__2+_0_3 , [ eq : ham0 ] where we work in @xmath9 units , @xmath10s are pauli matrices , @xmath11 are the momenta in the @xmath4 plane , @xmath12 , @xmath13 are the effective mass along and the dirac velocity along @xmath8 and @xmath14 , and @xmath15 is the energy gap parameter . absent a potential , the solutions of ( [ eq : ham0 ] ) correspond to quasiparticles that disperse quadratically and linearly along @xmath8 and @xmath14 respectively . in order to extract the dependence of landau - zener tunneling across a p - n junction on its orientation , we model the junction potential as a uniform field @xmath16 , in a rotated coordinate system @xmath17 where where @xmath18 is a rotation about the @xmath19-axis . as a first pass at the problem , we attempt a solution within the wentzel - kramers - brillouin ( wkb ) approximation . to that end , we assume @xmath5 is slowly varying and choose a semiclassical _ ansatz _ for the wavefunction , _ ( ) ~ ( c _ 1 + _ 2 ) e^ip_y ywhere @xmath20 is a slowly varying function of @xmath21 and we have used translational invariance in @xmath22 to fix the wavefunction to be a plane wave parallel to the barrier . using this in ( [ eq : ham0 ] ) we find that for a solution at energy @xmath23 , @xmath24 satisfies ( e - v(x))^2 & = & ( 2m)^-2 ( p_x(x)-p_y)^4 + & & + v^2 ( p_x(x)+p_y)^2+_0 ^ 2 . [ eq : energy ] in order to illustrate the angular dependence , let us choose normal incidence ( @xmath25 ) for simplicity and work at threshold ( @xmath26 ) . then , examining ( [ eq : energy ] ) , we find that the effective wkb action @xmath27 in the barrier region defined by @xmath28 varies with the orientation of the barrier potential . specifically , the four solutions to ( [ eq : energy ] ) lead to wkb actions of the form @xmath29 , where the signs are chosen independently and = _ -1 ^ 1 d ^1/2 , [ mon1 ] where @xmath30 , @xmath31 and @xmath32 are dimensionless , @xmath33 is an intrinsic length scale of the material , and @xmath34 is the characteristic energy scale of the junction . @xmath35 is real and positive for all @xmath7 , but @xmath36 sensitively depends on the orientation of the barrier . above a bias - independent critical angle @xmath37 , @xmath36 becomes purely imaginary , so that all four wkb solutions correspond to monotonically decaying exponentials , as in conventional p - n junctions . in contrast , for @xmath38 , @xmath36 has a non - zero real part , @xmath39 , where @xmath40 and @xmath41 are real and positive . the wkb solutions then take the form @xmath42 $ ] where @xmath43 ; since we are interested in physical solutions that decay into the barrier , we focus on those with @xmath44 , leading to a wkb wavefunction in the barrier region of the form = a e^- ( s_0 + s_1)+is_1+a^ e^-(s_0-s_1)-is_1 , where @xmath45 is a complex number , and we have used the fact that @xmath46 . .[tab : params]microscopic parameters for semi - dirac materials , with representative gaps and corresponding critical angles . @xmath47 denotes the free electron mass@xcite . [ cols="^,^,^,^,^,^",options="header " , ] the wkb phase shift @xmath48 may be determined analytically by examining classical turning points ; we instead obtain it numerically ( see below ) , and find @xmath49 . the total transmission is given by t ( ) = ||^2=2 |a|^2e^-2 s_0 . [ tran ] since @xmath35 and @xmath36 are monotonic functions of bandgap and field strength , the transmission through the p - n junction oscillates as these are varied ( fig . [ fig : transmission](a ) ) . the oscillations resemble those in blg @xcite , but differ in that the nodes of transmission are now angle - dependent , occurring when @xmath50 with @xmath51 an odd integer . so far , we have focused on @xmath25 ; we now comment on wkb solutions for finite @xmath52 . for @xmath53 , we obtain four complex solutions for all values of @xmath7 . as before , two of these solutions correspond to waves decaying into the barrier ; focusing on these , we obtain wkb amplitudes with different decay lengths and different phase factors , in contrast to the @xmath25 case . for small @xmath52 and @xmath54 , these differences in decay lengths and phases are small , and we still see oscillations in the transmission @xcite . however for larger angles and transverse momenta , the differences are sufficiently appreciable that the interference is much less effective due to the mismatch in amplitudes . a knowledge of the full dependence on @xmath52 is essential in order to compute @xmath3 characteristics of the junction ; we determine this numerically @xcite . table [ tab : params ] provides estimates of parameters for some semi - dirac systems and the corresponding critical angles . for different barrier angles @xmath7 . ( inset ) comparison between numerical and wkb results at @xmath55 ; a fit yields wkb phase shift @xmath56 . ( b ) transmission for semi - dirac systems with a single node(dashed line ) and with two perpendicular nodes ( solid line ) , as a function of barrier angle @xmath7 . at a critical angle @xmath57 , common - path interference gives way to conventional tunneling , for a single node , reflected by a dip in transmission . for two nodes , common - path interference is absent for @xmath58 . we take @xmath59 , achievable with @xmath60 ev in ( tio@xmath0)@xmath61/(vo@xmath0)@xmath2 . ] _ mapping to landau - zener problem. _ an alternative approach ( restricted to the uniform - field model ) uses the fact that ( [ eq : ham0 ] ) can be mapped into a landau - zener problem by working in momentum space , where @xmath62 . we then find the first - order equation ifp_x=(h(p_x , p_y)-e ) , [ eq : schr ] that describes the unitary evolution of a particle with @xmath63 playing the role of `` time '' . the zener tunneling probability is the square of the amplitude for an eigenstate of @xmath64 at @xmath65 with negative energy to evolve to one at @xmath66 with positive energy . it is therefore convenient to work in the basis of eigenstates of ( [ eq : ham0 ] ) for @xmath67 , given by @xmath68 where @xmath69 is the matrix transpose . substituting @xmath70 ( with @xmath71 ) into ( [ eq : schr ] ) , we find a set of coupled differential equations for the complex amplitudes @xmath72 . integrating these equations with @xmath73 chosen much bigger than @xmath74 , the momentum scale characterizing the region of landau - zener transitions . ] with the initial condition @xmath75 , we find @xmath76 and hence the transmission probability . fig . [ fig : transmission](a , inset ) shows a representative comparison between wkb and numerical results for a small range of @xmath77 , and for @xmath78 , indicating good agreement . by fitting the parameters of the wkb and numerical results , we find @xmath79 . the numerical results are unstable for large @xmath77 and for some higher @xmath52 values , but as their primary utility is in computing the wkb phase shift , this does not pose a serious problem . in the remainder , we use wkb results with @xmath49 unless otherwise stated . _ i - v characteristics. _ within the landauer formalism , the contribution of a single node to the net tunneling current across a junction of lateral width @xmath80 and orientation @xmath7 with applied voltage @xmath81 is given by i&=&_-^ dt_(,f ) , + t_(,f)&= & _ -^ t(p_y,,f ) dp_y , [ eq : gencur ] where @xmath82 is the fermi - dirac distribution at temperature @xmath69 , and @xmath83 counts the spin degeneracy . we have used the fact that the transmission is dominated by small-@xmath52 values to ignore the dependence of the fermi occupation factors on momentum . ignoring the energy - dependence of @xmath69 ( valid in the uniform - field model ) and writing the effective field in the barrier region as @xmath84 , where @xmath85 is the ` built in ' gate - induced electrostatic potential difference across the junction and @xmath86 is the size of the depletion region , the tunneling current at low temperatures is i v _ . through the p - n junction with energy gap @xmath87 ( shown schematically , bottom inset ) as a function of bias voltage @xmath88 for different values of @xmath7 , clearly showing n - shaped features and negative differential conductivity . we take @xmath89 appropriate to ( tio@xmath0)@xmath61/(vo@xmath0)@xmath2 , @xmath90 ev and @xmath91 . units are @xmath92 and @xmath93 . ( top inset ) zoomed - out view of @xmath94-@xmath95 curve . ] it is convenient to define @xmath92 , @xmath93 , as natural units for the current and bias voltage . fig . [ fig : tunnelcur ] shows the results of numerical integration of ( [ eq : tunncur1 ] ) for a reasonable value of the built - in potential , @xmath96 . note the n - shaped features in the tunnel current , with negative differential conductivity similar to an esaki diode @xcite ( though with a different origin ) , as well as zener - like threshold behavior where the current rises sharply above a breakdown voltage . these features were previously predicted for blg p - n junctions with the crucial difference here being the sensitive angular dependence : the @xmath97-shaped features and negative @xmath98 are absent when the junction angle exceeds the critical @xmath57 for common - path interference . this angular dependence is particularly salient to devices built from multi - valley semi - dirac materials , to which we now turn . , for fixed bias @xmath99 and parameters are appropriate to ( tio@xmath0)@xmath61/(vo@xmath0)@xmath2 [ table [ tab : params ] ] . for @xmath38 ( @xmath100 ) the current is dominated by common - path interference contributions from node @xmath101 ( node @xmath102 ) ; these oscillate with applied bias . for @xmath103 , the tunneling is solely due to conventional tunneling and does not oscillate with bias . ( b ) tunneling current contributions of nodes @xmath104 as a function of bias , for a barrier whose normal is parallel to the quadratic axis of node @xmath101 . as only the node @xmath101 current ( black curve ) oscillates with bias , the current is valley - filtered ( i.e. , dominated by node @xmath102 current , red curve ) when tunneling electrons in node @xmath101 experience destructive common - path interference . the small residual node @xmath101 current at the minima is due to electrons at non - normal incidence . as in fig . [ fig : tunnelcur ] , we choose a slightly larger gap @xmath90ev for illustrative reasons . ] _ multiple valleys. _ in order to facilitate a more direct comparison with experimental systems , and to study valley - selective transport it is useful to discuss situations with multiple semi - dirac nodes ( ` valleys ' ) in the brillouin zone ( bz ) . as minimal model consider a pair of valleys , denoted @xmath104 with quadratic axes rotated by 90@xmath105 relative to each other ; the generalization to multiple such pairs , or to different relative orientations , is straightforward . we take the tunnel barrier to be oriented at an arbitrary angle @xmath7 with respect to the quadratic axis of valley a , and hence at @xmath106 with respect to that of valley @xmath102 . the transmission through the junction is then given by summing contributions from the two valleys , viz . : [ eq : tottran ] t _ ( ) = t()+ t(2 - ) . recall that the dominant contribution to common - path interference is from electrons incident normally on the barrier ( i.e. , with @xmath107 ) . in this limit , the presence ( absence ) of this effect is sensitive to whether @xmath108 ( @xmath109 ) : for @xmath110 , ( [ eq : tottran ] ) always includes a common - path interference contribution from at least one node for any @xmath7 , whereas if @xmath111 common - path interference is absent for @xmath112 . as a consequence , in many cases the tunnel current is strongly suppressed in this window ( fig . [ fig : multivalley]a ) . since @xmath57 depends sensitively on material parameters , it may be possible to achieve ` switching ' behavior of the tunnel current e.g. by tuning @xmath15 . furthermore , if we choose the barrier angles @xmath113 , then only one of the nodes experiences common - path interference and hence has a complex @xmath94-@xmath95 characteristic ; therefore , for an appropriate choice of bias the current through the junction is valley - filtered , i.e. dominated by one of the nodes ( fig . [ fig : multivalley]b ) . this quite remarkable property of the semi - dirac system enables valley - selective functionality with a relatively conventional device design . we have focused here on the two - valley example ; in the case of tio@xmath0/vo@xmath0 multilayers , there are actually four nodes , one pair of which is oriented at 90@xmath105 to the other pair . in such situations , the valley - selective behavior of the junction is restricted to producing a current dominated by electrons from one pair of nodes or the other . _ concluding remarks. _ we have studied the tunneling across barriers between p- and n - doped regions of semi - dirac materials , and examined the impact of their anisotropic dispersion on transport through the junction . we find , first , that the role of common - path interference in the landau - zener tunneling problem is sensitively dependent on the angle of the tunneling barrier relative to the quadratic axis of the material : for instance , electrons incident normally on the barrier do not experience interference once the barrier angle exceeds a critical value @xmath57 . this is in turn reflected in the behavior of the @xmath3 characteristic , and specifically in the presence or absence of bias regimes with negative differential conductance . our analysis was twofold : first , we used a wkb treatment of tunneling under the barrier that can be readily generalized to more complicated barrier potential profiles ; second , we performed a landau - zener analysis , restricted to the case of a uniform barrier field , that served as a check on the wkb results . we also generalized these ideas to the multi - valley setting , where we propose that the angular dependence of the landau - zener tunneling may be leveraged to build a `` valley valve '' , as we find bias regimes where the current flowing across the junction is valley - filtered . although the basic ingredients of the devices proposed here are quite conventional , the semi - dirac system remains relatively unexplored from a device engineering perspective , so there may be unforeseen difficulties in building good p - n junctions and controlling their properties . however , we note that our analysis is quite general and applies to any material with semi - dirac dispersion ; individual materials may have widely differing critical angles and bias regimes for the effects considered here , making them more or less suitable for specific applications . farther afield , we note that there are also proposals to engineer semi - dirac dispersion in photonic crystals ; in such situations , our analysis suggests a promising new route to controlling the flow of light through photonic devices . _ akcnowledgements. _ we thank w. pickett and i. krivorotov for useful discussions . we acknowledge support from nsf grant dmr-1455366 and a uc president s research catalyst award ca-15 - 327861 ( sap , ks ) . 99 c. zener . proc . r. soc . london a * 137 * , 696 ( 1932 ) ; l. d. landau , phys . z. * 2 * , 46 ( 1932 ) . r. nandkishore and l. levitov , pnas * 108 * , 14021 , ( 2011 ) . s. johri , r nandkishore , rn . bhatt and e. j. mele , , 235413 ( 2013 ) . m. lasia , e. prada , and l. brey , , 245320 ( 2012 ) . p. delplace and g. montambaux , , 035438 ( 2010 ) . s. banerjee , r. r. p. singh , v. pardo , and w. e. pickett , 016402 ( 2009 ) . s. banerjee and w. e. pickett , , 075124 ( 2012 ) . f. zhai and j. wang , j. appl . phys . * 116 * 063704 ( 2014 ) v. pardo and w. e. pickett , , 166803 ( 2009 ) . s. katayama , a. kobayashi , and y. suzumura , j. phys . soc . jpn . * 75 * , 054705 ( 2006 ) . p. dietl , f. piechon , and g. montambaux , , 236405 ( 2008 ) . y. wu , opt . express * 22 * , 1906 ( 2014 ) . m. g. kaplunove , b. yagubsklii , p. rosenberg , and yu . g. b. borodko , phys . . sol . * 89 * , 609 ( 1985 ) . l. esaki , r. tsu , ibm j. res . dev . , * 4 * , 61 ( 1970 ) . supplementary material .
since its discovery the josephson effect @xcite has been studied for a variety of superconducting weak links @xcite . the research has recently entered a new phase with the experimental realization of quantum dot weak links exploiting electronic properties of finite - length carbon nanotubes coupled to superconducting leads @xcite . in particular , for the first time since its theoretical prediction @xcite resonant josephson tunneling through discrete electronic states has been observed in carbon nanotube quantum dots @xcite . as demonstrated in refs . , the novel type of weak links exhibits transistor - like functionalities , e.g. a periodic modulation of the critical current with a gate voltage tuning successive energy levels in the dot on- and off - resonance with the fermi energy in the leads . this property has already been implemented in a recently proposed carbon nanotube superconducting quantum interference device ( cnt - squid ) @xcite with possible applications in the field of molecular magnetism . motivated by the experiments on resonant josephson tunneling , in this paper we investigate theoretically how robust it is with respect to pair - breaking perturbations in the superconducting leads . cooper pair breaking can be induced by a number of factors , e.g. by paramagnetic impurities @xcite , an external magnetic field @xcite or by structural inhomogeneities producing spatial fluctuations of the superconducting coupling constant @xcite . it can cause a drastic distortion of the bardeen - cooper - schrieffer ( bcs ) superconducting state , which manifests itself in the smearing of the bsc density of states leading to gapless superconductivity @xcite . while the pair - breaking effect on bulk superconductivity is now well understood , its implications for quantum superconducting transport have been studied to a much lesser extent [ see , e.g. refs . ] which to our knowledge does not cover josephson tunneling through quantum dots . on the other hand , in low - dimensional systems pair - breaking effects may be observable in a common experimental situation when , for instance , a carbon nanotube weak link is subject to a magnetic field . since the orbital field effect in the quasi - one - dimensional channel is strongly suppressed , pair breaking in the superconducting leads can be the main source of the magnetic field dependence of the josephson current . this situation is addressed in our work . = 0.5 the influence of pair breaking on the josephson current can not , in general , be accounted for by mere suppression of the order parameter in the superconducting leads . as was pointed out in ref . , it is a more subtle effect involving the modification of the spectrum of current carrying states in the junction , in particular , the subgap states usually referred to as andreev bound states ( abs ) @xcite . we illustrate this idea for quantum dot junctions in the simple model of a short superconducting constriction with a scattering region containing a single breit - wigner resonance near the fermi energy . the josephson current is calculated using the normal - state scattering matrix of the system and the andreev reflection matrix @xcite . unlike refs . we focus on dirty superconductors for which the andreev matrix can be quite generally expressed in terms of the quasiclassical green functions @xcite , allowing us to treat pair breaking in the superconducting leads nonperturbatively . although we account for all energies ( below and above the abrikosov - gorkov gap @xmath2 ) , it turns out that the behavior of the josephson current can be well understood in terms of a pair - breaking - induced modification of the abs , which depends sensitively on the relation between the breit - wigner resonance width @xmath3 and the superconducting pairing energy @xmath4 . both the critical supercurrent and the josephson current - phase relation are analyzed under experimentally realizable conditions . we consider a junction between two superconductors @xmath5 and @xmath6 adiabatically narrowing into quasi - one - dimensional ballistic wires @xmath7 and @xmath8 coupled to a normal conductor @xmath9 [ fig . [ cons ] ] . the transformation from the superconducting electron spectrum to the normal - metal one is assumed to take place at the boundaries @xmath10 and @xmath11 , implying the pairing potential of the form @xcite : @xmath12 for @xmath13 , @xmath14 for @xmath15 and @xmath16 for @xmath17 with the order parameter phase difference @xmath18 and the junction length @xmath19 ( @xmath20 is the fermi velocity in @xmath21 ) . the josephson coupling can be interpreted in terms of the andreev process @xcite whereby an electron is retro - reflected as a fermi - sea hole from one of the superconductors with the subsequent hole - to - electron conversion in the other one . such an andreev reflection circle facilitates a cooper pair transfer between @xmath5 and @xmath6 . normal backscattering from disordered superconducting bulk into a single - channel junction is suppressed due to the smallness of the junction width compared to the elastic mean free path @xmath22 . the @xmath9 region in the middle of the junction is thus supposed to be the only source of normal scattering . in such type of weak links the josephson current is conveniently described by the scattering matrix expression of refs . that can be written at finite temperature @xmath23 as the following sum over the matsubara frequencies @xmath24 [ ref . ] : @xmath25_{e = i\omega_n . } \label{i}\end{aligned}\ ] ] here @xmath26 is a @xmath27 unitary matrix relating the incident electron and hole waves on the @xmath9 region to the outgoing ones [ fig . [ cons ] ] . it is diagonal in the electron - hole space : @xmath28 , \ , s_{ee}(e)= \left [ \begin{array}{cc } r_{11}(e ) & t_{12}(e)\\ t_{21}(e ) & r_{22}(e ) \end{array } \right ] . \nonumber\end{aligned}\ ] ] the matrix @xmath29 describes electron scattering in terms of the reflection and transmission amplitudes , @xmath30 and @xmath31 , for a transition from @xmath32 to @xmath33 ( @xmath34 ) . the hole scattering matrix is related to the electron one by @xmath35 . the andreev scattering matrix @xmath36 is off - diagonal in the electron - hole space : @xmath37 , \label{a}\end{aligned}\ ] ] where the @xmath38 matrices @xmath39 and @xmath40 govern the electron - to - hole and hole - to - electron scattering off the superconductors . equation ( [ i ] ) is valid for all energies as long as normal scattering from the superconductors is absent @xcite . in ref . the andreev matrix ( [ a ] ) was obtained by matching the solutions of the bogolubov - de gennes equations in the wires @xmath41 to the corresponding solutions in impurity - free leads . gorkov s green function formalism in combination with the quasiclassical theory @xcite allows one to generalize the results of ref . to dirty leads with a short mean free path @xmath42 . in the latter case the matrices @xmath39 and @xmath40 can be expressed in terms of the quasiclassical green functions of the superconductors as follows @xcite : @xmath43 , \ , s_{he}= \left [ \begin{array}{cc } \frac{-f^\dagger_1(e)}{g_1(e)+1 } & 0\\ 0 & \frac{-f^\dagger_2(e)}{g_2(e)+1 } \end{array } \right].\end{aligned}\ ] ] here @xmath44 and @xmath45 ( @xmath46 ) are , respectively , the normal and anomalous retarded green functions in @xmath21 . these matrices are diagonal in the electrode space due to a local character of andreev reflection in our geometry . neglecting the influence of the narrow weak link on the bulk superconductivity , we can use the green functions of the uncoupled superconductors @xmath21 described by the position - independent usadel equation @xcite , @xmath47=0 , \label{usadel}\end{aligned}\ ] ] with the normalization condition @xmath48 for the matrix green function @xmath49 , \,\ , \hat\delta_j=\left [ \begin{array}{cc } 0 & \delta { \rm e}^{i\varphi_j}\\ -\delta { \rm e}^{-i\varphi_j } & 0 \end{array } \right ] , \,\ , \end{aligned}\ ] ] here @xmath50 and @xmath51 are the unity and pauli matrices , respectively , and @xmath52 $ ] denotes a commutator . equation ( [ usadel ] ) accounts for a finite pair - breaking rate @xmath53 whose microscopic expression depends on the nature of the pair - breaking mechanism . for instance , for thin superconducting films in a parallel magnetic field , @xmath54 [ ref . ] where @xmath55 is the film thickness and @xmath56 is the flux quantum . for paramagnetic impurities , @xmath57 coincides with the spin - flip time @xcite . in the case of the spatial fluctuations of the superconducting coupling , @xmath53 is proportional to the variance of the fluctuations @xcite . from eq . ( [ usadel ] ) one obtains the green functions @xmath58 where , following refs . , we introduce a dimensionless pair - breaking parameter @xmath59 . the matrices @xmath60 and @xmath61 can be expressed using eqs . ( [ green ] ) as follows : @xmath62 , \,\ , s_{he}=\alpha \left [ \begin{array}{cc } { \rm e}^{-i\varphi_1 } & 0\\ 0 & { \rm e}^{-i\varphi_2 } \end{array } \right ] , & \label{sa}\\ & \alpha = u-\sqrt{u^2 - 1}. & \label{a}\end{aligned}\ ] ] we note that pair breaking modifies the energy dependence of the andreev reflection amplitude @xmath63 according to the non - bcs green functions ( [ green ] ) and ( [ u ] ) . a few words concerning the applicability of this result are due here . first of all , there is no restriction on energy @xmath64 , e.g. for @xmath65 , equations ( [ sa ] ) and ( [ a ] ) are valid both below and above the reduced ( abrikosov - gorkov ) quasiparticle gap @xmath66 . in particular , for @xmath67 one can show that @xmath68 is real and @xmath69 [ ref . ] , corresponding to perfect andreev reflection with @xmath70 . since in the usadel limit @xmath71 , normal scattering from the superconductors is suppressed due to the smallness of the junction width also in the presence of pair breaking . the absence of normal transmission at @xmath67 is consistent with the abrikosov - gorkov approach assuming no impurity states inside the gap and the validity of the born approximation @xcite . for @xmath72 the relevant solution of eq . ( [ u ] ) is complex and has positive @xmath73 related to the density of states of the superconductor @xcite . equations ( [ u ] ) and ( [ a ] ) are thus the generalization of the known result @xmath74 [ ref . ] for transparent point contact , where @xmath75 is the bcs gap . it is convenient to measure all energies in units of @xmath76 for which equations ( [ green])([a ] ) should be complemented with the self - consistency equation for @xmath4 . at @xmath77 , the case we are eventually interested in , this equation can be written as @xcite : @xmath78 with @xmath59 being now a function of a new pair - breaking parameter @xmath79 ranging from zero to the critical value @xmath80 at which @xmath81 and @xmath82 @xcite . inserting eqs . ( [ a ] ) and ( [ sa ] ) for @xmath36 into eq . ( [ i ] ) and taking the limit @xmath83 we obtain the josephson current for an arbitrary @xmath26 as @xmath84 \right\}_{e = i\omega . } \nonumber\end{aligned}\ ] ] = 0.8 let us assume that the @xmath9 region is a small quantum dot and electrons can only tunnel via one of its levels characterized by its position @xmath85 with respect to the fermi level and broadening @xmath3 . for the simplest breit - wigner scattering matrix with @xmath86 and @xmath87 , equation ( [ i0 ] ) reads @xmath88 where @xmath89 is the breit - wigner transmission probability at the fermi level . the parameter @xmath90 accounts for the energy dependence of the resonant superconducting tunneling . in eq . ( [ i1 ] ) the integrand has , in general , poles given by the equation @xmath91 along with eq . ( [ u ] ) they determine the energies of the andreev bound states ( abs ) localized below the abrikosov - gorkov gap @xmath66 . it is instructive to understand how the pair breaking modifies the abs spectrum since this is reflected on both the current - phase relation @xmath1 and the critical current @xmath92 . we start our analysis with an analytically accessible case of an infinitely broad resonant level , @xmath93 , where eq . ( [ eq ] ) reduces to @xmath94 , yielding the abs energies @xmath95 [ see , eq . ( [ u ] ) ] : @xmath96 . \label{broad}\end{aligned}\ ] ] requiring @xmath97 we find that the abs exist in the phase interval where @xmath98 and only if @xmath99 . the numerical solution of eqs . ( [ u ] ) , ( [ op1 ] ) and ( [ eq ] ) confirms that the interval of the existence of abs gradually shrinks from @xmath100 to a narrower one with increasing pair breaking [ see , fig . [ e15 ] ] . outside this interval the josephson current is carried by the continuum states ( @xmath101 ) alone , which is automatically accounted for by eq . ( [ i1 ] ) . an equation of the same form as eq . ( [ broad ] ) was derived earlier for a nonresonant system and by a different method @xcite . = 0.8 by contrast , the abs spectrum for a narrow resonant level turns out to be much less sensitive to pair breaking . indeed , under condition @xmath102 equations ( [ u ] ) and ( [ eq ] ) reproduce the known result , @xmath103 [ refs . ] . in particular , for @xmath104 the abs exist within the resonance width , @xmath105 and are separated from the continuum by a gap @xmath106 . solving eqs . ( [ u ] ) , ( [ op1 ] ) and ( [ eq ] ) numerically , we find that until this gap closes at a certain value of @xmath107 , the abs spectrum remains virtually intact [ see , fig . [ e03 ] ] . for bigger @xmath107 , the spectrum gets modified in a way similar to the previous case [ cf . , third panels in figs . [ e15 ] and [ e03 ] ] . in the case of a very narrow resonance , the characteristic value of @xmath107 is @xmath108 , corresponding to @xmath109 , i.e. to the onset of gapless superconductivity @xcite . for numerical evaluation of the josephson current ( [ i1 ] ) we first put @xmath110 in eq . ( [ u ] ) and then make the transformation @xmath111 , yielding @xmath112 . using this relation , in eq . ( [ i1 ] ) we change to the integration over @xmath113 with the jakobian @xmath114 $ ] : @xmath115 positiveness of @xmath116 in eq . ( [ i1 ] ) enforces the choice of the lower integration limit : @xmath117 for @xmath65 and @xmath118 for @xmath119 . using eqs . ( [ op1 ] ) , ( [ op2 ] ) and ( [ i2 ] ) we are able to analyze the critical current @xmath92 in the whole range of the pair - breaking parameter , @xmath120 [ see , fig . [ iz ] ] . in line with the discussed behavior of the andreev bound states , for a narrow resonance , @xmath121 , the critical current starts to drop significantly only upon entering the gapless superconductivity regime @xmath122 . on the other hand , for a broad resonance , @xmath123 , the suppression of @xmath0 is almost linear in the whole range . we note that in both cases the behavior of @xmath0 strongly deviates from that of the bulk order parameter ( red curve ) @xcite largely due to the pair - breaking effect on the abs . in practice , the @xmath124 dependence can be measured by applying a magnetic field [ the case where @xmath125 and @xmath126 in an experiment similar to ref . where a quantum dot , defined in a single - wall carbon nanotube , was strongly coupled to the leads with the ratio @xmath127 . carbon nanotube quantum dots with lower @xmath128 values are accessible experimentally , too @xcite . we also found that the crossover between the gapped and gapless regimes is accompanied by a qualitative change in the shape of the josephson current - phase relation @xmath1 as demonstrated in fig . [ i_f ] for the on - resonance case @xmath129 and @xmath130 . the @xmath1 relation is anharmonic as long as the junction with @xmath131 supports the abs ( black and blue curves ) . the vanishing of the abs upon entering the gapless regime leads to a nearly sinusoidal current - phase relation ( red curve ) . a closely related effect is demonstrated in fig . [ i_e ] showing the modification of the critical current resonance lineshape with the increasing pair - breaking strength . in the absence of pair breaking it is nonanalytic near @xmath129 ( black curves ) reflecting the anharmonic @xmath1 due to the abs in a transparent channel @xcite . on approaching the gapless regime this singularity is smeared out ( red curves ) , which is accompanied by the suppression of the @xmath0 amplitude . at finite temperatures @xmath132 the pair - breaking - induced smearing of the resonance peak will enhance the usual temperature effect . in conclusion , we have proposed a model describing resonant josephson tunneling through a quantum dot beyond the conventional bcs picture of the superconducting state in the leads . it allows for nonperturbative treatment of pair - breaking processes induced by a magnetic field or paramagnetic impurities in diffusive superconductors . we considered no coulomb blockade effects , assuming small charging energy in the dot @xmath133 , which was , for instance , the case in the experiment of ref . . our predictions , however , should be qualitatively correct also for weakly coupled dots with @xmath134 at least as far as the dependence of the ctitical supercurrent on the pair - breaking parameter is concerned . indeed , for a narrow resonance the andreev bound states begin to respond to pair breaking only when the gap @xmath2 becomes sufficiently small [ see , fig . [ e03 ] ] so that for a finite @xmath135 one can expect a sharp transition to the resistive state , too , similar to that shown in fig . [ iz ] for @xmath121 . we thank d. averin , c. bruder , p. fulde , a. golubov , m. hentschel , t. novotny , v. ryazanov and c. strunk for useful discussions . financial support by the deutsche forschungsgemeinschaft ( grk 638 at regensburg university ) is gratefully acknowledged .
we propose a model for resonant josephson tunneling through quantum dots that accounts for cooper pair - breaking processes in the superconducting leads caused by a magnetic field or spin - flip scattering . the pair - breaking effect on the critical supercurrent @xmath0 and the josephson current - phase relation @xmath1 is largely due to the modification of the spectrum of andreev bound states below the reduced ( abrikosov - gorkov ) quasiparticle gap . for a quantum dot formed in a quasi - one - dimensional channel , both @xmath0 and @xmath1 can show a significant magnetic field dependence induced by pair breaking despite the suppression of the orbital magnetic field effect in the channel . this case is relevant to recent experiments on quantum dot josephson junctions in carbon nanotubes . pair - breaking processes are taken into account via the relation between the andreev scattering matrix and the quasiclassical green functions of the superconductors in the usadel limit .
ceramic materials have applications in many fields of science and engineering @xcite . in particular , the efficient production of dense ceramics with nanosized grains using ultrafine powders is a subject of increased attention . the most common method for achieving densification in ceramics is solid - state sintering , a thermally activated process that involves the diffusion and redistribution of mass by means of various physical processes , leading to a compactified and consolidated final structure . in the case of nanosized powder compacts , a major concern is the considerable grain growth during the densification process . special consolidation techniques such as hot isostatic pressing , spark plasma and microwave sintering in many cases successfully limit this undesirable phenomenon . the addition of dopants able to limit grain boundary migration is another effective option . a very effective way to limit grain growth is to promote densification at a lower sintering temperature @xcite . mechanical activation of powders by high - energy ball milling can contribute greatly toward this direction . high - energy mechanical treatment of solids usually leads to the accumulation of excess energy in the structure resulting in plastic deformation and fracturing of crystals @xcite . the process usually involves initial reduction of grain sizes , followed by plastic deformation of the crystalline structure , whereas excessive milling time may finally lead to grain aggregation @xcite . grain size reduction leads to an increase of surface energy @xcite . under deformation the lattice structure is transformed , perturbed or completely destroyed . this effect is evidenced by the broadening , decrease of intensity , and finally total disappearance of xrd reflections @xcite . this means that the deformation and increase of structural disorder can bring the solid into a nanocrystalline or even completely amorphous state [ 6 , 11 - 13 ] . finally , the possible aggregation of grains imposes significant pressure on the grains increasing their contact surface areas @xcite . this lowers the activation energy for the diffusion processes involved in sintering , thus accelerating the entire process and reduces the sintering temperatures @xcite . the alteration of the state of solid materials due to mechanical activation can lead to significant changes in physical and chemical properties . order - disorder transitions and phase transformations can then take place at much lower temperatures . additionally , nuclei of new phases can often appear at the interfaces possibly followed by subsequent growth of crystallites @xcite . thus , crystallization from the amorphous state and chemical reactions can be easily accelerated with subsequent treatment , resulting in materials with improved or better - controlled properties @xcite . therefore , a central issue for mechanically activated powders is the relations between the properties of the initial powder and the structure of the resulting ceramic material . improved performance is in many cases achieved by an optimal balance between particle size and mechanical activation effects @xcite . theoretical and computational investigations can help to understand the impact of mechanical activation on the sintering behaviour of nanosized powders , as well as on sintering temperature dependence . even though theories of the sintering process has been intensely researched during the last fifty years , there is a lack of quantitative , predictive models for the microstructural evolution during sintering @xcite . in the final stage of sintering the densification process is accompanied by grain growth leading to the final pore structure ( grain size distribution , geometry and topology of pores ) . the physical mechanisms responsible for these effects involve various transport processes acting in the bulk and on the surface of grains , at the boundary between grains , as well as the evaporation - condensation mechanism . given the complexity of the physical processes and the initial micro - structure , it is not surprising that a variety of mathematical models have been proposed that focus on different stages of the sintering process , based on various simplifying assumptions regarding the number of grains , the geometry and topology of the pore structure , etc . the existing models include continuum diffusion or transport equations @xcite that are solved either analytically ( in simplified cases ) or numerically , and kinetic monte carlo ( potts model ) simulation approaches @xcite . more specifically in the case of grain growth , computational approaches involve vertex , monte carlo potts , phase field , and cellular automata models @xcite . we introduce a model of grain growth applying to the final stage of sintering . the model describes mass transfer between grains in terms of a balance equation . such a model should account for differences in the initial distribution of grain sizes , which contribute to differential growth of the grains . a stochastic aspect is introduced in the problem , due to the fact that the initial grain configuration is not known . instead of attempting to model explicitly the various physical mechanisms that contribute to mass transfer , we opt for an effective approach that does not distinguish between the different mechanisms . the amorphization degree represents the fraction of the grain that is in the amorphous state . grains with higher degree of amorpization are more active , and tend to transfer mass more effectively . smaller grains ( e.g. , such as those created by mechanical activation ) are expected to have a higher degree of amorphization . since the lattice formed by the centres of the grains is irregular , the approach to a continuum limit is not well - defined . thus , a discrete mass transfer equation is proposed . the considerations in the paragraph above imply that transition rates should depend on grain size . the locations of the grain centers are denoted by @xmath0 , and those of the nearest neighbors of point @xmath0 by @xmath1 . the grain masses at time @xmath2 are denoted by @xmath3 and @xmath4 is the grain radius . the grain mass density ( assumed constant ) is denoted by @xmath5 . we introduce a local ` order factor ' @xmath6 , which is inversely proportional to the ` activation ' of each grain . we assume that the order factor increases with the grain mass , since the fraction of amorphous areas is smaller for larger grains . the order factor influences the local activation energy . to our knowledge , there are no experimental measurements of activation energies as a function of grain amorpization . hence , it is necessary to surmise the functional dependence of the order factor based on intuition and physical constraints . based on these considerations , we propose the following system of equations : @xmath7 @xmath8 @xmath9 in eq . ( [ mass - trans ] ) @xmath10 represents the transition rate at time @xmath2 for mass transfer into the grain located at @xmath0 from the neighbouring grains . similarly , @xmath11 represents the transition rate at time @xmath2 for mass transfer from the grain at @xmath0 into the neighbouring grains . the mass and the radius are related via @xmath12 where @xmath13 is a geometric factor that on the dimension @xmath14 . the order factor defined by eq . ( [ order ] ) , where @xmath15 is a reference mass and @xmath16 a dimensionless constant , satisfies the conditions @xmath17 for @xmath18 and @xmath19 for @xmath20 . the first condition implies that the activation energy tends to vanish for small grains , which are expected to have a significant amorphous fraction , while the second condition implies a very large activation energy for big grains that are predominantly ordered . in eq . ( [ rate ] ) , @xmath21 is a constant that determines the scale of the transition rates , and @xmath22 is the gas constant . based on eqs . ( [ rate ] ) and ( [ order ] ) , the following equation is obtained for the transition rates : @xmath23 \right\}.\ ] ] although the mass transfer equation ( [ mass - trans ] ) looks similar to the master equation used in the analysis of diffusion processes , it differs from the latter in the following : first , the conserved quantity in the master equation is a probability function , while in eq . ( [ mass - trans ] ) it is the total mass of the grains . secondly , the classical master equation is linear in the probability , while eq . ( [ mass - trans ] ) is nonlinear in the mass due to the variations in grain activity . finally , in classical diffusion the transitions tend to generate a uniform steady state , while this is not necessarily the case for eq . ( [ mass - trans ] ) , as shown below . for a time - independent steady state to exist , the time derivative of @xmath3 must vanish asymptotically . this is accomplished if the ` detailed balance ' condition is satisfied , i.e. , @xmath24 \right\}.\ ] ] if a steady state exists in which some grains have @xmath25 , while for at least one of their neighbours @xmath26 , the left hand - side of eq . ( [ det - bal ] ) tends to zero . then the right hand - side becomes @xmath27 $ ] . since @xmath28 , a steady state with vanishing grain masses can exist only if @xmath29 . this condition is necessary but not sufficient for an asymmetric steady state . using as an average estimate for the activation energy the value of @xmath30 kj / mol @xcite and a typical sintering temperature ( e.g. , 1500@xmath31 c ) , @xmath32 , giving an order of magnitude estimate . the value of @xmath16 determines the variation of the activation energy due to mechanical activation effects . while a large value of @xmath33 is necessary for achieving a non - uniform steady state , very large values slow down considerably the sintering process . we expect that @xmath34 , to allow for reduction of the activation energy by mechanical activation . at the opposite limit , if @xmath35 the ratio of the transition rates in eq . ( [ det - bal ] ) will be close to one , and a uniform steady - state is expected . we use the length ( l ) - mass ( m ) - time ( t ) system of units for dimensional analysis . the governing parameters @xcite of the grain mass equation are @xmath36 , and @xmath37 . to these one should add the parameters that determine the initial grain size distribution . if this distribution is approximately gaussian , the parameters include the mean value @xmath38 and standard deviation @xmath39 . then , the governing parameters and their dimensions are as follows : @xmath40=m , \ , [ \lambda]=t^{-1 } , \ , [ \tilde{q}]=[r\ , t]=m\,l^{2}\,t^{-2 } , \ , [ x]=[\overline{r}_{0}]=[\sigma_{r,0}]=l , \ , [ t]=t , \ , [ \rho]= m\,l^{-3}. $ ] there are three variables with independent dimensions , i.e. , @xmath41 . according to the @xmath42-theorem of dimensional analysis @xcite , the equation ( [ mass - trans ] ) can be expressed in terms of six dimensionless combinations of the governing parameters , e.g. , @xmath43 using the notation @xmath44 , @xmath45 , @xmath46 , @xmath47 , @xmath48 , @xmath49 for the dimensionless variable combinations , the mass - transfer equation is expressed as follows : @xmath50 note that the dependence on the scaled variables @xmath51 , @xmath52 and @xmath53 is not explicit in eq . ( [ mass - trans2 ] ) , since these variables involve the initial conditions . equation ( [ mass - trans2 ] ) can be solved numerically using a forward finite - difference discretization of the time derivative . this leads to the following updating scheme : @xmath54,\ ] ] where the transition rate @xmath55 is given by @xmath56 $ ] . if @xmath57 is the maximum transition rate at time @xmath58 , the time increment @xmath59 should satisfy the condition @xmath60 . a very small time step would slow down the evolution . since the transition rates change dynamically during the process , the time step is adaptively updated . we solve eq . ( [ mass - disc ] ) for a @xmath61 chain of @xmath62 grains with periodic boundary conditions . the initial grain distribution is assumed to be gaussian with @xmath63 in arbitrary units and coefficient of variation @xmath64 . the reference mass @xmath15 is taken equal to the mean of the initial mass distribution . first , we consider the case of low activation energy using a dimensionless activation parameter @xmath65 . the initial grain size distribution and the evolved distribution after 100000 steps are shown in fig . ( [ figure1 ] ) . in the evolved state the radius of all the grains is approximately equal to the initial mean radius . in this case , the low activation energy leads to a diffusive behaviour that asymptotically drives the system toward a uniform steady state . next , we consider the case of high activation energy . the grain size distribution for @xmath66 is shown in fig . ( [ figure2 ] ) . here , the larger grains grow at the expense of the smaller ones . the grain radius distribution develops a bimodal structure that includes a fraction of very small grains . the evolution of the grain radius coefficient of variation @xmath67 , skewness coefficient , @xmath68 and the maximum radius , @xmath69 versus time are shown in fig . ( [ figure3 ] ) . the time is calculated based on @xmath70 . the coefficient of variation @xmath67 increases in magnitude with time as a result of the evolving asymmetry of the distribution . the skewness coefficient @xmath68 also develops a non - zero value as the distribution evolves away from the gaussian . the @xmath69 also increases albeit slowly . the plots shown in fig . ( [ figure3 ] ) seem to indicate a smooth approach to a time - independent steady state as the time increases . also note that if the evolution of the moments were plotted versus the number of steps ( instead of the time @xmath58 , the plots would exhibit discontinuities due to the fact that certain steps may correspond to significantly larger time increments than others . the spatial configuration of the grains is illustrated in fig . ( [ figure4 ] ) by plotting the initial and final ( after 100000 steps ) radius of the first 30 grains : the smaller grains tend to shrink , while the larger grains grow at a considerably slower rate . the asymmetry is due to mass conservation , i.e. , the fact that an increase of the grain radius by @xmath71 increases the mass by an amount that exceeds the respective reduction in mass due to a decrease @xmath71 of the radius . a number of grains , the initial size of which is close to the mean radius , do not change appreciably their size . we note that the model presented here does not include grain coalescence . this can be addressed by solving equation ( [ mass - disc ] ) iteratively , with the first iteration ending when the radius of one ( or more ) of the smaller grains drops below a certain threshold ; at this point the smaller grains would coalesce with the largest nearest neighbours . the number of grains would thus be reduced , and the resulting radius distribution would be used as an initial condition for the next iteration of eq . ( [ mass - disc ] ) . this mechanism leads to a coarse - graining of the grain radius distribution . the renormalization group approach may be a suitable framework for investigating the asymptotic grain size distribution under this coarse - graining procedure . we presented a discrete , nonlinear equation for grain growth in sintered grain aggregates , using transition rates that depend on the degree of amorphization of each grain . the model was solved numerically for a chain of grains with periodic boundary conditions . the activation energy was shown to be a crucial factor for the asymptotic grain size distribution , since it leads to a transition from a bimodal steady state to a uniform one ( diffusion regime ) . various aspects of the model require further study , including the existence of a well - defined threshold between the two regimes , the impact of governing parameters on the grain size evolution , the roles of the initial grain size distribution and the initial grain configuration . finally , grain coalescence needs to be incorporated in the model ( in the spirit discussed in the previous section ) in order to observe realistic grain growth . research supported by the ec grant ` super high energy milling in the production of hard alloys , ceramic and composite materials ' ( activation fp6 ) .
the evolution of grain structures in materials is a complex and multiscale process that determines the material s final properties @xcite . understanding the dynamics of grain growth is a key factor for controlling this process . we propose a phenomenological approach , based on a nonlinear , discrete mass transfer equation for the evolution of an arbitrary initial grain size distribution . transition rates for mass transfer across grains are assumed to follow the arrhenius law , but the activation energy depends on the degree of amorphization of each grain . we argue that the magnitude of the activation energy controls the final ( sintered ) grain size distribution , and we verify this prediction by numerical simulation of mass transfer in a one - dimensional grain aggregate . [ multiblock footnote omitted ]
deconvolution is an ubiquitous data processing method that arises in a variety of applications , _ e.g. _ , biomedical imaging @xcite , astronomy , remote sensing @xcite and video and photo enhancement @xcite . formally , this problem amounts to recovering a signal @xmath0 from _ blurred _ and _ corrupted _ observations @xmath1 : @xmath2 where @xmath3 stands for the _ convolution _ operation between @xmath0 and some other signal @xmath4 , while @xmath5 is a general function accounting for some corrupting noise , _ e.g. _ , @xmath6 under an additive corruption model with some ( gaussian ) noise @xmath7 . for instance , when a scene is captured by an optical instrument , the observation is a blurred or degraded version of the original image , corrupted by noise and by some effects due to the instrument ( motion , out - of - focus , light scattering , ... ) . in such a case , the imaging system is usually assumed to be well represented by the sensing model ( [ eq : blur - model ] ) where the ideal image @xmath0 is blurred by a point spread function ( psf ) @xmath4 . implicitly , the psf describes the response of the imaging device to a dirac point source , _ i.e. _ , the impulse response of the instrument , while @xmath5 can distinguish different sources of noise contaminating the , is an astronomical event where the moon or a planet move across the disk of the sun , hiding a part of it , as seen from the telescope . since the euv emission of transiting celestial bodies is zero , any apparent emission is known to be caused by the instrument . therefore , this type of event provides strong prior information that allows the regularization of the blind deconvolution problem . let us note that all these parametric methods present the disadvantage of being specific for a given instrument and depending on a good characterization of the psf , which is not always possible @xcite . central to our work is the information provided by the transit of a celestial object whose apparent boundary on the recorded image can be predicted with high ( sub - pixel ) accuracy . this is typically the case for the observation of moon or venus transits for which ephemeris allows us to precisely know their apparent boundaries at the recording time , provided of course that _ ( i ) _ the small variations of the object topography around a perfect disk ( _ e.g. _ , mountains , craters ) are smaller than a pixel width , and _ ( ii ) _ , that such a transitting object has no atmosphere that could blur its apparent boundary ( _ e.g. _ , as for an earth transit ) . as will be clear below , these hypotheses have been respected for at least two cases : for the instrument secchi / euvi during the 2007 lunar transit and for sdo / aia during the 2012 venus transit . for these two situations , and for any other transit respecting the conditions above , we know a priori the values for a set of pixels in the image and also their exact location . our work uses such information as constraint in the function to be optimized for blindly deconvolving images . notice that our approach could also be applied to other situations where pixel values and exact location are known a priori ( _ e.g. _ , dark calibration patterns in computer vision applications ) . contrarily to previous works in solar physics , the blind deconvolution method demonstrated in this paper is not based on a specific model of the psf , but infers it from the observed data . this allows the method to be used for estimating the psf of any instrument provided it has a prior information on a set of pixel values . since no model is imposed on the psf , the large number of unknowns makes the problem computationally intractable . we thus focus on the estimation of the psf core , which is defined as the central pixels encompassing at least @xmath8 of the total psf energy . note that this thresholding follows the radial energy characterization of the sdo / aia psf in the paper of poduval et al . @xcite and a similar study of the available psfs of secchi / euvi ( _ e.g. _ , @xcite ) . the psf core accounts for charge spreading , optical aberrations , and diffraction effects . note that the same algorithm would be able to handle a parametric model of the psf provided a precise instrument characterization is available . for simplicity , we consider an average additive gaussian noise model that gathers all sources of noise present in the observation . a detailed discussion on how to handle more general noise models is provided in section [ sec : discussion ] . the proposed method is first tested through simulated realistic scenarios . results show the ability to estimate a given psf with high quality . we also show the importance of considering multiple transit observations in order to provide a better conditioning to the filter estimation problem . we validate our method on aia and euvi observations using solar transits . the validation of the recovered filters is based on the reduction of the celestial body apparent emissions . the estimated psfs were also tested for images containing active regions . results show that , when recovering the psf core , the proposed non - parametric scheme can achieve similar results to parametric methods . also , we consider the case where the parametric psf is incorporated in the imaging model as an additional filter convolving the actual image @xcite . we show that the resulting parametric / non - parametric psf provides better results than both parametric and non - parametric psfs estimated independently . let us note that , since parametric methods are based on the physics of the instrument , they tend to provide a more precise psf containing both the core and the diffraction peaks . however , when the telescope presents unknown or unexpected behavior , an accurate model of the filter can not be built . in such cases , we require non - parametric models , as the one presented in this paper , which due to their flexibility can handle better some instrument s properties . in section [ sec : model ] we describe the discrete forward model and the noise estimation . in section [ sec : bid ] we present the formulation of the blind deconvolution problem based on the image and filter prior information . section [ sec : am ] describes the alternating minimization method used for estimating the filter and the image . it briefly presents the method , including the initialization , the parameter estimation , the stopping criteria and the numerical reconstruction . in section [ sec : nbid ] we present a non - blind deconvolution method that is used for the experimental validation . in section [ sec : results ] we present the results on both synthetic and experimental data for the aia and euvi telescopes . finally , section [ sec : discussion ] presents a discussion of the obtained results , providing some perspectives for future research work . we denote by @xmath9 , @xmath10 and @xmath11 the sets of natural , integer and real numbers , respectively . the set of the non - negative real numbers is denoted by @xmath12 . most of domain dimensions are denoted by capital roman letters , _ e.g. _ , @xmath13 vectors and matrices are associated with bold symbols , _ e.g. _ , @xmath14 or @xmath15 , while lowercase light letters are associated with scalar values . the @xmath16 component of a vector @xmath17 reads either @xmath18 or @xmath19 , while the vector @xmath20 may refer to the @xmath16 element of a vector set . the vector of ones in @xmath21 is denoted by @xmath22 . the set of indices in @xmath21 is @xmath23 = \{1 , \cdots , d\}$ ] . the support of a vector @xmath24 is defined as @xmath25 : u_i \neq 0 \}$ ] . the cardinality of a set @xmath26 , measuring the number of elements of the set , is denoted by @xmath27 . the convolution between two vectors @xmath28 for some dimension @xmath29 is denoted equivalently by @xmath30 . we denote @xmath31 the class of proper , convex and lower - semicontinuous functions from a finite dimensional vector space @xmath32 ( _ e.g. _ , @xmath33 ) to @xmath34 $ ] @xcite . the ( convex ) indicator function @xmath35 of the set @xmath26 is equal to @xmath36 if @xmath37 and @xmath38 otherwise . the 2-d discrete delta function , denoted as @xmath39 , is equal to 1 if @xmath40 , and @xmath36 otherwise . the transposition of a matrix @xmath41 reads @xmath42 . for any @xmath43 , the @xmath44-norm of @xmath17 is @xmath45 . the frobenius norm of @xmath41 is given by @xmath46 . the gaussian distribution of mean @xmath36 and variance @xmath47 is denoted by @xmath48 . the telescope imaging process can be mathematically modeled as the instrument s psf @xmath4 convolving the true image @xmath0 , _ i.e. _ , @xmath49 . the convolution operation , represented by @xmath3 , consists of integrating portions of the actual scene weighted by the psf . in the following , we consider a discrete setting where the convolution integration is represented by a discrete sum . let us consider that , during the transit , the euv telescope acquires a set of @xmath50 images containing a celestial body . in order to limit the computation time , the effective field - of - view ( fov ) of each observation is restricted to an image patch centered on the transiting celestial body with a size several times bigger than the body apparent diameter . the effects of this fov truncation will be discussed in detail in section [ sec : bid ] and section [ sec : results ] . in our work , all @xmath51 images are represented as @xmath52 vectors so that linear transformations of images are identifiable with matrices . each observed patch is a collection of values gathered in a vector @xmath53 , with @xmath54 . the observed values are modeled as a true image @xmath55 convolved by the instrument s psf ( @xmath56 ) . the psf is assumed to be spatially and temporally invariant , thus is the same for each observed patch . for simplicity , we consider that all sources of noise present in the observation can be gathered in an average noise model , which is assumed to be additive , white and gaussian . the observations are then assumed to be affected by an additive white gaussian noise ( awgn ) , @xmath57 , with @xmath58 . the acquisition process of the telescope is thus modeled as follows : @xmath59 this model assumes that the observed images @xmath60 are uncompressed and have been corrected for ccd effects ( dark current removal , flat fielding , despiking ) . in the case of sdo / aia , this corresponds to level 1 data processing . in 1-d , the discrete circular convolution @xmath61 of a vector @xmath62 with a filter @xmath63 can always be described by the product of @xmath17 with a circulant matrix @xmath64 _ i.e. _ , a matrix where every row is a right cyclic shift of the _ kernel _ @xmath65 @xcite . in 2-d and in particular for the model ( [ eq : discrete_model_vector ] ) , the convolution of @xmath66 by @xmath4 can still be represented by the multiplication of @xmath66 by matrix @xmath67 that is now _ block - circulant with circulant blocks _ , _ i.e. _ , a matrix made of @xmath51 blocks , each block being a @xmath68 circulant matrix representing the action of one row of the 2-d filter @xmath4 in ( [ eq : discrete_model_vector ] ) . if we gather the @xmath50 true images in a single matrix @xmath69 , the acquisition model can be transformed into the following matrix form : @xmath70 with @xmath71 two matrices that similarly gather the @xmath50 observed images and their noises , respectively . for the sake of simplicity and to keep fast numerical methods , the model ( [ eq : discrete_model_matrix ] ) implicitly considers an additive noise model . a way to generalize our method to poisson noise would be to include a vst @xcite in the acquisition model in ( [ eq : discrete_model_matrix ] ) , both on the observations @xmath72 and on the convolution result @xmath73 ( see section [ sec : discussion ] for more details ) . under the awgn assumption , the energy of the residual noise is known to be bounded using the chernoff - hoeffding bound @xcite : @xmath74 with high probability for @xmath75 . a first guess of the noise variance @xmath47 can be estimated using the robust median estimator ( @xmath76 ) @xcite , which is based on the assumption that the noise is an additional high frequency component in the observed signal . more details on this variance estimation are provided in section [ sec : expdata ] . let us note that the variance could also be obtained from the physical noise characteristics using , for instance , dark areas in the images to estimate the dark noise . however , such computations may not consider all the noise sources present in the observations , providing an underestimation of the actual variance . in this work , we prefer to adopt another strategy . since ( [ eq : epsilon ] ) considers an average noise model and since ( [ eq : discrete_model_matrix ] ) is an awgn approximation of the actual data noise corruption , we aim to find an adaptive value of @xmath77 that optimizes the awgn assumption , _ i.e. _ , which minimizes somehow the proximity of this simple model to the true corruptions . section [ sec : expdata ] describes in detail this adaptive strategy and demonstrates its advantage over the fixed choice @xmath78 . in this section , we aim at reconstructing both @xmath79 and @xmath4 from the noisy observations @xmath72 . since the data is corrupted by an awgn , we can estimate the image and the filter using a least squares minimization , _ i.e. _ , by finding the values that minimize the energy of the noise : @xmath80 with @xmath81 and @xmath82 . the blind deconvolution problem in ( [ eq : ls ] ) is ill - posed and can have infinite possible solutions . by regularizing this problem we aim to reduce the amount of possible solutions to those that are meaningful . in the coming sections , we first describe the prior information and constraints on the image and the filter and then use this information to formulate a regularized blind deconvolution problem . in the following , we present in detail the available prior information on the image candidate . [ [ a - zero - disk - intensity ] ] * ( a ) zero disk intensity : * + + + + + + + + + + + + + + + + + + + + + + + + + + each image of the transit contains the observation of a celestial body , venus or the moon . for each observation @xmath60 , we know that the celestial bodies euv emission is zero and therefore can be represented as a black disk of constant radius @xmath83 and center @xmath84 , both being known from external astronomical observations : @xcite for sdo / aia and ( j .- p . wuelser _ private communication _ ) for secchi / euvi . the set of pixels of @xmath66 inside the disk is denoted by @xmath85 . in the minimization problem we can set to zero the pixels of the @xmath86 image that are inside the set @xmath87 , _ i.e. _ , @xmath88 for all @xmath89 . this prior information is crucial in the regularization of the blind deconvolution problem as it allows us to remove the issue of interchangeability between the filter and the image . furthermore , it prevents `` oppositely '' translated solutions of the image and the filter , a problem occurring because the convolution process is blind to such translations , _ i.e. _ , @xmath90 , with @xmath91 a translation by @xmath92 . [ [ b - analysis - based - sparsity - model ] ] * ( b ) analysis - based sparsity model : * + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + as commonly done with ill - posed inverse problems associated to image restoration tasks , we regularize our method with a 2-d wavelet prior . compared to a frequency representation , as obtained by the fourier transform , a 2-d wavelet transform allows to represent images with a _ multi - resolution _ scheme @xcite , _ i.e. _ , with a linear combination of localized _ wave atoms _ , with variable sizes and locations , whose number is much smaller than the initial number of pixels . most inverse problems in imaging are regularized using a synthesis wavelet prior , _ i.e. _ , promoting the synthesis of the estimated image with as few ` wavelets ' as possible ( sparsity criterion ) . more recently , better inversion results have been obtained using _ analysis_-based sparsity models @xcite , which rather promote sparse _ projections _ of the estimated image over a redundant system of wave - functions , _ i.e. _ , a _ dictionary_. this set can be much larger than an orthonormal basis and hence it can efficiently capture much more different image features . following this recent trend , we decide to adopt such an analysis prior that , in opposition to the synthesis framework , also allows us to work directly on the image domain without increasing the dimensionality of the optimization problem . as a dictionary , we use the system associated to the undecimated discrete wavelet transform ( udwt ) @xcite . the udwt can also be seen as the union of all translations of an orthonormal dwt . conversely to the dwt , this makes the udwt translation invariant , _ i.e. _ , it enables an efficient characterization of all image features whatever their location @xcite , hence providing a better image reconstruction than the traditional dwt . in this context , we assume that , if we represent the image candidate on a redundant wavelet dictionary @xmath93 made of @xmath94 vectors in @xmath95 , the wavelet coefficients are sparse , _ i.e. _ , the coefficients matrix @xmath96 has few important values and its @xmath97-norm ( computed over all its entries ) is expected to be small . however , the wavelet coefficients whose support touches the boundary of the moon or venus disk are not sparse . hence , we do not consider in this prior model those coefficients that are affected by the occulting body . also , we follow a common practice in the field which removes the ( unsparse ) scaling coefficients @xcite from the @xmath97-norm computation . the set of _ detail _ coefficients that are not affected by the black disk is denoted by @xmath5 , with @xmath98 . see section [ sec : synthdata ] for further details on how to compute the set @xmath5 . the matrix @xmath99 is the corresponding selection operator that extracts from any coefficients vector @xmath100 only its components indexed by @xmath5 . the global prior information can thus be exploited by promoting a small @xmath97-norm on the wavelet coefficients belonging to @xmath5 . the rationale of this prior is to enforce noise canceling and avoid artifacts in the reconstructed image . [ [ c - image - non - negativity ] ] * ( c ) image non - negativity : * + + + + + + + + + + + + + + + + + + + + + + + + + + + note that the image corresponds to a measure of a photon emission process , therefore , its pixel values must be non - negative everywhere . since the filter is not known a priori and it changes from one instrument to another , we are interested in adding soft constraints that are common to most solar euv telescopes . we are not aiming at reconstructing all of the psf s components but rather at estimating its core , which is responsible for most of the diffused light . we first consider that , since the psf corresponds to an observation of a point , it is non - negative everywhere . additionally , by assuming that the amount of light entering in the instrument is preserved , the @xmath97-norm of the filter candidate must be equal to one . this is explained by taking the sum over the pixels of one observed patch in ( [ eq : discrete_model_matrix ] ) , which is equal to @xmath101 in a noiseless process . since the filter is non - negative , taking its @xmath97-norm equal to one is equivalent to @xmath102 , which results in @xmath103 , _ i.e. _ , the light is preserved . therefore , the filter candidate must belong to the probability simplex @xcite , defined as @xmath104 . one important assumption in our proposed method is that the filter candidate is of size @xmath105 , for any @xmath106 , and is centered in the spatial origin of the discrete grid . this means that the filter candidate has a limited support inside a set @xmath107 of size @xmath108 , _ i.e. _ , the filter has only important values in the central pixels and is negligible beyond the considered support . this allows us to work with a patch ( only a part of the observation ) and not with the complete fov of the instrument in order to reduce the computation time . euv images have a high dynamic range . hence , due to long - range effects , a given pixel value can be affected by the value of another high intensity pixel located as far as 100 arcsec away for sdo / aia @xcite and 1500 arcsec away for secchi / euvi @xcite . in such cases , despite a psf that can rapidly decay far from its center , the high intensity pixel can induce a long - range effect that is not taken into account by a filter with truncated support . let us define the complementary set of @xmath107 as @xmath109 \backslash \gamma$ ] , with @xmath110 . we assume that the actual psf is composed by two different filters : the psf core with support on @xmath107 , denoted by @xmath111 , which accounts for short - range effects , and another one , denoted @xmath112 with support on @xmath113 , accounting for long - range effects , _ i.e. _ , @xmath114 . an accurate estimation of the long - range psf is out of the scope of this work . its effect inside the disk of the celestial body is modeled as a constant @xmath115 , _ i.e. _ , @xmath116 ( see appendix [ sec : long_range_psf ] for more details ) . for a given patch , the acquisition model in ( [ eq : discrete_model_vector ] ) can then be approximated as : @xmath117 in this work , we aim at estimating only the psf core , which is called @xmath4 hereafter . the effect of the long - range psf , denoted by @xmath115 , is computed in a preprocessing stage by the average value inside the center of several observed transit disks ( see section [ sec : expdata ] ) . this simple estimation is motivated by the presence of a systematic intensity background at the center of each observed disk transit . this one is quite independent of the sun activity around such disk and its mean intensity ( few tens of dns ) is also far above the estimated noise level ( few dns ) . in a future work we plan to replace this rough evaluation by a joint estimation of the value of @xmath115 with the filter and the image ( see section [ sec : discussion ] for some discussions ) . notice that our method could allow the addition of other convex constraints , _ e.g. _ , if we know that the filter is sparse or that @xmath118 , for some upper bound @xmath119 on @xmath120 such as a specific power law decay . however , we will not consider this possibility here as we want to stay as agnostic as possible on the properties of the filter to reconstruct . from sections [ sec : imageprior ] and [ sec : filterprior ] , we can formulate the following regularized blind deconvolution problem : @xmath121 with @xmath81 , @xmath122 and @xmath123 cancels the long - range part of the filter . the regularization parameter , denoted by @xmath124 , controls the trade - off between the sparsity of the image projection in a wavelet dictionary and the fidelity to the observations . this essential parameter is estimated in appendix [ sec : param ] . it is important to note that the prior information related to the darkness of the occulting body ( @xmath125 if @xmath89 ) , which is crucial in the regularization of the blind deconvolution problem , is taken into account in the first constraint in ( [ eq : bid2 ] ) . the problem of estimating the image and the filter in ( [ eq : bid2 ] ) can also be written as @xmath126 where @xmath127 is the objective function , with @xmath128 the modified observations . the convex sets @xmath129 and @xmath130 are defined as follows : @xmath131 ; @xmath132 . in this section , we describe the alternating minimization method used for solving ( [ eq : bid3 ] ) and hence estimating the image and the psf from the noisy observations . the problem defined in ( [ eq : bid3 ] ) is non - convex with respect to both @xmath79 and @xmath4 , but it is convex with respect to @xmath79 ( resp . @xmath4 ) if @xmath4 ( resp . @xmath79 ) is known . such a problem is usually solved iteratively , by estimating the image and the filter alternately @xcite . the general behavior of this type of algorithm is as follows : @xmath133 it has been shown in @xcite that this algorithm is able to provide fairly good results . however , there are not theoretical guarantees for its convergence . recently , @xcite proposed a proximal alternating minimization algorithm where cost - to - move functions are added to the common alternating algorithm . these functions are quadratic costs that penalize variations between two consecutive iterations . they guarantee the algorithm convergence provided some mild conditions are met on the regularity of the objective function @xmath134 and the constraints in ( [ eq : bid3 ] ) . the algorithm iterates as follows : @xmath135 where @xmath136 and @xmath137 are the cost - to - move penalization parameters that control the variations between two consecutive iterations . section [ sec : lambda ] describes how these parameters are tuned . the non - convexity of ( [ eq : bid3 ] ) prevents attaining a global minimizer . however , it can be shown that , since the objective function @xmath138 and the constraints in ( [ eq : bid3 ] ) meet the conditions stated in @xcite , the algorithm ( [ eq : alternating_attouch ] ) converges to a critical point of the problem . later in this section we describe how to stop the iterations so that the vicinity of a critical point of ( [ eq : bid3 ] ) is reached . the presence of cost - to - move parameters ( @xmath136 , @xmath137 ) different than zero ensures the convergence of the algorithm ( [ eq : alternating_attouch ] ) to a critical point of @xmath139 @xcite . however , their values are not crucial in the reconstruction results . in the following , we describe the tuning criteria used in this work which is based on the works of puy and vandergheynst @xcite . the iterations start with high values of @xmath136 and @xmath137 , keeping the image and the filter estimations closer to the initial values . when the number of iterations @xmath140 increases , the filter and the image estimates become more and more accurate and the parameters @xmath136 and @xmath137 can be progressively decreased . it is important to find an automatic criterion for stopping the iterations in ( [ eq : alternating_attouch ] ) when the solution reaches the vicinity of a critical point of ( [ eq : bid3 ] ) . a usual stopping criterion is based on the quality of the reconstruction with respect to the ground truth image and filter , which are in general unavailable . as proposed by almeida and figueiredo @xcite , we use the spectral characteristics of the noise to analyze the quality of the estimation . this allows stopping the minimization when a high quality solution has been obtained . let us define each residual image at iteration @xmath140 as the difference between the observed image and the blurred estimate : @xmath141 since the observation model is degraded by an additive white noise , we know that the residual image is spectrally white if the estimation at iteration @xmath140 has a good quality , otherwise the residual contains structured artifacts . therefore , the iterations can be stopped when the residual is spectrally white , _ i.e. _ , when the 2-d autocorrelation @xmath142 of each residual image @xmath143 is approximately the function @xmath144 . let us note that , for the computation of the 2-d autocorrelation function , each residual vector @xmath143 needs to be previously transformed into a matrix of @xmath51 pixels and normalized to zero mean and unit variance . considering a @xmath145 window , we compute the distance between the autocorrelation and the kronecker function @xmath146 as follows : @xmath147 which is higher for whiter residuals . as suggested by almeida and figueiredo @xcite , in our experiments we have used @xmath148 . we then average over the measures obtained for each residual ] in practice , we observe that @xmath149 has large negative values at the beginning of the iterations when the image and the filter are not properly estimated . then , the value of @xmath149 increases with @xmath140 until it reaches a maximum close to zero . finally , it starts to decrease again after the algorithm has converged to the best estimations for a fixed value of @xmath124 and starts overfitting the noise . therefore , the iterations can be stopped when the maximum of the whiteness measure ( @xmath150 ) is reached . a similar behavior has also been observed in @xcite . the proximal alternating minimization algorithm is summarized in algorithm [ alg : alternatemin ] . this algorithm requires to set the initial values of the image , the filter and the regularization parameter . the initialization of the image and the filter is discussed in appendix [ sec : init_imagefilter ] , and the regularization parameter is tuned iteratively as explained in appendix [ sec : param ] . _ _ : @xmath151 _ _ : @xmath152 _ _ : @xmath153 @xmath154 _ _ : @xmath155 compute @xmath149 using ( [ eq : mrp ] ) _ _ : break . return @xmath156 and @xmath157 . the first step in algorithm [ alg : alternatemin ] estimates the image by minimizing a sum of three convex functions with the image constrained to the set @xmath129 . this constraint can be handled by adding the convex indicator function on the set , _ i.e. _ , @xmath158 . the resulting optimization , containing a sum of two smooth and two non - smooth convex functions , can be solved using the primal - dual algorithm of chambolle and pock @xcite summarized in appendix [ sec : appendix_cp ] . the second step in algorithm [ alg : alternatemin ] estimates the filter by minimizing a sum of two smooth convex functions with the filter constrained to the set @xmath130 . the optimization problem can be solved using the accelerated proximal gradient ( apg ) method @xcite . this algorithm is able to minimize a sum of a non - smooth and a smooth convex function , which allows us to handle the constraint as a convex indicator function on the set , _ i.e. _ , @xmath159 . the apg algorithm is briefly described in appendix [ sec : appendix_apg ] . the algorithms used to solve the minimization problems described above were chosen because they are well suited to solve each problem optimally . since the cost functions are different in each step , the same algorithm could not be used to optimally find a solution of both problems . the presence of two non - smooth functions in the first step , prevents the use of the apg algorithm . for the second step , the algorithm of chambolle and pock @xcite could be used to solve the optimization problem , however , this algorithm presented a slower convergence than apg , which is optimal when the optimized cost contains a differentiable function . let us note that , in the numerical reconstruction , the convolution operator is implemented using the fast fourier transform ( fft ) . this operation assumes that the boundaries of the image are periodic , a condition that is far from reality . in actual imaging , no condition can be assumed on the boundary pixels , _ i.e. _ , they can not be assumed to be zero , neither periodic , nor reflexive . not taking care of the boundaries conditions would introduce ringing artifacts . based on the work of almeida and figueiredo @xcite , in the numerical experiments we consider unknown boundary conditions , _ i.e. _ , , the pixels belonging to the image border are not observed . to take this into account , the problem formulation needs to be appropriately modified as discussed in appendix [ sec : appendix_ubc ] . this section is dedicated to the validation of the filter estimated using algorithm [ alg : alternatemin ] . since the true emissions of the moon and venus are zero , the deconvolution with a correct filter should remove all the celestial bodies apparent emissions . therefore , the validation process consists of taking an image from a transit observation that was not used for filter estimation ( _ i.e. _ , @xmath160 ) , deconvolving this image using the estimated filter @xmath4 , and verifying that , in the reconstructed image @xmath161 , the pixels inside the celestial body disk are zero . to obtain this reconstructed image , we formulate a least squares minimization problem regularized using the prior information available on the image , that is , that the image is non - negative and has a sparse representation on a suitable wavelet basis @xmath162 . the proposed non - blind deconvolution method reads as follows : @xmath163 where @xmath164 is the selection operator of the set @xmath165 , with @xmath166 , which contains the detail coefficients . the convex set @xmath167 is defined as @xmath168 . the non - blind image estimation is solved using an extension of the primal - dual algorithm of chambolle and pock @xcite ( see section [ sec : appendix_cp ] for more details ) . since this algorithm is able to minimize a sum of three non - smooth convex functions , the constraint can be handled as a convex indicator function . notice that other algorithms could be used such as the generalized forward backward algorithm @xcite . in this section , we first present some synthetic results that allow us to validate the effectiveness of the proposed blind deconvolution method to recover the correct filter . later we present experimental results on images taken by the sdo / aia and secchi / euvi telescopes . all algorithms were implemented in matlab and executed on a 3.2 ghz intel i5 - 650 cpu with 3.7 gib of ram , running on a 64 bit ubuntu 14.04 lts operating system . three synthetic images are selected to test the reconstruction ( @xmath169 ) . they are defined on a @xmath170 pixel grid ( @xmath171 ) . to simulate actual images from the celestial transit , we take an image of the sun ( the image observed by sdo / aia at 00:00 ut on june 6th 2012 ) and select three different cutouts of @xmath172 pixels each . the center of the cutouts are arbitrarily selected such that the corresponding regions are sufficiently distant from each other and correspond to a sample of a full venus trajectory . in each image , we add a black disk of radius @xmath173 pixels , centered at the pixel @xmath174 $ ] , which represents the celestial body . figure [ fig : syntheticimages ] depicts the images from the simulated transit . , ( center ) second image @xmath175 and ( right ) third image @xmath176 from the simulated transit.,title="fig:",height=141 ] , ( center ) second image @xmath175 and ( right ) third image @xmath176 from the simulated transit.,title="fig:",height=141 ] , ( center ) second image @xmath175 and ( right ) third image @xmath176 from the simulated transit.,title="fig:",height=141 ] during the solar transit , the position and size of the celestial body are known at all moments . however , when this event is imaged , the coordinates of the center represented in the discrete image grid may contain an error of maximum one pixel . this uncertainty is simulated in our synthetic experiments by considering the sets @xmath85 and @xmath5 to be defined using disks of radius @xmath177 and @xmath178 , respectively , with one pixel difference with respect to the actual object s radius , _ i.e. _ , @xmath179 and @xmath180 . two kinds of discrete filters are selected and they have a limited support on a @xmath181 pixel grid ( @xmath182 ) . the first filter is simulated by an anisotropic gaussian function with standard deviation of 2 pixels horizontally and 4 pixels vertically , rotated by 45 degrees ( see figure [ fig : syntheticfilters]-(left ) ) . the second filter is simulated by a x shape ( see figure [ fig : syntheticfilters]-(right ) ) , hereafter called the x filter . these functions help to demonstrate the capacity of the proposed method to recover not only diffusion filters such as the anisotropic gaussian , but also diffraction filters such as the x example . let us note that , in these synthetic experiments , there is no need to test the long - range assumption on the psf . window with @xmath183 . ( left ) anisotropic gaussian filter . ( right ) x filter.,title="fig:",height=154 ] window with @xmath183 . ( left ) anisotropic gaussian filter . ( right ) x filter.,title="fig:",height=154 ] the measurements are simulated according to ( [ eq : discrete_model_vector ] ) , where additive white gaussian noise @xmath184 is added in order to simulate a realistic scenario . the noise variance ( @xmath47 ) is set such that the blurred signal to noise ratio @xmath185 is equal to 30 db , which corresponds to a realistic bsnr in the actual observations . during the experiments , the wavelet transformation used for the sparse image representation is the redundant daubechies wavelet basis with two vanishing moments and three levels of decomposition @xcite . other mother wavelets and more vanishing moments can be considered , however , due to the dictionary redundancy , the choice does not have a significant impact on the reconstruction results . higher levels can also be considered but the computational time is notably higher and the reconstruction quality does not significantly improve . as discussed in section [ sec : imageprior ] , the wavelet coefficients whose support touches the boundary of the occulting body are not sparse as they spread over all the scales of the wavelet transform . thus , we need to define the set @xmath5 containing the coefficients that are not affected by the disk . to determine this set , we first generate a set of images with constant background that contain disks of radius @xmath178 centered as the celestial body . all the elements inside the disks are set as random values with a standard uniform distribution ( @xmath186 ) . we then compute the wavelet coefficients of this image on the basis @xmath187 . finally , the set @xmath5 is composed by the indexes of the _ detail _ coefficients that are equal to zero . the reconstruction quality of @xmath188 with respect to the true image @xmath189 is measured using the increase in snr ( isnr ) defined as @xmath190 . the reconstruction quality of @xmath191 with respect to the true filter @xmath192 is measured using the reconstruction snr ( rsnr ) , where @xmath193 let us first analyze the behavior of the algorithm when we increase the number of observations @xmath50 used for the reconstruction . table [ tab : synthresults_gauss2 ] presents the reconstruction quality of the results using the blind deconvolution problem for the anisotropic gaussian filter of figure [ fig : syntheticfilters]-(left ) . the results correspond to an average over four trials and they are presented for @xmath50 = 1 , 2 and 3 observations . for @xmath194 we use a warm start by initializing the algorithm with the results obtained for @xmath195 . .reconstruction quality for different number of simulated transit observations using the synthetic anisotropic gaussian filter . [ cols="<,^,^,^,^",options="header " , ] figure [ fig : expresults_ar_euvi ] shows that the estimated non - parametric psf is able to enhance the image and provide more details . similarly to sdo / aia , we compare these results with the ones from the other psfs by taking the ratio between the deconvolved images and the observation ( computing a pixel by pixel division ) . results are depicted in figure [ fig : expresults_ar_euvi_ratio ] . we observe that the non - parametric psf provides similar results to those obtained by the parametric psf given by the euvi_psf.pro procedure . these two psfs are able to provide higher details than the parametric psf given by the euvi_deconvolve.pro procedure . the deconvolution resulting from the combined parametric / non - parametric psfs present a higher correction from the observations than the ones obtained using the other three filters . this corresponds to what was observed for the moon transit images in table [ tab : expresults_moon_diskintensity ] . the figure shows the ratio ( computed using a pixel by pixel division ) between the deconvolved and the observed images along @xmath196 . ( top , best viewed in color ) ratio for filters @xmath197 , @xmath198 , and @xmath199 . ( bottom , best viewed in color ) ratio for filters @xmath200 and @xmath201 . the horizontal axis is in correspondence with the horizontal axis of figure [ fig : expresults_ar_euvi]-(right).,title="fig:",height=132 ] + -(left ) . the figure shows the ratio ( computed using a pixel by pixel division ) between the deconvolved and the observed images along @xmath196 . ( top , best viewed in color ) ratio for filters @xmath197 , @xmath198 , and @xmath199 . ( bottom , best viewed in color ) ratio for filters @xmath200 and @xmath201 . the horizontal axis is in correspondence with the horizontal axis of figure [ fig : expresults_ar_euvi]-(right).,title="fig:",height=132 ] we proposed a blind deconvolution method that allows one to recover the psf of a given astronomical instrument provided the latter captures a celestial body transit ( as observed in figure [ fig : transit ] ) . the proposed non - parametric approach is comparable to previous parametric ones with respect to the quality of the psf core . it presents , however , some limitations in terms of the considered noise model and the estimation of the psf long - range effects . fortunately , the optimization techniques we use in this work are flexible enough to open perspectives of improvement in future works . the incidence of photon flux on a euv telescope is converted to digital numbers ( dn ) through a series of steps , each potentially introducing some noise . the beam of photons impinges the optical system where the psf acts as a blurring operator . simultaneously , a spectral selection is performed on the signal before it reaches the ccd detector . the latter has an heterogeneous response across its surface . finally , the camera electronics convert electrons into dn , adding the read - out noise . the pixels in the resulting image can be modeled as the realization from a random variable @xmath202 , whose noise part can be decomposed into additive , poissonian , and multiplicative degradations . the expectation ( @xmath203 ) and the variance ( @xmath204 ) of @xmath202 then verify : @xmath205 = x \ \ \textrm{and } \ \ { \mathbb{v}}[y ] = \sigma^2 + \beta x + \alpha x^2,\ ] ] where @xmath77 is the standard deviation of the additive component and @xmath206 are parameters . shearer et al . @xcite handled poisson corrupted data through a variance stabilization transform ( vst ) which provides an approximated gaussian noise distributed data . in order to generalize our awgn model to models such as ( [ eq : noise_model ] ) , a vst could be included in the @xmath207-fidelity term in ( [ eq : bid2 ] ) and ( [ eq : nbid ] ) , on both the observations and the convolution result @xcite . this can be done provided that we know the conversion between dn and the photon counts ( _ e.g. _ , from instrument specifications ) . the algorithm described in section [ sec : am ] is adaptable to this stabilized fidelity term through specific proximal operators or gradient descent @xcite . notice , however , that such a stabilization can not be applied only on the observations as a mere preprocessing of the data , while using afterwards other deconvolution methods assuming additive gaussian noise corruption . indeed , the vst being a non - linear process of the form @xmath208 for some @xmath209 @xcite , it breaks the convolutive model ( [ eq : discrete_model_vector ] ) on which those deconvolution methods rely . in other words , the vst of the convolution of @xmath4 by @xmath66 in ( [ eq : discrete_model_vector ] ) is not equivalent to ( and hardly approximable by ) the convolution of the variance stabilized vectors , which breaks the applicability of those techniques . as an alternative to using a vst , the true poisson distribution could also be used to design a specific fidelity term , which is based on the negative log - likelihood of the posterior distribution induced by the observation model , _ i.e. _ , the kl divergence of this distribution @xcite . however , it is unclear how to adapt the method proposed by attouch et al . @xcite to the resulting framework . our paper aimed at preventing strong assumptions on the shape of the central part of the psf . we provide a general deconvolution method for different types of telescopes , the generality of our method being of particular interest for telescopes exhibiting optical aberrations . this strategy differs from the one adopted by most parametric psf estimations . nonetheless , additional convolutive filter regularizations could be included in our scheme by , for instance , promoting the sparsity ( in synthesis or in analysis ) of this filter in an appropriate basis ( _ e.g. _ , wavelet basis ) , or by enforcing a certain decaying law of its amplitude in function of the radial distance . both kind of priors can be expressed as convex costs ( _ e.g. _ , with a @xmath97-norm or a weighted @xmath210-ball constraint ) with closed - form ( or _ simple _ ) proximal operators . these adaptations can thus be integrated in algorithm [ alg : alternatemin ] with additional efforts for limiting the computational time of the more complex deconvolution procedure . moreover , such additional priors can help in enlarging the support @xmath107 where the filter is truly estimated , hence reaching an estimation of the long - range psf . regarding the approximation of the long - range psf impact by a constant , we notice that , instead of estimating the constant @xmath115 in a pre - processing step ( see section [ sec : filterprior ] ) , in future work we could let @xmath115 be a free parameter of the deconvolution problem ( [ eq : bid2 ] ) . the value of @xmath115 can then be optimized jointly with the image and the filter as this does not break the convexity of the sub - problems detailed in section [ sec : am ] . finally , inspired by the works of shearer et al . @xcite , a convolutive combination of a known parametric filter with an unknown non - parametric one was considered in this work to further stabilize the non - convexity of the blind deconvolution problem . however , this construction is limited since , from the convolution theorem , no correction of the parametric psf can be made in the part of its spectrum where it vanishes . future works should therefore consider additive correction of the parametric psf as suggested by poduval et al . @xcite , a modification that algorithm [ alg : alternatemin ] could also include . we have demonstrated how a non - parametric blind deconvolution technique is able to estimate the core of the psf of an optical instrument with high quality . the quality of the estimated psf core is comparable with the one provided by parametric models based on the optical characterization of the imaging instrument . we also demonstrate that , if the parametric psf is incorporated in the acquisition model , the blind deconvolution approach is able to provide a ` corrected ' psf such that most of the apparent emissions inside the disk of a celestial body during a solar transit can be removed . let us note that non - parametric techniques can not outperform the accuracy of parametric methods , however in situations where the telescope imaging model can not be obtained due to some instrument s properties ( _ e.g. _ , picard / sodism @xcite ) , the non - parametric method presents a great advantage . moreover , the proposed method is not specific to a given instrument but can be applied to any optical instrument provided that we have strong knowledge on some image pixels values and their exact location , such as the information available during the transit of the moon or a planet . we have also shown that the use of the proximal alternating minimization technique proposed by attouch et al . @xcite , allows to efficiently solve the non - convex problem of blind deconvolution with theoretical convergence guarantees . furthermore , we show the importance of considering multiple observations for the same filter in order to provide a better conditioning to the filter estimation problem . part of this work is funded by the aoc spw project ( skywin - 6894 ) . lj is supported by the belgian frs - fnrs fund . vd acknowledges support from the belgian federal science policy office through the esa - prodex program , grant no . the authors would like to thank craig deforest and jean - franois hochedez for inspiring discussions and jean - pierre wuelser for having kindly provided us moon trajectory data in secchi / euvi images . thanks are extended to the two anonymous referees for their comments , which substantially improved the quality of the paper . also , we would like to thank the anonymous reviewer who provided the matlab code that computes only the mesh diffraction component for the psf of sdo / aia . in this work , in order to compensate the restriction of the estimated filter to its core , we assume that , for moderate solar intensities , the long - range effect of the psf can be approximated by a constant inside a limited patch . this assumption is compatible with the observations made on the convolution of a solar image with a parametric filter that only contains the long - range effect , _ i.e. _ , a filter built by taking a standard parametric psf containing long - range patterns and setting to zero the central pixels inside a window of size @xmath105 with @xmath211 . for sdo / aia , we use the parametric filter estimated by @xcite . the validation was performed using a @xmath212 level 1 image from the venus transit recorded by the 19.3 nm channel of aia at 00:00:01 ut on june 6th 2012 . these results are presented in figure [ fig : constant_mu_aia ] . , with @xmath211 . ( center ) observed image inside a window containing the disk of venus . ( right ) convolved image inside a window containing the disk of venus . the intensity within the window is almost constant.,title="fig:",height=141 ] , with @xmath211 . ( center ) observed image inside a window containing the disk of venus . ( right ) convolved image inside a window containing the disk of venus . the intensity within the window is almost constant.,title="fig:",height=141 ] , with @xmath211 . ( center ) observed image inside a window containing the disk of venus . ( right ) convolved image inside a window containing the disk of venus . the intensity within the window is almost constant.,title="fig:",height=141 ] for secchi / euvi , we use the parametric filter given by the euvi_psf.pro procedure of _ solarsoft_. the validation was performed using a @xmath213 image from the moon transit recorded by the 17.1 nm channel of euvi at 14:00:00 ut on february 25th 2007 . the convolution results are presented in figure [ fig : constant_mu_euvi ] . , with @xmath211 . ( center ) observed image inside a window containing the lunar disk . ( right ) convolved image inside a window containing the lunar disk . the window intensity is almost constant.,title="fig:",height=141 ] , with @xmath211 . ( center ) observed image inside a window containing the lunar disk . ( right ) convolved image inside a window containing the lunar disk . the window intensity is almost constant.,title="fig:",height=141 ] , with @xmath211 . ( center ) observed image inside a window containing the lunar disk . ( right ) convolved image inside a window containing the lunar disk . the window intensity is almost constant.,title="fig:",height=141 ] the resulting low frequency images in figure [ fig : constant_mu_aia]-(right ) and figure [ fig : constant_mu_euvi]-(right ) validate the use of a constant @xmath115 to approximate the long - range effect . let us note that this is just a first approach to estimate the long - range effect and that more sophisticated methods can be used as explained in section [ sec : discussion ] . algorithm [ alg : alternatemin ] requires an initial value of the image ( @xmath214 ) , the filter ( @xmath215 ) and the regularization parameter ( @xmath216 ) . in this section we describe in details how we proceed with this initialization . for the first value of @xmath124 , the alternated algorithm ( algorithm [ alg : alternatemin ] ) is initialized with the trivial solution , where the image is given by the modified observations @xmath217 and the filter is given by the delta function @xmath218 . for the subsequent iterations on @xmath124 , the image and the filter are initialized using the value of the previous iteration . the regularization parameter has an important role in solving the first step in algorithm [ alg : alternatemin ] , since it determines the trade - off between the image regularization and the fidelity to the observations . several works have studied the choice of this parameter when solving general ill - posed inverse problems . in basis pursuit denoising , @xcite proposed a universal value for @xmath124 given by @xmath219 . since this value increases with the amount of samples ( @xmath172 ) , it tends to provide overly smoothed images . another state - of - the - art choice , proposed by @xcite , is based on the minimization of stein s unbiased risk estimate ( sure ) @xcite . this method provides accurate denoising results , however it requires complete knowledge of the degradation model which makes it unsuitable for our blind deconvolution problem . in this work , we use the bayesian approach proposed by @xcite , to estimate the parameter @xmath124 using the noise characteristics and the probability distribution function of the wavelet coefficients : @xmath220 where @xmath221 is the standard deviation of the wavelet coefficients of the signal to reconstruct . when we assume that they follow a laplacian probability distribution with zero mean , the value of @xmath221 can be estimated as : @xmath222 , where @xmath223 is the matrix containing the image wavelet coefficients inside the set @xmath5 . since the ground truth signal @xmath79 is usually not available , the value of @xmath221 in ( [ eq : rhomap ] ) is determined using the modified observations @xmath217 , _ i.e. _ , @xmath224 , with @xmath225 . since this value is higher than the actual @xmath124 , it will lead to over - regularized images . therefore , we propose to refine the value of @xmath124 by an iterative update based on the discrepancy principle between the actual noise energy in ( [ eq : epsilon ] ) and the energy of the residual image ( [ eq : residual ] ) . by performing a maximum of five updates of the parameter @xmath124 , we assure an appropriate image regularization . the iterative estimation of @xmath124 is summarized in algorithm [ alg : alternatemin_ext ] . _ _ : estimate @xmath226 and @xmath227 using algorithm [ alg : alternatemin ] with @xmath228 , @xmath229 and @xmath230 _ _ : @xmath231 compute @xmath232 using ( [ eq : mrp ] ) _ _ : @xmath233 @xmath234 _ _ : break . return @xmath235 and @xmath236 for the sake of completeness , in this section , we briefly describe the proximal algorithms that we use to solve the two subproblems of algorithm [ alg : alternatemin ] . we refer the reader to @xcite for a comprehensive understanding . proximal algorithms rely on a key element in convex signal analysis , the proximal operator @xmath237 which is uniquely defined for any function @xmath238 , for some @xmath239 @xcite . proximal algorithms allow the minimization of non - smooth functions such as the @xmath97-norm and the indicator functions present in the image and filter subproblems . in the first subproblem , we are interested in finding the image candidate that minimizes @xmath240 a problem that contains a sum of four functions belonging to @xmath241 , for some dimensions @xmath242 . for this , we use the chambolle - pock ( cp ) primal - dual algorithm @xcite , which is commonly used for solving the minimization of two functions in @xmath243 . in order to take into account the sum of more than two functions , the algorithm is extended as in @xcite . if we set @xmath244 for @xmath245 , @xmath246 for @xmath247 , @xmath248 for @xmath249 and @xmath250 for @xmath251 , the cp iterations are : @xmath252\right\},\\ \bar{{\mathbf}u}^{(t+1)}\hspace{-2.5mm}&= 2\,{\mathbf}u^{(t+1 ) } - { \mathbf}u^{(t)},\hspace{7.2 cm } \end{aligned}\ ] ] with @xmath253 tending to a minimizer @xmath254 of ( [ eq : step1 ] ) for @xmath255 . the functions @xmath256 are the convex conjugates of functions @xmath257 , with @xmath258 , and their proximal operators are computed via the proximal operator of @xmath257 in ( [ eq : prox ] ) by means of the conjugation property @xcite : @xmath259 . the step sizes @xmath260 and @xmath261 are adaptively selected using the procedure described in @xcite . in the second subproblem , we are interested in finding the filter candidate that minimizes @xmath262 with @xmath263 a circulant matrix of kernel @xmath264 . this problem has the following shape @xmath265 with @xmath266 and @xmath267 . since @xmath268 is convex and differentiable with gradient @xmath269 , and @xmath270 to @xmath271 belongs to @xmath272 , the problem can be solved using the accelerated proximal gradient ( apg ) method @xcite . the apg iterations are : @xmath273 with @xmath274 tending to a minimizer @xmath275 of ( [ eq : step2 ] ) for @xmath276 . the step size @xmath260 is adaptively selected using the line search of @xcite . as explained in section [ sec : model ] , our blind deconvolution approach is realized in a restricted fov of size @xmath68 . for all our computations , as those realized in algorithm [ alg : alternatemin ] , fast convolution methods exploiting the fft must be performed . therefore , it is important to properly handle the frontiers of our images and avoid both implicit fft frontier periodizations and the influence of unobserved image values integrated by the filter extension . inspired by the work of @xcite , we implement their _ * unknown * boundary conditions _ , where instead of expanding the observation using zero - padding as in , the boundaries are considered unknown in the convolution process and then , by properly selecting the pixels inside the known boundaries , the observed image is obtained . in a nutshell , the unknown boundary conditions method proceeds by considering every convolution in a bigger space obtained by expanding the original one by the radius of the filter in all directions , _ i.e. _ , by a border of width @xmath277 . later , the optimization methods can freely set the values in this border . we only impose that the correct convolution values are those selected in the former domain . mathematically , we consider a larger image @xmath278 and a larger filter @xmath279 , with @xmath280 . the observation is obtained by selecting the pixels inside the boundaries of the convolved image using a selection operator @xmath281 . the problem in ( [ eq : discrete_model_vector ] ) can be rewritten as : @xmath282 the regularized blind deconvolution problem in ( [ eq : bid2 ] ) transforms as follows : @xmath283 with @xmath284 and @xmath285 . the actual image @xmath188 and filter @xmath191 are obtained by applying the operator @xmath286 to both the estimated image @xmath287 and filter @xmath288 , respectively , _ i.e. _ , @xmath289 , @xmath290 . h. attouch , j. bolte , p. redont and a. soubeyran . proximal alternating minimization and projection methods for nonconvex problems : an approach based on the kurdyka - ojasiewicz inequality . , 35(2):438457 , may 2010 . b. anconelli , m. bertero , p. boccacci , m. carbillet and h. lanteri . reduction of boundary effects in multiple image deconvolution with an application to lbt linc - nirvana . , 448(3):12171224 , march 2006 . j. a. p. oliveira , m. a. t. figueiredo , and j. m. bioucas - dias . blind estimation of motion blur parameters for image deconvolution . , 4478:604611 . springer berlin heidelberg , 2007 . 10.1007/978 - 3 - 540 - 72849 - 8_76 . w. d. pence , r. seaman and r. l. white . a comparison of lossless image compression methods and the effects of noise . in astronomical data analysis software and systems asp conference series , vol . 411 , 25 , 2009 .
_ context : _ characterization of instrumental effects in astronomical imaging is important in order to extract accurate physical information from the observations . the measured image in a real optical instrument is usually represented by the convolution of an ideal image with a point spread function ( psf ) . additionally , the image acquisition process is also contaminated by other sources of noise ( read - out , photon - counting ) . the problem of estimating both the psf and a denoised image is called blind deconvolution and is ill - posed . _ aims : _ we propose a blind deconvolution scheme that relies on image regularization . contrarily to most methods presented in the literature , our method does not assume a parametric model of the psf and can thus be applied to any telescope . _ methods : _ our scheme uses a wavelet analysis prior model on the image and weak assumptions on the psf . we use observations from a celestial transit , where the occulting body can be assumed to be a black disk . these constraints allow us to retain meaningful solutions for the filter and the image , eliminating trivial , translated and interchanged solutions . under an additive gaussian noise assumption , they also enforce noise canceling and avoid reconstruction artifacts by promoting the whiteness of the residual between the blurred observations and the cleaned data . _ results : _ our method is applied to synthetic and experimental data . the psf is estimated for the secchi / euvi instrument using the 2007 lunar transit , and for sdo / aia using the 2012 venus transit . results show that the proposed non - parametric blind deconvolution method is able to estimate the core of the psf with a similar quality to parametric methods proposed in the literature . we also show that , if these parametric estimations are incorporated in the acquisition model , the resulting psf outperforms both the parametric and non - parametric methods .
the wheeler - feynman@xciteelectrodynamics developed from the schwarzschild - tetrode - fokker@xcite direct - interaction functional . equations of motion are derived from hamilton s principle for the action integral @xmath5 where the four - vector @xmath6 represents the four - position of particle @xmath7 parametrized by arc - length @xmath8 , double bars indicate quadri - vector modulus @xmath9 and the dot indicates the usual minkowski relativistic scalar product of four - vectors . ( integration is to be carried over the hole particle trajectories , at least formally ) . the above action integral describes an interaction at the advanced and retarded light - cones with an electromagnetic potential given by half the sum of the advanced and retarded linard - wierchert potentials @xcite . wheeler and feynman showed that electromagnetic phenomena can be described by this direct action - at - a - distance theory in complete agreement with maxwell s theory as far as the classical experimental consequences@xcite . this direct - interaction formulation of electrodynamics was developed to avoid the complications of divergent self - interaction , as there is no self - interaction in this theory , and also to eliminate the infinite number of field degrees of freedom of maxwell s theory @xcite . it was a great inspiration of wheeler and feynman in 1945 , that followed a lead of tetrode @xcite and showed that with the extra hypothesis that the electron interacts with a completely absorbing universe , the advanced response of this universe to the electron s retarded field arrives _ at the present time of the electron _ and is equivalent to the local instantaneous self - interaction of the lorentz - dirac theory@xcite . the action - at - a - distance theory is also symmetric under time reversal , as the fokker action includes both advanced and retarded interactions . dissipation in this time - reversible theory becomes a matter of statistical mechanics of absorption@xcite . the area of wheeler - feynman electrodynamics has been progressing slowly but steadily since 1945 : quantization was achieved by use of the feynman path integral technique and the effect of spontaneous emission was successfully described in terms of interaction with the future absorber , in agreement with quantum electrodynamics@xcite . it was also shown that it is possible to avoid the usual divergencies associated with quantum electrodynamics by use of proper cosmological boundary conditions@xcite . as far as understanding of the dynamics governed by the equations of motion , the state of the art is as follows : the exact circular orbit solution to the attractive two - body problem was proposed in 1946@xcite and rediscovered by schild in 1962@xcite . the 1-dimensional symmetric two - electron scattering is a special case where the equations of motion simplify a lot and it has been studied by many authors , both analytically and numerically @xcite . in this very special case the initial value functional problem surprisingly requires much less than an arbitrary initial function to determine a solution manifold with the extra condition of bounded manifold for all times . it was shown that the solution is uniquely determined by the interelectronic distance at the turning point if this distance is large enough ( this minimum distance curiously evaluates to 0.49 bohr radii by the action - at - a - distance theory@xcite , much larger than about one classical electronic radius that one would naively guess ) . as a result of this theorem , there is a single continuous parameter ( the positive energy ) describing the unique non - runaway symmetric orbit at that given positive energy . the noether s four - constant of motion derived from the fokker lagrangian involves an integral over the past history@xcite . for example in the case of a hydrogen atom this four - momentum constant evaluates to @xcite @xmath10 where @xmath11 represents the derivative of the delta function@xcite . notice that because of this delta function , only finite portions of the trajectory are involved : actually an extent of length @xmath12 approximately . this non - local constant will behave very differently from the local coulombian energy , that is known to confine orbits of a negative energy within a maximum separation distance . in the case where the particles acquire a large separation ( unbound state ) , the hole past history is involved ( @xmath13 ) in the determination of the non - local energy constant . as regards the mathematical structure of the equations of motion , for the case of a two - electron atom the acceleration of electron 1 is given by@xcite @xmath14 where @xmath15 and @xmath16 are the electronic charge and mass , @xmath17 @xmath18and @xmath19 and @xmath20 are the total electric and magnetic fields produced by electron 2 and the nucleus . in the action - at - a - distance theory these fields are given by the average of the retarded and advanced linard - wiechert fields , calculated with the instantaneous position of the stationary nucleus and the retarded and advanced positions of electron 2 at the times @xmath21 , which is defined by the implicit condition @xmath22 where the minus and plus signs are the conditions for the retarded and advanced times respectively . the partial electric fields of electron 2 acting on electron 1 at time @xmath23are @xcite @xmath24 , \end{aligned}\ ] ] where @xmath25 , @xmath26 and @xmath27 is the speed of light . the advanced field @xmath28 is obtained from the above expression by replacing @xmath29 by @xmath30 and @xmath27 by @xmath31 the partial magnetic fields of electron 2 are @xmath32 where the @xmath33 is to ensure an outgoing poynting vector @xmath34 for the retarded fields and an incoming poynting vector for the advanced fields . the total electric field in equation ( [ motion1 ] ) must include also the instantaneous coulomb electric field of the stationary nucleus . equation ( [ motion1 ] ) can suggest a paradox about causality , as the force depends on the future of particle 2 . in the following , and to finish this introduction , we show that equation ( [ motion1 ] ) , when written properly , becomes a functional differential equation with _ delayed argument only _ , as first observed in@xcite . to outline the essentials of the explanation , let us first ignore the field of the nucleus and take the nonrelativistic limit of ( [ motion1 ] ) ( @xmath35 ) in this approximation the electric field @xmath19 entering in equation ( [ motion1 ] ) evaluates to @xmath36 . then we note that one can use equation ( [ motion1 ] ) as an equation of motion for _ particle 2 , _ by solving the rearranged form of ( [ motion1 ] ) , @xmath37 for the most advanced acceleration of particle 2 , @xmath38 in the above form it is clear that the right hand side involves only functions evaluated at times prior to the most advanced time , defined by @xmath39 , and no further advanced information is necessary , eliminating the ghost of dependence on the future . in the same way , the causal equation of motion of particle 1 is to be produced from the equation for particle 2 by solving for the most advanced acceleration of particle 1 . for the special case of 1-dimensional motion of two electrons , @xmath40 depends only on the advanced velocity , and ( [ causal ] ) can easily be solved for this advanced velocity as a function of the past history . in the 3-dimensional case there is an extra complexity , as the acceleration appears in the linard - wiechert partial field @xmath28 in the form @xmath41 . the bad news is that the component of the acceleration along the advanced normal can not be solved for from the value of the double - vector - product only . because of this degeneracy , equation ( [ causal ] ) is an algebraic - differential equation , and the null direction of the left hand side of ( [ causal ] ) is a constraint to be satisfied by the right hand side ( the scalar product with @xmath42 must vanish ) . the numerically correct way to integrate this type of equation is by use of the modern integrators for algebraic - differential equations like dassl @xcite adapted for retarded equations ( which has never been done yet ) or by dealing directly with the algebraic constraint @xcite . according to the standard classification of g. a. kamenskii@xcite , equation ( [ causal ] ) belongs to the class of differential - difference equations of neutral type . even though more complex , the motion is still causally determined by the past trajectory , as we wanted to demonstrate , the price being an algebraic neutral delay equation . as far as initial conditions go , the general theory on delay equations @xcite tells us that we need to provide an initial @xmath43 function describing the position of particle 2 from @xmath44 up to the initial instant @xmath45 . the information on particle 1 needed is also to be provided over twice the retardation lag seen by particle 1 . this is a short piece of trajectory for bound nonrelativistic atomic orbits , but for a ionized state or a runaway orbit this can be the whole past history ! unless further simplifications or conditions are added , this is the generic problem at hand . the 3-dimensional cases of atomic interest ( e.g. helium ) have never been studied , and they are more complex than the 1-d scattering because one can have negative energy bound states for example . most relevant for physics is the question of the conditions for the existence of a bounded manifold solution , which still needs to be understood in the general case ( it would be very curious if they turned out to be a discrete set of negative energies ) . the only existing analytical result in the 3-dimensional case is the linear stability of the schonberg - schild circular orbits@xcite , resulting in an infinite number of unstable solutions to the characteristic equation . the numerical treatment ofthe exact neutral equations displays instabilities and is generally difficult . in the following we resort to the darwin approximation not as much as a mathematical approximation to the action - at - a - distance electrodynamics , but as a physical approximation of lorentz - invariant dynamics in the atomic ( shallow ) energy range . to introduce our numerical calculations , we start from the scale - invariant coulomb limit of the tetrode - fokker - wheeler - feynman interaction : let @xmath15 and @xmath16 be the electronic charge and mass respectively and @xmath46 the nuclear charge of our two - electron atom , which in this work is assumed to have an infinite mass . all our numerical work uses a scaling which exploits the scale invariance of the coulomb dynamics : given a negative energy , there is a unique circular orbit at that energy with frequency @xmath47 and radius@xmath48 related by @xmath49 . we scale distance , momentum , time and energy as @xmath50 , @xmath51 , @xmath52 and @xmath53 , respectively . in these scaled units , the coulomb dynamics of the two - electron atom is described by the scaled hamiltonian @xmath54 where @xmath55(single bars represent euclidean modulus ) and @xmath56 . for a generic non - circular orbit , @xmath57plays the role of a scale parameter , and we recover the value of the energy in ergs through @xmath58 . notice that @xmath57 does not appear in the scaled hamiltonian , which is the scale invariance property . from the scaled frequency @xmath59 and scaled angular momentum @xmath60 we can recover the actual values in cgs units by the formulas @xmath61 the only other analytic constant of the coulomb dynamics , besides the energy ( [ hamitwo ] ) is the total angular momentum , and this dynamics in chaotic and displays arnold diffusion , as proved in @xcite for a similar three - body system . the numerical calculations were performed using a 9th - order runge - kutta embedded integrator pair@xcite . we chose the embedded error per step to be @xmath62 , and after ten million time units of integration the percentage changes in energy and total angular momentum were less than @xmath63 . as a numerical precaution we performed the numerical calculations using the double kustanheimo coordinate transformation to regularize single collisions with the nucleus@xcite . as these alone are not enough for faithful integration , we checked that there was never a triple collision , as the minimum inter - electronic distance was about @xmath64 units while the minimum distance to the nucleus was @xmath65 units for all the orbits considered in this work . we also checked that along stable non - ionizing orbits we can integrate forward up to fifty thousand time units , reverse the integration , go backwards another fifty thousand units and recover the initial condition with a percentile error of @xmath66 . for longer times this precision of back and forth integration degenerates rapidly , which is due to the combined effect of numerical truncation and stochasticity . the question of how far in time the numerical trajectories approximate shadowing trajectories in the present system is far from trivial @xcite , but we assume it to be a time at least of the order of these one hundred thousand units . ( energy conservation of one part in a million is achieved for much longer times , even one billion time units ) . the study of orbits of a two - electron atom was greatly stimulated by the recent interest in semiclassical quantization , and these studies discovered two types of stable zero - angular - momentum periodic orbits for helium ( @xmath67 ) : the langmuir orbit and the frozen - planet orbit @xcite . a detailed study of the non - ionizing orbits of coulombian helium was initiated in reference @xcite for plane orbits , and we describe some of their results below . there are basically two types of non - ionizing orbits : symmetric if @xmath68 for all times and asymmetric if @xmath69 generically . symmetric orbits are produced by symmetric initial conditions like for example @xmath70 and @xmath71 or @xmath70 and @xmath72 with @xmath73 @xcite because ( [ hamitwo ] ) is symmetric under particle exchange , these orbits satisfy @xmath68 at all times , and therefore can not ionize if @xmath74 ( both electrons would have to ionize at the same time , which is impossible at negative energies ) . for example the double - elliptical orbits ( two equal ellipses symmetrically displaced along the x - axis ) discussed in @xcite are in this class . double - elliptical orbits are known to be unstable @xcite and we find that they ionize in about one hundred turns because of the numerical truncation error . most symmetric plane orbits are very unstable to asymmetric perturbations , with the exception of the langmuir orbit for a small range of @xmath75 values around @xmath67 @xcite@xmath76 the simplest way to produce an asymmetric non - ionizing plane orbit is from the initial condition @xmath77 , @xmath78 , @xmath79 , @xmath80 , as suggested in @xcite . in figure 1 we show the electronic trajectories for the first three hundred scaled time units along a two - dimensional non - ionizing orbit of @xmath81 with @xmath82 and @xmath83 in the above defined condition . we used a numerical refining procedure to finely adjust @xmath84as to maximize the non - ionizing time and this condition of figure 1 does not ionize for one million time units . the orbit survives that far only for a very sharp band of values of @xmath85 , other neighboring values producing quick ionization . this orbit was named double - ring torus in @xcite . the other possible type of non - ionizing orbit resulting from the above initial condition , depending on @xmath86 is what was named braiding torus in reference @xcite , with both electrons orbiting within the same region . a search over @xmath87 was conducted in @xcite , and it was found that most values of @xmath87 produce quick ionization except for a zero - measure set of @xmath87 values where braiding tori or double ring orbits are found . this suggests the general result that non - ionizing orbits are rare in phase space . to search for general tridimensional non - ionizing orbits in phase space , it is convenient to introduce canonical coordinates @xmath88 and @xmath89 @xmath90 initial conditions with @xmath91 @xmath92 describe double - elliptical orbits ( and circular as a special case ) . to generate an elliptical initial condition , we exploit the scale invariance and set the energy to minus one . it is easy to check that elliptical orbits of the hamiltonian ( [ hamitwo ] ) with an energy of minus one must have a total angular momentum of magnitude ranging from zero to two . to exploit the rotational invariance of ( [ hamitwo ] ) , we can choose the plane defined at @xmath91 @xmath92 by the angular momentum @xmath93 @xmath94 to be the @xmath95 plane . on this @xmath95 plane a single number @xmath96 ( the angular momentum ) , determines completely the elliptical orbit . the next step in producing a generic orbit is to add all possible perturbations along @xmath97 and @xmath98 to the chosen elliptical orbit . these are six directions and once we are looking for bound oscillatory orbits , we can choose @xmath99 , once @xmath100 has to cross the @xmath95 plane at some point . these are five numbers to vary and plus the angular momentum of the elliptical orbit it totals six parameters . our numerical search procedure consists in varying these six parameters over a fine grid , integrating every single initial condition until the distance from one electron to the nucleus is greater than twenty units , which is our ionization criterion . this criterion fails if the orbit has a very low angular momentum because these can go far away from the nucleus and come back , and therefore our search possibly misses low - angular - momentum non - ionizing orbits . as the majority of the initial conditions ionize very quickly , this search procedure is reasonably fast . we first perform a coarse search for ionization times above one thousand units and then refine in the neighborhood of each surviving condition to get conditions that do not ionize after one million time units . using the above numerical search procedure we found the tridimensional non - ionizing initial condition of figure 2 for helium , a tridimensional double - ring orbit generated by the initial condition @xmath101 which does not ionize before ten million turns . ( after the search and refinement , we scaled this orbit s energy to minus one , for later convenience ) . we also found the non - ionizing orbit orbit of figure 3 for h - minus ( @xmath102 ) , a tridimensional orbit generated by the condition @xmath103 which does not ionize before one million turns ( coulombian energy of this condition is also minus one ) . this last orbit is fragile and numerically harder to find : as the first electron has an orbit very close to the positive @xmath102 charge , there remains only a dipole field to bind the second electron . as the outer electron is much slower in the scaled units , we had to plot the first @xmath104 time units of evolution to display the generic features of the trajectory . non - ionizing orbits of @xmath105 are very rare in phase space , which is reminiscent of the quantum counterpart , as the @xmath106 ion is known to have only one quantum bound state at @xmath107 , very close to the ionization threshold @xmath108@xcite . one remarkable fact about these non - ionizing orbits is that they all have a very sharp fourier transform . this property makes them approximately quasi - periodic orbits . for example in figure 4 we plot the fast fourier transform of the orbit of figure 2 , performed using @xmath109 points . ( it seems that there are at least two basic frequencies in the resonance structure of figure 4 ) . even though these orbits look like quasi - periodic tori , there seems to be a thin stochastic tube surrounding each orbit , as evidenced by a small positive maximum lyapunov exponent . we calculated numerically this maximum lyapunov exponent by doubling the integration times up to @xmath110 and found that the exponent initially decreases but then saturates to a value of about @xmath111 for the orbits of figures 1 , 2 and 3 . the gravitational three - body problem has recently been proved to display arnold diffusion@xcite , and this numeriacally calculated positive lyapunov exponent suggests that the same is true for the two - electron coulombian atom . the numerical integrations in this section are performed using the darwin approximation . the darwin equations of motion are a @xmath112 perturbation of the coulomb dynamics , of size @xmath113 for atomic energies . in the scaled units of section ii the darwin hamiltonian is the following @xmath112 perturbation of hamiltonian ( [ hamitwo ] ) @xmath114 \nonumber \\ & & -\frac{\beta ^{2}}{8}[|\vec{p}_{1}|^{4}+|\vec{p}_{2}|^{4 } ] , \label{darwin}\end{aligned}\ ] ] where @xmath115 . the second line represents the biot - savart magnetic interaction plus the first relativistic correction to the static electric field and the last line describes the relativistic mass correction . notice that these are both proportional to the small parameter @xmath112 , which makes them a small scale - dependent perturbation on the scale invariant coulomb hamiltonian ( first line ) . it is possible to regularize the darwin equations with the same double - kustanheimo transformation@xcite , only that here one needs to define the regularized time using the higher powers @xmath116 , instead of the lower powers @xmath117 used to regularize the coulomb equations@xcite . the main question we address numerically in this section is the dependence of the stability of a non - ionizing orbit with the energy scale of the orbit . here we use the word stability to mean ionization - stability : we call an initial condition ionization - stable if any small perturbation of it produces another non - ionizing orbit . the scale - dependent darwin terms ( of size @xmath118 produce significant deviations from the coulomb dynamics only in a time - scale of order @xmath119 which we find numerically to be the typical time for a non - ionizing coulombian initial condition to ionize along the darwin vector field . this poses a numerical difficulty if @xmath57 is too small because one has to integrate the orbit for very long times to investigate the stability . it turns out that ionization - stable orbits can be found at larger values of @xmath57 for larger values of @xmath120 . here the dynamical stability mechanism is reminiscent of quantum atomic physics , where the values of @xmath57 vary with the nuclear charge as @xmath121 . large values of @xmath120 facilitate the numerical procedure and in the following we present the numerical investigation of the stability of non - ionizing orbits starting from the large @xmath120 case . let us start with the @xmath122 calcium ion two - electron system along the non - ionizing orbit of figure 1 by fixing @xmath82 and @xmath123 in the condition defined in section ii . to test the stability of the orbit at each value of @xmath57 we add a random perturbation of average size @xmath124 to the initial condition and integrate the darwin dynamics until either we find ionization or the time of integration is greater than @xmath125 time units we repeat this for at least twelve randomly chosen perturbations ( because of the twelve degrees of freedom ) and the minimum time to ionization is plotted in figure 5 as a function of @xmath57 . it can be seen that only for a narrow set of values around @xmath126 this minimum time to ionization was greater than @xmath127 . or the other values it decreases rapidly to a value of about @xmath128 . one could argue that for the other values of @xmath57 the non - ionizing initial condition has shifted away from the @xmath83 initial condition and this being the reason that our orbit ionized . to test this , we fixed @xmath57 at a bad value for example @xmath129 and varied the plane initial condition in the neighborhood of this condition of figure 1 . we found that the minimum time to ionization was always about @xmath128 ( also the maximum time before ionization was about @xmath128 ) . we also searched in a bigger neighborhood , of size proportional to @xmath130 . this suggests the interpretation that for the special resonant value of @xmath131 the net diffusive effect of the scale - dependent term vanishes , allowing a non - ionizing perturbed manifold . in order to have a direct interpretation ( in atomic units ) of the scale parameter @xmath130 , it is convenient to scale to minus one the energy of the initial condition of figure 1 ( by exployting the coulombian scale invariance ) . after this , the energy of the orbit in ergs evaluates to @xmath132 , and for @xmath131 this is approximately @xmath133 atomic units . the total angular momentum of this orbit is @xmath134 . this orbit s energy is above the ionization continuum of the ion , @xmath135 atomic units , but it is still in the quantum range . it serves nevertheless to demonstrate that this dynamical system might exhibit non - ionizing stable orbits only at very sharply defined energy values . for the orbits of figures 2 and 3 , the above procedure becomes prohibitively slow , as the value of @xmath130 are much smaller and one must integrate for very long times , much beyond the estimated shadowing time . to partially overcome this we used a larger amplitude random perturbation ( of average size @xmath136 ) , to produce faster ionization . the drawback with this is that the minimum ionization time does not show pronounced peaks , only the average ionization time still showing a signature of scale dependence . in figure 6 we show this average time for the orbit of figure 3 . this property of sharply defined energies can possibly be found for the lower - lying energies below the ionization threshold as well . these orbits would involve configurations where the electrons come very close to the nucleus and acquire a large velocity . even though our integrator is regularized , the correct physical electronic repulsion is greatly amplified when one electron has a relativistic velocity and the darwin approximation can not describe the physics then . actually , it is known that the darwin interaction can produce unphysical effects when pushed to relativistic energies@xcite . we therefore do not expect to find these low - lying atomic energy scales with the present darwin approximation and shall be contempt with these interesting result already . for the same reason given above , we do not study here the frozen - planet periodic orbit ( the two electrons performing one - dimensional periodic motion on the same side of the nucleus , with the inner electron rebounding from the origin , an artifact of regularization ) . the main problem being the failure of the darwin approximation , as the inner particle goes to the speed of light@xcite . the correct relativistic dynamics can actually produce a new _ physical _ inner turning point very close to the origin but not _ at _ the origin as the regularized motion , and we discuss elsewhere@xcite . last , we consider the non - ionizing symmetric periodic orbit called the langmuir orbit , where the two electrons perform symmetric bending motion shaped approximately like a semi - circle@xcite . for the coulomb two - electron atom with @xmath67 this orbit was found to have a zero maximum lyapunov exponent@xcite . the orbit is therefore neutrally stable , which is the best one can expect from a periodic orbit of a hamiltonian vector field . ( absolute stability violates the symplectic symmetry , which says that to every stability exponent @xmath137 one should have a @xmath138 exponent ) . it is a simple matter to obtain the langmuir - like orbit for the darwin hamiltonian at any given value of @xmath57 : all it takes is a little adjusting in the neighborhood of the coulombian langmuir condition . we attempted to investigate numerically any scale - dependent diffusion away from this darwin - langmuir condition for @xmath57 in the atomic range , but again the numerics is prohibitively slow at the time of writing this work . the simplified dynamical mechanism behind resonant non - ionization seems to go intuitively as follows : the peculiar scale - invariant coulomb dynamics determines the non - ionizing orbits within narrow stochastic tubes . the next step is the action of the small scale - dependent relativistic corrections that produce a slow diffusion of the orbit out of the thin tube in a time of the order of @xmath139 . after this , quick ionization follows . only at very special resonant values of @xmath57 the relativistic terms leave the orbit within the tube , a resonant effect that depends on @xmath140 , fixing the energy scale . in the literature , the escape to infinity from simpler to understand two - degree - of - freedom systems has been attributed to cantori , which , as is well known , can trap chaotic orbits near regular regions for extremely long times@xcite . in the present larger dimensional case it appears that resonances are also controlling the escape to infinity of one electron by the existence of extra resonant constants of motion@xcite . this seems to be in agreement with the numerical results of very sharp peaks for the minimum ionization time . we have tried to concentrate on the physics described by this combination of chaotic dynamics on a two - electron atom with inclusion of relativistic correction , while discussing this highly nontrivial result of nonlinear dynamics . in references @xcite we noticed that a simple resonant normal form criterion gives a surprisingly good prediction for the discrete atomic energy levels of helium . the resonant structure was calculated using the darwin interaction ( [ darwin ] ) , which is the low - velocity approximation to both maxwell s @xcite and wheeler - feynmans@xcite electrodynamics . as we saw in section ii , the coulombian non - ionizing orbits are far from circular , and these orbits would radiate even in dipole according to the time - irreversible maxwell s electrodynamics ( circular orbits radiate only in quadrupole but are linearly unstable ) . it becomes then clear that the heuristic results of @xcite can only have a physical meaning in the context of a time - reversible theory ( as the action - at - a - distance electrodynamics for example ) . the combination of chaotic dynamics with relativistic invariance has never been explored numerically , and most known lorentz - invariant dynamical systems are for one particle and possess trivially integrable dynamics . the situation gets unexpectedly much more complicated for more than one particle ( apart from the trivial non - interacting many - particle system ) : due to the famous no - interaction theorem@xcite , the relativistic description of two directly interacting particles is impossible within the hamiltonian formalism and its set of ten canonical generators for the poincare group @xcite . description of interacting particles is possible only in the context of constraint dynamics , with eleven canonical generators and with the dirac bracket replacing the poisson bracket . for example the relativistic action - at - a - distance equations for two interacting electrons are non - local and possess only infinite - dimensional constrained hamiltonian representations@xcite . the interested reader should consult some recently found two - body direct - interaction relativistic lagrangian dynamical systems@xcite as well as the constraint - dynamics direct - interaction models recently used in chromodynamics and two - body dirac equations@xcite . the nonlinear dynamics of these models could display interesting and so far inexplored dynamical behaviour . it would be natural to wonder if one can find an analogous scale - dependent dynamics for a dynamical system describing the hydrogen atom , apparently the simplest example of lorentz - invariant two - body relativistic dynamics of atomic interest . it turns out that hydrogen is not simpler than helium at all , but it appears to us that there is an essential difference which has actually made the interesting dynamics of a two - electron atom amenable to study already within the darwin approximation : in a two - electron atom orbits with a negative energy can ionize , while in hydrogen this might be possible only if one includes all orders of the relativistic action - at - a - distance interaction . ( as we saw in section i , the noether s energy constant involves a segment of the past trajectory , and a negative value does not forbid ionization ) . ionization with a negative energy would be impossible for hydrogen within the darwin approximation ( unless the electron goes to the speed of light ) . this is indication that in hydrogen the essential physics described by the action - at - a - distance electrodynamics is of non - perturbative character . the paradoxical result of the infinite linear instability of circular orbits in atomic hydrogen@xcite is another warning of this non - perturbative dynamics . the author acknowledges the support of fapesp , proc . 96/06479 - 9 and cnpq , proc . 301243/94 - 8(nv ) . s. bleher , c. grebogi , e. ott and r. brown , _ phys . a _ , * 38 * , 930 ( 1988 ) , f. t. arecchi , r. badii and a. politi,_phys . rev . a_. * 32 * , 402 , ( 1985 ) , f. c. moon and g .- x.li , _ lett . _ * 55 * , 1439 ( 1985 ) , e. g. gwinn and r. m. westervelt , _ phys lett . _ * 54 * , 1613 ( 1985 ) , c. grebogi , s. mcdonald , e. ott and j. yorke , _ phys . _ * 99a * , 415 , ( 1983 ) . r. n. hill , _ relativistic action at a distance : classical and quantum aspects _ , proceedings , barcelona , spain 1981 , edited by j. llosa , lecture notes in physics * 162 * , 104 , ( springer , new york 1982 ) . k. richter and d. wintgen , _ j. phys . b * 23 * , _ l197 ( 1990 ) , k. richter and d. wintgen _ phys . lett . _ * 65 * , 1965 ( 1990 ) , k. richter , g. tanner and d. wintgen , _ phys . a _ * 48 * , 4182 ( 1993 ) , d. wintgen , a. brgers , k. richter and g. tanner,._progress of theoretical physics supplement * 116 * , 121 ( 1994 ) . _ d. g. currie , t. f. jordan and e. c. g. sudarshan , _ rev . of mod phys _ , * 35 * , 350 ( 1963 ) , also in _ the theory of action - at - a - distance in relativistic particle dynamics , _ edited by edward h. kerner , ( gordon and breach , new york 1972 ) .
we study numerically the dynamical system of a two - electron atom with the darwin interaction as a model to investigate scale - dependent effects of the relativistic action - at - a - distance electrodynamics . this dynamical system consists of a small perturbation of the coulomb dynamics for energies in the atomic range . the key properties of the coulomb dynamics are : ( i ) a peculiar mixed - type phase space with sparse families of stable non - ionizing orbits and ( ii ) scale - invariance symmetry , with all orbits defined by an arbitrary scale parameter . the combination of this peculiar chaotic dynamics ( ( i ) and ( ii ) ) , with the scale - dependent relativistic corrections ( darwin interaction ) generates the phenomenon of scale - dependent stability : we find numerical evidence that stable non - ionizing orbits can exist only for a discrete set of resonant energies . the fourier transform of these non - ionizing orbits is a set of sharp frequencies . the energies and sharp frequencies of the non - ionizing orbits we study are in the quantum atomic range . the coulomb dynamical system of the helium atom is a very peculiar chaotic system that exhibits arnold diffusion@xcite , and with a typical trajectory having an infinity of possible time - asymptotic final states . for example , almost all negative - energy trajectories of coulombian helium display the generic phenomenon of ionization , namely , the ejection of one electron@xcite . several nonlinear dynamical systems share this property of having more than one time - asymptotic final state , with the respective basins for each outcome having a complicated structure in initial condition space@xcite . the numerical work on this paper is based on stable coulombian orbits of a two - electron atom that do not ionize for several millions of turns of one electron around the nucleus . it is a property of the coulomb dynamics of a two - electron atom that most initial conditions with a negative energy ionize very quickly in about 20 turns@xcite . then there are the very special initial conditions that do not ionize due to a precise phase balance between the two electrons . these rare non - ionizing orbits are defined very sharply in phase space and were first studied in reference @xcite for plane orbits . here we also develop a numerical procedure to search for non - ionizing orbits among a large number of possible tridimensional initial conditions . the coulomb hamiltonian exhibits the scale invariance degeneracy : if we scale time and space as @xmath0 , @xmath1 , for @xmath2 , the equations of motion are left invariant . for this reason , the behavior of the coulomb dynamics is the same in all scales , a degeneracy which is broken by the relativistic effects of electrodynamics . the phenomenon of breaking the scale invariance in electrodynamics was explored analytically in @xcite for the darwin interaction , which is the low - velocity approximation to the wheeler - feynman action - at - a - distance electrodynamics @xcite . it was found in @xcite that a simple resonant normal form approximation theory predicts a discrete set of quantized scales very close to the quantum atomic energies . using these preliminary findings as guide , we present a numerical investigation of the stability of non - ionizing orbits for the darwin dynamics and its dependence on the energy scale . it turns out that for energies of atomic interest , the darwin equations of motion approximate the coulomb equations plus a perturbation of size @xmath3 with @xmath4 . therefore , non - ionizing stable orbits of the darwin dynamics should exist in the neighborhood of non - ionizing stable coulombian orbits if the perturbation does not force ionization . for these , our numerical results with the darwin dynamics indicate that the non - ionizing property plus stabilty require sharply defined discrete energies . the darwin interaction is not exactly a lorentz invariant interaction@xcite , so we study it as an approximation to the relativistic action - at - a - distance electrodynamics , for the sake of including the present approach into an underlying physical theory . maxwell s theory would seem to be the natural candidate for the comprehensive physical theory , but it lacks time - reversibility and dipolar dissipation would forbid the orbits studied in this paper . there is also the choice of other more recent lorentz - invariant lagrangian@xcite systems and constrained hamiltonian dynamical systems@xcite , whose exact forms are actually more amenable to numerical treatment than the wheeler - feynman electrodynamics , but we shall not consider them here . the interested reader should consult reference @xcite , where a covariant approximation to wheeler - feynman electrodynamics is attempted by the two - body todorov equation of constraint dynamics . this paper is organized as follows : in section i we review the state of the art of the time - reversible action - at - a - distance electrodynamics , and if the reader wants to skip this part the rest of the paper makes full sense as a nonlinear dynamics study , except for the discussion at the end . in section ii we describe the numerical calculations with the coulomb limit of the darwin interaction , find some non - ionizing orbits and their fourier transforms . in section iii we include the scale dependent darwin terms and investigate the possibility of stable non - ionizing orbits . in section iv we put the conclusions and discussion .
consider @xmath0 equipped with coordinates @xmath1 and its standard symplectic form @xmath2 let @xmath3 denote the ball of radius @xmath4 in @xmath5 . gromov s nonsqueezing theorem states that there is no symplectic embedding of @xmath6 into @xmath7 for @xmath8 , @xcite . the following analogous problem remains open . * question 1 . * _ what , if any , is the smallest value of @xmath9 such that there exists a symplectic embedding of @xmath10 into @xmath11 ? _ prior to the current paper , the most that could be said is that if such an emebedding exists , then @xmath4 must be at least @xmath12 . this bound is implied by the second ekeland - hofer capacity from @xcite . in @xcite , l. guth constructs new symplectic embeddings of polydiscs which represent a major breakthrough in our understanding of the following _ bounded _ version of question 1 . * question 2 . * _ what , if any , is the smallest value of @xmath9 such that there exists a symplectic embedding of @xmath13 into @xmath11 for arbitrarily large @xmath14 ? _ among other things , guth s work settles the existence issue for this question . exists . ] in this setting , the second ekeland - hofer capacity again implies that @xmath4 must be at least @xmath12 . the following improvement of this bound is the main result of this paper . [ thm ] for any @xmath15 there are no symplectic embeddings of @xmath13 into @xmath16 when @xmath17 is sufficiently large . in fact we prove a slightly stronger result . for convenience , identify @xmath5 with @xmath18 using the complex coordinates @xmath19 . let @xmath20 and denote by @xmath21 the complex projective plane equipped with the symplectic form @xmath22 where @xmath23 is the standard fubini - study symplectic form . [ thm2 ] for any @xmath15 there are no symplectic embeddings of @xmath24 into @xmath25 when @xmath17 is sufficiently large . * a brief outline of the proof . * as in @xcite , theorem [ thm2 ] is proved via an existence theorem for certain holomorphic curves . suppose that for any @xmath14 there is a symplectic embedding @xmath26 of @xmath27 into @xmath25 . it suffices to show that for every positive integer @xmath28 , there exists a holomorphic plane of degree @xmath29 in the ( negative ) symplectic completion of @xmath30 whose negative end covers the shortest simple reeb orbit on the boundary of @xmath31 a total of @xmath32 times . in particular , the symplectic area of such a curve is both positive and equal to @xmath33 and so the existence of these curves implies that @xmath34 for all @xmath35 . to prove this existence result we use a cobordism argument to reduce it to an equivalent problem for curves in a four dimensional symplectic manifold . starting with well known results concerning holomorphic spheres of degree @xmath29 in @xmath36 , we then settle this alternative existence problem using automatic regularity theorems , the compactness theorem for splittings from @xcite , and several new techniques , many of which involve families of holomorphic curves with varying point constraints . ( in fact , we prove the existence problem in dimension four first , and then we apply it in the manner described above . ) * sharpness . * we also prove that theorem [ thm ] is sharp in the following sense . [ embed2 ] for any @xmath37 there exist symplectic embeddings of @xmath13 into @xmath38 for all @xmath14 . the proof of theorem [ embed2 ] involves a refinement of guth s embedding procedure from @xcite . we do not know whether there exists a symplectic embedding of @xmath13 into @xmath39 for all large @xmath17 . our approach also allows us to settle the following related problem . * question 3 . * _ what are the smallest value of @xmath40 for which there are symplectic embeddings of @xmath13 into @xmath41 for all @xmath14 ? _ more precisely , by compactifyng @xmath42 to @xmath43 , and replacing the study of holomorphic curves of high degree , say @xmath29 , in @xmath36 by curves of degree @xmath44 in @xmath45 , an identical argument to the one used to prove theorem [ thm ] yields the following result . [ eh ] if @xmath46 then there are no symplectic embeddings of @xmath13 into @xmath47 when @xmath17 is sufficiently large . this restriction is again stronger than those imposed by the ekeland - hofer capacities which imply that no such embeddings exist , for large enough @xmath14 , when @xmath48 . moreover , guth s embedding results again imply ( this time directly ) that theorem [ eh ] is sharp in the following sense . ( @xcite)[embed ] for any @xmath49 there exist symplectic embeddings of @xmath13 into @xmath50 for all @xmath14 . the next section contains some background material and the proof of the crucial existence result , theorem [ key ] , which concerns special holomorphic curves in certain four dimensional symplectic manifolds . theorem [ thm2 ] is then proved in section @xmath51 . our proof of theorem [ eh ] is completely similar and is thus omitted . in section @xmath52 , we construct the embeddings of theorems [ embed2 ] and [ embed ] . the authors would like to thank dusa mcduff for some helpful comments and suggestions . throughout this section we fix a real number @xmath53 and a positive integer @xmath29 . let @xmath54 be the space of smooth almost - complex structures on @xmath36 which are tame with respect to the symplectic form @xmath23 ( and hence @xmath22 ) . for a fixed @xmath55 in @xmath54 and an ordered collection of points @xmath56 in @xmath36 , we consider the moduli space @xmath57 defined as @xmath58 = dl\big\ } / g,\ ] ] where @xmath59 is an @xmath60-tuple of pairwise distinct points in @xmath61 , @xmath62 is the class of a complex line in @xmath36 and @xmath63 is the reparameterization group @xmath64 . as is well known , see for example section 7.4 of @xcite , for suitably generic @xmath55 the space @xmath57 is a smooth orientable manifold of dimension @xmath65 we will also make use of the following refinement of this statement . [ number ] for generic @xmath55 in @xmath54 , the space @xmath66 is a compact , zero - dimensional manifold which contains a fixed number , @xmath67 , of points , independently of @xmath55 . for suitably generic @xmath55 , the moduli space @xmath66 consists of immersed curves , that is , there are no singular curves or cusp curves . thus the relevant gromov - witten invariant counting curves through the @xmath32 points is computed by looking at immersed curves , and if these curves are counted with multiplicity then the total is independent of @xmath55 . on the other hand , the automatic regularity phenomena for holomorphic spheres in dimension four ( see lemma 3.3.3 of @xcite ) implies that all such curves appear with the same multiplicity @xmath68 , and our conclusion follows . for a generic choice of @xmath55 , @xmath69 is a two - dimensional manifold which can be compactified by adding finitely many cusp - curves that consist of the union of two holomorphic spheres each passing through a subset of the @xmath70 points . in particular , if the points @xmath71 are in generic position , then one does not need to add cusp - curves with more than two components , nor does one need to add multiply - covered curves . [ connected ] for @xmath72 and a generic choice of @xmath55 and the point constraints @xmath71 , the moduli space @xmath69 is connected . lemma 2.3.3 of @xcite yields a version of automatic regularity for holomorphic spheres of nonnegative index in a @xmath52-manifold , which includes the possibility of singularities . in particular , the manifold @xmath69 is diffeomorphic to @xmath73 , where @xmath74 is the standard complex structure on @xmath36 . now , the space of all rational algebraic curves of degree @xmath29 in @xmath36 is dominated by an open set of triples of homogeneous polynomials on @xmath75 and so is irreducible . sets of curves through @xmath32 points are given by intersecting this variety with an appropriate plane . it then follows from the uniform position principle that any two such curves can be connected with curves passing through any subset of @xmath70 points , see appendix e of @xcite . but any two curves of degree @xmath76 have @xmath77 points in common and so the conclusion follows . for @xmath78 , let @xmath79 denote the ellipsoid @xmath80 if @xmath81 , then the standard liouville form on @xmath82 restricts to @xmath83 as a contact form @xmath84 which has exactly two simple reeb orbits , @xmath85 and @xmath86 , which lie in the planes @xmath87 and @xmath88 , respectively . the action of @xmath85 is @xmath89 and the action of @xmath86 is @xmath90 . the conley - zehnder indices of these orbits , with respect to the natural trivialization of the @xmath91 and @xmath92 coordinate planes , are given by @xmath93 and @xmath94 . more generally , if @xmath95 is the @xmath96-fold cover of @xmath97 , then we have @xmath98 and @xmath99 these trivialization along closed reeb orbits will be used throughout , see remark [ chernclass ] below . it follows from theorem 2 of @xcite that for @xmath100 the ellipsoid @xmath101 can be symplectically embedded into the interior of @xmath102 and hence @xmath21 . for a fixed degree @xmath29 we now fix such an embedding of @xmath101 for an @xmath17 satisfying @xmath103 in what follows , we will denote @xmath101 by @xmath104 and will identify @xmath104 with its image in @xmath21 . in this section we recall from @xcite the structure of the limits of holomorphic spheres in @xmath36 as one splits this manifold ( _ stretches the neck _ ) along the boundary of the embedded ellipsoid @xmath104 . let @xmath106 be the vector field defined near @xmath105 as the symplectic dual of the liouville form . an almost - complex structure @xmath55 on @xmath36 is said to be compatible with @xmath104 if the contact structure @xmath107 on @xmath105 is equal to @xmath108 , and @xmath109 is equal to the reeb vector field of @xmath110 . denote by @xmath111 the set of @xmath112 which are compatible with @xmath104 . for every natural number @xmath113 , the union of the three pieces @xmath114 , d(e^{\tau } \alpha_e)\right),\ , \text { and } \left ( { { \mathbb c}}p^2 \smallsetminus e , e^{n}\omega_{fs } \right),\ ] ] attached along their appropriate boundary components , is a symplectic manifold @xmath115 . for a @xmath55 in @xmath111 , let @xmath116 be the continuous almost - complex structure on @xmath117 which equals @xmath55 on the disjoint union of @xmath118 and @xmath104 , and is translation invariant on @xmath119 $ ] . to make each of the @xmath116 smooth one must perturb @xmath55 near @xmath105 . as noted in 3.4 of @xcite , the choice of this perturbation is irrelevant for the compactness theorem below . accordingly , we will henceforth assume that this choice has been made , and that the almost - complex structures @xmath116 are smooth . fix points @xmath56 in @xmath120 and consider a sequence @xmath121 in the space @xmath122 . fix a representative curve @xmath123 for each @xmath121 . the following compactness theorem is proved by bourgeois , eliashberg , hofer , wysocki and zehnder in @xcite . ( theorem 10.6 of @xcite ) there exists a subsequence of the @xmath123 which converges to a holomorphic building , @xmath124 . we now describe the relevant aspects of this theorem starting with the structure of the limiting building @xmath124 . the reader is referred to @xcite for the precise definitions , statements and proofs of the general compactness theorem for holomorphic curves under splittings . we begin by describing a holomorphic building @xmath125 which arises as the limit of a subsequence of the curves @xmath123 . the domain of @xmath125 is a nodal riemann sphere @xmath126 with punctures . the building @xmath125 then consists of a collection of finite energy holomorphic maps from the collection of punctured spheres of @xmath127 to one of the following three symplectic manifolds : * @xmath128 * @xmath129 * @xmath130 , d ( e^{\tau } \alpha_e)\right ) . $ ] more precisely , these target manifolds are equipped with the almost - complex structures which are induced by @xmath55 and are translation invariant on the subsets which are cylinders over @xmath105 . the curves of @xmath124 are then holomorphic with respect to these almost - complex structures and the complex structures on the punctured spheres induced by the structure @xmath131 on @xmath132 . since each curve of @xmath124 has finite energy , they are all asymptotically cylindrical near each puncture , to some multiple of either @xmath85 or @xmath86 . if @xmath133 is a curve of @xmath125 with image in @xmath134 , then the punctures in its domain are all negative , i.e as @xmath135 approaches a puncture on the domain of @xmath133 , @xmath136 takes values in @xmath137 $ ] and its @xmath138$]-component converges to @xmath139 . similarly , each curve of @xmath125 with image in @xmath140 has only positive punctures , and curves of @xmath125 with image in @xmath141 have both negative and positive punctures ( but not only negative punctures ) . the limiting building @xmath125 is also equipped with a _ level structure_. for a building @xmath125 of level @xmath142 , this structure is encoded by a labeling of the punctured riemann spheres of @xmath127 by integers from @xmath143 to @xmath144 , called levels , such that the levels of two components which share a node , differ at most by @xmath68 . let @xmath145 denote the union of components of level @xmath96 and denote by @xmath146 the holomorphic curve of @xmath124 with ( possibly disconnected ) domain @xmath145 . then @xmath147 , @xmath148 , for @xmath149 , and @xmath150 . moreover , each node shared by @xmath145 and @xmath151 is a positive puncture for @xmath146 and a negative puncture for @xmath152 , each asymptotic to the same reeb orbit . as well , @xmath146 extends continuously across each node within @xmath145 . lastly , we recall that the curves of @xmath124 have two collective properties . the first of these is the fact that the sum , over components , of their virtual indices ( see below ) is equal to @xmath153 , the deformation index of the curves @xmath123 . to state the second collective property , we must first recall the definition of a _ compactifcation _ of a curve of @xmath124 . this definition depends on the target of the curve . for a curve @xmath63 of @xmath124 with image in @xmath154 one can write @xmath155 where @xmath156 maps the domain of @xmath63 to @xmath105 . the map @xmath156 then extends to a continuous map @xmath157 , the compactification of @xmath63 , which takes the oriented blow - up of the domain of @xmath63 to @xmath105 such that the circle corresponding to each puncture is mapped to the closed reeb orbit on @xmath105 which determines the asymptotic behavior of @xmath63 near that puncture , ( see proposition 5.10 of @xcite ) . let @xmath133 be a curve of @xmath125 with image in @xmath134 . as described in section 3.2 of @xcite , one can identify @xmath158 with @xmath159 via a diffeomorphism @xmath160 which is the identity map away from an arbitrarily small tubular neighborhood of @xmath161 in @xmath162 . one can then extend the map @xmath163 to a smooth map @xmath164 which takes the oriented blow - up of the domain of @xmath133 to @xmath162 such that each boundary circle goes to the appropriate closed reeb orbit . a choice of this extension @xmath164 is a compactification of @xmath133 . similarly , for a curve @xmath165 of @xmath124 with image in @xmath166 , one can use a diffeomorphism @xmath167 to define a compactification @xmath168 of @xmath165 as a smooth extension of @xmath169 which takes the oriented blow - up of the domain of @xmath165 to @xmath104 , and again takes each boundary circle to the corresponding closed reeb orbit on @xmath105 . if one fixes a compactification for each of the curves in @xmath125 , then these maps must fit together to form a continuous map @xmath170 . let @xmath124 be a holomorphic building of level @xmath142 , as above , whose domain is the riemann surface with nodes @xmath126 . if @xmath125 is the limit of the holomorphic spheres @xmath123 in the sense of @xcite , then there exist maps @xmath171 and sequences @xmath172 , @xmath173 , such that : 1 . the @xmath174 are diffeomorphisms except that they may collapse a finite collection of circles in @xmath61 to nodes in @xmath132 . moreover , @xmath175 converges to @xmath131 away from the nodes of @xmath132 , where @xmath74 is the standard complex structure on @xmath61 . the sequences of maps @xmath176 and @xmath177 converge in the @xmath178-topology to the maps @xmath179 and @xmath180 , respectively . for @xmath181 the maps @xmath182 converge to @xmath146 in the @xmath178-topology where @xmath183 is the diffeomorphism of @xmath184 which translates the @xmath185-component by @xmath186 . here , as is necessary , we are identifying @xmath187 with an increasing sequence of domains in @xmath141 , @xmath188 with an increasing sequence of domains in @xmath140 , and @xmath189 \subset ( { { \mathbb c}}p^2)^{n}$ ] with an increasing sequence of domains in @xmath158 . when the number of point constraints @xmath60 is equal to @xmath32 , one can use the special nature of the reeb flow of @xmath110 on @xmath105 , together with the regularity results of dragnev from @xcite , to obtain more restrictions on the curves comprising our limiting holomorphic buildings . throughout this section we will consider curves of a limiting building which have connected domains . [ index ] let @xmath124 be a holomorphic building which is the limit of a sequence of curves @xmath123 representing classes in @xmath190 . for a generic choice of @xmath55 in @xmath111 and the constraint points @xmath191 we have : 1 . the curves in @xmath125 whose image lies in @xmath134 are somewhere injective , regular , and have deformation index equal to zero . their ( negative ) ends are all asymptotic to multiples of @xmath85 . the curves in @xmath125 whose image lies in @xmath141 are all multiple covers of holomorphic cylinders over @xmath85 , and they have virtual index zero . the curves in @xmath125 whose image lies in @xmath140 are somewhere injective , regular , and have deformation index equal to zero ( with the point constraints ) . their ( positive ) ends are all asymptotic to multiples of @xmath85 . by @xcite , we may choose @xmath55 in @xmath111 so that the corresponding almost - complex structures on @xmath134 , @xmath141 , and @xmath140 are regular in the sense that for all somewhere injective curves which have finite energy and genus zero , the deformation index of the curve is equal to its virtual index . we choose the points @xmath191 so that they are in general position , as constraints , for the spaces @xmath190 as well as for the moduli spaces of somewhere injective finite energy pseudo - holomorphic curves of genus zero in @xmath140 . consider a curve @xmath133 of @xmath125 with image in @xmath134 . suppose that @xmath133 has @xmath192 negative ends asymptotic to multiples of @xmath85 , and @xmath193 negative ends asymptotic to multiples of @xmath86 . if the @xmath194 negative end covering @xmath85 does so @xmath195 times , and the @xmath194 negative end covering @xmath86 does so @xmath196 times , then the virtual deformation index of @xmath133 is @xmath197 [ chernclass ] the index formulas used in this paper all involve chern numbers of the finite energy curves . finite energy curves do not give closed cycles in the ambient manifold , say @xmath106 , however they are asymptotic at their ends to periodic orbits of reeb flows and so can be compactified to give @xmath198-dimensional cycles with boundary , which can always be perturbed to be defined by immersions . the definition of the chern number depends upon a choice of trivialization of @xmath199 along this boundary , and that trivialization must of course be compatible with the one used in section [ thinellipse ] to define our conley - zehnder indices . to be precise , in all cases considered in this paper our reeb orbits @xmath200 will lie in the boundary of an ellipsoid @xmath104 in @xmath18 . thus we have a standard symplectic trivialization of @xmath201 coming from restriction of the standard trivialization on @xmath18 . the action of the derivative of the reeb flow ( extended to act trivially on the normal vector to @xmath202 ) induces a family of symplectomorphisms @xmath203 for @xmath204 , where @xmath62 is the length of the reeb orbit . we then define @xmath205 following @xcite , as was done above in section [ thinellipse ] . suppose that @xmath106 has real dimension @xmath206 and is equipped with an almost - complex structure @xmath55 . the determinant line bundle @xmath207 has a standard section @xmath17 over @xmath104 , again using our trivialization . in the above formula @xmath208 can be defined to be the number of zeros ( counted with multiplicity ) of a section of @xmath209 which agrees with @xmath17 over @xmath200 . when applied to finite energy curves lying in @xmath140 or @xmath141 these choices imply automatically that the chern numbers vanish . for curves lying in @xmath210 the chern number is the same as the usual chern number , @xmath211 \rangle$ ] , where @xmath212 $ ] is the homology class represented by the cycle @xmath213 formed by gluing the appropriate discs in @xmath104 to the compactification of @xmath133 . in the case when @xmath214 , by poincar duality this is just @xmath51 times the intersection number of @xmath133 with the line at infinity in the complex projective space . by the iteration formulas and , the index formula above simplifies to @xmath215 as described above , the chern number @xmath208 is equal to three times the intersection number of @xmath133 with the line at infinity . but by positivity of intersection this intersection number is bounded above for each component by the intersection number of degree @xmath29 curves with the line at infinity . hence , @xmath208 is at most @xmath216 , the chern class of degree @xmath29 curves in @xmath36 . as well , the bound on @xmath17 from implies that @xmath217 , and so @xmath218 together with , this expression leads immediately to the following result . [ fpos ] if a holomorphic curve @xmath133 of @xmath125 with image in @xmath134 has nonnegative virtual index , then @xmath219 , @xmath220 and @xmath221 . moreover , @xmath222 the fredholm theory gives regularity for somewhere injective curves with respect to generic almost - complex structures . we will need the following proposition to describe possible finite energy curves which are not somewhere injective . [ multiplecover ] let @xmath106 be a symplectic manifold perhaps having cylindrical ends and equipped with a compatible almost - complex structure as described above . a finite energy holomorphic curve @xmath223 is either somewhere injective or there exists a proper holomorphic map @xmath224 and a somewhere injective curve @xmath225 such that @xmath226 . this follows almost exactly as in the case of closed holomorphic curves , see for example @xcite , proposition @xmath227 , at least if we assume that the reeb flow is nondegenerate . then our finite energy curves converge asymptotically to cylinders over closed reeb orbits and in particular have only finitely many critical points , see @xcite . the proof has already been adapted to finite energy planes in @xcite , theorem @xmath228 , in which case the map @xmath229 is shown to be polynomial . at least in the nondegenerate case this proof applies equally well to finite energy curves with multiple ends and perhaps higher genus ( although of course in the higher genus case @xmath229 may not necessarily be polynomial ) . in the situation at hand we may therefore suppose that @xmath133 is an r - fold cover of a simple curve @xmath230 . it follows from the regularity of @xmath55 , that @xmath231 . lemma [ fpos ] then implies that @xmath232 , @xmath233 , @xmath234 , and @xmath235 hence , for @xmath133 we have @xmath236 and @xmath237 thus , the hypothesis of lemma [ fpos ] also holds for each curve @xmath133 of @xmath124 with image in @xmath134 and we have [ f ] let @xmath133 be a curve of @xmath125 with image in @xmath134 . then @xmath219 and the ends of @xmath133 are all asymptotic to some multiple of @xmath85 and cover @xmath85 at most @xmath32 times . the virtual index of @xmath133 is nonnegative and is strictly positive when @xmath133 is a multiple cover . now consider a curve @xmath63 of @xmath125 whose image lies in the symplectization @xmath141 . since none of the curves of @xmath125 in @xmath134 have negative ends asymptotic to multiples of @xmath86 , it follows from the existence of the map @xmath238 that , at least if @xmath63 is of the top level @xmath142 , the positive ends of @xmath63 can only be asymptotic to multiples of @xmath85 . suppose that @xmath63 has @xmath239 such ends , and that the @xmath194 one covers @xmath85 a total of @xmath240 times . as established above , each curve of @xmath125 in @xmath134 has at most @xmath32 negative ends when counted with multiplicity , and there are at most @xmath29 such curves since each one has a positive chern number . hence @xmath241 suppose that @xmath63 has @xmath242 negative ends asymptotic to multiples of @xmath85 with the @xmath194 such end covering this orbit @xmath196 times , and @xmath63 has @xmath243 negative ends asymptotic to multiples of @xmath86 with the @xmath194 such end covering @xmath86 a total of @xmath244 times . then , by stokes theorem we have @xmath245 our choice of @xmath17 satisfying implies that @xmath246 and so @xmath220 . integrating @xmath247 over @xmath63 once again , we now have @xmath248 hence the total number of positive ends of @xmath63 is no less than the total number of its negative ends and by and we have @xmath249 hence , the virtual index of @xmath63 is strictly positive unless @xmath250 and @xmath251 . this condition is equivalent to the curve being a multiple cover of a cylinder over @xmath85 . as @xmath63 has no negative ends asymptotic to @xmath86 then same conclusions apply by induction to lower level curves mapping to @xmath141 . to summarize , we have [ g ] let @xmath63 be a curve of @xmath124 with image in the symplectization @xmath141 . the positive and negative ends of @xmath63 are all asymptotic to some multiple of @xmath85 and the positive ends cover @xmath85 at least as many times as the negative ends . the virtual index of @xmath63 is nonnegative and is strictly positive unless @xmath63 has one positive end and is a multiple cover of a cylinder over @xmath85 . finally , we consider a curve @xmath165 of @xmath125 whose image is in @xmath140 . our analysis of the curves @xmath63 above implies that none of the positive ends of @xmath165 are asymptotic to multiples of @xmath86 . suppose that @xmath165 has @xmath252 positive ends asymptotic to multiples of @xmath85 with the @xmath194 such end covering this orbit @xmath253 times . then @xmath254 where @xmath255 is the number of point constraints achieved by the curve @xmath165 . the fact that the negative ends of the collection of curves of @xmath124 in @xmath134 cover @xmath85 at most @xmath256 times , together with the fact that positive ends of the curves in @xmath141 cover @xmath85 at least as many times as their negative ends , implies that @xmath257 hence @xmath258 arguing as above , the curve @xmath165 can be realized as an @xmath96-fold cover of a somewhere injective curve @xmath259 with the same point constraints . by regularity @xmath260 since @xmath261 and @xmath262 , we then have @xmath263 [ h ] let @xmath165 be a curve of @xmath125 with image in @xmath140 . the positive ends of @xmath165 are asymptotic to some multiple of @xmath85 . the virtual index of @xmath165 is nonnegative and is strictly positive if @xmath165 is multiply covered . note that the deformation index of the curves in @xmath190 is zero , and so the sum of the virtual indices of the curves of @xmath124 is also zero . it follows from lemmas [ f ] , [ g ] , and [ h ] , that every curve of @xmath125 must have virtual index zero . the same three lemmas then yield the three statements of proposition [ index ] . [ points exist ] the third statement of proposition [ index ] implies that the curves in @xmath125 with image in @xmath140 must each pass through at least one of the point constraints . in particular , these curves are regular with deformation index equal to zero , and unconstrained regular curves must have deformation index at least two . more generally , curves in @xmath140 asymptotic to @xmath264 have an unconstrained index @xmath265 and so must pass through at least @xmath96 of the point constraints . for any integer @xmath266 , one can symplectically embed @xmath267 disjoint balls of radius one into the ellipsoid @xmath268 , @xcite . hence , for @xmath17 satisfying , we may assume that the points @xmath191 lie at the center of disjoint balls in @xmath269 which have radii slightly less than @xmath270 and whose closures lie in the interior of @xmath104 . let @xmath271 be an almost - complex structure which agrees with the standard complex structure in the balls around the points @xmath272 , and induces almost - complex structures on @xmath134 , @xmath141 , and @xmath140 which are regular for somewhere injective , finite energy curves of genus zero . consider again a sequence of curves @xmath123 which represent classes in @xmath190 and converge , in the sense of @xcite , to a holomorphic building @xmath124 . for this new choice of @xmath55 we get the following refinements of proposition [ index ] . [ limit ] with @xmath55 as above , there is exactly one curve of @xmath125 whose image lies in @xmath134 . it has exactly @xmath32 negative ends when counted with multiplicity . moreover , the curves of @xmath125 with image in @xmath141 each have exactly one positive puncture and the curves of @xmath125 with image in @xmath140 are all holomorphic planes . we first show that the negative ends of the curves of @xmath124 with image in @xmath134 cover @xmath85 at least @xmath32 times . let @xmath273 be the collection of curves of @xmath125 with images in either @xmath141 or @xmath274 . denote by @xmath275 the map to @xmath104 formed by fitting together the compactifications of the curves of @xmath273 . by the existence of the map @xmath276 , the ends of @xmath275 cover @xmath85 the same number of times , say @xmath96 , as the curves of @xmath124 with image in @xmath134 . now , the symplectic area of @xmath275 is @xmath277 . on the other hand , since the compactifications of curves @xmath274 are holomorphic away from @xmath105 , the monotonicity theorem and our choice of @xmath55 imply that the intersections of @xmath275 with the balls centered at the points @xmath278 have symplectic area arbitrarily close to @xmath279 . hence , @xmath280 . so , the curves of @xmath124 with image in @xmath134 have a total chern number of @xmath216 and , collectively , their negative ends cover @xmath85 at least @xmath32 times . as shown in the proof of proposition [ index ] , the deformation index of a curve @xmath133 in @xmath124 whose image lies in @xmath134 is given by @xmath222 where @xmath192 represents the number of negative ends and @xmath195 is the number of times the @xmath194 negative end covers @xmath85 . if there are @xmath281 such curves , then their total deformation index is at most @xmath282 since the total index must be nonnegative we have @xmath283 . hence , there is exactly one curve , say @xmath133 , of @xmath124 in @xmath134 , and @xmath133 has index zero and exactly @xmath32 negative ends , when counted with multiplicity . the remaining statements of lemma [ limit ] follows easily from the first as @xmath133 is a limit of genus @xmath143 curves . in this section we prove two results involving families of ( moduli spaces of ) curves with varying point constraints . let @xmath284 be a set of paths in @xmath104 for @xmath285 and @xmath286 . for each @xmath113 set @xmath287\},\ ] ] where @xmath116 is the almost - complex structure on @xmath288 defined by @xmath271 in the manner described in section [ split ] . [ path1 ] for a given fixed @xmath289 and almost - complex structure @xmath271 , there exists an open , dense set of paths @xmath284 for which the moduli spaces @xmath290 are compact @xmath68-dimensional manifolds and the natural projection onto @xmath291 $ ] is a covering map of degree @xmath67 . the set of nonimmersed curves is closed , and nonempty only with respect to a codimension @xmath198 collection of point constraints . therefore for an open dense set of paths we may assume that all curves in @xmath290 are immersed . automatic regularity for holomorphic spheres in dimension four , see lemma 3.3.3 of @xcite , implies that immersed spheres with nonnegative virtual deformation index are in fact regular . the same is true for moduli spaces of curves satisfying point constraints , as can be seen by identifying curves passing through the points with unconstrained curves in a suitable blow - up at the points . alternatively , and more directly , we can apply theorem @xmath292 of @xcite which gives regularity for all curves in a generic path . the following result will be used several times in the proof of theorem [ thm ] . [ path2 ] for a suitable choice of almost - complex structure @xmath293 , and for a generic family of paths @xmath284 , the conclusions of proposition [ path1 ] hold for all @xmath289 . further , let @xmath294 , for @xmath295 $ ] , be a generic @xmath68-parameter family of curves representing a component of @xmath290 . if the subsequences @xmath296 and @xmath297 both converge , then the top level curves of the two corresponding limiting holomorphic buildings , that is the curves which map into @xmath298 , have identical images . we begin by describing our choice of generic data . we choose @xmath293 so that the corresponding limiting almost - complex structures on @xmath298 and @xmath299 are regular for somewhere injective finite energy pseudo - holomorphic curves of genus zero . we then choose the paths @xmath284 so that the moduli spaces @xmath300 are all regular . it follows from proposition [ path1 ] , that this corresponds to choosing the paths @xmath284 from a countable , over @xmath289 , intersection of open dense subsets of families of paths . we then restrict the choice of paths again so that for all @xmath295 $ ] the points @xmath284 are in general position , as constraints , for the moduli spaces of somewhere injective finite energy pseudo - holomorphic curves of genus zero in @xmath140 . we note that the relevant moduli spaces of curves in @xmath140 have even unconstrained indices . therefore the general position condition can be stated equivalently as follows , we require that for any @xmath295 $ ] and any finite energy curve in @xmath299 of genus zero and deformation index @xmath301 , the curve passes through at most @xmath142 of the @xmath302 . we now verify that this last condition is also generic , that is , is satisfied by families of @xmath32 paths in an open , dense subset of the space of all such families of paths . since finite intersections of dense open sets are dense and open , it suffices to show that there is an open dense set of paths @xmath303 such that no genus @xmath143 finite energy curve in @xmath140 of index @xmath301 passes through all @xmath284 for any @xmath304 . consider a moduli space @xmath305 of such curves . there is a smooth evaluation map @xmath306 \to e^{k+1 } \times [ 0,1]$ ] acting trivially on the second factor , where @xmath307 denotes the corresponding moduli space of curves equipped with @xmath144 marked points and @xmath308 is @xmath144 copies of @xmath104 . now , @xmath307 has dimension @xmath309 while @xmath308 has dimension @xmath310 . thus by sard s theorem the image misses a generic submanifold of codimension @xmath311 , that is , dimension @xmath68 . thus an open , dense set of paths @xmath312 will avoid the image of the evaluation map , which is the same as saying that no curve in the moduli space will pass through all of the points , for any @xmath304 . with these choices in place , the conclusions of proposition [ index ] hold for any holomorphic building which arises as the limit of a subsequence of curves the form @xmath313 . assume now that the top level curves of the limiting buildings for @xmath296 and @xmath297 have distinct images . below , we derive from this assumption the existence of a subsequence of curves of the form @xmath313 that converges to a holomorphic building with a top level curve of virtual index greater than zero . this will contradict the first assertion of proposition [ index ] . it follows from our choice of @xmath55 that every finite energy @xmath55-holomorphic curve in @xmath298 with genus zero and virtual index zero is regular . ( for this , we recall from equation [ nonneg ] that for regular @xmath55 any curve which is a nontrivial multiple cover necessarily has deformation index at least @xmath198 . ) hence there are only finitely many images of such curves of degree bounded by @xmath29 , say @xmath314 . set @xmath315 by the unique continuation principle , @xmath316 are distinct compact subsets of @xmath162 . let @xmath317 and @xmath318 be the limiting buildings of @xmath296 and @xmath297 , respectively . the first statement of proposition [ index ] implies that the top levels curves of @xmath317 and @xmath318 are all somewhere injective finite energy curves of genus zero with image in @xmath298 . by our assumption , the intersections of the images of these top level curves with @xmath162 are distinct subsets , say @xmath319 and @xmath320 , of @xmath321 . fixing a metric on @xmath162 we consider the corresponding hausdorff metric on compact subsets of @xmath162 . for large @xmath131 , the sets @xmath322 are arbitrarily close to @xmath319 for @xmath304 near zero , and arbitrarily close to @xmath320 for @xmath304 near one . since @xmath323 , we can then choose , for all large @xmath131 , a time @xmath324 such that @xmath325 is a fixed distance , say @xmath326 , from @xmath327 . passing to a subsequence , if necessary , we may then assume that the @xmath324 converge to some @xmath328 $ ] and that @xmath329 converges to a holomorphic building , @xmath330 . by the choice of the @xmath331 , the images of the curves of @xmath330 with target @xmath298 intersect @xmath162 in a set whose distance from @xmath332 is at least @xmath333 . hence , at least one of the curves of @xmath330 with image in @xmath298 is not rigid and so has virtual index at least @xmath198 . this is the desired contradiction of proposition [ index ] . [ closeimage ] the above proof gives the following information on the images of the @xmath294 which is a convenient way to apply proposition [ path2 ] . let @xmath334 be a compact subset with smooth boundary @xmath335 such that the limiting component of the @xmath296 in @xmath134 , say @xmath133 , intersects @xmath335 transversally . define @xmath336 and @xmath337 . then given any @xmath326 and sequence @xmath338 $ ] there exists a @xmath339 such that if @xmath340 then there is a diffeomorphism @xmath341 such that @xmath342 is @xmath333 close to @xmath133 in a @xmath343-norm . ( we can define this norm by fixing metrics on @xmath344 and @xmath281 . ) indeed , if the @xmath345 converge to a building with top level component identical to @xmath133 then we can take the @xmath346 to be a small perturbation of the maps @xmath347 implicit in the definition of convergence in section [ compact ] . we notice that if @xmath348 is sufficiently @xmath343 close to @xmath133 and @xmath133 intersects @xmath335 transversally then by the implicit function theorem @xmath348 can be slightly perturbed such that the preimage of @xmath335 coincides with that of @xmath133 . on the other hand , if the @xmath345 do not converge to a building with top level component identical to @xmath133 then after taking a subsequence we will derive a contradiction to proposition [ path2 ] . the key to the proof of theorem [ thm ] is the following existence result . [ key ] for any integer @xmath349 and a suitable choice of almost - complex structure @xmath293 , there exists a regular , finite energy holomorphic plane of degree @xmath29 in @xmath134 whose negative end covers the periodic orbit @xmath85 precisely @xmath32 times . here , as before , @xmath104 is the ( the image of the ) ellipsoid @xmath101 and we assume that @xmath350 . our proof of this theorem is quite lengthy . the equivalent result for embeddings of @xmath104 into @xmath351 is that there exists a finite energy plane in @xmath352 whose negative end covers @xmath85 precisely @xmath353 times , and gluing a @xmath353 times cover of the disk @xmath354 to the plane produces a homology class @xmath44 . this turns out to have a simpler proof . on the one hand , the fact that the corresponding gromov - witten invariant for holomorphic spheres in the class @xmath44 through @xmath353 points is @xmath68 allows a proof as described in @xcite . alternatively , the fact that such spheres are all embedded allows an application of embedded contact homology , see @xcite , @xcite . the index formulas in ech imply immediately that when we stretch the neck there can only be a single component in @xmath352 , and it has a single negative end . unfortunately degree @xmath29 curves in @xmath36 are not embedded , and neither is the relevent gromov - witten invariant equal to @xmath68 . thus we were led to follow the argument described below . the rest of this section is devoted to the proof of theorem [ key ] . we begin by fixing points @xmath355 in @xmath120 and an almost - complex structure @xmath271 as in lemma [ limit ] . in particular , we assume that @xmath55 agrees with the standard complex structure in @xmath32 disjoint darboux balls of radius slightly less than @xmath270 which are contained in @xmath104 and centered around the points @xmath272 . we also assume that @xmath55 induces almost - complex structures on @xmath134 , @xmath141 , and @xmath140 which are regular for somewhere injective , finite energy curves of genus zero . consider the set of holomorphic buildings which occur as the limit of a sequence of curves representing classes in the spaces @xmath190 . by lemma [ limit ] these limit buildings each have a single curve with image in @xmath134 and the negative ends of this curve cover @xmath85 a total of @xmath32 times . let @xmath356 be the maximum number of times a single negative end of one of these limit curves in @xmath134 covers @xmath85 . to prove theorem [ key ] it suffices to show that @xmath357 . we will start by showing that @xmath358 . arguing by contradiction , we will then assume that @xmath359 and derive some consequences of this assumption on the curves of our limit buildings , expressed in the language of partitions . ultimately , we derive a contradiction from these consequences by using them to detect a limit building whose curve with image in @xmath134 has a negative end which covers @xmath85 at least @xmath360 times . [ mnotone ] @xmath358 . assume that this is false and that for any limit @xmath124 , as above , the unique curve @xmath133 of @xmath125 in @xmath134 has precisely @xmath32 negative ends each of which covers @xmath85 once . by lemma [ limit ] these limits each have precisely @xmath32 planar curves with image in @xmath140 , and by remark [ points exist ] each of these curves passes through exactly one of the point constraints . asymptotic convergence to a cylinder over the reeb orbit implies that we can find a closed subset @xmath361 such that @xmath362 is topologically a sphere with @xmath32 open disks removed . for each @xmath289 , we take a family of paths @xmath284 in @xmath104 with @xmath363 and @xmath364 for all @xmath74 , where @xmath365 is in @xmath104 . we can not guarantee that the families of point constraints @xmath366\}$ ] are generic in the sense of proposition [ path1 ] . indeed , the moduli space of degree @xmath29 curves passing through the @xmath367 has high dimension . nevertheless , we may choose the families @xmath368 generically so that as in proposition [ path1 ] there exist corresponding families of curves @xmath294 , for @xmath369 , which pass through the points @xmath284 and satisfy @xmath370 . a limit may not exist as @xmath371 . however by gromov compactness there exists a sequence @xmath372 such that the curves @xmath373 converge . the limit is a possibly nodal @xmath116-holomorphic sphere @xmath374 of degree @xmath29 with @xmath32-marked points @xmath375 mapping to @xmath365 . if we think in terms of stable holomorphic maps from riemann surfaces with marked points , the limiting curve @xmath376 here could certainly include ghost bubble components mapping onto @xmath365 . for example , in the case of degree @xmath377 and points @xmath378 and @xmath379 , after perhaps taking a subsequence curves @xmath380 in the homology class of a line will converge to a nodal curve consisting of a degree @xmath68 curve through the point @xmath365 and a ghost bubble containing the marked points @xmath381 and @xmath382 . which nontrivial curve arises in the limit depends upon the ( complex ) angle of approach of @xmath378 and @xmath379 to @xmath365 in our subsequence . more precisely , there is a @xmath198-parameter family of degree @xmath68 curves in @xmath36 passing through @xmath365 , parameterized by their tangent plane in @xmath383 , that is , by a @xmath75 . if we fix @xmath384 and choose @xmath379 to converge when lifted to the blow - up of @xmath36 at the point @xmath365 , then the tangency of the limit is determined by the limit of the @xmath379 in the exceptional divisor . we do not aim to specify the limit @xmath376 precisely , but for our fixed @xmath289 and any suitable sequence @xmath372 we make the following claim . [ differnt ] the complement of @xmath385 does _ not _ have @xmath32 separate components each containing a marked point . ( component here is meant topologically , it does not refer to components of a nodal curve . ) we recall that by our choice of @xmath281 and the convergence of @xmath386 to @xmath125 we have that for @xmath289 sufficiently large the complement of @xmath387 _ does _ have @xmath32 components each containing a single marked point . given claim [ differnt ] , by the convergence to @xmath376 we can find a @xmath388 in @xmath389 such that @xmath390 also has the property that the complement of @xmath391 does _ not _ have @xmath32 components each containing a marked point . for convenience let us change our parameterizations and call @xmath390 instead just @xmath392 . ( in any case , there is a family of curves connecting it to @xmath386 . ) repeating this construction for each @xmath289 , we get paths @xmath294 which connect curves that pullback @xmath281 to domains with different topology . now we make the following claim . [ same ] for these families of curves , @xmath294 with @xmath393 , the conclusion of proposition [ path2 ] still holds . that is , if the subsequences @xmath296 and @xmath297 both converge , then the curves with image in @xmath298 of the corresponding limiting holomorphic buildings , are identical . proposition [ mnotone ] follows immediately from the two claims . for if the limiting components in @xmath298 are identical , as they are according to claim [ same ] , then for large @xmath331 the pullbacks @xmath394 and @xmath395 must have the same topology , but this contradicts claim [ differnt ] . _ proof of claim [ differnt ] . _ if the claim is false then there are @xmath32 distinct components of @xmath396 each of which contains a marked point that is mapped by @xmath376 to @xmath365 . on the other hand , for each @xmath289 there exists a ( unique ) @xmath116-holomorphic sphere @xmath344 in @xmath117 which has degree one and passes through @xmath365 and another fixed point @xmath397 . by positivity of intersection we then have @xmath398 , a contradiction . _ proof of claim [ same ] . _ we argue as in the proposition [ path2 ] , see also remark [ closeimage ] . the new difficulty is that we can not necessarily assume that the point constraints are regular . on the other hand the many components in @xmath140 will be enough to ensure that curves outside of @xmath104 can not have strictly positive deformation index . assume , arguing by contradiction , that the limits of @xmath296 and @xmath297 in @xmath298 are distinct . consider the compact subsets @xmath399 of @xmath281 defined similarly to proposition [ path2 ] , i.e. , each @xmath400 is the intersection of @xmath281 with of the image of a rigid curve in @xmath298 . ( by compactness there are only finitely many such genus @xmath143 curves of bounded degree . ) fixing an @xmath326 sufficiently small , for @xmath131 large we can choose a @xmath324 such that @xmath401 is at least a distance @xmath333 from @xmath327 and @xmath402 still has @xmath32 components in a compact neighborhood of @xmath104 , each passing through a marked point . ( we observe that this number of components can not decrease without the component in @xmath281 itself moving . ) passing to a subsequence , if necessary , we may then assume that the @xmath324 converge to some @xmath328 $ ] and that @xmath329 converges to a holomorphic building which has a nonrigid curve in @xmath298 , and which has at least @xmath32 curves with image in @xmath299 . the curves in @xmath299 have unconstrained deformation index at least @xmath198 and therefore the total unconstrained index of limiting curves in @xmath299 is at least @xmath403 . on the other hand , degree @xmath29 curves in @xmath36 have deformation index @xmath404 . so , as our almost - complex structure @xmath55 induces a regular almost structure on @xmath298 , any curves of this limit with image in @xmath298 must have index exactly @xmath143 . this is a contradiction since , by construction , there is a curve of the limit in @xmath298 which is not rigid . in order to establish theorem [ key ] , that is , that @xmath357 , we argue by contradiction . given proposition [ mnotone ] , this means that we assume @xmath359 . let @xmath124 be a limiting holomorphic building for a subsequence of the curves @xmath123 with @xmath405 \in { \mathcal m}_d(j^n , p_1 , \dots , p_{3d-1})$ ] , such that the unique curve of @xmath124 with image in @xmath134 has a negative end which covers @xmath85 exactly @xmath356 times . we will use this particular limit @xmath124 to derive our contradiction . denote the level of @xmath125 by @xmath142 and its domain by @xmath126 . for simplicity , we will suppress the choice of a subsequence , and assume that the @xmath123 converge to @xmath125 in the sense of @xcite . * partitions associated to @xmath124 . * the limiting building @xmath124 can be used to define partitions of the set of point constraints @xmath406 as follows . let @xmath273 be the collection of curves of @xmath125 with images in either @xmath141 or @xmath274 and denote by @xmath275 the map to @xmath104 formed by fitting together the compactifications of the curves of @xmath273 , as described in section [ lliimm ] . it follows from lemma [ limit ] , that the domain of @xmath275 is a union of discs . the intersections of the images of these discs with the point constraints determines a partition @xmath407 of @xmath408 . in particular , we say that two points lie in the same block of @xmath407 if they lie in the image of the same disc of the domain of @xmath275 . our assumption that @xmath409 implies that there are at least two discs in the domain of @xmath275 . so , we may assume that @xmath410 lies in the image of a disc whose end covers @xmath85 a total of @xmath356 times , and that @xmath411 lies in a different disc . since @xmath358 , it follows from monotonicity and our choice of @xmath55 , see remark [ points exist ] , that the disc through @xmath410 intersects at least one other point constraint , say @xmath412 . in other words we can label the point constraints so that @xmath410 and @xmath412 lie in the same block of @xmath407 and this block does nt contain @xmath411 . similarly , for any integer @xmath413 $ ] we can define a partition @xmath414 of the set @xmath408 by considering the maps to @xmath104 constructed from the compactifications of the curves of @xmath125 with level at most @xmath96 , e.g. , @xmath415 . it follows from lemma [ limit ] that for @xmath416 the partition @xmath417 is a ( possibly trivial ) refinement of @xmath414 . hence @xmath410 and @xmath411 lie in different blocks of each @xmath414 . * a distinguished level of @xmath124 . * let @xmath418 in @xmath419 $ ] be the lowest level such that for all @xmath420 satisfying @xmath421 , the curve of @xmath273 at level @xmath96 that corresponds to the block of @xmath414 containing @xmath410 has exactly one positive and one negative end , each of which cover @xmath85 exactly @xmath356 times . equivalently , @xmath418 can be defined as the smallest integer in @xmath419 $ ] such that the block of @xmath414 which contains @xmath410 is the same for all @xmath422 . throughout the proof of theorem [ key ] the cases @xmath423 and @xmath424 will need to be dealt with separately . the latter case will require a significant amount of additional technical work . in preparation for this , we henceforth assume that the constraint points have been relabeled , if necessary , so that the point @xmath412 lies in a different block to @xmath410 for all partitions @xmath414 with @xmath425 . we also assume that the maps @xmath123 converging to @xmath124 have been chosen so that @xmath426 , @xmath427 and @xmath428 , where @xmath429 is the line at infinity in @xmath36 . for each @xmath289 , this choice is not unique since our curves intersect the line at infinity in @xmath29 points . however , we note that ( as required later ) such a choice can be made continuously over generic @xmath68-parameter families of curves , @xmath294 . * a special curve of @xmath124 of level @xmath418 . * we now identify a special curve of @xmath124 of level @xmath418 in terms of its domain . to do this we consider the graphs @xmath430 of the curves @xmath123 . for these maps , we have the following convergence statement . [ graph](theorem 10.6 of @xcite ) a subsequence of the maps @xmath431 converges to a holomorphic building ( see section @xmath432 ) . in this setting , the components of the limiting holomorphic building map to either @xmath433 , @xmath434 or @xmath435 . these are almost - complex manifolds with cylindrical ends , but now each end comes with morse - bott @xmath61-families of closed reeb orbits . ignoring again the passage to a subsequence , let us assume that the maps @xmath431 converges to a holomorphic building with domain @xmath126 . recalling the sequence of maps @xmath174 in the statement of the compactness theorem and looking at the @xmath61-components of the curves , it follows that the sequence of maps @xmath436 converges to a continuous map , say @xmath437 , which has removable singularities at the nodes of @xmath132 . for the @xmath117-components of the graphs , the sequence of maps @xmath438 converges to a holomorphic building , say @xmath125 , with the same domain @xmath126 ( at least if we allow trivial components ) . the map @xmath439 is actually holomorphic of degree @xmath68 since it is a limit of such maps . hence , it is nonconstant on exactly one component of @xmath440 . let @xmath441 be the unique component of @xmath126 on which @xmath439 is nonconstant , and denote the curve of @xmath125 with this domain as @xmath442 and its level by @xmath267 . [ param ] the curve @xmath442 maps into either @xmath141 or @xmath140 , that is , @xmath443 . in fact , with @xmath412 chosen as above , @xmath442 is the curve of level @xmath444 whose compactification lies in the component of @xmath445 containing the points @xmath410 and @xmath412 . if @xmath442 maps into @xmath141 then the values of @xmath439 at the punctures of @xmath441 include @xmath446 . if @xmath423 then there is a curve of @xmath125 with image in @xmath140 that passes through @xmath410 and @xmath412 this curve must be @xmath442 since @xmath439 is nonconstant on its domain which includes both the values @xmath143 and @xmath68 . now suppose that @xmath424 and consider the curve of level @xmath418 that corresponds to the block of @xmath447 containing @xmath410 . this curve has one positive puncture asymptotic to @xmath448 and at least two negative punctures , one connected through a chain of curves to a curve through @xmath410 , and another connected similarly to a curve through @xmath412 ( these punctures are different by the choice of @xmath412 ) . the positive puncture is connected through a chain of curves to a curve through @xmath429 . now , @xmath439 takes the values @xmath143 , @xmath68 and @xmath449 on the domains of the curves of @xmath124 which pass through @xmath410 , @xmath412 and @xmath429 , respectively . therefore , as the function @xmath439 can be nonconstant on at most one of these three distinct chains of spheres , the value of @xmath439 at ( at least ) two of the punctures on our curve at level @xmath418 are different . hence , @xmath439 is nonconstant on the domain of this curve , which must therefore be @xmath442 . it also follows that @xmath439 must be constant on the three chains of curves connected to the curves passing through @xmath410 , @xmath412 and @xmath429 . thus , we see that @xmath442 has punctures at @xmath143 , @xmath68 and @xmath449 . * a first technical aside . * in this aside we show that theorem [ graph ] applies equally well to the graphs of maps satisfying the holomorphic curve equation with respect to a domain dependent almost - complex structure , provided that the domain - dependence is carefully controlled near the neck region . to be precise , let @xmath450 be a sequence of maps which satisfy the holomorphic curve equation with respect to a domain dependent almost - complex structure @xmath451 . that is , in coordinates @xmath452 on @xmath61 , we have @xmath453 in our case , we will always be able to assume that in fact @xmath454 is domain independent unless @xmath455 lies in a fixed subset , say @xmath456 $ ] , of the cylindrical portion of @xmath117 . fix almost - complex structures @xmath457 on @xmath458 such that for @xmath459 and @xmath460 we have @xmath461 where @xmath74 is the standard complex structure on @xmath462 and @xmath463 is a complex structure on @xmath464 . ( so for @xmath465 $ ] we have @xmath466 . ) then the idea , going back to corollary @xmath467 of @xcite , is that @xmath468 satisfies the domain - dependent holomorphic curve equation if and only if its graph @xmath469 is @xmath457-holomorphic . as @xmath470 we require the @xmath457 to converge to an almost - complex structure @xmath471 . the convergence is uniform in the smooth topology on @xmath104 and @xmath159 . in the remaining portion of @xmath117 , @xmath119 $ ] , at least up to translation , we require our @xmath457 to be independent of both @xmath472 and @xmath289 and translation invariant on @xmath473 \cup [ 1 , n))$ ] , and uniformly convergent on @xmath474 $ ] . in this setting , as @xmath470 the complex structure is now stretched as before by gluing copies of @xmath119 $ ] to @xmath159 where the almost - complex structure is extended to be translationally invariant outside of @xmath456 $ ] . this is equivalent to stretching along two copies of @xmath475 ( or one if we have translation invariance on all of @xmath476 ) . then the compactness theorem of @xcite holds as before [ depgraph](@xcite ) a subsequence of the maps @xmath477 converges to a holomorphic building . the curves of the limiting building map into one of @xmath433 , @xmath434 or @xmath435 , where the almost - complex structure on @xmath434 is now either translation invariant or else translation invariant outside of a compact set . moreover , the curves which map into the copy of @xmath434 whose almost - complex structure is not translationally invariant , all have the same level . the sequence of maps @xmath436 converges to a continuous map @xmath478 , and the sequence of maps @xmath479 converges to a holomorphic building , say @xmath480 , with the same domain @xmath481 ( again allowing for trivial components ) . in local coordinates , the curves of @xmath480 now satisfy equation with @xmath451 replaced by its limit @xmath55 , and @xmath472 replaced by @xmath482 . * partitions associated to the @xmath123 . * we now define partitions of the @xmath408 which are intrinsically associated to the sequence of curves @xmath123 converging to @xmath124 ( rather than referring to the limiting building @xmath124 , itself ) . this can be done by selecting a sequence of subsets @xmath483 . [ partitiondef ] for sequences @xmath123 and @xmath484 as above , the partition @xmath485 of the set @xmath408 is defined by saying that two points @xmath278 and @xmath486 lie in the same block if for all large @xmath289 there exists a continuous path in @xmath487 which starts at @xmath488 and ends at @xmath489 . here the @xmath490 are as in the definition of @xmath491 from section [ ddeeff ] . as we now describe , the partition @xmath447 ( in fact any of the @xmath414 ) corresponds to certain partitions defined as in definition [ partitiondef ] . for the case @xmath424 , consider the sequence of neighborhoods of the form @xmath492 where @xmath493 and the @xmath494 are the shifts that determine the convergence of the @xmath123 at level @xmath495 ( see section [ compact ] ) . for large enough @xmath289 , each @xmath484 lies in the cylindrical portion @xmath496 of @xmath117 . the sets @xmath497 converge to a subset of @xmath498 , the image of the level @xmath495 curves of @xmath125 . in particular , modulo translations , the sets @xmath484 can be identified with the set @xmath499 in the copy of @xmath141 which acts as the target of the level @xmath495 curves of @xmath125 . choosing @xmath500 to be sufficiently large , we may assume that the sets @xmath501 converge to a collection of annuli , one in each component of @xmath502 , which are arbitrarily close to the positive asymptotic puncture in each component . that is , the components of the complements of these annuli that contain the positive punctures are precisely punctured disks . the convergence of @xmath123 to @xmath125 implies that @xmath503 . if @xmath504 , then @xmath505 and we recover @xmath447 , as above , by simply setting @xmath506 for sufficiently large @xmath507 . * another building realizing @xmath356 . * we now observe that there exists another limiting building @xmath508 whose curve in @xmath134 agrees with that of @xmath124 , but which determines a different family of partitions of @xmath509 . to see this , recalling section [ varypoint ] , consider the space @xmath290 determined by a family of paths @xmath284 in @xmath104 which switches @xmath410 and @xmath411 and leaves the other points fixed . more precisely , suppose that @xmath363 for all @xmath74 , @xmath510 , @xmath511 , and @xmath512 for all @xmath513 . let @xmath294 be a @xmath68-parameter family of curves representing a component of @xmath290 such that @xmath514 = [ f_n]$ ] . set @xmath515 . passing to subsequences we may assume that the @xmath516 also converge to some @xmath508 . by proposition [ path2 ] , the curves of @xmath124 and @xmath508 with image in @xmath134 are the same . on the other hand , by our choice of the paths @xmath284 , it follows from proposition [ path1 ] that the points @xmath410 and @xmath412 are in different blocks of the partition @xmath517 of @xmath509 which corresponds to @xmath508 . hence @xmath518 . more importantly , we observe the following . [ different ] for @xmath519 the partition @xmath520 is not equal to @xmath521 . in particular , the points @xmath410 and @xmath412 lie in different blocks of @xmath520 . [ diffferent ] here we are using the same subsets @xmath484 to define both partitions ( even though the subsets were originally defined in terms of @xmath123 , not @xmath516 ) . to see this , we argue by contradiction . define @xmath334 such that the level @xmath144 component @xmath133 of @xmath125 intersects @xmath335 transversally and restricted to the preimage of @xmath522 the map @xmath133 is , up to parameterization , @xmath343 close to trivial holomorphic cylinders over the reeb orbit . assume there exists a path @xmath523 in @xmath524 which connects @xmath525 and @xmath526 . we observe that in particular @xmath200 has image lying in @xmath527 since in any @xmath117 the sets @xmath484 separate @xmath104 from @xmath159 . let @xmath294 be the family of curves used to define @xmath508 above . by proposition [ path2 ] , see also remark [ closeimage ] , for large @xmath289 the sets @xmath528 have the same topology relative to the marked points . therefore the path @xmath529 can be included in a homotopy of paths @xmath530 connecting @xmath531 to @xmath532 within @xmath533 . but then @xmath534 connects @xmath535 and @xmath532 in @xmath536 , which contradicts the fact that @xmath411 and @xmath412 are in different blocks of @xmath407 . * strategy of the proof of theorem [ key ] . * for the sake of clarity , we briefly outline the plan for the remainder of the proof . at present , we have two convergent sequences of curves , @xmath123 and @xmath516 , whose limiting buildings have the same curve with image in @xmath134 , and this curve has a negative end that covers @xmath85 exactly @xmath356-times , the maximum covering number among all such curves in such limits . moreover , by construction , the corresponding partitions @xmath521 and @xmath520 of the constraint points differ , in that @xmath410 and @xmath412 lie in the same block of the former partition , and in different blocks of the latter . we now forget the constraint point @xmath411 . the sequences of maps @xmath123 and @xmath516 , and the sets @xmath484 , still induce different partitions of the remaining point constaints @xmath537 . by lemma [ connected ] , for each @xmath289 the maps @xmath123 and @xmath516 can be connected through a path of maps @xmath538 , for @xmath286 , such that @xmath539 \in { \mathcal m}_d(j^n , p_1 , p_3 , \dots , p_{3d-1})$ ] . for all @xmath289 and @xmath304 , the subset @xmath540 is disjoint from @xmath541 . since the partitions of @xmath542 determined by components of the complements of the @xmath543 are different for @xmath544 and @xmath545 , we conclude that the topology of @xmath543 must change with @xmath304 . this change will be the eventual source of our contradiction to the assumption that @xmath356 is both maximal and less than @xmath32 . to exploit this change we must first arrange that @xmath442 , the curve of @xmath125 of level @xmath418 that corresponds to the block of the partition containing @xmath410 , is isolated in @xmath141 or @xmath546 . ( recall that level @xmath418 curves include the limits of the curves through the @xmath484 . ) proposition [ index ] implies that @xmath442 must have deformation index @xmath143 . however , in the case @xmath424 this does not imply that @xmath442 is isolated . indeed , multiple covers of cylinders in @xmath141 with positive and negative ends covering @xmath85 an equal number of times have deformation index zero but are never isolated as they admit a free @xmath547-action . the underlying trivial cylinders themselves also have deformation index zero and therefore can not be removed by generic perturbations of the almost - complex structure on @xmath141 . to avoid this possibility we use almost - complex structures which are domain - dependent on the preimages of the @xmath484 to obtain a rigid perturbation of @xmath442 . this procedure is the subject of the forthcoming technical aside . ( the fact that the resulting rigid curve appears in a limiting holomorphic building corresponding to a sequence of holomorphic spheres , follows from theorem [ depgraph ] of our first technical aside . ) now , by compactness , there are only finitely many images of isolated curves ( with energy bounded by a constant ) in a copy of @xmath141 or @xmath546 . it follows that for large @xmath289 and some @xmath548 that @xmath549 must lie a fixed distance ( in a sense to be made precise ) from the images of isolated curves intersected with @xmath484 ( thinking now of @xmath484 as a subset of an @xmath141 ) . otherwise we could not have the change of topology detected above . the limiting building of a convergent subsequence of the @xmath390 therefore includes a curve which lies a fixed distance from all isolated curves and hence is not isolated itself . therefore it will have positive deformation index , and adding back the point constraint @xmath411 to this curve we still have a curve with nonnegative constrained deformation index . but the curve already corresponds to the block of the partition @xmath550 that includes @xmath410 and has @xmath356 elements . adding @xmath411 it now generates a partition containing a block of size @xmath360 . this will give a contradiction to the hypothesis that @xmath356 is maximal . * a second technical aside . * to complete the proof of theorem [ key ] we will need the distinguished curve @xmath442 discussed in lemma [ param ] to be rigid , that is , isolated and regular , and in particular of deformation index zero . this is automatic if @xmath442 maps to @xmath140 ( i.e. , @xmath423 ) and the constraint points are in generic position . let us assume then that @xmath424 in which case @xmath442 maps to @xmath141 and is a multiply covered cylinder with one positive and at least two negative punctures . ( it is not a trivial cylinder by the definition of @xmath418 above . ) to achieve rigidity for @xmath442 we will perturb it so that the resulting curve @xmath551 is somewhere injective , and is a regular solution of a cauchy - riemann type equation with some domain dependence . this process is carried out carefully in the following paragraphs , and is possible precisely because @xmath442 is the unique curve of @xmath124 on which the map @xmath439 is nonconstant . the curve @xmath551 will have the same asymptotic behavior as @xmath442 , and the other curves of @xmath125 will be unchanged by this procedure . in particular , the new limiting building will still be a limit of holmorphic spheres ( holomorphic with respect to a domain dependent almost - complex structure ) in the sense of theorem [ depgraph ] . it will also have the same curve of level @xmath144 as @xmath125 ( that is , the same curve mapping to @xmath134 ) with a negative end covering the orbit @xmath85 precisely @xmath356 times . let @xmath552 . as above , up to translation we identify @xmath553 with the sets @xmath554 and with the subset of the copy of @xmath141 that corresponds to the target of the level @xmath418 curves of @xmath125 . for large @xmath555 the set @xmath556 is arbitrarily close to the positive puncture in the domain of @xmath442 . by lemma [ param ] , @xmath442 is asymptotically cylindrical to @xmath448 at this puncture . since @xmath557 is a diffeomorphism on the domain of @xmath442 , we will use @xmath439 to identify this domain with its image in @xmath61 , whose points will be denoted by @xmath472 . let @xmath551 be a small perturbation of @xmath442 on the preimage of @xmath553 such that @xmath558 near the boundary of @xmath556 but @xmath551 is not a multiple cover . we can then define a domain dependent almost - complex structure @xmath559 on @xmath560 , ( that is , a map @xmath561 defined for @xmath562 and @xmath563 ) such that @xmath551 becomes @xmath559-holomorphic . in other words , for local coordinates @xmath452 on @xmath564 , as in equation we have @xmath565 this is possible since on intersecting branches of @xmath551 the @xmath472 coordinates are necessarily different . we extend @xmath551 such that @xmath566 outside of @xmath553 , and extend @xmath559 such that @xmath567 for @xmath568 . in particular @xmath567 if @xmath472 is close to a puncture , for example if @xmath472 is close to @xmath143 , @xmath68 or @xmath449 . since @xmath439 is constant on the domains of all other curves of our limiting building @xmath124 they can also be viewed as solutions of the @xmath559-holomorphic equation , which are unaffected by the domain dependence . the @xmath559-holomorphic curve @xmath551 has the same asymptotic limits as @xmath442 and thus still has virtual deformation index zero , this time in the space of parameterized spheres whose set of punctures includes the points @xmath143 , @xmath68 and @xmath449 . we note that the reparameterization group @xmath569 does not act on parameterized holomorphic maps , so to remain in a moduli space of deformation index zero we need to impose the three point constraints . however , @xmath551 is no longer multiply covered and so , for a generic choice of @xmath559 , it will be rigid and hence isolated from other maps of the sphere to @xmath141 which satisfy the new cauchy - riemann equation , and have bounded energy , deformation index equal to zero and punctures at @xmath143 , @xmath68 and @xmath449 . we recall from lemma [ param ] that limits of parameterized holomorphic curves on which @xmath439 is nonconstant do have punctures at @xmath143 , @xmath68 and @xmath449 , thus @xmath551 is isolated from other such limits . let @xmath570 be the holomorphic building obtained from @xmath125 by replacing @xmath442 by @xmath551 . by the nature of their convergence as parameterized maps to the original @xmath442 and @xmath55 , both @xmath123 and @xmath116 can be perturbed analogously to obtain a sequence of curves @xmath468 which converge to @xmath551 in the sense of theorem [ depgraph ] and are holomorphic with respect to a sequence of domain dependent almost - complex structures @xmath451 which converge to @xmath559 on @xmath553 . more precisely , for @xmath571 we can set @xmath572 and @xmath573 , and for @xmath574 we perturb @xmath468 to approach @xmath551 and define @xmath451 to approach @xmath559 and such that the @xmath468 are @xmath451-holomorphic . note that this can be done to give smooth ( domain dependent ) almost - complex structures @xmath451 . a concrete way to do this would be to define @xmath551 by adding to @xmath442 a multi - valued section @xmath156 of the normal bundle to the image of @xmath442 , where all branches of the section have support in @xmath553 . for @xmath289 large the images of the @xmath123 , up to translation , are themselves multi - valued sections of this same normal bundle and can therefore be perturbed by adding the same @xmath156 once we identify the relevent branches to ensure convergence . the curves @xmath575 which converge to the limiting building @xmath576 can now be obtained from the @xmath468 as before ( by following paths of points @xmath284 ) but now using the domain dependent almost - complex structures @xmath451 ( see the first technical aside ) . as we work now with parameterized curves , for the indices to correspond we must fix parameterizations as above so that the @xmath577 satisfy @xmath578 @xmath579 and @xmath580 . by proposition [ path2 ] , the limit @xmath576 of the @xmath575 will still have the same curve as @xmath570 in @xmath134 . * back to the proof of theorem [ key ] . * to simplify the notation we now drop the tildes from the perturbed versions of @xmath55 , @xmath442 , and @xmath124 defined in our second technical aside . we just retain the fact that the level @xmath418 curve @xmath442 of @xmath125 which corresponds to the block of @xmath447 containing @xmath410 , is now regular , isolated , and has deformation index zero . as described previously , for sufficiently large @xmath289 the components of @xmath581 and @xmath582 determine partitions @xmath521 and @xmath520 of the set @xmath583 . we also recall ( see lemma [ different ] and remark [ diffferent ] ) that these partitions are distinct since @xmath410 and @xmath412 are in the same block of @xmath521 but are in different blocks of @xmath520 . hence , if we forget the point @xmath411 , then @xmath521 and @xmath520 still determine distinct partitions , say @xmath407 and @xmath517 , of @xmath537 . by proposition [ connected ] , we can connect @xmath123 and @xmath516 by a continuous path of holomorphic curves @xmath294 through the points @xmath584 . we may assume that the @xmath294 are all parameterized such that @xmath143 , @xmath68 and @xmath449 map to @xmath410 , @xmath412 and @xmath429 as before . when @xmath289 is large enough , the components of @xmath585 determine a family of partitions @xmath586 of @xmath537 such that @xmath587 and @xmath588 . let @xmath589 be the restriction of @xmath294 to the component of @xmath585 which corresponds to the block of @xmath586 that contains @xmath410 . since the partitions @xmath590 and @xmath591 have different blocks containing @xmath410 , the @xmath589 must vary nontrivially with @xmath304 . in fact , the topology of the domain of the @xmath589 must change for some @xmath304 . we now use this transition to obtain a contradiction to the assumption that @xmath592 . henceforth , we only consider values of @xmath289 which are large enough for the maps @xmath589 to be defined as above . for values of @xmath304 near @xmath143 , each @xmath589 is defined on an annulus whose boundary components are mapped to different components of @xmath593 . if we make an identification between their domains and the standard annulus @xmath594 \times s^1 $ ] , then these @xmath589 define elements in the space @xmath595 \times s^1 , a_n ) \mid a(\{\pm1\ } \times s^1 ) \subset \{s_n\pm1\ } \times \partial e \right\}.\ ] ] consider the topology on this space determined by the distance function @xmath596 where the infimum is over smooth diffeomorphisms @xmath597 of @xmath594 \times s^1 $ ] . note that this also determines a well - defined ( yet degenerate ! ) distance function on the maps @xmath589 , for values of @xmath304 near @xmath143 , since it is independent of how we identify their domains with the standard annulus . as described above , for some @xmath304 the domain of @xmath589 must no longer be an annulus . hence , for all sufficiently small @xmath598 $ ] there is a minimal @xmath599 such that the distance between @xmath600 and @xmath601 is at least @xmath333 . ( in order for the topology of its domain to change , the derivative of @xmath589 at the boundary must at some time be tangent to @xmath593 . if we assume that @xmath601 is transverse to this boundary then at that time we can bound below the distance between @xmath589 and @xmath601 . ) let us fix such an @xmath602 $ ] and set @xmath603 . taking a convergent subsequence of the @xmath390 as @xmath604 , we obtain a holomorphic building @xmath605 . the domain of the building again comes with a map @xmath606 to @xmath61 which is constant on all but one component of the domain and extends continuously across the nodes . some care must be taken with the meaning of the deformation index of the components of @xmath605 . according to lemma [ param ] , the component of a limit on which @xmath606 is nonconstant has punctures at @xmath607 . therefore when we compute the index of the curve on whose domain @xmath606 is nonconstant , we work in a moduli space consisting only of curves with punctures at @xmath607 . ( the value of @xmath606 at other punctures is allowed to vary . ) for a building to arise as a limit it is necessary that the coordinate function be continuous . therefore , once @xmath606 is given on the nonconstant component it is determined on all other components ( since the domain @xmath132 is connected ) . the virtual indices of these other components ( where there is a @xmath569 action ) are defined as before . with these conventions the index formulas for components of @xmath605 are given by the same formulas as before . in particular , the indices of all component curves of @xmath605 must sum to @xmath198 , since this is the index of curves in the original moduli space . ( we have forgotten one point constraint . ) hence , in the generic case when all components have nonnegative index , at most one curve of @xmath605 has index @xmath198 . we observe here that by our choice of @xmath388 , a curve of @xmath605 in @xmath134 has a negative end which connects to a component in @xmath608 that passes through @xmath356 fixed points . in particular , for @xmath609 , the domain of @xmath589 is still an annulus and therefore still bounds a disc in the domain of @xmath294 which intersects the same @xmath356 marked points as in the block of @xmath407 containing @xmath410 . let @xmath610 be the curve of @xmath605 with image in @xmath141 or @xmath140 that includes the limits of the @xmath600 . ( as this copy of @xmath141 , or @xmath140 , contains @xmath553 , the complex structure is not translation invariant , so @xmath610 is in fact well defined as a map to @xmath141 or @xmath140 , not just up to translation . ) as @xmath611 , we have @xmath612 . nevertheless , we can assume that @xmath610 is close enough to @xmath442 that it s tangent planes can not be holomorphic for a domain independent almost - complex structure ( for example , because there may be self - intersection points of negative index ) , in other words @xmath610 will be the unique component of @xmath605 on which the coordinate function @xmath606 is nonconstant . when @xmath423 , we define @xmath613 to be the set of holomorphic maps of genus @xmath143 into @xmath140 which have index zero , energy less than @xmath614 and the following third property : the preimage of @xmath553 under the map has an annular component so that the restriction of the map to this component lies within distance @xmath29 of the @xmath601 ( as defined above ) for all @xmath289 large . for the case @xmath424 , we define @xmath613 to be the the space of genus @xmath143 maps into @xmath141 which have punctures at @xmath615 , are holomorphic with respect to the domain dependent almost - complex structure , and have the same three additional properties . since we have arranged the @xmath601 not to be multiply covered in both cases , we can choose @xmath29 so that the third property above implies that the curves in @xmath613 are somewhere injective . thus , @xmath613 is a finite set since it is compact and has dimension zero . for a suitably choice of @xmath616 ( and any choice of @xmath35 ) , we may assume that @xmath613 contains the curve @xmath442 . for any curve in @xmath613 which does not coincide with @xmath442 , its restriction to the annular components in the preimage of @xmath553 all lie more than some fixed distance @xmath617 away from the corresponding restriction of @xmath442 . choosing the parameter @xmath333 in the definition of @xmath610 above to be smaller than @xmath618 , we conclude that @xmath610 is not in @xmath613 . however , @xmath610 satisfies all of the defining criteria for @xmath613 except that of having deformation index zero . thus its deformation index is nonzero , @xmath610 is not rigid , and @xmath619 . [ endofproof ] the sequence converging to @xmath605 seems to already give a contradiction as there is a single component in @xmath273 now with ( unconstrained ) deformation index @xmath620 . however we must be a little careful . the number @xmath356 was defined as a maximum only over limits of curves passing through the original points @xmath621 ; in the situation at hand it may not be clear even that the domain of our relevent component of @xmath273 is a topological disk . thus the additional argument supplied below is perhaps necessary . we now introduce a point @xmath622 to the image of @xmath600 . in the limit we may assume that these points converge to a point @xmath623 in the image of @xmath610 which lies a distance @xmath333 from all rigid curves . then in the moduli space of curves passing through @xmath32 points ( when we include @xmath622 or @xmath624 ) the curves @xmath390 and the limiting curve @xmath610 can be assumed to be rigid . there is now a component of each @xmath390 restricted to @xmath625))$ ] passing through @xmath360 points . to finally derive a contradiction we must find a limit through the original set of points @xmath626 . to do this , we proceed as follows . we extend @xmath622 to a path @xmath627 such that @xmath628 . this can be done generically for each @xmath289 and to satisfy the following conditions : 1 . for @xmath289 large all @xmath629 remain a bounded distance from the images of rigid curves in @xmath484 ( when identified with @xmath553 ) ; 2 . for @xmath289 large all @xmath629 remain a bounded distance from the intersection of rigid curves through a subset of the @xmath630 with @xmath104 ; 3 . for @xmath289 large all @xmath629 remain a bounded distance from @xmath631 $ ] in @xmath474 $ ] . the sets above that the @xmath629 must avoid are all of codimension @xmath198 and so such paths @xmath629 do exist . we can assume by exponential convergence that rigid curves in the copy of @xmath141 containing @xmath553 are all very close to the cylinders @xmath632 outside of @xmath553 ( replacing @xmath553 by a longer cylinder if necessary ) . there are corresponding families of curves @xmath294 for @xmath286 through @xmath633 connecting @xmath390 to another curve @xmath634 that passes through @xmath635 . [ endpf22 ] let @xmath636 $ ] be a sequence and suppose that the sequence of curves @xmath637 converges in the sense of section @xmath432 . then the limiting holomorphic building has the same components as @xmath605 in @xmath134 . claim [ endpf22 ] implies that for the @xmath68-parameter family of curves connecting @xmath390 and @xmath634 , the intersections of the images of this family with any fixed compact subset @xmath334 ( which can be canonically identified with a subset of each @xmath117 for @xmath289 large ) are approximately constant when @xmath289 is large , see remark [ closeimage ] . hence , the components of the intersection of these curves with the complement of @xmath281 determine identical partitions of the sets @xmath638 . for @xmath289 sufficiently large , it follows that the image of the curve @xmath634 has a component in @xmath639 passing through @xmath360 points . looking at the limit of the @xmath634 and choosing @xmath281 such that our limits are approximately cylindrical on @xmath522 , we find a component in @xmath608 that passes through @xmath360 disjoint balls , implying that it must cover @xmath85 at least @xmath360 times . this contradicts the definition of @xmath356 as the maximum such covering number . _ proof of claim [ endpf22 ] . _ this follows proposition [ path2 ] closely . we just need to ensure that in any limit of @xmath637 the sum of the unconstrained indices of the components in @xmath608 is at least @xmath403 . the components intersecting @xmath630 have indices summing to at least @xmath640 . suppose that , following the notation of section [ compact ] , @xmath641 converges to a point @xmath472 in the limiting building . then by our choice of @xmath642 ( avoiding cylinders over reeb orbits , condition @xmath643 ) the point @xmath472 is not a node . if @xmath472 lies in a component mapping to @xmath140 which passes through @xmath142 other constraint points , then by condition @xmath644 the limiting component has index at least @xmath645 and our sum is as required . if @xmath472 lies in a component mapping to the copy of @xmath141 containing @xmath553 , then by condition @xmath646 this component has index at least @xmath198 . recall from lemma [ g ] that any curves of deformation index zero in a copy of @xmath141 with its translation invariant almost - complex structure are multiple covers of the trivial cylinder over @xmath85 . thus by condition @xmath643 if @xmath472 lies in a component mapping to a copy of @xmath141 with its translation invariant almost - complex structure then again the component has index at least @xmath198 and in all cases we can conclude as required . let @xmath647 be the product manifold @xmath648 equipped with the split symplectic form @xmath649 . when convenient , we will equip the @xmath650factor of @xmath60 with complex coordinates @xmath651 . we will also consider the @xmath652-dimensional torus , @xmath653 , acting on @xmath650 in the standard way by rotations in the @xmath654directions . let @xmath655 be the ellipsoid @xmath656 suppose that for any @xmath14 there exists a symplectic embedding @xmath657 to prove theorem [ thm2 ] we must show that this implies that @xmath658 . fix an integer @xmath659 , and a positive real number @xmath17 such that @xmath61 is irrational and @xmath660 . set @xmath661 . for @xmath662 $ ] , let @xmath663 be a smooth family of symplectic embeddings @xmath664 such that : * for @xmath304 in some neighborhood of @xmath68 the embeddings @xmath663 are just the restrictions of @xmath229 to @xmath665 ; * @xmath666 coincides with the inclusion of @xmath667 into @xmath668 . in particular , the image of @xmath666 is chosen to be invariant under the action of @xmath653 on @xmath60 . in what follows , we will identify @xmath665 with its image @xmath669 . the standard contact form on @xmath670 has @xmath671 simple closed reeb orbits , whose images correspond to the intersections @xmath672 for @xmath673 for simplicity , we will denote the closed reeb orbit on @xmath670 with the smallest action by @xmath85 , i.e. , we suppress the dependence on @xmath304 . the action of @xmath85 is @xmath674 and the conley - zehnder index of its @xmath96-fold cover , @xmath675 , is given by @xmath676 let @xmath677 be the space of smooth almost - complex structures on @xmath60 which are compatible with the symplectic form on @xmath60 and are also compatible with @xmath665 in the sense of section [ split ] . as described in that same section , @xmath678 determines an almost - complex structure on @xmath679 which we will denote by the same symbol . consider a @xmath680holomorphic curve in @xmath679 which has genus zero , degree @xmath29 , and @xmath681 negative ends asymptotic to multiples of @xmath85 such that the @xmath194 such end covers @xmath85 a total of @xmath682 times . the virtual dimension of the moduli space represented by this curve is @xmath683 now , let @xmath680 be a smooth family of almost - complex structures such that @xmath680 belongs to @xmath684 for all @xmath662 $ ] . we define @xmath685 to be the moduli space of somewhere injective @xmath680holomorphic planes in @xmath679 which have finite energy , degree @xmath29 , and whose negative end is asymptotic to @xmath686 . since @xmath687 , the index formula above implies that each @xmath685 has virtual dimension @xmath688 here we prove the following result . [ nonempty ] for sufficiently large @xmath14 and every regular almost - complex structure @xmath690 in @xmath691 , the corresponding moduli space @xmath689 is an oriented , compact , zero - dimensional manifold whose oriented cobordism class is nontrivial . it suffices to prove that the conditions on @xmath689 are satisfied for any regular almost - complex structure @xmath690 in @xmath691 . we will restrict our attention to the subset @xmath692 of @xmath691 consisting of almost - complex structures which are invariant under the @xmath653action on @xmath60 . note that for @xmath693 the submanifold @xmath694 is @xmath690-holomorphic . hence , theorem [ key ] implies that , for a suitable choice of the restriction of @xmath690 to @xmath695 , the space @xmath689 is nonempty . in particular , the theorem yields a curve with image in @xmath696 that represents a class in @xmath689 . [ reg ] for @xmath690 in @xmath692 , let @xmath133 be an immersed @xmath690-holomorphic curve in @xmath697 whose image is contained in the subset @xmath698 . the linearized normal bundle @xmath699 for the image of @xmath133 splits as a sum of holomorphic line bundles . in particular , @xmath700 where @xmath165 is the subbundle of vectors parallel to @xmath701 and @xmath702 is the subbundle of vectors along @xmath133 parallel to the @xmath654 factor . let @xmath703 denote the natural projection @xmath704 . this is of course holomorphic with respect to the standard integrable complex structure @xmath74 . let @xmath690 be our ( stretched ) almost - complex structure on @xmath697 and also the induced complex structure on @xmath699 . it is clear that @xmath165 is a holomorphic subbundle . since the derivative of the projection is holomorphic along @xmath133 the fibers of the @xmath702 are complex . but we can choose a connection on @xmath702 whose horizontal subspaces are all invariant under rotations in the @xmath654 plane . then if @xmath705 is the lift of a vector from the base @xmath133 , the projection of @xmath706 to the fibers is also invariant under rotation . but the component of this projection in the @xmath707 factors varies linearly with @xmath708 and is fixed under rotation , and so must be identically zero . it follows that @xmath702 is a holomorphic subbundle as required . now , for @xmath690 in @xmath692 which restrict to generic almost - complex structures on @xmath709 , curves which represent classes in @xmath689 and whose images lie in @xmath709 are immersed , and by proposition [ reg ] the linearized operator along such curves splits . thus we can apply the ( four - dimensional ) automatic transversality results of @xcite to conclude that such curves are regular ( in particular our curves satisfy the hypotheses of @xcite theorem @xmath68 since the @xmath710 term there is negative ) . we recall that to compute the difference in orientation of two curves in the @xmath143 dimensional moduli space @xmath689 we must identify their normal bundles and examine a family of linear almost - complex structures interpolating between the two induced structures . the difference is given by a sum of crossing numbers evaluated at points where the associated cauchy - riemann operators have nontrivial cokernel ( see for example @xcite , remark 3.2.5 ) . but the automatic regularity here means that there are no such irregular almost - complex structures and thus all curves have the same orientation . observe that if @xmath711 is regular for all curves representing classes in @xmath689 , then the images of these curves are _ all _ contained in @xmath712 this is because any such curve whose image is not contained in @xmath709 would have deformation index zero but , by invariance , would appear in a family of curves of dimension at least one . hence , it follows from theorem [ key ] and proposition [ reg ] that to prove proposition [ nonempty ] it suffices for us to establish the existence of a @xmath690 in @xmath692 which is regular for @xmath689 . a curve which represents a class in @xmath689 will be called _ orbitally simple _ if it intersects at least one orbit of the @xmath653action exactly once . there is an open and dense subset of @xmath713 for which all orbitally simple curves for the corresponding spaces @xmath689 are regular . this follows from the standard methods , exactly as in , say , @xcite @xmath714 , here , the condition of orbital simplicity replaces the assumption in @xcite that all curves are simple . in particular , the analogue of proposition @xmath715 of @xcite allows one to construct sections of bundles over @xmath697 which are both @xmath653invariant and , when restricted to the image of a curve , have support contained in the neighborhood of a single point . thus , to detect the desired almost - complex structure @xmath716 it will suffice to find an open subset of @xmath713 such that every curve for the corresponding spaces @xmath689 is either contained in @xmath709 or is orbitally simple . in fact , as orbital simplicity is an open property ( by compactness ) , it will suffice to construct a single such almost - complex structure . for a suitable @xmath716 , all curves which represent classes in @xmath689 and which do not lie entirely in @xmath698 must intersect some orbit of the @xmath653action exactly once . for a real number @xmath717 slightly larger than @xmath68 , let @xmath718 denote the hypersurface @xmath719 which divides @xmath697 into two regions ; @xmath720 which contains the cylindrical concave end corresponding to @xmath721 , and @xmath722 . we will work with @xmath653invariant almost - complex structures on @xmath697 for which the projection onto @xmath723 is holomorphic on @xmath724 . let @xmath725 denote a sequence of such almost - complex structures which are stretched to a length @xmath289 along @xmath718 . we observe that @xmath718 contains a @xmath726 parameter family of reeb orbits corresponding to translations of @xmath727 . arguing by contradiction , suppose that for all such almost - complex structures @xmath725 there exist curves @xmath123 which represent classes in @xmath689 and intersect @xmath653orbits in multiple points , or not at all . then on @xmath724 the curves @xmath123 project with a degree @xmath728 onto curves @xmath634 lying in @xmath723 . the sequence of @xmath729 is bounded by @xmath29 , as the @xmath123 intersect the line at infinity at least @xmath729 times . so , taking a subsequence , we may assume that @xmath730 is independent of @xmath289 . now , taking a limit as @xmath470 our curves @xmath123 will converge to a holomorphic building whose component in @xmath724 has negative ends asymptotic to translations of @xmath731 . since the component in @xmath720 has a negative end asymptotic to @xmath686 we may assume , since our curves have positive area , that for @xmath717 sufficiently close to @xmath68 the negative ends in @xmath724 cover the orbits a total of at least @xmath32 times . the projected curves @xmath634 have bounded area and so will also converge to a holomorphic building which includes a curve @xmath597 with image in @xmath732 working with @xmath653invariant structures we may still assume that the limiting almost - complex structure on @xmath733 is regular ( indeed , any almost - complex structure preserving @xmath723 can be extended to @xmath724 as a @xmath653invariant one ) . suppose that @xmath597 has degree @xmath614 and @xmath681 negative ends with the @xmath194 such end covering @xmath727 a total of @xmath734 times . using formula , the deformation index of @xmath597 is @xmath735 which is nonnegative only if @xmath736 . now , the degree @xmath614 of @xmath597 is equal to @xmath737 , so the negative ends of the limits of our curves @xmath123 cover the translates of @xmath727 a total of @xmath738 times . thus , if @xmath739 then we have a contradiction as required . for a generic choice of the family @xmath680 , the set @xmath740\}$ ] is an oriented @xmath68-dimensional manifold with boundary . by @xcite , @xmath741 is compact modulo convergence to equivalence classes of holomorphic buildings in the @xmath679 . a priori , such buildings could include multiply covered curves , that is , curves not included in the definition of @xmath741 . in this section we prove [ compactt ] the family @xmath741 is compact . let @xmath124 be a holomorphic building in @xmath679 representing a limit point of @xmath741 . then @xmath124 consists of a finite collection of holomorphic curves with connected domains and images in either @xmath679 or @xmath742 . each of these curves has virtual index zero and their compactifications fit together to form a continuous map @xmath238 from the unit disc to @xmath743 which takes the boundary circle to @xmath686 . arguing as in section [ restrictions ] , one can also show that the curves of @xmath124 with image in @xmath679 have negative ends which are asymptotic only to multiples of @xmath85 , and the curves with image in @xmath742 are multiply covered holomorphic cylinders over @xmath85 . moreover , as inherited by the curves in @xmath741 , there is a unique curve of @xmath124 at the bottom level , and it has a single negative end which covers @xmath85 precisely @xmath32 times . this is because any curves without negative ends necessarily have positive deformation index . our first claim is that there is a single curve of @xmath124 with image in @xmath679 , and that it is somewhere injective . arguing by contradiction , suppose that there are @xmath744 such curves . then they have degrees @xmath745 with @xmath746 and we may assume that the ends of the @xmath194 component cover @xmath85 a total number @xmath747 times . here we allow for some @xmath748 , which corresponds to rational curves without ( nonremovable ) punctures which bubble off in @xmath679 . since the curves of @xmath124 with image in @xmath742 have at least as many positive ends as negative ends , when counted with multiplicity , and the curve of lowest level has a single negative end which covers @xmath85 a total of @xmath32 times , we have @xmath749 . the assumption that @xmath744 then implies that there must exist a @xmath750 $ ] for which @xmath751 . suppose the corresponding curve @xmath133 of @xmath124 is an @xmath96-fold cover of a somewhere injective curve @xmath230 ( if @xmath133 is not a multiple cover then we allow @xmath752 and @xmath753 here ) . then @xmath230 has degree @xmath754 and has some number @xmath755 of negative ends , the @xmath194 of which covers @xmath85 a total of @xmath756 times . hence , @xmath757 and the index formula ( [ indexx ] ) for @xmath230 becomes @xmath758 for generic choices of almost - complex structure we then have a contradiction . thus , there is a single curve @xmath63 of @xmath124 with image in @xmath679 . suppose that this curve has a total of @xmath681 negative ends , each covering @xmath85 a number @xmath734 times . if @xmath63 is an @xmath96-fold cover of a somewhere injective curve @xmath759 with @xmath755 negative ends then , as above , the index formula ( [ indexx ] ) gives @xmath760 but as @xmath761 , we notice that the index can exceed @xmath762 only if @xmath763 and @xmath752 and @xmath764 . in other words , the unique curve @xmath63 of @xmath124 with image in @xmath679 is somewhere injective and does indeed lie in @xmath741 . now assume that @xmath741 is not compact and that @xmath124 represents a limit point not contained in @xmath741 . then @xmath124 is a holomorphic building of height @xmath765 . as described above , the unique curve of @xmath124 of level one has a single negative end which covers @xmath85 precisely @xmath32 times . since @xmath765 , this curve , say @xmath165 , has image in @xmath742 . the compactifications of @xmath63 , @xmath165 , and the other curves of @xmath124 with image in @xmath742 fit together to form a curve of genus zero . hence , the curves with level in @xmath766 must each have one positive and one negative end . if follows that the curves of @xmath124 with image in @xmath742 are all cylindrical , and so @xmath124 and @xmath63 represents the same element of @xmath741 . this contradiction affirms that @xmath741 is compact . theorem [ thm2 ] now follows almost immediately from propositions [ nonempty ] and [ compactt ] . for any @xmath14 with @xmath767 , and any positive integer @xmath29 satisfying @xmath768 , it follows from proposition [ compactt ] that for a generic family @xmath680 the space @xmath740\}$ ] is a compact , oriented , @xmath68-dimensional manifold whose boundary is @xmath769 . proposition [ nonempty ] then implies that the moduli space @xmath770 is nonempty . hence , there exists a holomorphic plane in @xmath771 whose negative end is asymptotic to @xmath686 . the symplectic area of this curve is positive and equal to @xmath33 . thus , @xmath772 , and taking the limit as @xmath29 ( and hence @xmath17 ) goes to @xmath449 we have @xmath773 . in this section we prove theorems [ embed2 ] and [ embed ] . to establish theorem [ embed ] it suffices to find an embedding of @xmath778 into @xmath779 , where @xmath780 is fixed and @xmath333 can be arbitrarily small . the existence of such an embedding is essentially contained in lemma @xmath781 of @xcite . we review the construction here for the sake of completeness and because it will play an important role in the proof of theorem [ embed2 ] . [ push ] for every @xmath784 there is a nonnegative function @xmath165 whose support is contained in @xmath785 , whose maximum value is less than @xmath786 , and whose time-@xmath304 hamiltonian flow , @xmath787 , satisfies @xmath788 and @xmath789 let @xmath790 be any set whose closure is contained in the interior of the square @xmath791 \times [ 0 , \sqrt{\pi + \delta } ] \subset \r^2.$ ] the time-@xmath304 hamiltonian flow of the function @xmath792 on @xmath793 is given by @xmath794 ( we use the convention that the hamiltonian vectorfield , @xmath795 , of @xmath281 is defined by the equation @xmath796 . ) hence , for all @xmath797 the set @xmath798 is contained in the interior of the rectangle @xmath791 \times [ 0 , ( 1+t)\sqrt{\pi + \delta } ] $ ] and @xmath799 . cutting @xmath281 off appropriately near @xmath800 } \phi^t_k(u)$ ] , we get a nonnegative function @xmath801 whose hamiltonian flow still has these properties , but is now supported in @xmath791 \times [ 0 , 2\sqrt{\pi + \delta } ] $ ] and satisfies @xmath802 . one can construct a symplectic diffeomorphism @xmath803 of @xmath793 which maps @xmath791 \times [ 0 , 2\sqrt{\pi + \delta } ] $ ] into @xmath785 and for @xmath295 $ ] maps arbitrarily large subsets of each rectangle @xmath791 \times [ 0 , ( 1+t)\sqrt{\pi + \delta}]$ ] onto balls centered at the origin . ( such maps are described and illustrated explicitly in section @xmath781 of @xcite . ) we choose these arbitrarily large subsets of the rectangles @xmath791\times [ 0 , ( 1+t)\sqrt{\pi + \delta}]$ ] so that they contain @xmath798 for all @xmath295 $ ] . setting @xmath804 and @xmath805 , we are done . [ error1 ] it is clear from the definition of @xmath165 in terms of @xmath801 that for all @xmath806 $ ] the distance between @xmath807 and the boundary of @xmath808 is greater than zero and of order @xmath809 . * the double points are concentrated in an arbitrarily small region around the origin @xmath811 . * the vertical and horizontal sections crossing at @xmath811 are arbitrarily thin ( and hence arbitrarily long if need be ) . * the areas of the regions @xmath553 and @xmath812 are both equal to @xmath813 . let @xmath816 be the symplectic immersion of @xmath778 into @xmath817 which acts as @xmath810 on the first factor and as inclusion on the second . we now alter the image of @xmath818 to obtain the desired embedding . in what follows , we will use coordinates @xmath819 on the plane containing the first copy of @xmath785 and coordinates @xmath820 on the plane containing the second copy . the projection from @xmath821 to the @xmath822-plane will be denoted by @xmath823 the self - intersections of @xmath816 project under @xmath824 to the self intersections of @xmath810 . the @xmath825-coordinates of these points take values in an interval of the form @xmath826 $ ] . for @xmath827 , let @xmath815 be the horizontal portion of the image of @xmath810 which passes through the origin and whose first coordinates satisfy @xmath828 $ ] . to remove the intersections of @xmath816 , we consider the hamiltonian @xmath829 where @xmath165 is the hamiltonian from lemma [ push ] and @xmath830 is a bump - function which equals @xmath68 for @xmath831 and equals @xmath143 for @xmath832 . the time-@xmath68 hamiltonian flow of @xmath833 is @xmath834 let @xmath835 be the symplectic immersion of @xmath778 into @xmath817 obtained by applying @xmath836 to @xmath837 clearly , @xmath835 agrees with @xmath816 away from @xmath838 , and shares none of the original double points of @xmath816 . the immersion @xmath835 can only have new double points in @xmath839 where @xmath840 and @xmath841 are the portions of @xmath815 corresponding to points with @xmath825-values in @xmath842 $ ] and @xmath843 $ ] , respectively . we now show that , for an appropriate choice of @xmath810 and @xmath830 , @xmath835 can be adjusted on @xmath839 so that no new double points occur . in the @xmath822-plane , @xmath836 only moves points in @xmath844 and does so only in the @xmath381direction . moreover , the maximum displacement in this direction is bounded from above by @xmath845 hence , the image of @xmath846 under @xmath824 is contained in the set @xmath847 \times \left[0 , \max(|\chi'|)(\pi+\delta)\right]\right),\ ] ] and the projection of @xmath848 is contained in @xmath849 \times \left [ -\max(|\chi'|)(\pi+\delta ) , 0\right]\right).\ ] ] choosing the width of @xmath815 to be sufficiently small , and @xmath850 sufficiently close to @xmath851 , we may assume that both @xmath839 project to regions in the @xmath822-plane whose area is less than @xmath852 and hence less than the area of each of the regions @xmath553 and @xmath812 . acting on @xmath853 and @xmath854 by another symplectic diffeomorphism which acts nontrivially only in the @xmath822directions , we may then ensure that they are mapped by @xmath824 into @xmath553 and @xmath812 , respectively . the resulting symplectic immersion therefore has no double points and is the desired embedding of theorem [ embed ] . scaling things appropriately , the argument above yields a symplectic embedding of @xmath855 into @xmath856 for any @xmath857 provided that @xmath333 is sufficiently small . in this section , we show that the previous embedding procedure can be refined to obtain a symplectic embedding of @xmath855 into @xmath858 . this will prove theorem [ embed2 ] which implies that theorem [ thm ] is sharp . the bi - disc @xmath856 can be symplectically embedded into @xmath859 , by inclusion . the second ekeland - hofer capacity implies that this is optimal in the sense that @xmath856 can not be symplectically embedded into a smaller ball . as in the proof of theorem [ embed ] , we start with a symplectic immersion @xmath860 of @xmath855 into @xmath856 which acts by an immersion @xmath861 in the first factor , and by inclusion on the second factor . the immersion @xmath74 is chosen so that for some @xmath862 we have : * the vertical and horizontal crossing portions of the image have length equal to @xmath863 . * the regions @xmath553 and @xmath812 have areas in the interval @xmath864 and all but an arbitrarily small amount of this area is concentrated within a distance @xmath865 of the horizontal crossing portion , @xmath866 . set @xmath867 , so that @xmath868 . defining @xmath830 and @xmath833 as in the proof of theorem [ embed ] , we remove the double points of @xmath860 by applying the time-@xmath68 flow of @xmath833 to @xmath869 to obtain a new immersion @xmath870 . choosing @xmath865 and @xmath871 to be sufficiently small , we may assume that the projection @xmath824 maps @xmath872 to the interior of the following region of the @xmath822-plane @xmath873 \times \left [ -\lambda , 0\right ] ) \cup ( [ \lambda,\pi/\lambda ] \times \left[0 , \lambda \right ] ) \right\}.\ ] ] when the width of @xmath866 is small enough , the area of @xmath874 is less than @xmath875 . as in the proof of theorem [ embed ] we can then apply a suitable symplectic map to @xmath872 to shift the relevant parts of its projection into @xmath553 and @xmath812 and hence obtain a symplectic embedding of @xmath855 into @xmath856 . we now refine this embedding procedure by choosing a new immersion of @xmath774 . we begin by analyzing the fibers of the projection map @xmath824 acting on @xmath876 . the points of @xmath876 not in @xmath872 belong to fibres which can all be identified with @xmath877 . the points in @xmath872 belong to fibres of @xmath824 determined by the @xmath825-component of their projections . that is , the fibres of @xmath824 corresponding to a fixed value of @xmath825 can all be identified with a fixed subset of @xmath878 , which we denote by @xmath879 . for @xmath880 , @xmath879 is a fixed subset of the interior of @xmath878 . for @xmath881 , each @xmath879 is contained in the interior of @xmath882 ( see lemma [ push ] ) . we can choose the bump function @xmath883 so that on @xmath884 \cup [ \lambda , \pi/\lambda]$ ] it is arbitrarily @xmath885-close to the function @xmath886 . for all sufficiently small @xmath887 we may then assume that @xmath879 is contained in the interior of @xmath888 for all @xmath889 $ ] . by remark [ error1 ] , the distance from @xmath879 to the boundary of @xmath882 is bounded from below by a positive constant which is independent of @xmath865 and which goes to zero as @xmath96 approaches @xmath12 . for @xmath830 as above , and @xmath865 sufficiently small , we may assume the same is true of the distance from @xmath879 to the boundary of @xmath888 . in this case we denote the lower bound for this distance by @xmath890 . this winding is nearly tight but includes small gaps ( represented by the thicker lines in figure @xmath51 ) to accommodate the winding of the rest of @xmath893 . since the area of @xmath874 is less than @xmath875 , for sufficiently small @xmath326 we may assume that the image of @xmath894 under the winding map still lies in the ball @xmath878 . replacing the embedding @xmath74 in the previous construction with the composition @xmath895 , we get a new symplectic embedding @xmath896 we now show that the image of @xmath897 is contained in @xmath858 . the first approximation here comes from equating the area of the portion of @xmath866 determined by @xmath825 , @xmath904 , with @xmath905 the error terms of correspond , respectively , to the discrepancy caused by the width of @xmath874 , and the discrepancy caused by the gap in the winding . f. bourgeois , a morse - bott approach to contact homology . symplectic and contact topology : interactions and perspectives ( toronto , on / montreal , qc , 2001 ) , 5577 , fields inst . commun . , 35 , amer . soc . , providence , ri , 2003 . h. hofer , k. wysocki and e. zehnder , properties of pseudoholomorphic curves in symplectisations iii : fredholm theory , _ topics in nonlinear analysis _ , 381 - 475 , prog . nonlinear differential equations appl . , 35 , birkhuser , basel , 1999 .
in this paper we establish new restrictions on symplectic embeddings of certain convex domains into symplectic vector spaces . these restrictions are stronger than those implied by the ekeland - hofer capacities . by refining an embedding technique due to guth , we also show that they are sharp .
the rare @xmath2 decays are being measured at @xmath2 factories in kek and slac . the generalized factorization approach has been applied to the theoretical treatment of non - leptonic @xmath2 decays @xcite . it is a great success in explaining many decay branching ratios @xcite . the factorization approach ( f.a . ) is a rather simple method . some efforts have been made to improve their theoretical application @xcite and to understand the reason why the f.a . has gone well @xcite . one of these method is the perturbative qcd approach ( pqcd ) , where we can calculate the annihilation diagrams as well as the factorizable and nonfactorizable diagrams . the rare decays @xmath5 are pure annihilation type decays . in the usual f.a . , this decay picture is described as @xmath2 meson annihilating into vacuum and the @xmath6 and @xmath7 mesons produced from vacuum then afterwards . to calculate this decay in the f.a . , one needs the @xmath8 form factor at very large time like momentum transfer @xmath9 . however the form factor at such a large momentum transfer is not known in f.a . this makes the f.a . calculation of these decays unreliable . the annihilation amplitude is a phenomenological parameter in qcd factorization approach ( qcdf)@xcite , and the qcdf calculation of these decays is also unreliable . in this paper , we will try to use the pqcd approach , where the annihilation amplitude is calculable , to evaluate the @xmath5 decays . by comparing the predictions with the experimental data , we can test the pqcd evaluation of the annihilation amplitude . a @xmath10 boson exchange causes @xmath11 or @xmath12 , and the @xmath13 quarks included in @xmath14 are produced from a gluon . this gluon attaches to any one of the quarks participating in the @xmath10 boson exchange . this is shown in figure [ fig : need1gluon ] . in the rest frame of @xmath2 meson , both @xmath15 and @xmath16 quarks included in @xmath14 have @xmath17 momenta , and the gluon producing them also has @xmath18 . this is a hard gluon . one can perturbatively treat the process where the four quark operator exchanges a hard gluon with @xmath19 quark pair . it is just the picture of pqcd approach . in the next section , we explain the framework of pqcd briefly . in section [ sc : formula ] , we give the analytic formulas for the decay amplitude of @xmath20 decays . in section [ sc : neval ] , we show the predicted branching ratio from the analytic formulas and discuss the theoretical errors . it is found that the prediction are is in good agreement with the data , and pqcd approach correctly gives the annihilation amplitude in consequence . finally , we conclude this study in section [ sc : concl ] . pqcd approach has been developed and applied in the non - leptonic @xmath2 meson decays @xcite for some time . in this approach , the decay amplitude is separated into soft(@xmath21 ) , hard(@xmath22 ) , and harder(@xmath23 ) dynamics characterized by different scales . it is conceptually written as the convolution , @xmath24 , \label{eq : convolution1}\ ] ] where @xmath25 s are momenta of light quarks included in each mesons , and @xmath26 denotes the trace over dirac and color indices . @xmath27 is wilson coefficient which results from the radiative corrections at short distance . in the above convolution , @xmath27 includes the harder dynamics at larger scale than @xmath28 scale and describes the evolution of local @xmath29-fermi operators from @xmath30 , @xmath10 boson mass , down to @xmath31 scale , where @xmath32 . @xmath22 describes the four quark operator and the spectator quark connected by a hard gluon whose @xmath33 is on the order of @xmath34 , and includes the @xmath35 hard dynamics . therefore , this hard part @xmath22 can be perturbatively calculated . @xmath36 is the wave function which describes hadronization of the quark and anti - quark to the meson @xmath37 . while the @xmath22 depends on the processes considered , @xmath36 is independent of the specific processes . determining @xmath36 in some other decays , we can make quantitative predictions here . we consider the @xmath2 meson at rest for simplicity . it is convenient to use light - cone coordinate @xmath38 to describe the meson s momenta , where @xmath39 and @xmath40 . using these coordinates we can take the @xmath2 , @xmath6 , and @xmath7 mesons momenta as @xmath41 , @xmath42 , and @xmath43 , respectively , where @xmath44 and we neglect the @xmath7 meson s mass @xmath45 . putting the light ( anti-)quark momenta in @xmath2 , @xmath6 and @xmath7 mesons as @xmath46 , @xmath47 , and @xmath48 , respectively , we can choose @xmath49 , @xmath50 , and @xmath51 . then , integration over @xmath52 , @xmath53 , and @xmath54 in eq.([eq : convolution1 ] ) leads to @xmath55 , \label{eq : convolution2}\end{gathered}\ ] ] where @xmath56 is the conjugate space coordinate of @xmath57 , and @xmath58 is the largest energy scale in @xmath22 , as the function in terms of @xmath59 and @xmath56 . the large logarithms ( @xmath60 ) coming from qcd radiative corrections to four quark operators are included in the wilson coefficients @xmath27 . the large double logarithms ( @xmath61 ) on the longitudinal direction are summed by the threshold resummation@xcite , and they lead to @xmath62 which smears the end - point singularities on @xmath59 . the last term , @xmath63 , contains two kinds of logarithms . one of the large logarithms is due to the renormalization of ultra - violet divergence @xmath64 , the other is double logarithm @xmath65 from the overlap of collinear and soft gluon corrections . this sudakov form factor suppresses the soft dynamics effectively @xcite . thus it makes perturbative calculation of the hard part @xmath22 applicable at intermediate scale , i.e. , @xmath28 scale . we calculate the @xmath22 for @xmath20 decays in the first order in @xmath66 expansion and give the convoluted amplitudes in next section . in order to calculate analytic formulas of the decay amplitude , we use the wave functions @xmath67 decomposed in terms of spin structure . in general , @xmath67 having dirac indices @xmath68 are decomposed into 16 independent components , @xmath69 , @xmath70 , @xmath71 , @xmath72 , @xmath73 . if the considered meson @xmath37 is @xmath2 or @xmath6 meson , to be pseudo - scalar and heavy meson , the structure @xmath72 and @xmath73 components remain as leading contributions . then , @xmath67 is written by @xmath74 where @xmath75 is color s degree of freedom , @xmath76 is the corresponding meson s momentum , and @xmath77 are lorentz scalar wave functions . as heavy quark effective theory leads to @xmath78 , then @xmath2 meson s wave function can be expressed by @xmath79 \phi_b(x , b).\ ] ] according to ref . @xcite , a pseudo - scalar meson moving fast is parameterized by lorentz scalar wave functions , @xmath80 , @xmath81 , and @xmath82 as @xmath83 where @xmath84 . we ignore the difference between @xmath85 quark s mass and @xmath6 meson s mass in the perturbative calculation . this means , putting @xmath86 , the terms proportional to @xmath87 are neglected . in this approximation , the contributions of eq.([pt ] ) are of higher power than those of eqs.([pv ] , [ ps ] ) by @xmath88 because of the factor @xmath89 in eq.([pt ] ) , and we neglect the @xmath90 component in the @xmath6 meson s wave function . in addition , the eq.([pv ] ) , eq.([ps ] ) , and the relations @xmath91 with equations of motion lead to @xmath92 therefore the @xmath6 meson s wave function can be expressed by one lorentz scalar wave function , @xmath93 \phi_{d_s}(x , b),\ ] ] where @xmath94 is defined by @xmath95 the wave function @xmath96 for @xmath97 meson is normalized by its decay constant @xmath98 @xmath99 in contrast to the @xmath2 and @xmath100 mesons , for the @xmath7 meson , being light meson , the @xmath101 component remains because the factor corresponding to @xmath89 in eq.([pt ] ) is @xmath102 . then , @xmath7 meson s wave function is parameterized by lorentz scalar wave functions @xmath103 as @xmath104_{\alpha\beta}\end{aligned}\ ] ] where @xmath105 , @xmath106 , and @xmath107 . in the numerical analysis we will use @xmath103 which were calculated from qcd sum rule @xcite . they will be shown in section [ sc : neval ] . we first consider the neutral @xmath109 decay @xmath108 . the effective hamiltonian at the scale lower than @xmath110 related to this decay is given as @xcite @xmath111 , \\ o_1 = ( \bar{b}d)_{v - a } ( \bar{u}c)_{v - a } , \quad o_2 = ( \bar{b}c)_{v - a } ( \bar{u}d)_{v - a},\end{gathered}\ ] ] where @xmath112 are wilson coefficients at renormalization scale @xmath113 , and summation in @xmath114 color s index @xmath115 and chiral projection , @xmath116 , are abbreviated to @xmath117 . the lowest order diagrams contributing to @xmath108 are drawn in fig.[fig : diagrams1 ] according to this effective hamiltonian . as stated above , @xmath5 decays only have annihilation diagrams . we get the following analytic formulas by calculating the hard part @xmath22 at first order in @xmath66 . together with the meson wave functions , the amplitude for the factorizable annihilation diagram in fig.[fig : diagrams1](a ) and ( b ) results in @xmath118 , @xmath119 , \label{eq : fa}\end{gathered}\ ] ] where @xmath120 is the group factor of @xmath114 gauge group , and @xmath121 , and the functions @xmath122 , @xmath123 , @xmath124 are given in the appendix . the explicit form for the wave functions , @xmath96 , is given in the next section . the amplitude for the nonfactorizable annihilation diagram in fig.[fig : diagrams1](c ) and ( d ) results in @xmath125 , \label{eq : ma1}\end{gathered}\ ] ] where @xmath126 dependence in the numerators of the hard part are neglected by the assumption @xmath127 . because such a @xmath126 behaves as the cut off , the resultant branching ratio is smaller than it given in ref . @xcite where the whole @xmath126 in the hard part are neglected . ] . the total decay amplitude for @xmath108 decay is given as @xmath128 where the overall factor is included in the decay width with the kinematics factor . the decay width is expressed as @xmath129 the decay width for cp conjugated mode , @xmath130 , is the same value as @xmath108 , just replacing @xmath131 with @xmath132 . since there is only one kind of ckm phase involved in the decay , there is no cp violation in the standard model . the effective hamiltonian related to @xmath1 decay is given as @xmath133 , \\ o_1 = ( \bar{b}d)_{v - a } ( \bar{c}u)_{v - a } , \quad o_2 = ( \bar{b}u)_{v - a } ( \bar{c}d)_{v - a}.\end{gathered}\ ] ] the amplitude for the factorizable annihilation diagram results in @xmath134 . the amplitude for the nonfactorizable annihilation diagram results in @xmath135 . \label{eq : ma2}\end{gathered}\ ] ] thus , the total decay amplitude @xmath136 and decay width @xmath137 for @xmath1 decay is given as @xmath138 the decay width for cp conjugated mode , @xmath139 , is the same value as @xmath1 . similar to the @xmath109 decay , there is also no cp violation in this decay within standard model . in this section we show numerical results obtained from the previous formulas . at the beginning , we give the branching ratios predicted from the same parameters and wave functions that are adopted in the other works . secondly , we show the theoretical errors due to uncertainty of some parameters . for the @xmath2 meson s wave function , there is a sharp peak at the small @xmath140 region , we use @xmath141,\ ] ] which is adopted in ref . this choice of @xmath2 meson s wave function is almost a best fit from the @xmath142 , @xmath143 decays . for the @xmath6 meson s wave function , we assume the same form as @xmath100 meson s one except for the normalization @xcite @xmath144 since @xmath85 quark is much heavier than @xmath15 quark , this function is peaked at @xmath85 quark side , i.e. small @xmath140 region , too . the wave functions @xmath103 of the @xmath7 meson are expanded by gegenbauer polynomials , @xmath145 where @xmath146 . the coefficients @xmath147 calculated from qcd sum rule are given in ref.@xcite , and their values are @xmath148 for @xmath149 . in addition , we use the following input parameters : @xmath150 with these values and eq.([eq : normalization ] ) we get the normalization factor @xmath151 gev . we show the decay amplitudes calculated with the above parameters at table [ tb : amplitudes ] . for the neutral decay @xmath108 , the dominant contribution is the nonfactorizable annihilation diagrams , where the contribution @xmath152 is proportional to the wilson coefficient @xmath153 , which is of order one . the factorizable annihilation diagram contribution is proportional to @xmath154 , which is one order magnitude smaller . for the charged decay @xmath155 , it is the inverse situation . the wilson coefficient in @xmath156 is @xmath157 , which is smaller than the one in @xmath158 , @xmath159 . both amplitudes @xmath160 and @xmath136 are at the same order magnitude here . the propagators of inner quark and gluon in figure 2 and 3 are usually proportional to @xmath161 . one may suspect that these amplitudes are enhanced by the endpoint singularity around @xmath162 . this can be explicitly found in eq.([eq : propagator1 ] , [ eq : propagator2 ] ) , where the bessel function @xmath163 diverges at @xmath162 or @xmath164 . however this is not the case in our calculation . first we introduce the transverse momentum of quark , such that the propagators become @xmath165 . secondly , the sudakov form factor @xmath166 $ ] suppresses the region of small @xmath167 . therefore there is no singularity in our calculation . the dominant contribution is not from the endpoint of the wave function . as a proof , in our numerical calculations , for example , an expectation value of @xmath66 in the integration for @xmath152 results in @xmath168 , which gives the dominant contribution to @xmath108 decay . therefore , the perturbative calculations are self - consistent . .amplitudes(@xmath169 gev ) with parameters eqs . ( [ eq : parm1]-[eq : shapewv ] ) . [ cols="^,>,^ , > " , ] it is found that uncertainty of the predictions on pqcd is mainly due to @xmath170 and @xmath171 . below we show the limits of the branching ratio within the suitable ranges on @xmath170 and @xmath171 . the appropriate extent of @xmath170 can be found in ref.@xcite , @xmath172 @xmath173 in the wave functions @xmath174 are given as functions with respect to @xmath170 , @xmath175 , and some input parameters , @xmath176 , @xmath177 in ref.@xcite . within eq . ( [ eq : m0k ] ) , the branching ratios normalized by the decay constants and the ckm matrix elements result in @xmath178 from the @xmath179 transition form factor @xmath180 , the appropriate extent of @xmath171 can be obtained . @xmath180 calculated from pqcd at @xmath181 gev in the region @xmath182 is consistent with @xmath180 by qcd sum rules given in ref . the branching ratios calculated at the region of eqs . ( [ eq : wb ] ) are found within @xmath183 in the factorizable contribution , the @xmath2 meson s wave function is integrated and normalized by the decay constant @xmath184 . thus , in the factorizable dominant decay , @xmath1 , its branching ratio is insensitive to the change of @xmath171 . in two - body hadronic @xmath2 meson decays , the final state mesons are moving very fast , since each of them carry more than 2 gev energy . there is not enough time for them to exchange soft gluons . the soft final state interaction is not important in the two - body @xmath2 decays . this is consistent with the argument based on color - transparency @xcite . we thus neglect the soft final state interaction in the pqcd approach . the pqcd with sudakov form factor is a self - consistent approach to describe the two - body @xmath2 meson decays . although the annihilation diagrams are suppressed comparing to other spectator diagrams , but their contributions are not negligible in pqcd approach @xcite . in this paper , we calculate the @xmath108 and @xmath1 decays , which occur purely via annihilation type diagrams . the branching ratios are still sizable . the @xmath108 decay has been observed in the @xmath2 factories @xcite . this is the first channel measured in @xmath2 decays via annihilation type diagram . the fact that the predicted branching ratio is in good agreement with the data means that pqcd approach gives the annihilation amplitude correctly , the soft final state interaction is probably small in @xmath2 decays , and it is one of the evidences to justify pqcd approach . we thank our pqcd group members : c .- h . chen , y .- y . keum , e. kou , t. kurimoto , h .- li , t. morozumi , s. mishima , m. nagashima , a.i . sanda , n. sinha , r. sinha , m.z . yang and t. yoshikawa for fruitful discussions . this work is partly supported by national science foundation of china under grant ( no . 90103013 and 10135060 ) . the research of k.u . is supported by the japan society for promotion of science under the predoctoral research program . this work was supported by a grand - in aid for special project research ( physics of cp violation ) and by a grand - in - aid for science research from the ministry of education , culture , sports , science and technology , japan ( no . 12004276 ) . the definitions of some functions used in the text are presented in this appendix . in the numerical analysis we use one loop expression for strong coupling constant , @xmath185 where @xmath186 and @xmath187 is number of active flavor at appropriate scale . @xmath188 is qcd scale , which we use as @xmath189 mev at @xmath190 . we also use leading logarithms expressions for wilson coefficients @xmath191 presented in ref.@xcite . then , we put @xmath192 gev , @xmath193 gev , and @xmath194 gev . the function @xmath195 , @xmath196 , and @xmath197 including wilson coefficients are defined as @xmath198 where @xmath199 and @xmath200 , @xmath201 , and @xmath202 result from summing both double logarithms caused by soft gluon corrections and single ones due to the renormalization of ultra - violet divergence . the above @xmath203 are defined as @xmath204 where @xmath205 , so - called sudakov factor , is given as @xcite @xmath206 , \label{eq : sudakovexpress}\end{aligned}\ ] ] @xmath207 is euler constant , and @xmath208 is the quark anomalous dimension . the functions @xmath124 , @xmath209 , and @xmath210 in the decay amplitudes consist of two parts : one is the jet function @xmath62 derived by the threshold resummation@xcite , the other is the propagator of virtual quark and gluon . they are defined by @xmath211 where @xmath212 , and @xmath213s are defined by @xmath214 we adopt the parametrization for @xmath215 of the factorizable contributions , @xmath216^c,\quad c = 0.3,\ ] ] which is proposed in ref . @xcite . in the nonfactorizable annihilation contributions , @xmath215 gives a very small numerical effect to the amplitude@xcite . therefore , we drop @xmath215 in @xmath209 and @xmath210 . the hard scale @xmath58 s in the amplitudes are taken as the largest energy scale in the @xmath22 to kill the large logarithmic radiative corrections : @xmath217 m. wirbel , b. stech , m. bauer , z. phys . c29 , 637 ( 1985 ) ; m. bauer , b. stech , m. wirbel , z. phys . c34 , 103 ( 1987 ) ; l .- l . chau , h .- y . cheng , w.k . sze , h. yao , b. tseng , phys . d43 , 2176 ( 1991 ) , erratum : d58 , 019902 ( 1998 ) . keum , h .- li and a. i. sanda , phys . b504 , 6 ( 2001 ) ; phys . d63 , 054008 ( 2001 ) . l , k. ukai and m .- z . yang , phys . d63 , 074009 ( 2001 ) ; c .- d . l , page 173 - 184 , proceedings of international conference on flavor physics ( icfp 2001 ) , world scientific , 2001 , hep - ph/0110327 . h . v. chang and li , phys . d55 , 5577 ( 1997 ) ; t .- w . yeh and h .- n . li , phys . d56 , 1615 ( 1997 ) . li , phys . d64 , 014019 ( 2001 ) ; s. mishima , phys . b521 , 252 ( 2001 ) ; e. kou and a.i . sanda , phys . b525 , 240 ( 2002 ) ; c .- h . chen , y .- y . keum , and h .- n . li , phys . d64 , 112002 ( 2001 ) ; c .- d . l and m.z . yang , eur . j. c23 , 275 ( 2002 ) ; a.i . sanda and k. ukai , 107 , 421 ( 2002 ) ; c.d . j. c24 , 121 ( 2002 ) ; c .- h . chen , y .- y . keum , and h .- n . li , phys . d66 , 054013 ( 2002 ) ; m. nagashima and h .- n . li , hep - ph/0202127 ; y .- y . keum , hep - ph/0209002 ; hep - ph/0209208(to appear in prl ) ; hep - ph/0210127 ; y .- y . keum and a. i. sanda , hep - ph/0209014(to appear in prd ) .
the rare decays @xmath0 and @xmath1 can occur only via annihilation type diagrams in the standard model . we calculate these decays in perturbative qcd approach . we found that the calculated branching ratio of @xmath0 agreed with the data which had been observed in the kek and slac @xmath2 factories . while the decay @xmath1 has a very small branching ratio at @xmath3 , due to the suppression from ckm matrix elements @xmath4 . ( 0,0)(-320,-320 ) ( 0,32)*hep - ph/0210206 * ( 0,15)bihep - th-2002 - 47 ( 0,0)dpnu-02 - 31
fidelity , a concept emerging from quantum information theory , has recently become an attractive approach towards the study of critical phenomena in condensed matter physics . in a quantum many - body system , the quantum phase transition is completely driven by the quantum fluctuation in the ground state and is incarnated by an abrupt change in the qualitative structure of the ground state wavefunction as the system varies across the critical point @xcite . therefore , being a measure of the similarity between two states , the fidelity is excepted to show a dramatic change across the transition points . this motivated people to start exploring its role played in quantum phase transitions @xcite . moreover , as the fidelity can be viewed as a space geometrical quantity , no a priori knowledge of the order parameter and symmetry breaking of the system is required . this is thus a great advantage to the study of quantum phase transitions using fidelity approaches . following the streamline of fidelity , some alternative schemes , like the fidelity susceptibility @xcite , fidelity per site @xcite , operator fidelity @xcite , and density - functional fidelity @xcite , have been proposed . as to establish a closer picture to condensed matter physics , we follow the concept of fidelity susceptibility in this paper . mathematically , the fidelity susceptibility is just the leading term of the fidelity . it defines the response of the fidelity to the driving parameter . as a result , the singularity of the fidelity across the transition points could thus be reflected in the divergence of the fidelity susceptibility . in fact , this argument has been consolidated by the results in a number of one - dimensional quantum many - body systems @xcite ( see also a review article @xcite ) . in this paper , we investigate the behavior of the fidelity susceptibility in two two - dimensional ( 2d ) models , namely the 2d transverse field ising model and the xxz model numerically . our results show that the fidelity susceptibility as a function of the driving parameter diverges at the quantum phase transition points in both of the models . the scaling behavior of the extremum of the fidelity susceptibility at the transition point and that of the second derivative of the ground state energy are also compared . from those results , the theoretical argument that the fidelity susceptibility is a more sensitive indicator than the second derivative of the ground state energy in searching for a second order quantum phase transition is testified . besides , it is also found that the fidelity susceptibility shows a scaling behavior in the vicinity of the critical point and its critical exponents for both models are also obtained through finite - size scaling analysis . for a general form of the hamiltonian , @xmath0where @xmath1 is the driving hamiltonian and @xmath2 denotes its strength . the fidelity is the modulus of the overlap between two ground states @xmath3 and @xmath4 @xcite , @xmath5since our focus is on continuous quantum phase transitions , the ground state of the hamiltonian is non - degenerated for a finite system . @xmath6 can thus be obtained from the time - independent non - degenerated perturbation theory . extracting the leading term of the fidelity , the fidelity susceptibility can be expressed as @xcite @xmath7^{2 } } , \label{eq : fspert}\]]where @xmath8 is a set of orthogonal basis satisfying @xmath9 . on the other hand , consider the second derivative of the ground state energy with respect to @xmath2 @xcite , @xmath10one may easily realize that the above expression is very similar to the perturbation form of fidelity susceptibility in eq . ( [ eq : fspert ] ) except having different exponent in the denominator . therefore , one may expect that both the singularity of fidelity susceptibility and the second derivative of the ground state energy are intrinsically due to the vanishing of the energy gap in the thermodynamic limit @xcite . however , the difference in the exponent of the dominator makes fidelity susceptibility a more sensitive quantity in searching for quantum phase transitions . that is to say , while the fidelity susceptibility shows a divergence at the critical point , the second derivative of the ground state energy may still be a continuous function . which can be placed on a square lattice with periodic boundary conditions.,width=302 ] furthermore , to study the scaling behavior of the fidelity susceptibility around the critical point , we may perform finite - size scaling analysis lcvenuti07,guscal . let s consider a system consisting of @xmath11 sites such that @xmath12 , where @xmath13 is the real dimension of the system . around the critical point @xmath14 , the fidelity susceptibility behaves as @xmath15 where @xmath16(@xmath17 ) is the critical exponent of the fidelity susceptibility above ( below ) the critical point , @xmath18 is the quantum adiabatic dimension and hence @xmath19 is an intensive quantity . for a finite system , if the fidelity susceptibility shows a peak at a certain point @xmath20 , it s maximum value scales like @xmath21 where @xmath22 is the critical adiabatic dimension . the above two asymptotic behaviors satisfy @xcite @xmath23 where @xmath24 is a constant , b is a non - zero function of @xmath25 and both of them are independent of the system size . from eq . ( [ eq : scafun ] ) , one can find that the rescaled fidelity susceptibility is an universal function of @xmath26 , @xmath27,\end{aligned}\ ] ] where @xmath28 is the critical exponent of the correlation length . the critical exponent of the fidelity susceptibility can then be obtained as guscal , sjguqad @xmath29 . the inset shows @xmath30 as a function of @xmath31 . the @xmath32-intercept of the line is @xmath33 . , width=302 ] . the inset shows the scaling behavior of the minimum of the second derivative of the averaged ground state energy . the slope of the line is approximately @xmath34.,width=302 ] the hamiltonian of the 2d transverse field ising model @xcite defined on a square lattice reads @xmath35where @xmath36 , @xmath37 , @xmath38 are spin-@xmath39 operators at site * i * , @xmath40 is the transverse field strength in unit of the ising coupling , and the sum @xmath41 runs over the nearest - neighboring pairs on the lattice . periodic boundary conditions are assumed . this model was originally introduced by de gennes to describe potassium - dihydrogen - phosphate type ferroelectrics @xcite and has been studied extensively via various approaches , like real - space renormalization group @xcite , density - matrix renormalization @xcite , numerical diagonalization@xcite , and entanglement @xcite . obviously , the hamiltonian of the model commutes with the parity operator @xmath42 . thus each eigenstate of the hamiltonian is also an eigenstate of @xmath43 . the hilbert space can then be decomposed into subspace @xmath44 where @xmath45 is the eigenvalue of @xmath43 and is specified in each subspace . for a finite system , the ground state of the ising model is non - degenerate in each subspace , thus the perturbation expansion as introduced in the previous section is valid as long as the lattice is finite . in the thermodynamic limit , the model exhibits a quantum phase transition at @xmath46 @xcite . for @xmath47 , the transverse field dominates and the ground state is a paramagnetic phase , with spins almost fully polarized in the @xmath48-direction . for @xmath49 , the ground state is a ferromagnetic phase and is doubly degenerated . to study a model on a two - dimensional square lattice with periodic boundary conditions , we need to construct proper lattice structures that are suitable for exact diagonalization . in this paper , we will diagonalize two models with system sizes @xmath50 , whose structures are shown in fig . fig:2dconfig . the effective length @xmath51 might then be a real number instead of an integer . [ fig : ising_fs_h ] shows the numerical result of the fidelity susceptibility of the ising model on a square lattice for various system sizes . it can be seen that on both sides around the critical point , the averaged fidelity susceptibility is an intensive quantity , i.e. @xmath52 and we have @xmath53 . more importantly , the averaged fidelity susceptibility for different @xmath11 all show a peak at @xmath30 . this peak position of the fidelity susceptibility @xmath30 is plotted as a function of @xmath31 , as shown in the inset of fig . [ fig : ising_fs_h ] . the linear fitting gives @xmath54 in the thermodynamic limit , we obtain @xmath55which gives @xmath56 . comparing this value to the critical point @xmath57 obtained in previous works @xcite , our result here is consistent with them up to two digits . moreover , we can also see that the averaged fidelity susceptibility peaks sharper for a larger @xmath11 and is in fact scales approximately with @xmath58 as shown in the inset of fig . fig : ising_scaling . physically , as the ground state wavefunction of the model changes abruptly across the transition point , the fidelity susceptibility , as a measure of the leading response of the fidelity to the driving parameter , is intuitively expected to show a divergence at the critical point . here , we have shown numerically this is in fact the case and verified the significance of fidelity susceptibility in signaling for the quantum phase transition in the 2d ising model . fig . [ fig : ising_gse ] shows the second derivative of the averaged ground state energy of the ising model for several system size of a square lattice as a function of @xmath40 . as it is well - known that the ising model exhibits a second order phase transition , the second derivative of the averaged ground state energy is expected to show a minimum at the transition point . from the inset of fig . [ fig : ising_gse ] , it is found that the minimum value of the second derivative of the averaged ground state energy scales approximately with @xmath59 . comparing this value with that of the fidelity susceptibility , which is about @xmath60 , we may conclude that the fidelity susceptibility is a more sensitive tool in detecting for a second order quantum phase transition . [ fig : ising_scaling ] shows the finite - size scaling analysis in the case of power - law divergence of the 2d transverse field ising model . the rescaled fidelity susceptibility collapsed to a single curve for various system sizes . the critical exponent of the correlation length can thus be obtained as @xmath61 . together with the slope of the line in the inset of fig . [ fig : ising_scaling ] and from eq . ( [ eq : exp ] ) , the critical exponent of the fidelity susceptibility is found to be @xmath62 only , with the critical exponent @xmath63 . the insert shows the scaling behavior of the maximum of the averaged fidelity susceptibility . the slope of the line is 0.51.,width=302 ] . the inset shows @xmath64 as a function of @xmath31 . the @xmath32-intercept of the line is @xmath65.,width=302 ] . the inset shows the scaling behavior of the minimum of the second derivative of the averaged ground state energy . the slope of the line is approximately @xmath66.,width=302 ] is a function of @xmath67 only , with the critical exponent @xmath68 . the insert shows the scaling behavior of the maximum of the averaged fidelity susceptibility . the slope of the line is 1.81.,width=302 ] for the 2d xxz model , the hamiltonian is given by @xmath69where @xmath70 is the dimensionless parameter characterizing the anisotropy of the model . the sum is over all nearest - neighbors on a square lattice . again , periodic boundary conditions are assumed . for the xxz model in two dimensions , there exists no exact solution . one has to use either approximate analytical approach such as the spin - wave theory or numerical approach such as exact diagonalization studies of a finite lattice . for the latter approach , to obtain results in the thermodynamic limit , finite - size scaling analysis must be performed @xcite . therefore , a physical quantity which is more sensitive to the system size than the traditional second derivative of the ground state energy would be very useful to study the critical phenomena numerically . from eq . ( [ eq : h_xxz ] ) , it can be easily seen that the hamiltonian of the xxz model commutes with the @xmath48-component of total spin operator @xmath71 . thus , each eigenstate of the hamiltonian is also an egienstate of @xmath72 . the hilbert space can then be decomposed into numerous subspaces @xmath73 , where @xmath74 is the eigenvalue of @xmath75 . for a finite sample , the ground state of the xxz model is non - degenerated in any of the admissible subspace @xmath73 @xcite . therefore , the perturbation expansion can also be applied to this model as long as the system is finite . in the thermodynamic limit , the quantum phase transition takes place at the isotropic point @xmath76 . this phenomenon can be understood by the picture of the first excited energy levels crossing at the transition point @xcite . for @xmath77 , the last term in the hamiltonian dominates and the ground state is an antiferromagnetic phase along the @xmath48-direction . for @xmath78 , the first two terms in the hamiltonian dominate and the ground state is also an antiferromagnetic phase , but in the @xmath79 plane . it is well known that long - range orders are present in both of the two phases . however , whether there s a long - range order at the critical point is still an open question . with the help of fidelity susceptibility , we may also find some hints towards this question . the numerical result of the averaged fidelity susceptibility for various system sizes of the 2d xxz model on a square lattice as a function of @xmath80 is shown in fig . [ fig : xxz_fs ] . the averaged fidelity susceptibility is an intensive quantity , meaning that @xmath81 , on both sides of the critical point . moreover , like the previous case of the ising model , the averaged fidelity susceptibility of the xxz model also shows a peak at @xmath82 . the inset of fig . [ fig : xxz_fs ] shows the peak position of the fidelity susceptibility @xmath83 as a function of @xmath31 . the linear fitting gives @xmath84 in the thermodynamic limit , we obtain @xmath85comparing to the theoretical critical point @xmath86 , our result here is consistent up to two digits . besides , from the slope of the straight line in the inset of fig . [ fig : xxz_scaling ] , it is found that the peak of the averaged fidelity susceptibility scales with the system size like @xmath87 . therefore , one may expect the fidelity susceptibility to show a singularity at the critical point in the thermodynamic limit . hence , the validity of the fidelity susceptibility as a seeker for the quantum phase transition is also verified in the 2d xxz model . nevertheless , following the idea of the implication of existence of long - range correlation from the divergence of the fidelity susceptibility @xcite , we argue that long - range correlation is in fact present at the transition point of the 2d xxz model . this is also in agreement with the previous conclusion drawn from the study of the 2d xxz model using entanglement guxxzentan . in comparison , the second derivative of the averaged ground state energy for various system sizes exhibits a minimum at the transition point , as shown in fig . [ fig : xxz_gse ] . from the inset of fig . [ fig : xxz_gse ] , it is also found that the minimum value of the second derivative of the averaged ground state energy scales approximately with @xmath88 , meaning that it shows a slower divergence at the critical point compared to the fidelity susceptibility . in other words , the fidelity susceptibility is again a more sensitive candidate in seeking for the quantum phase transition in the 2d xxz model . [ fig : xxz_scaling ] shows the finite - size scaling analysis in the case of power - law divergence of the 2d xxz model . the rescaled fidelity susceptibility almost collapsed to a single curve for a large enough system size , say @xmath89 . the exponent of the correlation lenght is obtained as @xmath90 . from the slope of the inset in fig . [ fig : xxz_scaling ] and using eq . ( [ eq : exp ] ) , the critical exponent of the fidelity susceptibility is calculated to be @xmath91 to conclude , through the numerical study of the fidelity susceptibility in the 2d transverse field ising model and the 2d xxz model , we found that the fidelity susceptibility as a function of the driving parameter diverges in both models at the critical point . by comparing the scaling behavior of the extremum of the fidelity susceptibility to that of the second derivative of the ground state energy , we also showed that fidelity susceptibility is a more sensitive indicator in detecting for a second order quantum phase transition . by performing finite - size scaling analysis , the critical exponent of the fidelity susceptibility in both models are also obtained . finally , due to the divergence of fidelity susceptibility in the 2d xxz model , we argued that the system shows a long - range correlation at the critical point of the model . s. j. gu , chin . * 26 * , 026401 ( 2009 ) . for examples : h. m. kwok , w. q. ning , s. j. gu , and h. q. lin , phys . e * 78 * , 032103 ( 2008 ) ; m. f. yang , phys . rev . b * 76 * , 180403(r ) ( 2007 ) ; j. o. fjrestad , j. stat . mech . p07011 ( 2008 ) ; l. gong and p. tong , phys . b * 78 * , 115114 ( 2008 ) ; y. c. li and s. s. li , phys . rev . b * 78 * , 184412 ( 2008 ) ; j. ren and s. zhu , eur . phys . j. d * 50 * , 103 ( 2008 ) ; k. w. sun , y. y. zhang , and q. h. chen , phys . b * 79 * , 104429 ( 2009 ) .
we study the fidelity susceptibility in the two - dimensional(2d ) transverse field ising model and the 2d xxz model numerically . it is found that in both models , the fidelity susceptibility as a function of the driving parameter diverges at the critical points . the validity of the fidelity susceptibility to signal for the quantum phase transition is thus verified in these two models . we also compare the scaling behavior of the extremum of the fidelity susceptibility to that of the second derivative of the ground state energy . from those results , the theoretical argument that fidelity susceptibility is a more sensitive seeker for a second order quantum phase transition is also testified in the two models .
there are lots of gravitational evidences suggesting the dominance of dark matter ( dm ) over baryonic visible matter in our universe , though little is known about its non - gravitational interaction . if dm can annihilate or decay into the standard model ( sm ) particles , gamma rays will be generated either directly or by the final state radiation , inverse compton scattering and bremsstrahlung of the final charged particles . as these gamma rays depend on the dm density as @xmath0 for annihilations and @xmath1 for decays , they might be detectable at locations with very high density of dark matter , such as the galactic center ( gc ) . recently hooper and goodenough @xcite investigated the gamma rays from the gc using the first two years of data from the fermi gamma ray space telescope . they modeled the backgrounds and found that the morphology and spectrum of the gamma rays between @xmath2 and @xmath3 from the gc can be well understood . however excess emission of gamma rays seems to be present within @xmath2 degree of the gc . this may well indicate that we have to improve our understanding on the astrophysical background . nevertheless , this additional component could also be interpreted by the dm annihilation into tau leptons , with a mass around @xmath4 gev and the annihilation cross section @xmath5s . there are also anomalous microwave emission from the inner galaxy in the wmap data @xcite which is known as the `` wmap haze '' . it was further observed by hooper and linden @xcite that , if the dm annihilates equally into @xmath6 , @xmath7 and @xmath8 , the wmap haze can be simultaneously accounted for by the synchrotron radiation emitted from the annihilation final state of electrons and positrons . notice that wmap haze could also be explained by astrophysical processes @xcite . coincidentally , the annual modulation effect measured by the dama / libra collaboration @xcite together with the recent results about excesses of events ( over the backgrounds ) reported by cdms - ii @xcite and cogent @xcite , if interpreted as elastically scattering dm , favor the dm mass about 5 - 10 gev @xcite . the null results from xenon100 @xcite and cdms - ii @xcite are in tension with the previous data . nevertheless , concerning the uncertainties of the dm velocity distribution @xcite and of the relative scintillation efficiency of the liquid xenon at low recoil energy @xcite , it is still possible to interpret the cogent and dama / libra observations as light dm with mass around @xmath9 gev @xcite , and at the same time , consistent with the null results of xenon100 and cdms - ii . such a light dm is beyond the conventional paradigm of weakly - interacting massive particle ( wimp ) with the mass around @xmath10 gev to several tev , though it is certainly possible to realize this scenario in susy models , e.g. in an effective minimal supersymmetric extension of the standard model @xcite . interestingly , in the wimpless models @xcite the dm particle naturally has the correct thermal relic density with a wide range of mass varying from tev down to subgev @xcite . the key observation in @xcite is that , the thermal relic density of a stable particle is proportional to @xmath11 , with @xmath12 and @xmath13 the typical mass and coupling entering the annihilation cross section . the correct thermal relic density can be easily realized for the dm particle x in the hidden sector if @xmath14 with the weak scale @xmath15 gev and @xmath16 . such relation can be satisfied for example in the gauge - mediated supersymmetry breaking models if there is only one supersymmetry(susy ) breaking sector , so that the susy - breaking soft scales in both the visible sector and the hidden sector are generated by the gauge interactions with the same susy - breaking sector . in the following , we will then study the possible implications of the above experimental results for the wimpless models . notice that we shall not attempt to explore the full parameter space of the wimpless models . instead , as illustration only benchmark models are examined in this paper . for example , following ref . @xcite , we only consider the wimpless dm annihilating democratically into @xmath6 , @xmath7 and @xmath8 to interpret the wmap haze . but it is definitely possible to explain the wmap haze with other choices of parameters provided the dm can significantly annihilate into @xmath6 . nevertheless , our main conclusions should remain qualitatively unchanged for general wimpless models . phenomenologically , the wimpless dm has been used to explain the dama signal @xcite . the indirect detection of fermion wimpless dm at the neutrino telescopes icecube and deepcore has been discussed in @xcite . the low energy constraints on wimpless dm can be found in @xcite . it is also worth noting that the particle physics implications for the gamma ray excess from the gc together with cogent and dama / libra has been studied in @xcite , though in a model independent way and without the inclusion of the wmap haze . this paper is organized as follows . in the next section , we adopt scalar and vector wimpless models to discuss the excess gamma rays from the gc , the wmap haze and the cogent and dama results . we also investigate the low energy constraints from the anomalous magnetic moment of leptons and from some lepton flavor violating decays . we then extend the discussions to fermion wimpless models in the third section . finally we conclude with a summary in section iv . in wimpless models , to couple the dm particle x to the visible sector , yukawa like interactions including a connector sector can be introduced . to be concrete , let s first consider the case of scalar dm which was exploited in @xcite to investigate the dama annual modulation signals . the interactions can be written as @xmath17 where @xmath18 and @xmath19 denote the left - handed and right - handed standard model ( sm ) fermions , @xmath20 are the connector fields which have both hidden and sm charges . notice that @xmath20 behave just like exotic fourth generation quarks and leptons ( or squarks and sleptons ) . the direct searches at tevatron constrain the masses of the exotic fourth generation quarks to be @xmath21 gev @xcite , while the searches at lep constrain the sleptons to be heavier than @xmath22 gev @xcite . therefore the connector fields y must be much heavier than the dm particle x which is about @xmath4 gev , and in the following calculations we will take the limit of large @xmath23 . @xmath24 can also be a vector field , correspondingly the interactions is @xmath25 with the above lagrangians , the dm particles can annihilate into the sm fermions as shown in fig . then @xmath26 it is shown in @xcite that the excess gamma rays from the gc can be explained by the dm annihilation primarily into tau lepton pairs , with the annihilation cross section @xmath27 in the range of @xmath28s to @xmath29s . taking @xmath30 gev , this requires @xmath31 it was also pointed out in @xcite that the dm x may annihilate into @xmath32 or @xmath33 final states up to @xmath34% of the time without spoiling the dm interpretation of the excess gamma rays from the gc . this means @xmath35 the feynman diagrams for ( a ) , ( b ) : the scalar dm annihilation process @xmath36 and ( c),(d ) : the scalar dm - bottom quark scattering . for the case of vector dm , the diagrams are the same except for the external scalar bosons to be replaced by the vector bosons ] the interaction terms in eqs.([eq : yukawa],[eq : yukawa - vector ] ) can also induce spin - independent x - nucleon scattering as shown in fig . 1c , d . to escape flavor changing neutral current ( fcnc ) constraints , we will assume following @xcite that only @xmath37 and @xmath38 are significant for x - y - quark couplings . the corresponding direct detection cross section @xcite is the same for both scalar x and vector @xmath39 , as @xmath40 where @xmath41 @xcite with the uncertainties coming mainly from @xmath42 . the cogent and dama / libra results , if interpreted as dm elastic scattering , require @xmath43 @xcite . this implies the following constraint @xmath44 which completely contradicts eq.([eq : gcconstraint ] ) . therefore it seems to be difficult , if not impossible , for the scalar or vector wimpless dm to account for the gamma ray excess from the gc together with the direct detection results of cogent and dama . furthermore , even if we forget about the direct detection signals , there are stringent constraints , if considering the wmap haze , from anomalous magnetic moment of leptons and some lepton flavor violating decays . as analyzed in @xcite , the wmap haze could be accounted for by the dm annihilation democratically into @xmath6 , @xmath7 and @xmath8 , which means @xmath45 but as shown in fig 2(a ) , the @xmath46 couplings also contribute to the anomalous magnetic moment of leptons . for the scalar case , it is found to be should be @xmath47 . ] @xmath48 while for the vector case , it is @xmath49 there is well known @xmath50 deviation for the muon anomalous magnetic moment between the sm expectation and experimental measurement as @xcite @xmath51 for the scalar case , this limits @xmath52 which strongly violates eq.([eq : hazeconstraint ] ) . for the vector case , assuming @xmath53 , the second term of eq.([vector - amu ] ) is about one order smaller than the first term , and we obtain @xmath54 which again severely violates eq.([eq : hazeconstraint ] ) . for the electron , the experimental data gives @xmath55 @xcite which agrees well with the sm calculation @xmath56@xcite . we will assume @xmath57 which is slightly larger than the theoretical error , to get a restriction on the corresponding @xmath58 couplings @xmath59 which also strongly violates eq.([eq : hazeconstraint ] ) . if one connector field @xmath60 can couple to different generations of leptons , namely @xmath61 , the @xmath46 interactions could also induce the lepton flavor violating ( lfv ) @xmath62 , @xmath63 and @xmath64 decays , as shown in fig . as the anomalous magnetic moment of leptons already exclude the possibility for scalar or vector wimpless dm to explain the wmap haze together with the excess gamma rays from the gc , we will not investigate further these lfv decays . but in the next section these lfv decays will turn out to be important in the case of fermion wimpless dm . the @xmath46 interactions contribute to ( a ) the anomalous magnetic moment of leptons and ( b ) the lepton flavor violating @xmath62 , @xmath63 and @xmath64 decays with scalar x. for the case of vector @xmath39 , the diagrams are the same except for the internal scalar line to be replaced by the vector line . ] therefore in scalar or vector wimpless models , neither the wmap haze nor the cogent and dama observations could be explained simultaneously with the excess gamma rays from the gc . if the dm particle @xmath24 is a fermion , the connector fields @xmath65 should be either scalar or vector fields . for the scalar @xmath65 case , the interactions can be written as @xmath66 where @xmath67 and @xmath68 are different scalar fields carrying different sm charges . for the vector @xmath65 case , the corresponding interactions are @xmath69 here @xmath67 and @xmath68 are different vector fields . with the above lagrangians , the dm particles can annihilate into the sm fermions as shown in fig . 3(a ) with the cross section @xmath70 assuming for simplicity @xmath71 and taking @xmath30 gev , the excess gamma rays from the gc restricts @xmath72 which corresponds to the dm annihilation into @xmath8 with the cross section @xmath73s . equivalently this means @xmath74 gev for scalar y and @xmath75 gev for vector @xmath76 . as @xmath77 behaves just like slepton , this does not contradict , if @xmath78 is not much smaller than one , with the direct searches at lep which constrain the sleptons to be heavier than @xmath22 gev @xcite . considering direct detection experiments , the dm - nucleon scattering , as shown in fig . 3(b ) , should be evaluated . in the large @xmath23 limit and after a fierz transformation , the dm - quark scattering can be written as effective operators @xmath79 notice that the spin - independent and velocity - independent contribution to dm - nucleon scattering comes only from the vector - vector interaction . assuming @xmath80 , the dm - nucleon scattering cross section is @xmath81 to obtain @xmath43 suggested by the cogent and dama results , one finds @xmath82 equivalently this means @xmath83 gev for scalar y and @xmath84 gev for vector @xmath76 . as @xmath85 behaves just like exotic fourth generation quark , the scalar y case seems to be in disagreement , if @xmath86 is smaller than one , with the tevatron direct searches which limit the exotic fourth generation quark to be heavier than @xmath87 gev @xcite . while for the vector @xmath76 case , it is only marginally consistent with the tevatron limit . the feynman diagrams for ( a ) the dm annihilation process @xmath36 , ( b ) the dm - quark scattering , ( c ) the anomalous magnetic moment of leptons and ( d ) the lepton flavor violating @xmath88 decay with the scalar connector fields . for the vector @xmath76 case , the diagrams are the same except for all internal scalar lines to be replaced by the vector lines . ] to explain the wmap haze in the same framework , the dm candidate x should annihilate democratically into @xmath6 , @xmath7 and @xmath8 , which means @xmath89 the anomalous magnetic moment of leptons , as shown in fig . 3(c ) , can be evaluated as @xmath90 for the case of scalar connector fields . similarly the contribution of vector connector fields to lepton s anomalous magnetic moment is @xmath91 notice that in both cases , the above contributions to lepton s anomalous magnetic moment are negative , which is opposite in sign to the @xmath50 deviation of the muon s anomalous magnetic moment @xmath92 . therefore it is impossible for the fermion wimpless models to account for the @xmath50 deviation of @xmath93 . instead the wimpless contribution should be small enough , say @xmath94 , so that the situation does not become significantly worse . then one can obtain the following limits for the scalar y case @xmath95 which are consistent with eq.([eq : fermion - haze ] ) . however for the vector connector fields , the limits are @xmath96 where unfortunately the muon s @xmath97 leads to a constraint too stringent to be satisfied by eq.([eq : fermion - haze ] ) . furthermore , if one connector field @xmath60 can couple to any charged leptons , namely @xmath98 , it could also induce the lepton flavor violating @xmath62 , @xmath63 and @xmath64 decays , as shown in fig . notice that the photon can also be attached to external lepton lines , but those diagrams are gauge - dependent and should be canceled by the gauge - dependent part of fig . 3(d ) . then taking the final state lepton to be massless and assuming for simplicity @xmath99 , the decay width is found to be @xmath100 experimentally @xmath101 , @xmath102 and @xmath103 @xcite lead to the constraints @xmath104 for the case of scalar ( vector ) connector fields . unfortunately all the above limits are in clear contradiction with the requirement of the wmap haze of eq.([eq : fermion - haze ] ) . in this paper we discuss in wimpless models the possibility to interpret the excess gamma rays from the galactic center , the wmap haze and the cogent and dama results . at the same time some low energy constraints must be satisfied , such as the anomalous magnetic moment of leptons and lepton flavor violating @xmath105 decays . notice that a connector sector y is introduced to couple the dm candidate x to the sm particles via yukawa - like interactions . as shown in @xcite , the excess gamma rays from the gc implies that the dm particles should annihilate dominantly into lepton pairs with @xmath32 or @xmath33 final states less than @xmath34% of the time . for scalar or vector wimpless dm , this limits the coupling strength of @xmath106 to be too small to account for the cogent and dama observations . to interpret the wmap haze , the @xmath107 coupling is required to be roughly universal for @xmath108 , @xmath109 and @xmath110 leptons , which will lead to too large contribution to anomalous magnetic moment of electron and muon . as to fermion wimpless dm with scalar or vector connector sector , @xmath111 a few @xmath112 could accommodate the excess gamma rays from the gc and the cogent and dama results . this corresponds to a scalar connector particle @xmath85 less than @xmath113 gev or a vector @xmath85 less than about @xmath114 gev . as @xmath85 couples to quarks , it behaves like an exotic fourth generation quark , which has been restricted by the tevatron to be heavier than @xmath87 gev @xcite . therefore the scalar y case seems to be disfavored , while the vector @xmath76 case is only marginally consistent with the tevatron limit . instead , to interpret the wmap haze in the same framework , the constraints of anomalous magnetic moment of leptons can only be satisfied for the scalar connector fields . furthermore , if there is only one connector field for charged leptons , the lepton flavor violating @xmath88 decays could happen in both cases with too large branching ratios severely violating the experimental bounds . therefore it is difficult , if not impossible , for the scalar or vector wimpless dm to interpret the excess gamma rays from the gc together with either the wmap haze or the direct detection experiments cogent and dama . while for fermion wimpless dm , it may be possible to accommodate the excess gamma rays from the gc and the cogent and dama results with vector connector fields , though it is just marginally consistent with the tevatron limit . on the contrary , only scalar connector fields could explain the wmap haze under the constraints of anomalous magnetic moment of the leptons . in addition , if one connector field can couple to any charged leptons , the lepton flavor violating @xmath88 decays could happen with the branching ratios surpassing the current experimental bounds by ( at most ) even several orders of magnitude . quantitatively , the above conclusions are only valid in the wimpless models . but going beyond this specific model , one may at least learn that the excess gamma rays from the galactic center , if interpreted as dm annihilation , could impose strong constraint on the dm couplings to quarks . then one should check , under this constraint , the interpretation of the cogent and dama results in terms of elastic dm - nucleon scattering . it is also interesting to investigate the interplay between the direct and indirect dm searches and low energy precision observables , such as anomalous magnetic moment of leptons and the lepton flavor violating decays . this work is supported in part by the national science foundation of china ( no . 11075139 and no.10705024 ) and national basic research program of china ( no . 2010cb833000 ) . g.z is also supported in part by the fundamental research funds for the central universities . 99 d. hooper and l. goodenough , arxiv:1010.2752 [ hep - ph ] . d. p. finkbeiner , astrophys . j. * 614 * ( 2004 ) 186 [ arxiv : astro - ph/0311547 ] . g. dobler and d. p. finkbeiner , astrophys . j. * 680 * ( 2008 ) 1222 [ arxiv:0712.1038 [ astro - ph ] ] . d. hooper and t. linden , arxiv:1011.4520 [ astro-ph.he ] . r. m. crocker , d. i. jones , f. aharonian , c. j. law , f. melia , t. oka and j. ott , arxiv:1011.0206 [ astro-ph.ga ] . c. e. aalseth _ [ cogent collaboration ] , arxiv:1002.4703 [ astro-ph.co ] . r. bernabei _ et al . _ , eur . j. c * 67 * ( 2010 ) 39 [ arxiv:1002.1028 [ astro-ph.ga ] ] . j. kopp , t. schwetz and j. zupan , jcap * 1002 * ( 2010 ) 014 [ arxiv:0912.4264 [ hep - ph ] ] . a. bottino , f. donato , n. fornengo and s. scopel , phys . d * 81 * ( 2010 ) 107302 [ arxiv:0912.4025 [ hep - ph ] ] . a. l. fitzpatrick , d. hooper and k. m. zurek , phys . rev . d * 81 * ( 2010 ) 115005 [ arxiv:1003.0014 [ hep - ph ] ] . s. chang , j. liu , a. pierce , n. weiner and i. yavin , jcap * 1008 * ( 2010 ) 018 [ arxiv:1004.0697 [ hep - ph ] ] . r. foot , phys . b * 692 * ( 2010 ) 65 [ arxiv:1004.1424 [ hep - ph ] ] . e. aprile _ et al . _ [ xenon100 collaboration ] , phys . rev . * 105 * ( 2010 ) 131302 [ arxiv:1005.0380 [ astro-ph.co ] ] . z. ahmed _ et al . _ [ cdms - ii collaboration ] , arxiv:1011.2482 [ astro-ph.co ] . z. ahmed _ et al . _ [ the cdms - ii collaboration ] , science * 327 * ( 2010 ) 1619 [ arxiv:0912.3592 [ astro-ph.co ] ] . m. lisanti , l. e. strigari , j. g. wacker and r. h. wechsler , arxiv:1010.4300 [ astro-ph.co ] . c. savage , g. gelmini , p. gondolo and k. freese , arxiv:1006.0972 [ astro-ph.co ] . j. i. collar , arxiv:1006.2031 [ astro-ph.co ] . j. i. collar , arxiv:1010.5187 [ astro-ph.im ] . d. hooper , j. i. collar , j. hall , d. mckinsey and c. m. kelso , phys . d * 82 * ( 2010 ) 123509 [ arxiv:1007.1005 [ hep - ph ] ] . a. bottino , f. donato , n. fornengo and s. scopel , phys . d * 68 * ( 2003 ) 043506 [ arxiv : hep - ph/0304080 ] . a. bottino , f. donato , n. fornengo and s. scopel , phys . d * 69 * ( 2004 ) 037302 [ arxiv : hep - ph/0307303 ] . n. fornengo , s. scopel and a. bottino , phys . d * 83 * ( 2011 ) 015001 [ arxiv:1011.4743 [ hep - ph ] ] . s. scopel , s. choi , n. fornengo and a. bottino , arxiv:1102.4033 [ hep - ph ] . j. l. feng and j. kumar , phys . * 101 * ( 2008 ) 231301 [ arxiv:0803.4196 [ hep - ph ] ] . j. l. feng , h. tu and h. b. yu , jcap * 0810 * ( 2008 ) 043 [ arxiv:0808.2318 [ hep - ph ] ] . m. y. khlopov , a. g. mayorov and e. y. soldatov , int . j. mod . d * 19 * ( 2010 ) 1385 [ arxiv:1003.1144 [ astro-ph.co ] ] . j. l. feng , j. kumar and l. e. strigari , phys . b * 670 * ( 2008 ) 37 [ arxiv:0806.3746 [ hep - ph ] ] . v. barger , j. kumar , d. marfatia and e. m. sessolo , phys . d * 81 * ( 2010 ) 115010 [ arxiv:1004.4573 [ hep - ph ] ] . d. mckeen , phys . d * 79 * ( 2009 ) 114001 [ arxiv:0903.4982 [ hep - ph ] ] . d. mckeen , arxiv:0912.1076 [ hep - ph ] . m. r. buckley , d. hooper and t. m. p. tait , arxiv:1011.1499 [ hep - ph ] . j. alwall , j. l. feng , j. kumar and s. su , phys . d * 81 * ( 2010 ) 114027 [ arxiv:1002.3366 [ hep - ph ] ] . lepsusywg , aleph , delphi , l3 and opal experiments , note lepsusywg/04 - 01 ( http://lepsusy.web.cern.ch/lepsusy/ ) . a. bottino , f. donato , n. fornengo and s. scopel , astropart . * 18 * ( 2002 ) 205 [ arxiv : hep - ph/0111229 ] . j. r. ellis , k. a. olive , y. santoso and v. c. spanos , phys . d * 71 * ( 2005 ) 095007 [ arxiv : hep - ph/0502001 ] . k. nakamura _ et al . _ [ particle data group ] , j. phys . g * 37 * ( 2010 ) 075021 . t. aoyama , m. hayakawa , t. kinoshita and m. nio , phys . d * 77 * ( 2008 ) 053012 [ arxiv:0712.2607 [ hep - ph ] ] .
in this paper we discuss the excess gamma rays from the galactic center , the wmap haze and the cogent and dama results in wimpless models . at the same time we also investigate the low energy constraints from the anomalous magnetic moment of leptons and from some lepton flavor violating decays . it is found that , for scalar or vector wimpless dark matter , neither the wmap haze nor the cogent and dama observations could be explained simultaneously with the excess gamma rays from the galactic center . as to fermion wimpless dark matter , it is only marginally possible to accommodate the cogent and dama results with the excess gamma rays from the galactic center with vector connector fields . on the other hand , only scalar connector fields could interpret the wmap haze concerning the constraints of anomalous magnetic moment of leptons . furthermore , if there is only one connector field for all the charged leptons , some lepton flavor violating decays could happen with too large branching ratios severely violating the experimental bounds .
understanding the properties and behaviour of the interstellar medium ( ism ) of galaxies in the early universe is a cornerstone of modern galaxy evolution studies . galaxies at early epochs show significantly elevated gas fractions relative to their local analogues @xcite , and it is these massive gas reservoirs that drive the enhanced star formation rates characteristic of the high-@xmath11 universe ( i.e. , @xcite ; @xcite ; ) . a variety of techniques have been used to observe the gas reservoirs in distant galaxies . traditionally , the detection of intergalactic molecular gas ( particularly at high-@xmath11 ) has relied on observations of various molecular emission lines of carbon monoxide ( @xmath6co ; @xcite ; @xcite ; @xcite ; @xcite ) . co exists in the centres of molecular clouds , and with the aid of a co - to - h@xmath7 conversion factor observations of the co line luminosity can be converted into a mass of molecular gas . this conversion is non - trivial however , with the abundance of co relative to h@xmath7 varying as a function of ism metallicity ( increased metal abundance results in larger quantities of dust to protect co from photodissociation ) and even galactic behaviour ( in merging systems the star - forming ism no longer consists of discrete molecular clouds , making co a more efficient tracer of molecular gas ) . moreover , recent results have suggested that co may also be destroyed by cosmic rays ( produced indirectly by star formation , via supernovae , as well as agn activity ) . the ism of intensely star - forming galaxies , which produce a large cosmic ray flux , may be a hostile environment for co molecules making it a less effective tracer of molecular gas than previously thought ( @xcite ; @xcite ) . another method for tracing molecular gas is via the long - wavelength dust emission . observations of the dust continuum can be converted into a total dust mass , which , with the aid of an assumed gas - to - dust ratio , can be used to calculate a gas mass ( see @xcite ; @xcite ; @xcite ) . this technique has some advantages over the use of co lines : dust continuum observations are generally less time intensive , allowing for larger samples to be assembled . the dust mass method is not without its disadvantages , however . the dust temperature must be constrained ( or assumed ) in order for a dust mass to be measured . in addition , the gas - to - dust ratio remains a relatively poorly studied quantity , which may vary by up to a factor of @xmath12 in bright star - forming galaxies @xcite , as well as potentially varying at high redshift ( @xcite ; @xcite ) . furthermore , observations of the dust continuum provide no kinematic information , which is available when observing co emission lines . in recent years , the emission line of atomic carbon ( ( @xmath13 ) ; ( @xmath0 ) hereafter ) has been found to be an excellent alternate tracer of the cold molecular ism , being closely associated with low-@xmath5 co emission across a wide range of environments and redshifts ( @xcite ; @xcite ; @xcite ; @xcite ) . this conclusion is supported by both detailed studies of nearby galactic molecular clouds ( in which co and are found to co - exist throughout the bulk of the cold molecular component ; @xcite ) , as well as hydrodynamic simulations @xcite . using as a molecular gas tracer offers a number of advantages , compared to observations of both co and the dust continuum . due to the simplicity of the quantum fine structure level of , many physical parameters ( including excitation temperature and total carbon mass ) can be calculated with minimal uncertainty . and while co becomes an increasingly poor tracer of molecular gas as metallicity decreases ( due to a lack of dust grains being available to shield co molecules from photodissociation ) , is affected far less severely significantly reducing the uncertainty on the molecular gas mass introduced by unknown metal abundances . furthermore , @xcite found that a high cosmic - ray ionisation rate will destroy co molecules , dissociating them into , while leaving the underlying h@xmath7 unaffected . additionally , observations of the line provide the same valuable kinematic information as co , offering a distinct advantage over dust - based methods . as a result can be a powerful and effective tracer of the molecular ism in distant galaxies . in this work , we present observations of the ( @xmath0 ) emission line in a sample of strongly lensed dusty star forming galaxies ( dsfgs ) which were identified via the south pole telescope ( spt ; @xcite ) wide - field survey ( @xcite ) . dsfgs are extremely luminous star - forming galaxies ( typical sfrs @xmath14 m@xmath4/yr ) , which are thought to be the high-@xmath11 progenitors of the most massive galaxies in the @xmath15 universe . the spt has proven to be an efficient machine for finding the brightest ( strongly - lensed ) dsfgs in the universe ( @xcite ; @xcite ) . the dsfgs in this work were all taken from the 26-galaxy sample targeted for spectroscopic redshift identification by the atacama large millimeter array ( alma ) in cycle 0 @xcite , which confirmed redshifts via the identification of a number of atomic and molecular emission lines . thirteen dsfgs were found to lie at redshifts @xmath16 , shifting the ( @xmath0 ) into alma band 3 . in this work we present an analysis of the properties of these 13 sources . this paper falls broadly into two halves . in the first half , we present an analysis of the properties of our sample . we describe the sample and the various ancillary data in 2 , and in 3 we present our analysis of the ( @xmath0 ) data ( including calculations of the mass and cooling contribution of atomic carbon , analysis of the kinematic properties of the sample , and discussions of the use of ( @xmath0 ) as a tracer of molecular gas and star formation mode ) . moving to the second half of the paper , in 4 we combine the ( @xmath0 ) line with a variety of atomic and molecular emission lines in order to constrain the physical conditions in the ism of our galaxies using a modern photodissociation region ( pdr ) modelling code , 3d - pdr . we present a discussion of our results in 5 , and we present our conclusions in 6 . throughout this work we adopt a standard @xmath17-cdm cosmology with parameters taken from @xcite ; @xmath18 , @xmath19 , and @xmath20 . the 13 sources presented in this work were originally targeted as part of our alma blind redshift search programme @xcite , in which 26 spt dsfgs were observed across the entirety of alma band 3 ( @xmath21 ghz ) as part of the cycle 0 ` early compact array ' setup . each observation consisted of 5 distinct 7.5 ghz tunings , spaced to cover the band . each source was observed for 120s in each tuning configuration . further details of the observing programme and data reduction can be found in @xcite . 13 of the 26 dsfgs lie in the redshift range @xmath16 , causing the ( @xmath0 ) emission line ( @xmath22 492.161 ghz ) to be redshifted into band 3 . of these 13 sources , just one ( spt0345 - 47 ) was not detected in ( @xmath0 ) . a further two sources ( spt0300 - 46 and spt2103 - 60 ) are tentatively detected at the @xmath23 level ( i.e. , 3 times the rms channel noise , which we measure to be 2.1 mjy and 2.6 mjy per 50 km / s channel respectively ) . the remaining 10 sources all show clearly detected ( @xmath0 ) emission at @xmath24 significance . for spt0345 - 47 , we calculate a @xmath25 upper limit on the intensity of the ( @xmath0 ) emission line : @xmath26 where @xmath27 is the rms channel noise in the spectrum of spt0345 - 47 ( which we measure to be 2.4 mjy per 50 km / s channel ) , @xmath28 is the mean linewidth of the detected sample ( = 410 km / s ) , and @xmath29 is the bin size in kms@xmath30 ( = 50 km / s ) . we calculate an upper limit on the line intensity for spt0345 - 47 of @xmath31(1 - 0 ) } < 1.03 $ ] jy km / s . figure [ fig : prop1 ] shows spectral cutouts at the position of the ( @xmath0 ) line for the 13 spt dsfgs analysed in this work . for reference , we have overlaid ( where available ) the co(@xmath32 ) emission line ( scaled arbitrarily in flux for ease of comparison ) . due to a lack of x - ray data for our sample we can not rule out a possible agn component in any of our 13 dsfgs . however it is unlikely that any of our sources contains a significant agn . dsfgs as a class are star - formation dominated ( i.e. , the bolometric output of the galaxy originates predominantly from young massive stars , rather than accretion onto a central compact object ) ; even dsfgs with some measurable agn activity tend to be primarily star - formation - driven objects @xcite . furthermore , chandra x - ray observations of the most compact and ir - luminous spt - dsfg , spt0346 - 52 , found no sign of agn activity ( ma et al . , 2016 ) . we proceed with the assumption that our the dsfgs in our sample are star - formation dominated objects . all 13 galaxies presented in this work are strongly gravitationally lensed , and have detailed lens models based on alma @xmath33 m observations , which allow their lensing magnifications to be calculated @xcite . we discuss the use of these lens models to remove the effects of gravitational lensing in 3.1 below . the spt - dsfg sample has been the target of several followup programmes designed to survey a number of ism diagnostics . in particular , in addition to the ( @xmath0 ) emission lines presented in this work , 9/13 sources have observations covering the co(@xmath32 ) emission line , 10/13 have observations covering the co(@xmath34 ) emission line ( the remaining 3 have observed co(@xmath35 ) emission , which can be converted to ( @xmath34 ) with the aid of an assumed co spectral line energy distribution ( @xcite ; @xcite ) ) . in addition , 10/13 have the emission line observed . the mid-@xmath5 lines of co come from the alma spectra used in this work . the observations were obtained from the apex first light apex submillimetre heterodyne receiver ( flash ) , and the herschel ` spire ' fts spectrometer . observations were made at 345 ghz ( for sources with redshifts @xmath36 ) and 460ghz ( @xmath37 ) , with system temperatures of 230 k and 170 k respectively . further details of the observations and data reduction can be found in @xcite . the low-@xmath5 co observations were taken by the australia telescope compact array ( atca ) , as part of a targeted program aimed at obtaining co(@xmath0 ) or co(@xmath32 ) for spt dsfgs with secure redshifts . atca was used in its h214 hybrid array configuration , with the compact array broadband backend in wide bandwidth mode . the mean rms noise for the co(@xmath32 ) observations used in this work was 0.5 mjy per 50 km / s channel . further details of the observations and data reduction can be found in @xcite . this suite of molecular emission lines enables a more detailed treatment of the conditions of the ism , as multiple line ratios can be used to independently constrain various parameters of interest . throughout this work , we calculate emission line luminosities using the formulae below . luminosities in solar units ( l@xmath4 ) , which represent the true energy output carried by the emission line ( used , for example , for calculating cooling contributions ) are calculated following @xcite : where @xmath39 is the velocity - integrated line flux in jy km s@xmath30 , @xmath40 is the rest frequency in ghz , and @xmath41 is the luminosity distance in mpc . alternatively , line luminosities in units of k km s@xmath30 pc@xmath42 are calculated using table [ tab : lum ] lists line luminosities ( in solar units ) for the dsfgs used in this work . in addition , we also list in table [ tab : lum ] the bolometric far - ir luminosity , calculated using greybody sed fits to far - ir / millimetre photometry ( greve et al . 2012 ) . note that all values in table [ tab : lum ] are observed quantities , which have not been corrected for the effects of gravitational lensing . .line luminosities , far - ir luminosity , /fir ratios , and lensing magnifications ( @xmath44 ) for the dsfgs studied in this work . all luminosities are given in solar units , and have not been corrected for gravitational lensing . where @xmath6co(@xmath34 ) transition line luminosities are not directly measured , they are inferred from the @xmath6co(@xmath35 ) line luminosity using the conversions derived by @xcite , and are shown in parentheses . the far - infrared luminosities are derived by integrating under a modified blackbody curve from 8 - 1000@xmath44 m . lensing magnifications are derived from visibility - based lens models fit to alma 870@xmath44 m observations ( spilker et al . data are taken from gullberg et al . co@xmath45 data are taken from aravena et al . ( 2016 ) other co lines are taken from the program described in weiss et al . ( 2013 ) . [ cols="<,^,^,^,^,^,^,^,^,^,^,^,^,^,^",options="header " , ]
we present alma ( @xmath0 ) ( rest frequency 492 ghz ) observations for a sample of 13 strongly - lensed dusty star - forming galaxies originally discovered at 1.4 mm in a blank - field survey by the south pole telescope . we compare these new data with available observations from the literature , allowing a study of the ism properties of @xmath1 extreme dusty star - forming galaxies spanning a redshift range @xmath2 . using the line as a tracer of the molecular ism , we find a mean molecular gas mass for spt - dsfgs of @xmath3 m@xmath4 . this is in tension with gas masses derived via low-@xmath5 @xmath6co and dust masses ; bringing the estimates into accordance requires either ( a ) an elevated co - to - h@xmath7 conversion factor for our sample of @xmath8 and a gas - to - dust ratio @xmath9 , or ( b ) an high carbon abundance @xmath10 . using observations of a range of additional atomic and molecular lines ( including , , and multiple transitions of co ) , we use a modern photodissociation region code ( 3d - pdr ) to assess the physical conditions ( including the density , uv radiation field strength , and gas temperature ) within the ism of the dsfgs in our sample . we find that the ism within our dsfgs is characterised by dense gas permeated by strong uv fields . we note that previous efforts to characterise pdr regions in dsfgs may have significantly under - estimated the density of the ism . combined , our analysis suggests that the ism of extreme dusty starbursts at high redshift consists of dense , carbon - rich gas not directly comparable to the ism of starbursts in the local universe . galaxies : high - redshift galaxies : ism gravitational lensing : strong galaxies : evolution galaxies : formation
one of key topics in finite - time thermodynamics is the efficiency at maximum power for heat engines , which has been widely investigated by many researchers @xcite . the most elegant result on this topic is @xmath7 , the efficiency at maximum power for endoreversible heat engines @xcite , where @xmath8 represents the carnot efficiency . recently , the efficiencies at maximum power for stochastic heat engines @xcite , the feynman ratchet @xcite , and quantum dot engines@xcite were respectively achieved , and their expressions look quite different from @xmath9 . it is this difference that attracts researchers attention to the following two issues : one is the universal efficiency at maximum power for tight - coupling heat engines operating between two baths at small temperature difference @xcite ; the other is the global bounds of efficiency at maximum power for heat engines operating between two baths with arbitrary temperature difference @xcite . in particular , van den broeck @xcite found that the universal efficiency at maximum power for tight - coupling heat engines is equal to @xmath10 up to the first order term of relative temperature difference between two baths . the universality up to the second order term of relative temperature difference was first observed in ref . @xcite , then proposed as a conjecture in ref . @xcite , and finally proved by esposito _ et al . _ for tight - coupling heat engines in the presence of left - right symmetry @xcite . in addition , esposito _ et al . _ also found the efficiency at maximum power of heat engines to be bounded between @xmath10 and @xmath11 under low - dissipation conditions @xcite . interestingly , these two bounds are also shared by the linear - irreversible heat engines @xcite . the accessibility of the bounds for different kinds of heat engines were also investigated in refs . @xcite . on the other hand , the maximum power point was proved to be unavailable for refrigerators @xcite , which leads to discussions @xcite on the proper optimization criterion for refrigerators and the corresponding coefficient of performance ( cop ) . yan and chen @xcite suggested to take @xmath12 as the target function , where @xmath13 and @xmath14 are the cop of refrigerators and the heat absorbed by the working substance from the cold bath per unit time , respectively . they also optimized this target function and found the corresponding cop to be @xmath15 for endoreversible refrigerators @xcite , where @xmath16 is the carnot cop for reversible refrigerators . recently , de toms _ et al . _ introduced a unified target function for heat devices including heat engines and refrigerators . this target function is now called @xmath0-criterion which is defined as the product of the energy conversion efficiency and the heat absorbed per unit time by the working substance @xcite . @xmath0-criterion reverts to the power output for heat engines since the energy conversion efficiency for heat engines is exactly the thermal efficiency which is defined as the ratio of the power output to the heat absorbed per unit time by the working substance ( from the hot bath ) . it also degenerates into the target function proposed by yan and chen for refrigerators since the energy conversion efficiency is exactly the cop of refrigerators and the the working substance of refrigerators absorbs heat @xmath14 from the cold bath per unit time . additionally , the cop at maximum @xmath0 was also found to be @xmath17 for symmetric low - dissipation refrigerators @xcite . based on the work by de toms _ et al . _ , one of the present authors and his coworkers derived that the cop at maximum @xmath0 was bounded between @xmath18 and @xmath19 for low - dissipation refrigerators @xcite . the observed cop s for real refrigerators are also located in the region between these bounds , which is in good agreement with their theoretical estimation . these bounds were also confirmed by izumida _ et al . _ with a minimally nonlinear irreversible model for refrigerators @xcite . the above researches on refrigerators support that @xmath0-criterion is an appropriate figure of merit for refrigerators . the results obtained from maximizing @xmath0 for refrigerators have counterparts in those derived from maximizing the power output ( equivalent to the @xmath0-criterion ) for heat engines . to begin with , the cop at maximum @xmath0-criterion ( @xmath17 ) for endoreversible refrigerators @xcite corresponds to the efficiency at maximum power ( @xmath7 ) for endoreversible heat engines @xcite . next , the bounds [ @xmath18 and @xmath19 ] of cop at maximum @xmath0 for low - dissipation refrigerators @xcite correspond to the bounds [ @xmath10 and @xmath11 ] of efficiency at maximum power for low - dissipation heat engines @xcite . in particular , the cop at maximum @xmath0-criterion ( @xmath17 ) for symmetric low - dissipation refrigerators also corresponds to the efficiency at maximum power ( @xmath7 ) for symmetric low - dissipation heat engines @xcite . however , there is still a lack of the counterpart of universal efficiency at maximum power for tight - coupling heat engines in the cop at maximum @xmath0-criterion for tight - coupling refrigerators when they operate between two baths at small temperature difference . our main goal in this work is to complement this theoretical imperfection . based on the expression of cop at maximum @xmath0 for endoreversible refrigerators and symmetric low - dissipation refrigerators , we conjecture that @xmath3 might be the universal cop at maximum @xmath0 for tight - coupling refrigerators operating between two baths at small temperature difference . this conjecture is proved to be true under certain symmetric conditions within the framework of linear irreversible thermodynamics by adopting a new convention ( eq . [ eq - newconv ] ) on the thermodynamic flux related to the heat transfer between two baths . the universal behavior of efficiency at maximum power for tight - coupling heat engines is also recalculated with the consideration of this new convention . as it was done by van den broeck @xcite for heat engines , here we consider a generic setup for a tight - coupling refrigerator shown in fig . [ fig - generic ] . an external force @xmath22 is applied on the system and inputs a power @xmath23 into the system , where @xmath24 is the thermodynamically conjugate variable of @xmath22 . the dot represents the derivative with respect to time . the corresponding thermodynamic force is @xmath25 , where @xmath26 is the temperature of the system which can be well defined due to the assumption of local equilibrium . the thermodynamic flux conjugated to @xmath27 is @xmath28 . then the power input can be expressed as @xmath29 in terms of the thermodynamic flux and force . assume that the system is in contact with a cold bath at temperature @xmath30 and a hot bath at temperature @xmath31 with @xmath32 . the positive parameters @xmath33 and @xmath34 should satisfy @xmath35 due to the constraint @xmath36 . their specific values depend on the coupling strengths between the model system and the cold or hot baths . , width=302 ] in steady state , since a certain amount of heat @xmath37 is released to the hot bath and heat @xmath14 is simultaneously absorbed from the cold bath by the refrigerator per unit time , the total entropy production rate can be expressed as @xmath38 , which can be further expressed as @xmath39 or @xmath40 due to the conservation of energy ( @xmath41 ) . considering @xmath31 and @xmath30 , we can transform ( [ eq - e11 ] ) and ( [ eq - e12 ] ) into @xmath42+\dot{q}_c(1/t_h-1/t_c)\label{eq - e13}\ ] ] and @xmath43+\dot{q}_h(1/t_h-1/t_c),\label{eq - e14}\ ] ] respectively , where @xmath44 represents the term in the same order of @xmath45 . multiplying ( [ eq - e13 ] ) by @xmath33 and ( [ eq - e14 ] ) by @xmath34 , respectively , then adding them up together , we can further derive @xmath46+(s_c\dot{q}_c+s_h\dot{q}_h)(1/t_h-1/t_c)\ ] ] with the consideration of @xmath35 and @xmath47 for refrigerators . the above equation enlightens us to take @xmath48 and @xmath49 respectively as the thermodynamic flux and force related to the heat transfer between two heat baths . the truncation error with the consideration of new convention ( [ eq - newconv ] ) is in the order of @xmath45 when the entropy production rate is expressed as @xmath50 . the other conventions such as @xmath51 or @xmath52 result in the lower accuracy with the truncation error in the order of @xmath53 when the entropy production rate is expressed as @xmath54 . now we turn to the generic setup for a tight - coupling heat engine as shown in fig [ fig - generic ] . as was done by van den broeck @xcite , we assume that the system outputs power @xmath55 against an external force @xmath22 , where @xmath24 is the thermodynamically conjugate variable of @xmath22 . the corresponding thermodynamic force is @xmath25 , where @xmath26 is the temperature of the system . the thermodynamic flux conjugated to @xmath27 is @xmath28 . then the power output can be expressed as @xmath56 in terms of the thermodynamic flux and force . assume that the system is in contact with a cold bath at temperature @xmath30 and a hot bath at temperature @xmath31 with @xmath32 . similar to the analysis of refrigerators , since a certain amount of heat @xmath37 is absorbed from the hot bath and heat @xmath14 is simultaneously released to the cold bath by the heat engine per unit time , the total entropy production rate @xmath57 can be expressed as @xmath58 or @xmath59 if considering @xmath31 and @xmath30 , we can transform ( [ eq - entgen5c ] ) and ( [ eq - entgen5d ] ) into @xmath60+\dot{q}_c ( { 1}/{t_c}-{1}/{t_h}),\label{eq - entgen5a}\ ] ] and @xmath61+\dot{q}_h({1}/{t_c}-{1}/{t_h}).\label{eq - entgen5b}\ ] ] from the above two equations , we obtain @xmath62+(s_c\dot{q}_c+s_h\dot{q}_h)({1}/{t_c}-{1}/{t_h})\label{eq - entgen5}\ ] ] with the consideration of @xmath35 and @xmath63 for heat engines . the above equation enlightens us to take the thermodynamic flux @xmath64 and the thermodynamic force @xmath65 such that the entropy production rate can be expressed as @xmath66 with a truncation error in the order of @xmath45 . we note that the expression of thermodynamic force @xmath67 for heat engines opposites to the thermodynamic force ( [ eq - force2ref ] ) for refrigerators while the thermodynamic flux has the same form as ( [ eq - newconv ] ) . it is easy to see that @xmath68 when @xmath69 . that is , endoreversible refrigerators and symmetric low - dissipation refrigerators share the same asymptotic behavior at small temperature difference @xmath70 . therefore , we may conjecture that a universal cop at maximum @xmath0-criterion , @xmath71 might exist for tight - coupling refrigerators working between two baths at small temperature difference since both kinds of refrigerators can be regarded as tight - coupling refrigerators @xcite . now we will adopt the high - precision convention ( [ eq - newconv ] ) to deal with tight - coupling refrigerators . according to linear irreversible thermodynamics , we write the linear relationship between the thermodynamic fluxes and forces : @xmath72 where the onsager coefficients satisfy @xmath73 , @xmath74 , @xmath75 and @xmath76 . it is necessary to point out that the expression of @xmath77 implies that @xmath78 is a quadratic form . thus @xmath37 and @xmath14 may contain quadratic terms due to the constraint @xmath79 . it is unsuitable to merely take either @xmath37 or @xmath14 as @xmath80 because @xmath80 is a linear form . our new convention ( [ eq - newconv ] ) ensure @xmath80 to be linear through a special combination between @xmath37 and @xmath14 , which is reflected in the truncation error in the order of @xmath45 when the entropy production rate is expressed as @xmath66 . furthermore , the tight - coupling condition @xmath81 leads to @xmath82 which indicates that the mechanical flux is proportional to the heat flux under the tight - coupling condition @xcite . with the consideration of ( [ eq - newconv ] ) and @xmath83 , the heat absorbed from the cold bath can be expressed as @xmath84 . thus , the cop can be expressed as @xmath85 due to @xmath47 , from which we have @xmath86 . substituting this equation into the definition of @xmath0-criterion , we obtain @xmath87/t(\varepsilon+s_h)^2.\ ] ] maximizing @xmath0 with respect to @xmath13 ( equivalently with respect to @xmath88 ) , we find that the cop at maximum @xmath0-criterion satisfies @xmath89 , which gives @xmath90 . considering @xmath91 and @xmath92 , we achieve @xmath93 if the model system is symmetrically coupled with both heat baths such that @xmath94 , we derive @xmath95 when @xmath96 from ( [ eq - coprefig23 ] ) . therefore , we have proved the conjecture on the universal cop at maximum @xmath0 for tight - coupling refrigerators symmetrically interacting with two heat baths at small temperature difference . in addition , in the extremely asymmetric cases of @xmath97 ( i.e. , @xmath98 ) and @xmath99 ( i.e. , @xmath100 ) , equation ( [ eq - coprefig23 ] ) leads to @xmath101 and @xmath102 , respectively , which surprisingly equate the global lower and upper bounds [ @xmath103 and @xmath104 of cop at maximum @xmath0 for low - dissipation refrigerators which are also found to be reached in the case of extremely asymmetric dissipations @xcite . indeed , equation ( [ eq - coprefig23 ] ) implies that @xmath105 is strictly bounded between @xmath101 and @xmath102 if the temperature difference is not too large . in this section , we adopt convention ( [ eq - newconv ] ) to recalculate the efficiency at maximum maximum @xmath0 ( i.e. , maximum power ) for tight - coupling heat engines . the heat absorbed from the hot bath can be expressed as @xmath106 with the consideration of @xmath107 . so the efficiency can be expressed as @xmath108 with the consideration of tight - coupling condition ( [ eq - tccond ] ) and @xmath63 for heat engines . on the other hand , the @xmath0-criterion degenerates into the power output for heat engines @xcite . within the linear irreversible thermodynamics , the power output can be expressed as @xmath109 with the consideration of linear relationship ( [ eq - lirrevt ] ) . maximizing @xmath110 with respect to @xmath88 , we obtain the optimized @xmath111 . substituting it into ( [ eq - effc ] ) , we obtain the efficiency at maximum power @xmath112 . by considering @xmath113 and @xmath114 , we can derive the efficiency at maximum power : @xmath115 thus @xmath116 is equal to @xmath10 up to the first order term in the limit of small relative temperature difference as it was obtained in ref . @xcite . interestingly , when the model system is symmetrically coupled with both heat baths such that @xmath117 , we arrive at @xmath118 which equates the universal efficiency at maximum power up to the second order term of @xmath8 for tight - coupling heat engines in the presence of left - right symmetry @xcite . in addition , in the extremely asymmetric cases of @xmath100 ( i.e. , @xmath99 ) and @xmath98 ( i.e. , @xmath97 ) , equation ( [ eq - effcc3 ] ) leads to @xmath119 and @xmath120 , respectively , which surprisingly equate the global lower and upper bounds [ @xmath121 and @xmath122 of efficiency at maximum power for low - dissipation heat engines which are also found to be reached in the case of extremely asymmetric dissipations @xcite . indeed , equation ( [ eq - effcc3 ] ) implies that @xmath116 is strictly bounded between @xmath119 and @xmath120 if the temperature difference is not too large . in the above discussions , we have obtained the universal cop ( [ eq - copsmalt ] ) at maximum @xmath0 for refrigerators tightly and symmetrically coupled with two baths at small temperature difference with the consideration of new convention ( [ eq - newconv ] ) as the thermodynamic flux related to the heat transfer between two baths . by adopting this convention , we have also derived the universal efficiency ( [ eq - univeta ] ) at maximum power up to the second order term for tight - coupling heat engines symmetrically interacting with two heat baths at small temperature difference . in this sense , asymptotic behavior ( [ eq - copsmalt ] ) for refrigerators tightly and symmetrically coupled with two baths at small temperature difference can be regarded as the counterpart of universal efficiency ( [ eq - univeta ] ) at maximum power for tight - coupling heat engines operating between two baths at small temperature difference in the presence of left - right symmetry . as a byproduct , we surprisingly find that the global bounds of efficiency at maximum power for low - dissipation heat engines obtained in ref . @xcite and the global bounds of cop at maximum @xmath0 for low - dissipation refrigerators obtained in ref . @xcite can be achieved when we adopt new convention ( [ eq - newconv ] ) to deal with tight - coupling heat engines or refrigerators in the extremely asymmetric cases . it is necessary for us to investigate whether this fact takes place merely by coincidence or due to some underlying reasons in the future work . the authors are grateful to the financial support of nature science foundation of china ( grant no . 11075015 ) . they also thank yang wang for carefully proofreading the manuscript . 99 yvon j 1955 the saclay reactor : two years experience on heat transfer by means of compressed gas _ proceedings of the international conference on peaceful uses of atomic energy _ report no . 387 novikov i 1957 efficiency of an atomic power generating installation _ soviet j. atomic energy _ * 3 * 1269 - 1272 chambadal p 1957 _ les centrales nucleaires _ ( paris : armand colin ) curzon f l and ahlborn b 1975 efficiency of a carnot engine at maximum power output _ am . j. phys . _ * 43 * 22 - 24 andresen b , salamon p and berry r s 1977 thermodynamics in finite time extremals for imperfect heat engines _ j. chem . phys . _ * 66 * 1571 - 1577 chen l and yan z 1989 the effect of heat - transfer law on performance of a two - heat - source endoreversible cycle _ j. chem . phys . _ * 90 * 3740 - 3743 chen j 1994 the maximum power output and maximum efficiency of an irreversible carnot heat engine _ j. phys . * 27 * 1144 - 1149 bejan a 1996 entropy generation minimization : the new thermodynamics of finite - size devices and finite - time processes _ j. appl . phys . _ * 79 * 1191 - 1218 chen l , wu c and sun f 1999 finite time thermodynamic optimization or entropy generation minimization of energy systems _ j. non - equil . thermody . _ * 24 * 327 - 359 van den broeck c 2005 thermodynamic efficiency at maximum power _ phys . lett . _ * 95 * 190602 jimnez de cisneros b and calvo hernndez a 2007 collective working regimes for coupled heat engines _ phys . lett . _ * 98 * 130602 schmiedl t and seifert u 2008 efficiency at maximum power : an analytically solvable model for stochastic heat engines _ europhys . lett . _ * 81 * 20003 tu z c 2008 efficiency at maximum power of feynman s ratchet as a heat engine _ j. phys . a : math . theor . _ * 41 * 312003 esposito m , lindenberg k and van den broeck c 2009 thermoelectric efficiency at maximum power in a quantum dot _ europhys . * 85 * 60010 esposito m , lindenberg k and van den broeck c 2009 universality of efficiency at maximum power _ phys . lett . _ * 102 * 130602 apertet y , ouerdane h , goupil c and lecoeur p 2012 efficiency at maximum power of thermally coupled heat engines _ phys . e _ * 85 * 041144 seifert u 2012 stochastic thermodynamics , fluctuation theorems and molecular machines _ rep . phys . _ * 75 * 126001 esposito m , kawai r , lindenberg k and van den broeck c 2010 efficiency at maximum power of low - dissipation carnot engines _ phys . * 105 * 150603 gaveau b , moreau m and schulman l s 2010 stochastic thermodynamics and sustainable efficiency in work production _ phys . _ * 105 * 060601 wang y and tu z c 2012 efficiency at maximum power output of linear irreversible carnot - like heat engines _ phys . e _ * 85 * 011127 wang y and tu z c 2012 bounds of efficiency at maximum power for linear , superlinear and sublinear irreversible carnot - like heat engines _ europhys . lett . _ * 98 * 40001 izumida y and okuda k 2012 efficiency at maximum power of minimally nonlinear irreversible heat engines _ europhys . lett . _ * 97 * 10004 wang j , he j and wu z 2012 efficiency at maximum power output of quantum heat engines under finite - time operation _ phys . rev 85 * 031145 wang j and he j 2012 efficiency at maximum power output of an irreversible carnot - like cycle with internally dissipative friction _ phys . rev . e _ * 86 * 051112 wang y and tu z c 2013 bounds of efficiency at maximum power for normal- , sub- and super - dissipative carnot - like heat engines _ commun . phys . _ * 59 * 175 - 178 chen l sun f and chen w 1995 optimization of the specific rate of refrigeration in combined refrigeration cycle _ energy _ * 20 * 1049 - 1053 chen l , ding z and sun f 2011 a generalized model of an irreversible thermal brownian refrigerator and its performance _ appl . math . model . _ * 35 * ( 2011 ) 2945 - 2958 chen j and yan z 1998 the effect of thermal resistances and regenerative losses on the performance characteristics of a magnetic ericsson refrigeration cycle _ j. appl . * 84 * 1791 - 1974 . lin b and chen j 2009 performance characteristics and parametric optimum criteria of a brownian micro - refrigerator in a spatially periodic temperature field _ j. phys . a : math . theor . _ * 42 * 075006 he j , chen j and hua b 2002 quantum refrigeration cycles using spin-1/2 systems as the working substance _ phys e _ * 65 * 036145 jimnez de cisneros b , arias - hernndez l a and calvo hernndez a 2006 linear irreversible thermodynamics and coefficient of performance _ phys . rev . e _ * 73 * 057103 de toms c , roco j m m , calvo hernndez a , wang y and tu z c 2013 low - dissipation heat devices : unified trade - off optimization and bounds _ phys . e _ * 87 * 012105 allahverdyan a e , hovhannisyan k and mahler g 2010 optimal refrigerator _ phys . e _ * 81 * 051129 velasco s , roco j m m , medina a and calvo hernndez a 1997 new performance bounds for a finite - time carnot refrigerator _ phys . * 78 * 3241 - 3244 yan z and chen j 1990 a class of irreversible carnot refrigeration cycles with a general heat transfer law _ j. phys . * 23 * 136 - 141 de toms c , calvo hernndez a and roco j m m 2012 optimal low symmetric dissipation carnot engines and refrigerators _ phys . e _ * 85 * 010104(r ) wang y , li m , tu z c , calvo hernndez a and roco j m m 2012 coefficient of performance at maximum figure of merit and its bounds for low - dissipation carnot - like refrigerators _ phys . e _ * 86 * 011127 izumida y , okuda k , calvo hernndez a and roco j m m 2013 coefficient of performance under optimized figure of merit in minimally nonlinear irreversible refrigerator _ europhys . lett . _ * 101 * 10005 the mapping from the endoreversible refrigerators and the symmetric low - dissipation refrigerators into tight - coupling refrigerators will be investigated in a full paper .
a unified @xmath0-criterion for heat devices ( including heat engines and refrigerators ) which is defined as the product of the energy conversion efficiency and the heat absorbed per unit time by the working substance [ de toms _ et al _ 2012 _ phys . rev . e _ * 85 * 010104(r ) ] is optimized for tight - coupling heat engines and refrigerators operating between two heat baths at temperatures @xmath1 and @xmath2 . by taking a new convention on the thermodynamic flux related to the heat transfer between two baths , we find that for a refrigerator tightly and symmetrically coupled with two heat baths , the coefficient of performance ( i.e. , the energy conversion efficiency of refrigerators ) at maximum @xmath0 asymptotically approaches to @xmath3 when the relative temperature difference between two heat baths @xmath4 is sufficiently small . correspondingly , the efficiency at maximum @xmath0 ( equivalent to maximum power ) for a heat engine tightly and symmetrically coupled with two heat baths is proved to be @xmath5 up to the second order term of @xmath6 , which reverts to the universal efficiency at maximum power for tight - coupling heat engines operating between two heat baths at small temperature difference in the presence of left - right symmetry [ esposito _ et al _ 2009 _ phys . rev . lett . _ * 102 * 130602 ] .
as first noticed by peierls @xcite , collective physical phenomena in an environment with a reduced number of dimensions can be dramatically changed with respect to our experience in three dimensions . the example of bose - einstein condensation in a uniform gas is a good illustration of the crucial role of dimensionality . in three dimensions ( 3d ) , condensation occurs at a finite temperature , and the phase of the macroscopic wave function exhibits long range order @xcite . in two dimensions , such long range order is destroyed by thermal fluctuations at any finite temperature , both for an ideal and for an interacting bose gas @xcite . in presence of repulsive interactions between particles , a uniform 2d bose gas can nevertheless undergo a phase transition from a normal to a superfluid state at a finite critical temperature . this transition was predicted by berezinskii @xcite and by kosterlitz and thouless @xcite ( bkt ) , and it has been observed in several macroscopic quantum systems , such as helium films adsorbed on a substrate @xcite . the superfluid state exhibits quasi - long range order , such that the one - body correlation function decays algebraically at large distance . by contrast the decay is exponential in the normal phase . the recent advances in the manipulation of quantum atomic gases have made it possible to address the properties of low - dimensional bose gases with novel tools and diagnostic techniques @xcite ( for recent reviews , see @xcite ) . a recent cold atom experiment also addressed the bkt problem by realizing a two - dimensional array of atomic josephson junctions @xcite . all these systems bring new questions , since one is now dealing with a harmonically trapped , instead of a uniform , fluid . in particular , due to a different density of states , even in 2d one expects to recover the bose - einstein condensation phenomenon in the ideal bose gas case @xcite . the total number of atoms in the excited states of the trap is bounded from above , and a macroscopic population of the ground state appears for large enough atom numbers . however real atomic gases do interact . it is therefore a challenging question to understand whether in presence of atomic interactions , a trapped bose gas will undergo a bkt superfluid transition like in the uniform case , or whether conventional bose - einstein condensation will take place , as for an ideal system . in recent experiments performed in our laboratory @xcite , a gas of rubidium atoms was trapped using a combination of a magnetic trap providing harmonic confinement in the @xmath1 plane , and an optical lattice , ensuring that the third degree of freedom ( @xmath2 ) of the gas was frozen . the analysis of the atomic density profile revealed a critical point , between a high temperature phase with a single component density distribution , and a low temperature phase with a clear bimodal distribution @xcite . this critical point also corresponded to the onset of clearly visible interferences between independent gases , which were used to study the coherence properties of the system @xcite . surprisingly , the density profile of the normal component was observed to be close to a gaussian all the way down ( in temperature ) to the critical point . this density profile is strikingly different from the one expected for the ideal gas close to the bec critical temperature . furthermore , if the width of the observed quasi - gaussian distribution is interpreted as an empirical measure of the temperature , this leads to a critical atom number at a given temperature which is about five times larger than that needed for conventional bose - einstein condensation in the ideal gas . these two facts showed that , in sharp contrast to the 3d case , interactions in 2d can not be treated as a minor correction to the ideal gas bec picture , but rather qualitatively change the behavior of the system . the main goal of the present paper is to analyze this critical point . we start in [ sec : ideal2d ] with a brief review of the properties of an ideal bose gas in the uniform case and in the case of harmonic confinement . in [ sec : ideallattice ] we adapt the ideal gas treatment to the experimental geometry of @xcite , and provide a detailed calculation showing that the experimental results can not be explained by this theory . next , in [ sec : interactions ] , we take interactions into account at the mean - field ( mf ) level and we combine this analysis with the numerically known threshold for the bkt transition in the uniform case @xcite . we first present a mf treatment for a strictly 2d gas . for the parameter range explored experimentally , it leads to a critical atom number in good agreement with the prediction of the most recent quantum monte carlo ( qmc ) calculations @xcite . however the predicted mf density profiles significantly differ from the qmc ones in the vicinity of the critical point . in a second step , we take into account the residual excitation of the @xmath2 motion in the mean - field model and we obtain an excellent agreement with the qmc calculation . the predicted density distribution near the critical point has a quasi - gaussian shape and the empirical " temperature extracted from this distribution is in fact somewhat lower than the real temperature . taking this into account we obtain good quantitative agreement between experimental results and theoretical predictions . finally we summarize our findings and discuss the connection between the bec and the bkt transition in a 2d gas . while in a uniform , infinite system only the latter can occur at finite temperature , in a trapped gas both are possible , and the bec transition can be viewed as a special , non - interacting limit of the more general bkt behavior . this section is devoted to a review of well known results concerning the ideal bose gas in two dimensions . we first address the case of a uniform system at the thermodynamic limit , and we then consider a gas confined in a harmonic potential . in the thermodynamic limit a uniform , ideal bose gas does not undergo bose - einstein condensation when the temperature @xmath3 is reduced , or the 2d spatial density @xmath4 is increased . bose - einstein statistics leads to the following relation between the phase space density @xmath5 and the fugacity @xmath6 @xmath7 here @xmath8 is the thermal wavelength , @xmath9 is the atomic mass , @xmath10 and @xmath11 is the chemical potential . the function @xmath12 is the polylogarithm , that takes the simple form @xmath13 for @xmath14 . because @xmath15 when @xmath16 , ( [ eq : uniform2d ] ) has a solution in @xmath17 for any value of @xmath18 . hence no singularity appears in the distribution of the population of the single particle levels , even when the gas is strongly degenerate ( @xmath19 ) . this is to be contrasted with the well known 3d case : the relation @xmath20 ceases to have a solution for a phase space density above the critical value @xmath21 , where the 3d bose - einstein condensation phenomenon takes place . we now consider an ideal gas confined in a harmonic potential @xmath22 . we assume that thermal equilibrium has been reached , so that the population of each energy level is given by bose - einstein statistics . since the chemical potential @xmath11 is always lower than the energy @xmath23 of the ground state of the trap , the number of atoms @xmath24 occupying the excited states of the trap can not exceed the critical value @xmath25 @xmath26 this expression can be evaluated in the so - called _ semi - classical limit _ @xmath27 by replacing the discrete sum by an integral over the energy ranging from @xmath28 to @xmath29 @xcite : @xmath30 with @xmath31 . this result also holds in the case of an anisotropic harmonic potential in the @xmath1 plane , in which case @xmath32 is replaced by the geometric mean @xmath33 , where @xmath34 are the two eigenfrequencies of the trap . the saturation of the number of atoms in the excited states is a direct manifestation of bose - einstein condensation : any total atom number @xmath35 above @xmath25 must lead to the accumulation of at least @xmath36 in the ground state of the trap . equation ( [ eq : idealtrapped2 ] ) is very reminiscent of the result for the harmonically trapped 3d gas , where the saturation number is @xmath37 . however an important difference arises between the 2d and the 3d cases for the spatial density profile . in 3d the phase space density in @xmath38 is given by @xmath39 in the limit @xmath40 . the threshold for bose - einstein condensation is reached for @xmath41 ; at this point @xmath35 is equal to the critical number @xmath42 and simultaneously the phase space density at the center of the trap @xmath43 equals the critical value @xmath44 . this allows for a simple connection between the bec thresholds for a homogenous gas and for a trapped system in the semi - classical limit @xmath40 . in 2d such a simple connection between global properties ( critical atom number @xmath25 ) and local properties ( critical density at center @xmath45 ) does not exist . indeed the semi - classical expression of the 2d phase space density is @xmath46 because @xmath47 this leads to a diverging value at the center of the trap when @xmath17 approaches 1 . therefore , although the integral of @xmath48 over the whole space converges for @xmath41 and allows to recover ( [ eq : idealtrapped2 ] ) , the semiclassical result ( [ eq:2didealphasespace ] ) can not be used to derive a local criterion for condensation at the center of the trap . one can go beyond the semi - classical approximation and calculate numerically the central phase space density as a function of the total number of atoms . we consider as an example the trap parameters used in @xcite , where @xmath49 hz , @xmath50 hz . in the typical case @xmath51 ( @xmath52 nk ) , the discrete summation of the bose - einstein occupation factors for @xmath41 gives @xmath53 ( the value obtained from the semi - classical result ( [ eq : idealtrapped2 ] ) is 4100 ) . using the expression of the energy eigenstates ( hermite functions ) , we also calculate the phase space density at the origin and we find @xmath54 . let us emphasize that this value is a mere result of the finite size of the system , and does not have any character of universality . in order to produce a quasi two - dimensional gas experimentally , one needs to freeze the motion along one direction of space , say @xmath2 . in practice this is conveniently done using the potential @xmath55 created by an optical lattice along this direction . a precise comparison between the measured critical atom number and the prediction for an ideal gas requires to properly model the confining potential and find its energy levels . this is the purpose of the present section . the optical lattice is formed by two running laser waves of wavelength @xmath56 , propagating in the @xmath57 plane with an angle @xmath58 with respect to the @xmath59 axis . the period @xmath60 of the lattice along the @xmath2 direction can be adjusted to any value above @xmath61 by a proper choice of the angle @xmath62 . for a blue - detuned lattice ( @xmath56 is smaller than the atomic resonance wavelength ) , @xmath63 is positive and the atoms accumulate in the vicinity of the nodal planes @xmath64 , @xmath65 , etc . the oscillation frequency at the bottom of the lattice wells is @xmath66 , where @xmath67 . in order for the quasi-2d regime to be reached , @xmath68 must notably exceed the typical thermal energy @xmath69 as well as the interaction energy per particle for a non ideal gas . for a blue detuned lattice an additional confinement in the @xmath1 plane must be added to the optical lattice potential . this is conveniently achieved using a magnetic trap , that creates a harmonic potential with frequencies @xmath34 . the magnetic trap also provides an additional trapping potential @xmath70 along the @xmath2 direction . the oscillation frequency @xmath71 created by the magnetic trap is usually much lower than the one created by the lattice @xmath72 . the main effect of the magnetic confinement along the @xmath2 direction is to localize the atoms in the @xmath73 central lattice planes , where the effective number of planes @xmath74 . as we see below this number is on the order of 2 to 4 for the range of parameters explored in @xcite . the fact that more than just one plane is populated is an important ingredient of the experimental procedure used in @xcite . it allows one to look for the interferences between planes , and to access in this way the spatial coherence of the quasi-2d gas . in order to extract thermodynamic information from the interference between planes , one must ensure that the various populated planes have the same temperature . this is achieved by using finite size lattice beams in the @xmath1 plane , so that atoms in the high energy tail of the thermal distribution can actually travel quasi freely from one plane to the other , thus ensuring thermalization . in the experiments described in @xcite , the waist @xmath75 of the lattice beams along the @xmath76 direction was chosen accordingly . the total trapping potential can then be written in the following way @xmath77 with @xmath78 note that we have included here the offset @xmath79 between the optical lattice and the bottom of the magnetic potential ; this quantity was not set to a fixed value in the experiments @xcite . we consider below two limiting situations : ( a ) @xmath80 , with two principal equivalent minima in @xmath81 , ( b ) @xmath82 , with one principal minimum in @xmath64 and two side minima in @xmath83 . at very low temperatures , we expect that a will lead to two equally populated planes whereas configuration b will lead to one populated plane . for the temperature range considered in practice , the differences between the predictions for a and b are minor , as we will see below . in order to use the bose - einstein statistics for an ideal gas , one needs to know the position of the single particle energy levels . for the potential ( [ eq : totalpot ] ) it is not possible to find an exact analytical expression of these levels . however if the extension of the atomic motion along the @xmath76 direction is smaller than the laser waist , an approximate expression can be readily obtained , as we show now . the frequencies of the magnetic trap used in @xcite are @xmath85 hz and @xmath86 hz . the optical lattice has a period @xmath87 m ( @xmath88 hz ) and a potential height at center @xmath89 khz ( @xmath90k ) . the lattice oscillation frequency at center ( @xmath91 ) is thus @xmath92 khz ( @xmath93 nk ) . when the atoms occupy the ground state of the @xmath2 motion , they acquire the zero - point energy @xmath94 from the @xmath95degree of freedom . the dependance on @xmath76 of @xmath96 , due to the gaussian term @xmath97 in the laser intensity , causes a renormalization of the @xmath76 frequency : @xmath98 the waist of the lattice beams is @xmath99 m which leads to @xmath100 hz . a similar effect should in principle be taken into account for the frequency @xmath101 . however the scale of variation of the laser intensity along the @xmath59 axis is the rayleigh length , which is much larger than the waist @xmath75 , and the effect is negligible . this simple way of accounting for the finiteness of the waist @xmath75 is valid when the extension of the motion along @xmath76 is small compared to @xmath75 . for @xmath102 nk , the width of the thermal distribution along @xmath76 is @xmath103 m , which is indeed notably smaller than @xmath75 . taking into account the finiteness of @xmath75 by a mere reduction of the trapping frequency along @xmath76 is therefore valid for the major part of the energy distribution . we note however that atoms in the high energy tail of the distribution ( @xmath104 for our largest temperatures ) can explore the region @xmath105 , where the influence of the lattice beams is strongly reduced . in this region the atoms can move from one lattice plane to the other . as explained above these atoms play an important role by ensuring full thermalization between the various planes . we now turn to an accurate treatment of the critical atom number required for bose - einstein condensation , taking into account these high energy levels for which the 2d approximation is not valid . in order to get the single particle energy eigenstates in the lattice + harmonic potential confinement , and thus the critical atom number , one could perform a numerical diagonalization of the 3d hamiltonian with the potential ( [ eq : totalpot ] ) . this is however a computationally involved task and it is preferable to take advantage of the well separated energy scales in the problem . we first note that the trapping potential ( [ eq : totalpot ] ) is the sum of a term involving the variables @xmath76 and @xmath2 , and a quadratic component in @xmath59 . the motion along the @xmath59 axis can then be separated from the @xmath106 problem , and it is easily taken into account thanks to its harmonic character . for treating the @xmath106 problem we use a ` born - oppenheimer ' type approximation . we exploit the fact that the characteristic frequencies of the @xmath2 motion are in any point @xmath76 notably larger than the frequency of the @xmath76 motion . this is of course true inside the lattice laser waist , since @xmath107 , and it is also true outside the laser waist as the @xmath76 direction corresponds to the weak axis of our magnetic trap . therefore we can proceed in two steps : 1 . for any fixed @xmath76 we numerically find the eigenvalues @xmath108 , @xmath109 of the @xmath2 motion in the @xmath110 potential . we determine the @xmath111 s up to the threshold @xmath112 above which the thermal excitation of the levels is negligible . the result of this diagonalization is shown in figure [ fig : eigenvalues ] for configurations @xmath113 and @xmath114 . + -motion in the magnetic+optical @xmath115 potential , for fixed values of the @xmath76 coordinate ( left : configuration a , right : configuration b).,title="fig:",width=264 ] -motion in the magnetic+optical @xmath115 potential , for fixed values of the @xmath76 coordinate ( left : configuration a , right : configuration b).,title="fig:",width=264 ] 2 . we then treat semi - classically the @xmath76-motion on the various potential curves @xmath108 . adding the result for the independent harmonic @xmath59 motion also treated semi - classically we obtain the surface density @xmath116 ( integral of the spatial density @xmath117 along the direction @xmath2 ) @xmath118 this procedure yields a result that is identical to the semi - classical prediction in two limiting cases : 1 . the pure 2d case , that is recovered for large waists and low temperatures . in this case the restriction to the closest - to - center lattice plane and to the first @xmath2-level is legitimate , and the sum over @xmath119 contains only one significant term corresponding to ( [ eq:2didealphasespace ] ) . the pure 3d harmonic case with zero lattice intensity where @xmath120 . in this case the sum over @xmath119 in ( [ eq : maglatresult ] ) leads to @xmath121 , which coincides with the 3d result @xmath39 when integrated along @xmath2 . of course this procedure also allows to interpolate between these two limiting cases , which is the desired outcome . the integral of @xmath4 in the @xmath1 plane for @xmath122 gives the critical atom number @xmath123 in the ideal gas model for this lattice geometry . it is shown in figure [ fig : nc_lattice]a for the two configurations a and b. the critical atom number @xmath124 can be compared with the result for a single plane @xmath25 with eigenfrequencies @xmath125 and @xmath101 . the ratio gives the effective number of planes @xmath126 , shown as a function of temperature in figure [ fig : nc_lattice]b for the two configurations a and b. this ratio increases with temperature , which means that @xmath124 increases faster than @xmath127 with temperature in the temperature domain considered here . for example in the range @xmath128 nk , the variation of @xmath124 is well represented by @xmath129 , with @xmath130 . three phenomena contribute significantly to this faster than @xmath127 " increase of @xmath124 . first in the lattice + harmonic potential geometry , the number of contributing planes increases with temperature , even if the atomic motion in each plane remains two - dimensional ( i.e. the atom number per plane increasing strictly as @xmath127 ) . second , we are exploring here a region of temperature where @xmath131 becomes non negligible with respect to @xmath132 ( the two quantities are equal for @xmath133 nk ) , and the thermal excitations of the @xmath2-motion in each lattice plane can not be fully neglected . third , for the largest considered temperatures , the extension of the atomic motion along @xmath76 becomes comparable to the laser waist , and the lattice strength is then significantly reduced . in the ideal gas model for the optical lattice + magnetic trap configuration , as a function of temperature . the points represent the experimental results of @xcite , and the error bars combine the systematic and statistical uncertainties on atom numbers . ( b ) effective number of planes @xmath134 as a function of temperature . in both panels the continuous ( dashed ) line is for configuration a ( b ) . the calculation is performed using the first 100 eigenvalues of the @xmath2 motion , and the first neglected levels @xmath108 are 22 khz ( @xmath135k ) above the bottom of the trap.,title="fig:",width=302 ] in the ideal gas model for the optical lattice + magnetic trap configuration , as a function of temperature . the points represent the experimental results of @xcite , and the error bars combine the systematic and statistical uncertainties on atom numbers . ( b ) effective number of planes @xmath134 as a function of temperature . in both panels the continuous ( dashed ) line is for configuration a ( b ) . the calculation is performed using the first 100 eigenvalues of the @xmath2 motion , and the first neglected levels @xmath108 are 22 khz ( @xmath135k ) above the bottom of the trap.,title="fig:",width=264 ] in @xcite the critical atom number in the lattice + magnetic trap configuration was measured for various effective " temperatures , deduced from the width of the quasi - gaussian atomic distribution . each critical point ( @xmath136 ) was defined as the place where a bimodal spatial distribution appeared , if the atom number was increased beyond this point at constant temperature , or the temperature reduced at constant atom number . the critical point also corresponded to the threshold for the appearance of interferences with a significant contrast between adjacent planes . the experimental measurements of critical points , taken over the effective temperature range 50110 nk , are shown as dots in figure [ fig : nc_lattice]a . the systematic + statistical uncertainties of the atom number calibration are 25% . assuming that the effective temperatures coincide with the true ones ( this point will be examined in [ subsec : latticemf ] ) we find @xmath137 . in addition to this large discrepancy between experiment and ideal gas model for the critical atom numbers , one also finds a strong mismatch concerning the functional shape of the column density @xmath138 that was measured in absorption imaging in @xcite . while the experimental result is quasi - gaussian , the column density profiles calculated for an ideal gas at the critical point are much ` peakier ' . an example is given in the appendix for a single plane , and we checked that a similar shape remains valid for our harmonic + lattice potential . we therefore conclude that the experimental results of @xcite can not be accounted for with this ideal gas prediction for ` conventional ' bec . to improve the agreement between the experimental results and the theoretical modeling we now take repulsive atomic interactions into account . in a first stage we present a 2d mean - field analysis , in which the motion along @xmath2 is assumed to be completely frozen whereas the @xmath1 motion is treated semi - classically . in order to model interactions in this case , we start from the 3d interaction energy @xmath139 , where @xmath140 and @xmath141 is the scattering length . the @xmath2-degree of freedom is restricted to the gaussian ground state of the confining potential , with an extension @xmath142 , and the interaction energy is @xmath143 we set @xmath144 , where the dimensionless parameter @xmath145 characterizes the strength of the 2d interaction ( for a more elaborate treatment of atomic interactions in a quasi-2d geometry , see @xcite ) . for the optical lattice used in @xcite we find @xmath146 . in a second stage we take into account the residual excitation of the @xmath2-motion in a ` hybrid ' 3d mean - field approximation . we calculate in a self - consistent way the quantum levels of the @xmath2 motion , whereas the motion in the @xmath1 plane is still treated semi - classically . we start our discussion with a brief reminder of the role of ( weak ) interactions in a trapped 3d bose gas @xcite . one often uses the mean - field hartree - fock approximation , that gives in particular a relatively accurate value for the shift of the critical temperature for bose - einstein condensation . in order to calculate this shift , one assumes that above the critical temperature , the atoms evolve in the effective potential @xmath147 . the phase space density in @xmath148 is thus a solution of @xmath149 . as for the ideal case this equation ceases to have a solution when the central phase space density goes above @xmath44 . the mere effect of repulsive interactions within the mean - field approximation is to increase the number of atoms for which this threshold is met . the increase is typically @xmath150 for standard trap and interaction parameters @xcite . for a trapped 2d gas this treatment based on a local criterion ( phase space density at center ) can not be used . indeed as explained in section [ subsec : ideal2dtrap ] , it is not possible to identify a critical phase space density at which bec of the 2d gas is expected . on the contrary the semiclassical approximation leads to an infinite central density at the critical point , and it is unclear whether one can achieve an arbitrarily large spatial density in presence of repulsive interactions . one could also look for a global criterion for criticality based on the total atom number . the starting point is the solution of the mean - field equation @xmath151 with @xmath152 . when @xmath153 , we saw in [ sec : ideal2d ] that the solution of ( [ eq:2dhartreefock ] ) can only accommodate a finite number of atoms ( [ eq : idealtrapped2 ] ) . however the situation is dramatically changed in presence of repulsive interactions . indeed for any non zero @xmath154 , a solution to ( [ eq:2dhartreefock ] ) exists for arbitrarily large atom numbers @xcite . consequently no critical point can be found by simply searching for a maximal atom number compatible with ( [ eq:2dhartreefock ] ) . in the following we will therefore turn to a different approach , starting from the known exact ( i.e. non mean - field ) results concerning the critical bkt point in a uniform interacting 2d bose gas . the mean - field approximation will be used in a second stage , in combination with the local density approximation ( lda ) , to determine the critical atom number in the trapped system . note that it is also possible to pursue the search for a critical point only within the mean - field approach , by looking for example whether its solution exhibits a thermodynamical or dynamical instability above a critical atom number @xcite . this instability would be an indication that the system tends to evolve towards a different kind of state , with a non - zero quasi - condensed and/or superfluid component , and quasi - long range order @xcite . in an infinite uniform 2d bose fluid , repulsive interactions have a dramatic effect since they can induce a transition from the normal to the superfluid state , when the temperature is lowered below a critical value . the superfluid density jumps from 0 to @xmath155 at the transition point @xcite . the microscopic mechanism of the 2d superfluid transition has been elucidated by berezinskii @xcite and by kosterlitz and thouless @xcite . for a temperature larger than the critical temperature , free vortices proliferate in the gas , destroying the superfluidity . below the transition , vortices exist only in the form of bound pairs involving two vortices of opposite circulations , which have little influence on the superfluid properties of the system . in a uniform system the phase space density @xmath18 is a function of the chemical potential and temperature @xmath156 . for any given @xmath3 , the superfluid transition occurs when @xmath11 is equal to a critical value @xmath157 . the corresponding critical value @xmath158 for the phase space density depends on the interaction strength as @xcite @xmath159 where @xmath160 is a dimensionless number . a recent monte - carlo analysis provided the result @xmath161 @xcite ( see also @xcite ) . for @xmath146 this gives a critical phase space density @xmath162 . we consider now a trapped gas whose size is large enough to be well described by the local density approximation ( lda ) . the phase space density in @xmath38 is given by @xmath163 and a superfluid component forms around the center of the trap if the central phase density @xmath164 is larger than @xmath158 @xcite . the edge of the superfluid region corresponds to the critical line where @xmath165 . the phase space density along this line is equal to @xmath158 , independently of the total number of atoms in the trap . this can be checked experimentally and constitutes a validation of the lda . the integration of the experimental data along the line of sight @xmath59 does not lead to any complication because the trapping potential is separable , @xmath166 . therefore the edges of the superfluid region along the @xmath76 axis are located in @xmath167 such that @xmath168 ( see figure [ fig : ldacheck]a ) , and the column density along the line of sight passing in @xmath169 is @xmath170 which is also independent of the total atom number @xmath35 . this is confirmed experimentally , as shown in figure [ fig : ldacheck]b where we plot @xmath171 as a function of @xmath35 . the slight increase ( 10% ) of @xmath171 for atom numbers larger than @xmath172 may be due to the fact that the population of additional planes becomes non - negligible for such large @xmath35 . the possibility to use the lda to study the berezinskii - kosterlitz - thouless critical point in a harmonically trapped quasi-2d bose gas has been checked recently by holzmann and krauth using a quantum monte - carlo analysis @xcite . for trap parameters close to the ones of @xcite they have shown that a superfluid component , characterized by a reduced moment of inertia , indeed appears at the center of the trap when the local phase space density reaches a critical value close to the prediction ( [ eq : prokofev ] ) . is measured at the edge @xmath169 of the central part ( in grey ) of the bimodal distribution . ( b ) @xmath171 is plotted as a function of the total atom number @xmath35 in the harmonic trap + lattice configuration . within the lda for a single plane , @xmath171 should be independent of @xmath35 , which is indeed nearly the case . the small variation of @xmath171 for large @xmath35 may be due to the appearance of a non negligible population in side planes of the optical lattice potential . the data have been taken for the effective temperature @xmath173 nk . each point is extracted from a single image.,title="fig:",width=264 ] is measured at the edge @xmath169 of the central part ( in grey ) of the bimodal distribution . ( b ) @xmath171 is plotted as a function of the total atom number @xmath35 in the harmonic trap + lattice configuration . within the lda for a single plane , @xmath171 should be independent of @xmath35 , which is indeed nearly the case . the small variation of @xmath171 for large @xmath35 may be due to the appearance of a non negligible population in side planes of the optical lattice potential . the data have been taken for the effective temperature @xmath173 nk . each point is extracted from a single image.,title="fig:",width=264 ] in this section we use the mean - field hartree - fock approximation ( [ eq:2dhartreefock ] ) to calculate the density profile of the trapped atomic cloud . as we mentioned above , this equation admits a solution for any value of the fugacity @xmath17 , and therefore for an arbitrarily large number of particles . rewriting ( [ eq:2dhartreefock ] ) as @xmath174 we see that the value of @xmath18 for any temperature and at any point in space depends only on the parameter @xmath175 defined by @xmath176 , where @xmath177 and @xmath178 . the total atom number is given by @xmath179 where @xmath180 is the solution of the reduced equation @xmath181 quite remarkably this result for @xmath180 neither depends on the trap parameters , nor on the temperature . the only relevant parameters are the fugacity @xmath17 and the reduced interaction strength @xmath182 . the scaling of the atom number @xmath35 with the temperature and the trap frequency in ( [ eq : atom_number ] ) is therefore very simple . in particular it does not depend on the trap anisotropy @xmath183 but only on the geometric mean @xmath184 . for atom numbers much larger than @xmath25 it is interesting to note that the radial density profile deduced from the mean - field equation ( [ eq : hfford ] ) exhibits a clear bi - modal shape , with wings given by @xmath185 and a central core with a thomas - fermi profile @xmath186 . however this prediction of a bi - modal distribution using the hartree - fock approximation can not be quantitatively correct . indeed the hartree - fock treatment assumes a mean - field energy @xmath187 . the factor 2 in front of this energy originates from the hypothesis that density fluctuations are those of a gaussian field @xmath188 . actually when the phase space density becomes significantly larger than 1 , density fluctuations are reduced and one approaches a situation closer to a quasi - condensate in which @xmath189 @xcite . taking into account this reduction could be done for example using the equation of state obtained from a classical field monte - carlo analysis in @xcite . . the dotted line represents the semi - classical prediction for the ideal gas . the dashed line indicates where the threshold for superfluidity ( [ eq : prokofev ] ) is met at the center of the trap . ( b ) column density for @xmath190 atoms in an isotropic trap ( @xmath191 , @xmath146 , @xmath192 ) . the dashed line is the result of the 2d mean - field analysis of [ sec : trapped_interacting ] . the continuous line is the 3d quantum monte - carlo result obtained in @xcite , with @xmath193 and @xmath194 . the dots are the result of the hybrid 3d mean - field calculation of [ subsec : hybridmf ] for the same parameters . inset : prediction of the hybrid 3d mean - field approach for the phase space density ( log scale ) as a function of @xmath195 ( a gaussian distribution leads to a straight line ) . continuous line : total phase space density ; dashed line : phase space density associated to the ground state @xmath196 of the @xmath2 motion ; dash - dot line : phase space density associated to all other states @xmath197 , @xmath198.,title="fig:",width=264 ] . the dotted line represents the semi - classical prediction for the ideal gas . the dashed line indicates where the threshold for superfluidity ( [ eq : prokofev ] ) is met at the center of the trap . ( b ) column density for @xmath190 atoms in an isotropic trap ( @xmath191 , @xmath146 , @xmath192 ) . the dashed line is the result of the 2d mean - field analysis of [ sec : trapped_interacting ] . the continuous line is the 3d quantum monte - carlo result obtained in @xcite , with @xmath193 and @xmath194 . the dots are the result of the hybrid 3d mean - field calculation of [ subsec : hybridmf ] for the same parameters . inset : prediction of the hybrid 3d mean - field approach for the phase space density ( log scale ) as a function of @xmath195 ( a gaussian distribution leads to a straight line ) . continuous line : total phase space density ; dashed line : phase space density associated to the ground state @xmath196 of the @xmath2 motion ; dash - dot line : phase space density associated to all other states @xmath197 , @xmath198.,title="fig:",width=302 ] we now use the solution of the mean - field equation ( [ eq : hfford ] ) to evaluate the critical atom number @xmath199 that is needed to reach the threshold ( [ eq : prokofev ] ) for the bkt transition at the center of the trap @xmath200 . for a given interaction strength @xmath182 we vary the fugacity @xmath17 and solve numerically ( [ eq : hfford ] ) at any point in space . examples of spatial density profiles at the critical point are given in the appendix , both before and after time - of - flight . the integration of the density profile over the whole @xmath1 plane gives the total atom number @xmath35 . from ( [ eq : atom_number ] ) and ( [ eq : intermediate ] ) , it is clear that the scaling of @xmath199 with the frequencies @xmath201 and with the temperature is identical to the one expected for an ideal trapped gas . we have plotted in figure [ fig : nat_fsd]a the variation of @xmath164 as a function of @xmath202 for various interaction strengths . for a given atom number the phase space density at center decreases when the strength of the interactions increases , as expected . the numerical result for @xmath199 is plotted as a dashed line in figure [ fig : nat_fsd]a . we find that it is in excellent agreement to better than 1% with the result of @xcite @xmath203 over the whole range @xmath204@xmath205 . this analytical result was initially derived in @xcite for @xmath206 using an expansion around the solution for the ideal bose gas , but this approximation can actually be extended to an arbitrary value of @xmath182 @xcite . the strongly interacting limit ( @xmath207 ) can be easily understood by noticing that in this case , the atomic distribution ( [ eq:2dhartreefock ] ) nearly coincides with the thomas - fermi profile @xmath208 . using the relation between the total atom number and the central density for this thomas - fermi distribution @xmath209 ( with @xmath210 ) , one then recovers the second term of the right - hand side of ( [ eq : ncanalytics ] ) . let us emphasize that figure [ fig : nat_fsd]a is a mix of two approaches : ( i ) the mean - field model , that does not lead in itself to a singularity along the dashed line of figure [ fig : nat_fsd]a . ( ii ) the bkt theory for a uniform system , which is a beyond mean - field treatment and which has been adapted to the trapped case using the local density approximation in order to obtain the critical number indicated by the dashed line . we now compare the 2d mf prediction with the results of the qmc calculation of @xcite , looking first at the critical atom number and then at the density profiles . for @xmath146 the mean field prediction for the critical number ( [ eq : ncanalytics ] ) is @xmath211 . this is in relatively good agreement with the qmc calculation of @xcite , which gives @xmath212 or equivalently @xmath213 . the qmc calculation has been performed for various atom numbers @xmath35 , for a 3d harmonic trap such that @xmath214 and a 3d scattering length @xmath215 , with @xmath216 . the agreement between the 2d mf and the qmc approaches is not as good for the density profiles close to the critical point . an example is shown in figure [ fig : nat_fsd]b , where we take @xmath217 ( @xmath218 ) . we choose @xmath190 and we plot the column density @xmath219 obtained by integrating the spatial density along the directions @xmath59 and @xmath2 . the mean field result is shown as a dashed line , and it notably differs from the qmc result , plotted as a continuous line . as we show below this disagreement is essentially a consequence of the residual excitation of the @xmath2 motion ( @xmath220 ) , that is neglected in the 2d mf approach , whereas it is implicitly taken into account in the 3d qmc calculation . in this section we extend the 2d mean - field treatment to take into account the residual excitation of the @xmath2 motion . as pointed out to us by the authors of @xcite , this is necessary for a quantitative analysis of the experiment @xcite , since the temperature and the chemical potential at the critical point were not very small compared to @xmath221 , but rather comparable to it . we follow here a method related to the one developed in section [ sec : ideallattice ] to analyze the ideal gas case . we start from a trial 3d density distribution @xmath117 . at any point @xmath222 , we treat quantum mechanically the @xmath2 motion and solve the eigenvalue problem for the @xmath2 variable @xmath223\varphi_j(z|x , y)= e_j(x , y)\;\varphi_j(z|x , y)\ , \ ] ] where @xmath224 and latexmath:[$ \int @xmath1 degrees of freedom , we obtain a new spatial density @xmath226 we then iterate this calculation until the spatial density @xmath117 reaches a fixed point @xcite . with this method , we fulfill two goals : ( i ) we take into account the residual thermal excitation of the levels in the @xmath2 direction . ( ii ) even at zero temperature we take into account the deformation of the @xmath2 ground state due to interactions . this ` hybrid 3d mean - field ' method is different from the standard mean - field treatment used to describe 3d bose gases . in the latter case , all three degrees of freedom are treated semi - classically , which is valid when the particle population is distributed smoothly over several quantum states . this standard 3d mean - field would not be applicable in our case , where a significant part of the total population accumulates in the lowest state @xmath196 . an example of a result is shown in figure [ fig : nat_fsd]b , where we plot the column density @xmath227 obtained with this hybrid mf method , taking into account the first 5 eigenstates @xmath197 . the agreement between the hybrid 3d mean - field prediction and the ` exact ' qmc prediction of @xcite is excellent . this shows that the predictions of this hybrid 3d mean - field approach are reliable as long as the superfluid transition has not been reached at the center of the trap . we show in the inset the variations of the phase space density @xmath228 . we plot @xmath229 as a function of @xmath195 , so that a gaussian distribution would appear as a straight line . the dashed line is the phase - space density associated to the ground state of the @xmath2-motion @xmath196 , and the dash - dot line corresponds to the excited states @xmath197 , @xmath198 . the continuous line is the total phase - space density . at the center of the trap , as a consequence of bose statistics , most of the population ( 80% ) accumulate in the ground state @xmath196 . for @xmath230 , the repartition of the population among the eigenstates of the @xmath2 motion follows the boltzmann law , and @xmath231 of the atoms occupy the excited states @xmath197 , @xmath198 . a practical consequence of this increasing influence of excited states of the @xmath2-motion with increasing @xmath232 is that the total phase space density profile is notably closer to a gaussian distribution than when only the ground state of the @xmath2-motion is retained in the calculation , as it is the case in the 2d mean - field approach . finally we mention that we have also developed a simpler version of this hybrid 3d mean - field analysis , in which the @xmath197 levels are not calculated self - consistently , but are assumed to coincide with the energy eigenstates in the potential @xmath70 ( see also @xcite ) . for the domain of parameters relevant for the experiment , the two approaches lead to very similar results . in order to compare the predictions of the mean - field approach with the experimental results , we now turn to the lattice geometry , corresponding to a stack of parallel planes located in @xmath233 , @xmath119 integer . for simplicity we assume that the laser waist @xmath75 is large compared to the spatial extent of the atomic cloud , so that we can treat the gas as a superposition of independent harmonically trapped systems . each system is located at the vicinity of a nodal plane of the optical lattice , and is treated as a harmonic trap with frequencies @xmath234 . note that we include here ` by hand ' the renormalization @xmath235 of the @xmath76 frequency due to the finiteness of the laser waist , that we discussed in [ subsec : renorm ] . the magnetic trap adds an extra confinement along the @xmath2 axis with a frequency @xmath71 so that the chemical potential for the plane @xmath236 is @xmath237 . we assume that the critical point is reached when the phase space density associated to the lowest eigenstate @xmath196 in the most populated plane reaches the critical value ( [ eq : prokofev ] ) . once the corresponding fugacity is determined , we calculate the spatial distribution in each plane , sum up the various contributions , and integrate the spatial distribution over the line of sight @xmath59 to obtain the column density @xmath227 . and @xmath238 atoms . for these parameters the phase space density associated to the lowest eigenvalue of the @xmath2 motion reaches the critical value @xmath158 at the center of the most populated planes . the dotted line is a gaussian fit which gives the effective temperature @xmath239 nk . ( b ) critical atom number as a function of the effective temperature obtained from a gaussian fit of the mf result . the continuous line ( resp . dashed ) line is for configuration a ( resp . the points are the experimental results of @xcite , already shown in fig . [ fig : nc_lattice ] . , title="fig:",width=264 ] and @xmath238 atoms . for these parameters the phase space density associated to the lowest eigenvalue of the @xmath2 motion reaches the critical value @xmath158 at the center of the most populated planes . the dotted line is a gaussian fit which gives the effective temperature @xmath239 nk . ( b ) critical atom number as a function of the effective temperature obtained from a gaussian fit of the mf result . the continuous line ( resp . dashed ) line is for configuration a ( resp . the points are the experimental results of @xcite , already shown in fig . [ fig : nc_lattice ] . , title="fig:",width=302 ] a typical result for @xmath240 is given in figure [ fig : mflattice]a for the temperature @xmath133 nk and for the lattice configuration a. the total atom number is @xmath241 . it is well fitted by a gaussian distribution @xmath242 ( dotted line ) , so that we can assign an effective temperature to this distribution @xmath243 . in the example of fig.[fig : mflattice]a , we find @xmath244 ( 96 nk ) . for the same @xmath3 and a lattice in configuration b , the effective temperature obtained with a gaussian fit is @xmath245 ( 103 nk ) and the total atom number is @xmath246 . we have repeated this procedure for temperatures @xmath3 in the range 100 - 200 nk and consistently found the ratio @xmath247 in the range @xmath248@xmath249 , with a quality of the gaussian fit similar to what is shown in figure [ fig : mflattice]a . we have plotted in figure [ fig : mflattice]b the calculated total number of atoms in the lattice at the critical point , as a function of the effective temperature deduced from the gaussian fit to the column density . we have also plotted the experimental points of @xcite already shown in figure [ fig : nc_lattice ] . we remind the reader that the experimental ` effective ' temperature is also deduced from a gaussian fit to the measured column density . one reaches in this way a good agreement between the experimental results and the hybrid 3d mean - field prediction . the predicted density profiles with the 3d mf approach therefore provide a satisfactory means for temperature calibration . they indicate in particular that for the experiment @xcite , the effective temperatures are typically @xmath250@xmath251 below the real ones . to improve on the comparison between theory and experiment , a more controlled setup will be needed with an accurate independent measurement of temperature , as well as the possibility of addressing only a single or a fixed number of planes . in this paper we have analyzed the critical point of a trapped quasi-2d bose gas . we have shown that the experimental results of @xcite are not in agreement with the ideal bose gas theory . the differences are found first at the qualitative level : the predicted shape for the ideal gas distribution is ` peaky ' around its center , which clearly differs from the quasi - gaussian measured profile . also the measured critical atom numbers @xmath252 do not agree with the predictions for the ideal gas . using the ` effective ' temperatures obtained by treating the gaussian profiles as boltzmann distributions , the measured @xmath252 are larger by a factor @xmath253 than the predicted ones . we then discussed the predictions of a hybrid approach based on the local density approximation . it combines the density profile calculated using a mean - field hartree - fock treatment , and the known result for the critical phase space density for the bkt transition in an infinite , uniform 2d bose gas @xcite . we compared the predictions of this approach with the results of a recent quantum monte - carlo calculation @xcite and reached the following conclusions : ( i ) if one is interested only in the critical atom number , it is sufficient to use a strictly 2d mean - field approach . it leads to the approximate analytical result ( [ eq : ncanalytics ] ) , in good agreement with the qmc prediction . for the experimental parameters of @xcite the critical atom number is @xmath254 . ( ii ) in order to calculate accurately the density profiles for the experimental temperature range ( @xmath69 between @xmath255 and @xmath256 ) , it is important to take into account the residual excitation of the @xmath2 degree of freedom ( the same conclusion has been reached in @xcite ) . we have presented a hybrid 3d mean - field approximation which leads to density distributions in excellent agreement with the qmc predictions close to the critical point . when generalized to the lattice geometry used in the experiment , the predicted density profiles are close to a gaussian distribution , and a good agreement between theory and experiment is reached concerning the critical number @xmath257 when the predicted density profile is used for temperature calibration . in dashed line and effective mean - field potential @xmath258 in continuous line , for @xmath146 and a central phase space density equal to the critical value ( [ eq : prokofev ] ) . ( b ) schematic representation of the condensed fraction in a finite 2d bose gas for a given interaction strength @xmath182 ( continuous line ) . two limits can be considered : ( 1 ) thermodynamic limit @xmath259 , @xmath260 , @xmath261 constant ; the condensed fraction tends to zero for any non zero value of @xmath182 . ( 2 ) ideal gas limit @xmath262 . , title="fig:",width=302 ] in dashed line and effective mean - field potential @xmath258 in continuous line , for @xmath146 and a central phase space density equal to the critical value ( [ eq : prokofev ] ) . ( b ) schematic representation of the condensed fraction in a finite 2d bose gas for a given interaction strength @xmath182 ( continuous line ) . two limits can be considered : ( 1 ) thermodynamic limit @xmath259 , @xmath260 , @xmath261 constant ; the condensed fraction tends to zero for any non zero value of @xmath182 . ( 2 ) ideal gas limit @xmath262 . , title="fig:",width=302 ] we now briefly discuss the nature of the critical point that appears in the trapped 2d bose gas and compare it with ` standard ' bose - einstein condensation . for a harmonically trapped ideal gas , we recall that conventional bose - einstein condensation is expected in the thermodynamic limit @xmath263 , @xmath260 , @xmath261 constant . this is a consequence of the density of states for a quadratic hamiltonian around the zero energy . the price to pay for this condensation in a 2d system is a diverging atomic density at the center of the trap . on the contrary when interactions are taken into account , the mean - field approximation leads to the potential @xmath264 that is flat at the origin ( figure [ fig : conclusion]a and @xcite ) . the ` benefit ' of the harmonic trapping potential is lost and the physics of the trapped interacting gas is very similar to that of a uniform system . in particular one expects in the thermodynamic limit the appearance of quasi - long range order only , with no true bose - einstein condensate @xcite . the transition between the ideal and the interacting case is explicit in equations ( [ eq : prokofev ] ) and ( [ eq : ncanalytics ] ) , where the limit @xmath262 gives @xmath265 and @xmath266 . in particular ( [ eq : ncanalytics ] ) can be used to separate a ` bec - dominated ' regime where @xmath267 and @xmath268 , and a ` bkt - dominated ' regime , where the contribution of @xmath269 is dominant and @xmath270 . in the latter case , the spatial distribution in the mean - field approximation is a thomas - fermi disk with radius @xmath271 and ( [ eq : ncanalytics ] ) is equivalent ( within a numerical factor ) to the bkt threshold ( [ eq : prokofev ] ) for a uniform gas with density @xmath272 . the rubidium gas studied in @xcite is at the border of the ` bkt - dominated ' regime ( @xmath273 ) , whereas previous experiments performed on quasi-2d gases of sodium atoms @xcite corresponded to @xmath274 , well inside the ` bec - dominated ' regime . finally , we must take into account the finite size of the gas in our discussion . it is known from simulations of 2d spin assemblies that for a finite size system , the average magnetization increases rapidly around the bkt transition @xcite . it is at first sight surprising that this magnetization can be used as a signature of bkt physics , since it would not exist in an infinite system where a genuine bkt transition takes place . however it is relevant for all practical 2d situations : as emphasized in @xcite one would need extremely large systems ( ` bigger than the state of texas ' ) to avoid a significant magnetization even just below the transition point . a similar phenomenon occurs for a finite size bose gas . a few states acquire a large population around the transition point , and this allows for the observation of good contrast interferences between two independent gases . in particular the condensed fraction @xmath275 ( largest eigenvalue of the one - body density matrix ) is expected to grow rapidly at the critical point , and this has been quantitatively confirmed by the qmc calculation of @xcite . to illustrate this point we have schematically plotted in figure [ fig : conclusion]b the expected variations of @xmath275 with the parameters of the problem . for given @xmath182 and @xmath35 , @xmath275 takes significant values for @xmath276 ( continuous line ) . if the strength of the interactions @xmath182 is kept constant , the condensed fraction @xmath275 tends to zero for any finite temperature if the thermodynamic limit is taken ( arrow 1 in figure [ fig : conclusion]b ) . note that the superfluid fraction should tend to a finite value in this limiting procedure . now one can also keep @xmath35 constant and decrease @xmath182 to zero ( arrow 2 in figure [ fig : conclusion]b ) . in this case one expects to recover the ideal gas result @xmath277 for any value of @xmath35 . therefore we are facing here a situation where two limits do not commute : @xmath278 . of course this does not cause any problem in practice since none of these limits is reached . in this sense the phenomenon observed in our interacting , trapped 2d bose gas is hybrid : the transition point is due to bkt physics ( the density of states of the ideal 2d harmonic oscillator does not play a significant role because of the flattening of the potential ) , but thanks to the finite size of the system , some diagnoses of the transition such as the appearance of interferences , take benefit of the emergence of a significant condensed fraction . we are indebted to markus holzmann and werner krauth for numerous discussions , for providing the quantum monte - carlo data shown in figure [ fig : nat_fsd ] , and for pointing out the significant role of the excitation of the @xmath2 motion in our experiments . we also thank yvan castin and pierre clad for helpful discussions . p.k . and s. p. r. acknowledge support from eu ( contract meif - ct-2006 - 025047 ) and from the german academic exchange service ( daad , grant d/06/41156 ) , respectively . this work is supported by rgion le de france ( ifraf ) , cnrs , the french ministry of research , anr ( project gascor , nt05 - 2 - 42103 ) and the eu project scala . laboratoire kastler brossel is a research unit of cole normale suprieure , universit pierre and marie curie and cnrs . we have emphasized in this paper that the measured density profiles differ from those calculated for an ideal gas or within the 2d mean - field theory . the profiles calculated in steady - state in the trap are found much ` peakier ' than the experimental ones . as the experimental profiles were actually measured after a time of flight of @xmath279 ms ( @xmath280 ) , it is important to check that this mismatch between predicted and observed profiles remain valid when the ballistic expansion of the atoms is taken into account . also the atom distributions were measured using an absorption imaging technique , with an imaging beam propagating along the @xmath59 axis . therefore the measurement gave access to the column density @xmath281 , obtained by integrating the total density along @xmath59 . in this appendix we take into account the time - of - flight and the integration along the @xmath59 direction , both for an ideal and for an interacting gas within the 2d mean - field approximation . the spatial distribution @xmath282 at time @xmath283 can be calculated from the phase space distribution @xmath284 at initial time using @xmath285 in the semi - classical approximation the in - trap phase space density is given by @xmath286\ ; -\;1 \right\}^{-1}\ , \ ] ] where @xmath287 , and @xmath288 is obtained by solving ( [ eq : hfford ] ) . the result for the column density can be written as @xmath289 the results for @xmath290 are shown in figure [ fig : tof]a for an ideal gas , and in figure [ fig : tof]b for an interacting gas in the mean field approximation . in the ideal gas case , the initial column density can be calculated analytically : @xmath291 and the column density after time of flight is deduced from the initial value by a simple dilatation @xmath292 in figure [ fig : tof]a , the fugacity is such that the atom number equals the critical number ( [ eq : idealtrapped2 ] ) . in the interacting case of figure [ fig : tof]b , the number of atoms is such that the criterion for superfluidity is met at the center of the trap . in all cases , it is clear that the observed profiles are very different from a gaussian , in clear disagreement with the experimental observation . in the trap ( dashed line ) and after a time of flight @xmath283 such that @xmath293 ( continuous line ) . ( a ) ideal gas case , for an atom number equal to the critical value ( [ eq : idealtrapped2 ] ) . ( b ) mean - field result for @xmath216 . the fugacity is chosen such that the threshold for superfluidity ( [ eq : prokofev ] ) is met at the center of the trap . , title="fig:",width=302 ] in the trap ( dashed line ) and after a time of flight @xmath283 such that @xmath293 ( continuous line ) . ( a ) ideal gas case , for an atom number equal to the critical value ( [ eq : idealtrapped2 ] ) . ( b ) mean - field result for @xmath216 . the fugacity is chosen such that the threshold for superfluidity ( [ eq : prokofev ] ) is met at the center of the trap . , title="fig:",width=302 ] a. grlitz , j. m. vogels , a. e. leanhardt , c. raman , t. l. gustavson , j. r. abo - shaeer , a. p. chikkatur , s. gupta , s. inouye , t. rosenband , and w. ketterle . realization of bose - einstein condensates in lower dimensions . , 87:130402 , 2001 .
we analyze the results of a recent experiment with bosonic rubidium atoms harmonically confined in a quasi - two - dimensional geometry . in this experiment a well defined critical point was identified , which separates the high - temperature normal state characterized by a single component density distribution , and the low - temperature state characterized by a bimodal density distribution and the emergence of high - contrast interference between independent two - dimensional clouds . we first show that this transition can not be explained in terms of conventional bose - einstein condensation of the trapped ideal bose gas . using the local density approximation , we then combine the mean - field ( mf ) hartree - fock theory with the prediction for the berezinskii - kosterlitz - thouless transition in an infinite uniform system . if the gas is treated as a strictly 2d system , the mf predictions for the spatial density profiles significantly deviate from those of a recent quantum monte - carlo ( qmc ) analysis . however when the residual thermal excitation of the strongly confined degree of freedom is taken into account , an excellent agreement is reached between the mf and the qmc approaches . for the interaction strength corresponding to the experiment , we predict a strong correction to the critical atom number with respect to the ideal gas theory ( factor @xmath0 ) . a quantitative agreement between theory and experiment is reached concerning the critical atom number if the predicted density profiles are used for temperature calibration .
the emergence of collective motion in active systems composed of ( self-)propelled entities is a highly complex self - organization phenomenon @xcite . examples include systems on vastly different length scales , ranging from micrometer - sized systems , made of biological constituents @xcite , over millimeter - sized vibrating hard granules @xcite , to large cooperative animal groups @xcite . while all these systems share the capacity to form collective patterns , the precise mechanisms how they are formed and their degree of universality remain elusive . historically , generic agent - based models @xcite constitute the first theoretical approach aiming to understand the minimal ingredients necessary for the emergence of collective motion . in the vicsek model @xcite collective motion is thought to be a consequence of a generic competition between a local tendency of ferromagnetic parallel alignment " and noise . specifically , particle alignment is implemented as an update rule in the spirit of a cellular automaton : each particle aligns parallel to the average of all particles orientations within some defined finite neighborhood . vicsek - like models have been instrumental in exploring the pattern forming capabilities of active systems @xcite . nevertheless , there are also drawbacks of such generic agent - based models . most importantly , they do not account for the physics of the interaction between active constituent particles . in the meantime there have emerged however a range of well characterized experimental model systems including actin and microtubule gliding assays @xcite , and shaken granular particles @xcite , which are amenable to a highly quantitative analysis at the scale of collisions between individual particles . these microscopic features can , in principle , be analyzed in terms of molecular dynamics simulations : while in cellular automata like the vicsek model particles move at constant speed , newtonian dynamics explicitly account for driving forces and dissipation . moreover , instead of update rules there are interaction forces between the constituent particles . the ensuing equations of motion allow for a more realistic description of individual particle trajectories as well as resolving the time and length scales of pairwise collisions . although previous studies of such models show an even richer spatio - temporal dynamics than vicsek - like models @xcite , a thorough analysis of how order builds up from microscopic particle interactions has never been undertaken . here we try to close this gap and ask , focusing on dilute conditions : can we understand the collective dynamics of active particle systems by solely considering _ binary _ interactions between the constituents ? we address this question by a combination of molecular dynamics ( md ) simulations and kinetic theory . specifically , we consider a system of active spherical particles interacting by a short - ranged and repulsive harmonic force . in the absence of interactions , these soft particles move at constant speed determined by a balance between a driving and a dissipative force . from md simulations for this system of _ soft active colloids _ we determine the phase boundary between the isotropic and the polarized state . to connect these numerical studies with a kinetic approach , we employ the md simulations to also analyze binary scattering of particles . thereby one can extract the underlying collision - rule , _ i.e. _ , the mapping of the pre - collisional to the post - collisional velocities , which lends itself as a starting point to set up a boltzmann equation for the one - particle distribution function . following previous approaches in the literature @xcite , the latter can also be used to determine the phase boundary . any mismatch between this phase boundary and the one obtained from the md simulations must be directly related to the assumptions underlying the boltzmann equation . since these assumptions concern the nature of the particle collisions and the ensuing correlations between the positions and orientations of the particles , a quantification of the mismatch will allow a deepened understanding of the ordering process in active systems and highlight the differences to thermodynamic systems . in particular , it will be interesting to see to what degree the molecular chaos assumption of classical boltzmann s theory remains valid for active systems . the structure of this paper is the following : the models for soft active colloids are described in detail in sect . [ sect : models ] . we present the phase boundary between the isotropic and polarized states obtained from the molecular dynamics simulations in section [ sect : multi_particle_simulation ] . section [ sect : binary_scattering ] is devoted to the md study of two particle interactions : the binary scattering study . specifically , the collision geometry is introduced in sect . [ sect : collision_geometry ] and the parameters characterizing the strength of alignment are described in sect . [ sect : alignment_strength ] . then , we discuss the results of the binary scattering study in sect . [ sect : scattering_study ] . section [ sect : collective ] deals with the collective dynamics of one of the models for soft active colloids . specifically , we explain in sect . [ sect : collective_coarse_grained_col_rule ] how the collision rule is implemented into the kinetic description , and set up the corresponding boltzmann equation in sect . [ sect : collective_kinetic_model ] . finally , in sect . [ sect : collective_comparison ] , we derive the phase boundary between the isotropic and polarized state from the boltzmann equation and compare it to the phase boundary obtained from the md simulations . in this section we also show how the kinetic approach has to be modified in order to account for correlations close to the phase transition to collective motion ( termed as _ precursor _ correlations ) . we close by a concise conclusion and outlook in the final sect . [ sect : conclusion_outlook ] . we study _ dynamic models _ for active colloids in two dimensions @xcite in terms of newtonian equations of motion including the following forces : ( i ) an active propelling force @xmath0 [ with @xmath1 : unit vector of the velocity ] capturing the internal propulsion mechanism that in turn is balanced by ( ii ) a dissipative force @xmath2 accounting for the particle s loss of kinetic energy . finally , ( iii ) particles interact by means of a two - body interaction force denoted as @xmath3 , which may be attractive , repulsive , or any combination thereof . the ensuing newtonian equations of motion read @xmath4 where we use units such that the mass of each particle is set to unity . the propelling and dissipative forces are commonly taken as @xcite @xmath5 with exponents @xmath6 and @xmath7 characterizing each force s respective dependence on the particle velocity . their choice must fulfill the condition @xmath8 to ensure a proper definition of a ( stable ) stationary velocity @xmath9 . to accommodate a more complex dependence on @xmath10 , the amplitudes @xmath11 and @xmath12 may in general be functions of the particle s velocity @xcite . the interaction force @xmath3 exerted by particle @xmath13 on particle @xmath14 is a function of the relative position @xmath15 @xcite , with @xmath16 denoting the position of particle @xmath14 . in addition , it may depend on the particles velocities , accounting for inelastic interactions between the particles @xcite . here , we restrict ourselves to short - ranged repulsive interactions between the particles : if two particles @xmath14 and @xmath13 of radius @xmath17 exhibit some finite overlap , @xmath18 , there is a harmonic repelling force @xmath19 acting on particle @xmath14 in the direction of @xmath20 . the coefficient @xmath21 denotes the stiffness of the harmonic spring . in the following , we focus on two specific _ dynamic models _ , where @xmath11 and @xmath12 are constants . in the first model , hereafter referred to as _ model a _ , there is a constant propulsion in the direction of motion , while the dissipation has a linear dependence on the particle s velocity ( @xmath22 , @xmath23 ) . the corresponding equations of motion are then given by @xcite : @xmath24 the second model considered , referred to as _ model b _ , features a driving force that scales linearly with the particle s velocity , and a dissipative force that is cubic in the velocity ( @xmath25 , @xmath26 ) @xcite : @xmath27 for deterministic particle motion governed by eq . in the absence of interactions , the direction of the velocity does not change , while the particle s speed @xmath28 exponentially approaches the stationary value @xmath29 as @xmath30 , with a characteristic relaxation time @xmath31 rescaling the particles spatial coordinate @xmath32 by the particle diameter @xmath33 and time by the relaxation time @xmath34 , we arrive at the following dimensionless equations : @xmath35 where @xmath36 is the dimensionless velocity and @xmath37 is the rescaled penetration depth . the rescaled interaction constant @xmath38 , and the rescaled stationary speed @xmath39 constitute the two key parameters of the model . for the short - ranged harmonic interaction potential , the duration of a particle encounter is of the order of @xmath40 . hence , the dimensionless parameter @xmath41 compares the characteristic relaxation time of the velocity @xmath42 to the duration of a particle interaction , and can be rewritten as @xmath43 therefore , @xmath44 signifies that the repulsive interaction force dominates the collision , whereas for @xmath45 the dissipative force is the major factor governing the dynamics during collisions . in the following we will refer to @xmath41 as the _ interaction parameter_. the dimensionless parameter @xmath46 can be interpreted as follows : consider the length @xmath47 that a particle moving with the stationary speed @xmath48 covers within a time interval equal to the relaxation time @xmath49 . the ratio of this length scale @xmath50 and the particle diameter @xmath51 then is equal to the parameter @xmath46 : @xmath52 since @xmath50 provides an estimate for the relaxation length in the system , _ i.e. _ , the distance traveled by a particle until its velocity is relaxed , small values of @xmath46 signify that particles are moving in a highly damped system , while for high values of @xmath46 damping is weak . therefore , we will hereafter refer to @xmath46 as the dimensionless relaxation length or in short _ relaxation length_. in the absence of interactions , the speed of particles relaxes according to eq . to the stationary speed @xmath53 as @xmath54^{-1/2},\ ] ] with a characteristic relaxation time @xmath55 instead of @xmath56 as for _ model a_. again , rescaling length by particle diameter @xmath32 and time by the relaxation time @xmath57 gives the following dimensionless equations for _ model b _ : @xmath58 where @xmath36 is the dimensionless velocity , and @xmath59 is the rescaled penetration distance . the two parameters of the model are given by @xmath60 and @xmath61 . the interpretation of @xmath41 as the squared ratio of the relaxation to the interaction timescale , and of @xmath46 as the ratio of the relaxation length @xmath62 to the particle diameter is similar to the corresponding parameters for _ model a_. this allows the comparison of the alignment capabilities for the two models arising from the difference in the form of the equations alone . in the following , we will not distinguish between the parameters @xmath41 and @xmath46 of _ model a _ and _ model b _ , respectively . for the study of systems consisting of a large number of particles , the dimensionless , deterministic equations of motion [ eq . or eq . ] are complemented by a stochastic element accounting for noise in the system . since brownian noise is irrelevant in most active systems @xcite , we restrict ourselves to a stochastic element that solely leads to fluctuations in the particles orientations ; see _ e.g. _ refs . specifically , we implement noise as an additive , uncorrelated stochastic force periodically changing the particles orientation with a frequency @xmath63 . denoting the velocity at time @xmath64 obtained from integrating the deterministic model equations as @xmath65 , this direction is rotated by a gaussian - distributed random angle @xmath66 leading to the stochastic velocity direction @xmath67 : @xmath68 the gaussian distribution of the random angle @xmath66 has zero mean and variance @xmath69 . in general , the parameters @xmath63 and @xmath70 together determine the strength of noise in the system . however , for all our later studies the frequency @xmath63 is not a central issue . as a matter of convenience , we choose the time between two stochastic events equal to the discrete time step @xmath71 used for the numerical integration of the equations of motion , _ i.e. _ , @xmath72 this leaves us with @xmath70 to determine the strength of noise in the multi - particle simulations presented in the next section . as a reference point for our subsequent bottom - up analysis of the emergence of collective motion , we performed molecular dynamics simulations of a large number of particles to determine the phase diagram . to this end , the deterministic equations of motion for _ model a _ [ eq . ] , supplemented by the angular noise described in eq . , were integrated numerically . for the parameters of the model , we chose @xmath73 and @xmath74 . : the employed values correspond to a maximal alignment capability of _ model a _ obtained from analyzing the scattering of two particles . ] particles moved in a square box of linear size @xmath75 with periodic boundary conditions . we considered particle numbers in the range @xmath76@xmath77 . as initial configuration particles were placed randomly in the simulation box . overlapping particles were relocated until any particles overlap had vanished . particle velocities were given by randomized directions , with their modulus set equal to the stationary velocity , which is given by @xmath46 in dimensionless units . in order to numerically determine the phase boundary , we computed typically @xmath78 realizations of different initial coordinates and velocity directions for a set of values of the single particle noise @xmath70 and the packing fraction @xmath79 running the simulations for sufficiently long times , we classify a point in @xmath80-@xmath70-parameter space to be macroscopically _ polarized _ if the system s polarization @xmath81 given as the average over all particles velocity directions @xmath82 , exceeded a value of @xmath83 for at least one realization . otherwise , the parameter set is termed as _ isotropic_. the resulting phase diagram is depicted in fig . [ plot : phase_diagram_simulation ] : red dots indicate the values of the control parameters ( @xmath84 ) where a transition to a polar state was observed , while grey squares correspond to isotropic states . close to the phase boundary the dynamics exhibits critical slowing down , _ i.e. _ , the time until the system builds up a polar state grows quickly beyond computationally accessible time - scales . therefore , for small packing fractions @xmath80 corresponding to small critical noise strengths @xmath85 , we prepared multiple realizations of systems in a perfectly polar state ( @xmath86 , but with randomized spatial initial configurations ) , and studied whether this polar state remains stable or disintegrates into an isotropic state ( only investigated for @xmath87\times[0^\circ,0.4^\circ]$ ] ) . those states where all realizations remained polarized throughout the simulation time interval of @xmath77 time units are indicated by blue diamonds in fig . [ plot : phase_diagram_simulation ] . and angular noise strength @xmath70 [ in degrees ] obtained from multi - particle simulations . red dots indicate polarized states while grey squares correspond to control parameter sets where no change in the initial isotropic state was observed . initially perfectly polar states ( @xmath86 ) that remained polarized over the entire simulation time are marked by blue diamonds . ] for packing fractions @xmath88 , the phase diagram shows the shape expected from the generic picture of a competition between aligning interactions and noise @xcite : the critical noise strength @xmath85 , at which a transition from an isotropic to a polarized state occurs , shows a monotonic increase with the value of the packing fraction @xmath80 . additionally , we find hysteresis as observed in the vicsek model @xcite : there is a small range of control parameters ( @xmath84 ) adjacent to the phase boundary where polar states are always stable , but polar ordering from an isotropic state could not be observed . finally , in the regime of large packing fractions polarized states do not develop from initial isotropic states , which has also been found in other agent - based models @xcite . in the previous section we have shown that polar order can emerge from an unordered , homogeneous state for large enough packing fraction @xmath80 and small enough angular noise strength @xmath70 . for dilute conditions ( @xmath89 ) , one expects that binary particle interactions are the dominant process leading to the alignment of particles . therefore , we performed binary scattering studies using the deterministic equations of motion given in section [ sect : models_det ] , and analyzed how the interplay between driving , dissipation and the strength of the repulsive interaction affects the ordering propensity of the active system . specifically , the aim of the scattering study is two - fold : ( i ) first , we systematically compare models with different driving and dissipation forces regarding their capabilities to induce alignment through binary collisions . in particular , we will work out conditions for optimal alignment , and compare the underlying principles leading to parallel alignment . ( ii ) second , we determine the `` collision - rule , '' _ i.e. _ , the mapping of pre - collision velocities and impact parameter onto post - collisional velocities . this mapping then constitutes the foundation for a mesoscopic description of the system in the framework of kinetic theory , detailed in section [ sect : collective ] . due to its microscopic origin the collision rule allows to scrutinize the predictions of kinetic theory for propelled particle systems through direct comparison with the multi - particle simulations presented in section [ sect : multi_particle_simulation ] . due to the short - ranged nature of the repulsive interaction potential [ in eq . and eq . ] one can give a precise definition of the instant when two particles come into contact [ fig . [ pic : coll_geometry](a ) ] . capturing all possible configurations for particle encounters then amounts to defining an appropriate set of parameters describing the geometry at this first moment of contact . denoting the particles positions as @xmath90 and @xmath91 , the inter - particle distance is @xmath92 at contact , and the spatial arrangement of the collision is appropriately described by the unit vector @xmath93 that defines the normal direction @xcite : @xmath94 in combination with the relative velocity @xmath95 , the unit vector @xmath93 completely determines the geometry of the collision at the moment of contact . instead of these vectorial quantities , however , it is more convenient to work with two equivalent scalar parameters . since the relative position of the particles only matters with respect to the direction of their relative velocity , we introduce the angle @xmath7 as a parametrization for the unit vector @xmath93 [ fig . [ pic : coll_geometry](b ) ] , @xmath96 in our dimensionless units where the particle diameter @xmath51 constitutes the unit length , the impact parameter @xmath97 is then defined as @xcite @xmath98 the impact parameter characterizes the type of collision : @xmath99 signifies a head - on collision in the relative frame or symmetric collision in the laboratory frame , whereas @xmath100 corresponds to glancing collisions where particles are barely touching each other . for @xmath101 there is no collision [ fig . [ pic : coll_geometry](b ) ] . finally , assuming that the particles before the encounter move with equal and constant speed , the relative angle [ fig . [ pic : coll_geometry](a ) ] @xmath102 suffices in place of the relative velocity . note that @xmath103 gives an equivalent collision geometry to @xmath104 differing only by an exchange of the particle indices . in summary , for identical particles moving with equal speed the configuration at the moment of contact , the collision geometry , is completely determined by the impact parameter @xmath97 and the relative angle @xmath104 . for the scattering studies we therefore prepare the two particles in an initial state with @xmath105 $ ] and @xmath106 $ ] . [ scale=0.7 , dashedlines/.style = thick , color = gray , densely dashed , lines/.style = thick , color = black , circles/.style = thick , color = black , ] at ( -5,3.2 ) * ( a ) * ; ( 0,0 ) node[coordinate ] ( zero ) + ( 180:4 ) circle ( 1 ) node[coordinate ] ( two ) + ( 140:4 ) circle(1 ) node[coordinate ] ( one ) ; ( zero ) ( two ) ( zero ) ( one ) ; ( two ) + ( 0:1.5 ) node[above ] @xmath107 ; ( one ) + ( -40:1.5 ) node[above ] @xmath108 ; ( zero ) + + ( 140:1.5 ) arc(140:180:1.5 ) ; ( zero ) + ( 160:1 ) node @xmath104 ; [ scale=0.8 , rotate=-20 , dashedlines/.style = thick , color = gray , densely dashed , lines/.style = thick , color = black , circles/.style = thick , color = black ] at ( -3.5,1.5 ) * ( b ) * ; ( 0,0 ) circle ( 1 ) node[coordinate ] ( two ) + ( 130:2 ) circle(1 ) node[coordinate ] ( one ) ; ( two ) + ( 180:-1.5 ) + ( 180:2.5 ) + ( 0,0 ) + ( -90:2 * 0.766 ) node[coordinate ] ( b ) ; ( two ) node[right=3pt ] @xmath109 + ( 130:1.5 ) ; ( one ) node[left , near start ] @xmath110 + ( -90:1.2 ) ; ( b ) + ( 90:0.4 ) arc(90:180:0.4 ) ; ( b ) + ( 135:0.25 ) circle(1pt ) ; ( two ) + + ( 130:0.8 ) arc(130:180:0.8 ) ; ( two ) + ( 155:0.55 ) node @xmath7 ; ( two ) + ( 0,-0.4 ) node[coordinate ] ( startb ) ( b ) + ( 0,-0.4 ) node[coordinate ] ( endb ) ; ( startb ) node[fill = white , below=2pt , pos=0.6 ] @xmath97 ( endb ) ; to characterize the strength of parallel alignment in binary collisions we choose the particles mean polarization , which is hereafter referred to as _ parallel alignment _ @xmath111 with @xmath112 $ ] . for two particles moving exactly in the same direction @xmath113 , whereas for particles moving in opposite directions @xmath114 . the initial value of @xmath115 is fully determined by the relative angle @xmath104 between the particles velocities and given by @xmath116 in order to determine the change in parallel alignment , we define the _ relative alignment _ @xmath117 where @xmath118 denotes the parallel alignment after the collision . @xmath119 depends on the collision geometry @xmath120 as well as the model parameters . obtaining a measure for the overall alignment tendency , without referring to specific values of @xmath104 and @xmath97 , requires integration over all possible collision geometries . as collision events for different collision geometries are in general not equally likely to occur in a given time interval , an appropriate integration weight is required when evaluating the relative alignment for each collision geometry ( @xmath121 ) . a two - dimensional system consisting of ballistically moving uncorrelated particles in an isotropic and homogeneous state , implies that impact parameters @xmath97 are equally likely , and relative angles are distributed according to @xmath122 @xcite . this leads to the following definition of the _ alignment integral _ , which characterizes the overall alignment tendency : @xmath123 the systematic derivation of the weight function in terms of the unit vector @xmath93 [ eq . ] and the relative velocity @xmath110 can be found in appendix [ appendix : alignment_integral ] . alignment integral _ @xmath124 has the following properties : * if every collision geometry @xmath120 results in parallel alignment : @xmath125 @xmath126 @xmath127 . * if every collision geometry @xmath120 results in anti - parallel alignment : @xmath128 @xmath126 @xmath129 . neither case is likely to be encountered in actual active systems since , in general , the dynamics of collisions lead to a variety of post - collisional relative angles deviating from the extremal cases of zero or @xmath130 . for fully elastic collisions of particles with initially equal speeds , the average relative alignment has a negative value , @xmath131 , as obtained by integration of the elastic collision rule ( cf . eq . , @xcite ) . within the framework of binary collisions , @xmath132 may be taken as a heuristic criterion constituting a prerequisite for the emergence of collective motion . in the following , we use @xmath133 to analyse the ordering capabilities for each model as a function of the model s control parameters . in this section we present the results of a binary scattering study conducted by numerically integrating the respective model equations eq . or eq . for two particles and varying collision geometry ( @xmath134 ) to characterize the average alignment behavior for _ model a _ we use the alignment integral @xmath133 defined in eq . . varying the interaction parameter @xmath41 for fixed relaxation length @xmath46 , we find three distinct regimes , which are depicted in fig . [ plot : modela_alignint](a ) : ( i ) a maximum in @xmath133 at around @xmath135 which corresponds to optimal alignment . ( ii ) for @xmath136 , which is equal to the overdamped limit , @xmath133 gradually vanishes , whereas ( iii ) for @xmath137 , @xmath133 approaches the elastic value @xmath138 . interestingly , the relaxation length @xmath46 has only a minor effect on the alignment : for the most part , the position of the maximum of @xmath133 is independent of @xmath46 , and its magnitude varies only weakly with the value of @xmath46 [ fig . [ plot : modela_alignint](a ) ] . the aforementioned elastic limit is expected since for large @xmath41 the impact of driving force and dissipation is negligible for the outcome of encounters [ cf . the maximum of the alignment integral always occurs for values of the interaction parameter @xmath139 and this value is largely independent of the relaxation length @xmath46 [ fig . [ plot : modela_alignint](a ) ] . as @xmath41 is given by the square of the ratio between the relaxation and the interaction timescale [ eq . ] , @xmath139 implies that there is approximately a factor of @xmath140 between the scales on which the driving and dissipation forces and the interaction force operate . from this we infer that achieving optimal alignment during a binary collision requires driving and dissipation as well as interaction forces being of similar importance , without either taking on a largely dominant role over the course of a collision . note that for fixed @xmath46 the accessible range of values for the interaction parameter @xmath41 is restricted . decreasing @xmath41 , the relative influence of both , the driving force as well as the initial velocity , grows in competition to the interaction force . this results in an increase of the maximal penetration distance @xmath141 during a collision and finally leads to unphysical behavior as the two particles are completely passing through each other ( in a central collision ) . for this reason , the curves in fig . [ plot : modela_alignint](a ) terminate each when the maximum overlap would exceed a threshold of @xmath142 . lowering the interaction parameter @xmath41 the alignment integral @xmath143 decreases to zero . this limit is accessible only if the relaxation length @xmath144 as well . therefore , in this overdamped limit both the relaxation time [ eq . ] as well as the relaxation length [ eq . ] become small in relation to the respective scales of the interaction time and particle diameter . to examine the behavior of @xmath143 in the overdamped limit ( @xmath145 ) , we introduce a new dimensionless parameter @xmath146 . it can be interpreted as the ratio of the maximal possible interaction force @xmath147 to the ( constant ) driving force @xmath11 . then , taking the relaxation length @xmath144 is equivalent to increasing the dissipation in the system while at the same time preserving the strength ratio of the forces determining the outcome of collisions by fixing @xmath148 . in this limit , the alignment integral @xmath124 decreases smoothly to zero [ fig . [ plot : modela_alignint](b ) ] . moreover , the variance @xmath149 = \langle { \delta a}^2 \rangle - { \langle \delta a \rangle}^2 $ ] vanishes identically as well [ inset in fig . [ plot : modela_alignint](b ) ] . from this we conclude that @xmath150 , _ i.e. _ , there is no change in the relative angle , not only on average but independently for any collision geometry @xmath151 in the overdamped limit @xmath145 . for _ model a _ we found that the behavior of the alignment integral @xmath152 is comprised of three characteristic regimes . in the following we ask whether _ model b _ exhibits the same characteristics . as fig . [ plot : ab_comparison_alignint ] shows , the functional form of @xmath133 as a function of the interaction parameter @xmath41 is indeed preserved for _ model b_. for @xmath137 , the value obtained from the elastic collision rule is recovered , @xmath153 . in the same way there is a convergence to a vanishing change in alignment , @xmath154 , in the overdamped case ( @xmath155 ) . furthermore , the maximum of the alignment integral @xmath133 indicating optimal alignment occurs for values of @xmath139 , with only little variation with the relaxation length @xmath46 , identical to the behavior for _ model a_. we therefore conclude that it is indeed the relative influence of the driving / dissipation forces and the interaction force that determines the aligning capabilities inherent in the model equations : good alignment can only arise if neither force is particularly dominant during collisions , indicated by an intermediate value of @xmath41 in fig . [ plot : ab_comparison_alignint ] . the only difference between the models is found in the magnitude of the alignment integral @xmath133 at the maximum [ fig . [ plot : ab_comparison_alignint ] ] : the amplitudes of the maxima for _ model b _ are lower for all values of @xmath46 . understanding the reason for this difference requires investigation of the change in alignment @xmath156 as a function of the collision geometry @xmath151 . this analysis is detailed in the next section and will also reveal the underlying mechanism responsible for the maximum in @xmath133 . for _ model a _ ( solid lines and filled symbols ) and _ model b _ ( dotted lines and open symbols ) , showing the same generic shape for both models . moreover , both models have a similar behavior in the approach to the elastic limit ( @xmath137 ) as well as in the overdamped case @xmath154 ( @xmath157 ) . additionally , the position of the maximum is approximately the same for both models , however _ model b _ exhibits generally lower values for @xmath133 . ] so far we have studied the binary scattering behavior on average by means of the alignment integral @xmath133 , which represents a mean value over all collision geometries @xmath151 . now , we take a closer look and study the change in alignment @xmath156 as a function of the specific collision geometry given by the relative angle @xmath104 , and the impact parameter @xmath97 . choosing a fixed value for the relaxation length @xmath158 , fig . [ plot : modelab_alignmaps ] depicts @xmath156 for _ model a _ ( a - d ) and _ model b _ ( e - h ) for values of the interaction parameter @xmath159 , respectively . the largest value of @xmath41 corresponds to mostly elastic collisions , and the smallest value is close to the overdamped limit . recall that the maximum in the alignment integral @xmath133 [ fig . [ plot : ab_comparison_alignint ] ] is reached for @xmath160 . the change in alignment @xmath156 for _ model a _ and _ model b _ features strong similarities ( see also a video in the supplemental material @xcite ) : for a large interaction parameter ( @xmath161 ) , @xmath156 exhibits balloon - like regions with rather large values , located at intermediate values of the impact parameter @xmath162 . at a relative angle @xmath163 , corresponding to orthogonal pre - collisional velocities , there is a discontinuous change from positive @xmath119 for smaller @xmath104 to negative values for larger @xmath104 [ figs . [ plot : modelab_alignmaps](d ) and [ plot : modelab_alignmaps](h ) ] . this is the result of a flip in the post - collisional velocity of one of the colliding particles ( see video in supplemental material @xcite , and @xmath164 , particles velocities are aligned in elastic collisions , with one particle s velocity being increased and the other s becoming small . the increase in the relative velocity @xmath165 when going to larger relative angles @xmath166 results in the latter particle s velocity being completely flipped . the former particle s scattering behavior changes in a continuous manner , resulting in the discontinuous change in @xmath119 observed in figs . [ plot : modelab_alignmaps](d ) and [ plot : modelab_alignmaps](h ) . ] ) . as discussed further below these results for @xmath161 are almost identical to those found for perfectly elastic collisions ; see fig . [ plot : collrule_alignmaps](a ) . for small values of @xmath167 [ plot : modelab_alignmaps](a ) and [ plot : modelab_alignmaps](e ) ] , all features present at larger values of @xmath41 essentially vanish , and @xmath168 for all collision geometries . this result has already been implied by the analysis of the alignment integral @xmath152 and its variance , which both approach zero in the overdamped limit [ fig . [ plot : modela_alignint ] ] . for intermediate values of the interaction parameter @xmath41 , there is a triangular - shaped region of positive @xmath119 across the whole range of relative angles @xmath169 $ ] [ see figs . [ plot : modelab_alignmaps](b ) and [ plot : modelab_alignmaps](f ) for @xmath160 , and to a lesser extent figs . [ plot : modelab_alignmaps](c ) and [ plot : modelab_alignmaps](g ) for @xmath170 . at two edges of this triangular region , there are pronounced peaks of positive @xmath119 , which for @xmath160 are the most prominent features in the graph , and therefore provide the dominant contribution to the alignment integral [ fig . [ plot : ab_comparison_alignint ] ] . ( image ) at ( 0,0 ) , @xmath171 , called scattering angle , as a function of the collision geometry @xmath172 . parameters : _ model a _ , @xmath158 , @xmath160 , corresponding to a maximum in @xmath133 . as both particles are identical spheres , the scattering behavior of the particle @xmath173 can be read off at the point @xmath174 in the plot . * ( b ) * schematic illustration of the dominant alignment principle termed as _ alignment of the first _ " . during the collision @xmath175 is aligned to @xmath176 with only marginal change of @xmath176 . , title="fig : " ] ; at ( 0.07,0.97 ) * ( a ) * ; at ( 1.05,0.97 ) * ( b ) * ; ( sketch ) at ( 1.18,0.7 ) ; ( sketch ) + + ( 0,0 ) node[coordinate ] ( zero ) + ( 180:2.5 ) circle ( 1 ) node[coordinate ] ( two ) + ( 100:1.9 ) circle(1 ) node[coordinate ] ( one ) ; ( zero ) ( two ) ( zero ) ( one ) ; ( two ) + ( 0:1.5 ) node[left ] @xmath107 ; ( one ) + ( -80:1.5 ) node[below ] @xmath108 ; ( zero ) + + ( 1.5,0.5 ) + ( 2,0 ) ; ( zero ) + + ( 5,0 ) node[coordinate ] ( zero ) + ( 180:0 ) circle ( 1 ) node[coordinate ] ( two ) + ( 35:2 ) circle(1 ) node[coordinate ] ( one ) ; ( two ) + ( -5:1.5 ) node[left ] @xmath177 ; ( one ) + ( -5:1.5 ) node[right ] @xmath178 ; understanding the detailed scattering behavior for parameters corresponding to the edges of the triangular - shaped region is hence vital for determining the underlying principle of alignment . let @xmath179 be the angle of the pre - collisional velocity for particle @xmath14 with respect to some reference axis , and @xmath180 the corresponding angle after the collision . then , the scattering angle @xmath181 describes the change in the particle s direction of motion as the result of a collision with another particle . figure [ plot : oneparticlescattering_map](a ) shows the scattering angle @xmath171 for particle @xmath182 as a function of the collision geometry ( @xmath121 ) for _ model a _ and parameters @xmath158 , @xmath160 . the scattering angle @xmath183 for particle @xmath173 in the same collision can be read off at the point ( @xmath184 ) in fig . [ plot : oneparticlescattering_map](a ) ; this can be seen by considering an exchange of indices in the definitions of the relative angle @xmath104 [ eq . ] and the impact parameter @xmath97 [ eq . ] . the scattering angle @xmath185 in fig . [ plot : oneparticlescattering_map](a ) exhibits the same kind of triangular structure as found for @xmath119 in fig . [ plot : modelab_alignmaps](c ) . for collision geometries at the edges of this triangular structure , one particle hardly changes its direction of motion [ white region in fig . [ plot : oneparticlescattering_map](a ) ] , while the orientation of the other particle changes by an angle close to the relative angle @xmath104 . this results in an alignment of the latter particle s velocity to that of its collision partner ( see supplementary material @xcite for a video ) . closer examination of the collision geometry reveals that it is the `` first '' particle s velocity that is aligned . here , `` first '' means that before the collision it is closer to the _ center of collision _ [ see fig . [ pic : coll_geometry](a ) ] , defined by the intersection point of the pre - collisional orientations [ particle @xmath173 for @xmath186 , see the sketch in fig . [ plot : oneparticlescattering_map](b ) ] . the orientation of the other particle [ particle @xmath182 in fig . [ plot : oneparticlescattering_map](b ) ] hardly changes because the repulsive force mostly affects the magnitude of its velocity , which is counteracted to some extent by the driving force . at the same time , the `` first '' particle s velocity is rotated quickly until both velocities become aligned . therefore , we term this mechanism `` _ _ alignment of the first__. '' interestingly , all collision geometries corresponding to the edges of the triangular structure lead to `` _ _ alignment of the first _ _ '' ( refer to the supplemental material @xcite for a video ) . following the same line of reasoning for _ model b _ , we find the identical principle of alignment . we conclude that the `` _ _ alignment of the first _ _ '' is the dominant mechanism giving rise to parallel alignment during binary collisions of soft active colloids . as a function of the collision geometry given by the relative angle @xmath104 and the impact parameter @xmath97 obtained by use of the inelastic collision rule [ eq . ] . the normal coefficient of restitution @xmath187 varies between a fully elastic collision [ @xmath188 , * ( a ) * ] and a fully inelastic collision [ @xmath189 , * ( d ) * ] . blue / black ( dark grey ) regions ( marked with a minus sign ) denote collisions with post - collisional alignment @xmath190 , while for red / yellow ( unmarked , lighter grey ) regions @xmath191 ( see mapping of @xmath192 to corresponding shading on the right ) . for a video please refer to the supplementary material @xcite . ] finally , we contrast the scattering behavior of particles governed by the equations of motion of _ model a _ or _ model b _ ( for all values of @xmath41 ) to that of classical inelastic collisions described by the inelastic collision rule with a fixed normal coefficient of restitution @xmath187 . denoting the velocities of two particles after a collision as @xmath193 and @xmath194 , the classical inelastic collision rule is given by @xcite : @xmath195 the normal coefficient of restitution @xmath187 determines the change in the normal component of the relative velocity @xmath110 : @xmath196 . the relative alignment @xmath119 for a collision with collision geometry ( @xmath121 ) is shown in fig . [ plot : collrule_alignmaps ] for values of @xmath187 ranging from fully elastic collisions ( @xmath188 ) , to collision where the normal component of the relative velocity is damped completely ( @xmath189 ) . similar to _ model a _ and _ model b _ [ figs . [ plot : modelab_alignmaps](b ) and [ plot : modelab_alignmaps](f ) ] we find for very small @xmath187 a triangle - like region of positive @xmath119 [ fig . [ plot : collrule_alignmaps](d ) ] , however the edges of the triangle are far less pronounced . for inelastic collisions , most of the contribution to the alignment integral @xmath152 comes from collision geometries corresponding to the inner region of the triangle , _ i.e. _ , small @xmath197 . for these small impact parameters the relative velocity is directed along the normal direction and vanishes in a fully inelastic collision . in contrast , _ model a _ and _ model b _ exhibit only small @xmath119 for small impact parameters [ figs . [ plot : modelab_alignmaps](a - d ) and [ plot : modelab_alignmaps](e - h ) ] . comparing inelastic collisions with _ model a _ at fixed @xmath152 corresponding to _ s maximum , it turns out that inelastic collisions still exhibit the balloon - like structure in @xmath156 , while _ model a _ already developed the triangle ( see video in the supplemental material @xcite ) . taken together , we conclude that _ model a _ ( and _ model b _ ) follow an alignment mechanism that is : alignment in the inelastic collision rule comes from damping of the normal component of the relative velocity , whereas for the dynamic models it is the presence of the propelling force that keeps one particle `` stuck '' against the repulsive force and enables the alignment of the other particle s velocity according to the _ alignment of the first _ " mechanism . in summary , the scattering study provided the means to systematically study the alignment properties of soft active colloids . we found that the nature of collisions is determined by two dimensionless parameters : the interaction parameter @xmath41 , which determines the influence of the driving force and dissipation relative to the interaction force during collisions , and the relaxation length @xmath46 , which gives the typical relaxation length relative to the particle size . we observed that each model comprises two distinct limits determined by these dimensionless parameters : the _ elastic limit _ ( @xmath198 ) , where particles obey an elastic collision rule , and the _ overdamped limit _ ( @xmath199 ) , where there is no change to relative orientations . further , we identified the model parameters for which parallel alignment is maximal , and found that this maximum occurs for all models at the same intermediate value of the interaction parameter @xmath41 , largely independent of @xmath46 . additionally , parallel alignment for all considered models followed the same principle , termed _ alignment of the first _ " , regardless of the type of driving or dissipating force . the principle states that those collision geometries that contribute dominantly to the particles parallel alignment exhibit the following typical characteristics in the particles dynamics : the first incoming particle [ with respect to the _ center of collision _ , fig . [ pic : coll_geometry](a ) ] is aligned parallel to the second of the colliding particles . identification of the universal fingerprints of colliding active colloids , as well as the conditions for maximal alignment , provides the starting point for a study of the collective properties of active colloids . on the basis of the `` collision - rule '' depicted in fig . [ plot : oneparticlescattering_map](a ) , we derive a mesoscopic description using the framework of kinetic theory for propelled particle systems . in order to connect the system s collective behavior studied in section [ sect : multi_particle_simulation ] with the results of the binary scattering study , we use kinetic theory for propelled particles moving at constant speed @xcite . kinetic theory aims to provide a description for the time evolution of the one - particle distribution function @xmath200 , which is a function of the spatial coordinates @xmath201 , the orientation of the velocity @xmath202 and time @xmath64 . it is conceptually restricted to binary interactions between the constituent particles , limiting its range of validity to dilute conditions ( packing fraction @xmath89 ) . the binary interactions are described by collision integrals with each kernel involving a measure for the rate of collisions , known as _ boltzmann collision cylinder _ , as well as a `` collision rule . '' the latter constitutes a mapping between the pre - collisional angles @xmath203 and @xmath204 and the post - collisional orientations of the two colliding particles . the corresponding distribution function required to compute the rate of binary collisions is the two - particle density @xmath205 . to obtain a closed equation for the time evolution of the one - particle density @xmath200 , called the _ boltzmann equation _ , the _ assumption of molecular chaos _ is commonly made @xcite , _ i.e. _ , one assumes that correlations in space and orientation are completely absent such that @xmath206 this constitutes a rather strong assumption concerning the system s dynamics and it is not clear to which degree this assumption holds for active systems , in particular at the onset of collective motion . for the following analysis we restrict ourselves to _ model a _ because , as we have shown in the last section , _ model a _ and _ model b _ are equivalent with respect to their qualitative alignment principle . moreover , we specify a certain parameter set , namely @xmath207 , was chosen to be close to the maximum in @xmath133 . the value of @xmath46 was chosen out of consideration for computational efficiency . ] which corresponds to the maximum of the alignment parameter @xmath133 [ see fig . [ plot : modela_alignint](a ) ] , and is equal to the parameter set used for the multi - particle simulations described in section [ sect : multi_particle_simulation ] . we expect that optimal alignment in a binary collision also optimizes the capability of a multi - particle system to develop a macroscopic polarized state . moreover , if polar order develops the critical packing fraction should be lowest for optimal alignment . this in turn improves the validity of a boltzmann description as the regimes of large packing fractions are expected to be captured insufficiently by this kinetic approach . the collision rule required to set up the boltzmann equation maps the pre - collisional orientations given by the angles @xmath203 and @xmath204 on the post - collisional orientations , denoted as @xmath208 and @xmath209 . denoting the angular change for particle @xmath210 by @xmath211 , the collision rule has the following general form @xmath212 in a collision with given relative pre - collisional angle @xmath104 , a scattering angle @xmath211 occurs with probability @xmath213 , where @xmath214 since @xmath215 is computed from @xmath216 ( see fig . [ plot : oneparticlescattering_map](a),(a ) depicts the scattering angle for a different value of @xmath46 . ] ) by integrating over the impact parameter @xmath97 , we have now turned from a deterministic description to a probabilistic treatment of the collision process . as shown in fig . [ plot : scatdist_charfunct](a ) , @xmath217 has a pronounced maximum at @xmath218 with @xmath219 for most of the range of relative angles @xmath220 $ ] . this maximum is close to specular reflection ( @xmath221 ) , yet is skewed towards a slightly smaller post - collisional relative angle . however , the maximum is far removed from a half - angle alignment rule where @xmath222 with @xmath223 @xcite , leading to @xmath224 [ see dotted line in fig . [ plot : scatdist_charfunct](a ) ] . overall , the distribution @xmath215 indicates that our model exhibits a bias towards alignment , even though it is rather weak . note that assigning indices @xmath173 and @xmath182 to the individual particles in a collision is a matter of convention yet affects the sign of @xmath104 . exchange symmetry between the identical particles then enforces that @xmath225 . additionally , consider collisions with relative angles @xmath104 and @xmath103 , respectively , each seen from the point of view of a specific particle , say particle @xmath182 . since these two pre - collisional states exhibit a mirror symmetry with respect to the velocity direction of particle @xmath182 , the outcome of the collisions for particle @xmath182 differs only by the sign of its scattering angle , _ i.e. _ , @xmath226 . the same argument applies for particle @xmath173 . taken together , we have : @xmath227 _ i.e. _ , given a collision with a relative angle @xmath104 , the respective distributions of the two particles are related by a change in the sign of the argument . later , in the analysis of the boltzmann equation in fourier space the characteristic functions of the distributions [ eq . , fig . [ plot : scatdist_charfunct](a ) ] will be required , which are defined as @xmath228 due to the symmetry between the distributions for the two particles [ eq . ] the corresponding characteristic functions are complex conjugates of each other : @xmath229 where @xmath230 denotes complex conjugation . the real and imaginary parts of @xmath231 are depicted in fig . [ plot : scatdist_charfunct](b ) for @xmath232 . instead of using the full characteristic function for the boltzmann equation , it is tempting to account solely for the average and the deviation computed for the distributions @xmath215 defined in eq . ( _ gaussian approximation _ ) . however , such a procedure fails because it misrepresents the actual scattering behavior by underestimating the impact of scattering events with large relative angles , and thus masks the true reason for the emergence of collective motion ( see appendix [ appendix : gaussian_approximation ] for details ) . for the scattering angle @xmath233 of particle @xmath182 as a function of the relative angle @xmath104 obtained by a reduction of variables with respect to the impact parameter @xmath97 [ see eq . ] . the maximum of the distribution deviates slightly from the case of specular reflection where particles exchange their orientations ( solid grey line ) . the dotted line marks the scattering angle of a half - angle alignment rule @xcite . * ( b ) * _ real and imaginary parts of the characteristic function _ @xmath231 of the distribution @xmath234 for @xmath232 as a function of the relative angle @xmath104 . parameters : _ model a _ with @xmath73 and @xmath160 . ] following refs . @xcite , the time evolution of the one - particle distribution function @xmath235 is determined by a boltzmann equation of the form [ eq : boltzmannequation_all ] @xmath236+\mathcal{c}[f^{(2 ) } ] .\ ] ] the second term on the l.h.s . in eq . is the streaming term accounting for movement of particles with the velocity @xmath237 , where @xmath238 . the speed @xmath239 is assumed to be a constant is small , according to its definition in eq . , particle velocities relax quickly to the stationary speed @xmath239 over a distance small compared to the particle diameter . thus , on the coarse - grained scale of a kinetic model a constant particle speed is a good approximation . ] and taken equal to the stationary speed @xmath48 , with @xmath11 and @xmath12 denoting the amplitudes of the driving and friction forces in the microscopic model , respectively [ eq . ] . the term @xmath240 $ ] describes single particle , angular fluctuations ( self - diffusion events ) occurring at a rate @xmath63 and reads @xcite @xmath241 = & - \lambda f({\mathbf{r}},\theta , t ) + \lambda \int_{-\pi}^{\pi } \text{d}\theta^\prime \ , f({\mathbf{r}},\theta^\prime , t ) \\ & \times \int_{-\infty}^{\infty } \text{d}\eta \ , p_0(\eta ) \ , \overline{\delta}(\theta^\prime+\eta-\theta ) , \end{split}\end{aligned}\ ] ] where the fluctuations are assumed to be distributed according to a gaussian @xmath242 with a standard deviation @xmath70 . the periodicity of angles is accounted for by a sum of @xmath243-functions : @xmath244 . the parameters @xmath63 and @xmath70 determine the strength of angular noise in the system . note that these parameters have already been introduced in the multi - particle md simulations detailed in section [ sect : multi_particle_simulation ] as a rate within the framework of our boltzmann approach . however , the periodic stochastic process used in our multi - particle simulations constitutes one of the possible choices . ] . the second term on the r.h.s . in eq . is the collision integral @xmath245 $ ] , which depends on the two - particle density @xmath246 and captures the effect of binary collisions . it can be split into a _ loss _ ( @xmath247 ) and a _ gain _ ( @xmath248 ) contribution : @xmath245=\mathcal{c}^{-}[f^{(2 ) } ] + \mathcal{c}^{+}[f^{(2 ) } ] $ ] . the respective contributions capture the scattering of particles out of or into an angle interval @xmath249 $ ] and read @xcite : @xmath250 = - \int_{-\pi}^{\pi } \text{d}\theta^\prime \ , \gamma(\theta^\prime , \theta ) f^{(2)}({\mathbf{r}},\theta , \theta^\prime , t ) , \\ \label{eq : collision_integral_gain } & \mathcal{c}^+[f^{(2 ) } ] = \int_{-\pi}^{\pi } \text{d}\theta_1 \int_{-\pi}^{\pi } \text{d}\theta_2 \ , \gamma(\theta_1 , \theta_2 ) f^{(2)}({\mathbf{r}},\theta_1,\theta_2,t ) \nonumber \\ & \quad \times \frac{1}{2 } \sum_{j=1}^{2 } \int_{-\infty}^{\infty } \text{d}\eta_j \ , p_j(\eta_j| \theta_{12 } ) \ , \overline{\delta}(\theta_j+\eta_j-\theta ) .\end{aligned}\ ] ] in the _ gain _ contribution @xmath251 $ ] , each of the the two terms ( @xmath252 ) accounts for the scattering of one of the particles in a binary collision using the respective distribution @xmath253 or @xmath254 [ eq . ] . the factor @xmath255 is required to avoid counting collisions twice . the function @xmath256 describes the rate of collisions for a given pre - collisional state determined by the distribution of particles orientations . the functional form of @xmath256 can be argued geometrically : consider a collision between two particles with orientations @xmath203 and @xmath204 . given short - ranged repulsive interactions , two particles collide if their relative distance becomes less than the particles diameter @xmath51 . changing into the reference frame of _ e.g. _ particle @xmath182 , the velocity of particle @xmath173 is given by the relative velocity @xmath257}$ ] . a collision between the two particles occurs within the time interval [ @xmath258 if particle @xmath173 can be found in a rectangle of length @xmath259 and width @xmath260 . back in the laboratory frame , this rectangle deforms into a parallelogram retaining its surface area given by @xmath261 @xcite . this function is commonly referred to as _ boltzmann collision cylinder_. in combination with the two - particle density @xmath246 it determines the rate of collisions in the pre - collisional state for spherical particles moving ballistically and with constant speed in two dimensions . the function @xmath262 only depends on the relative angle @xmath104 , and can be written as @xmath263 in other systems , the dependence on the relative orientation @xmath104 may be significantly different , like in the case of ballistically moving rod - shaped particles @xcite where we have @xmath264 ; here @xmath265 and @xmath51 are the lengths of the rods long and short axis . in a system of highly diffusive particles like microtubules transported by molecular motors , the dependence on the relative angle may even disappear altogether such that all collisions occur at a constant rate , _ i.e. _ , @xmath266 @xcite . finally , in order to turn eq . into a closed equation for the time evolution of the one - particle density @xmath267 , an expression for the two - particle density @xmath246 in terms of @xmath267 has to be postulated . in a monoatomic gas , elastic collisions prohibit on average a build - up of inter - particle correlations over time , thereby supporting the validity of the _ molecular chaos assumption _ [ eq . ] . in contrast , in a system consisting of actively propelled constituents , collisions quite generally result in orientational correlations as detailed in section [ sect : scattering_study ] , casting doubt on the validity of the molecular chaos assumption . to account for orientational correlations , we use a modified closure relations for the two - particle density @xmath268 where the function @xmath269 measures the magnitude of these correlations as a function of the relative angle @xmath104 solely depends on the relative angle @xmath104 . ] . the set of equations represents a generalized kinetic theory for propelled particles moving with constant speed , which is extended compared to refs . @xcite regarding the following two aspects : the collision rule is quantitatively determined by the results of the microscopic scattering study , and eq . accounts for angular correlations among the active particles . in the following we scrutinize whether these two modifications allow to quantitatively predict the phase boundary obtained from multi - particle md simulations [ see fig . [ plot : phase_diagram_simulation ] ] . to this end , the generalized boltzmann equation [ eq . ] is analyzed in terms of fourier modes , @xmath270 . projecting the resulting equation onto the @xmath271-th fourier mode @xmath272 yields : @xmath273=\\ & \quad -\lambda \big ( 1-\exp{(-k^2\sigma_0 ^ 2/2)}\big ) { \hat{f}}_k \\ & \quad -\frac{4 g d v_0}{\pi } \sum_{n=-\infty}^{\infty}{\hat{f}}_n{\hat{f}}_{k - n } \big[2\mathcal{i}_n\left[{\chi}\right ] - \mathcal{j}_{n , k}\left[{\chi}\right]\big ] , \end{split}\ ] ] where the dependence of @xmath274 on @xmath201 and @xmath64 was omitted for brevity . the coefficients in the collision term , @xmath275 and @xmath276 , depend on the function @xmath269 as an additional integration weight and are given by @xmath277 & = \int_{-\pi}^{\pi } \text{d}\theta_{12 } \ , { \chi}(\theta_{12 } ) \ , \cos{(n\theta_{12 } ) } \ , \left|\sin\left({\theta_{12}}/{2}\right)\right| , \\ \label{eq : coeff_j } \mathcal{j}_{n , k}\left[{\chi}\right ] & = \mathcal{j}^{(1)}_{n , k}\left[{\chi}\right ] + \mathcal{j}^{(2)}_{n , k}\left[{\chi}\right ] .\end{aligned}\ ] ] the two contributions to the coefficient @xmath276 result from the two terms in the _ gain _ contribution @xmath278 of the collision integral eq . . these two terms account for the change in the orientation resulting from a collision for particle @xmath173 and particle @xmath182 , respectively , as denoted by the upper index in brackets . we find @xmath279 = \int_{-\pi}^{\pi } \text{d}\theta_{12 } \ , { \chi}(\theta_{12 } ) \ , \text{e}^{-in\theta_{12 } } g_2(k|\theta_{12 } ) \ , \left|\sin\left({\theta_{12}}/{2}\right)\right| .\ ] ] and @xmath280 , where we have used that the characteristic functions @xmath231 and @xmath281 are related by complex conjugation [ eq . ] . ( image ) at ( 0,0 ) ; ( image_inset ) at ( 0.20,0.43 ) ; ( image_inset.south west ) rectangle ( image_inset.north east ) ; using the boltzmann equation in fourier space [ eq . ] as a starting point , one can determine the onset of polar order in a homogeneous system as follows . the first two fourier components determine the ( hydrodynamic ) particle density @xmath282 and the momentum density @xmath283 , where @xmath284 denotes the hydrodynamic velocity . since @xmath285 is conserved and the momentum field @xmath286 plays the role of the broken symmetry variable , both fields are slow hydrodynamic variables . near the onset of the instability of the isotropic state , @xmath287 , the hydrodynamic velocity @xmath288 is small compared to the microscopic driving velocity @xmath239 . since @xmath289\ll1 $ ] , we are thus able to truncate eq . by setting @xmath290 for all @xmath291 @xcite . since we are only interested in the location of the phase boundary marking the transition from the homogeneous isotropic to the polarized state , we neglect all spatial gradients @xmath292 , @xmath293 and @xmath294 , of the linear term . ] and find the following set of equations : [ eq : hydrodynamicequations ] @xmath295 the kinetic coefficient @xmath6 determines the linear stability of the initially isotropic state : for @xmath296 polar order may develop , whereas @xmath297 signals that the isotropic state is linearly stable . we find @xmath298 \frac{4 v_0}{\pi^2d } \phi , \ ] ] where the packing fraction @xmath299 and @xmath300 $ ] reads in terms of the scattering coefficients defined in eq . and eq . : @xmath301 = \mathcal{j}_{0,1}\left[{\chi}\right ] + \mathcal{j}_{1,1}\left[{\chi}\right ] - 2\mathcal{i}_0\left[{\chi}\right ] - 2\mathcal{i}_1\left[{\chi}\right ] .\ ] ] the function @xmath302 , which describes angular correlations , determines the sign and magnitude of @xmath300 $ ] , and is thereby a main determinant of the phase boundary between the isotropic and polarized states . note that for @xmath300<0 $ ] , the isotropic state is linearly stable for all values of the control parameters @xmath303 . the condition @xmath304 determines the phase boundary as a function of the control parameters , _ i.e. _ , single particle noise strength @xmath70 and packing fraction @xmath80 . solving for the critical single particle noise strength @xmath305 as a function of the packing fraction , we find : @xmath306=\bigg [ { -2 } \ln \bigg(1 - \overline{\nu}\left[{\chi}\right ] \frac{4 v_0}{\pi^2d\lambda}\phi\bigg)\bigg]^{1/2}.\ ] ] above this threshold noise strength @xmath307 the isotropic state is stable , whereas for @xmath308 a macroscopic polarized state can develop . for the specific parameters used in the md simulations in section [ sect : multi_particle_simulation ] one has @xmath309 . at this point , we have to specify the function @xmath269 in eq . to compute the phase boundary , and compare it to the one obtained from multi - particle md simulations [ see fig . [ plot : phase_diagram_simulation ] and section [ sect : multi_particle_simulation ] for corresponding discussion ] . let us first calculate the phase boundary by assuming that the initial states at the onset of collective motion are devoid of angular correlations , _ i.e. _ , the _ assumption of molecular chaos _ is fulfilled with @xmath310 in eq . . in this case we find that @xmath311<0 $ ] , implying that @xmath297 for all control parameters @xmath80 and @xmath70 [ see eq . ] . hence , one would conclude that the system s isotropic state remains stable for arbitrary values of the control parameters , which is obviously at odds with the phase diagram obtained from multi - particle simulations [ fig . [ plot : phase_diagram_simulation ] ] . this clearly indicates that the state of the system preceding a transition to a polarized state can not be free of angular correlations . to further scrutinize this finding , we ran multi - particle md simulations starting from an initially uncorrelated , homogeneous and isotropic state and studied @xmath312 as a function of time . for this purpose we selected a set of control parameters ( @xmath80,@xmath70 ) sufficiently close to the phase boundary obtained from md simulations for which the system remains isotropic . the function @xmath312 was computed by recording the relative angles @xmath104 of all collisions occurring in the simulation box of area @xmath313 within a sampling time interval @xmath314 $ ] . the length of the sampling time interval @xmath315 determines the number of recorded collisions , yet has to be chosen small enough to properly resolve the time evolution of @xmath312 [ values of @xmath315 are given in the caption of fig . [ plot : relativeangledistribution ] ] . noting that @xmath316 $ ] , with @xmath317 $ ] given in eq . , is equal to the total rate of binary collisions in a volume @xmath318 , the collision frequency @xmath319 as a function of the relative angle @xmath104 can be written as @xmath320 the above relation connects @xmath312 with the collision frequency @xmath319 , which can be measured in the multi - particle simulations . in our subsequent numerical studies we only consider states that remain approximately homogeneous and isotropic in time . therefore , we assume @xmath321 , which allows to considerably simplify the above equation to : @xmath322 the initial configuration in the md simulations was a homogenous and isotropic state . to this end , the initial positions of the particles and the orientations of their velocities were chosen randomly , which was followed by a relocation of overlapping particles until excluded volume had been enforced for all particles . in this initial state , angular correlations are absent and therefore @xmath323 . the assumption of molecular chaos implies @xmath310 , however the initial value in our system is expected to be larger . this is due to the finite size of our active spheres , which reduces the amount of free - volume in the system and in turn gives rise to spatial correlations between the particles . in kinetic theory for hard granular gases this is accounted for by the so called enskog factor @xmath324 @xcite . it depends on the packing fraction @xmath80 , and is given in two dimensions by the following approximation @xcite : @xmath325 measured in our system of active particles for a control parameter set ( @xmath80,@xmath70 ) where the system remains isotropic , fig . [ plot : relativeangledistribution](a ) depicts the time evolution of @xmath312 . the initial distribution of @xmath269 [ fig . [ plot : relativeangledistribution](a ) , black curve ] is indeed shifted to a numerical value slightly larger than @xmath173 , reflecting the presence of spatial correlations . however , for the packing fraction @xmath326 used in fig . [ plot : relativeangledistribution ] , the enskog factor [ eq . ] is @xmath327 and therefore slightly overestimates the actual increase of @xmath269 found in the md simulations , . we attribute this small discrepancy to the fact that our active colloids are soft , consistently leading to a smaller amount of decreased free - volume , and thereby a lower value of @xmath328 . as time progresses , we find that @xmath312 evolves to a distribution favoring smaller relative angles [ fig . [ plot : relativeangledistribution](a ) ] , while the system remains isotropic and uniform . the latter is reflected in a flat angular probability distribution @xmath329 [ fig . [ plot : relativeangledistribution](b ) ] , and a polarization @xmath330 ( fig . [ plot : relativeangledistribution](c ) , dark grey curve ; see also video in supplemental material @xcite ) . the deviation from a uniform @xmath269 clearly indicates that angular correlations in the system develop as the system approaches its stationary state . concomitantly , the coefficient @xmath331 shows ( i ) a rapid increase from an initial negative value @xmath332 ( corresponding to @xmath333 ) to ( ii ) a prolonged plateau at a positive value @xmath334 [ fig . [ plot : relativeangledistribution](d ) , dark grey curve ] . this sign change in @xmath331 is triggered by the emerging angular correlations and allows the generalized boltzmann approach to predict an ordering transition . a quantitatively similar behavior of @xmath331 as a function of time can be found for control parameter sets ( @xmath80,@xmath70 ) close to the phase boundary that give rise to a polar ordering transition [ red dots in fig . [ plot : phase_diagram_simulation ] ] . the only difference manifests in a subsequent third regime ( iii ) where the system begins to develop a polarized state with @xmath335 , which is reflected by a further increase of @xmath336 $ ] away from the plateau value . the prolonged plateau can be interpreted as a _ lag phase _ , in which the system remains isotropic [ @xmath337 , fig . [ plot : relativeangledistribution](c ) , red ( light grey ) curve ] and `` waits '' for the nucleation of a cluster of sufficiently large size @xcite . please refer to the supplemental material @xcite for a video depicting the time evolution of the system from the initial configuration to the fully polarized state . taken together , close to the phase boundary on the isotropic as well as the polar side orientational correlations exist which are the essential prerequisites for a subsequent transition to a polar state . these correlations are a _ precursor _ phenomenon that precedes the phase transition . now we would like to study the implications of these _ precursor _ correlations on the phase boundary [ eq . ] . to this end , we use the plateau value of @xmath338 [ see fig . [ plot : relativeangledistribution](d ) , dark grey curve ] , and assume that the underlying @xmath302 is valid for all packing fractions @xmath80 and noise strengths @xmath70 . the corresponding result for the phase boundary @xmath305 is depicted by the solid line in fig . [ plot : phase_boundary ] . it nicely agrees with the phase boundary obtained from multi - particle simulations for small packing fractions . this indicates that our extended kinetic theory for propelled particle systems constitutes a quantitative description for soft active colloids . further our findings stress the significance of correlations in active systems at the onset of collective motion . however , for larger packing fractions , polar order persists beyond the critical noise strength predicted by the boltzmann theory . this increased stability of polar order with respect to noise may be attributed to clustering processes in the regime of intermediate packing fractions . how such effects are properly accounted for within a kinetic theory is presently unclear . one will certainly need to go beyond eq . and account for higher - order correlations , or employ a multi - species formulation @xcite . and angular noise @xmath70 [ degrees ] : results obtained from multi - particle simulations are displayed by symbols , while the solid line corresponds to the predictions of kinetic theory for the phase boundary . _ multi - particle simulation : _ red dots indicate polarized states while grey squares correspond to control parameter sets where no change in the initial isotropic state has been observed . _ analytical results : _ the solid line indicates the phase boundary obtained from kinetic theory [ eq . ] using @xmath339 [ resulting in @xmath340 ; see fig . [ plot : relativeangledistribution](d ) ] for all @xmath80 and @xmath70 . ] in this paper we addressed the question focusing on dilute conditions whether we can understand the collective dynamics of soft active colloids by solely considering _ binary _ interactions between the constituent particles . to this end we performed binary scattering studies of active colloids interacting via soft repulsive interactions . from these studies we discovered that the dynamic models considered share the same principle of parallel alignment . the underlying principle was termed as `` _ _ alignment of the first _ _ , '' stating that the first incoming particle with respect to the center of collision is aligned to the second of the colliding particles . we showed that these types of collisions contribute the dominant part to the system s alignment tendency . additionally , this principle is genuinely different to alignment in systems of inelastic gases , which is mediated by damping of the relative particle velocities . moreover , from the binary scattering studies we deduced a non - linear collision rule mapping pre - collisional on post - collisional velocities that is devoid of any approximation . this collision rule was then connected to the system s collective behavior by kinetic theory for propelled particle systems . the microscopic origin of the collision rule allowed to quantitatively scrutinize the predictions of kinetic theory with regard to the phase boundary marking the instability of the isotropic unpolarized state . by comparing the resulting phase boundary with that obtained from multi - particle simulations of the underlying microscopic model for active colloids , we discovered that non - trivial modifications in the kinetic description are necessary to obtain a quantitative agreement in the phase boundary . specifically , we found that _ precursor _ orientational and spatial correlations exist close to the phase boundary . only if the kinetic description included these correlations , the analytic prediction for the phase boundary coincided quantitatively for small packing fractions with the one from multi - particle simulations . most importantly , if orientational correlations were neglected , kinetic theory for propelled particles failed , _ i.e. _ , it predicted that ordering is absent , which is at odds with corresponding molecular dynamics simulations . our findings clearly indicate that the framework of kinetic theory for propelled particle systems is flexible enough to accommodate the complex behavior of soft active colloids and allow a bottom - up understanding of how the microscopic dynamics of binary collisions is related to the system s behavior on large length and time - scales . the developed renormalized " kinetic theory , where the interaction kernel , _ i.e. _ , the collision rule and the correlations of the pre - collisional state , are determined from microscopic molecular dynamics simulations , could serve as the appropriate starting point for an extension of kinetic theory for propelled particle systems into the regime of intermediate packing fractions . moreover , we are convinced that our approach is also perfectly suited to bridge between microscopic experimental studies of propelled particle systems @xcite , in which _ precursor _ correlations are likely to exist , and their corresponding quantitative mesoscopic description . the authors would like to thank florian throff for discussions and critical reading of our manuscript . this project was supported by the deutsche forschungsgemeinschaft in the framework of the sfb 863 and the german excellence initiative via the program `` nanosystems initiative munich '' ( nim ) . ( 3.8,1 ) node[coordinate ] ( p ) + ( 10:1 ) arc(10:350:1 ) ; ( p ) + ( 10:2 ) arc(10:350:2 ) ; ( p ) + + ( 10:2 ) node[coordinate ] ( cylinder_start_left ) + ( 30:3 ) node[coordinate ] ( cylinder_end_left ) ; ( p ) + + ( 0:2 ) node[coordinate ] ( cylinder_start_middle ) + ( 30:2.5 ) node[coordinate ] ( cylinder_nearend_middle ) ; ( p ) + + ( -10:2 ) node[coordinate ] ( cylinder_start_right ) + ( 30:3 ) node[coordinate ] ( cylinder_end_right ) ; ( p ) ( cylinder_start_left ) ( cylinder_start_right ) cycle ; ( cylinder_start_left ) ( cylinder_end_left ) ( cylinder_end_right ) ( cylinder_start_right ) cycle ; ( p ) node[above=2 , pos=0.3 ] @xmath93 + ( 0:1.5 ) ; ( cylinder_nearend_middle ) ( cylinder_start_middle ) ; ( cylinder_nearend_middle ) node[fill = white , pos=0.4,sloped ] @xmath110 + ( 30:-1.9 ) ; ( cylinder_end_right ) + + ( 0,-1.5 ) + ( 0,-0.15 ) coordinate ( cylinder_brace_start ) ; ( cylinder_start_right ) + ( 0,-0.15 ) coordinate ( cylinder_brace_end ) ; ( cylinder_brace_start ) node[pos=0.45,below=3pt ] @xmath341 ( cylinder_brace_end ) ; ( p ) + + ( 0.6,1.9 ) node[fill = white ] ( text_1 ) @xmath342 ; ( cylinder_start_left ) + ( 0,-0.25 ) ( text_1.south ) ; ( p ) + + ( 1,-1.35 ) node[fill = white , pos=0.5 ] @xmath51 + ( 180:2 ) ; in the following we derive the integration weight used in eq . , which depends on the specific collision geometry . to this end we assume uniformly distributed positions and velocities and consider a collision of particles with diameter @xmath51 , relative velocity @xmath110 , and the point of contact defined by the unit vector @xmath93 [ see fig . [ pic : coll_geometry ] ] . the likelihood of such a collision is proportional to the number of particles in the boltzmann collision cylinder [ fig . [ pic : collision_cylinder ] ] with volume @xmath343 the number of particles with velocity @xmath175 in the collision cylinder is proportional to @xmath344 , where @xmath345 is the one - particle distribution function , which is independent of position in a homogeneous system . integration yields the total number of particles that collide with a given particle with velocity @xmath176 per time interval @xmath346 : @xmath347 where the heaviside function ensures that the particles collide [ see fig . [ pic : collision_cylinder ] ] . to rewrite eq . as an integral over the impact parameter @xmath97 , we employ the angle @xmath7 [ defined in fig . [ pic : coll_geometry](b ) ] as an intermediate step : @xmath348 where we used @xmath349 . using the definition of the impact parameter as @xmath350 [ eq . ] , we arrive at @xmath351 since there is no weight function depending on @xmath97 left in the integral , collisions with different impact parameters occur at the same rate provided that the system is homogeneous and devoid of correlations . note that this result is valid independent of the distribution of particle velocities . for the system considered here , we make two assumptions concerning the velocity distribution function : ( i ) particle velocities are relaxed to the stationary value @xmath239 before a collision , _ i.e. _ , @xmath352 where the relative angle @xmath104 denotes the direction of @xmath175 with @xmath176 defining the reference axis . ( ii ) directions of particle velocities are isotropically distributed , _ i.e. _ , @xmath353 . taken together , this yields for the rate of collision given in eq . @xmath354 where we used that the norm of the relative velocity @xmath355 . the intuitive interpretation for the weight function is that two particles moving in the same direction with equal speed will never collide , whereas particles moving in opposite directions have maximal relative velocity and therefore occur most often . particle exchange symmetry allows to restrict the range of integration to @xmath106 $ ] . therefore , we can define the following normalized weighted average over all collision geometries : @xmath356 using eq . for the relative alignment @xmath119 [ eq . ] , we arrive at the alignment integral given in eq . . ( solid red line ) and standard deviation @xmath357 ( dashed green line ) of the scattering distribution @xmath234 [ cf . fig . [ plot : scatdist_charfunct ] ] . since particles are identical , the corresponding deviations are equal and @xmath358 . * ( b ) * _ phase diagram _ as a function of packing fraction @xmath80 and angular noise @xmath70 [ degrees ] , as described in caption of fig . [ plot : phase_boundary ] . the solid line corresponds to the phase boundary obtained from the kinetic approach using a _ gaussian approximation _ _ ) for the scattering distribution and assuming _ molecular chaos _ , _ i.e. _ , @xmath359.,title="fig : " ] ( solid red line ) and standard deviation @xmath357 ( dashed green line ) of the scattering distribution @xmath234 [ cf . fig . [ plot : scatdist_charfunct ] ] . since particles are identical , the corresponding deviations are equal and @xmath358 . * ( b ) * _ phase diagram _ as a function of packing fraction @xmath80 and angular noise @xmath70 [ degrees ] , as described in caption of fig . [ plot : phase_boundary ] . the solid line corresponds to the phase boundary obtained from the kinetic approach using a _ gaussian approximation _ _ ) for the scattering distribution and assuming _ molecular chaos _ , _ i.e. _ , @xmath359.,title="fig : " ] in order to quantitatively scrutinize the predictions for the phase boundary obtained from kinetic theory , accounting for the full characteristic function [ eq . ] is crucial . this can be seen when considering a _ gaussian approximation _ for the scattering distribution @xmath234 [ cf . [ plot : scatdist_charfunct ] ] , _ i.e. _ , expressing the characteristic function through the moments of the distribution and keeping only the first two terms @xmath360 where @xmath361 denotes the mean scattering angle and @xmath362 is the standard deviation of the scattering distribution ; both are depicted in fig . [ plot : gaussian_approximation](a ) . using @xmath359 ( _ assumption of molecular chaos _ ) and @xmath363 to compute the scattering coefficients @xmath276 [ eq . ] , the phase boundary in the _ gaussian approximation _ can be determined ; see solid black line in fig . [ plot : gaussian_approximation](b ) . in section [ sect : collective_comparison ] we showed that _ precursor _ correlations exist at the onset of collective motion , and are essential to reproduce quantitatively the phase boundary obtained from md simulations . even though the _ gaussian approximation _ correctly predicts the existence of an ordering transition , this is only due to an inaccurate approximation : it misrepresents the actual scattering behavior by underestimating the impact of scattering events with large relative angles , and thus masks the true reason for the instability . 63ifxundefined [ 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 _ _ ( , , ) chap . link:\doibase 10.1140/epje / e2005 - 00002 - 5 [ * * , ( ) ] link:\doibase 10.4064/bc80 - 0 - 6 [ * * , ( ) ] , link:\doibase 10.1146/annurev - conmatphys-070909 - 104101 [ * * , ( ) ] @noop * * , ( ) http://stacks.iop.org/0034-4885/74/i=11/a=116601 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.85.1143 [ * * , ( ) ] link:\doibase 10.1074/jbc.m109.044792 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.93.098103 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1039/c0sm01063d [ * * , ( ) ] link:\doibase 10.1073/pnas.1107540108 [ * * , ( ) ] link:\doibase 10.1073/pnas.1001651107 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1073/pnas.1107281109 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.098001 [ * * , ( ) ] link:\doibase 10.1039/c2sm25186h [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.058001 [ * * , ( ) ] link:\doibase 10.1016/j.tics.2008.10.002 [ * * , ( ) ] link:\doibase 10.1016/j.anbehav.2008.02.004 [ * * , ( ) ] link:\doibase 10.1098/rsfs.2012.0033 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.75.1226 [ * * , ( ) ] link:\doibase 10.1016/s0378 - 4371(00)00013 - 3 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.92.025702 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physreve.77.046113 [ * * , ( ) ] link:\doibase 10.1103/physreve.80.050103 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.208001 [ * * , ( ) ] link:\doibase 10.1103/physreve.63.017101 [ * * , ( ) ] link:\doibase 10.1103/physreve.71.051904 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.104302 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) http://dx.doi.org/10.1007/s100510051104 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.230601 [ * * , ( ) ] http://stacks.iop.org/1367-2630/14/i=7/a=073033 [ * * , ( ) ] http://dx.doi.org/10.1140/epjst/e2012-01529-y [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.010602 [ * * , ( ) ] link:\doibase 10.1103/physreve.71.050901 [ * * , ( ) ] link:\doibase 10.1103/physreve.74.022101 [ * * , ( ) ] @noop * * , ( ) http://stacks.iop.org/0295-5075/99/i=5/a=58001 [ * * , ( ) ] @noop _ _ ( , , ) @noop http://stacks.iop.org/1367-2630/15/i=4/a=045014 [ * * , ( ) ] link:\doibase 10.1103/physreve.86.030901 [ * * , ( ) ] link:\doibase 10.1080/00268978200101711 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.268701 [ * * , ( ) ] @noop
collective motion in actively propelled particle systems is triggered on the very local scale by nucleation of coherently moving units consisting of just a handful of particles . these units grow and merge over time , ending up in a long - range ordered , coherently - moving state . so far , there exists no bottom - up understanding of how the microscopic dynamics and interactions between the constituents are related to the system s ordering instability . in this paper , we study a class of models for propelled colloids allowing an explicit treatment of the microscopic details of the collision process . specifically , the model equations are newtonian equations of motion with separate force terms for particles driving , dissipation and interaction forces . focusing on dilute particle systems , we analyze the binary scattering behavior for these models , and determine based on the microscopic dynamics the corresponding collision - rule " , _ i.e. _ , the mapping of pre - collisional velocities and impact parameter on post - collisional velocities . by studying binary scattering we also find that the considered models for active colloids share the same principle for parallel alignment : the first incoming particle ( with respect to the center of collision ) is aligned to the second particle as a result of the encounter . this behavior is distinctively different to alignment in non - driven dissipative gases . moreover , the obtained collision rule lends itself as a starting point to apply kinetic theory for propelled particle systems in order to determine the phase boundary to a long - range ordered , coherently - moving state . the microscopic origin of the collision rule offers the opportunity to quantitatively scrutinize the predictions of kinetic theory for propelled particle systems through direct comparison with multi - particle simulations . we identify local _ precursor _ correlations at the onset of collective motion to constitute the essential determinant for a qualitative and quantitative validity of kinetic theory . in conclusion , our renormalized " approach clearly indicates that the framework of kinetic theory is flexible enough to accommodate the complex behavior of soft active colloids and allows a bottom - up understanding of how the microscopic dynamics of binary collisions relates to the system s behavior on large length and time - scales .
with the discovery of the higgs boson @xcite , the particle content of the standard model ( sm ) is well established . the future aim of the large hadron collider ( lhc ) lies in uncovering clues for beyond the sm ( bsm ) physics - we expect these would manifest in the form of tev scale resonances . to make connection with theoretical models , we would need to measure both the mass of such a resonance and its coupling to the sm particles . however , it is not straightforward to measure the couplings of a massive particle in a hadron collider environment , where the momenta of the interacting partons are not known for any single event . generally , it is easier to measure ratios of couplings or production rate times branching ratios ( brs ) . from these , in principle , it is possible to measure any coupling if the total decay width of the particle is also known . however , since unlike the brs , measuring the width of a particle accurately can be quite difficult , this way of measuring couplings may lead to large uncertainties . in this note we present a method of extracting couplings at a hadron collider that depends only on the brs but not on the measured width or the production mechanisms of the particle in question . motivation for this method comes from the simple observation that the decay of an off - shell particle is sensitive to the coupling involved . this idea , in itself , is not new and has been used , for example , in studies about constraining the higgs width @xcite and also in ref . @xcite where it was hinted that it can be used to measure new physics couplings . what we propose in this note is a systematic way of extracting unknown couplings using this observation by identifying a set of new physical variables that are sensitive to the coupling of interest . when a massive unstable particle decays to lighter particles , the invariant mass of the daughter particles shows a distribution peaked about the mass of the particle . the shape of the distribution is well approximated by the famous breit - wigner distribution around the peak ( resonance ) and its width gives us the decay rate . discovery of a new unstable particle usually entails looking at this distribution , designing cuts to isolate regions close to the peak as its position gives us the mass of the particle . however , the same distribution can also be used to extract information about the coupling involved in the decay since both the width and the height of the distribution depend on it . if instead of looking near the resonance we look at regions away from it , the cross - sections increase with the coupling as the width grows . at the same time , the height of the peak decreases with the increasing coupling keeping the total cross - section roughly the same . in other words , when the coupling involved in the decay increases , the distribution spreads out keeping the area covered roughly constant . this tells us it may be possible to learn about the coupling involved in the decay if instead of the total cross - section we focus on parts of the phase - space . to begin with , let us consider the single production of a heavy particle @xmath0 within a toy model at the lhc . the @xmath0 decays to two sm particles , @xmath1 and @xmath2 , with the decay controlled by an unknown coupling @xmath3 . the actual method of production of @xmath0 does not concern us here as we are interested in its decay and the final state kinematics . our analysis would hold even if @xmath0 is produced in association with a different particle or produced in pairs . the width - dependent part of the amplitude for the process @xmath4 can be written as : m ( qq^ ) _ qq /,where @xmath5 is the momentum of @xmath0 . from this , it is clear that for @xmath6 , i.e. , near the resonance , the cross - section scales as @xmath7 . if @xmath8 is the dominant decay mode of @xmath0 , then @xmath9 and so @xmath10 . on the other hand , in the kinematic region where @xmath11 , the cross - section scales like @xmath12 . our minimal assumptions at this point are that both @xmath13 and br(@xmath14 ) are known fairly accurately from experiments . is equivalent to measuring its decay width . hence the recent studies about constraining the higgs width @xcite are actually similar in spirit to this study . ] below , we outline the main steps that one could follow to obtain @xmath3 . 1 . the first step is to numerically simulate the process for many different values of @xmath15 using @xmath16 obtained from the theoretically computed partial width and the experimentally measured br as @xmath17 . this way , one can avoid using the measured width but rely on the br that can be measured more accurately . the regions _ outside _ the resonance can be accessed with an invariant mass cut on @xmath18 defined by a new parameter @xmath19 that parametrizes the degree of off - shellness of the intermediate @xmath20 : |m(qq)-m_| & _ o m _ .[eq : phioff - cut ] we note that the above cut is similar to the one used in @xcite . for each @xmath21 and for various choices of @xmath19 , one could compute the following ratio : r(_qq^i)|__r,_o = , where @xmath22 is the _ on - shell _ cross - section , computed using a cut that isolates the resonance : |m(qq)-m_| & _ r m_.[eq : phion - cut ] the advantage of working with the ratio @xmath23 instead of either cross - section is that most uncertainties associated with the production mechanism involving pdfs etc . get canceled in the ratio . and more importantly , any other unknown coupling present in the production of @xmath20 , i.e. , @xmath24 , also gets canceled . now , for each combination of @xmath25 , one can prepare a `` calibration curve '' ( cc ) by interpolating between @xmath26 s and thus prepare a family of such curves . 3 . from the experiment one can measure the ratio @xmath27 for all the @xmath25 combinations considered to prepare the ccs . here , we have accounted for errors in the cross - section measurements by a factor @xmath28 . now , in principle , one could simply match @xmath29 for any single combination of @xmath25 with the corresponding theoretical cc and read off the coupling . however , there are some difficulties associated with this procedure . the efficiency of coupling extraction is not uniform for a particular cc since it depends on the steepness ( or slope ) of the curve - for instance , if a particular cc is almost flat , it is not possible to extract a unique value of the coupling by matching on to the experiment . _ a priori _ , it is not clear how to select the optimal combination of @xmath25 . one could also take the average of the couplings extracted from different ccs but the issue of assigning proper weight factors for averaging may become an ambiguous issue . instead , one can extract @xmath3 from a simultaneous fit using all the ccs . this can be done , e.g. , by maximizing a likelihood function defined as : l = _ k , [ eq : likelihood ] where the index @xmath30 runs over all the combinations of \{@xmath31}. the @xmath3 thus extracted will correspond to the value for which @xmath32 best describes @xmath33 for all the \{@xmath31 } combinations . for these outlined steps to work optimally , the process in question should have a large enough cross - section so that there are an appreciable number of events left over after the off - shell cuts . if the cross - section is small it may not be possible to measure @xmath33 reasonably accurately for many \{@xmath31 } combinations which in turn will increase the error in fitting . also , the method works better with a broad resonance . thus far , our discussion has been purely about the signal process . however , in practice , handling the sm backgrounds is an important issue . the preparation of the ccs should thus also include the effect of the sm background and its interference with the signal . inclusion of these effects could change the dependence of the ccs on the coupling ( due to the interference term ) . however , since the method does not rely on the nature of the dependence and only on the fact that there is one , we expect the outlined steps to work . we demonstrate this within the context of an illustrative example in the next section . to demonstrate the above method with an example , we consider a simple model with a new , heavy color triplet @xmath34 quark with electric charge @xmath35 . we make the simplifying assumption that it decays to @xmath36 via : @xmath37 in the lagrangian above , @xmath38 is the unknown coupling that we want to extract . at the lhc , if the @xmath39 is not too heavy , qcd mediated pair production will be the dominant production channel . there will also be the @xmath40 mediated single production channel of @xmath41 , like @xmath42 . following the steps described before one can use the single production process to probe @xmath38 . however , as this process is weak interaction mediated , its cross - section will be quite small and it will be more so if @xmath38 is also small . hence , to take advantage of the enhanced cross - sections , here we make use of pair production and simply apply off - shell cuts to one of the two final state particles - this reiterates our earlier point that this analysis depends only on the kinematics of the final state particles and we are free to choose the particular production process that would serve our purpose well . we note , however , that in our analysis below we will include all possible contributions to the @xmath43 process , although the pair production will dominate . in fig . [ fig : dist ] we show the invariant mass distributions of the @xmath36 pair in @xmath44 for different values of @xmath38 . as @xmath45 increases with increasing @xmath38 , the distribution spreads out and , with the increasing width , the height of the distribution decreases so that the area contained by the distribution remains approximately constant , i.e , the total cross - section is not very sensitive to @xmath38 , reiterating our point at the beginning of the last section . the invariant mass distribution of the @xmath36 pair in the process @xmath46 for different values of @xmath38 ( events generated with with madgraph5 @xcite ) . here , we fix @xmath47 tev and br@xmath48 . ] but as can be seen , there _ are _ parts of the phase space that are sensitive to @xmath38 . for example , if we look away from the mass peak , i.e. , when the @xmath36 pair is not coming from a close to on - shell @xmath39 , the cross - section becomes highly sensitive to @xmath38 , in keeping with our discussion earlier . of course , the price to pay for looking away from the resonance is the reduced cross - section compared to the resonant production . however , we can mitigate this to a large extent by considering `` mixed states '' where one @xmath39 is produced resonantly . to extract @xmath38 , we follow the steps outlined before . assuming @xmath49 and @xmath50 are known from experiments , we generate events for @xmath51 in madgraph5 @xcite ( including both @xmath39 and sm contributions as well as their interference ) for different values of @xmath38 where we compute @xmath52 as @xmath53 . for this example , we set @xmath54 tev and @xmath55 and generate events for the 14 tev lhc . since in this case we are considering pair production , rather than using the cut defined in eq . [ eq : phioff - cut ] , we modify it to include two cuts on the @xmath36 pairs , one on - shell and one off - shell : @xmath56 where @xmath57 or @xmath58 and the numbers imply that the particles are @xmath59-ordered . condition ( i ) reconstructs the resonant @xmath39 while condition ( ii ) accesses the @xmath38 sensitive off - shell region . our expression for @xmath60 is given by : r(_tw^i)|__r,_o = , [ eq : bprr ] where @xmath61 is now computed by applying cuts as in eq . [ eq : on - cut ] to reconstruct two resonant @xmath62s . we produce a family of ccs for @xmath63 s with various combinations of \{@xmath31}. the dependence of @xmath64 on @xmath38 with @xmath65 cuts defined in eqs . ( [ eq : on - cut ] ) & ( [ eq : off - cut ] ) . the curves marked with only @xmath66 show @xmath67 . ] in fig . [ fig : cali ] we plot @xmath68 ( including sm and bsm contributions ) with respect to @xmath38 before and after cuts for three different choices of @xmath25 . we also show @xmath69 in the same figure . the total cross - section @xmath68 without any cut is mostly insensitive to @xmath38 it remains almost constant for smaller @xmath38 , but slightly decreases in the higher @xmath38 region ( mostly ) because of the larger widths . in this example the conditions [ eq : on - cut ] and [ eq : off - cut ] cuts become competitive in nature as they are applied on different @xmath36 pairs simultaneously . applying the cut in eq . [ eq : on - cut ] we reconstruct the resonant @xmath39 , i.e. , we accept the @xmath36 pairs that fall _ inside _ the @xmath66 mass window and so an increasing @xmath70 means accepting more events ( larger cross - section ) . whereas , with eq . [ eq : off - cut ] we look away from the peak ( we accept the @xmath36 pairs that fall _ outside _ the @xmath19 mass window ) and so an increasing @xmath19 results in reduced cross - section . with this in mind , it becomes easier to understand the curves with the @xmath25 cuts . for example , for a fixed @xmath66 , we expect the cross - section to go down with increasing @xmath19 . this we can see by comparing the curves with \{0.10 , 0.01 } and \{0.10 , 0.10 } the second one is smaller than the first for all values of @xmath38 . similarly for a fixed @xmath19 , if we decrease @xmath66 we should get a smaller cross - section which can be seen from the curves with \{0.10 , 0.01 } and \{0.04 , 0.01}. we also observe that with increasing @xmath38 , the @xmath68 curves first increase then decrease . this is because for a fixed @xmath66 , as we increase @xmath38 ( and hence @xmath71 ) , we miss more and more @xmath36 pairs that come from the resonant @xmath39 , i.e. , reconstruction efficiency of the resonant @xmath39 decreases and as a result the cross - section decreases . this also explains why the @xmath72 curves , for which we are reconstructing two resonant @xmath39 s , fall ( and fall faster compared to @xmath68 curves ) with increasing @xmath38 . when the coupling @xmath73 0 , the difference between the @xmath74 curves and the total cross - section is accounted for by the sm background . to see how the method performs we have considered four different test couplings @xmath75 and use @xmath76 in place of @xmath33 ( eq . [ eq : bprr ] ) , assuming all @xmath76 s have uniform 10% errors for all combinations of @xmath25 . we have considered eight different values of @xmath70 between @xmath77 and @xmath78 and ten different values of @xmath19 between @xmath79 and @xmath80 . we present the results of our analysis in table [ tab : fit ] . we see that we are able to extract the couplings with percent - level accuracies . we note , however , that ours is a parton level analysis . doing full detector simulations and considering sm backgrounds properly are likely to modify the extraction efficiency . these issues are currently under investigation @xcite . .[tab : fit ] the different choices for @xmath81 and the corresponding values obtained by maximizing the likelihood function defined in eq . [ eq : likelihood ] . the errors in the fitted values correspond to @xmath82 . [ cols="^,^,^",options="header " , ] in this note , we presented a novel method of extracting the couplings of new , heavy states to sm particles . the essential point underlying the method is a rather simple observation that when a particle decays inside a collider , the invariant mass distribution of its decay products retains information of the coupling involved in the decay . with the off - shell cuts , this coupling extraction procedure actually becomes sensitive to the shape of the invariant mass distribution . the method is largely free of modeling assumptions apart from the fact it uses theoretically computed partial decay width to avoid the use of the measured total width of the new particle , and thus avoids the errors associated with width measurement . of course , computing the total width using the theoretically computed partial width and the experimentally measured br means that some errors could come from these sources and ideally one should account for both . however , since generally these errors are expected to be much less severe than the ones in measuring widths experimentally , we neglect these for the time being . similarly we neglect any error in measuring the mass of the new particle . the method does not depend on the exact production mechanism of the new particle . the coupling extraction method outlined here is completely general and applicable to any new fermionic or bosonic state . it also has the advantage of being independent of collider details , or the actual values of the mass and the brs of the particle in question . in this note , we have ignored the issue of considering the complete sm backgrounds which in reality is a very crucial one . since realistic background can only be considered in a case - by - case basis , inclusion of simple parton level background in the example shown should only be considered as an illustration . however , we hope that after the discovery of a new particle , its background will also be fairly well known and thus can be dealt with . since our aim here is simply to outline the methodology , we postpone a more complete demonstration with detector level simulations for both signal and background to a future publication @xcite . to extract any coupling the cross - section of the process in question needs to be large enough to start with so we have enough events left after all the cuts . it is interesting to contrast this method with the usual discovery procedure . if the new state has a significant width ( comparable to its mass ) , discovery is rendered more difficult as we would not have the typical sharp peak . on the other hand , a large width guarantees that regions away from resonance do not have negligibly small cross - sections and hence coupling extraction becomes more feasible . of course , one need not restrict attention to new physics alone . in principle , this method can be applied to extract the sm ckm matrix element @xmath83 from top pair production . in fact , it can be applied to extract any coupling , be it sm or bsm , as long as enough events remain after the cuts . we thank heather logan and biswarup mukhopadhyaya for helpful comments on the draft , shrihari gopalakrishna and erich varnes for useful discussions . bc is supported by the department of energy under grant de- fg02 - 13er41976 . tm is partially supported by funding from the dae , for the recapp , hri . sm acknowledges financial support from the cnrs . bc acknowledges the hospitality of imsc where this project was started . 99 g. aad _ et al . _ [ atlas collaboration ] , phys . b * 716 * , 1 ( 2012 ) [ arxiv:1207.7214 [ hep - ex ] ] ; s. chatrchyan _ et al . _ [ cms collaboration ] , phys . b * 716 * , 30 ( 2012 ) [ arxiv:1207.7235 [ hep - ex ] ] ; f. caola and k. melnikov , phys . d * 88 * , 054024 ( 2013 ) [ arxiv:1307.4935 [ hep - ph ] ] . j. m. campbell , r. k. ellis and c. williams , jhep * 1404 * , 060 ( 2014 ) [ arxiv:1311.3589 [ hep - ph ] ] . j. m. campbell , r. k. ellis and c. williams , phys . d * 89 * , 053011 ( 2014 ) [ arxiv:1312.1628 [ hep - ph ] ] . n. kauer and g. passarino , jhep * 1208 * , 116 ( 2012 ) [ arxiv:1206.4803 [ hep - ph ] ] . v. khachatryan _ et al . _ [ cms collaboration ] , phys . b * 736 * , 64 ( 2014 ) [ arxiv:1405.3455 [ hep - ex ] ] . c. englert and m. spannowsky , phys . d * 90 * , 053003 ( 2014 ) [ arxiv:1405.0285 [ hep - ph ] ] . s. gopalakrishna , t. mandal , s. mitra and g. moreau , jhep * 1408 * , 079 ( 2014 ) [ arxiv:1306.2656 [ hep - ph ] ] . s. gopalakrishna , t. mandal , s. mitra and r. tibrewala , phys . d * 84 * , 055001 ( 2011 ) [ arxiv:1107.4306 [ hep - ph ] ] ; r. contino , l. da rold and a. pomarol , phys . d * 75 * , 055014 ( 2007 ) [ hep - ph/0612048 ] ; c. anastasiou , e. furlan and j. santiago , phys . d * 79 * , 075003 ( 2009 ) [ arxiv:0901.2117 [ hep - ph ] ] ; n. vignaroli , jhep * 1207 * , 158 ( 2012 ) [ arxiv:1204.0468 [ hep - ph ] ] ; a. de simone , o. matsedonskyi , r. rattazzi and a. wulzer , jhep * 1304 * , 004 ( 2013 ) [ arxiv:1211.5663 [ hep - ph ] ] ; t. han , h. e. logan , b. mcelrath and l. -t . wang , phys . d * 67 * , 095004 ( 2003 ) [ hep - ph/0301040 ] ; m. s. carena , j. hubisz , m. perelstein and p. verdier , phys . rev . d * 75 * , 091701 ( 2007 ) [ hep - ph/0610156 ] ; s. matsumoto , t. moroi and k. tobe , phys . rev . d * 78 * , 055018 ( 2008 ) [ arxiv:0806.3837 [ hep - ph ] ] ; j. berger , j. hubisz and m. perelstein , jhep * 1207 * , 016 ( 2012 ) [ arxiv:1205.0013 [ hep - ph ] ] ; j. kang , p. langacker and b. d. nelson , phys . d * 77 * , 035003 ( 2008 ) [ arxiv:0708.2701 [ hep - ph ] ] ; p. w. graham , a. ismail , s. rajendran and p. saraswat , phys . d * 81 * , 055016 ( 2010 ) [ arxiv:0910.3020 [ hep - ph ] ] ; s. p. martin , phys . d * 82 * , 055019 ( 2010 ) [ arxiv:1006.4186 [ hep - ph ] ] ; j. alwall , m. herquet , f. maltoni , o. mattelaer and t. stelzer , jhep * 1106 * , 128 ( 2011 ) [ arxiv:1106.0522 [ hep - ph ] ] . b. coleppa , t. mandal , and s. mitra ( work in progress ) .
in this note , we present a novel method of extracting the couplings of a heavy particle to the standard model states . contrary to the usual discovery process which involves studying the on - shell production , we look at regions away from resonance to take advantage of the simple scaling of the cross - section with the couplings . we illustrate the procedure with the example of a heavy quark .
recently the evidences of neutrino oscillations are strongly supported by both of the atmospheric @xcite and the solar neutrino experiments @xcite . the former suggests an almost maximal lepton flavor mixing between the 2nd and the 3rd generations , while the favorable solution to the solar neutrino deficits is given by large mixing angle solution between the 1st and the 2nd generations ( lma , low or vo ) @xcite . neutrino oscillation experiments indicate that the neutrinos have tiny but finite masses , with two mass squared differences @xmath1 . since the neutrino mass is much smaller than other quarks and leptons , the origin of the neutrino mass is expected to be different from that of others . the most popular mechanism to induce the small neutrino masses is the seesaw mechanism @xcite , which needs heavy right - handed neutrinos . however , it is important to consider also other possible scenarios which can explain the small neutrino masses , especially low - energy extensions of the standard model . the zee model @xcite is such an alternative , which induces small neutrino masses by radiative corrections , where the lepton number is violated explicitly . generally , if there are lepton number violating interactions , care has to be taken for these interactions not to be too strong at the temperatures above the electroweak scale , since the baryon asymmetry generated in the early universe might be washed out by the `` sphaleron '' process @xcite , which violates a linear combination of baryon ( @xmath2 ) and lepton ( @xmath3 ) number , @xmath4 . if the rate of the lepton number violating process becomes faster than the hubble expansion rate @xmath5 during the epoch when the sphaleron process is in thermal equilibrium , the baryon asymmetry generated at higher temperature would be washed out , and there would be no matter anti - matter asymmetry in the present universe , which conflicts with the observation . in this letter , we analyze the cosmological condition in order for the baryon asymmetry generated in the early universe not to be washed out in the zee model . ( we will not discuss the production mechanism of the baryon asymmetry . instead , we just assume that the desired amount of the baryon asymmetry is generated in the early universe and discuss whether or not this baryon asymmetry can be preserved against the lepton number violating interaction in the zee model together with the sphaleron process . ) in most part of the analysis , we take the low solution to the solar neutrino problem . ( we shall note on the cases of other solutions in the last section . ) we find that the baryon asymmetry is _ not _ washed out , although it has been said that the zee model can not preserve the baryon asymmetry @xcite . this can be easily shown from the view point of a new lepton number @xmath0 . let us give a brief overview of the zee model @xcite at first . the zee model is a simple extension of the standard model , which has two higgs doublets @xmath6 ( @xmath7 ) and one @xmath8-singlet charged higgs field ( zee singlet ) @xmath9 . the zee model has the following interactions in addition to the standard model ones ; @xmath10 where @xmath11 denote the left - handed lepton doublets with flavor indices @xmath12 . notice that the coupling @xmath13 is anti - symmetric for the flavor index . for simplicity , we have omitted the higgs potential @xmath14 which are irrelevant to the following discussion . as for the yukawa interaction , we assume only @xmath15 couples to lepton fields as , @xmath16 where @xmath17 . as can be seen from eqs . ( [ zee_lag ] ) and ( [ l - yukawa ] ) , the lepton number @xmath3 is explicitly violated in the zee model . and @xmath18 , @xmath19 , in @xmath20 . ] in the charged higgs sector , the zee singlet @xmath21 is mixed with the charged higgs boson @xmath22 through the @xmath23 interaction in eq . ( [ zee_lag ] ) ; @xmath24 here , @xmath25 is the charged higgs boson which is orthogonal to the would - be goldstone boson after the neutral higgs fields acquire vacuum - expectation values ( vevs ) @xmath26 , where @xmath27/@xmath28 . we denote the mass eigenstates of the charged higgs sector as @xmath29 and their mass @xmath30 . ( 5,5 ) ( -2,0 ) ( -2,1.7)@xmath31 ( 0,.5)@xmath32 ( 2,-.5)@xmath33 ( 1.2,.7)@xmath34 ( 3.1,.7)@xmath35 ( 4.2,.5)@xmath36 ( 5.7,1.7)@xmath37 ( 2.3,1.5)@xmath36 ( 2,4.5)@xmath38 ( .5,3)@xmath39 ( 3.8,3)@xmath40 ( 2.3,2.7)@xmath41 in the zee model , neutrino masses are generated by radiative corrections , as shown in fig . [ fig : zee_mass ] , and hence this model could provide an explanation of the smallness of neutrino masses . the element of the mass matrix , generated by radiative correction at one loop level , is given by @xmath42 here @xmath43 are the charged lepton masses . in eq . ( [ zee_mass ] ) , we have used @xmath44 . since the coupling constants @xmath45 are antisymmetric for the indices @xmath46 , the mass matrix eq . ( [ zee_mass ] ) is traceless . the above zee mass matrix has been analyzed @xcite in the light of recent neutrino - oscillation experiments , and it was shown that it must be the following bi - maximal form to explain the experimental results ; @xmath47 where @xmath48 , @xmath49 , @xmath50 and @xmath51 . the eigenvalues and the mns matrix induced from the neutrino mass matrix eq . ( [ zee_mass_matrix ] ) is given by @xmath52 and @xmath53 the neutrino mass hierarchy is shown in fig . [ fig : mass ] . the mixing angle between the 1st and the 2nd generations is maximal as well as the mixing angle between the 2nd and the 3rd generations ( @xmath54 ) . from oscillation experiments , the eigenvalues of the mass matrix have to satisfy the following relations , @xmath55 ( 6,5 ) ( 0,0)@xmath56 ( 1,.1)(1,0)4 ( 0,3.1)@xmath57 ( 1,3.2)(1,0)4 ( 0,3.7)@xmath58 ( 1,3.8)(1,0)4 ( 5.3,3.2)(5.5,3.5)(5.3,3.8 ) ( 6.5,3.5)(7.2,1.8)(6.5,.1 ) ( 5.5,3.4)@xmath59 ( 7.2,1.7)@xmath60 then , the relations among @xmath61 , @xmath62 and @xmath63 should be @xmath64 from eq . ( [ zee_mass ] ) and explicit values of @xmath65 , @xmath66 , @xmath60 and @xmath67 . here we take the low solution to the solar neutrino problem . these relations induce the ratio of @xmath68 and @xmath69 as @xmath70 the phenomenological constraints on @xmath45 from various experimental bounds @xcite are given by @xmath71 where @xmath72 is fermi constant and @xmath73 equations ( [ f_emu ] ) and ( [ other ] ) mean that @xmath74 can not be of @xmath75 unless the charged higgs boson masses are of order 10 tev . notice that the constraints for @xmath76 and @xmath77 in eq . ( [ other ] ) are automatically satisfied as long as the relation in eq . ( [ hi-2 ] ) and the constraint for @xmath61 in eq . ( [ f_emu ] ) are satisfied . actually , the conditions eqs . ( [ hi-2 ] ) and ( [ f_emu ] ) give rise to the following constraint on the smallest coupling @xmath76 ; @xmath78 which is much severer constraint than that of eq . ( [ other ] ) . now let us turn to discuss the cosmological constraint on the zee model . a crucial point is that a linear combination of the lepton flavors , @xmath79 , is approximately conserved in the zee model . this is because the @xmath80-number is violated only through the coupling @xmath63 , which is much smaller than the other couplings , as can be seen in eq . ( [ hi-2 ] ) . ( here , we assign the @xmath80 number of the zee singlet @xmath21 to be zero . ) notice that the sphaleron preserves not only the total @xmath81 but each @xmath82 ( @xmath83 ) . thus , the number @xmath84 is also conserved under the sphaleron process , and is violated only by the tiny coupling @xmath63 . when the temperature @xmath85 is higher than the mass of @xmath21 , @xmath86 , the relevant interactions which violate @xmath80 are ; @xmath87 , @xmath88 , and @xmath89 . we consider the three - body process @xmath90 , since it gives the severest constraint on the coupling @xmath63 . , the four - body process @xmath91 can give a severer constraint on @xmath63 than that of the above three - body process . in this case , we must take care the possibility that the condition @xmath92 might make the physical higgs masses , @xmath93 , be negative . if the condition @xmath94 and @xmath95 is satisfied , we can analyze the out - of - equilibrium condition of this four - body process in a similar way to the three - body case . however , the following conclusion does not change , since the coupling @xmath63 is suppressed by @xmath96 for @xmath97 [ see eq . ( [ f_mutau ] ) ] and hence the out - of - equilibrium condition is satisfied more easily than the three - body case . ] the rate of this process is given by @xmath98 this @xmath80-violating interaction is out of equilibrium if the above rates is slower than the hubble parameter of the expanding universe , @xmath99 . ( @xmath100 gev is the reduced planck scale and @xmath101 is the number of relativistic degrees of freedom . ) namely , the out - of - equilibrium condition is given by @xmath102 on the other hand , as discussed in the previous section , the coupling @xmath63 is very small in order to explain the neutrino oscillation experiments . hereafter , we assume @xmath103 for simplicity . in this case , we could consider @xmath104 . then , from eqs . ( [ zee_mass]),([mass_atm ] ) , and ( [ mass_sol ] ) , the coupling @xmath63 is given by @xmath105 where we take @xmath106 and @xmath107 . here we take the explicit values of @xmath41 and @xmath108 , for example , as @xmath109 and @xmath110 . in this case ( [ ooe - condition ] ) and ( [ f_mutau ] ) show that the @xmath80-violating interactions are really out of equilibrium during @xmath111 . this means that the baryon asymmetry is _ not _ washed out in the case of low solution . this result is not changed , unless @xmath86 is much heavier than @xmath112 or @xmath41 is extremely small . at the lower temperature @xmath113 , the rate of the three - body process is reduced because the number density of the @xmath21 particle is suppressed by a boltzmann factor @xmath114 , and the four - body interactions are also suppressed by a factor @xmath115 . as we have seen in this section , the @xmath80-violating interactions have been out of equilibrium since the birth of our universe . this is because @xmath63 must be small in order to explain the neutrino oscillation experiments . actually , the analysis of chemical potentials in the high - temperature phase gives the following relation between the baryon asymmetry @xmath2 and the number @xmath116 ; @xmath117 therefore the baryon asymmetry in our universe can be preserved , once the asymmetry of @xmath116 is produced in the early universe . recent neutrino experiments indicate that neutrino masses are tiny and there are two mass squared differences as @xmath118 . it is well known that the zee model induces small neutrino masses by radiative corrections , where the lepton number is violated explicitly . if the lepton number violating interaction becomes faster than the hubble parameter before the electroweak phase transition , the baryon asymmetry generated at higher temperature ( in the early universe ) would be washed out . therefore , in order not to destroy the baryon asymmetry , the lepton number violating interaction must be out - of - equilibrium . in this letter we have analyzed the cosmological condition in order for the baryon asymmetry generated in the early universe not to be washed out in the zee model , taking the low solution to the solar neutrino problem . in this case , we find the baryon asymmetry is _ not _ washed out , although it has been said that the zee model can not preserve the baryon asymmetry @xcite . this can be easily shown from the view point of the new lepton number @xmath0 . the lepton number @xmath80 is almost conserved quantity which is violated only by the tiny coupling @xmath76 . from the viewpoint of @xmath80 , the sphaleron process preserves @xmath84 . the @xmath80 violating interaction through @xmath63 is out - of - equilibrium in the zee model when the results from neutrino - oscillation experiments and natural mass scales of higgs masses are used for the input parameters . therefore , we can conclude that once the asymmetry of @xmath119 is produced , the baryon asymmetry is preserved in our universe . we have shown that the baryon asymmetry is preserved against the lepton - number violating interaction in the zee model . however , it is very difficult to explain the origin of the asymmetry of @xmath84 within the zee model . therefore , some mechanism which can generate a @xmath84 asymmetry is necessary . finally , we comment on the cases of other solar neutrino solutions ( lma and vo ) . first , we consider the lma case . it has been pointed out @xcite that the zee model could not explain the lma solution , since the zee model induces a nearly maximal mixing of solar - neutrino oscillation ( @xmath120 ) , which is in poor agreement with the observed data @xcite . if one would still apply the present analysis to the lma solution , the out - of - equilibrium condition of @xmath80 violating interaction is marginally satisfied [ see eqs . ( [ ooe - condition ] ) and ( [ f_mutau ] ) ] , although it requires a more detailed analysis including a numerical solution of the boltzmann equation . the second case is the vo solution . as can be seen in eq . ( [ f_mutau ] ) , the coupling @xmath63 which violates @xmath80 is proportional to the @xmath59 . therefore , in the case of the vo solution , where @xmath59 is much smaller than that of the low solution , the rate of the @xmath80-violating interaction is also small enough , and the baryon asymmetry is preserved as discussed in previous section . we would like to thank t. yanagida for the suggestion of this work . we also thank t. moroi for the collaboration in the early stage of this work , and thank m. tanimoto , y. koide and s. kanemura for useful discussions . nh is supported in part in part by the grant - in - aid for science research , ministry of education , science and culture , japan ( no . 12740146 ) . the work of kh was supported by the japanese society for the promotion of science . y. fukuda _ et al . _ [ kamiokande collaboration ] , phys . rev . * 77 * ( 1996 ) 1683 . y. fukuda _ et al . _ [ kamiokande collaboration ] , phys . b * 335 * ( 1994 ) 237 ; + y. fukuda _ et al . _ [ super - kamiokande collaboration ] , phys . * 81 * ( 1998 ) 1562 [ hep - ex/9807003 ] ; + t. kajita [ super - kamiokande collaboration ] , in _ neutrino physics and astrophysics _ , proceedings of the xviiith international conference on neutrino physics and astrophysics ( neutrino 98 ) , june 4 - 9 , 1998 , takayama , japan , edited by y. suzuki and y. totsuka , ( elsevier science b.v . , amsterdam , 1999 ) page 123 ; nucl . suppl . * 77 * , 123 ( 1999 ) [ hep - ex/9810001 ] . y. fukuda _ et al . _ [ super - kamiokande collaboration ] , + phys . rev . lett . * 81 * ( 1998 ) 1158 [ hep - ex/9805021 ] ; erratum - ibid . * 81 * ( 1998 ) 4279 ; + phys . * 82 * ( 1999 ) 2430 [ hep - ex/9812011 ] ; + phys . * 82 * ( 1999 ) 1810 [ hep - ex/9812009 ] . q. r. ahmad _ et al . _ [ sno collaboration ] , nucl - ex/0106015 . v. barger , d. marfatia and k. whisnant , hep - ph/0106207 . + g. l. fogli , e. lisi , d. montanino and a. palazzo , hep - ph/0106247 ; + j. n. bahcall , m. c. gonzalez - garcia and c. pena - garay , hep - ph/0106258 ; + a. bandyopadhyay , s. choubey , s. goswami and k. kar , hep - ph/0106264 . t. yanagida , _ `` horizontal symmetry and masses of neutrinos '' _ , prog . * 64 * ( 1980 ) 1103 , and in proceedings of the _ `` workshop on the unified theory and the baryon number in the universe '' _ , tsukuba , japan , feb 13 - 14 , 1979 , eds . o. sawada and a. sugamoto , kek report kek-79 - 18 , p. 95 ; + m. gell - mann , p. ramond and r. slansky , _ in _ `` supergravity '' ( north - holland , amsterdam , 1979 ) _ eds . _ d.z . freedman and p. van nieuwenhuizen , print-80 - 0576 ( cern ) .
it is well known that the zee model induces small neutrino masses by radiative corrections , where the bi - maximal flavor mixing is possible . we analyze the cosmological condition in order for the baryon asymmetry generated in the early universe not to be washed out in the zee model . since the lepton number is violated explicitly in the zee model , the baryon asymmetry might be washed out through the sphaleron processes together with the lepton - number violating interactions . in this letter , we will show that the baryon asymmetry is _ not _ washed out , although it has been said that the zee model can not preserve the baryon asymmetry generated in the early universe . this can be seen by considering an approximately conserved number , @xmath0 . hep - ph/0108013 + dpnu-01 - 18 + ut-956 +
it is believed that one of the signatures of a hot , dense medium expected to appear in most central high - energy collisions of heavy ions , will be modification of properties of jets highly collimated streams of particles , originating from hard scattering of partons and therefore produced early in a collision as a result of their interaction with that medium @xcite . unfortunately , direct reconstruction of jets in the presence of a large background typical of such events is an extremely challenging task . one possible answer to this problem is studying two - particle azimuthal correlations , which allow one to extract the jet signal from the soft background by taking advantage of the fact jet particles are strongly correlated in azimuth . in recent years rhic experiments have made this approach highly successful by not only observing the expected signatures , but also providing , through further investigations , evidence which has led to a major change of expectations regarding properties of the hot medium ( the `` perfect liquid '' description ; see _ e.g. _ @xcite ) . however , until recently ( @xcite ) no analyses of this sort were performed in the energy range of the cern sps . the na49 detector is a large - acceptance spectrometer dedicated to the study of hadron production in fixed - target nucleon - nucleon , nucleon - nucleus and nucleus - nucleus collisions at a wide range of energies offered by the cern sps . during its eight years of running it has been used to register pb+pb collisions at 20@xmath1 , 30@xmath1 , 40@xmath1 , 80@xmath1 and 158@xmath1 gev , c+c and si+si collisions at 40@xmath1 and 158@xmath1 gev , along with p+p and p+pb interactions at 158@xmath1 gev . the main components of the detector are four time projection chambers , used for tracking as well as particle identification by de / dx . two of the chambers , the so - called vertex tpcs , are located inside two superconducting magnets ( with field intensity of 1.5 and 1.1 t , respectively ) positioned along the beam axis right downstream of the target , with the other two ( `` main '' tpcs ) placed further downstream on both sides of the beam line . two time of flight walls located beyond the mtpcs complement particle identification in the intersection region of bethe - bloch de / dx bands . centrality selection in nucleus - nucleus collisions is based on the projectile spectator energy deposited in the veto calorimeter , located at the downstream end of the experiment . the present analysis is based on 3 million central pb+pb collisions at 158@xmath1 gev recorded by na49 in the year 2000 . the two - particle correlation function is calculated independently in four event centrality bins : 0 - 5 % , 5 - 10 % , 10 - 15 % , and 15 - 20 % . the transverse momentum bins are @xmath2 for trigger particles and @xmath3 for associates . following the prescription of the phenix collaboration @xcite we define the correlation function as a ratio of two normalised distributions of @xmath4 , where @xmath5 is the azimuthal angle : @xmath6 , in which both particles come from the same event , and @xmath7 , in which the trigger and the associate originate from different events . division by the mixed - event spectrum accounts for non - uniform detector acceptance . @xmath8 in this analysis the mixing is accomplished with a sliding window of up to 50 most recent events in each centrality bin . the resulting correlation functions are shown in the top row of fig . [ fig : c2phi_yield ] . although the correlation function as defined in eq . [ eqn : c2phi ] is free of acceptance effects , it still includes all physical correlations : conservation laws , resonance decays , flow , quantum statistics , parton fragmentation _ etc . _ the _ two - source model _ @xcite , a widely - used description of azimuthal correlations in nucleus - nucleus collisions , postulates that the function @xmath9 can be decomposed into only hard - scattering and flow contributions : @xmath10 $ ] . the two flow terms @xmath11 and @xmath12 have been obtained from an independent reaction - plane analysis @xcite , see table [ tbl : v2values ] for the values . finally , we obtain the factor @xmath13 by employing the zero yield at minimum ( zyam ) assumption @xcite . the flow contribution to the correlation function is illustrated in the top row of fig . [ fig : c2phi_yield ] by the solid black line , with the dashed coloured lines indicating modulation due to statistical uncertainties of @xmath14 and @xmath15 . | r | c | c | centrality & @xmath11 & @xmath12 + 0 - 5 % & 0.022 @xmath16 0.003 & 0.01395 @xmath16 0.00027 + 5 - 10 % & 0.073 @xmath16 0.003 & 0.04079 @xmath16 0.00029 + 10 - 20 % & 0.117 @xmath16 0.003 & 0.06570 @xmath16 0.00026 + having subtracted the flow one can use the hard - scattering component of the correlation function to calculate per - trigger conditional yield of associate particles : @xmath17 where @xmath18 is the number of trigger particles , and @xmath19 the number of same - event trigger - associate pairs . yields obtained this way are shown in the bottom row of fig . [ fig : c2phi_yield ] . gev , for centrality bins 0 - 5 % ( left ) , 5 - 10 % ( middle ) and 10 - 20 % ( right ) . the solid black lines illustrate zyam - normalised flow contribution to the function , with the dashed coloured lines indicating modulation due to statistical uncertainties on flow coefficients . * bottom * : per - trigger conditional yield of associate particles obtained by normalising the flow - subtracted correlation function , again in three centrality bins . all errors are statistical only . ] next , fig . [ fig : chargeselection ] shows correlation functions obtained by imposing additional constraints on the electric charge of trigger and associate particles . like- and unlike - sign pairs have been considered separately and compared to the correlation functions without charge selection , for centrality bins 0 - 5 % and 10 - 20 % . finally , fig . [ fig : expcomparison ] compares the conditional yield of na49 with that from pb+au collisions at 158@xmath1 gev as observed by the ceres experiment at the sps @xcite and that from au+au collisions at @xmath20 = 200 gev as observed by the phenix experiment at rhic @xcite . phenix results have been scaled down by the factor 0.4 to match the amplitude of sps yields near the minimum . all three yields have been obtained for the same centrality bin as well as trigger and associate particle @xmath0 ranges . ) nucleus - nucleus collisions . black full circles : pb+pb at 158@xmath1 gev ( na49 preliminary ) . red open squares : pb+au at 158@xmath1 gev ( ceres preliminary ) . blue open triangles : au+au at @xmath20 = 200 gev ( phenix ) , scaled down to match sps results around the minimum.,scaledwidth=90.0% ] na49 has measured two - particle azimuthal correlation functions in central and mid - central pb+pb collisions at 158@xmath1 gev . the away side of the correlation function in the most central collisions shows plateau - like structure , visible even before subtraction of flow and significantly different from the away - side shape in more peripheral bins . such behaviour is consistent with qualitative expectations for a hot , dense medium in the most central high - energy nucleus - nucleus interactions . moreover , na49 results agree with those from the ceres experiment at the sps and are in qualitative agreement with rhic results . finally , the observed difference in amplitude of the near - side peak of the correlation function of like- and unlike - sign pairs , is consistent with local charge conservation which could be associated with fragmentation of partons . wang , m. gyulassy , phys . lett . * 68 * , 1480 - 1483 ( 1992 ) . j. adams _ et al . _ ( star collaboration ) , phys . lett . * 95 * , 152301 ( 2005 ) . c. zhang for the phenix collaboration , j. phys . g * 34 * , s671-s674 ( 2007 ) . m. ploskon for the ceres collaboration , acta phys . hung . * a27 * , 255 - 258 ( 2006 ) . s. afanasev _ et al . _ ( na49 collaboration ) , nucl . instrum . * a430 * , 210 - 244 ( 1999 ) . _ ( phenix collaboration ) , phys . lett . * 97 * , 052301 ( 2006 ) . ajitanand , j.m . alexander , p. chung , w.g . holzmann , m. issah , r.a . lacey , a. shevel , a. taranenko , p. danielewicz , phys . rev . * c72 * , 011902 ( 2005 ) . a.m. poskanzer , s.a . voloshin , phys . rev . * c58 * , 1671 - 1678 ( 1998 ) . s. kniege for the ceres collaboration , ismd2007 talk .
1.0 cm 1.0 cm two - particle azimuthal correlations of high-@xmath0 hadrons can serve as a probe of interactions of partons with the dense medium produced in high - energy heavy - ion collisions . first na49 results on such correlations are presented for central and mid - central pb+pb collisions at 158@xmath1 gev beam energy , for different centrality bins and charge combinations of trigger and associate particles . these results feature a flattened away - side peak in the most central collisions , which is consistent with expectations of the medium - interaction scenario . a comparison with ceres pb+au results at the same energy , as well as with phenix au+au results at the top rhic energy , is provided .
since the launch in march 2009 , the _ kepler _ mission has discovered a few thousand planetary candidates , called kepler objects of interests ( kois ) , by detecting the flux deficit as a planet transits in front of its star @xcite . while some of the stars are observed to have one transiting planet ( called tranet from now on , following * ? ? ? * ) , others show up to 6 @xcite . a natural question to ask is , do all of these systems share the same intrinsic orbital structures ? for observing transiting planets , the two most relevant orbital parameters are the dispersion in orbital inclinations , and the typical spacing between adjacent planets . a number of groups have studied the inclination dispersion of kepler planets and reached the common conclusion that this must be small and is of order a few degrees @xcite . however , it has been pointed out that models with a single inclination dispersion falls short in explaining the number of single tranets relative to higher multiples , by a factor of three or more @xcite . this suggests that all kepler planets are not the same , and motivates models where the inclination dispersion itself is broadly distributed ( rayleigh of rayleigh , * ? ? ? * ; * ? ? ? * ) . however , the relative occurrences of different kepler multiples ( denoted here as 1p , 2p , 3p ... by the number of tranets seen in a system ) are sensitive to both the inclination dispersion and the intrinsic planet spacing . larger spacing between adjacent planets will raise the relative number of single tranet systems , as so will larger inclination dispersion . it is difficult to disentangle the two without the aid of further information . therefore we turn to a new measure , the ttv fraction . if a tranet is accompanied by another planet , its transit times deviate from strict periodicity ( transit - time - variation , ttv , * ? ? ? * ; * ? ? ? many studies have used ttv to confirm the planetary nature of kepler candidates . furthermore , it is realized that when the companion is near a mean - motion resonance ( mmr ) with the tranet , the ttv is particularly strong and exhibits a characteristic sinusoidal form @xcite . the amplitude and phase of this sinusoid have been simply related to the perturber s mass , as well as the orbital eccentricities @xcite , thereby allowing us to infer the interior composition and orbital parameters of these objects @xcite . just as the ttv signal can be used to infer the presence of unseen ( non - transiting ) companions around specific candidates ( e.g. * ? ? ? * ; * ? ? ? * ) , the number of tranets that exhibit sinusoidal ttvs provides constraints on near - mmr companions . since the period ratios of adjacent kepler pairs do not much prefer mmrs @xcite , these near - mmr companions can be taken as a proxy for companions that lie close to and inward of the 2:1 mmr . to be quantitative , we shall define the `` intrinsic ttv fraction '' as half the probability that a planet induces a sufficiently large ttv amplitude for detection in another planet in the system . the reason for the factor of a half is that when one planet has a large ttv , then typically so does its ttv partner , and we do not wish to double - count such a pair . when trying to measure this quantity observationally , we shall first count the number of observed tranets with measured ttv s , and then subtract one each time two tranets are ttv partners . dividing by the total number of tranets yields the `` measured ttv fraction , '' which is our estimate for the intrinsic fraction . our ability to measure ttv is affected by the noise level in the transit signals , which is in turn determined by a range of parameters including stellar brightness , the size of the planet relative to its host star , the orbital period and the transit duration . however , if we split the planet candidates into different groups , and if these groups share the same noise properties , then one can argue that the relative ttv fractions measured for different groups represent the relative differences in their intrinsic ttv fractions . in the following , we proceed to measure the relative ttv fractions among 1p , 2p , 3p and 4p+ systems , where 4p+ stands for systems that have four or more transiting planets . we carry out the analysis for all kois that have suitable light - curves , which include more than 2600 kois . we interpret the significance of our results in [ sec : discussion ] . we use the publicly available q0-q12 long cadence ( lc , pdc ) data for 2740 kois ( kepler objects of interest , * ? ? ? * ) . out of these , 134 kois have fewer than 7 transit time measurements , either because the transit periods are very long , or the signal - to - noise ratios ( snr ) are too small . these spread evenly across all multiplicities . this leaves us with 2606 kois , out of which there are , 1488 , 571 , 320 , 227 systems that are designated as 1p , 2p , 3p and 4p+ , respectively . we refer to this sample of 2606 kois as the ` full ' sample . we have also selected a ` reduced ' sample by excluding those kois for which timing measurements are less accurate . these include those with large noise ( @xmath0 ) , and short transit duration ( less than an hour ) . we include only kois with intermediate planet sizes ( @xmath1 ) as they likely have the lowest false positive rate @xcite . this reduced sample contains a total of 1989 kois , with 1097 , 446 , 253 , and 193 systems designated as 1p , 2p , 3p , and 4p+ , respectively . the pipeline used to measure the transit times has been developed and described in @xcite . we compare our transit time measurements to the published ones @xcite , and found good consistency ( see , e.g. , fig.[fig : ttv ] ) . our measurements for the ttv candidates are publicly available at http://www.astro.utoronto.ca/@xmath2jwxie/ttv . from the above transit time measurements , we derive ttv , which are the residuals after a best linear fit . we then search for a sinusoidal signal by obtaining a lomb - scargle ( ls ) periodogram @xcite on these residuals and identify the highest peak that has a period longer than twice the orbital period , as well as @xmath3 days . the former threshold comes about because twice the orbital period is the nyquist frequency for sampling ttv . the latter threshold is enforced because ttv at shorter periods can be significantly polluted by noise from chromospheric activities , as stellar rotation periods fall typically in the range from a few to a few tens of days @xcite . sinusoidal ttv caused by a perturber near a mmr has a `` super - period '' @xcite , @xmath4 where @xmath5 and @xmath6 are the orbital periods of the two planets that are near a first - order @xmath7 mmr by a fractional distance @xmath8 . for a planet pair with period @xmath9 days , @xmath10 , @xmath11 , we find @xmath12 days . a planet pair with a larger @xmath13 will have a shorter super - period , however their ttvs also become increasingly difficult to detect as the ttv amplitudes scale inversely with @xmath13 @xcite . all reported cases @xcite have @xmath13 falling between @xmath14 . this consideration , coupled with the above concern for chromospheric noise , leads us to discard sinusoids short - ward of 100 days . we also adopt an upper limit of @xmath15 days . this comes about because the data ( q0-q12 ) stretch only @xmath16 days . however , some ttv systems show strong , identifiable sinusoids even before a full ttv cycle is observed . this constraint is later relaxed and is found not to impact the conclusion . many of the sinusoids thus identified are false , caused by random alignment of noisy data . it is an important task to exclude these . we adopt the following strategy from @xcite , originally applied to detect planets from radial velocity data . for each koi , we scramble the time stamps of the original ttv data for @xmath17 times , perform a ls periodogram analysis on each set of data , obtaining the amplitude and frequency on the highest peak . the fap ( false alarm probability ) of the original ttv peak is estimated as the fraction of permutations that have higher sinusoids than the original ttv . we assign an fap value of @xmath18 if not a single random realization exceeds the observed sinusoid amplitude . our fap estimates compare well with those from @xcite . fig.[fig : pttv ] shows the fap and @xmath19 for each of the kois in our full sample , as well as those using a scrambled time series for the same kois . the latter set is equivalent to random noise and so acts as a control sample . the true data show a significant excess of objects at very low faps , when compared to those of the scrambled data . we adopt the following ` standard ' criterion ( the region labelled as ` s ' in fig . [ fig : pttv ] ) for identifying our ttv candidates : * \(1 ) @xmath20 and * \(2 ) ttv period between 100 and 1000 d. objects that have fap @xmath21 exhibit ttv amplitudes that range from one to hundreds of minutes , with ttv sensitivity higher for larger snr objects . our above fap criterion is on the conservative side : for an fap of @xmath22 , there should only be @xmath23 false positives among our ttv candidates , much fewer than the actual number of candidates ( @xmath24 ) . we experiment by relaxing the above criterion , either by raising the fap threshold to @xmath25 ( adding the ` f ' region in fig . [ fig : pttv ] ) , or by removing the 1000 day upper limit ( adding the ` p ' region ) . we report our results below . [ cols="^,^,^,^,^,^,^,^,^,^,^ " , ] @xmath26 see [ sec : method ] for definitions of various samples and ttv thresholds . the ` reduced ' sample is selected based on snr , transit duration , and luminosity ; the s , p , and f criteria are illustrated in fig . [ fig : pttv ] . + @xmath27 number of identified ttv candidates , versus numbers of kois in that category . the numbers in parentheses are the raw ttv candidates , while the the corrected ones after parentheses result from removing one of the two ttv candidates from the count whenever a ttv pair is seen . + @xmath28 the measured ttv fraction using the corrected ttv count . we observe a remarkable rise of the ttv fraction with transit multiplicity . in table 1 , we list the number of ttv candidates for different transiting multiplicities , for four different combinations of sample and ttv selection criteria . for ease of comparison against theory , we list the `` measured '' ttv fractions , obtained by removing one candidate from the raw count whenever both it and its ttv partner have observed ttvs . such a method is justified in [ sec : discussion ] . from now on , we focus on these `` measured '' fractions ( illustrated in fig . [ fig : frac ] ) in fact , we focus on the relative `` measured '' fractions , the ttv fractions normalized by that in 4p+ systems . the choice for the normalization is arbitrary . however , since the errorbars for these relative fractions are taken to be quadratic sums of the individual errorbars , which type of system one normalizes against does not affect the statistical conclusion . these results are presented in fig . [ fig : frac ] . except for case 3 , all other combinations give very similar results for the relative ttv fractions : 1p systems have about five times lower ( values from case 1 : @xmath29 ) ttv fraction than 4p+ , and 2p and 3p systems are about twice lower ( @xmath30 , and @xmath31 ) . results from case 3 are less reliable as the ttv selection criterion is too relaxed and allows for too many false positives . we have also used the ttv data from @xcite , published while we are editing our final draft , to confirm the above results ( fig . [ fig : frac ] ) . , we plot ttv fractions from their catalog ( their table 3 ) after applying our ` s ' selection criterion . the good agreement is encouraging , as we extract ttvs using a different method . ] we first discuss what potential bias may affect the absolute ttv fractions that we obtain , then move on to discuss biases that may affect the relative ttv fractions among groups of different transit multiplicity . it becomes clear that by focusing only on the relative ttv fractions , we can eliminate many , if not all , observational bias . we compare properties of the set of ttv candidates against the koi sample . the top panels of fig . [ fig_dis_ttv ] display four transit properties : signal - to - noise ratio , planet radius , orbital period and transit duration . the ttv sample have in general larger transit snr , larger planet radii ( disfavouring small planets ) and slightly longer transit durations than the average kois . in addition , they are also concentrated around orbital periods @xmath32 days . these characteristics allow for optimum ttv detections , as is demonstrated in recent ttv studies by @xcite . for instance , longer orbital periods generally lead to larger ttv amplitudes @xcite , yet too long orbital periods permit only a small number of transits to be observed . as such , we expect that the intrinsic ttv fraction , quantified as half the fraction of planets that have comparable ttv amplitudes as the ones detected here , should be higher than our reported values ( table 1 ) . on the other hand , we find no significant difference between the ttv and koi samples in terms of stellar mass , effective temperature , metallicity and stellar brightness ( stellar parameters from * ? ? ? ks tests performed to compare these two populations always return p - values greater than @xmath33 . this suggests that ttv candidates live in all possible systems . however , this deserves further study as currently there are large uncertainties in stellar parameters , and our ttv sample is relatively small . while the ttv sample as a whole are a biased representation of the koi sample , we find that the different sub - samples , separated by their transiting multiplicity , share similar distributions in both the transit parameters ( fig.[fig_dis_ttv ] ) and the stellar parameters ( not shown here ) . the large p - values returned from ks tests ( fig.[fig_dis_ttv ] ) do not support the hypothesis that the different subgroups experience different selection effects . moreover , we have confirmed that our reduced koi samples , when separated into groups of different transit multiplicities , are statistically similar in their transit and stellar parameters . since the ability to detect ttv above a certain threshold amplitude , only depends on these transit and stellar parameters , these two results then argue that the relative measured ttv fractions reflect the relative intrinsic ttv fractions . in other words , the significant correlation between ttv fraction and transit multiplicity that we observe ( fig.3 ) is unlikely to be caused by systematic biases on stellar / transit parameters . another potential bias could arise during transit detection transiting planets with significant ttvs can be systematically missed , or cataloged as false positives by the kepler pipeline @xcite . to remove this bias , @xcite designed an algorithm ( qats ) that can simultaneously detect transits and measure their ttvs . searches using qats have only found a handful of new planetary candidates ( private communication j. a. carter ) , which might indicate that the kepler catalogue is not significant impacted by this bias . nevertheless , we caution that there could be another possibility , namely , the qats could not fully remove the bias , which deserves further study but is out of the scope of this paper . last but not least , we note that the transit multiplicity of a given system is evolving as the catalog updates . for example , a 1p system may become a 2p system when a new transit candidate is found , either due to accumulation of new data and/or improvement of the pipeline / algorithm for planet detection . to see how these factors affect our results , we took an older version koi catalog from @xcite and performed the same analysis as done in the standard case ( case 1 in table 1 ) . we obtained similar results for the relative ttv fractions : ( @xmath34 , ( @xmath35 , ( @xmath36 and ( @xmath37 for the 1p , 2p , 3p and 4p systems , respectively . this suggests that the correlation between ttv fraction and transit multiplicity observed in fig . 3 may remain unaffected as more improved catalogues are published . , satisfies a rayleigh distribution @xmath38 , and where the planets mutual inclinations are described by an independent rayleigh distribution with @xmath39 . the results presented here are insensitive to the value of @xmath40 . we identify ttv planets as those that transit their host stars and have a companion within a fractional distance of @xmath41 from a first - order mmr . the solid lines indicate the measured fractions , while the dotted lines are the raw fractions ( i.e. , the fraction before removing doubly - counted ttv pairs ) , plotted as functions of @xmath42 , for different multiplicity groups . the raw fractions in the 1p systems can drop to half of that in the 4p systems , because one is likely to observe both planets in the same ttv pair in the latter case . in contrast , in the corrected form ( our so - called measured ttv fraction ) , the relative ttv fractions remain close to unity , and are largely independent of either transit multiplicity or model parameters . in contrast , the observed relative fractions ( case 1 in fig . [ fig : frac ] , marked here as arrows ) fall much below unity . , scaledwidth=45.0% ] we discuss the significance of our results by contrasting them against predictions from a simple toy model . assume all kois , independent of transit multiplicity , are drawn from the same intrinsic distribution , with similar dispersions in mutual inclinations and planet spacing ( with no preference for mmrs ) . in this case , single systems are the ones where the viewing angles are less favourable and we miss most of the planets in the system , while the higher multiples are ones where more planets are caught . one can estimate the ttv fraction for the theoretical population as half the fraction of planets that both transit and have companions within a certain distance from a first - order mmr . as one naively expects and as is confirmed by monte carlo simulations ( fig . [ fig : ttv_toymodel ] ) , the relative ttv fractions cluster around @xmath43 , and are largely independent of the model parameters and transit multiplicities . the lower ttv fraction observed for singles is unlikely to be completely explained by the higher false positive rates in koi singles . the reported false positive rate is of order @xmath44 @xcite . more importantly , our reduced sample , which is expected to have a lower false positive rate than the full sample , yields the same relative ttv fractions . moreover , since ttv amplitudes are strongly boosted by eccentricities as small as a few percent @xcite , the lower ttv fraction can be explained if higher multiple systems have higher eccentricities . however , this is likely excluded by the tight spacing observed among high multiples . a simple explanation for our results is that the basic assumption in our toy model is not true , namely , all kois can not be treated as the same intrinsic population @xcite . for example , there could be at least two distinct populations of kepler planets , different in their intrinsic frequencies of close companions . the high multiples ( 4p+ ) are dominated by a population that has a higher companion frequency , while the 1p systems may be dominated by a population that have a lower frequency of close companions . in other words , there are at least two populations of kepler planets , one that are closely spaced , and one that is sparsely spaced . in an upcoming publication , we will use ttv fractions obtained in this paper , together with a variety of other observational facts , to constrain the properties of these two populations of kepler planets . this will yield important constraints on the process of planet formation . , c. j. , bryson , s. , christiansen , j. , mullally , f. , rowe , j. , science office , k. , & kepler science team . 2013 , in american astronomical society meeting abstracts , vol . 221 , american astronomical society meeting abstracts , 216.02
a transiting planet exhibits sinusoidal transit - time - variations ( ttvs ) if perturbed by a companion near a mean - motion - resonance ( mmr ) . we search for sinusoidal ttvs in more than 2600 kepler candidates , using the publicly available kepler light - curves ( q0-q12 ) . we find that the ttv fractions rise strikingly with the transit multiplicity . systems where four or more planets transit enjoy four roughly five times higher ttv fraction than those where a single planet transits , and about twice higher than those for doubles and triples . in contrast , models in which all transiting planets arise from similar dynamical configurations predict comparable ttv fractions among these different systems . one simple explanation for our results is that there are at least two different classes of kepler systems , one closely packed and one more sparsely populated .
adaptive mesh refinement(amr ) cosmological hydrodynamics simulations play an important role in the study of structure formation of different scales from galaxy clusters to first stars in the past ten years . its ability to achieve very high resolution in large scale simulations with relatively small computer resources has helped us to understand the first stars formed in our universe . the possible effects of magnetic fields have been largely ignored . it is well established that magnetic fields are present on different scales , from intracluster medium to interstellar medium . the origin and evolution of these magnetic fields and their role on the structure formation are still unclear . so , it is imperative to include magnetic fields into current hydrodynamics amr cosmology code . in this paper , we present the newly developed mhd version of the enzo , which is wildly used in the study of first stars @xcite . the mhd equations in the comoving coordinates are : @xmath0 & = & - \frac{\dot{a}}{a}(\rho v^{2}+3p+\frac{b^{2}}{2 } ) \nonumber \\ & & - \frac{\rho}{a } \bf{v}\cdot \nabla \phi \\ \frac{\partial \bf{b}}{\partial t } - \frac{1}{a } \nabla \times ( \bf{v } \times \bf{b } ) & = & -\frac{\dot{a}}{2a}\bf{b}\end{aligned}\ ] ] with @xmath1 where all variables have their usual meaning , a is the expansion parameter . to track the pressure more accurately in the supersonic region , we have also implemented the modified entropy equation given in @xcite and the internal energy equation given in @xcite in our code . the mhd solver used for all the test problems here is a high - order godunov - type finite - volume numerical solver developed by s.t . li @xcite . this solver was recently successfully used to study magnetic jet problems @xcite . we used a constrained transport(ct ) scheme flux ct @xcite to maintain divergence - free magnetic fields . for the amr hierarchy in enzo , we used a modified divergence - free reconstruction scheme original proposed by @xcite including second order accurate divergence - free restriction and prolongation for magnetic fields . details of the ct and amr of magnetic fields in the mhd enzo can be found in @xcite . the mhd module has been tested and shown to be compatible with other physics packages installed in enzo , such as radiative cooling , star formation and feedback . the mhd caustic test is taken from @xcite which generalizes the hd test of @xcite . the initial sinusoidal velocity field in the x - direction has the peak value @xmath2 , the initial density and pressure has been set to be uniform with @xmath3 and @xmath4 . then caustics will be formed because of the compression by the velocity field . an initial uniform magnetic field of @xmath5 in code units in the y direction was added to the simulation . figure [ fig : caustic ] compares the density and @xmath6 at @xmath7 of unigrid and amr runs . amr run had 256 cells in the root grid and 2 level refinements by 2 . the amr solution is indistinguishable from a uniform grid solution with 1024 cells . pancake is another standard test problem of cosmological hydrodynamics simulation @xcite . we have run the collapse of a one - dimension pancake in a purely baryonic universe with @xmath8 and @xmath9 . initially , at @xmath10 , which corresponds to @xmath11 in this test , a sinusoidal velocity field with the peak value @xmath12 in the normalized unit has been imposed in a box with the comoving size @xmath13 , so that the shock forms at @xmath14 . the initial baryonic density and pressure have been set to be uniform with @xmath3 and @xmath15 in the normalized code units . we applied initial uniform magnetic fields @xmath16 in the simulation . we did the calculations with unigrid with 1024 cells and amr run with 256 cells of root grid and 2 level refinements . figure [ fig : pancake ] shows the density and @xmath6 at @xmath17 . .,title="fig : " ] .,title="fig : " ] , title="fig : " ] , title="fig : " ] cluster formation is one of the problems most widely studied by enzo @xcite . we have done this problem and compared with results from enzo - ppm to test our new code in large scale structure formation . the simulation is a @xmath18cdm model with parameters @xmath19 , @xmath20 , @xmath21 , @xmath22 , @xmath23 . the survey volume is 256 @xmath24 mpc on a side . the simulations were computed from a @xmath25 root grid and 2 level nested static grids in the lagrangian region where the cluster forms which gives an effective root grid resolution of @xmath26 cells ( 0.5 @xmath24 mpc ) and dark matter particles of mass @xmath27 . amr is allowed only in the region where the galaxy cluster forms , with a total of 8 levels of refinement beyond the root grid , for a maximum spatial resolution of 7.8125 @xmath24kpc . we first present the results of no initial magnetic fields and compare them with the results from enzo - ppm . the cluster parameters from the mhd code is almost identical to those from enzo - ppm . the virial radii are 2.229mpc from hydro and 2.226mpc from mhd while the virial masses are @xmath28 for hydro and @xmath29 for mhd . figure [ fig : cluster2 ] compare the projections of the baryon density and temperature . of adiabatic simulations . the images cover the inner 4 @xmath24mpc of cluster centers . the left panels show results from the ppm solver and the right panels show results from the mhd solver . the units of density and temperature are @xmath30 and kelvin respectively . ] we have performed simulation with initial magnetic fields , @xmath31 , @xmath32 with radiative cooling , star formation and star formation feedback . figure [ fig : cluster4 ] shows the images of baryon density , temperature , magnetic energy density and faraday rotation measurement of the cluster center . the rotation measurement is integral along the projection direction . ) , temperature ( logarithmic , kelvin ) , magnetic energy density ( logarithmic , @xmath33 ) and rotation measurement ( @xmath34 ) of the inner 1@xmath24mpc of cluster center . the initial magnetic field is @xmath32 at @xmath17 . ] another simulation we performed is without initial magnetic field but with the biermann battery effect . the induction equation is modified by adding an additional battery term@xcite : @xmath35 where c is speed of light , @xmath36 is the pressure of electron , @xmath37 is the electron number density , e is the electron charge , @xmath38 is the hydrogen mass and @xmath39 is the ionization fraction . we took @xmath40 constant in space in our simulation . we performed this computation with radiative cooling . figure [ fig : cluster5 ] shows the projection of logarithmic baryon density and the magnetic energy density of the cluster center . ) and magnetic energy density ( logarithmic , @xmath33 ) of the simulation with biermann battery effect of the inner 1@xmath24mpc of cluster center . ] we have introduced our mhd module for enzo and presented some test results . with this new module , we have the ability to perform amr mhd cosmology simulations . this module uses high accuracy tvd mhd solver and employs ct and amr divergence - free magnetic fields reconstruction scheme to guarantee divergence - free of magnetic fields . mhd simulations using this mhd module can use all the exist physics packages in enzo . it is widely believed that the magnetic fields should have little effects in the formation of the first stars , since maybe there were no magnetic fields at all . but even there are no magnetic fields from the early universe , the biermann battery effect should have generated some very small fields during the collapse to form those stars . if the first generation stars spin very fast , these small seed fields could be maintained and amplified by dynamo effect . then the magnetic fields in the first stars may pollute the environment by their explosion and play a part in the formation of next generation stars . this would be investigated in the near future .
magnetic fields play an important role in almost all astrophysical phenomena including star formation . but due to the difficulty in analytic modeling and observation , magnetic fields are still poorly studied and numerical simulation has become a major tool . we have implemented a cosmological magnetohydrodynamics package for enzo which is an amr hydrodynamics code designed to simulate structure formation . we use the tvd solver developed by s. li as the base solver . in addition , we employ the constrained transport ( ct ) algorithm as described by d. balsara . for interpolation magnetic fields to fine grids we used a divergence free quadratic reconstruction , also described by balsara . we present results from several test problems including mhd caustics , mhd pancake and galaxy cluster formation with magnetic fields . we also discuss possible applications of our amr mhd code to first star research . address = department of physics and cass , university of california , san diego , san diego , ca , altaddress = theoretical astrophysics , los alamos national laboratory , los alamos , nm address = department of physics and cass , university of california , san diego , san diego , ca address = department of physics and cass , university of california , san diego , san diego , ca address = mathematical modeling and analysis , los alamos national laboratory , los alamos , nm address = theoretical astrophysics , los alamos national laboratory , los alamos , nm
pattern speed ( @xmath6 ) is defined as an angular rotation velocity of spiral pattern . according to the spiral density wave theory @xcite , @xmath6 determines the location and even the existence of resonances , which would greatly affect the morphology and kinematics of a galaxy . the results of numerical simulations also show the large dependence of spiral structure on @xmath6 @xcite . despite its importance on the study of galaxies , it can not be determined directly from observations , since the pattern structure is not a material but a density wave propagating through the disk . several methods have been proposed to date for the determination of pattern speed . @xcite presented the method using the continuity equation for the surface brightness of galaxies . @xcite showed that the residual velocity field , obtained by subtracting the axisymmetric component from observed velocity field , has a difference in pattern between the inside and outside of the corotation . @xcite pointed out that where the arm to interarm ratio of star formation efficiency has a sharp dip could be thought as the corotation . these methods work well for some galaxies , but there are still inconsistencies or large uncertainties . in this letter , we examined offsets between h@xmath2 and co arms of a spiral galaxy . we suggest that we can use them as a new method for the determination of pattern speed and typical timescale for star formation . the h@xmath2 emission is the tracer of hii region , which surrounds newly born massive stars , while the co emission is the tracer of molecular gas . according to the galactic shock wave theory @xcite , most molecular clouds are formed and then condensed by the compression owing to the spiral shock wave and thus belong to the spiral arm . since galactic scale star formations take place in these molecular arms , some molecular gases are used to form stars , and the others are dissociated by uv emission from massive stars born in that star formation . in consequence , molecular clouds disappear soon after their formation and their position appears to be fixed to the spiral pattern , though individual molecular clouds move at the speed of gas . therefore , the position of spiral pattern and h@xmath2 arm can be measured by using that of a molecular arm and an ensemble of hii regions respectively . in other words , offset between the arms of h@xmath2 and co represents the difference between the speed of gas and spiral pattern , as well as the time needed for star formation from molecular clouds . this offset has been found in many spiral galaxies ( eg . @xcite ) . we can easily see it in a @xmath11-band image as a displacement of an arm s peak and a dust lane , since the light of @xmath11-band is mainly emitted by the young ob star and the distribution of dust lane and molecular gas are almost the same . however , this offset has not been studied quantitatively yet , since the resolution of the co image was too large to examine detail structures of spiral arms . we assume that the pattern of spiral is rigid , and that the gas rotates in circular orbit . then we define a timescale @xmath7 as an _ average _ time for h@xmath2 arm to develop from a galactic - shock - compressed molecular arm . if the physical processes of star formation are not extremely different in the spiral disk , this timescale can be regarded as a constant parameter representing a typical value of the entire disk . ( 80mm,80mm)draw.ps figure [ fig.draw ] illustrates this idea for the inside of the cr . if we observe a face - on spiral galaxy at @xmath12 ( the left panel ) , at @xmath13 the same galaxy will be observed as the right panel and the offset distance between the arms of h@xmath2 and co , @xmath14 , can be written as @xmath15 } , \label{eq.d}\ ] ] where @xmath16 is the velocity of gas , and @xmath17 is the velocity of pattern . we adopt this expression , in which @xmath14 becomes a positive value , since most of molecules exist in a rather central part of the disk and thus they seem to be in the inside of the cr . if molecular gas is more extended and the outside of the cr should be taken into account , h@xmath2 arm will be seen on the concave side of co arm and @xmath14 will be negative . dividing both sides of equation ( [ eq.d ] ) by radius @xmath18 [ kpc ] , we obtain @xmath19 } , \label{eq.ofst}\end{aligned}\ ] ] where @xmath20 , @xmath21 , and @xmath3 is the azimuthal offset . this equation shows the relation between two observables , @xmath4 and @xmath3 , and we can rewrite it as @xmath22 where @xmath3 is in degree and @xmath7 is in @xmath23 yr . if we assume that @xmath7 is constant in a certain range of radius in a galaxy , @xmath3 becomes a linear function of @xmath4 , since we assume the rigid pattern , in other words , constant @xmath6 . therefore , by plotting @xmath3 against @xmath4 and fitting them with a line , both @xmath6 and @xmath7 can be determined at the same time . in addition , the linearity of this plot can verify the validity of the assumptions we made . we discuss this validity in section 4 . we observed a sac galaxy ngc 4254 in the @xmath0co ( @xmath1=1 - 0 ) line using nobeyama millimeter array during a long - term project , `` virgo high resolution co survey '' @xcite . the spatial resolution is about 2 arcsec , which corresponds to @xmath24 pc at a distance from the virgo cluster of 16.1 mpc . this resolution is small enough to trace the spiral arms with typical width of 1 kpc . for more detailed information about this galaxy and the co observation , see @xcite . @xcite presented broadband _ r _ and narrowband h@xmath2 images of 63 spiral galaxies in the virgo cluster . the resolution of the image is 1.2 arcsec , comparable to that of our co data . we determined the coordinate of the h@xmath2 image using the galactic stars seen in the _ r _ band image , since these two images had the same field of view . the typical uncertainty of this coordinate fitting is about 1 arcsec , smaller than the resolution of the images . ( 80mm,80mm)ds9.ps we show an overlaid image of co and h@xmath2 in figure [ fig.coha ] . as we see in this figure , it is nearly face - on ( we derived an inclination angle as 34@xmath25 for central region . see the following paragraph for detail . ) and has well - ordered spiral arms ( classified as arm class 9 by @xcite ) , so that the spiral structure can be easily traced . this is why we selected this galaxy to examine offsets between h@xmath2 and co arms . position angle ( p.a . ) and inclination ( _ i _ ) are also important parameters of a galaxy . we determined them applying the task gal of aips to the velocity field of co. this task cuts annuli out from a velocity field , and fits them with a pure circular rotation velocity field in order to derive p.a . and _ i _ at each radius . we averaged the derived values at @xmath26 , corresponding to 0.4 kpc @xmath27 1.2 kpc , where the values do not change largely , and then obtained p.a . = @xmath28 and @xmath29 . in table [ tb.pa_i ] , these p.a . and _ i _ are listed with values from other papers . .position angle and inclination of ngc 4254[tb.pa_i ] [ cols="^,^,^",options="header " , ] although only the central region was used in this work , the determined values are not so different from previous work , in most of which the whole image of the galaxy was used for the determination . using the obtained p.a . and _ i _ , we calculated the rotation velocity of gas by the use of gal again . ( 60mm,60mm)f2.eps the resultant rotation curve ( figure [ fig.rc ] ) is not perfectly the same with that derived by the iteration method @xcite , since they used the different value of p.a . and _ i_. these p.a . and _ i _ were also used to deproject the image into a face - on view and to transform it into a polar - coordinate image , a phase diagram ( figure [ fig.phase ] ) . ( 90mm,90mm)phase.ps azimuth is taken to be zero at the nw side of minor axis and to increase clockwise , in the same direction of this galaxy s rotation assuming the trailing arms . a range between the two horizontal lines in this figure ( 1 kpc @xmath27 2.6 kpc ) show the region used in the following analysis . filled boxes and circles are the peaks of h@xmath2 and co intensity , respectively . see the following section for detail . we divided the phase diagram into strips of 200 pc radial width , averaged the intensity with respect to radius at each azimuth in each strip , and plotted the averaged intensity against azimuth . we defined an offset angle @xmath3 as an azimuthal angular separation of intensity peaks between the h@xmath2 and co arms . since arms of this galaxy are not totally continuous and have some breaks , we could not find corresponding peaks at some radii . in addition , we only used the range of 1 kpc @xmath27 2.6 kpc , since the spiral feature is not clear in the inner region and the rotation curve has a large dispersion in the outer region ( see figure [ fig.rc ] ) . in figure [ fig.phase ] , this range is indicated by two horizontal lines . the peak of co and h@xmath2 arm is marked by a filled and labeled circle and box , respectively , with a label from a to j , corresponding to that in figure [ fig.chi_fit ] . the largest factor for the error in @xmath3 is the resolution of co data , which is about 2 arcsec . thus this error is written as @xmath30}. \label{eq.delta}\ ] ] in addition , the uncertainty of the coordinate fitting of h@xmath2 image and the fact that the determination of offset angles is somewhat subjective would also be possible sources of error in @xmath3 . we , however , neglected these errors for the simplicity of calculations , so that the error in @xmath3 could be larger than the value obtained from equation ( [ eq.delta ] ) by a factor of 2 at most . we plotted the derived offsets against @xmath4 , derived from rotation curve ( figure [ fig.chi_fit ] ) . ( 80mm,80mm)fit.eps errorbars in offset angle are @xmath31 , calculated from equation ( [ eq.delta ] ) at each radius . we used the @xmath32-square fitting method and obtained @xmath9 yr and @xmath10 . the dashed line in figure [ fig.chi_fit ] is the fitted line , and the gradient and horizontal - axis - intercept of this line corresponds to the resultant value of @xmath7 and @xmath6 , respectively . @xcite showed that the corotation radius for the same galaxy is about 7.5 kpc from numerical simulations , and corresponding @xmath6 is about 20 @xmath33 . they assumed the distance to the galaxy as 20 mpc . if we adjust this result to the distance of 16.1 mpc , @xmath6 should be about 25 @xmath33 . the typical timescale of star formation has been suggested as about @xmath23 yr from the calculation of jeans time in molecular clouds . our results are in good agreement with these previous studies . in our method , we assumed a pure circular rotation and a constant time delay for star formation from molecular clouds , and we did not do any correction for extinction in h@xmath2 data . we however know that deviations from these assumptions are not critically large . the streaming motion and the velocity dispersion would generate some non - circular motion , but in a spiral disk , these are usually about 10 km s@xmath34 @xcite , which is very small compared to the circular rotation velocity of @xmath35 km s@xmath34 . the validity of the latter assumption , constant @xmath7 , can be checked by the linearity of @xmath36 plot ( figure [ fig.chi_fit ] ) . since this plot is well fitted with a line , dependence of @xmath7 on the environment should be small and included in the error of the derived value . the extinction of h@xmath2 is surely significant to luminosity and this might lead us to overestimate the offset values , since the amount of extinction becomes larger as hii regions get closer to molecular arms and an h@xmath2 arm thus seems to shift to the downstream side from where it really is . @xcite showed azimuthal brightness profiles of ngc 4254 in @xmath37- , @xmath38- , and @xmath39-band image at r=@xmath40 ( figure 20 in their paper ) . though the shape of profile is different from one to another , the positions of peak brightness are well consistent among them . hence , we can assume that the extinction would hardly change the position of h@xmath2 arms and neglect the effect of extinction on the offset values . we examined the offset between the arm of h@xmath2 and co of sac galaxy ngc 4254 , derived ten offset values and found the linear relation between the offset angle and the angular rotation velocity . this linearity implies that a rigid spiral pattern exists and that the physical processes of star formation can be regarded as not to extremely change in the disk . we emphasize that this is the first work to show this relationship . with this relation , we derived the pattern speed @xmath41 , and the time delay for star formation @xmath9 yr . we again emphasize that we can derive both @xmath6 and @xmath7 simultaneously . since the obtained values are consistent with previous studies , the way of our analysis can be a new method for the determination of the pattern speed and the typical timescale needed for star formations . thus , this method can investigate not only the kinematics , but also the star formation mechanisms quantitatively . moreover , this method can be applied to many spiral galaxies , since offsets between h@xmath2 and co arms have been found in many spirals . we are very grateful to rebecca a. koopmann , jeffrey d. p. kenney , and judith young for kindly providing us with their @xmath18 and h@xmath2 image of ngc 4254 . h. n. is financially supported by a research fellowship from the japan society for the promotion of science for young scientists . adler , d. s. , & westpfahl , d. j. 1996 , , 111 , 735 canzian , b. 1993 , , 414 , 487 cepa , j. & beckman , j. e. 1990 , , 349 , 497 combes , f. , & becquaert , j - f . 1997 , , 326 , 554 de vaucouleurs , g. , de vaucouleurs , a. , corwin , h. g. jr . , buta , r. j. , paturel , g. , & fouqu p. 1991 , third reference catalog of bright galaxies ( new york : springer - verlag ) elmegreen , d. m. , & elmegreen , b. g. 1987 , , 314 , 3 ferrarese , l. , freedman , w. l. , hill , r. j. , saha , a. , madore , b. f. , kennicut , r. c. jr . , stetson , p. b. , ford , h. c. , et al . 1996 , , 464 , 568 fujimoto , m. 1968 , iaus , 29 , 453 gonzlez , r. a. , & graham , j. r. 1996 , , 460 , 651 gerssen , j. , kuijken , k. , & merrifield , m. r. 1999 , , 306 , 926 iye , m. , okamura , s. , hamabe , m. , & watanabe , m. 1982 , , 256 , 103 koopmann , r. a. , kenney , j. d. , & young , j. 2001 , , 135 , 125 kranz , t. , slyz , a. , & rix , h .- w . 2001 , , 562 , 164 lin , c. c. & shu , f. h. 1964 , , 140 , 646 phookun , b. , vogel , s. n. , & mundy , l. g. 1993 , , 418 , 113 rand , r. j. & kulkarni , s. r. 1990 , , 349 , l43 roberts , w. w. 1969 , , 158 , 123 sempere , m. j. , garca - burillo , s. , combes , f. , & knapen , j. h. 1995 , , 296 , 45 schweizer , f. 1976 , , 31 , 313 sofue , y. , koda , j. , nakanishi , h. , onodera , s. , kohno , k. , tomita , a. , & okumura , s. k. 2003 , , 55,17 sofue , y. , koda , j. , nakanishi , h. , & onodera , s. 2003 , , 55 , 59 sofue , y. , koda , j. , nakanishi , h. , & hidaka , m. 2003 , , 55 , 75 takamiya , t. , & sofue , y. 2002 , , 576 , l15 tremaine , s. & weinberg , m. d. 1984 , , 282 , l5 tully , r. b. 1974 , , 27 , 449
we examined offsets between hii regions and molecular clouds belonging to spiral arms of a late type spiral galaxy ngc 4254 ( m99 ) . we used a high resolution @xmath0co ( @xmath1=1 - 0 ) image obtained by nobeyama millimeter array ( nma ) and an h@xmath2 image . we derived angular offsets ( @xmath3 ) in the galactic disk , and found that these offsets show a linear dependence on the angular rotation velocity of gas ( @xmath4 ) . this linear relation can be expressed by an equation : @xmath5 , where @xmath6 and @xmath7 are constant . here , @xmath6 corresponds to the pattern speed of spiral arms and @xmath7 is interpreted as the timescale between the peak compression of the molecular gas in spiral arms and the peak of massive star formation . we may thus determine @xmath6 and @xmath7 simultaneously by fitting a line to our @xmath8 plot , if we assumed they are constant . from our plot , we obtained @xmath9 yr and @xmath10 , which are consistent with previous studies . we suggest that this @xmath8 plot can be a new tool to determine the pattern speed and the typical timescale needed for star formations .
recently , many researchers are interested in the transmission eigenvalue problem @xcite . the transmission eigenvalue problem arises in the study of the inverse scattering for inhomogeneous media which not only has theoretical importance @xcite , but also can be used to estimate the properties of the scattering material @xcite since they can be determined from the scattering data . in the past few years , significant progress of the existence of transmission eigenvalues and applications has been made . we refer the readers to the recent papers @xcite . meanwhile , there are also many papers to give the numerical treatment for the transmission eigenvalue problem and the associated interior transmission problem @xcite . but there are few papers providing the corresponding theoretical analysis for their numerical methods due to the difficulty that the problem is neither elliptic nor self - adjoint . the paper @xcite presents an accurate error estimate of the eigenvalue and eigenfunction approximations for the helmhotz transmission eigenvalue problem based on the iterative methods ( bisection and secant ) from @xcite . the first aim of this paper is to give a theoretical analysis of the finite element method for the transmission eigenvalue problem with the inhomogeneous media . in the past few years , a new type of multilevel correction method is proposed to solve the eigenvalue problem @xcite . in the multilevel correction scheme , the solution on the finest mesh can be reduced to a series of solutions of the eigenvalue problem in a very low dimensional space and a series of solutions of the boundary value problem on the multilevel meshes . this multilevel correction method gives a way to construct a type of multigrid scheme for the eigenvalue problem @xcite . the second aim of this paper is to propose a multilevel correction method for the transmission eigenvalue problem based on the obtained error estimate results . the rest of this paper is organized as follows . in section [ transmission_eigenvalue_problem ] , we introduce the transmission eigenvalue problem and the corresponding theoretical results about the eigenvalue distribution . the finite element method and the corresponding error estimates are given in section [ fem_method ] . section [ multilevel_correction_method ] is devoted to introducing a type of multilevel correction method for the transmission eigenvalue problem . in section [ numerical_results ] , four examples are presented to validate the theoretical results and the efficiency of the proposed numerical methods . some concluding remarks are given in the last section . first , we introduce some notation and the transmission eigenvalue problem . the letter @xmath0 ( with or without subscripts ) denotes a generic positive constant which may be different at its different occurrences through the paper . for convenience , the symbols @xmath1 , @xmath2 and @xmath3 will be used in this paper . notations @xmath4 and @xmath5 , mean that @xmath6 , @xmath7 and @xmath8 for some constants @xmath9 and @xmath10 that are independent of mesh sizes . in this paper , we are concerned with the transmission eigenvalues corresponding to the scattering of acoustic waves by a bounded simply connected inhomogeneous medium @xmath11 ( @xmath12 ) . the transmission eigenvalue problem is to find @xmath13 , @xmath14 such that @xmath15 where @xmath16 is the unit outward normal to the boundary @xmath17 . there exists a real number @xmath18 such that the symmetric matrix @xmath19 and the index of refraction @xmath20 satisfy that @xmath21 values of @xmath22 such that there exists a nontrivial solution @xmath23 to ( [ problem ] ) are called transmission eigenvalues . [ remark_condition ] as in @xcite , the numerical method and analysis can be extended to the case that there exists a real number @xmath24 such that the symmetric matrix @xmath19 and the index of refraction @xmath20 satisfy that @xmath25 obviously , the eigenvalue problem ( [ problem ] ) can be transformed into the following version : find @xmath26 , @xmath27 such that @xmath28 where @xmath29 . in the following of this paper , we mainly consider this eigenvalue problem . there are some papers @xcite being concerned with the distribution of the eigenvalues for the eigenvalue problem ( [ problem_2 ] ) . in this paper , in order to give the analysis , we define the function spaces @xmath30 and @xmath31 as follows @xmath32 equipped with the norms @xmath33 respectively , where @xmath34 . for the simplicity of notation , we define two sesquilinear forms @xmath35 where @xmath36 . the associated variational form for ( [ problem_2 ] ) can be defined as follows : find @xmath37 such that @xmath38 and @xmath39 then the corresponding adjoint eigenvalue problem is : find @xmath40 such that @xmath41 and @xmath42 in order to analyze the properties of the eigenvalue problem ( [ problem_weak ] ) , we introduce the so - called @xmath43-coercivity ( inf - sup condition ) for the bilinear form @xmath44 ( see , e.g. , @xcite ) . in this paper , the notation @xmath43 denotes an isomorphic operator from @xmath30 to @xmath30 which is defined as follows @xmath45 similarly to @xcite , in order to give the eigenvalue distribution of ( [ problem_2 ] ) , we also state the following @xmath43-coercivity properties ( inf - sup conditions ) . [ inf_sup_theorem ] the bilinear forms @xmath44 and @xmath46 have the following inf - sup conditions ( @xmath43-coercivities ) : @xmath47 and @xmath48 for some positive constants @xmath49 and @xmath50 . from the conditions of the matrix @xmath51 and the refraction index @xmath20 , the following estimates hold @xmath52 since @xmath18 , we can choose @xmath53 such that @xmath44 is @xmath43-coercive which means there exists a positive constant @xmath49 such that the desired result ( [ inf - sup_left ] ) holds . in the same way , we can also prove the result ( [ inf - sup_right ] ) . similarly , it is easy to prove that @xmath54 since @xmath18 , we can choose @xmath53 such that @xmath46 is @xmath43-coercive in @xmath55 which means the inf - sup conditions ( [ inf - sup_left_b ] ) and ( [ inf - sup_right_b ] ) hold for some positive constant @xmath50 . we introduce the operators @xmath56 defined by the equations @xmath57 from theorem [ inf_sup_theorem ] , it is easy to know the operator @xmath58 and @xmath59 are linear bijective operators . then the eigenvalue problem ( [ problem_weak ] ) can be written as an operator form for @xmath60 ( denoting @xmath61 ) : @xmath62 with @xmath63 for the adjoint eigenvalue problem ( [ eigenvalue_problem_weak_adjoint ] ) . the @xmath43-coercivity conditions ( [ inf - sup_left])-([inf - sup_right ] ) and ( [ inf - sup_left_b])-([inf - sup_right_b ] ) guarantee that every eigenvalue @xmath64 is nonzero . from ( [ inf - sup_left ] ) and ( [ inf - sup_right ] ) and the compact embedding theorem of sobolev spaces , it is well known that the operators @xmath58 and @xmath59 are compact . thus the spectral theory for compact operators gives us a complete characterization of the eigenvalue problem ( [ problem_weak ] ) . there is a countable set of eigenvalues of ( [ problem_weak ] ) . let @xmath64 be an eigenvalue of problem ( [ problem_weak ] ) . there exists a smallest integer @xmath65 such that @xmath66 where @xmath67 denotes the null space and we use the notation @xmath68 . let @xmath69 denote the algebraic and geometric eigenspaces , respectively . the subspaces @xmath70 and @xmath71 are finite dimensional . the numbers @xmath72 and @xmath73 are called the algebraic and the geometric multiplicities of @xmath74 ( and @xmath64 ) . the vectors in @xmath70 are generalized eigenvectors . the order of a generalized eigenvector is the smallest integer @xmath75 such that @xmath76 ( vectors in @xmath77 being generalized eigenvectors of order @xmath78 ) . let us point out that a generalized eigenvector @xmath79 of order @xmath75 satisfies @xmath80 where @xmath81 is a generalized eigenvector of order @xmath82 . similarly we define the spaces of ( generalized ) eigenvectors for the adjoint problem @xmath83 note that @xmath74 is an eigenvalue of @xmath58 ( @xmath64 is an eigenvalue of problem ( [ problem_weak ] ) ) if and only if @xmath84 is an eigenvalue of @xmath59 ( @xmath85 is an eigenvalue of adjoint problem ( [ eigenvalue_problem_weak_adjoint ] ) ) with the ascent @xmath65 and the algebraic multiplicity @xmath86 for both eigenvalues being the same . now , let us define the finite element approximations for the problem ( [ problem_weak ] ) . first we generate a shape - regular decomposition of the computational domain @xmath87 into triangles or rectangles for @xmath88 ( tetrahedrons or hexahedrons for @xmath89 ) . the diameter of a cell @xmath90 is denoted by @xmath91 . the mesh diameter @xmath92 describes the maximum diameter of all cells @xmath90 . based on the mesh @xmath93 , we construct a finite element space denoted by @xmath94 . the same argument as in the beginning of this section illustrates that the following discrete inf - sup conditions also hold @xmath95 the standard finite element method for the problem ( [ problem_weak ] ) is defined as follows : find @xmath96 such that @xmath97 and @xmath98 similarly , the discretization for the adjoint problem ( [ eigenvalue_problem_weak_adjoint ] ) can be defined as : find @xmath99 such that @xmath100 and @xmath101 by introducing galerkin projections @xmath102 with the following equations @xmath103 the equation ( [ eigenvalue_problem_weak_discrete ] ) can be rewritten as an operator form with @xmath104 ( @xmath105 is a bounded operator ) , @xmath106 similarly for the adjoint problem ( [ eigenvalue_problem_weak_discrete_adjoint ] ) , we have @xmath107 let @xmath74 be an eigenvalue ( with algebraic multiplicity @xmath86 ) of the compact operator @xmath58 . if @xmath58 is approximated by a sequence of compact operators @xmath108 converging to @xmath58 in norm , i.e. , @xmath109 , then for @xmath92 sufficiently small @xmath74 is approximated by exactly @xmath86 eigenvalues @xmath110 ( counted according to their algebraic multiplicities ) of @xmath108 , i.e. , @xmath111 the space of generalized eigenvectors of @xmath58 is approximated by the subspace @xmath112 where @xmath113 is the smallest integer such that @xmath114 . we similarly define the space @xmath115 and the counterparts @xmath116 , @xmath117 for the adjoint problem . now , we describe a computational scheme to produce the algebraic eigenspace @xmath118 from the geometric eigenspace @xmath119 corresponding to eigenvalues @xmath120 , which converge to the same eigenvalue @xmath64 . starting from all eigenfunctions in the geometric eigenspace @xmath121 ( of order @xmath78 ) , we use the following recursive process to compute algebraic eigenspaces ( c.f . @xcite ) @xmath122 where @xmath123 , @xmath124 is the general eigenfunction of order @xmath75 and @xmath125 for @xmath126 . with the above process , we generate the algebraic eigenspace @xmath127 corresponding to eigenvalues @xmath128 , which converge to the same eigenvalue @xmath64 . similarly , we can produce the adjoint algebraic eigenspace @xmath116 from the geometric eigenspace @xmath117 . for two linear spaces @xmath51 and @xmath129 , we denote @xmath130 and define gaps between @xmath51 and @xmath129 in @xmath131 as @xmath132 and in @xmath133 as @xmath134 before introducing the convergence results of the finite element approximation for nonsymmetric eigenvalue problems , we define the following notations @xmath135 in order to derive error bounds for eigenpair approximations in the weak norm @xmath133 , we need the following error estimates in the weak norm @xmath133 of the finite element approximation . [ negative_norm_estimate_lemma ] ( ( * ? ? ? * lemma 3.3 and lemma 3.4 ) ) @xmath136 and @xmath137 base on the general theory of the error estimates for the eigenvalue problems by the finite element method ( * ? ? ? * section 8) , we have the following results for the transmission eigenvalue problem . [ error_estimate_theorem ] when the mesh size @xmath92 is small enough , we have @xmath138 where @xmath139 with @xmath140 converging to @xmath64 . in this section , we introduce a type of multilevel correction method for the transmission eigenvalue problem . this multilevel correction method consists of solving some auxiliary linear problems in a sequence of finite element spaces and an eigenvalue problem in a very low dimensional space . for more discussion about the multilevel correction method , please refer to @xcite . in order to do multilevel correction scheme , we first generate a coarse mesh @xmath141 with the mesh size @xmath142 and the coarse linear finite element space @xmath143 is defined on the mesh @xmath141 @xcite . then we define a sequence of triangulations @xmath144 of @xmath11 determined as follows . suppose @xmath145 ( produced from @xmath141 by regular refinements ) is given and let @xmath146 be obtained from @xmath147 via regular refinement ( produce @xmath148 subelements ) such that @xmath149 where the integer @xmath150 denotes the refinement index @xcite . it always equals @xmath151 in the first three numerical experiments with quasi - uniform refinement . based on this sequence of meshes , we construct the corresponding linear finite element spaces such that @xmath152 before designing the multilevel correction method , we first introduce a type of one correction step which can improve the accuracy of the given eigenpair approximation by solving a linear problem and an eigenvalue problem in a very low dimensional space . assume that we have obtained the algebraic eigenpair approximations @xmath153 and the corresponding adjoint ones @xmath154 for @xmath155 , where eigenvalues @xmath156 converge to the desired eigenvalue @xmath157 of ( [ problem_weak ] ) with multiplicity @xmath86 . now we introduce a correction step to improve the accuracy of the current eigenpair approximations . let @xmath158 be the conforming finite element space based on a finer mesh @xmath159 which is produced by refining @xmath146 . we start from a conforming linear finite element space @xmath143 on the coarsest mesh @xmath141 to design the following one correction step . [ correction_step ] one correction step 1 . for @xmath155 do + solve the following two boundary value problems : + @xmath160 find @xmath161 such that @xmath162 @xmath160 find @xmath163 such that @xmath164 end do 2 . define two new finite element spaces @xmath165 and @xmath166 + solve the following two eigenvalue problems : + @xmath160 find @xmath167 such that @xmath168 and @xmath169 @xmath160 find @xmath170 such that @xmath171 and @xmath172 3 . choose @xmath173 eigenpairs @xmath174 and @xmath175 to define two new geometric eigenspaces @xmath176 and @xmath177 based on these two geometric eigenspaces @xmath178 and @xmath179 , compute two algebraic eigenspaces @xmath180 and @xmath181 in order to simplify the notations and summarize the above three steps , we define @xmath182 since in step 1 of algorithm [ correction_step ] , the solving processes for the boundary value problems are independent of each other for different @xmath183 , we can implement them in parallel . [ error_estimate_one_correction_theorem ] assume the given eigenpairs @xmath184 in algorithm [ correction_step ] have the following error estimates @xmath185 then after one correction step , the resultant eigenpair approximations + @xmath186 have the following error estimates @xmath187 where @xmath188 from ( [ recurisve_law ] ) , there exist the basis functions @xmath189 of @xmath190 such that @xmath191 where @xmath192 denotes a polynomial of degree no more than @xmath65 for @xmath193 with @xmath194 and @xmath195 for @xmath196 . we can define a matrix @xmath197 such that @xmath198 where @xmath199 . it is easy to know that the matrix @xmath200 is nonsingular providing @xmath201 . for each @xmath202 , from the definitions of @xmath203 and @xmath204 , there exist a vector @xmath205 such that @xmath206 for any @xmath207 , we have @xmath208 from ( [ inf - sup_left ] ) and ( [ error_estimate_1 ] ) , the following estimate holds @xmath209 combining with the error estimate @xmath210 we have @xmath211 after step 3 , from the definition of @xmath212 and ( [ error_estimates_tilde_u_j ] ) , we derive @xmath213 where @xmath214 . similarly , we have @xmath215 then from the error estimate results stated in theorem [ error_estimate_theorem ] for the eigenvalue problem ( see , e.g. , ( * ? ? ? * section 8) ) and ( [ definition_varepsilon_k+1])-([definition_varepsilon_k+1_adjoint ] ) , the following error estimates hold @xmath216 these are the desired estimates ( [ estimate_u_u_h_2 ] ) and ( [ estimate_u_u_h_2_adjoint ] ) . furthermore , @xmath217 where @xmath218 then we obtain ( [ estimate_u_h_2_nagative ] ) . a similar argument leads to ( [ estimate_u_h_2_nagative_adjoint ] ) . now , based on the _ one correction step _ defined in algorithm [ correction_step ] , we introduce a multilevel correction scheme for the transmission eigenvalue problem . [ multi_correction ] multilevel correction scheme 1 . construct a coarse conforming finite element space @xmath219 on @xmath145 such that @xmath220 and solve the following two eigenvalue problems : + @xmath160 find @xmath221 such that @xmath222 and @xmath223 @xmath160 find @xmath224 such that @xmath225 and @xmath226 + choose @xmath173 eigenpairs @xmath227 and @xmath228 which approximate the desired eigenvalue @xmath157 and its geometric eigenspaces of the eigenvalue problem ( [ initial_eigen_problem ] ) and its adjoint one ( [ initial_eigen_problem_adjoint ] ) . based on these two geometric eigenspace , we compute the corresponding algebraic eigenspaces @xmath229 and @xmath230 . then do the following correction steps . 2 . construct a series of finer finite element spaces @xmath231 on the sequence of nested meshes @xmath232 ( c.f . 3 . do @xmath233 + obtain new eigenpair approximations @xmath186 by algorithm [ correction_step ] @xmath182 end do finally , we obtain eigenpair approximations @xmath234 . [ error_estimate_final_theorem ] after implementing algorithm [ multi_correction ] , the resultant eigenpair approximations @xmath235 have the following error estimates @xmath236 where @xmath237 , @xmath238 and + @xmath239 . first , we set @xmath240 and @xmath241 . then the following estimates hold @xmath242 by recursive relation and theorem [ error_estimate_one_correction_theorem ] , we derive @xmath243 and @xmath244 these are the estimates ( [ multi_correction_err_fun ] ) and ( [ multi_correction_err_fun_weak ] ) and the estimates ( [ multi_correction_err_fun_adjoint ] ) and ( [ multi_correction_err_fun_weak_adjoint ] ) can be proved similarly . from theorem [ error_estimate_theorem ] , ( [ multi_correction_err_fun ] ) and ( [ multi_correction_err_fun_adjoint ] ) , we can obtain the estimate ( [ multi_correction_err_eigen ] ) . in order to give the final error estimate results for the eigenpair approximations by the multilevel correction method , we assume the following properties for the error estimates hold @xcite @xmath245 when the mesh sizes @xmath246 , @xmath247 satisfy the relation ( [ mesh_size_relation ] ) and the eigenfunctions have the corresponding regularities . after implementing algorithm [ multi_correction ] , the resultant eigenpair approximations @xmath235 have the following error estimates @xmath248 when the mesh size @xmath142 is small enough , the conditions ( [ relations ] ) , @xmath249 and @xmath250 hold for the hidden constant @xmath0 . when the mesh size @xmath142 is small enough , the conditions ( [ relations ] ) and @xmath249 hold , we have the following inequalities @xmath251 combining the above estimate and theorem [ error_estimate_final_theorem ] , we can obtain the desired results ( [ multi_correction_err_fun_2 ] ) and ( [ multi_correction_err_fun_weak_2 ] ) . similar argument can lead to the results ( [ multi_correction_err_fun_adjoint_2 ] ) and ( [ multi_correction_err_fun_weak_adjoint_2 ] ) . then the result ( [ multi_correction_err_eigen_2 ] ) can be derived from ( [ multi_correction_err_fun_2])-([multi_correction_err_fun_weak_adjoint_2 ] ) and the proof is complete . in this section , we present four examples to validate the efficiency of the proposed multilevel correction scheme defined by algorithm [ multi_correction ] . the conforming linear finite element is used in the discretization for all the examples . let @xmath252 be a unit disk , @xmath253 for some constant @xmath254 and @xmath255 is also a constant . the solutions of can be written as @xmath256 where @xmath257 is the first kind bessel function of order @xmath86 . in order to satisfy the boundary condition @xmath258 on @xmath259 , one can choose @xmath260 of course , the trigonometric functions in - can also be chosen as @xmath261 . ignoring the trigonometric functions and using the recursive identity @xmath262 we obtain that @xmath263 thus the boundary condition for the normal derivative implies that @xmath264 from , we can get the exact transmission eigenvalues and the corresponding eigenfunctions by using - . in this example , we take @xmath265 and @xmath266 . the eigenvalues have multiplicity 2 for @xmath267 due to the trigonometric terms @xmath268 and @xmath261 . we use the quasi - uniform meshes @xmath269 and solve the transmission eigenvalue problem by using algorithm [ multi_correction ] . let @xmath270 be the @xmath183-th transmission eigenvalue computed on the mesh @xmath146 , while @xmath271 be the @xmath183-th exact eigenvalue obtained from . the left part of figure [ fig - circle ] shows the @xmath272-@xmath273 curve , where @xmath274 is the number of degrees of freedom(dofs ) with the mesh @xmath146 which is almost double of the number of nodes . it indicates that the sum of the errors for the first six transmission eigenvalues decreases as @xmath275 or has the @xmath276 convergence order as implied in theorem [ error_estimate_final_theorem ] . table [ tab - circle ] depicts the first six transmission eigenvalues computed by algorithm [ multi_correction ] on the finest mesh . .the first six transmission eigenvalues computed on the unit disk . [ cols="^,^,^,^,^,^,^,^",options="header " , ] in this paper , we give the error estimates for the transmission eigenvalue problem by the finite element method . furthermore , based on the obtained error estimates in theorem [ error_estimate_theorem ] , a type of multilevel correction method is proposed to solve the transmission eigenvalue problem . in the multilevel correction method , we transform the transmission eigenvalue solving in the finest finite element space to a sequence of linear problems and some transmission eigenvalue solving in a very low dimensional space . since the main computational work is to solve the sequence of linear problems , the multilevel correction method can improve the overfull efficiency of the transmission eigenvalue solving . the numerical results also show the efficiency of the proposed numerical scheme . the work of hehu xie is supported in part by the national natural science foundations of china ( nsfc 91330202 , 11371026 , 11001259 , 11031006 , 2011cb309703 ) , the national center for mathematics and interdisciplinary science the national center for mathematics and interdisciplinary science , cas . the work of xinming wu is supported in part by the national natural science foundations of china ( nsfc 91330202 , 11301089 ) . 99 i. babuka and j. e. osborn , eigenvalue problems . in : p.g . ciarlet , j.l . lions ( eds . ) handbook of numerical analysis , vol . ii , finite element methods ( part 1 ) , pp . 641 - 787 . north - holland , amsterdam , 1991 .
in this paper , we give a numerical analysis for the transmission eigenvalue problem by the finite element method . a type of multilevel correction method is proposed to solve the transmission eigenvalue problem . the multilevel correction method can transform the transmission eigenvalue solving in the finest finite element space to a sequence of linear problems and some transmission eigenvalue solving in a very low dimensional spaces . since the main computational work is to solve the sequence of linear problems , the multilevel correction method improves the overfull efficiency of the transmission eigenvalue solving . some numerical examples are provided to validate the theoretical results and the efficiency of the proposed numerical scheme . transmission eigenvalue problem , finite element method , error estimates , multilevel correction method 65n30 , 65n25 , 65l15 , 65b99 .
optical properties of structures composed of noble metals have long been attracting a considerable attention due to unique features of such systems in the visible spectrum.@xcite recent advances in fabrication techniques@xcite along with a tremendous progress in laser technologies opened new venues for application of plasmonic materials in biology,@xcite integrated optics,@xcite nanoscale imagining,@xcite and single molecule manipulation.@xcite physics of surface plasmon phenomenon is relatively simple and has long been studied.@xcite in brief , coherent oscillations of conductive electrons in a skin - layer of metal known as plasmons are capable of producing strong local electromagnetic ( em ) fields in the near - field region . it has been reported that such `` hot '' spots can be localized within @xmath0 nm or less . this along with a great sensitivity to initial conditions and geometry makes plasmonic structures so attractable for atom / molecule manipulations . a natural combination of nanoplasmonics and molecular response to the generated field started to appear as molecular nanopolaritonics,@xcite which studies molecular influence on field propagation , and as a tool for developing molecular switches.@xcite the latter utilizes nonadiabatic alignment of a molecule on semiconductor surface under a tip of scanning tunneling microscope . recent developments in experimental techniques capable of measuring optical response of current - carrying molecular junctions@xcite lead to theoretical formulations suitable for simultaneous description of both transport and optical properties of molecular devices.@xcite while experimental data are measured in real time , theoretical description of both transport and optical response so far has mostly been focused on a steady - state description . time - dependent transport usually is treated either within kinetic theory@xcite or within time - dependent density functional approach.@xcite the former generally misses broadening of molecular states due to coupling to macroscopic contacts@xcite and information on coherence,@xcite although interesting generalizations started to appear.@xcite limitations of the latter are due to absence of developed pseudopotentials and fundamental necessity to treat finite ( closed ) systems ( see e.g. ref . for discussion ) . an alternative approach , based on non - equilibrium green function ( negf ) technique , was initially formulated in refs . . this approach is a natural choice for description of open non - equilibrium systems . moreover it provides possibility to describe response of a molecular junction initially under bias to external time - dependent perturbation ( e.g. laser field ) . here we consider influence of external field specific for particular geometry on transport properties and optical response of molecular junction . while formulation of time - dependent transport within negf is general,@xcite all the applications so far were restricted to resonant single level models only . we propose a variant of the scheme capable of dealing with many - level systems . the exact calculations are compared to adiabatic pumping regime , frequent in the literature on time - dependent transport,@xcite were at the lowest order the problem is reduced to a set of quasi - steady - state solutions with time dependent ( slow timescale ) parameters . also we generalize our previous consideration of steady - state optical response of current - carrying junctions@xcite to a time - dependent situation . the paper is organized as follows . section [ model ] presents a model of molecular junction . section [ fdtd ] describes methodology of em field calculation . section [ td ] describes methodology for simulating transport through molecular junction subjected to external time - dependent field . adiabatic pumping version is discussed in section [ atd ] . numerical results are presented in section [ numres ] . section [ conclude ] concludes . we consider a two - level system @xmath1 , representing highest occupied ( homo ) and lowest unoccupied ( lumo ) molecular orbitals ( or ground and excited states in the many - body language ) , coupled to two macroscopic electrodes @xmath2 and @xmath3 . the electrodes are considered to be each in its own equilibrium with electrochemical potentials @xmath4 and @xmath5 , respectively . we assume that the driving ( laser field ) frequency is smaller than the plasma frequency , so that usual division of the junction into non - equilibrium molecule coupled to free electron reservoirs ( metallic contacts ) is relevant ( for a thorough discussion of the assumptions see ref . ) . local field at the position of the molecule is calculated within finite difference time domain technique ( see section [ fdtd ] for details ) , and is assumed to be an external time - dependent driving force causing ( de)excitation in the molecule . following ref . in addition to charge transfer between contacts and molecule we introduce also energy transfer ( coupling of molecular excitations to electron - hole excitations in the contacts ) . molecular excitations are coupled to a bath of free photon modes ( accepting modes ) , which serve as a measurement device of molecular optical response . hamiltonian of the system is @xmath6 here @xmath7 ( @xmath8 ) and @xmath9 ( @xmath10 ) are creation ( annihilation ) operators for an electron in the state @xmath11 of the molecule and state @xmath12 of the contact , respectively . @xmath13 ( @xmath14 ) is creation ( annihilation ) operator for a photon in the state @xmath15 , @xmath16 is external time - dependent field , and @xmath17 is matrix element of the molecular ( vector ) dipole operator between states @xmath11 and @xmath18 of the molecule ( @xmath19 ) . we assume @xmath20 ( or alternatively one can think about these contributions being included into definition of the state energies @xmath1 ) . @xmath21 and @xmath22 are matrix elements for electron and energy transfer between molecule and contacts , and @xmath23 represents optical response of the molecule . below we consider two approaches to transport and optical response simulations within the model : exact solution of the time - dependent dyson equation and adiabatic pumping regime . the former is similar to the procedure described in refs . , however it is presented in a form convenient for treating a multi - level molecular system ( see section [ td ] for discussion ) . the latter assumes that @xmath16 can be represented as a product of an oscillation of frequency @xmath24 with a slowly varying in time ( on the timescale of @xmath24 ) envelope @xmath25 . in the spirit of the born - oppenheimer approximation @xmath26 is considered as a parameter when solving electronic part of the problem . in this case the form of molecule - field interaction becomes ( within rotating wave approximation ) @xmath27 details of the approach are presented in section [ atd ] . as usual , we treat the perturbation @xmath28 , eq.([v ] ) , at the second order and within noncrossing approximation.@xcite self - energy due to energy transfer ( on the keldysh contour ) is@xcite @xmath29\end{aligned}\ ] ] where @xmath30 are molecular green functions in the lowest order of expansion associated with the hamiltonian @xmath31 , eq.([h0 ] ) , and @xmath32 are green functions of free electrons in the contacts . self - energy due to coupling to photon bath is@xcite @xmath33\ ] ] where @xmath34 is green function for free photon and @xmath35 is non - equilibrium reduced density matrix . below we discuss methods for calculating external field for different geometries , and present approaches to calculate time - dependent current and optical response of driven molecular junction . among various numerical techniques that allow one to predict optical properties of plasmonic systems the finite - difference time - domain approach ( fdtd ) is considered to be the most efficient and yet relatively simple . fdtd yields data in perfect agreement with experimental measurements and results obtained within other techniques.@xcite we simulate optical response of metal structures utilizing fdtd approach , in which maxwell equations are discretized in space and time following yee s algorithm@xcite . dispersion of dielectric constant of metal , @xmath36 , is taken in the form of the drude model @xmath37 with numerical parameters describing silver for the wavelengths of interest @xmath38 , @xmath39 rad / sec , @xmath40 rad / sec . for simulations of open systems , one needs to impose artificial absorbing boundaries in order to avoid reflection of outgoing em waves back to the simulation domain . among various approaches that address this numerical issue , the perfectly matched layers ( pml ) technique@xcite is considered to be the most adequate . it reduces the reflection coefficient of outgoing waves at the simulation region boundary to @xmath41 . essentially , the pml approach surrounds the simulation domain by thin layers of non - physical material that efficiently absorbs outgoing waves incident at any angle . we implement the most efficient and least memory intensive method , convolution perfectly matched layers ( cpml)@xcite absorbing boundaries , at all six sides of the 3d modeling space . through extensive numerical experimentation , we have empirically determined optimal parameters for the cpml boundaries that lead to almost no reflection of the outgoing em waves at all incident angles . spatial steps , @xmath42 , along all axes are fixed at @xmath43 nm to assure numerical convergence and the temporal step is @xmath44 , where @xmath45 is the speed of light in vacuum . numerical integration of maxwell equations on a grid within the fdtd framework was performed at the local asu home - built supercomputer utilizing 120 processors . an average execution time for our codes is around 20 minutes . a particular advantage of the fdtd method is its ability to obtain the optical response of the structure ( assuming linear response ) in the desired spectral range in a single run.@xcite the system is excited with an ultra - short optical pulse constructed from fourier components spanning the frequency range of interest . next , maxwell s equations are propagated in time for several hundred femtoseconds and the components of the em field are detected at the point of interest ( for our purposes we consider the detection point where a molecule is located ) . fourier transforming the detected em field on the fly yields intensities that can be easily processed into the spectral response . since we also have access to the field components , we can evaluate the intensity enhancement relative to the incident field . this provides the capability for straightforward evaluation of ` coupling efficiency ' of our plasmonic structures in the spectral range of interest . we are interested in calculating time - dependent current and optical response of the junction . expression for the current at the interface @xmath46 ( @xmath47 ) between molecule and contact is@xcite @xmath48 \nonumber\end{aligned}\ ] ] where @xmath49 is self - energy due to coupling to contact @xmath46 @xmath50_{ij}=\sum_{k\in k } v_{ik } g_k(\tau_1,\tau_2 ) v_{kj}\ ] ] and @xmath51 , @xmath52 , @xmath53 , @xmath54 are retarded , advanced , lesser , and greater projections respectively . in the wide band limit , when escape rate matrix @xmath55_{ij } = 2\pi \sum_{k\in k}v_{ik}v_{kj}\delta(e-\varepsilon_k)\ ] ] is assumed to be energy independent and real part of the self - energy ( [ set ] ) is disregarded , and when time modulation is restricted to molecular subspace only , expression ( [ ikt ] ) can be reduced to@xcite @xmath56 \\ \label{ikout } i_k^{out}(t ) = & + \frac{e}{\hbar } \mbox{re}\mbox{tr}\left[\mathbf{\gamma}_k\mathbf{\rho}(t)\right]\end{aligned}\ ] ] where @xmath57 is fermi - dirac distribution in contact @xmath46 and @xmath58 is time - dependent ( one - sided ) fourier transform of the retarded green function @xmath59 . @xmath60 in the absence of time - dependent driving @xmath58 reduces to usual fourier transform for retarded green function @xmath61^{-1}$ ] . in general @xmath62 has contributions ( additive within noncrossing approximation ) from all the processes involved . @xmath63 in ( [ ikout ] ) is reduced density matrix @xmath64 lesser and greater green functions are calculated from the time dependent dyson equation @xmath65 where @xmath66 and @xmath58 is defined in eq.([ar ] ) . contrary to our previous consideration@xcite optical response of molecular junction is calculated as a true photon flux into modes @xmath67 , rather than corresponding electronic current between molecular orbitals . we start from general expression for time - dependent photon flux into mode @xmath15 ( the derivation follows the corresponding procedure for electronic current , the latter can be found in e.g. ref . ) @xmath68 \nonumber\end{aligned}\ ] ] here @xmath69 is two - particle green function @xmath70 where @xmath71 is molecular de - excitation operator . for empty accepting mode @xmath15 expression ( [ jat ] ) reduces to @xmath72 as in ref . we approximate the two - particle green function by zero - order ( in interaction ) expression @xmath73\ ] ] note that if envelope change in time is slow ( on the timescale of @xmath74 ) second term on the right of ( [ gf2approx ] ) can be safely disregarded . in this case expression ( [ jatacc ] ) becomes equivalent to approximate expression used in ref . . below we calculate frequency resolved @xmath75 and total @xmath76 photon fluxes . here @xmath77,and in simulations we use@xcite @xmath78 to calculate time - dependent charge , eq.([iktwbl ] ) , and photon , eq.([jatacc ] ) , fluxes one needs time - dependent fourier transform of retarded green function , eq.([ar ] ) . the dyson equation for retarded green function is @xmath79 its one - sided fourier transform leads to equation for @xmath58 in the form @xmath80\right)\mathbf{a}^r(t , e ) \\ & \quad -\int_{-\infty}^{+\infty}dt_1\,\mathbf{\sigma}^r(t - t_1)\ , \mathbf{a}^r(t_1,e ) = \mathbf{i } \nonumber\end{aligned}\ ] ] we consider situation when time - dependent external field is applied at time @xmath81 to a biased molecular junction initially at steady - state . in this case differential equation ( [ dysa ] ) can be solved numerically starting from known initial condition @xmath82^{-1}$ ] . alternatively , splitting @xmath83 into time - independent @xmath84 and time - dependent @xmath85 parts ( average over time of the time - dependent part can be included into the time - independent hamiltonian ) , one can rewrite dyson equation ( [ dysg ] ) in the integral form @xmath86 one - sided fourier transform of ( [ dysgint ] ) leads to integral equation for @xmath58 @xmath87 where lower limit of the integral in the right is set to @xmath81 since @xmath88 . its solution is @xmath89\end{aligned}\ ] ] effective evolution operator @xmath90 can be obtained by variety of methods available in the literature ( see e.g. ref . and references therein ) . one of the simplest schemes is cumulant ( or magnus ) expansion.@xcite note that although our consideration is restricted to the case when time - dependent driving takes place in the molecular subspace only , generalization to driving in the contacts or at the molecule - contact interface is straightforward . when time evolution of an envelope @xmath25 , eq.([adiab ] ) , is slow on the timescale of the field frequency @xmath24 , consideration of the time dependent transport is simplified by invoking adiabatic assumption ( treating @xmath25 as a parameter ) . we start with hamiltonian ( [ h ] ) in which interaction with driving field is written in the form presented in eq.([adiab ] ) . transforming the hamiltonian into rotating frame of the field@xcite @xmath91 where @xmath92 ( @xmath93 ) , leads to @xmath94 where @xmath95 within rotating wave approximation only diagonal elements of the self - energy due to coupling to the contacts ( electron transfer ) @xmath96 , eq.([set ] ) , and self - energy due to coupling to electron - hole excitations ( energy transfer ) @xmath97 , eq.([sen ] ) , survive @xmath98 for self - energy due to coupling to photon bath @xmath99 , eq.([sp ] ) , we neglect non - diagonal terms , since they contribute to retarded ( advanced ) projection only and coupling to the bath is assumed to be small relative to coupling to the contacts . the self - energy becomes diagonal @xmath100 resulting green functions @xmath101 depend parametrically on slow time variable @xmath102 through time dependence of the envelope @xmath25 , eq.([h0bar ] ) . transforming to wigner coordinates , taking fourier transform in the relative coordinate @xmath103 , and using gradient expansion,@xcite leads to the following expressions for charge @xmath104 \nonumber\end{aligned}\ ] ] and photon @xmath105 fluxes . ( [ ikbar ] ) and ( [ jabar ] ) are main results of this section . they are to be compared with general expressions ( [ ikt ] ) and ( [ jatacc ] ) , respectively . we calculate time - dependent transport and optical response by envoking runge - kutta scheme with adaptive stepsize control@xcite to solve numerically system of differential equations ( [ dysa ] ) . ( color online ) current on the left , @xmath106 , and right , @xmath107 , interfaces vs. time for single level model . numerical results ( dashed line , red ) are compared to analytical expression ( solid line , blue ) . also shown is sum of the currents , @xmath108 at the two interfaces ( dotted line , black ) . see text for parameters . ] ( color online ) results of fdtd simulations . left panel shows intensity enhancement as a function of the incident wavelength ( in nm ) in logarithmic scale for two spheres of @xmath109 nm in diameter with a gap of @xmath0 nm ( solid line , black ) and bowtie antenna with a gap of @xmath0 nm ( dashed line , red ) , and @xmath110 nm ( dash - dotted line , blue ) . top right inset represents steady - state intensity enhancement distribution in logarithmic scale for two spheres system at the resonant wavelength of @xmath111 nm . lower right inset shows intensity distribution for the bowtie antennas with a gap of @xmath110 nm at @xmath112 nm . ] to check accuracy of our numerical approach we start from a test calculation for a single level model . analytical solution is available for the latter.@xcite in a biased junction ( @xmath113 ev and @xmath114 ev ) the level is set below both chemical potentials ( @xmath115 ev ) , so that initially the level is occupied and current through the junction is negligible ( escape rates are @xmath116 ) . at time @xmath81 position of the level is shifted to @xmath117 ev ( steplike modulation ) . here and below we assume fermi distributions in the leads corresponding to room temperature @xmath118 k. figure [ fig1 ] presents transient current at the two interfaces ( direction from contact into the system is taken to be positive for both currents ) calculated numerically ( dashed line ) and with analytical solution ( solid line ) . also shown is sum of the currents at the two interfaces ( dotted line ) . outflux of electrons from initially fully populated level into the right contact leads to ringing effect . eventually the current achieves steady - state . our numerical procedure is seen to give good correspondence with the analytical result . below we use similar parameters for calculation of time - dependent response of the two - level system . we consider two geometries of a junction : a bowtie antenna like electrodes and electrodes in the form of metallic spheres . large single - molecule fluorescence measurements were reported recently for the former.@xcite the latter ( molecule between two metallic nanoparticles ) is customary in experimental setups . both structures are excited by a plane wave polarized along the axis of symmetry ( i.e. along the axis connecting centers of two spheres , for instance ) . the electric field amplitude is then detected as a function of time . recorded amplitudes are fourier transformed and normalized with respect to the incident field amplitude leading to enhancement as a function in the frequency domain . results of our simulations for both geometries are presented in fig . [ fig2 ] showing intensity enhancements in the main panel . as expected bowtie structures result in noticeably higher enhancements reaching @xmath119 centered at @xmath120 nm for a bowtie antenna with a gap of @xmath110 nm . two spheres also show significant enhancement of @xmath121 around @xmath122 nm . we note that the bowtie antenna in comparison to two spheres system exhibits two resonances . the `` blue '' resonance located at low wavelength corresponds to rod lightning effect with high enhancement localized primarily at the edges of each triangle . this feature disappears from the spectrum once sharp corners are replaced with smooth edges.@xcite top and bottom insets show intensity enhancement distributions at resonant conditions for the two spheres and bowtie antennas , respectively . we place molecular junction in the hot spot regions . ( color online ) comparison of exact numerical solution ( solid line , red ) to adiabatic approximation ( dashed line , blue ) for the two - level ( homo - lumo ) model . shown are ( a ) levels populations and ( b ) current at the left interface vs. time . see text for parameters . ] ( color online ) the two - level ( homo - lumo ) model . shown are ( a ) current and ( b ) total optical response , ( [ jtot ] ) , vs. time for bowtie nanoantennas ( the strongest signal , blue ) and two spheres junction geometries . in the latter case the response is calculated for two positions of the molecule in the junction : in the middle between the spheres ( the weakest signal , red ) and closer to one of the spheres ( intermediate signal , ( a ) white silhouette and ( b ) solid line , black ) . figure ( c ) shows contour map of optical flux , ( [ jwa ] ) , for bowtie geometry vs. outgoing frequency and time . see text for parameters . ] figure [ fig3]a shows time - dependent populations of molecular junction driven by external electromagnetic field for the ground , @xmath123 , and excited , @xmath124 , states . time - dependent current at the left interface , @xmath106 , is shown in fig . parameters of the calculation are @xmath118 k , @xmath125 ev , @xmath126 ev , @xmath127_{mm}=0.1 $ ] ev and @xmath127_{12}=[\mathbf{\gamma}_{k}]_{21}=0 $ ] ( @xmath128 and @xmath47 ) . for interaction with electromagnetic field we take @xmath129 ev , where @xmath130 is amplitude of the external laser field before enhancement . bias @xmath131 is applied symmetrically @xmath132 , and the fermi energy is @xmath133 . results presented in fig . [ fig3 ] are obtained for bowtie geometry with @xmath0 nm gap at bias @xmath134 v. exact numerical calculation ( solid line ) is compared with adiabatic approximation data ( dashed line ) . one sees , that the adiabatic approximation for realistic parameters provides qualitatively correct results . it misses however delay ( memory ) effects and overestimates response signal . electromagnetic pulse depletes ground state and populates excited state , which for the chosen bias leads to increase of current through the junction due to increase in transmission of the excited state channel ( see also fig . [ fig5 ] below ) . we compare response of the two molecular junction geometries in fig . bowtie geometry provides stronger local enhancement , and consequently stronger molecular response . in the case of spherical nanoparticles we consider two possible positions of molecule between the electrodes : symmetric and asymmetric ( @xmath135 nm shift from the center , where the field enhancement for the geometry is strongest ) . these yield weakest and intermediate signal , respectively . note , that it is natural to expect that local field enhancement is stronger for a structure with uneven surface . [ fig4]a presents time - dependent current for the three cases . total optical response , eq.([jtot ] ) , is shown in fig . [ fig4]b . we choose @xmath136 and @xmath137 ev , other parameters are as in fig . note much more sensitive character of optical response to resonant conditions . it results from our choice of @xmath138 , eq.([gamma ] ) , so that most of the electronic excitation contributes to current . while the choice is arbitrary , it indicates importance of the environment ( bath spectral density ) . [ fig4]c shows time - dependent optical spectrum , eq.([jwa ] ) , for the bowtie geometry . the signal follows ( with a delay ) the pulse of the external field . asymmetric character of the spectrum relative to resonance , @xmath139 ev , stems from overlap of lorentzians ( levels boradening due to coupling to the contacts ) centered on ground and excited states . ( color online ) current vs. time for the two - level ( homo - lumo ) model calculated at two different biases . see text for parameters . ] figure [ fig5 ] shows time - dependent current response to external driving at two different constant biases . the calculation is done for bowtie geometry with a gap of @xmath0 nm , parameters are the same as in fig . [ fig3 ] . for pre - resonant bias , @xmath140 v , optical excitation is effective in depleting the ground and populating the excited states of the molecule , which results in increased current through both channels . at post - resonant bias , @xmath141 v , the charge transfer channels are open . here optical excitation contributes mostly to decrease in conductance of the ground state and appearance of leakage current to the left contact in the excited state . this leads to overall decrease in current through the junction ( see also discussion below ) . ( color online ) role of energy transfer process . shown are ( a ) total optical response vs. time with ( dotted line , red ) and without ( solid line , blue ) electron - hole excitations and ( b ) difference between current calculated with and without electron - hole excitations vs. time and bias . calculations are performed within adiabatic approximation scheme . see text for parameters . ] calculations so far disregarded influence of both energy transfer , eq.([sen ] ) , and external photon bath , eq.([sp ] ) , ( except its contribution to optical rate ) on electronic distribution in the molecule . while the latter can indeed be disregarded due to smallness of the reasonable coupling parameter ( see ref . for discussion ) , the former can make a difference . here we illustrate influence of energy transfer process on time - dependent response of the junction within adiabatic approximation ( full numeric calculation is straightforward but time - consuming ) . figure [ fig6]a shows total optical response calculated with ( dashed line ) and without ( solid line ) energy transfer included . calculation is done for bowtie geometry with @xmath0 nm gap at pre - resonant constant bias @xmath140 v. other parameters are as in fig . as expected , energy transfer diminishes optical response of the junction , since both energy transfer from molecule to contacts and fluorescence compete for the same excess electronic population in the excited state . current change upon including electron - hole excitations into consideration is more interesting . interplay between channel blocking and resonant pathways for electron transfer may lead to increase in current through the junction as is illustrated in fig . [ fig6]b ) . this effect is similar to the situation presented in fig . we consider a two - level ( homo - lumo ) model of molecular junction driven by external time - dependent laser field . finite difference time domain technique is used to calculate field distribution for two junction geometries . resulting local field at the molecule is considered to be the driving force . we assume that the junction is initially in a nonequilibrium steady - state resulting from applied constant bias . at time @xmath81 driving force ( laser pulse ) starts to influence the system . time - dependent transport ( charge flux through the junction ) and optical response ( photon flux from the molecule into accepting modes ) are calculated for a set of geometries and applied biases . we rewrite a nonequilibrium green function technique for time - dependent calculation in a form convenient for treating many - level molecular systems . results of the simulations within the approach are compared to approximate scheme for an adiabatic pumping regime . note that while our present consideration is restricted to driving force applied to the molecule only , generalization of the approach to situations of time - dependent bias and/or coupling between molecule and contacts is straightforward . extension of the consideration to realistic molecular devices , taking into account time - dependent non - equilibrium distribution in the contacts and spatial profile of the field , and considering interplay of time - dependencies of bias and laser field are goals of future research . is grateful to asu financial and technical support ( startup funds ) . m.g . gratefully acknowledges support by the ucsd ( startup funds ) , the uc academic senate ( research grant ) , and the u.s .- israel binational science foundation . 99 s.a.maier and h.a.atwater . phys . * 98 * , 011101 ( 2005 ) . science * 311 * , 189 ( 2006 ) . w.a.murray and w.l.barnes . adv . mat . * 19 * , 3771 ( 2007 ) . a.m.schwartzberg and j.z.zhang . j. phys . chem . c * 112 * , 10323 ( 2008 ) . e.hutter and j.h.fendler . adv . mat . * 16 * , 1685 ( 2004 ) . a.j.haes and r.p . van duyne . expert rev . diagn . * 4 * , 527 ( 2004 ) . r.charbonneau , n.lahoud , g.mattiussi , and p.berini . * 13 * , 977 ( 2005 ) . a.k.sarychev and v.m.shalaev . _ electrodynamics of metamaterials_. world scientific ( 2007 ) . h.yokota , k.saito , and t.yanagida . lett . * 80 * , 4606 ( 1998 ) . h.raether . _ surface plasmons on smooth and rough surfaces and on gratings_. springer , berlin ( 1988 ) . u.kreibig and m.vollmer . _ optical properties of metal clusters_. springer , new york ( 1995 ) . k.lopata and d.neuhauser . j. chem . 130 * , 104707 ( 2009 ) . k.lopata and d.neuhauser . j. chem . 131 * , 014701 ( 2009 ) . m.g.reuter , m.sukharev , and t.seideman . * 101 * , 208303 ( 2008 ) . s.w.wu , g.v.nazin , and w.ho , phys . b * 77 * , 205430 ( 2008 ) . d.r.ward , n.j.halas , j.w.ciszek , j.m.tour , y.wu , p.nordlander and d.natelson . nano lett . * 8 * , 919 ( 2008 ) . z.ioffe , t.shamai , a.ophir , g.noy , i.yutsis , k.kfir , o.cheshnovsky and y.selzer . nature nanotechnology * 3 * , 727 ( 2008 ) . m.galperin and a.nitzan . j. phys . * 124 * , 234709 ( 2006 ) . m.galperin , m.a.ratner , and a.nitzan . j. chem . phys . * 130 * , 144109 ( 2009 ) . e.g.petrov , v.may , and p.hnggi . * 319 * , 380 ( 2005 ) . e.g.petrov , v.may , and p.hnggi . b * 73 * , 045408 ( 2006 ) . m.koentopp , c.chang , k.burke , and r.car . j. phys . : condens . matter * 20 * , 083203 ( 2008 ) . g.stefanucci and c .- o.almbladh . . lett . * 67 * , 14 ( 2004 ) . s.kurth , g.stefanucci , c .- o.almbladh , a.rubio , and e.k.u.gross . b * 72 * , 035308 ( 2005 ) . j.n.pedersen and a.wacker . phys . b * 72 * , 195330 ( 2005 ) . i.v.ovchinnikov and d.neuhauser . * 122 * , 024707 ( 2005 ) . m.esposito and m.galperin . b * 79 * , 205303 ( 2009 ) . m.g.schultz and f. von oppen . b * 80 * , 033302 ( 2009 ) . m.leijnse and m.r.wegewijs . b * 78 * , 235424 ( 2008 ) . m.galperin and s.tretiak . j. chem . phys . * 128 * , 124705 ( 2008 ) . a.jauho , n.s.wingreen and y.meir . b * 50 * , 5528 ( 1994 ) . m.p.anantram and s.datta . b * 51 * , 7632 ( 1995 ) . b.wang , j.wang , and h.guo . textbf82 , 398 ( 1999 ) . a.prociuk and b.d.dunietz . b * 78 * , 165112 ( 2008 ) . a.r.hern ' andez , f.a.pinheiro , c.h.lewenkopf , and e.r.mucciolo . phys . rev . b * 80 * , 115311 ( 2009 ) . m.galperin and a.nitzan . phys . lett . * 95 * , 206802 ( 2005 ) . m.galperin and a.nitzan . j. chem . phys . * 124 * , 234709 ( 2006 ) . m.besbes , j.p.hugonin , p.lalanne , s. van haver , o.t.a.jansse , a.m.nugrowati , m.xu , s.f.pereira , h.p.urbach , a.s . van de nes , p.bienstman , g.granet , s.helfert , m.sukharev , t.seideman , f.i.baida , b.guizal , and d. van labeke . soc . : rapid publications * 2 * , 07022 ( 2007 ) . a.taflove and s.c.hagness . _ computational electrodynamics : the finite - difference time - domain method _ , 3rd ed . , artech house , boston ( 2005 ) . _ perfectly matched layer ( pml ) for computational electromagnetics _ in _ synthesis lectures on computational electromagnetics _ ed . c.a.balanis , morgan and claypool publ . j.a.roden and s.d.gedney . * 27 * , 334 ( 2000 ) . m.sukharev , j.sung , k.g.spears , and t.seideman . phys . rev . b * 76 * , 184302 ( 2007 ) . g.d.mahan . _ many - particle physics_. kluwer academic / plenum publishers ( 2000 ) . h.haug and a .- p.jauho . _ quantum kinetics in transport and optics of semiconductors_. springer ( 2008 ) . _ chemical dynamics in condensed phases_. oxford university press ( 2006 ) . d.lauvergnat , s.blasco , x.chapuisat , and a.nauts . j. chem . phys . * 126 * , 204103 ( 2007 ) . w.magnus . pure appl . math . * 7 * , 649 ( 1954 ) . _ principles of nonlinear optical spectroscopy_. oxford university press ( 1995 ) . d.prato and p.w.lamberti . j. chem . phys . * 106 * , 4640 ( 1997 ) . p.zhang , q .- k.xue , and x.c.xie . * 91 * , 196602 ( 2003 ) . j.fransson and j .- x.zhu . b * 78 * , 113307 ( 2008 ) . w.h.press , s.a.teukolsky , w.t.vetterling , and b.p.flannery . _ numerical recipes in c _ , cambridge university press ( 1994 ) . a.kinkhabwala , z.yu , s.fan , yu.avlasevich , k.m " ullen , and w.e.moerner . nature photonics * 3 * , 654 ( 2009 ) .
we consider a biased molecular junction subjected to external time - dependent electromagnetic field . the field for two typical junction geometries ( bowtie antennas and metal nanospheres ) is calculated within finite - difference time - domain technique . time - dependent transport and optical response of the junctions is calculated within non - equilibrium green s function approach expressed in a form convenient for description of multi - level systems . we present numerical results for a two - level ( homo - lumo ) model , and discuss influence of localized surface plasmon polariton modes on transport .
[ sec : intro ] alexandrov spaces are geodesic metric spaces with curvature bounds in the sense of local triangle comparisons . specifically , let @xmath5 denote the simply connected 2-dimensional riemannian space form of constant curvature @xmath4 . for curvature bounded below ( cbb ) by @xmath4 , the distance between any two points of a geodesic triangle is required to be @xmath6 the distance between the corresponding points on the `` model '' triangle with the same sidelengths in @xmath5 . for curvature bounded above ( cba ) , substitute `` @xmath7 '' . examples of alexandrov spaces include riemannian manifolds with sectional curvature @xmath2 or @xmath3 . a crucial property of alexandrov spaces is their preservation by gromov - hausdorff convergence ( assuming uniform injectivity radius bounds in the cba case ) . moreover , cbb spaces are topologically stable in the limit @xcite , a fact at the root of landmark riemannian finiteness and recognition theorems . ( see grove s essay @xcite . ) cba spaces are also important in geometric group theory ( see @xcite ) and harmonic map theory ( see , for example , @xcite ) . in lorentzian geometry , _ timelike _ comparison and rigidity theory is well developed . early advances in timelike comparison geometry were made by flaherty @xcite , beem and ehrlich @xcite , and harris @xcite . in particular , a purely timelike , global triangle comparison theorem was proved by harris @xcite . a major advance in rigidity theory was the lorentzian splitting theorem , to which a number of researchers contributed ; see the survey in @xcite , and also the subsequent warped product splitting theorem in @xcite . the comparison theorems mentioned assume a bound on sectional curvatures @xmath8 of timelike 2-planes @xmath9 . note that a bound over _ all _ nonsingular 2-planes forces the sectional curvature to be constant @xcite , and so such bounds are uninteresting . this project began with the realization that certain lorentzian warped products , which may be called minkowski , de sitter or anti - de sitter cones , possess a global triangle comparison property that is not just timelike , but is fully analogous to the alexandrov one . the comparisons we mean are on signed lengths of geodesics , where the timelike sign is taken to be negative . in this paper , _ length _ of either geodesics or vectors is always signed , and we will not talk about the length of nongeodesic curves . the _ model spaces _ are @xmath5 , @xmath10 or @xmath11 , where @xmath10 is the simply connected @xmath12-dimensional lorentz space form of constant curvature @xmath4 , and @xmath11 is @xmath5 with the sign of the metric switched , a space of constant curvature @xmath13 . the cones mentioned above turn out to have sectional curvature bounds of the following type . for any semi - riemannian manifold , call a tangent section _ spacelike _ if the metric is definite there , and _ timelike _ if it is nondegenerate and indefinite . write _ @xmath0 _ if spacelike sectional curvatures are @xmath2 and timelike ones are @xmath3 ; for @xmath1 , reverse `` timelike '' and `` spacelike '' . equivalently , @xmath0 if the curvature tensor satisfies @xmath14 and similarly with inequalities reversed . the meaning of this type of curvature bound is clarified by noting that if one has merely a bound above on timelike sectional curvatures , or merely a bound below on spacelike ones , then the restriction @xmath15 of the sectional curvature function to any nondegenerate @xmath16-plane @xmath17 has a curvature bound below in our sense : @xmath18 ( as follows from @xcite ; see [ sec : algebra ] below ) . then @xmath0 means that @xmath19 may be chosen independently of @xmath17 . spaces satisfying @xmath0 ( or @xmath1 ) are abundant , as warped product constructions show . they include , for example , the big bang cosmological models discussed by hawking and ellis @xcite ( see [ sec : examples ] below ) . since there are many warped product examples satisfying @xmath0 for all @xmath4 in a nontrivial finite interval , then by stability , there are many non - warped product examples . searching the literature for this type of curvature bound , we found it had been studied earlier by andersson and howard @xcite . their paper contains a riccati equation analysis and gap rigidity theorems . for example : a geodesically complete semi - riemannian manifold of dimension @xmath20 and index @xmath21 , having either @xmath22 or @xmath23 and an end with finite fundamental group on which @xmath24 , is @xmath25 @xcite . their method uses parallel hypersurfaces , and does not concern triangle comparisons or the methods of alexandrov geometry . subsequently , daz - ramos , garca - ro , and hervella obtained a volume comparison theorem for `` celestial spheres '' ( exponential images of spheres in spacelike hyperplanes ) in a lorentz manifold with @xmath26 or @xmath27 @xcite . does this type of curvature bound always imply local triangle comparisons , or do triangle comparisons only arise in special cones ? in this paper we prove that curvature bounds @xmath0 or @xmath1 are actually equivalent to local triangle comparisons . the existence of model triangles is described in the realizability lemma of [ sec : modellem ] . it states that any point in @xmath28 represents the sidelengths of a unique triangle in a model space of curvature @xmath29 , and the same holds for @xmath30 under appropriate _ size bounds for @xmath4_. we say @xmath31 is a _ normal neighborhood _ if it is a normal coordinate neighborhood ( the diffeomeorphic exponential image of some open domain in the tangent space ) of each of its points . there is a corresponding distinguished geodesic between any two points of @xmath31 , and the following theorem refers to these geodesics and the triangles they form . if in addition the triangles satisfy size bounds for @xmath4 , we say @xmath31 is _ normal for @xmath4_. all geodesics are assumed parametrized by @xmath32 $ ] , and by _ corresponding _ points on two geodesics , we mean points having the same affine parameter . [ thm : comparison]if a semi - riemannian manifold satisfies @xmath33 , and @xmath31 is a normal neighborhood for @xmath4 , then the signed length of the geodesic between two points on any geodesic triangle of @xmath31 is at least ( at most ) that for the corresponding points on the model triangle in @xmath5 , @xmath10 or @xmath11 . conversely , if triangle comparisons hold in some normal neighborhood of each point of a semi - riemannian manifold , then @xmath33 . in this paper , we restrict our attention to local triangle comparisons ( i.e. , to normal neighborhoods ) in smooth spaces . in the riemannian / alexandrov theory , local triangle comparisons have features of potential interest to semi - riemannian and lorentz geometers : they incorporate singularities , imply global comparison theorems , and are consistent with a theory of limit spaces . our longer - term goal is to see what the extension of the theory presented here can contribute to similar questions in semi - riemannian and lorentz geometry . we begin by mentioning some intuitive barriers to approaching theorem [ thm : comparison ] . in resolving them , we are going to draw on papers by karcher @xcite and andersson and howard @xcite , putting them to different uses than were originally envisioned . first , a fundamental object in riemannian theory is the locally isometrically embedded interval , that is , the unitspeed geodesic . these are the paths studied in @xcite and @xcite . however , in the semi - riemannian case this choice constrains consideration to fields of geodesics all having the same causal character . by contrast , our construction , which uses affine parameters on @xmath32 $ ] , applies uniformly to all the geodesics radiating from a point ( or orthogonally from a nondegenerate submanifold ) . secondly , a common paradigm in riemannian and alexandrov comparison theory is the construction of a curve that is shorter than some original one , so that the minimizing geodesic between the endpoints is even shorter . in the lorentz setting , this argument still works for _ timelike _ curves , under a causality assumption . however , spacelike geodesics are unstable critical points of the length functional , and so this argument is forbidden . thirdly , while the comparisons we seek can be reduced in the riemannian setting to @xmath34-dimensional riccati equations ( as in @xcite ) , the semi - riemannian case seem to require matrix riccati equations ( as in @xcite ) . such increased complexity is to be expected , since semi - riemannian curvature bounds below ( say ) have some of the qualities of riemannian curvature bounds both below _ and _ above . let us start by outlining karcher s approach to riemannian curvature bounds . it included a new proof of local triangle comparisons , one that integrated infinitesimal rauch comparisons to get distance comparisons without using the `` forbidden argument '' mentioned above . such an approach , motivated by simplicity rather than necessity in the riemannian case , is what the semi - riemannian case requires . in this approach , alexandrov curvature bounds are characterized by a differential inequality . namely , @xmath35 has cbb by @xmath4 in the triangle comparison sense if and only if for every @xmath36 and unit - speed geodesic @xmath37 , the differential inequality @xmath38 is satisfied ( in the barrier sense ) by the following function @xmath39 : @xmath40 the reason for this equivalence is that the inequalities ( [ eq : cb ] ) reduce to equations in the model spaces @xmath5 ; since solutions of the differential inequalities may be compared to those of the equations , distances in @xmath35 may be compared to those in @xmath5 . the functions @xmath41 then provide a convenient connection between triangle comparisons and curvature bounds , since they lead via their hessians to a riccati equation along radial geodesics from @xmath42 . we wish to view this program as a special case of a procedure on semi - riemannian manifolds . for a geodesic @xmath37 parametrized by @xmath32 $ ] , let @xmath43 thus @xmath44 in this paper , we work with normal neighborhoods , and set @xmath45 where @xmath46 is the geodesic from @xmath47 to @xmath42 that is distinguished by the normal neighborhood . ( in a broader setting , one may instead use the definition @xmath48 under hypotheses that ensure the two definitions agree locally . in ( [ eq : edef2 ] ) , @xmath49 if @xmath47 and @xmath42 are not connected by a geodesic . ) now define the _ modified distance function @xmath50 at @xmath42 _ by @xmath51 here , the formula remains valid when the argument of cosine is imaginary , converting @xmath52 to @xmath53 . in the riemannian case , @xmath54 . the cbb triangle comparisons we seek will be characterized by the differential inequality @xmath55 on any geodesic @xmath37 parametrized by @xmath32 $ ] . the self - adjoint operator @xmath56 associated with the hessian of @xmath57 may be regarded as a _ modified shape operator_. it has the following properties : in the model spaces , it is a scalar multiple of the identity on the tangent space to @xmath35 at each point ; along a nonnull geodesic from @xmath42 , its restriction to normal vectors is a scalar multiple of the second fundamental form of the equidistant hypersurfaces from @xmath42 ; it is smoothly defined on the regular set of @xmath58 , hence along null geodesics from @xmath42 ( as the second fundamental forms are not ) ; and finally , it satisfies a matrix riccati equation along every geodesic from @xmath42 , after reparametrization as an integral curve of @xmath59 . we shall also need semi - riemannian analogues to the three basic triangle lemmas on which alexandrov geometry builds , namely , the realizability , hinge and straightening lemmas . the analogues are intuitively surprising , both in one of the quantities considered , and also in the fact that monotonicity statements persist even though the model space may change . the straightening lemma is an indicator that , as in the standard riemannian / alexandrov case , there is a singular counterpart to the smooth theory developed in this paper . we begin in [ sec : modellem ] with the triangle lemmas just mentioned . in [ sec : mode ] , it is shown that the differential inequalities ( [ eq : diffineq ] ) become equations in the model spaces , and hence characterize our triangle comparisons . comparisons for the modified shape operators under semi - riemannian curvature bounds are proved in [ sec : comparericcati ] , and theorem [ thm : comparison ] is proved in [ sec : comparison ] . in [ sec : algebra ] , semi - riemannian curvature bounds are related to the analysis by beem and parker of the pointwise ranges of sectional curvature @xcite , and to the `` null '' curvature bounds considered by uhlenbeck @xcite and harris @xcite . finally , [ sec : examples ] considers examples of semi - riemannian spaces with curvature bounds , including robertson - walker `` big bang '' spacetimes . say three numbers _ satisfy the strict triangle inequality _ if they are positive and the largest is less than the sum of the other two . denote the points of @xmath60 whose coordinates satisfy the strict triangle inequality by @xmath61 , and their negatives by @xmath62 . a triple , one of whose entries is the sum of the other two , will be called _ degenerate_. denote the points of @xmath28 whose coordinates are nonnegative degenerate triples by @xmath63 , and their negatives by @xmath64 . in figure 1 , the shaded cone is @xmath63 , and the interior of its convex hull is @xmath65 . say a point is _ realized _ in a model space if its coordinates are the sidelengths of a triangle . as usual , set @xmath66 if @xmath67 . [ lem : realiz ] points of @xmath28 have unique realizations , up to isometry of the model space , as follows : _ . _ a point in @xmath65 is realized by a unique triangle in @xmath5 , provided the sum of its coordinates is @xmath68 . a point in @xmath62 is realized by a unique triangle in @xmath11 , provided the sum of its coordinates is @xmath69 . a point in @xmath63 is realized by unique triangles in @xmath5 and @xmath10 , provided the largest coordinate is @xmath70 . a point in @xmath64 is realized by unique triangles in @xmath11 and @xmath10 , provided the smallest coordinate is @xmath71 . a point in the complement of @xmath72 is realized by a unique triangle in @xmath73 . for @xmath74 , if the largest coordinate is @xmath70 , the point is realized by a unique triangle in @xmath10 . for @xmath75 , if the smallest coordinate is @xmath76 , the point is realized by a unique triangle in @xmath10 . part 1 is standard , as is part 2 for @xmath77 . now consider a point not in @xmath78 , and denote its coordinates by @xmath79 . to realize this point in @xmath73 , suppose @xmath80 and take a segment @xmath37 of length @xmath81 on the @xmath82-axis . since distance `` circles '' about a point @xmath47 are pairs of lines of slope @xmath83 through @xmath47 if the radius is @xmath29 , and hyperbolas asymptotic to these lines otherwise , it is easy to see that circles about the endpoints of @xmath37 intersect , either in two points or tangentially , subject only to the condition that @xmath84 if @xmath85 , namely , the point is not in @xmath65 . thus our point may be realized in @xmath86 , uniquely up to an isometry of @xmath86 . on the other hand , if @xmath87 then @xmath88 , so by switching the sign of the metric , we have just shown there is a realization in @xmath89 . for @xmath74 , @xmath10 is the simply connected cover of the quadric surface @xmath90 in minkowski @xmath16-space with signature @xmath91 . suppose @xmath92 , and take a segment @xmath37 of length @xmath81 on the quadric s equatorial circle of length @xmath93 in the @xmath94-plane . a distance circle about an endpoint of @xmath37 is a hyperbola or pair of lines obtained by intersection with a @xmath12-plane parallel to or coinciding with the tangent plane . two circles about the endpoints of @xmath37 intersect , either in two points or tangentially , if the vertical line of intersection of their @xmath12-planes cuts the quadric . this occurs subject only to the condition that @xmath84 if @xmath85 , namely , the point is not in @xmath65 . on the other hand , if @xmath87 then @xmath88 . take a segment @xmath37 of length @xmath95 in the quadric , where @xmath37 is symmetric about the @xmath94-plane . circles of nonpositive radius about the endpoints of @xmath37 intersect if the horizontal line of intersection of their @xmath12-planes cuts the quadric , and this occurs subject only to the condition that @xmath96 , namely , the point is not in @xmath62 . since @xmath97 , switching the sign of the metric completes the proof . let us say the points of @xmath28 for which lemma [ lem : realiz ] gives model space realizations _ satisfy size bounds for @xmath4 _ ( for @xmath98 , no size bounds apply ) . such a point may be expressed as @xmath99 , where @xmath100 is a realizing triangle in a model space of curvature @xmath4 , the geodesic @xmath46 is a side parametrized by @xmath32 $ ] with @xmath101 , and we write @xmath102 . by the _ nonnormalized angle @xmath103 _ , we mean the inner product @xmath104 . in our terminology , @xmath103 is the _ included _ , and @xmath105 and @xmath106 are the _ shoulder _ , nonnormalized angles for @xmath99 . this terminology is welldefined since the realizing model space and triangle are uniquely determined except for degenerate triples . the latter have only two realizations , which lie in geodesic segments in different model spaces but are isometric to each other . an important ingredient of the alexandrov theory is the hinge lemma for angles in @xmath5 , a monotonicity statement that follows directly from the law of cosines . part 1 of the following lemma is its semi - riemannian version . a new ingredient of our arguments is the use of nonnormalized shoulder angles , in which both the `` angle '' and one side vary simultaneously . not only do we obtain a monotonicity statement that for @xmath30 is not directly apparent from the law of cosines ( part 2 of the following lemma ) , but we find that monotonicity persists even as the model space changes . [ lem : hinge ] suppose a point of @xmath28 satisfies size bounds for @xmath4 , and the third coordinate varies with the first two fixed . denote the point by @xmath99 where @xmath100 lies in a possibly varying model space of curvature @xmath4 . _ . _ the included nonnormalized angle @xmath103 is a decreasing function of @xmath107 . each shoulder nonnormalized angle , @xmath105 or @xmath106 , is an increasing function of @xmath107 . suppose @xmath98 . then the model spaces are semi - euclidean planes , and the sides of a triangle may be represented by vectors @xmath108 , @xmath109 and @xmath110 . set @xmath111 and @xmath112 , so @xmath113 since @xmath95 is an increasing function of its sidelength , part @xmath34 in any fixed model space is immediate by taking @xmath114 and @xmath115 in ( [ eq : ip1 ] ) to be fixed . for part @xmath12 in any fixed model space , it is only necessary to rewrite ( [ eq : ip1 ] ) as @xmath116 where @xmath114 and @xmath95 are fixed . a change of model space occurs when the varying point in @xmath28 moves upward on a vertical line @xmath117 , and passes either into or out of @xmath65 by crossing @xmath63 ( the same argument will hold for @xmath62 and @xmath64 ) . see figure [ fig : hinge ] . thus @xmath117 is the union of three closed segments , intersecting only at their two endpoints on @xmath63 . we have just seen that the included angle function is decreasing on each segment , since the realizing triangles are in the same model space ( by choice at the endpoints and by necessity elsewhere ) . since the values at the endpoints are the same from left or right , the included angle function is decreasing on all of @xmath117 . similarly , each shoulder angle function is increasing . suppose @xmath74 . the vertices of a triangle in the quadric model space are also the vertices of a triangle in an ambient @xmath12-plane , whose sides are the chords of the original sides . the length of the chord is an increasing function of the original sidelength . thus to derive the lemma for @xmath74 from ( [ eq : ip1 ] ) and ( [ eq : ip2 ] ) , we must verify the following : if a triangle in a quadric model space varies with fixed sidelengths adjacent to one vertex , and @xmath118 are the tangent vectors to the sides at that vertex , then @xmath119 is an increasing function of @xmath120 where the @xmath121 are the chordal vectors of the two sides . indeed , all points of a distance circle of nonzero radius in the quadric model space lie at a fixed nonzero ambient distance from the tangent plane at the centerpoint . thus @xmath121 is a linear combination of @xmath122 and a fixed normal vector @xmath123 to the tangent plane , where the coefficients depend only on the sidelength @xmath124 . the desired correlation follows . by switching the sign of the metric , we obtain the claim for @xmath75 . the law of cosines in a semi - riemannian model space with @xmath98 is ( [ eq : ip1 ] ) . if @xmath30 , the law of cosines for @xmath100 may be written in unified form as follows : @xmath125 here we assume @xmath100 satisfies the size bounds for @xmath4 . then each sidelength is @xmath70 if @xmath74 , and @xmath126 if @xmath75 . part 1 of lemma [ lem : hinge ] can be derived from ( [ eq : coslaw ] ) as follows . fix @xmath127 and @xmath128 , and observe that @xmath129 is decreasing in @xmath95 if @xmath74 , regardless of the sign of @xmath95 and even as @xmath95 passes through @xmath29 , and increasing in @xmath95 if @xmath75 . the size bounds imply that the factors @xmath130 become either @xmath130 for @xmath131 , or @xmath132 , depending on the signs of @xmath4 and @xmath81 , and hence are nonnegative . now we are ready to prove a semi - riemannian version of alexandrov s straightening lemma , according to which a triangle inherits comparison properties from two smaller triangles that subdivide it . it turns out that the comparisons we need are on nonnormalized shoulder angles . moreover , the original and `` subdividing '' triangles may lie in varying model spaces , so that geometrically we have come a long way from the original interpretation in terms of hinged rods . since geodesics are parametrized by @xmath32 $ ] , a point @xmath133 on a directed side of a triangle inherits an affine parameter @xmath134 $ ] . [ lem : snap ] suppose @xmath135 is a triangle satisfying size bounds for @xmath4 in a model space of curvature @xmath4 . let @xmath136 be a point on side @xmath137 , and set @xmath138 . let @xmath139 and @xmath140 be triangles in respective model spaces of curvature @xmath4 , where @xmath141 , @xmath142 , @xmath143 , @xmath144 , and @xmath145 . assume @xmath146 if @xmath74 , and @xmath147 if @xmath75 . if @xmath148 then @xmath149 the same statement holds with all inequalities reversed . by the definition of nonnormalized angles , @xmath150 therefore , by hypothesis , either @xmath151 or @xmath152 . by lemma [ lem : hinge].2 , the inequality @xmath153 , and hence for both . but then by lemma [ lem : hinge].1 , the claim follows . in this section we give a unified proof that in the model spaces of curvature @xmath4 , the restrictions to geodesics @xmath37 of the modified distance functions @xmath154 defined by ( [ eq : mode ] ) satisfy the differential equation @xmath155 we begin by constructing the @xmath4-affine functions on the model spaces . for intrinsic metric spaces the notion of a @xmath4-affine function was considered in @xcite and their structural implications were pursued in @xcite . for semi - riemannian manifolds the definition should be formulated to account for the causal character of geodesics , as follows . a _ @xmath4-affine function _ on a semi - riemannian manifold is a real - valued function @xmath156 such that for every geodesic @xmath37 the restriction satisfies @xmath157 we say @xmath156 is _ @xmath4-concave _ if `` @xmath158 '' holds in ( [ eq : kaffine ] ) , and _ @xmath4-convex _ if `` @xmath159 '' holds . ( elsewhere we have called the latter classes _ @xmath160-concave / convex_. ) as in the riemannian case , the @xmath161-dimensional model spaces of curvature @xmath4 carry an @xmath162-dimensional vector space of @xmath4-affine functions , namely , the space of restrictions of linear functionals in the ambient semi - euclidean space of a quadric surface model . specifically , let @xmath163 be the semi - euclidean space of index @xmath21 . for @xmath30 , set @xmath164 , with the induced semi - riemannian metric , so that @xmath165 is an @xmath161-dimensional space of constant curvature @xmath4 . ( the @xmath12-dimensional model spaces @xmath10 are the universal covers of such quadric surfaces . ) for @xmath166 , let @xmath167 be the restriction to @xmath165 of the linear functional on @xmath168 dual to the element @xmath42 , namely , @xmath169 . define @xmath58 on @xmath165 by ( [ eq : edef2 ] ) . [ prop : kaffine]for @xmath30 , the function @xmath170 on @xmath165 is @xmath4-affine . for any @xmath47 that is joined to @xmath42 by a geodesic in @xmath165 , @xmath171 where the argument of cosine may be imaginary . we use the customary identification of elements of @xmath168 with tangent vectors to @xmath168and @xmath165 . then the gradient of the linear functional @xmath172 on @xmath168 is @xmath42 , viewed as a parallel vector field . for @xmath173 , projection @xmath174 is given by @xmath175 . in particular , @xmath176 . it is easily checked that @xmath177 . the connection @xmath178 of @xmath165 is related to the connection @xmath179 of @xmath168 by projection , that is , @xmath180 for @xmath181 . writing @xmath182 , @xmath183 for a geodesic @xmath37 of @xmath165 , then @xmath184 thus @xmath170 is @xmath4-affine . since @xmath42 is orthogonal to the tangent plane @xmath185 , the derivatives of @xmath170 at @xmath42 are all @xmath29 . along a geodesic @xmath37 in @xmath165 that starts at @xmath42 , the initial conditions for @xmath186 are @xmath187 , @xmath188 , so the formula for @xmath189 is @xmath190 . for the case @xmath191 we consider the quadric surface model to be a hyperplane not through the origin , so that the affine functions on it are trivially the restrictions of linear functionals . on a model space @xmath165 of curvature @xmath30 , the modified distance function @xmath154 defined by ( [ eq : mode ] ) may be written on its domain as @xmath192 and satisfies the same differential equation along geodesics as @xmath170 except for an additional constant term , that is , @xmath154 satisfies ( [ eq : modkaffine ] ) . it is trivial to check that this equation holds when @xmath191 and @xmath193 . in a given semi - riemannian manifold @xmath35 , set @xmath194 ( as in ( [ eq : mode ] ) ) for some fixed choice of @xmath4 and @xmath42 . define the _ modified shape operator _ @xmath56 , on the region where @xmath195 is smooth , to be the self - adjoint operator associated with the hessian of @xmath195 , namely , @xmath196 the form of @xmath195 was chosen so that in a _ model space _ @xmath165 , @xmath197 is always a scalar multiple of the identity . indeed , at any point in @xmath165 , @xmath198 where the latter equality is by proposition [ prop : kaffine ] and ( [ eq : hess ] ) . below , our riccati equation ( [ eq : riccati ] ) along radial geodesics @xmath199 from @xmath42 differs from the standard one in @xcite and @xcite , being adjusted to facilitate the proof of theorem [ thm : comparison ] . thus it applies even if @xmath199 is null ; it concerns an operator @xmath197 that is defined on the whole tangent space ; when @xmath199 is nonnull , the restriction of @xmath197 to the normal space of @xmath199 does not agree with the second fundamental form of the equidistant hypersurface but rather with a rescaling of it ; and we do not differentiate with respect to an affine parameter along @xmath199 , but rather use the integral curve parameter of @xmath200 . the gradient vector field @xmath201 is tangent to the radial geodesics from @xmath42 . note that @xmath202 is nonzero along null geodesics radiating from @xmath42 even though @xmath195 vanishes along such geodesics . specifically , @xmath202 may be expressed in terms of @xmath203 on a normal coordinate neighborhood via ( [ eq : mode ] ) . here @xmath204 , where @xmath9 is the image under @xmath205 of the position vector field @xmath206 on @xmath207 ( see @xcite ) . if @xmath98 , then @xmath208 , and an affine parameter @xmath209 on a radial geodesic from @xmath42 is given in terms of the integral curve parameter @xmath210 of @xmath202 by @xmath211 with @xmath212 at @xmath29 . if @xmath30 , then @xmath213 , so @xmath202 agrees with @xmath9 up to higher order terms , and the dominant term at @xmath42 in the integral curve expression is an exponential . let @xmath214 be the self - adjoint _ ricci operator _ , @xmath215 . we are going to establish comparisons on modified shape operators , governed by comparisons on ricci operators . since we are interested in comparisons along two given geodesics , each radiating from a given basepoint , the effect of restricting to normal coordinate neighborhoods in the following proposition is merely to rule out conjugate points along both geodesics . [ prop : riccati ] in a semi - riemannian manifold @xmath35 , on a normal coordinate neighborhood of @xmath42 , the modified shape operator @xmath197 satisfies the first - order pde @xmath216 before verifying proposition [ prop : riccati ] , we shift to the general setting of systems of ordinary differential equations in order to summarize all we need about jacobi and riccati equations . [ lem : jacobiriccati ] for self - adjoint linear maps @xmath217 on a semi - euclidean space , suppose @xmath218 satisfies @xmath219 for @xmath220 $ ] , where @xmath221 , @xmath222 is invertible , and @xmath218 is invertible for all @xmath223 $ ] . for a given function @xmath224\to{\mathbf{r}}$ ] with @xmath225 , @xmath226 , and @xmath227 on @xmath228 $ ] , define @xmath197 by @xmath229,\ ] ] and @xmath230 then @xmath197 is self - adjoint , smooth on @xmath231 $ ] , and satisfies @xmath232 self - adjointness of @xmath197 follows from ( [ eq : jacobi ] ) and self - adjointness of @xmath233 ( see @xcite ) . by ( [ eq : sdef2 ] ) and ( [ eq : jacobi ] ) , on @xmath228 $ ] we have @xmath234 multiplying the first and last expressions by @xmath235 on the right yields ( [ eq : generalriccati ] ) . on @xmath231 $ ] we have @xmath236 where @xmath237 , and @xmath238 where @xmath239 is invertible . then ( [ eq : sdef2 ] ) gives @xmath240 on @xmath228 $ ] . by ( [ eq : sdef2(0 ) ] ) , @xmath241 on @xmath231 $ ] , so @xmath197 is smooth there . comparisons of solutions of ( [ eq : generalriccati ] ) will be in terms of the notion of positive definite and positive semi - definite self - adjoint operators @xcite . a linear operator @xmath242 on a semi - euclidean space is _ positive definite _ if @xmath243 for every @xmath244 , _ positive semi - definite _ if @xmath245 . we then write @xmath246 if @xmath247 is positive definite , and similarly for @xmath248 . note that the identity map @xmath249 is not positive definite if the index is positive ; however , the eigenvalues of a positive definite operator @xmath242 are real . if @xmath250 and @xmath251 , then @xmath252 . in @xcite , a comparison theorem for the shape operators of tubes in semi - riemannian manifolds is stated without proof . for the proof of theorem [ thm : comparison ] we require a stronger version of the special case in which the central submanifolds are just points , so the shape operators of distance - spheres are compared ; the strengthening comes from the extension to modified shape operators . since it is a key result for us , we now show how this version can be derived from a modification of the comparison theorem proved in @xcite , together with a taylor series argument to cover the behavior at the base - point singularity . [ thm : ahcompare ] let @xmath253 and @xmath254 ( @xmath255 ) be as in lemma [ lem : jacobiriccati ] , and assume @xmath256 . if @xmath257 for all @xmath258 $ ] , then @xmath259 on @xmath231 $ ] . if @xmath260 , then @xmath261 on @xmath231 $ ] . first we show that ( [ eq : generalriccati ] ) and the initial data for @xmath253 imply @xmath262 and @xmath263 to see this , differentiate ( [ eq : generalriccati ] ) , obtaining @xmath264 applying the initial data for @xmath253 and @xmath265 gives ( [ eq : a ] ) . now cancel the @xmath266 terms and differentiate again : @xmath267 setting @xmath268 gives ( [ eq : b ] ) . now for @xmath269 , let @xmath270 , where @xmath271 is a positive definite self - adjoint operator , constant as a function of @xmath209 . the solutions @xmath272 of @xmath273 with @xmath274 and @xmath275 depend continuously on the parameter @xmath276 , approaching the solution @xmath277 of @xmath278 . in particular , @xmath279 is invertible for all @xmath220 $ ] if @xmath276 is sufficiently small . define @xmath280 as in ( [ eq : jacobi ] ) , ( [ eq : sdef2 ] ) with @xmath281 . since @xmath282 , setting @xmath283 and @xmath284 in ( [ eq : b ] ) implies @xmath285 . since @xmath286 , and @xmath287 by ( [ eq : a ] ) , then @xmath288 for all @xmath289 , where @xmath80 depends on @xmath276 . but then @xmath288 for @xmath290 $ ] . our argument for this follows @xcite , except for showing that the additional linear term in ( [ eq : generalriccati ] ) is harmless . namely , assume the statement is false . then there exists @xmath291 $ ] for which @xmath292 , @xmath293 is not positive definite , and @xmath288 for @xmath294 . hence there is a nonzero vector @xmath295 such that @xmath296 , and so @xmath297 . for @xmath298 , then by ( [ eq : generalriccati ] ) , @xmath299 this contradicts @xmath300 , which is true because @xmath301 on @xmath302 and @xmath303 . since @xmath288 for all @xmath290 $ ] , and @xmath304 for all @xmath258 $ ] , we have @xmath305 $ ] . returning to the geometric setting , let us verify proposition [ prop : riccati ] . _ proof of proposition [ prop : riccati ] . _ let @xmath123 be the unit radial vector field tangent to nonnull geodesics from @xmath42 . by continuity , it suffices to verify ( [ eq : riccati ] ) at every point that is joined to @xmath42 by a nonnull geodesic @xmath199 . first we check that ( [ eq : riccati ] ) holds when applied to @xmath306 . note that the modified shape operator @xmath197 satisfies @xmath307 indeed , the form of @xmath308 along a unitspeed radial geodesic from the basepoint is the same in all manifolds , hence the same in @xmath35 as in a model space . but in a model space , ( [ eq : models ] ) and ( [ eq : modelh ] ) imply @xmath309 . therefore @xmath310 as required . now we verify that ( [ eq : riccati ] ) holds on @xmath311 . if @xmath35 has dimension @xmath161 and index @xmath21 , consider an isometry @xmath312 . for a nonnull , unitspeed geodesic @xmath199 in @xmath35 radiating from @xmath42 , identify @xmath313 with @xmath314 by parallel translation to the base point composed with @xmath315 . thus we identify linear operators on @xmath313 and @xmath314 , and likewise on @xmath316 and the corresponding @xmath317-dimensional subspace of @xmath314 . if we restrict to @xmath318 , and set @xmath319 and @xmath320 , then ( [ eq : jacobi ] ) becomes the jacobi equation for normal jacobi fields , and the operator defined by ( [ eq : sdef2 ] ) is @xmath321 , the weingarten operator , for @xmath322 : @xmath323 ( see @xcite , which uses the opposite sign convention for @xmath324 . ) if instead we set @xmath319 as before but @xmath325 where @xmath201 , so that @xmath326 and @xmath327 for @xmath328 , then the operator @xmath329 defined by ( [ eq : sdef2 ] ) and ( [ eq : sdef2(0 ) ] ) is the restriction to @xmath17 of the modified shape operator , for @xmath330 . indeed , ( [ eq : sdef2 ] ) implies @xmath331 for @xmath322 , hence @xmath332 which agrees with the definition ( [ eq : sdef1 ] ) of the modified shape operator . and the modified shape operator is the identity at @xmath42 by ( [ eq : radial ] ) , since @xmath123 can be chosen to be any unit vector at @xmath42 . then it is straightforward from ( [ eq : generalriccati ] ) that the restriction to @xmath17 of the modified shape operator satisfies ( [ eq : riccati ] ) . the proof of the rigidity statement proceeds just as in @xcite . to summarize , ( * ? ? ? * theorem 3.2 ) applies to the weingarten operator of the equidistant hypersurfaces from a _ hypersurface_. in that case , both @xmath233 and @xmath333 are perturbed in order to obtain a strict inequality on operators ; if instead we considered the modified weingarten operator @xmath334 , so @xmath335 , we would perturb @xmath233 and @xmath336 . on the other hand , theorem [ thm : ahcompare ] above applies to @xmath337 , where @xmath324 is the weingarten operator of the equidistant hypersurfaces from a _ point_. here we had @xmath265 and @xmath338 , and showed that merely perturbing @xmath233 implied a desired perturbation of @xmath339 and hence of @xmath340 for small @xmath81 . the theorem stated without proof in @xcite applies to the intermediate case of equidistant hypersurfaces from any submanifold @xmath117 . except for changes in details , our proof above works for that case as well . now let us compare modified shape operators via theorem [ thm : ahcompare ] . we say two geodesic segments @xmath199 and @xmath341 in semi - riemannian manifolds @xmath35 and @xmath342 _ correspond _ if they are defined on the same affine parameter interval and satisfy @xmath343 . [ cor : scompare1 ] for semi - riemannian manifolds @xmath35 and @xmath342 of the same dimension and index , suppose @xmath199 and @xmath341 are corresponding nonnull geodesic segments radiating from the basepoints @xmath36 and @xmath344 and having no conjugate points . identify linear operators on @xmath313 with those on @xmath345 by parallel translation to the basepoints , together with an isometry of @xmath207 and @xmath346 that identifies @xmath347 and @xmath348 . if @xmath349 at corresponding points of @xmath199 and @xmath341 , then the modified shape operators satisfy @xmath350 at corresponding points of @xmath199 and @xmath341 . the modified shape operators split into direct summands , corresponding to their action on the one - dimensional spaces tangent to the radial geodesics and on the orthogonal complements @xmath17 . the first summand is the same for both @xmath351 and @xmath342 . the second summand is as described in lemma [ lem : jacobiriccati ] with @xmath319 and @xmath325 . ( since our identification of @xmath313 and @xmath345 identifies @xmath202 and @xmath352 , we denote both of these by @xmath202 . ) furthermore , @xmath353 by ( [ eq : radial ] ) , so @xmath256 by ( [ eq : mode ] ) . therefore the corollary follows from theorem [ thm : ahcompare ] . [ cor : scompare2 ] suppose @xmath35 is a semi - riemannian manifold satisfying @xmath0 , and @xmath354 has the same dimension and index as @xmath35 and constant curvature @xmath4 . then for any @xmath355 that is joined to @xmath42 by a geodesic that has no conjugate points and such that a corresponding geodesic segment in @xmath342 has no conjugate points , the modified shape operator @xmath56 satisfies @xmath356 the same statement holds with inequalities reversed . let @xmath199 be the given geodesic from @xmath42 to @xmath357 , and @xmath341 be a corresponding geodesic from @xmath358 to @xmath359 . if @xmath199 is nonnull , then by corollary [ cor : scompare1 ] , ( [ eq : models ] ) and ( [ eq : modelh ] ) , we have @xmath360 where @xmath361 denotes the identity operator on @xmath362 , and @xmath363 are identified by parallel translation to @xmath364 followed by an isometry identifying @xmath365 . corollary [ cor : scompare1 ] applies here because the righthand side of ( [ eq : def ] ) is @xmath366 , and so @xmath367 at corresponding points of @xmath199 and @xmath341 . since @xmath368 , then ( [ eq : scompare2 ] ) holds at @xmath47 . therefore ( [ eq : scompare2 ] ) holds everywhere by continuity . now we are ready to prove that in a semi - riemannian manifold @xmath35 , triangle comparisons hold in any normal neighborhood @xmath31 in which there is a curvature bound @xmath4 and triangles satisfy size bounds for @xmath4 . by the realizability lemma , such a @xmath100 has a model triangle @xmath369 , which in this section we embed in @xmath370 , where @xmath165 is taken of the same dimension and index as @xmath31 . there are several equivalent formulations of the triangle comparisons we seek : [ prop : equivcomp ] the following conditions on all triangles in @xmath31 are equivalent : _ . _ the signed distance between any two points is @xmath6 ( @xmath7 ) the signed distance between the corresponding points in the model triangle . the signed distance from any vertex to any point on the opposite side is @xmath6 ( @xmath7 ) the signed distance between the corresponding points in the model triangle . the nonnormalized angles are @xmath7 ( @xmath6 ) the corresponding nonnormalized angles of the model triangle . @xmath34 obviously implies @xmath12 . conversely , for @xmath100 in @xmath31 , suppose @xmath133 is on side @xmath371 and @xmath161 is on side @xmath46 , and @xmath372 and @xmath373 are the corresponding affine parameters . let @xmath374 be the model triangle for @xmath100 , @xmath375 be the model triangle for @xmath376 , and @xmath377 be the model triangle for @xmath378 . let @xmath379 on @xmath380 and @xmath381 on @xmath382 have affine parameters @xmath372 and @xmath373 , and similarly for @xmath383 on @xmath384 . by 2 , @xmath385 . therefore by lemma [ lem : hinge].1 ( hinge ) , @xmath386 again by 2 , latexmath:[$|{\widetilde}{m}'\,{\widetilde}{q}'|=|mq|\ge @xmath376 , together with ( [ eq : comp1 ] ) , we have @xmath388 again by hinge , @xmath389 is a direct consequence of the first variation formula ( see @xcite ) : @xmath390 ( note that our definition of @xmath391 and oneill s differ by a factor of @xmath12 . ) conversely , using the same triangle notation as above , 3 gives @xmath392 , and similarly @xmath393 . since @xmath394 , we have @xmath395 . by lemma [ lem : snap ] ( straightening ) , @xmath396 . therefore by hinge , @xmath397 , consider @xmath100 in @xmath31 , and its model triangle @xmath398 , which we regard as lying in @xmath354 . taking @xmath42 and @xmath399 as base points gives modified distance functions @xmath400 and @xmath401 . for any @xmath402 , the signed distance @xmath403 is a monotone increasing function of @xmath404 , and distances from @xmath399 in @xmath165 have exactly the same relation with @xmath405 . thus the following proposition shows that curvature bounds imply triangle comparisons in the sense of proposition [ prop : equivcomp].2 , thereby proving the `` only if '' part of theorem [ thm : comparison ] . [ prop : onlyif]set @xmath406 and @xmath407 . if @xmath26 in @xmath31 , then @xmath408 . if @xmath1 in @xmath31 , then @xmath409 . assume @xmath26 . aside from reversing inequalities the proof for @xmath1 is just the same . set @xmath410 and @xmath411 . for @xmath412 , by corollary [ cor : scompare2 ] , the modified shape operator @xmath56 satisfies @xmath413 since , by definition , @xmath414 is the second derivative of @xmath154 along the geodesic with velocity @xmath415 , then @xmath416 that is , along @xmath37 , @xmath154 satisfies the differential inequality @xmath417 on the other hand , the above inequalities become equations in @xmath165 , so @xmath418 but @xmath419 since @xmath420 is a model segment for @xmath37 . hence the difference @xmath421 is @xmath422-concave : @xmath423 moreover , at @xmath29 and @xmath34 the values of @xmath195 and @xmath424 are the same since @xmath425 and @xmath426 , so the end values of @xmath156 are just @xmath427 . by concavity @xmath156 is bounded below by the @xmath422-affine function with those end values , which is just @xmath29 . that is , @xmath428 , or @xmath429 . next we verify the `` if '' part of theorem [ thm : comparison ] : if signed distances between pairs of points on any triangle in @xmath31 are at least ( at most ) those between the corresponding points of the comparison triangle , then @xmath26 ( @xmath1 ) . let @xmath199 be a nonnull geodesic segment in @xmath31 , let @xmath430 be nonnull and perpendicular to @xmath347 , and let @xmath431 be the jacobi field along @xmath199 such that @xmath432 . in the @xmath12-dimensional model space @xmath433 of curvature @xmath4 and of the same signature as the section spanned by @xmath347 and @xmath415 , choose a geodesic @xmath434 and vector @xmath435 at @xmath436 perpendicular to @xmath347 such that @xmath437 and @xmath438 . let @xmath439 be the jacobi field on @xmath434 such that @xmath440 . write @xmath441 and similarly for @xmath442 . since @xmath443 , then @xmath444 is equal to the corresponding expression in @xmath342 . but then our triangle comparison assumption , in the form given in proposition [ prop : equivcomp].3 , and lemma [ lem : hinge].1 ( hinge ) combine to give @xmath445 and similarly in @xmath342 , we conclude @xmath446 now we calculate the third order taylor expansion of @xmath431 . @xmath447 @xmath448 and hence @xmath449 where @xmath450 is parallel translation from @xmath451 to @xmath452 and the primes indicate @xmath453 . then we get an expansion @xmath454 and a similar expansion for @xmath455 . since the @xmath456-terms are the same , we must have the inequality for the @xmath457-terms : @xmath458 since @xmath459 and @xmath415 span an arbitrary nonnull section , @xmath26 follows . curvature bounds of the type studied in this paper are clarified by the analysis by beem and parker of the pointwise ranges of sectional curvature @xcite , as we now explain . we go further , to relate our curvature bounds to the `` null '' curvature bounds considered by uhlenbeck @xcite and harris @xcite . since in a semi - riemannian manifold with indefinite metric , a spacelike section always lies in a lorentz or anti - lorentz @xmath16-plane @xmath17 , the range of sectional curvature may be studied by restricting to such @xmath16-planes @xmath17 . on @xmath17 , unless the curvature is constant , both the time - like and space - like sections have infinite intervals as their range , and either both are the entire real line or both are rays which overlap in at most a common end ( see theorem [ thm : kvalues ] ) . then as we vary @xmath17 in the tangent bundle , either the separation between the two rays can be lost or we can have numbers that separate all pairs of intervals , namely , a curvature bound in our sense . in this section , @xmath17 always denotes a lorentz or anti - lorentz @xmath16-plane . following @xcite , consider a curvature tensor @xmath233 on @xmath17 . express @xmath233 as a homogeneous quadratic form @xmath460 on @xmath461 . if @xmath462 is a frame for which @xmath463 and @xmath464 have the same signature , then @xmath465 is a frame for @xmath461 with signature @xmath466 with respect to the natural extension of the inner product . every nonzero element @xmath467 of @xmath461 is decomposable , and so represents a oriented section of @xmath17 , so the projective plane @xmath468 of all nonorientable sections of @xmath17 has homogeneous coordinates @xmath469 . the inner product quadratic form on @xmath461 has the coordinate expression @xmath470 , and the sectional curvature function is @xmath471 . we also identify @xmath472 and @xmath473 with the quadratic functions on @xmath474 given in terms of the corresponding nonhomogeneous coordinates @xmath475 by @xmath476 and @xmath477 . for various curvature tensors there is no restriction on @xmath472 ; that is , for a given point @xmath47 in any @xmath161-dimensional manifold @xmath35 , and a given @xmath16-dimensional subspace @xmath17 of @xmath478 , a semi - riemannian metric with indefinite restriction to @xmath17 can be specified in a neighborhood of @xmath47 in terms of normal coordinates so as to realize any curvature tensor on @xmath17 . the _ null conic _ @xmath123 is given by @xmath479 , and represents those sections of @xmath17 on which the inner product is degenerate and @xmath480 is undefined . the _ homaloidal ( flat ) conic _ @xmath481 is given by @xmath482 . the inclusion @xmath483 is equivalent to @xmath484 being constant on the sections of @xmath17 , which is to say , @xmath472 being @xmath473 multiplied by that constant value ( which may be @xmath29 so the inclusion could be proper ) . otherwise , @xmath481 and @xmath123 intersect in at most @xmath485 points , counting multiplicities . the points of odd multiplicity are precisely the points where @xmath481 and @xmath123 cross . since the interior and exterior of @xmath123 are connected sets on which @xmath484 is continuous , the ranges of @xmath484 on time - like sections and space - like sections of @xmath17 are intervals , @xmath486 and @xmath487 . the following theorem characterizes the possible ranges . it implies , in particular , that if on @xmath17 either timelike or spacelike curvatures are bounded , then both are , and there exists a curvature bound in our sense . [ thm : kvalues]for a curvature tensor on a lorentz or anti - lorentz @xmath16-plane : _ . _ @xmath484 is constant if @xmath488 . @xmath489 if @xmath481 and @xmath123 cross . @xmath487 and @xmath486 are oppositely directed closed half - lines , separated by a nontrivial open interval of curvature bounds , if @xmath481 does not intersect @xmath123 ( including the cases when @xmath481 is empty or a point not in @xmath123 ) . @xmath487 and @xmath486 are oppositely directed half - lines with a common endpoint otherwise , namely , when @xmath481 and @xmath123 have a point of tangency but never cross . more specifically , @xmath487 and @xmath486 are both open , both closed , or complementary , according as @xmath481 and @xmath123 intersect in a single point of order @xmath12 , two points of order @xmath12 , or a single point of order @xmath485 . in a semi - riemannian manifold with indefinite metric , @xmath0 holds if and only if the restriction of the curvature tensor to each lorentz or anti - lorentz @xmath16-plane @xmath17 satisfies @xmath0 ( and similarly for @xmath1 ) . equivalently , on each @xmath17 , either @xmath484 is constantly @xmath4 , or @xmath486 is a semi - infinite interval in @xmath490 $ ] and @xmath487 is a semi - infinite interval in @xmath491 . theorem [ thm : kvalues ] leads us to consider a weaker condition , which we denote by @xmath18 , in which the interval betweeen @xmath486 and @xmath487 varies with the indefinite @xmath16-plane @xmath17 , and there may be no @xmath4 common to all . write @xmath492 if @xmath493 for any null vector @xmath494 and non - zero vector @xmath415 perpendicular to @xmath494 . it is shown in ( * ? ? ? * proposition 2.3 ) ( or see ( * ? ? ? * proposition a.7 ) ) that if @xmath495 at a point , then the range of timelike sectional curvatures at that point is unbounded below ( above ) . the following proposition gives precise information . [ prop : nullcurv ] a semi - riemannian manifold with indefinite metric satisfies @xmath496 if and only if @xmath18 , and similarly with signs reversed . in a given lorentz or anti - lorentz @xmath16-plane @xmath17 , the condition @xmath497 is equivalent to @xmath498 on the null conic @xmath123 . in turn this implies that @xmath123 and @xmath481 do not cross , and hence cases 1 , 3 or 4 of theorem [ thm : kvalues ] hold . in case 1 , obviously there is a lower curvature bound . in cases 3 and 4 , there are points of @xmath123 at which @xmath499 . approaching @xmath123 from the spacelike side gives @xmath500 , so @xmath487 is unbounded above and again @xmath17 has a lower curvature bound . conversely , suppose there is a lower curvature bound for @xmath17 , so case 2 is ruled out . in case 1 , @xmath501 on @xmath123 . in cases 3 or 4 , since @xmath487 is bounded below , there can not be points of @xmath123 at which @xmath502 . the condition @xmath503 plus a `` growth condition '' was used in @xcite to prove a hadamard - cartan theorem for lorentz manifolds . it seems interesting to investigate the relation between @xmath23 and these hypotheses ; uhlenbeck comments about the growth condition,``it is to be hoped that a similar condition that does not depend on coordinates can be found '' @xcite . the condition @xmath504 ( or @xmath505 ) isolates case 3 of theorem [ thm : kvalues ] . now let us show how a strengthening of this condition bounds below the length of the interval of curvature bounds in each lorentz or anti - lorentz @xmath16-plane @xmath17 . while sectional curvature is undefined for null sections , harris has used a substitute , relative to a choice of null vector @xmath494 . namely , for a null section @xmath506 containing @xmath494 , define the _ null curvature of @xmath506 with respect to @xmath494 _ by @xmath507 for any non - null vector @xmath508 in @xmath506 @xcite . while there is no _ a priori _ way to normalize the null vector @xmath494 , it is still possible to strengthen proposition [ prop : nullcurv ] . this is because , in the presence of an interval of curvature bounds larger than a single point , the algebra of the curvature operator @xmath509 selects a distinguished timelike unit vector @xmath209 , or `` observer '' , and hence a distinguished circle of null vectors @xmath494 . in the following proposition , we suppose @xmath17 is lorentz ( that is , has signature @xmath510 ) . there are obvious sign changes if @xmath511 is lorentz . suppose there is an interval @xmath512 $ ] of curvature bounds below on the lorentz @xmath16-plane @xmath17 , where @xmath513 . then @xmath514 is diagonalizable . let @xmath209 be a unit timelike vector perpendicular to the spacelike eigenbivector of @xmath514 . then @xmath515 where @xmath415 runs over unit vectors perpendicular to @xmath209 , and @xmath494 and @xmath506 are the null vector and null section @xmath516 and @xmath517 respectively . for curvature bounds above , substitute @xmath518 for ( [ eq : bdinterval ] ) . we consider the case of curvature bounds below . first observe that , while self - adjoint linear operators in indefinite inner product spaces are not always diagonalizable , our hypotheses imply diagonalizability . indeed , the unit eigenbivectors of @xmath514 , of which one is spacelike and two are timelike , are the critical points of the corresponding quadratic form on unit bivectors . the values of this quadratic form are sectional curvatures , up to sign . therefore @xmath519 , the minimum spacelike sectional curvature , and @xmath520 , the maximum timelike sectional curvature , are eigenvalues , which are distinct by hypothesis . the corresponding eigenbivectors span a nondegenerate @xmath12-dimensional subspace of @xmath461 ; a bivector perpendicular to both is an eigenbivector by self - adjointness . thus our eigenbivectors diagonalize @xmath514 . let @xmath521 be a frame of vectors perpendicular to the eigensections , so that @xmath522 and @xmath523 are the timelike eigenbivectors . then the null vectors @xmath516 have the form @xmath524 , and the null curvatures @xmath525 have the form @xmath526 where @xmath527 . thus the minimum is @xmath528 . if @xmath271 and @xmath529 are riemannian manifolds , @xmath530 will denote the product manifold with the warped product metric @xmath531 . the sectional curvature @xmath484 of @xmath530 , in terms of the sectional curvatures @xmath532 and @xmath533 , may be calculated for a frame @xmath534 , for @xmath535 and @xmath536 . without loss of generality , suppose @xmath537 . let @xmath202 be the gradient of @xmath156 . then @xmath538\\ & + f^{-2}(p)\left[\mathcal{k}_f(v\wedge w ) - < g(p),g(p)>\right ] < v , v > < w , w>.\end{aligned}\ ] ] therefore : [ prop : wp ] consider riemannian manifolds @xmath271 and @xmath529 , and a smooth function @xmath539 . then @xmath540 is a semi - riemannian manifold satisfying @xmath0 ( @xmath1 ) if and only if the following three conditions hold : _ . _ @xmath156 is @xmath541-concave ( @xmath541-convex ) . @xmath542 or @xmath271 has sectional curvature @xmath543 ( @xmath2 ) , @xmath544 , or for all points @xmath545 and @xmath12-planes @xmath546 tangent to @xmath529 , @xmath547 ( @xmath7 ) @xmath548 . taking @xmath271 to be an interval @xmath249 in proposition [ prop : wp ] , we easily construct a rich class of lorentz examples : [ cor : wp]if @xmath549 is @xmath541-concave and @xmath529 is a riemannian manifold of sectional curvature @xmath550 , then @xmath551 satisfies @xmath552 for any @xmath4 in the interval @xmath553.\ ] ] if @xmath549 is @xmath541-convex and @xmath529 is a riemannian manifold of sectional curvature @xmath554 , then @xmath551 satisfies @xmath555 for any @xmath4 in the interval @xmath556.\ ] ] following @xcite , by a _ robertson - walker space _ we mean a warped product @xmath557 where @xmath529 is @xmath16-dimensional spherical , hyperbolic or euclidean space , say with curvature @xmath558 . then the sectional curvatures of sections containing @xmath559 are @xmath560 , and those of sections @xmath506 tangent to the fiber are @xmath561 . by corollary [ cor : wp ] , @xmath35 satisfies @xmath0 if and only if @xmath562 . it is easy to check that a robertson - walker space satisfies the _ strong energy condition _ , @xmath563 for all timelike vectors @xmath209 , if and only if the curvature restricted to each tangent @xmath485-plane has a nonpositive curvature bound below in our sense ( see ( * ? ? ? * exercise 10 , p. 362 ) ) . by the einstein equation , taking the cosmological constant @xmath564 , the stress - energy tensor of any robertson - walker space has the form of a perfect fluid whose energy density @xmath565 and pressure @xmath47 are functions of @xmath209 given by ( see @xcite ) : @xmath566 as discussed in @xcite , the conditions @xmath567 , @xmath568 for some constants @xmath81 and @xmath242 , and positive hubble constant @xmath569 for some @xmath570 , correspond to an initial big bang singularity . then @xmath571 , hence @xmath572 . therefore by ( [ eq : rw ] ) , these big bang robertson - walker spaces all satisfy @xmath22 . suppose the interval @xmath249 in these models is maximal . if @xmath573 , then @xmath249 is semi - infinite and @xmath574 , hence also @xmath575 , so @xmath29 is the only curvature bound for the entire space . however , every point has a neighborhood which has an interval of curvature bounds having @xmath29 as an interior point . if @xmath576 , then @xmath156 reaches a maximum followed by a big crunch , and @xmath577 takes a positive minimum . thus when @xmath576 , the entire space has an interval of curvature bounds with @xmath29 as an interior point . taking @xmath578 here does not change the existence of curvature bounds , but shifts them to the right by @xmath579 . in particular , a _ friedmann model _ is the special case in which @xmath564 and @xmath580 . then one can solve explicitly for @xmath156 , obtaining ( see @xcite ) : @xmath581 the first two of these solutions satisfy @xmath22 , and the third satisfies @xmath0 for all @xmath582 $ ] . vacuum spacetimes ( @xmath583 ) only have curvature bounds when they are flat . more generally , any @xmath485-dimensional einstein lorentz space with a curvature bound has constant curvature , since perpendicular sections always have the same curvature by a theorem of thorpe @xcite . we may also generate examples with higher index , that is , higher - dimensional base . the following examples ( a ) and ( b ) of curvature bounds for @xmath530 are from @xcite : \(a ) @xmath584 : take a cartesian product @xmath585 ( so @xmath586 ) , with sectional curvature@xmath3 in @xmath271 and @xmath2 in @xmath529 ( or the reverse ) . \(b ) @xmath587 : take @xmath588 , @xmath589 , and @xmath529 of sectional curvature @xmath590 . note that to achieve @xmath591 when @xmath271 is not @xmath34-dimensional , @xmath271 must have curvature @xmath592 . such a @xmath271 carries many @xmath593-convex functions , but by proposition [ prop : wp ] , we need the warping function @xmath156 on @xmath271 to be @xmath593-concave . a solution is to take @xmath588 and @xmath156 to be @xmath593-affine . example ( b ) fits this pattern , with the righthand side of the inequality in proposition [ prop : wp].3 equal to @xmath34 . other constructions in this pattern are : \(c ) @xmath594 : take @xmath588 , @xmath595 , and @xmath529 of sectional curvature @xmath596 . \(d ) @xmath597 : take @xmath598 , @xmath599 , and @xmath529 of sectional curvature @xmath600 . examples ( a ) - ( d ) are all geodesically complete . reversing the sign on an example that satisfies @xmath0 and is negative definite on the base , gives one that satisfies @xmath601 and is negative definite on the fiber . we thank yuri burago for the picture that triggered this project ( @xcite ) . l. andersson , g. galloway , r. howard , _ a strong maximum principle for weak solutions of quasi - linear elliptic equations with applications to lorentzian and riemannian geometry _ pure appl . math.*51 * ( 1998 ) , 819 - 877 .
a semi - riemannian manifold is said to satisfy @xmath0 ( or @xmath1 ) if spacelike sectional curvatures are @xmath2 and timelike ones are @xmath3 ( or the reverse ) . such spaces are abundant , as warped product constructions show ; they include , in particular , big bang robertson - walker spaces . by stability , there are many non - warped product examples . we prove the equivalence of this type of curvature bound with local triangle comparisons on the signed lengths of geodesics . specifically , @xmath0 if and only if locally the signed length of the geodesic between two points on any geodesic triangle is at least that for the corresponding points of its model triangle in the riemannian , lorentz or anti - riemannian plane of curvature @xmath4 ( and the reverse for @xmath1 ) . the proof is by comparison of solutions of matrix riccati equations for a modified shape operator that is smoothly defined along reparametrized geodesics ( including null geodesics ) radiating from a point . also proved are semi - riemannian analogues to the three basic alexandrov triangle lemmas , namely , the realizability , hinge and straightening lemmas . these analogues are intuitively surprising , both in one of the quantities considered , and also in the fact that monotonicity statements persist even though the model space may change . finally , the algebraic meaning of these curvature bounds is elucidated , for example by relating them to a curvature function on null sections .
regge calculus was proposed to formulate einstein s general relativity on piecewise linear manifolds @xcite . it provides coordinate - free lattice formulation of gravitation and has been used in investigations of classical as well as quamtum gravity . as is the case of qcd , lattice theoretical approach provides a powerfull framework in nonperturbative studies of quantum gravity @xcite . however , before moving to detailed quantum study , it is desired to investigate the formalism at the classical level . any lattice regularized theory should reproduce basic results of corresponding continuum theory . taking classical continuum limit is relatively easy in the case of lattice gauge theories . the reason of this is obvious . in lattice gauge theory space - time itself is not dynamical . we usually consider hypercubic regular lattice for the space - time . dynamical variables sit on sites for matter fields and on links for gauge fields . classical lattice actions are written in manifestly gauge invariant form by using plaquette variables and covariant differences whose classical continuum limits can be taken easily . though the lorentz invariance is absent from the beginning , lattice action is still invariant under discrete rotations . the lorentz invariance is considered to recover in the continuum limit . in regge calculus , however , the space - time itself should be considered as dymnamical . we replace it by a piecewise linear manifold , which is composed of a set of simplices . the basic variables are the edge lengths of the simplices . for a given piecewise linear manifold we can write the regge action . in general there is no natural choice of piecewise linear manifold . furthermore , the regge action is written in coordinate - free form . hence , the recoverance of general coordinate invariance in classcal continuum limit becomes much more complicated than that of lorentz invarinace in nonabelian lattice gauge theories . in this note we investigate friedmann - lematre - robertson - walker ( flrw ) universe of three - dimensional einstein gravity with positive cosmological constant in regge calculus . the contimuum action is given by @xmath2 it is well - known in three dimensions that the vacuum einstein equation leads to flat space - time without the cosmological term . in the case of negative cosmological constant the theory admits black hole solution @xcite and has been investigated within the context of conformal field theory @xcite . as in four dimensions the einstein equations have evolving universe as a solution for the flrw metric ansatz @xmath3 where @xmath4 for the space being spherical , euclidean and hyperspherical , respectively , and @xmath5 is the so - called scale factor . it is subject to the friedmann equations @xmath6 of these our concern is the spherical universe , which can be approximated by convex polyhedron of a finite volume . regge calculus has been applied to four - dimensional closed flrw universe by collins and williams @xcite . they considered regular polytopes as the cauchy surfaces of discrete flrw universe and used , instead of simplices , truncated world - tubes evolving from one cauchy surface to the next as the building blocks of piecewise linear space - time . their method , called the collins - williams ( cw ) formalism , is based on the idea of the @xmath7 decomposition of space - time and plays a similar role of well - known adm formalism @xcite . recently liu and williams has extensively studied the discrete flrw univese @xcite . they found that the universe with regular polytopes as the cauchy surfaces can reproduce the continuum flrw universe to a certain degree of precision . their solutions fairly agree with the continuum when the size of the universe is small , whereas the deviations from the exact results become large for the large universe because of the finite edge length . since the regge action heavily depends on the choice of polytopes to approximate the cauchy surface , it seems to be hard to take the continuum limit keeping the action simple . this motivates us to investigate simpler but less realistic three - dimensional model . in four dimensions there are six types of regular polytopes . restricting to those obtained by tessellating the three - dimensional sphere by regular tetrahedrons , there are only three with 5 , 16 and 600 cells @xcite . the foregoing investigations are mainly restricted to these regular polytopes . the regge equations , however , are still very complicated . as we shall show the situation becomes much simpler in three dimensions , where every geometric calculations can be done without complications coming from higher dimensions . this is the reason why we consider three dimensions , where the spherical cauchy surfaces are replaced by regular polyhedrons . there are five types of polyhedrons . we treat them in a unified way and give generic expressions for the regge equations , which is convenient to analyze the continuum limit . let us briefly summarize the essence of regge calculus . in regge calculus , an analog of the einstein - hilbert action is given by the regge action @xcite @xmath8 where @xmath9 is the volume of a hinge , @xmath10 the deficit angle around the hinge @xmath11 and @xmath12 the volume elements of the building blocks of the piecewise linear manifold . in three dimensions the hinges are the links , or equivalently the edges of the 3-simplices , and @xmath9 is nothing but the edge length @xmath13 . regge s original derivation is concerned with simplicial lattice , so that it describes the gravity as simplicial geometry . this formalism can easily be generalized to arbitrary lattice geometries . we can fully triangulate the non - simplicial flat blocks by adding extra hinges with vanishing deficit angles @xcite without affecting the regge action . the fundamental variables in regge calculus are the edge lengths @xmath13 . varying the regge action with respect to @xmath13 , we obtain the regge equations @xmath14 note that there is no need to carry out the variation of the deficit angles owing to the schlfli identity @xcite @xmath15 in three dimensions the regge equations simply relate the deficit angle around an edge to the total rate of variation of the volumes having the edge in common with respect to the edge length . in particular the space - time becomes flat in the absence of the cosmological term as it should be . this paper is organized as follows : in the next section we setup the regular polyhedral universe in the cw formalism and introduce the regge action . derivation of the regge equations is given in section [ sec : req ] . in the continuum time limit the regge equations are reduced to differential equations . applying the wick roation , we arrive at the equations , regge calculus analog of the freedmann equations , describing the evolution of the polyhedral universe . this is done in section [ sec : ctl ] . in section [ sec : nsol ] we solve the differential regge equation numerically and compare the scale factors of the polyhedral universe with the continuum solution . to have better approximations we introduce geodesic domes as the cauchy surface . in section [ sec : gdu ] we propose a geodesic dome universe as a pseudo - regular polyhedral universe with a fractional schrfli symbol and show that the continuum solution can be recovered in the infinite frequency limit . section [ sec : sum ] is devoted to summary and discussions . the flrw metric ( [ eq : flrwm ] ) describes expanding or contracting universe with a maximally symmetric space as the cauchy surface . surfaces of maximally symmetric compact space are spheres . it is geometrically characterized by the radius @xmath5 , the scale factor in cosmology . in this paper we shall be concerned with the regge calculus of closed flrw universe in three dimensions , which describes an evolution of two - dimensional sphere . following the cw formalism , we replace the spherical cauchy surfaces by regular polyhedrons . the fundamental building blocks of space - time are world - tubes of truncated pyramids or frustums as depicted in figure [ fig : fbb ] . in this section we restrict ourselves to regular polyhedrons as the cauchy surfaces . then every edge has equal length in each cauchy surface and so does any strut between two adjacent cauchy surfaces . evolution of the universe can be seen by focusing our attention only to expanding or shrinking of a face of the polyhedron . this considerably reduces the number of dynamical variables . it is well - known that there are only five types of regular polyhedrons , a tetrahedron , cube , octahedron , dodecahedron and icosahedron . let us denote the numbers of vertices , edges and faces of a polyhedron by @xmath16 , @xmath17 and @xmath18 , respectively . then they are constrained by euler s polyhedron formula @xmath19 a regular polyhedron is specified by the schlfli symbol @xmath20 , where @xmath21 is the numbers of sides of each face and @xmath22 the number of faces meeting at each vertex . they give rise to further constraints @xmath23 and @xmath24 . these together with ( [ eq : epf ] ) completely determine @xmath25 . in table [ tab : rph ] we summarize the properties of regular ployhedrons for reader s reference . .five regular polyhedrons in three dimensions [ cols="^,^,^,^,^,^",options="header " , ] to avoid the difficulty we regard the geodesic dome as pseudo - regular polyhedron of an average edge length @xmath26 and assume that ( [ eq : hcphu ] ) and ( [ eq : mcphu ] ) still hold true . how about the schlfli symbol @xmath20 ? since all the faces of the geodesic dome is triangle , we use @xmath27 . as for @xmath22 , we employ the average number of faces meeting at a vertex . we thus assign the fractional schlfli symbol for the geodesic dome of frequency @xmath28 as @xmath29 this recovers icosahedron for @xmath30 and approaches @xmath31 as @xmath28 gets large , which corresponds to the tessellation of flat plane with regular triangles . if we introduce average dihedral angle for the geodesic dome universe by ( [ eq:31 ] ) as in the case of regular polyhedron univese and define the scale factor as ( [ eq : rc ] ) , then the evolution of the universe can be described by ( [ eq:33 ] ) with the parameters ( [ eq : ssprp ] ) . the initial edge length is given by @xmath32 in contrast to the cases of regular polyhedrons this can be arbitrarily small as @xmath33 , or equivalently @xmath34 . the initial scale factor ( [ eq:35 ] ) , however , approaches to the value of the continuum theory as can be easily seen from @xmath35 there is no difficulty to numerically integrate ( [ eq:33 ] ) for fractional @xmath22 . we give plots of dihedral angle in figure [ fig : dhagdm ] and scale factor in figure [ fig : sfgdm ] as functions of time for @xmath36 . as we have mentioned before , the dihedral angle is a monotone decreasing function of time . the era of almost constant dihedral angle for small @xmath37 gets longer with the frequency of geodesic dome . one easily see that the results for the polyhedral universe given in sect . [ sec : ctl ] is more and more improved as @xmath28 increases . in figure [ fig : lflthgdm ] we give the dihedral angle for @xmath38 . it apporaches to a constant solution @xmath39 in the infinite frequency limit . indeed , @xmath40 satisfies ( [ eq:32 ] ) and ( [ eq:33 ] ) as one can verify directly . the scale factor for the geodesic dome universe with @xmath38 is plotted in figure [ fig : lflargdm ] . during the era of almost constant dihedral angle the agreement with the continuum scale factor can be seen immediately . the velocity of expansion @xmath41 still approaches to a constant value ( [ eq:36 ] ) as @xmath42 for finite @xmath28 . in the infinite frequency limit @xmath43 becomes arbitrarily large . this property is required for the geodesic dome universe to correctly converge to the flrw universe . : the broken line corresponds to the solution in the infinite frequency limit @xmath44 . ] : the broken curve is the scale factor of the continuum theory . ] it can be shown that the scale factor of the geodesic dome universe @xmath45 approaches to that of the continuum theory in the infinite frequency limit . the easiest way to see this is to come back to ( [ eq : hcphu ] ) and ( [ eq : mcphu ] ) . rewriting them in terms of @xmath46 and then taking the limit @xmath47 , we obtain the continuum equations of the scale factor ( [ eq : cc ] ) . this justifes the approach regarding ( [ eq : hcphu ] ) and ( [ eq : mcphu ] ) as the equations of motion of discretized flrw universe . we have investigated closed flrw universe in tree dimensions using the cw formalism in regge calculus . we have given unified expressions applicable to any regular polyhedron as the cauchy surfaces . we have shown that the regge equations in the continuum lorentzian time limit , one corresponding to hamiltonian constraint and the other to momentum constraint , describe the evolution of universe . in spite of the simplest approximations of spheres by regular polyhedrons the coincidence with the continuum solution is appreciable when the size of the universe is around the minimum , where the nonlinearity of the evolution equation ( [ eq : mcphu ] ) can be neglected . the discrepancies , however , become larger and larger with the size of the universe . in particular , the expansion speed of the universe approaches to a constant as time goes to infinity , whereas the dihedral angle around the time - like struts appoarches to zero . we have seen that the geodesic dome model can remedy the shortcoming of the regular polyhedral universe . by choosing the frequency sufficiently large the expansion velocity of the geodesic dome universe can be arbitrarirly large and the dihedral angle approaches to the natural value @xmath48 as time goes to infinity . we found that the continuum solution can be obtained in the infinite frequency limit . strictly speaking , we have not carried out the regge calculus for the geodesic domes . instead of solving exact regge equations , we have only assumed that ( [ eq : hcphu ] ) and ( [ eq : mcphu ] ) are valid for an average edge length and the fractional schlfli symbol ( [ eq : ssprp ] ) . it is desired to give further justification , numerically or analytically , to the assumption . it is also of great interest to extend our approach to four - dimensional flrw universe . we have considered closed compact universe . in the continuum theory there is an oscillating solution for negative cosmological constant , where the cauchy surface is not compact . application of our approach to hyperspherical universe might be interesting . our concern in this work was the vacuum solution . inclusion of matter is also interesting . the authors would like to thank y. hyakutake , n. motoyui , m. sakaguchi and s. tomizawa for useful discussions . this work was supported in part by grants - in - aid for scientific research numbers 24540247 from the japan society for the promotion of science . m. baados , c. teitelboim and j. zanelli , phys . * 69 * ( 1992 ) 1894 [ arxiv : hep - th/9204099 ] ; + m. baados , m. henneaux , c. teitelboim and j. zanelli , phys . rev . * d48 * ( 1993 ) 1506 [ arxiv : gr - qc/9302012 ] .
closed flrw universe of three - dimensional einstein gravity with positive cosmological constant is investigated by using the cillins - williams formalism in regge calculus . spherical cauchy surface is replaced with regular polyhedrons . the regge equations are reduced to differential equations in continuum time limit . numerical solution to the regge equations well approximate the continuum flrw universe during the era of small edge length . the deviation from the continuum solution becomes larger and larger as time goes to infinity . unlike the continuum the expansion speed of universe approaches to a constant in the infinite future . to remedy the shortcoming of the model universe we introduce geodesic domes . regarding the geodesic domes as pseudo - regular polyhedorons with fractional schlfli symbol , we can describe the behaviors of the universe more acurately in regge calculus . it is shown that the scale factor of the continuum theory can be reproduced from the regge equations in the infinite frequency limit . * expanding polyhedral universe in regge calculus * ren tsuda@xmath0 and takanori fujiwara@xmath1 _ @xmath0 graduate school of science and engineering , ibaraki university , mito 310 - 8512 , japan + @xmath1 department of physics , ibaraki university , mito 310 - 8512 , japan _
the yang - baxter equation was introduced independently by c.n . yang(1967)@xcite and r.j . it is well known that a certain solution of yang - baxter equation give rise to the jones polynomial @xcite . in 2004 , carter , elhamdadi and saito defined a ( co)homology theory for set - theoretic yang - baxter operators , from which they gave a way to generate link invariants , cocycle invariants@xcite . in 2012 , przytycki gave a graphical definition of homology for a pre - yang - baxter operator @xcite . we provide the definitions of two homology theories for set - theoretic yang - baxter operators in section 2 and show their equivalence in section 3 . in section 4 , we give definitions of one - term and two - term homology of pre - yang - baxter operators , and show examples , in particular , we find torsion in two - term homology of yang - baxter operator that yields the jones polynomial . we start from basic definitions . [ definition 1.1 ] let @xmath0 be a commutative ring and @xmath1 be a set . consider @xmath2 the free @xmath3module generated by @xmath1 . if a @xmath3linear map , @xmath4 @xmath5 , satisfies the following equation + @xmath6 + then we say @xmath7 is a pre - yang - baxter operator . the equation above is called a yang - baxter equation . if , in addition , @xmath7 is invertible , then we say @xmath7 is a yang - baxter operator . from now on , we assume @xmath0 is a commutative ring with identity whenever we deal with yang - baxter operators . [ definition 1.2 ] let @xmath1 be a set . if @xmath4 @xmath8 is a function that satisfies + @xmath9 + then we say @xmath7 is a pre - set - theoretic yang - baxter operator and the equation above is a set - theoretic yang - baxter equation . if , in addition , @xmath7 is invertible , then we say @xmath7 is a set - theoretic yang - baxter operator . pre - set - theoretic yang - baxter operator leads to yang - baxter operator by putting @xmath2 and extending @xmath4 @xmath8 to @xmath4 @xmath5 . this yang - baxter operator is still called pre - set - theoretic yang - baxter operator . given a pre - set - theoretic yang - baxter operator @xmath10 we have two approaches to homology built on @xmath7 . the algebraic " version defined in @xcite and the graphic " version in @xcite . we discuss them in the next two subsections . we prove the equivalence of them in section 3 . we first review the algebraic " version of set - theoretic yang - baxter homology theory based on@xcite and then introduce the graphic " version of set - theoretic yang - baxter homology theory . [ definition 2.1.1 ] the set @xmath1 together with a set - theoretic yang - baxter operator @xmath7 , @xmath11 is called in @xcite a yang - baxter set . we represent a function @xmath7 by @xmath12 we use the following notation . let @xmath13 be the @xmath14dimensional cube @xmath15 ( @xmath16 $ ] ) regarded as a cw ( cubical ) complex , where @xmath17 is a positive integer . where @xmath18 and co - face maps @xmath19 defined by @xmath20 ; they satisfy @xmath21 where @xmath22 denote the @xmath3skeleton by @xmath23 with orientation given by the order of coordinate axes . in particular , every @xmath24face can be written as @xmath25 for some @xmath26 where @xmath27 or @xmath28 and @xmath29 @xmath30 denote two copies of @xmath31 at the @xmath32th , @xmath33th positions , respectively . [ definition 2.1.2 ] the yang - baxter coloring of @xmath13 by a yang - baxter set @xmath34 is a map @xmath35 , where @xmath36 denotes the set of edges of @xmath37 with each edge oriented as above , such that if @xmath38 @xmath39 then @xmath40 @xmath41 + + [ definition 2.1.3 ] the initial path in @xmath13 is the sequence of edges of @xmath13,@xmath42 where @xmath43 @xmath44 @xmath45 @xmath46 [ lemma 2.1.4 ] ( s.carter , m. elhamdadi , and m.saito , 2004 ) : let @xmath34 be a yang - baxter set , and @xmath47 be the initial path of @xmath13 . for any n - tuple of elements of @xmath1 , @xmath48 there exists a unique yang - baxter coloring @xmath49 of @xmath13 by @xmath34 such that @xmath50 for all @xmath51 this lemma gives the following two properties . 1 . each edge has the color uniquely induced by the @xmath14tuple asociated to the initial path of @xmath52 2 . each @xmath3face @xmath53 of @xmath13 has its induced initial path determined by the order of coordinates . therefore , we can associate to it the @xmath3tuple @xmath54 determined by colors on its induced initial path . denote this situation by @xmath55 from these two facts , we have a way to map an @xmath14tuple to @xmath56tuple through the face maps in cubic homology theory . recall that @xmath57 denotes the @xmath14dimensional boundary map in the cubical homology theory . thus @xmath58 where @xmath59 is an @xmath56face and @xmath60 depending on whether the orientation of @xmath59 matches the induced orientation . for the induced orientation , we take the convention that the inward pointing normal to an @xmath56face appears last in a sequence of vectors that specifies an orientation , and the orientation of the @xmath56face is chosen so that this sequence agrees with the orientation of the @xmath14cube . let @xmath34 be a yang - baxter set . let @xmath61 be the free abelian group generated by @xmath14tuples @xmath62 of elements of @xmath63 define a homomorphism @xmath64 by @xmath65 we have @xmath66 and @xmath67 is a chain complex . as usual , we can define @xmath68 to be the algebraic " version of yang - baxter homology group @xcite . in this homology theory , the chain groups are the same as before , that is @xmath69 we define the boundary homomorphism @xmath70 as follow , @xmath71 where @xmath72 . we can interpret the face maps through figure 2.1 + + the meaning of @xmath73 is illustrated in figure 2.2 ; @xmath74 can be described similarly . + + we have an @xmath14tuple as an input and each strand carries the corresponding element of the @xmath14tuple . we track down the graph from top to bottom , and at each crossing we apply the fixed yang - baxter operator with input the ordered pair consists of two elements carried by the two strands right above the crossing(as in figure 2.3 ) . + then the left strand after the crossing carries the @xmath75 function value and the right strand after the crossing carries the @xmath76 function value . in the end , we ignore the element carried by the left most strand , and this procedure generates an @xmath56tuple consisting of the @xmath77 elements carried by the other @xmath77 strands at the bottom . one can easily check @xmath78 form a pre - cubic set , which implies that @xmath79 is a chain complex . we define @xmath80 as the graphic " version of yang - baxter homology group . the algebraic " and graphic " definitions of yang - baxter homology coincide . more generally , chain complexes leading to above homologies are isomorphic . consider an @xmath14dimensional cube @xmath15 and denote it as @xmath81 for any coloring , say @xmath82 and an @xmath56face @xmath83 we are going to demonstrate that @xmath84 to see this , we need to calculate the coloring of the initial path @xmath85 of this @xmath56face . by definition , + @xmath86 @xmath87 @xmath45 @xmath88 @xmath89 @xmath90 @xmath45 @xmath91 @xmath92 + we need another sequence @xmath93 where + @xmath94 @xmath95 @xmath45 @xmath96 @xmath97 + for example , for @xmath98 the edges of the square are + @xmath99 @xmath100 @xmath96 @xmath88 + since + @xmath101 @xmath102 + we have + @xmath103 @xmath104 + once we know the color of @xmath105 we know the colors of @xmath106 and @xmath107 ( at each iteration we deal with a square in figure 3.1 ) . thus , recursively , we know all the colors of @xmath108 i.e. we get an @xmath56tuple . + + in general , we have @xmath109 @xmath110 we can see that this @xmath56tuple is the same as the one given by @xmath111(compare with figure 2.2 ) . similarly , if we consider @xmath112 we can show @xmath113 . as for the sign , we can directly calculate that , for @xmath114 the sign is @xmath115 thus , the boundary map of algebraic " version and the boundary map of graphic " version only differ by a global sign , therefore , the considered chain complexes are isomorphic and they give isomorphic homology groups . [ example 3.1 ] comparison of the face maps corresponding to the @xmath24face @xmath116 of a cube and @xmath117 . + + + from the figure above , we can see a curtain - like " object similar to the figure on the right hand side climbing " on the faces of the cube . for higher dimensions , the similar consideration holds . therefore this also gives a way to visualize the equivalence . this graphic " definition of yang - baxter homology was motivated by the homology theory of self - distributive systems @xcite , for example shelves , racks , and quandles . similarly , we can define one - term and two - term homology not only for set - theoretic yang - baxter operators but also for pre - yang - baxter operators . furthermore , we can extend the definition of pre - yang - baxter homology to pre - yang - baxter operators with yang - baxter wall ( see definition 4.2 ) . we will give the general definitions below and discuss some properties of these homology theories . let @xmath0 be a commutative ring with identity , @xmath118 be a free @xmath0 module with basis @xmath1 and @xmath119 be a right @xmath3module . we define in definition 4.2 the one - term pre - yang - baxter chain complex @xmath120 from the pre - simplicial module @xmath121 . [ definition 4.1 ] the pre - simplicial module @xmath122 is a collection of modules @xmath123 , @xmath124 , together with maps , called face maps or face operators , @xmath125 such that : @xmath126 we define a chain complex with chain modules @xmath123 and a boundary map @xmath127 given by : @xmath128 [ proposition 4.3 ] let @xmath135 be the chain complex of the one - term pre - yang - baxter homology . for a fixed element @xmath137 consider the map @xmath138 , defined by @xmath139 , where @xmath140 then this is a chain map . we have we construct a chain homotopy @xmath144 between @xmath145 and zero map , in particular showing that @xmath143 is a chain map . the chain homotopy @xmath146 is defined by @xmath147 we check : @xmath148 @xmath149 follows because @xmath150 is chain homotopic to the @xmath151 map , therefore , @xmath152 is the @xmath151 map on homology . @xmath153 follows since if @xmath143 is invertible , so is @xmath154 thus @xmath155 is isomorphic to @xmath155 through zero map . this shows that @xmath156 in the case that @xmath157 and of @xmath158 acting on @xmath0 trivially , the condition making @xmath159 a pre - simplicial module is equivalent to that the sum of each column of the @xmath7 matrix is @xmath160 , which we call the column unital condition ( e.g. stochastic matrices satisfy the condition ) . let @xmath157 @xmath158 act on @xmath0 trivially , and @xmath7 be a set - theoretic yang - baxter operator . if for any pair @xmath161 there is a unique @xmath162 such that @xmath163 then conditions in proposition 4.3 hold . in particular , biracks satisfy this condition ( see definition 3.1 condition 3 , that is right invertibility , in @xcite ) . we need to show that for any @xmath14tuple @xmath164 in @xmath165 there exists unique @xmath14tuple @xmath166 in @xmath167 such that @xmath168 since we know @xmath169 and @xmath170 , we get the values of @xmath171 and @xmath172 uniquely . once we have @xmath173 together the value of @xmath174 we get the value of @xmath175 and @xmath176 uniquely . thus by this iteration , we get the @xmath14tuple @xmath166 uniquely and this shows @xmath143 is invertible . [ example 4.2.1 ] ( compare @xcite ) let @xmath177 be a commutatinve ring with identity . let @xmath178/(1-s)(1-t),$ ] then @xmath179 is a set - theoretic yang - baxter operator satisfying the conditions in corollary 4.4 . this holds because for any given @xmath180 and @xmath181 , we can solve @xmath182 thus the one - term homology of this operator is trivial . let @xmath0 be a commutative ring with identity , @xmath118 be a free @xmath0 module with basis @xmath1 , @xmath119 be a right @xmath3module and @xmath185 be a left @xmath3module . we define in definition 4.8 the two - term pre - yang - baxter chain complex @xmath186 from the pre - cubical module @xmath187 . [ definition 4.1 ] the pre - cubical module @xmath122 is a collection of modules @xmath123 , @xmath124 , together with maps , called face maps or face operators , @xmath188 such that : @xmath189 we define a chain complex with chain groups @xmath123 and a boundary map @xmath127 given by : @xmath190 [ definition 4.2](@xcite ) consider a linear map @xmath191 such that @xmath192 and a linear map @xmath193 such that @xmath194 we call them the left and right wall conditions(graphically see figure 4.3 and figure 4.4 ) . + in the case that @xmath204 and the action of @xmath158 on @xmath0 is the trivial action , the conditions making @xmath205 a pre - cubical module is equivalent to saying that @xmath7 has the column unital condition . [ example 4.2.1 ] we give a family of unital yang - baxter operator @xmath206 where @xmath207 @xmath208 $ ] and @xmath209 is a positive integer . for any given @xmath209 , @xmath7 can be represented by its coefficients , @xmath210 what fascinates us is that this family of yang - baxter operators come from the yang - baxter operators giving @xmath221 polynomial invariants of links ( substitutions to the homflypt polynomial ) . for example , when @xmath211 , the matrix is if we divide elements in each column by the sum of those elements and make the substitution @xmath223 we will get the matrix in example 4.9 . in general , we can get our family in example 4.9 in a similar way ( normalizing columns ) and again they are yang - baxter operators . more interestingly , this new family of yang - baxter operators also provide @xmath221 polynomial invariants of links @xcite . this fact is implicit in @xcite . in the paper @xcite where carter , elhamdadi and saito define their set - theoretic yang - baxter homology , they also define a 2-cocycle link invariant from the set - theoretic yang - baxter homology . this can also be done in the case of column unital yang - baxter operators . in @xcite , it demonstrates the third reidemeister move preserve the ( co)homology of the ( co)cycle constructed form a knot diagram . see figure 4.6 and @xcite for details . pape-1 r. j. baxter , partition function of the eight - vertex lattice model _ annals of physics _ , 70 , 1 , 1972 , 193 - 228 j. s. carter , m . elhamdadi , m . saito , homology theory fpr the set - theoretic yang - baxter equation and knot invariants from generalizations of quandles , _ fund . math _ , 184(2004),31 - 54 v. f. r. jones , on knot invariants related to some statistical mechanical models , _ pacific j. math . _ , 137 ( 1989 ) , no . 2 , 311 - 334 . v. f. r. jones , hecke algebra representations of braid groups and link polynomials , _ annals of mathematics _ vol . 126 , no . 2 ( sep . , 1987 ) , pp . 335 - 388 v. lebed , homologies of algebraic structures via braidings and quantum shuffles _ j. algebra _ , 391 : 152 - 192(2013 ) ; + e - print : arxiv:1204.3312 [ math.ct ] j. h. przytycki , distributivity versus associativity in the homology theory of algebraic structures , _ demonstratio math . _ , 44(4 ) , 2011 , 823 - 869 ; + e - print : arxiv:1109.4850 [ math.gt ] j. h. przytycki , knots and distributive homology : from arc colorings to yang - baxter homology , chapter in : _ new ideas in low dimensional topology _ , world scientific , vol . 56 , march - april 2015 , 413 - 488 . e - print : arxiv:1409.7044 [ math.gt ] j. h. przytycki , curtain homology : 2-(co)cycle invariants from yang - baxter operators , in preparation . v. g. turaev , the yang - baxter equation and invariants of links , _ invent . _ , 92,1988 , 527 - 553 . x. wang , phd thesis george washington university . in progress . c. n. yang , some exact results for the many - body problem in one dimension with repulsive delta - function interaction _ phys . _ 1967 , 1312
in 2004 , carter , elhamdadi and saito defined a homology theory for set - theoretic yang - baxter operators(we will call it the algebraic " version in this article ) . in 2012 , przytycki defined another homology theory for pre - yang - baxter operators which has a nice graphic visualization(we will call it the graphic " version in this article ) . we show that they are equivalent . the graphic " homology is also defined for pre - yang - baxter operators , and we give some examples of it s one - term and two - term homologies . in the two - term case , we have found torsion in homology of yang - baxter operator that yields the jones polynomial .
there are today five detectors of gravitational waves ( gw ) in operation @xcite , all of them of the resonant type . it is thus important to study in detail the problem of the coincidence search . in the past , after the initial works of weber , three papers on coincidence search have been published @xcite . these coincidence search was made under two hidden assumptions : + a)the signal - to - noise ratio ( snr ) was considered to be very large , + b)the event time was considered to be equal to the signal time . + since we expect very tiny signals , the study of the problem when dealing with small snrs is fundamental . this is our object here using simulated signals but with real noise measured with the explorer and nautilus detectors . in order to clarify the distinction between @xmath0 and @xmath1 let us recall how an @xmath1 is defined . we describe the procedure adopted by the rome group , but a similar procedure is adopted also by the allegro , auriga and niobe groups . for nautilus and explorer the data have a sampling time of 4.544 ms and are filtered with a filter matched to short bursts @xcite for the detection of delta - like signals . the filter makes use of power spectra obtained during periods of two hours . be @xmath2 the filtered output of the electromechanical transducer which converts the mechanical vibrations of the bar in electrical signals . this quantity is normalized , using the detector calibration , such that its square gives the energy innovation @xmath3 for each sample , expressed in kelvin units . in absence of signals , for well behaved noise due only to the thermal motion of the bar and to the electronic noise of the amplifier , the distribution of @xmath2 is normal with zero mean . the variance ( average value of the square of @xmath2 ) is called @xmath4 and is indicated with @xmath5 . the distribution of @xmath2 is f(x)=e^ [ normal ] after the filtering of the raw - data , @xmath6 are extracted as follows . a threshold is set in terms of a critical ratio defined by cr== [ creq ] where @xmath7 is the standard deviation of @xmath8 and snr_f= [ teffeq ] @xmath5 is determined by taking the average of the filtered data during the ten minutes preceeding each considered event . the threshold is set at cr=6 , in order to have about one or two hundred events per day . this corresponds to an energy @xmath9 . when the filtered data go above this threshold , the time behaviour is considered until the filtered data go below the threshold for more than ten seconds . the maximum amplitude and its occurrence time define the @xmath1 . by the word @xmath0 here we mean the response of the detector to an external excitation in absence of noise . it is then evident that an @xmath1 is a combination of signal and noise . in the following we shall use snr to indicate the ratio between the @xmath0 energy , which we denote with @xmath10 and the noise @xmath5 , snr= [ snr ] the effect of the noise on the signal has been discussed in @xcite and it turns out to be larger that one could erroneously think . for example , with snr=20 ( for nautilus ) , one could think that most of the signals would be detected above the threshold @xmath11 . it turns out that the detection efficiency is of the order of 50% , as the noise might be in phase with the signal , pushing it even higher over the threshold or in counter - phase , pushing it below the threshold . this means that the detection efficiency for @xmath12 coincidences with @xmath13 detectors , in the case @xmath14 , is of the order of @xmath15 . the noise acts also in producing an @xmath16 different from the time the @xmath0 was applied . this influences the choice of the coincidence time window . we use two sets of experimental data , obtained with explorer in 1991 and with nautilus in 1998 . this is because the two detectors had their best performance respectively in 1991 and 1998 , and also because their detection bandwidth is very different in the two cases . the main characteristics of these two detectors are given in table [ carac ] . 0.1 in . main characteristics of explorer and nautilus for the data used in the present analysis . [ cols="^,^,^,^,^,^",options="header " , ] [ effic1n ] the efficiency is also shown in fig.[effi ] the theoretical probability to detect a signal with a given snr , in presence of a well behaved gaussian noise , is calculated as follows . we put @xmath17 where @xmath18 is the signal we look for and @xmath19 is the gaussian noise . we obtain easily @xcite probability(snr)=_snr_t^ e^- cosh()dy [ papou ] where we put @xmath20 for the present explorer and nautilus detectors . the theoretical efficiency as deduced from eq.[papou ] is reported in tables [ effic1e ] and [ effic1n ] and in fig.[effi ] . we notice a deviation between experimental and theoretical efficiencies at small snr . this is due to the non gaussian character of the real noise . the time when the event due to a signal is observed deviates from the time the signal is applied . we show in fig . [ deviation ] the standard deviation against snr for explorer 1991 and for nautilus 1998 . the lines are the best fits with the following equations : + explorer @xmath21 + nautilus @xmath22 . we can write the empirical formula = [ formula ] we see , as expected , that the time deviation decreases linearly with increasing bandwidth . if we extrapolate to a snr=100 and with a target bandwidth for resonant detectors of the order of @xmath23 we find a possible time resolution of the order of less than one millisecond , as already recognized with room temperature experiments @xcite . the delay distributions for signals with snr=30 and snr=10 are shown in fig . [ distris30_10 ] . we note for nautilus a few events with delay greater than 1 s with respect to the time of the applied signals . we have asked ourselves how it is possible to have a time deviation over 1 s for signals with snr=30 . this is due to the fact that the noise , although the data were selected so to have small @xmath24 , has not completely a gaussian character . we have considered the particular case of the event ( fig.[distris30_10 ] , snr=30 ) detected with the nautilus data 1.422 s @xmath25 the signal was applied . in order to understand this result we plot in fig.[delta1s ] the behaviour of the signal with zero noise , of the noise alone and of the signal added to the noise . for this particular case if we raise the signal to snr=50 the corresponding event has a time delay of -64 ms ( still not quite zero ) . if an additional filter is applied to the data , such as to require i.e. that the detected event behaved in a gaussian way , the signal with snr=30 is lost , in spite of being a @xmath26 signal . we remark also that the events have energy different from that of the signal . this is shown in tables [ effic1e],[effic1n ] and in fig.[distrisnr ] where we give the distributions of the ratio @xmath27 for snr=30 and snr=10 . finally we make the following consideration for the case when multiple coincidences with @xmath28 detectors are searched for . from tables [ effic1e ] and [ effic1n ] we deduce that when the signal is near the threshold the efficiency of detection in nearly 50% . this means that for these signals the total efficiency for @xmath29 coincidences is @xmath30 . since we want an efficiency near unity ( because of the very few possible gw signals ) we must consider only signals with @xmath31 at least twice the value @xmath32 of the threshold . we have studied the @xmath6 generated in a resonant gw detector when excited by gw bursts with @xmath31 near the threshold @xmath32 used for defining the events . for @xmath33 the detection efficiency is nearly @xmath34 . the efficiency goes to 100% for @xmath35 , and it is still @xmath36 for @xmath37 . the time of the event might be different from that of the signal , with standard deviation depending on the snr and on the bandwidth of the experimental apparatus . in this analysis we have applied delta signals at the exact time of the samples . if the delta signals are applied randomly , as in the real case , the efficiency will be smaller and the time dispersion larger . delta - like signals can be lost if the requirement to satisfy the theoretical behaviour expected for a delta signal is imposed on the detected events , even for @xmath38 , as shown in fig.[delta1s ] . this can jeopardize a search looking for very rare gravitational wave signals . m.cerdonio et al . , first edoardo amaldi conference on gravitational wave experiments , pag . 176 , frascati , 14 - 17 june 1994 e.amaldi el al . , astron.astrophys . 216,325 ( 1989 ) p.astone et al , astroparticle physics 10 ( 1999)83 - 92 p.astone et al . , d,59 , 122001 ( 1999 ) p.astone , c.buttiglione,s.frasca , g.v.pallottino and g.pizzella , il nuovo cimento 20,9 ( 1997 ) s.boughn et al . , astrophys . j. 261 , 119 ( 1982 ) p.astone , g.p.pallottino and g.pizzella , general relativity and gravitation , 30 , 105 ( 1998 ) p.astone , g.v.pallottino and g.pizzella , class . quantum grav . 14 ( 1997 ) 2019 - 2030 a.papoulis `` probability , random variables and stochastic processes '' , mcgraw - hill book company ( 1965 ) , pag 126 . s.vitale et al . , first edoardo amaldi conference on gravitational wave experiments , pag . 220 , frascati , 14 - 17 june 1994
using simulated signals and measured noise with the explorer and nautilus detectors we find the efficiency of signal detection and the signal arrival time dispersion versus the signal - to - noise ratio .
the small - scale photospheric magnetic field of the quiet sun , the _ magnetic carpet _ , is complex and constantly evolving . this evolution is due to underlying photospheric flows produced by convection on many scales ( _ e.g. _ granulation and supergranulation , , , , ) and the flux evolution processes of emergence , cancellation , coalescence and fragmentation . small - scale magnetic features classified as ephemeral regions @xcite , internetwork features @xcite and network features @xcite continually interact with one another , resulting in a photospheric recycle time of just @xmath1 hr ( hours ) @xcite . this is the time taken for all flux within the quiet sun photosphere to be replaced . since magnetic fields from the magnetic carpet extend up into the solar chromosphere and lower corona , it is expected that the quiet sun corona is also highly dynamic . complex interactions of magnetic features on the photosphere may result in significant heating of the corona , for example through braiding and reconnection of magnetic field lines ( _ e.g. _ , , , , , ) . therefore , it is of interest to simulate the small - scale coronal field resulting from the evolution of the solar magnetic carpet . previous magnetic carpet coronal field models using extrapolation methods have studied , for example , flux topology and connectivity @xcite ; the number density and locations of coronal null points @xcite ; coronal remap times @xcite ; and whether the solar wind can be driven by reconnection in the magnetic carpet @xcite . however , each of these studies considered only potential field extrapolations of the small - scale coronal field . within this paper , we present preliminary results of 3d simulations of the network - scale coronal field of the magnetic carpet . in contrast to the models described above , which produce independent potential field extrapolations , we model a continuous evolution of a non - linear force - free coronal field . a non - linear force - free field satisfies the conditions @xmath2 and @xmath3 , where @xmath4 is a scalar function of position ( but constant along a given field line ) describing the twist of the field . this approximation is a step up in complexity from a potential field as it allows for the existence of electric currents and free magnetic energy . in ( hereafter paper i ) we presented a realistic model for the photospheric evolution of the magnetic carpet , that reproduced many observational properties . synthetic magnetograms produced from this model will provide the photospheric boundary condition to drive the evolution of the full 3d coronal field . the coronal field is evolved through a series of quasi - static , non - linear force - free states in response to the evolution of the photospheric magnetic field , using a magnetofrictional technique that is described in ( hereafter paper ii ) . there are several advantages to using synthetic magnetograms as opposed to actual observed data : ( i ) the total magnetic flux within the synthetic magnetograms is always known @xmath5 there is no noise or instrumental limitations . ( ii ) we have complete control over the photospheric evolution @xmath5 magnetic features do not drift into or out of the field of view , which is unavoidable in real magnetograms . ( iii ) the synthetic magnetograms are always in flux balance ( total positive flux = total negative flux ) . this is a requirement for the coronal model due to our set - up . ( iv ) we know exactly where and when each of the processes of emergence , cancellation , coalescence and fragmentation occur within the synthetic magnetograms , as well as the exact flux involved in each event . many feature tracking techniques exist to follow the evolution of magnetic features in real magnetogram data , _ . however , these are limited by factors such as spatial resolution and time cadence . in a future study , we will consider in detail how various photospheric events in our synthetic magnetograms affect the evolution of the coronal field . in this paper we consider four simulations in total , each with the same photospheric boundary condition : a @xmath6 hr set of synthetic magnetograms . three of the simulations have a uniform , overlying coronal magnetic field ( of varying strength ) which points in the positive @xmath7direction ; this simulates the influence of larger scale magnetic features , such as exist on the sun , especially during solar maximum . the fourth simulation has no overlying field and thus represents a very quiet region of the sun , such as may be found during solar minimum . we consider the magnetic energy both stored and dissipated within the coronal volume , along with the square of the current density , @xmath8 . some of the properties of these quantities that we consider are : how they evolve in time , where they are located spatially with respect to the underlying photospheric magnetic field , and the effect of varying the strength of the overlying coronal magnetic field . section [ sec : model ] describes the lower boundary condition and the set - up of the 3d simulations . section [ sec : result ] provides the results , while section [ sec : conc ] gives a discussion and conclusions . * * we choose the most realistic simulation from paper i to provide the lower boundary condition for 3d modelling . this simulation , which covers an area of @xmath9 mm@xmath10 , is of length @xmath11 hr with a cadence of 1 min , and has a flux emergence range of @xmath12 mx . flux emergence is determined by the probability distribution of , giving an emergence rate of @xmath13 mx @xmath14 s@xmath15 . the smallest magnetic features that occur within this simulation have a flux of @xmath16 mx . the magnetograms are all in flux balance and are periodic in the @xmath7 and @xmath17directions . for full details of the simulation , see paper i. for the 3d simulations , we select a @xmath6 hr window of the synthetic magnetograms from the full @xmath11 hr set , as explained below , providing a series of @xmath18 synthetic magnetograms . figure [ fig : lower](a ) shows the total absolute flux for the @xmath6 hr section of the 2d simulation that is used for 3d modelling . the curve oscillates about a mean value of @xmath19 mx , with a standard deviation of @xmath20 mx , or @xmath21 , indicating that the 2d simulation is in a steady state at this time . figures [ fig : lower](b ) shows an example of a synthetic magnetogram , taken at @xmath22 hr , where red and blue contours represent positive and negative magnetic field , respectively . a movie showing the photospheric evolution of @xmath23 for the first @xmath24 hr is available ( ) . the magnetic elements are mainly located around the boundaries of the supergranules , forming the magnetic network ( an image of the simulated supergranules can be seen in figure 3(b ) of paper i ) . the supergranular flow profile is not time - evolving , so the general shape of the magnetic network does not vary much . despite this limitation , the exact distribution of magnetic elements does change significantly throughout the @xmath6 hr period . in paper i it was determined that the photospheric recycle time for this series of synthetic magnetograms was @xmath25 hr , in excellent agreement with s recycle time of @xmath1 hr . therefore the synthetic magnetogram series realistically simulates the dynamic nature of the magnetic carpet , making it a suitable lower boundary condition for 3d modelling . we now discuss the set - up of the 3d model . we choose a numerical box of size @xmath26 mm@xmath27 , composed of @xmath28 grid cells . the box is periodic in the @xmath7 and @xmath17directions and closed at the top . since magnetic flux may only enter and exit the box through the lower boundary , the synthetic magnetograms are required to be in complete flux balance . the periodicity of the side boundaries gives the effect of the region being surrounded by similar regions of small - scale magnetic carpet features . the initial condition for each simulation is a potential field extrapolated from the first magnetogram . we subsequently evolve the coronal magnetic field in response to photospheric boundary motions using a magnetofrictional relaxation technique @xcite . the magnetofrictional technique produces a continuous evolution of the coronal magnetic field , so that a ` memory ' of flux connectivity and electric current systems is maintained from one step to the next . the coronal field induction equation is given by @xmath29 where @xmath30 is the magnetic field and @xmath31 its the associated vector potential . the plasma velocity , @xmath32 , is given by @xmath33 where @xmath34 is the coefficient of friction , @xmath35 km@xmath10 s@xmath15 , and @xmath36 . this velocity describes the relaxation of the coronal magnetic field towards a non - linear force - free state in response to perturbations , and takes advantage of the fact that the lorentz force , @xmath37 , is the dominant force within the corona . the term @xmath38 within equation [ eqn : ind ] represents _ hyperdiffusion _ ( _ e.g. _ , ) and is chosen to be of the form @xmath39 where , @xmath40 km@xmath41 s@xmath15 and @xmath42 is the scalar coefficient from the definition of a non - linear force - free field such that @xmath3 . it is calculated as @xmath43 hyperdiffusion aids the stability of the code by smoothing gradients in @xmath42 and allows reconnection to occur . it acts to reduce the magnetic field to a linear force - free state , however the time scales of the present simulations are too short for such a state to be reached . the evolution of the photospheric boundary perturbs the coronal field , which responds through equations [ eqn : ind ] and [ eqn : v ] . this continual stressing and relaxing of the coronal field , in response to the lower boundary motions , evolves the field through a series of quasi - static , non - linear force - free equilibria . the lower boundary treatment is the same as is described in paper ii , where a full description may be found . since equation [ eqn : ind ] is specified in terms of the vector potential @xmath31 , the lower boundary is also required in terms of @xmath31 . a linear interpolation of the @xmath44 and @xmath45 corresponding to @xmath23 at @xmath46 mm is carried out between each synthetic magnetogram . @xmath23 at @xmath46 mm is analytically specified at each time step , rather than advected numerically ( see paper i ) . this avoids certain undesirable numerical effects such as numerical overshoot or pile - up at cancellation sites . between each analytical specification of the photospheric magnetic field , @xmath47 interpolation steps are taken , corresponding to @xmath48 s each . the magnetofrictional technique is applied during each of these interpolation steps so that the relaxation is gradual . in total , four simulations are run , each with the same photospheric boundary evolution . three of the simulations have an overlying , uniform magnetic field of strength @xmath49 g , 3 g or 10 g , which points in the positive @xmath7direction . the fourth simulation has no overlying field . for the cases with a uniform overlying field , we add this field to the initial potential field as follows : the potential field is computed in terms of the vector potential @xmath31 . we add the term @xmath50 to @xmath45 throughout the volume , where @xmath51 is the strength of the overlying field and @xmath52 is the height above the photosphere . then , when we compute @xmath30 , since the @xmath7component is @xmath53 this has the effect of adding a constant @xmath51 to @xmath54 throughout the volume . we also add a contribution @xmath55 to the @xmath45 component of all of the synthetic magnetograms that provide the lower boundary condition . investigated an eit wave during solar minimum and estimated the strength of the quiet sun coronal magnetic field to be @xmath56 g. a variety of authors have attempted to estimate the strength of active region coronal loops using coronal seismology , with values obtained anywhere in the range @xmath57 g @xcite , although stress that coronal seismology does not necessarily reliably determine the strength of the magnetic field . indeed , they find that when they apply the method of to their 3d model of a coronal loop , it overestimates the strength of the magnetic field by @xmath58 % . however , our no overlying field and 1 g overlying field cases can be taken to represent the quiet sun during solar minimum , while the 3 g and 10 g cases represent a quiet region of the sun influenced by nearby active regions during solar maximum . the following sections give preliminary results of these 3d non - potential simulations . in section [ sec : result ] , we briefly discuss field line connectivity between magnetic elements . section [ sec : free ] focuses on the free magnetic energy , section [ sec : j ] the square of the current density ( @xmath8 ) and section [ sec : q ] on the energy dissipated . for both energies , we consider the time evolution of the quantity integrated over the whole volume , and also how it is affected by varying the overlying field strength . we then consider where the energy is located spatially . * * * * figures [ fig : fl](a ) and ( b ) show images from the simulation with no overlying field , ( c ) and ( d ) the 3 g overlying field simulation . the images on the left are both taken at @xmath59 hr and the images on the right are shown at @xmath60 hr . in all four images , contours of @xmath23 at @xmath46 mm are plotted on the base ( red = positive , green = negative ) . a selection of magnetic field lines are also over - plotted in each case , where dark blue , magenta and pale blue field lines reach a maximum height of @xmath61 mm , @xmath62 mm and @xmath63 mm respectively . for comparison , field lines in figures [ fig : fl](c ) and ( d ) are plotted from the same photospheric starting points as field lines in ( a ) and ( b ) respectively . as expected , we find that connections between magnetic elements within the simulation with no overlying field may reach much greater heights than those in the 3 g simulation , as the overlying field suppresses the extension of the magnetic elements field into the corona . also , the connectivity is quite different between the images at @xmath59 hr and @xmath60 hr , showing that the coronal field has changed significantly during this interval , as expected . the images here are intended only to give an indication of what the connectivity is like between the magnetic elements . a more in depth analysis will be carried out in future . * * * * * * the presence of free magnetic energy within our 3d model is a significant difference to the potential field ( minimum energy ) models of the magnetic carpet coronal field discussed in the introduction . figure [ fig : free1](a ) shows a plot of the free magnetic energy ( ergs ) as a function of time , defined as @xmath64 where @xmath65 is the non - linear force - free magnetic field and @xmath66 is the magnetic field of the corresponding potential field . results are shown for the no overlying field ( green ) , 1 g ( black ) , 3 g ( blue ) and 10 g ( red ) simulations . for each simulation , the free energy initially increases rapidly as the coronal field diverges from a potential state due to surface motions . for the no overlying field and 1 g cases , the free energy then levels off and oscillates around a mean value of @xmath67 ergs and @xmath68 ergs , respectively , with standard deviations of @xmath69 ergs and @xmath70 ergs . for the 3 g and 10 g cases , the free energy also levels off to a lesser extent , but is less steady , with mean values of @xmath71 ergs and @xmath72 ergs and standard deviations of @xmath73 ergs and @xmath74 ergs respectively . the mean and maximum values of free energy per unit area for each simulation are given in table [ tab : free ] . the trend is that a stronger overlying field leads to a greater build - up of free energy . in each case , the individual peaks in free energy differ between the simulations , however the general shape of the curve is the same for all four overlying field strengths . therefore the overall behaviour of the free magnetic energy is largely dependent on the evolution of the photospheric magnetic field , rather than the strength of the overlying coronal field . .mean and maximum values of free magnetic energy for each simulation . [ cols="^,^,^ " , ] the variation of free energy with height is similar for all simulations , so here we just discuss the results from the 3 g overlying field simulation . figure [ fig : free1](b ) shows plots of the free magnetic energy integrated over @xmath75 and @xmath76 , as a function of height , for the simulation with a 3 g overlying field . this is computed as follows : @xmath77 where @xmath78 mm , @xmath79 mm and @xmath80 mm is the length of a cell in @xmath52 . the total free magnetic energy as a function of height is shown at times spaced evenly throughout the simulations , at @xmath59 hr ( black ) , @xmath81 hr ( blue ) , @xmath82 hr ( red ) , @xmath83 hr ( green ) , @xmath84 hr ( yellow ) and @xmath60 hr ( purple ) . in figure [ fig : free1](a ) , which shows the total free magnetic energy as a function of time , the vertical yellow dashed lines are over - plotted on the graph at intervals of 8 hr , indicating the times at which the lines in figure [ fig : free1](b ) are taken . in figure [ fig : free1](b ) , it can be seen that the ordering of the curves is not time dependent . once the coronal field has evolved away from its initial potential state , the total amount of free magnetic energy within the volume depends upon how much is both built up and stored due to surface motions . for each of the curves in figure [ fig : free1](b ) , there is a peak between roughly @xmath85 mm and @xmath86 mm , indicating that this is where the majority of the free magnetic energy is stored . the free energy then drops off rapidly after @xmath87 mm . as in paper ii , we find that the field departs most from a potential state low down in the corona , as this is close to where we are driving the evolution of the field by photospheric motions . in addition , most closed connections between magnetic elements ( as opposed to connections from the magnetic elements to the overlying field ) are found to be low lying ( _ e.g. _ , ) , and it is along these connections that free energy is stored . a movie showing the time evolution of the free magnetic energy as a function of height for the 3 g simulation is available ( ) . the movie shows that the free energy is highly dynamic and rapidly evolving . the height of the curve is continually changing , however it can be seen that the peak in the curve tends to remain between @xmath85 mm and @xmath86 mm . figures [ fig : free1](c ) and ( d ) show the free magnetic energy density , @xmath88 , integrated in @xmath52 , in the @xmath89 plane for the 3 g simulation at @xmath59 hr and @xmath60 hr , respectively . this is computed as follows : @xmath90 where @xmath91 , @xmath92 mm is the length of a cell in @xmath75 and @xmath93 mm is the length of a cell in @xmath76 . white patches indicate where the free energy density is positive , _ i.e. _ where @xmath94 , black where the free energy density is negative , @xmath95 . note that the total free magnetic energy integrated over the volume is always positive ( see figure [ fig : free1](a ) ) . positive ( red ) and negative ( green ) contours of @xmath23 at @xmath46 mm are over - plotted . we define free magnetic energy to be ` stored ' at locations where the line - of - sight ( los ) integrated free magnetic energy density is positive . from these images , it can be seen that free magnetic energy may be stored both at the boundaries between supergranular cells and within the cells themselves . in paper ii , it was found that when the evolution of the magnetic elements disturbed a larger volume of the overlying field , a greater amount of free energy was built up . it was also found that closed connections between the magnetic elements are required , along which the free energy may be stored . in agreement with paper ii , we find that free energy is stored mainly in two locations . firstly , we see that white patches are located around supergranule cell boundaries where the magnetic network is formed . large numbers of magnetic elements are swept to these locations by supergranular flows , and continually interact with one another , cancelling , coalescing and fragmenting . this continual evolution of the magnetic elements results in a large build up of free energy , which may then be stored along the multiple connections that form between nearby magnetic elements that lie in the network . the second location where we see free energy stored is along long - lived , far - reaching , twisted connections between magnetic elements . such connections may stretch across supergranule cells , between magnetic elements located at opposite boundaries ( examples of this will be given later , in figure [ fig : free2 ] ) . longer connections will clearly disturb a larger volume of the surrounding coronal magnetic field , hence building up free energy . a movie showing the free magnetic energy density , integrated in @xmath52 , in the @xmath89 plane is available ( ) , with contours of @xmath23 at @xmath46 mm over - plotted . within the movie , one can see that regions of positive free energy density are continually evolving in response to photospheric motions . in particular , occasional large patches of positive free energy density develop around the magnetic network . one can also see long - lived bands of positive free magnetic energy density stretching across supergranules , between magnetic elements . many of the regions of free magnetic energy density last for several hours . for readers unable to view the movie , six still images spaced 1 hr apart are included in appendix [ sec : app1 ] ( figure [ fig : app1 ] ) , to give an impression of the time - scale of the evolution . figures [ fig : free1](e ) and ( f ) show @xmath96 plane images of the los integrated free magnetic energy density , saturated at @xmath97 ergs , for the 3 g simulation at @xmath59 hr and @xmath60 hr , respectively . a movie of the free magnetic energy density in the @xmath96 plane , integrated in @xmath76 , is available ( ) . in both the movie and the still images , it can be seen that the free energy is mainly located low down , with the bulk of it being below @xmath98 mm . this is where many closed connections exist between the magnetic elements and a larger departure of the magnetic field from a potential state is found . one can see that the locations of positive free energy density are highly dynamic and there exist long - lived ` bulbs ' of positive free energy density , where it is stored within the corona along closed connections between magnetic elements . similar evolution of the los integrated free magnetic energy density is seen in the @xmath99 plane and for different strengths of overlying field . * * * * figures [ fig : free2](a ) and ( b ) show two zoomed in sections of the @xmath89 plane images of the los integrated free energy density taken at @xmath59 hr and @xmath60 hr , respectively , from figures [ fig : free1](c ) and ( d ) . a selection of closed field lines has been over - plotted in blue in each case . figure [ fig : free2](a ) shows the band of positive free energy density that can be seen lying across the lower left supergranule in figure [ fig : free1](c ) , while figure [ fig : free2](b ) shows the band of positive free energy density across the lower right supergranule in figure [ fig : free1](d ) . in both zoomed images , twisted magnetic field lines connect between various magnetic elements on either side of the supergranule . the free energy is stored in these regions of complex connections . figure [ fig : free2](c ) shows an @xmath96 plane image of free energy density integrated in @xmath76 at @xmath59 hr , and is a side view of the band of positive free energy density in figure [ fig : free2](a ) . similarly , figure [ fig : free2](d ) is a side view of the band of positive free energy density in figure [ fig : free2](b ) and shows a @xmath99 plane image of free energy density integrated in @xmath75 at @xmath60 hr . for each of the cases , a complex structure of the magnetic field can be seen . * * * * * * the square of the current density , @xmath8 , is of interest because it indicates locations of possible ohmic heating , @xmath100 . however , it should be noted that in the present simulations , due to the simplified form used , we do not have an energy equation . figures [ fig : j](a ) and ( b ) show images in the @xmath89 plane of @xmath8 integrated in @xmath52 , at @xmath59 hr and @xmath60 hr , respectively . this is computed as follows : @xmath101 note that the colour table has been reversed , so that darker regions correspond to higher values of @xmath102 . contours of @xmath23 at @xmath46 mm are over - plotted at the same levels as in figures [ fig : free1](c ) and ( d ) . on comparison with the @xmath89 plane images of free magnetic energy in figures [ fig : free1](c ) and ( d ) , the locations of high @xmath8 and of positive free magnetic energy density seem to match very well . the regions of high @xmath8 appear to be strongest in the magnetic network . within a force - free field , by definition , @xmath103 , hence @xmath104 is parallel to @xmath105 . this means that we can express @xmath104 as a scalar multiple of @xmath105 , @xmath106 , where @xmath42 is a scalar representing the twist of the magnetic field with respect to the corresponding potential field . therefore , it makes sense that @xmath8 is at its largest near the magnetic sources , where @xmath105 is largest . we also see fainter bands of @xmath8 stretching across supergranules , often in the same places as bands of positive free magnetic energy density . again , it makes sense for @xmath8 to be high in such locations as free magnetic energy is built up in regions of high non - potentiality ( large @xmath107 ) which arise due to non - zero @xmath104 ( @xmath108 ) . the evolution of @xmath8 , integrated in @xmath52 , in the @xmath89 plane can be seen in the movie , , for the 3 g simulation . six still images from this movie , spaced 1 hr apart , are included in appendix [ sec : app1 ] ( figure [ fig : app2 ] ) , to give an impression of the evolution for those who can not view the movie . although regions of @xmath8 are often co - located with regions of positive free energy density , the regions of @xmath8 appear to be more rapidly evolving than those of free energy . the spatial distribution of @xmath8 is quite different from one hour to the next . figures [ fig : j](c ) and ( d ) show images of the logarithm of @xmath8 in the @xmath96 plane at @xmath109 mm , at @xmath59 hr and @xmath60 hr , respectively . the regions of @xmath8 appear to be well structured , and obviously follow the shape of the magnetic field ( as @xmath104 is parallel to @xmath105 ) . figures [ fig : j](e ) and ( f ) show images of the free magnetic energy density at @xmath109 mm , shown at the same times as ( c ) and ( d ) respectively . it can be seen that regions of positive free energy density tend to be co - located with regions of high @xmath8 . * * * * * * in addition to the free magnetic energy stored within the system , we consider energy that is being continually dissipated due to the relaxation processes within the model applied . this is described by the energy dissipation term , @xmath110 the first term , @xmath111 , represents energy dissipation due to magnetofriction , which is released as the coronal magnetic field relaxes towards a force - free state @xcite . the second term , @xmath112 , represents energy dissipation due to hyperdiffusion @xcite . it has previously been interpreted as the rate at which magnetic energy is converted into heat during the relaxation of the magnetic field @xcite . for the derivation of @xmath113 and a full description of the terms @xmath111 and @xmath112 , see paper ii . cccccccc + & @xmath51 & 0 g & 1 g & 3 g & 10 g + @xmath114 & mean & 2.66 & 2.67 & 2.74 & 3.25 + ( @xmath115 ergs @xmath14 s@xmath15 ) & max & 4.71 & 4.78 & 4.91 & 5.60 + + @xmath116 & mean & 0.44 & 0.45 & 0.53 & 0.95 + ( @xmath115 ergs @xmath14 s@xmath15 ) & max & 0.58 & 0.62 & 0.77 & 1.47 + + @xmath117 & mean & 3.10 & 3.12 & 3.27 & 4.20 + ( @xmath115 ergs @xmath14 s@xmath15 ) & max & 5.26 & 5.30 & 5.49 & 6.87 + + typical @xmath118 ( g ) & & 6.5 & 6.6 & 7.2 & 12.1 + @xmath119 ( @xmath115 ergs @xmath14 s@xmath15 ) & & 5.96 & 6.05 & 6.60 & 11.09 + + @xmath120 ( @xmath121 ergs ) & & 1.35 & 1.37 & 1.43 & 1.83 + @xmath122 ( @xmath123 ergs @xmath14 ) & & 5.43 & 5.47 & 5.73 & 7.31 + figure [ fig : q1](a ) shows a plot of the rate of energy dissipation per unit area , due to magnetofriction , @xmath114 , as a function of time , for the no overlying field ( green ) , 1 g ( black ) , 3 g ( blue ) and 10 g ( red ) simulations . cm@xmath10 is the area of the photospheric boundary surface . it can be seen that @xmath111 is not strongly dependent on the overlying field strength . a stronger overlying field leads to slightly higher @xmath126 , but the variation of values of @xmath111 within each curve is much larger than the variation of values between the curves for different overlying field strengths . table [ tab : q ] shows the mean and maximum values of @xmath114 for each overlying field strength simulation . the difference in mean values between the 0 g and 10 g cases is just @xmath127 ergs @xmath14 s@xmath15 . however , from the plot in figure [ fig : q1](a ) , each curve has a variation of around @xmath128 ergs @xmath14 s@xmath15 between its maximum and minimum . therefore , the energy dissipation due to magnetofriction is predominantly dependent upon the evolution of the photospheric magnetic field driving change within the coronal field . figure [ fig : q1](b ) shows the rate of energy dissipation per unit area due to hyperdiffusion , @xmath116 , as a function of time , where lines are coloured as in figure [ fig : q1](a ) . very little difference can be seen between the curves for the no overlying field and 1 g cases . the 3 g case results in slightly higher @xmath112 , while the 10 g case results in significantly higher @xmath112 . also , for the 10 g case there is a larger variation in the values of @xmath112 than in the no overlying field or 1 g cases . thus , @xmath112 is clearly dependent on the strength of the overlying field . while this is the case , the general shape of all of the curves are the same , implying that like @xmath111 , @xmath112 is also predominantly dependent on the evolution of the photospheric magnetic field driving changes within the coronal volume . the mean and maximum values for @xmath116 for each strength of overlying field are given in table [ tab : q ] . for each case , the mean values for @xmath112 are @xmath129 times smaller than the mean values for @xmath111 , and the maximum values for @xmath112 are @xmath130 times smaller than the maximum values for @xmath111 . figure [ fig : q1](c ) shows a plot of the total rate of energy dissipation per unit area , @xmath131 , as a function of time , with lines coloured as in figure [ fig : q1](a ) . since throughout each simulation , @xmath111 is larger than @xmath132 , the curves for @xmath113 follow the same trend as those for @xmath133 . a stronger overlying field leads to slightly higher @xmath113 , but the variation of @xmath113 within each curve ( @xmath134 ergs @xmath14 s@xmath15 ) is larger than the variation between the simulations with different overlying field strengths ( @xmath135 ergs @xmath14 s@xmath15 difference between the mean values for the no overlying field and 10 g cases ) . therefore , the overall energy dissipation is determined mainly by the photospheric evolution of the magnetic field . the mean and maximum values of @xmath117 for each simulation are given in table [ tab : q ] . figure [ fig : q1](d ) shows the cumulative energy dissipated , per unit area , as a function of time , @xmath136 , for each strength of overlying field , obtained by integrating @xmath113 over both the volume and time : @xmath137 dt.\ ] ] we see that a stronger overlying field leads to a greater cumulative amount of energy dissipated . the slopes of these curves are given by the mean values of @xmath117 in table [ tab : q ] , and can be related to the overlying field strength ( @xmath51 ) by a quadratic : @xmath138 the values for the total energy dissipated by the end of each simulation ( @xmath139 ) are also given in table [ tab : q ] . the rate of energy dissipation per unit area ( ergs @xmath14 s@xmath15 ) in each simulation is on average @xmath140 ergs @xmath14 s@xmath15 . these values are too low to explain the radiative losses of the quiet sun corona , being only @xmath141 of the @xmath142 ergs @xmath14 s@xmath15 calculated by and @xmath143 of the @xmath144 ergs @xmath14 s@xmath15 calculated by . due to the fact that the values are too low , it is of interest to compare the energy dissipated to the poynting flux through the photospheric boundary , as clearly we can not dissipate more energy than has been injected . considering the plots of free magnetic energy as a function of time ( figure [ fig : free1](a ) ) , they are roughly steady , at least for the @xmath145 g , @xmath146 g and no overlying field cases . this implies that the energy dissipated ( @xmath117 ) should roughly balance the poynting flux injected . the poynting flux through the photospheric boundary is given by @xmath147 where @xmath148 and @xmath149 is the area of the photospheric boundary surface with unit normal vector @xmath150 . from ohm s law , @xmath151 so equation [ eqn : poynt ] simplifies to @xmath152 ds.\ ] ] the individual terms in equation [ eqn : poynt2 ] may be split into two distinct groups . those involving @xmath153 and @xmath154 represent energy flow into or out of the domain due to horizontal boundary flows . the terms involving @xmath155 represent energy flow into or out of the domain due to flux emergence or cancellation . as computed in , an order of magnitude estimate of the poynting flux due to horizontal boundary flows is : @xmath156 where @xmath118 and @xmath23 are typical values for horizontal and vertical magnetic field at the photosphere and @xmath157 is a typical value for horizontal velocities at the photosphere . in all simulations , the mean vertical field strength is @xmath158 g over a photospheric area of @xmath124 cm@xmath10 and the mean value of our supergranular velocity profile at the photosphere is @xmath159 km s@xmath15 . the only other parameter is the typical value for the horizontal field strength , @xmath118 , which varies from one simulation to the next due to the overlying field . the mean values of @xmath118 for each simulation are listed in table [ tab : q ] , along with the mean energy dissipation , @xmath117 , and the value of @xmath160 calculated for each simulation . the values in table [ tab : q ] show that our order of magnitude estimate for energy injected due to horizontal flows is within a factor of @xmath161 of the energy dissipated for each simulation , so are in good agreement . while we are able to estimate the poynting flux resulting from horizontal flows , we can not calculate the poynting flux from vertical flows as , due to our special boundary treatment ( section [ sec : lower ] ) , there is no @xmath155 defined on the photospheric boundary of our model . the component of the poynting flux due to @xmath155 would likely lead to both injection and removal of energy respectively during emergence and cancellation events . as the flux domain is in a steady state , with the rates of flux emergence and cancellation roughly identical ( see paper i ) , we may assume that these processes lead to no net injection of energy . in addition , have estimated , using observational data , that the poynting flux injected due to the emergence of magnetic flux is significantly smaller than that injected due to the horizontal motions of existing magnetic flux . assuming equal rates of energy input and loss due to emergence and cancellation , this calculation shows that the energy injected does in fact match the energy dissipated within an order of magnitude . this implies that the reason why the energy dissipated within our simulations is not high enough to account for coronal radiative losses is in part due to the fact that not enough energy has been injected in the first place . therefore , in future simulations , increasing @xmath157 , @xmath118 or @xmath23 may lead to an increase in energy dissipation . we note that other possible reasons for the energy dissipation rate being too low are that this is not a true physical dissipation , and many simplifications have been made for our model . for example , our model contains no plasma and we do not resolve wave motions . the relative simplicity of the synthetic magnetograms compared to observed magnetograms may also be a factor . while the values calculated for @xmath117 are too low to explain coronal heating , it is of interest for us to consider the location and structure of the energy dissipation term , @xmath113 . figures [ fig : q1](a)@xmath5(d ) consider the volume integrated rate of energy dissipation over the entire 3d simulation for each strength of overlying field . we now consider where @xmath113 is spatially located within individual frames of the 3 g simulation . although results are presented here only for the 3 g simulation , similar results are found for other strengths of the overlying field . figure [ fig : q1](e ) shows the rate of energy dissipation ( integrated in @xmath75 and @xmath76 ) as a function of height . this is computed with units of ergs @xmath14 s@xmath15 as follows : @xmath162 six curves are plotted at intervals of 50 min from @xmath163 hr , the other six are plotted at intervals of 8 hr from @xmath164 hr . colours span from blue to red with increasing time . the curve at the earliest time ( @xmath165 hr ) is lowest , as the coronal magnetic field is still close to potential . the height of the curves increases with increasing time until a near steady rate of dissipation is reached , where very little difference is seen between the curves . at each time , the greatest rate of energy dissipation is found low down , near the photosphere . this is not surprising when we consider the equation describing @xmath113 ( equation [ eqn : q ] ) . from this equation we see that @xmath111 and @xmath112 are both proportional to magnetic field strength , @xmath166 , @xmath111 is also proportional to velocity , @xmath167 , and @xmath112 to gradients in @xmath42 . figure [ fig : q1](f ) shows a plot of the total absolute flux through surfaces of constant @xmath52 , as a function of height . this is shown at @xmath60 hr , but similar curves are seen throughout the simulation . this is computed as : @xmath168 the flux is greatest low down , near the magnetic sources , then drops of rapidly with increasing height . this indicates that most connections between magnetic features close low down . from figure [ fig : q1](e ) , the rate of energy dissipation also rapidly drops with increasing height , having decreased by more than an order of magnitude by @xmath169 mm . therefore , the energy dissipation term @xmath113 has its largest effect close to the photosphere . in contrast to energy dissipated ( which from equation [ eqn : q ] is always @xmath170 ) , localised values of free magnetic energy density can be either positive or negative . in figure [ fig : free1](e ) and ( f ) black patches can be seen low down , indicating locations of negative free energy density , most likely occurring due to the cancellation of magnetic features on the photosphere removing energy . higher up , the free energy density tends to be only positive ( white patches ) . the combined effect of positive and negative regions low down results in the peak in total free energy density as a function of height occurring higher up , just below 1 mm ( figure [ fig : free1](b ) ) , where the free energy density becomes solely positive . the movie , , shows the rate of energy dissipation as a function of height for the first @xmath171 hr of the 3 g simulation . at the start of the movie , at greater heights , the rate of energy dissipation gradually increases until the curve becomes more or less steady . one can see occasional kinks in the curve at low @xmath52 , which then propagate upward . * * * * * * figures [ fig : q2](a ) and ( b ) show images of @xmath113 integrated in @xmath52 , in the @xmath89 plane , from the 3 g simulation . this is computed with units of ergs @xmath14 s@xmath15 as follows : @xmath172 as in figure [ fig : j ] , the colour table has been reversed , so that darker regions correspond to higher values of @xmath173 . the images are shown at ( a ) @xmath59 hr and ( b ) @xmath60 hr , and are saturated at a level of @xmath174 ergs @xmath14 s@xmath15 . as in figures [ fig : free1](c ) and ( d ) , contours of @xmath23 at @xmath46 mm are over - plotted . one can see that @xmath113 is more localised than the free magnetic energy . patches of @xmath113 tend to lie around the magnetic network , but not necessarily at the same locations as free magnetic energy seen in figures [ fig : free1](c ) and ( d ) . in addition , we do not see far - reaching bands of @xmath113 across the supergranules as we did with the free energy . in paper ii , it was determined that @xmath113 is mainly located at sites of changing magnetic connectivity and low down , near the magnetic elements . therefore , it is unsurprising that in these simulations , @xmath113 is seen mainly at the magnetic network , as this is where large numbers of magnetic elements lie , and the magnetic connectivity is constantly changing due to the continual interaction of these elements with one another . a movie showing @xmath113 in the @xmath89 plane for the 3 g simulation is available ( ) . it shows @xmath113 integrated in @xmath52 , and saturated at @xmath174 ergs @xmath14 s@xmath15 , with contours of @xmath23 at @xmath46 mm over - plotted . the movie shows that the locations and evolution of @xmath113 are very different from those of the positive free magnetic energy density ( ) . indeed , while many of the patches of positive free magnetic energy density were seen to be long - lived , regions of @xmath113 are seen to be very short - lived , occurring in rapidly evolving ` bursts ' . within the free energy density movie , patches of positive free energy density are often seen stretched across the supergranular cells , whereas @xmath113 tends to be much more localised , occurring predominantly within the magnetic network where large magnetic elements lie . several large bursts of @xmath113 can be seen throughout the movie , in regions where many magnetic elements are emerging and interacting with one another . for readers unable to view the movie , six still images from the movie spaced 1 hr apart are included in appendix [ sec : app1 ] ( figure [ fig : app3 ] ) . it can be seen that the spatial distribution of @xmath113 integrated in the los changes significantly from one hour to the next . note that the evolution of @xmath113 that we see in these movies is dominated by @xmath113 low down , since the rate of energy dissipation rapidly decreases with increasing @xmath52 ( figures [ fig : q1](e ) and ( f ) ) . we now consider the spatial location of @xmath113 as a function of height above the photosphere . figures [ fig : q2](c ) and ( d ) show @xmath89 plane images of @xmath113 at @xmath98 mm in the 3 g simulation , at @xmath59 hr and @xmath60 hr respectively . @xmath113 at height @xmath98 mm is given by @xmath175 . as in figures [ fig : q2](a ) and ( b ) , many patches of @xmath113 can be seen located above the boundaries between supergranules . however , in contrast to figures [ fig : q2](a ) and ( b ) , within figures [ fig : q2](c ) and ( d ) one can also see much more fine - scale structure to @xmath113 when it is not integrated along the line of sight . also in contrast to figures [ fig : q2](a ) and ( b ) , where @xmath113 integrated over @xmath52 is predominantly located at the magnetic network , here we see long strands of @xmath113 which lie across the supergranules at @xmath98 mm . these are localised regions of energy dissipation that are found mainly at sites of changing magnetic connectivity . a movie of @xmath113 above the photosphere is available ( ) showing @xmath113 in the @xmath89 plane , integrated between @xmath98 mm and @xmath176 mm : @xmath177 in units of ergs @xmath14 s@xmath15 . the movie is saturated at @xmath178 ergs @xmath14 s@xmath15 . at this height , the rate of energy dissipation is much lower than at the photosphere , as the coronal field is evolving more slowly and is less non - potential . the result of this is that @xmath113 is also less rapidly evolving than it is lower down . in addition , between these heights , @xmath113 is seen to occur anywhere within the @xmath89 plane and not just above the magnetic network . we also see more fine - scale , further reaching structures . six images of @xmath113 at @xmath98 mm spaced 1 hr apart are also included in appendix [ sec : app1 ] ( figure [ fig : app4 ] ) , to illustrate the time - scale of evolution at this height . figures [ fig : q2](e ) and ( f ) show @xmath96 plane images of @xmath113 integrated in @xmath76 , saturated at @xmath174 ergs @xmath14 s@xmath15 , at @xmath59 hr and @xmath60 hr , respectively . in agreement with figures [ fig : q1](e ) and ( f ) , the energy dissipation is seen to be greatest low down . finer - scale structure can then be seen between @xmath179 mm and @xmath180 mm . two movies of @xmath113 in the @xmath96 plane for the 3 g simulation are included with this paper . similar evolution is seen in the @xmath99 plane and for other strengths of overlying field . the first movie , , shows @xmath113 integrated in @xmath76 and saturated at the same level as in figures [ fig : q2](e ) and ( f ) . the second movie , , shows the logarithm of the first so that the energy dissipation can be seen for a wider range of values . from this it can be seen that fine - scale structures also exist higher up in the corona , where the energy dissipation is much weaker . in both movies , occasional ` bursts ' can be seen , where a feature will drift upwards and disappear ( or rather , become too faint to be shown at the current level of saturation ) . from these movies , we find that the energy dissipated and summed along the line of sight provides a fine - scale dynamic structure that is in qualitative agreement with what is observed on the sun low down . the aim of this paper was to carry out a preliminary analysis of a set of small - scale , non - linear force - free field simulations . the simulations were driven by synthetic magnetograms produced by the model described in paper i. four simulations were run , each driven by the same lower boundary data , three with different strengths of overlying field : 1 g , 3 g and 10 g , and one with no overlying field . the lower boundary data consisted of a @xmath6 hr series of synthetic magnetograms of area @xmath181 mm@xmath10 and of cadence 1 min . the initial coronal magnetic field for each simulation was potential . this field was then evolved through a series of quasi - static , non - linear force - free equilibria , via a magnetofrictional relaxation technique , in response to photospheric boundary motions . the continuous nature of this coronal evolution technique means that current systems are maintained within the corona from one step to the next and the evolution is smooth . this allows for the build - up and storage of free magnetic energy @xmath5 one of the quantities studied within this paper . the presence of free magnetic energy within our model shows a significant departure from previous models of the magnetic carpet coronal field , which use potential fields . the other quantities considered were the energy dissipated and the square of the electric current density , @xmath8 . initially , for each simulation , both the free magnetic energy and energy dissipation rate rapidly increase , before levelling off and oscillating about a mean value . the mean free magnetic energy per unit area for each simulation is @xmath182 ergs @xmath14 , whilst the mean energy dissipation rate is @xmath183 ergs @xmath14 s@xmath15 , resulting in @xmath184 ergs @xmath14 ( @xmath185 ergs ) being cumulatively dissipated over each @xmath6 hr simulation . for both the free and dissipated energies , a stronger overlying field results in higher values , although the effect is more significant for the free energy . it is also clear that the evolution of both the free and dissipated energies is highly dependent upon the evolution of the photospheric magnetic field . while there are similarities between the evolution of the two types of energy integrated over the volume , they are seen to be less alike when we consider their location within each simulation . the bulk of the free magnetic energy is located above the photosphere , between @xmath186 mm . this is stored along closed connections between magnetic elements . regions of positive free energy density can be seen both in the magnetic network and across supergranular cells ; such regions may also be long - lived . in contrast , the largest amount of energy dissipation is found low down , near the magnetic sources , and values decrease rapidly with increasing height . regions of increased energy dissipation are seen predominantly in the magnetic network , although weaker , fine - scale strands are also seen above the photosphere at sites of changing magnetic connectivity . also unlike the free magnetic energy density , the large regions of energy dissipation seen in the @xmath89 plane are much more rapidly evolving ( compare movies and ) . the amount of free magnetic energy built up and stored in each simulation ( @xmath187 ergs ) is sufficient to account for such small - scale transient phenomena as nanoflares ( @xmath188 ergs , parker , @xcite ) and x - ray bright points ( @xmath189 ergs s@xmath15 , habbal and withbroe , @xcite ; longcope , @xcite ) . the energy dissipation rate is not high enough to be able to explain the heating of the quiet corona , providing a contribution of around @xmath141 to the required @xmath142 ergs @xmath14 s@xmath15 of or @xmath143 to the required @xmath144 ergs @xmath14 s@xmath15 of . in fact , the discrepancy may be even larger since did not include the energy losses due to thermal conduction . however , the location and structure of regions of energy dissipation within our model are at least in qualitative agreement with what is observed on small scales on the sun . the lower rate of energy dissipation may in part be due to the simplified magnetofrictional model that has been applied , but also due to an insufficient rate of energy being injected in the first place . an order of magnitude estimate of the poynting flux injected showed that this was within a factor or @xmath161 of the mean energy dissipated in each simulation . this indicates that increasing the magnitude of the horizontal velocities or the horizontal or vertical magnetic field components at the photosphere , hence increasing the poynting flux , could result in higher energy dissipation . another reason for our poynting flux ( and hence energy dissipated ) being lower than observed values is that in the present model , our magnetic features are treated as ` rigid ' bodies . if the magnetic field of an actual feature were instead contained within many intense kilogauss flux tubes , the movement of features would be more fluid ( as is observed ) as the intense flux tubes have high local velocities . this would lead to a greater injection of poynting flux . it should also be noted that our model does not include a chromosphere . determined chromospheric radiative losses to be @xmath190 ergs @xmath14 s@xmath15 , while the value determined by is an order of magnitude higher at @xmath191 ergs @xmath14 s@xmath15 . a more complex model that included a chromosphere with the effects of , for example , alfvn waves and turbulence would likely result in a much higher rate of energy dissipation ( _ e.g. _ ) . cccc + quantity & location & variation & effect of o / l + & & in height & field strength + free & magnetic network and & mostly stored @xmath186 mm & significant + energy & across s / g cells . & then rapid decrease & increase in + & & with increasing @xmath52 . & total with + & & & increasing @xmath51 + & & & + @xmath113 & magnetic network and & rapid decrease with & slight + & sites of changing & increasing @xmath52 & increase in + & magnetic connectivity . & & total with + & & & increasing @xmath51 + & & & + @xmath8 & magnetic network and & & + & non - potential regions . & @xmath5 & @xmath5 + & follows field lines . & & + locations of increased @xmath8 are found to be co - located with regions of positive free magnetic energy density , as both are dependent on the magnetic field being non - potential . visually , @xmath8 is seen to follow the shape of the magnetic field where @xmath42 is non - zero . table [ tab : summary ] gives a summary of results for each of the quantities studied in this paper , indicating their locations and the effect of increasing the strength of the overlying field . there are several avenues for future work using the non - linear force - free coronal modelling technique described here . the simulations presented in this paper will be studied in more detail . for example , a more in depth study of the connectivity of the magnetic field will be conducted , similar to the study of , who analysed potential coronal fields . it would also be of interest to investigate in detail the effect of each of the photospheric flux evolution processes on the evolution of the coronal field and on each of the energy quantities discussed in this paper . we already know the flux , location and time of occurrence of each of the processes within the synthetic magnetogram series . further 2d simulations will also be run to produce synthetic magnetograms that include an evolving supergranular flow profile and magnetic features on smaller scales . the added complexity of evolving photospheric flows and smaller magnetic features will likely result in a larger build up of free magnetic energy and greater energy dissipation , when the coronal field evolution is simulated . the rate of energy dissipation per unit area produced by our present simulations was found to be too low to explain quiet sun radiative losses . this is unsurprising , however , as many simplifications have been made for our coronal model and it is not a true physical energy dissipation @xmath5 we do not have currently have an energy equation . for example , the present simulations do not include the effects of small - scale magnetic braiding and mhd waves , which are believed to be important for heating the chromosphere and corona in active regions @xcite . in future we intend to extend the present magnetofrictional model to contain plasma by including the pressure and density terms in the equations of magnetohydrodynamics . we will then be able to follow the corresponding plasma processes associated with energy dissipation and give an estimate of the resultant plasma heating . the magnetofrictional technique will also be applied to real magnetogram data , such as from hinode / sot or sdo / hmi . a study will be carried out to compare regions of interest within a simulated non - linear force - free coronal field driven by observed magnetograms to events occurring in corresponding coronal images ( _ e.g. _ using various wavelengths of sdo / aia ) . what is clear is that within this single , relatively ` simple ' simulation , a wide range of processes and dynamics is occurring . a careful and detailed analysis of these and other non - linear force - free field simulations of the magnetic carpet coronal field will be carried out in future . kam and dhm gratefully acknowledge the support of the leverhulme trust and the stfc . dhm would like to thank the royal society for their support through the research grant scheme . dhm and cep acknowledge support from the eu under fp7 . this appendix contains four figures , each with six images spaced 1 hr apart from @xmath82 hr to @xmath192 hr , taken from some of the @xmath89 plane movies included with this paper .
+ this paper is the third in a series of papers working towards the construction of a realistic , evolving , non - linear force - free coronal field model for the solar magnetic carpet . here , we present preliminary results of 3d time - dependent simulations of the small - scale coronal field of the magnetic carpet . four simulations are considered , each with the same evolving photospheric boundary condition : a 48 hr time series of synthetic magnetograms produced from the model of . three simulations include a uniform , overlying coronal magnetic field of differing strength , the fourth simulation includes no overlying field . the build - up , storage and dissipation of magnetic energy within the simulations is studied . in particular , we study their dependence upon the evolution of the photospheric magnetic field and the strength of the overlying coronal field . we also consider where energy is stored and dissipated within the coronal field . the free magnetic energy built up is found to be more than sufficient to power small - scale , transient phenomena such as nanoflares and x - ray bright points , with the bulk of the free energy found to be stored low down , between @xmath0 mm . the energy dissipated is presently found to be too small to account for the heating of the entire quiet sun corona . however , the form and location of energy dissipation regions are in qualitative agreement with what is observed on small scales on the sun . future mhd modelling using the same synthetic magnetograms may lead to a higher energy release .
in this letter , we present _ dark mediator dark matter _ ( dmdm ) to address two important gaps in the dm literature : exploring mediators with dark charge , and non - standard interaction topologies for scattering off nuclei . additional details and constraints will be explored in a companion paper @xcite . the existence of dark matter is firmly established by many astrophysical and cosmological observations @xcite , but its mass and coupling to the standard model ( sm ) particles are still unknown . weakly interacting massive particles ( wimps ) are the most popular dm candidates since they arise in many theories beyond the sm , including supersymmetry , and may naturally give the correct relic abundance @xcite . however , improved experimental constraints from collider searches , indirect detection and direct detection @xcite begin to set tight limits ( with some conflicting signal hints ) on the standard wimp scenario with a contact interaction to quarks . this makes it necessary to look for a more complete set of dm models which are theoretically motivated while giving unique experimental signatures . process at tree - level . right : the loop - induced @xmath6 process . the arrows indicate flow of dark charge . , width=302 ] given its apparently long lifetime , most models of dm include some symmetry under which the dm candidate is charged to make it stable . an interesting possibility is that not only the dm candidate , but also the mediator connecting it to the visible sector is charged under this dark symmetry . such a ` dark mediator ' @xmath0 could only couple to the sm fields in pairs . as a simple example , consider real or complex sm singlet scalars @xmath11 coupled to quarks , along with yukawa couplings to a dirac fermion dm candidate @xmath12 . the terms in the effective lagrangian relevant for direct detection are @xmath13 where @xmath14 stands for @xmath15 mass terms , as well as the rest of the dark sector , which may be more complicated than this minimal setup . this interaction structure can be enforced by a @xmath16 symmetry . to emphasize the new features of this model for direct detection , we focus on the minimal case with a single mediator @xmath17 ( omitting the @xmath18-index ) . however , the actual number of dark mediators is important for interpreting indirect constraints @xcite . the leading order process for dm - nucleus scattering is @xmath19 if @xmath20 . however , an elastic scattering @xmath21 is always present at loop - level since it satisfies all possible symmetries , see fig . [ f.feynmandiagram ] . which of the two possibilities dominates direct detection depends on the size of the yukawa couplings @xmath22 as well as the dark mediator masses . previous modifications to wimp - nucleon scattering kinematics include the introduction of a mass splitting @xcite ; considering matrix elements @xmath23 with additional velocity- or momentum transfer suppressions ( for a complete list see e.g. @xcite ) , especially at low dm masses close to a gev @xcite ; light scalar or ` dark photon ' mediators ( see e.g. @xcite which give large enhancements at low nuclear recoil ) ; various forms of composite dark matter @xcite which may introduce additional form factors ; dm - nucleus scattering with intermediate bound states @xcite which enhances scattering in a narrow range of dm velocities ; and induced nucleon decay in asymmetric dark matter models @xcite . notably missing from this list are alternative process topologies for dm - nucleus scattering . this omission is remedied by the dmdm scenario . dmdm is uniquely favored to produce a detectable @xmath24 scattering signal at direct detection experiments . this is because it contains two important ingredients : ( 1 ) a light mediator with non - derivative couplings to enhance the cross section , compensating for the large suppression of emitting a relativistic particle in a non - relativistic scattering process , and ( 2 ) a scalar as opposed to a vector mediator , allowing it to carry dark charge ( without a derivative coupling ) . this imposes selection rules which make the @xmath25 process subleading in @xmath26 . these ingredients are difficult to consistently implement in other model constructions without violating constraints on light force carriers . the effect of strong differences between proton and neutron coupling to dm have been explored by @xcite . to concentrate on the kinematics we shall therefore assume the operator @xmath1 is flavor - blind in the quark mass basis . above the electroweak symmetry breaking scale this operator is realized as @xmath27 . it can be generated by integrating out heavy vector - like quarks which couple to the sm and @xmath0 @xcite , giving @xmath28 . this uv completion allows for large direct detection cross sections without being in conflict with collider bounds , but may be still probed at the 14 tev lhc . [ cols= " < " , ] + in a silicon , germanium and xenon target . the dashed lines are spectra of standard wimp scattering ( via operator @xmath29 , with @xmath30 ) shown for comparison . dmdm spectra computed with ` madgraph5 ` @xcite and ` feynrules1.4 ` @xcite . , title="fig:",width=264 ] we start by examining the novel @xmath3 regime of dmdm . the dm - nucleus collision is inelastic , not by introducing a new mass scale like a splitting , but by virtue of the process topology . the nuclear recoil spectrum is different compared to previously explored scenarios . this is illustrated in fig . [ f.partonleveldistribution ] , where we compare nuclear recoil spectra of standard wimps to dmdm for fixed velocity and different nucleus mass , _ before _ convolving with various form factors and the ambient dm speed distribution . the observable dmdm differential cross section is independent of @xmath10 for @xmath31 and can be well described by the function @xmath32 where @xmath33 @xmath34 and @xmath35 , same as the wimp case for a given dm velocity . ( we emphasize that this is a phenomenological description , the actual spectra were produced in madgraph , see section [ s.directdetection ] . ) the first factor comes from the light mediator propagator @xmath36 as well the integrated phase space of the escaping @xmath0 . the cross section suppression ( second factor ) is more pronounced as the dm becomes lighter or slower , and as the nucleus becomes heavier , both of which reduces @xmath37 . this is because massless @xmath0 emission carries away a more significant fraction of the total collision energy if the heavy particle momenta are smaller . the maximum kinematically allowed nuclear recoil is then less likely . when @xmath38 , the @xmath6 process will dominate direct detection for yukawa coupling @xmath26 above some threshold , or if @xmath39 . for the purpose of calculating the matrix element , the loop diagram in fig . [ f.feynmandiagram ] ( right ) is equivalent to the operator @xmath40 where @xmath41 is the momentum transfer in the scattering . effectively , this is identical to a standard wimp with a @xmath42 contact operator , but with an additional @xmath43 suppression in the cross section . this gives a similar phenomenology as a light mediator being exchanged at tree - level with derivative coupling . note that the relative importance of these two scattering processes is highly model dependent . for example , if @xmath44 the dominant scalar - dm coupling could be @xmath45 . in that case , the @xmath25 operator above is @xmath46 and can be suppressed without reducing the @xmath24 rate by taking @xmath47 . the scattering behavior of both the @xmath3 and @xmath6 regimes necessitates a re - interpretation of all dm direct detection bounds . we will do this below . direct detection experiments probe the ratio @xmath48 and @xmath49 for @xmath24 and @xmath25 scattering respectively . however , indirect constraints on dmdm from cosmology , stellar astrophysics and collider experiments are sensitive to the yukawa coupling and @xmath50 separately . in @xcite we conduct an extensive study of these bounds , including the first systematic exploration of constraints on the @xmath1 operator with light scalars @xmath0 . since these constraints ( in particular , eqns . [ e.nscoolingbound ] and [ e.thermalrelic ] below ) provide important context for our results on direct detection , we summarize the two most important results here . for details we refer the reader to @xcite . the scalar mediator(s ) of dmdm are most stringently constrained from stellar astrophysics and cosmology : * avoiding overclosure requires @xmath51 @xcite , so we take the heaviest stable @xmath0 to be essentially massless , making it a very subdominant dark matter component . this also satisfies structure formation , computed for light sterile neutrinos in @xcite . measurements by the planck satellite @xcite restrict the number of light degrees of freedom during big bang nucleosynthesis , enforcing the bound @xmath52 for real scalars . * the coupling of @xmath0 to the sm is most constrained from stellar astrophysics . for @xmath17 , observational data on neutron star cooling essentially rules out any directly detectable dmdm model @xcite . however , this bound is easily relaxed for @xmath44 if @xmath53 , @xmath54 , with a cosmologically unstable @xmath55 . the dominant interaction to the sm is assumed to be @xmath56 . in that case , @xmath55 emission in the neutron star is boltzmann suppressed due to its core temperature of @xmath57 , and @xmath58 emission proceeds via a loop process . the bound on @xmath50 is then weakened to @xmath59 * in supernovae , emission of light invisible particles can truncate the neutrino burst @xcite . however , if these particles interact with the stellar medium more strongly than neutrinos they are trapped and do not leak away energy from the explosion . the temperature of supernovae @xmath60 is large enough to produce @xmath61 at tree - level in the above @xmath44 scenario , and the scattering cross section with nuclei is much larger than for neutrinos if @xmath62 . therefore this setup is compatible with supernovae constraints . * the lhc can set constraints on heavy dark vector quarks in a possible uv completion of dmdm . the cms @xmath63 di - jet + met search @xcite search sets a lower bound on the heavy quark to be @xmath64 tev . the physics of direct detection for this @xmath44 setup is identical to the minimal @xmath17 model . this is because the typical momentum transfer is @xmath65 , making the intermediate @xmath55 mediator massless for the purposes of direct detection . we are therefore justified in examining the direct detection phenomenology of the @xmath17 model in detail , applying the @xmath50 bound eqn . [ e.nscoolingbound ] and with the understanding that the full realization of dmdm requires a slightly non - minimal spectrum . the dark matter yukawa coupling is constrained from observations on large scale structure and ( under certain assumptions ) from cosmology : * dark matter self - interaction bounds from bullet cluster observations constrain the dm yukawa coupling to be @xmath66 @xcite . * a thermal relic @xmath12 with @xmath67 requires @xmath68 if there is no significant @xmath69 term . this also satisfies the above self - interaction bounds . interestingly , the range of @xmath70 to mev that is relevant for its novel direct detection signal also makes dmdm a realization of self - interacting dm ( sidm ) @xcite . a yukawa interaction consistent with @xmath12 being a thermal relic can then help resolve the `` core / cusp '' and `` too - big - to - fail '' inconsistencies between dwarf galaxy observations and many - body simulations @xcite . we compute dmdm nuclear recoil spectra at direct detection experiments by simulating the parton - level process in ` madgraph5 ` @xcite , and derive the event rates according to @xmath71 where @xmath72 is the local dm speed distribution ( approximate maxwell - bolzmann with @xmath73 km / s and a @xmath74 km / s cutoff , boosted into the earth frame @xmath75 @xcite ) , while @xmath76 is the local dm density @xcite , and @xmath77 is the target number density per kg . @xmath78 includes the usual helm nuclear form factor @xcite , the @xmath79 coherent scattering enhancement as well as the quark - nucleon form factor for scalar interactions ( see @xcite for a review ) . we validated our monte carlo pipeline by reproducing analytically known @xmath6 results . [ f.ddspectracdms ] shows some nuclear recoil spectra for a silicon and xenon target . ( we henceforth assume an effectively massless @xmath0 . ) dmdm is compared to standard wimps ( velocity- and recoil - independent contact - interaction ) for different dm masses . an important feature of our model is apparent : a @xmath80 dmdm candidate looks like a @xmath81 @xmath82 ) wimp when scattering off silicon ( xenon ) . moreover , the shape of @xmath78 is insensitive to @xmath83 unless @xmath83 is much smaller than @xmath84 ( see eq . [ eq : recoil ] ) . this makes it much more difficult to measure the dm mass using the shape of the spectrum . signals at two detectors with different target materials are required . + m6.3cmm6.3 cm this plot shows the wimp mass @xmath85 which gives the same spectral shape at xenon100 ( @xmath86 with 6% light gathering efficiency , dashed red line ) , lux ( @xmath87 with 14% light gathering efficiency , dash - dotted black line ) , cdmsii silicon ( @xmath88 , solid blue line ) , and cdmslite ( germanium , @xmath89 , dotted purple line ) before selection cuts . * right : * the ` observed ' wimp - nucleon cross section for each dmdm mass @xmath90 , assuming the best - fit @xmath91 from the left . the dmdm parameters are @xmath92 . , title="fig:",width=238 ] & this plot shows the wimp mass @xmath85 which gives the same spectral shape at xenon100 ( @xmath86 with 6% light gathering efficiency , dashed red line ) , lux ( @xmath87 with 14% light gathering efficiency , dash - dotted black line ) , cdmsii silicon ( @xmath88 , solid blue line ) , and cdmslite ( germanium , @xmath89 , dotted purple line ) before selection cuts . * right : * the ` observed ' wimp - nucleon cross section for each dmdm mass @xmath90 , assuming the best - fit @xmath91 from the left . the dmdm parameters are @xmath92 . , title="fig:",width=238 ] we can make this observation more concrete by mapping dmdm parameters to wimp parameters . this is possible because both sets of nuclear recoil spectra look roughly like falling exponentials . for each dmdm spectrum with a given mass there is a closely matching wimp spectrum with some different ( lower ) mass . to find the @xmath93 corresponding to each @xmath94 we compare binned wimp and dmdm distributions and minimize the total relative difference in each bin . the resulting mapping is shown in fig . [ f.compare2to2to2to3 ] ( left ) . even very heavy dmdm candidates mimic light wimps of different masses at different experiments . a corresponding cross section remapping ( right ) shows that experiments with heavier nuclei are more sensitive to dmdm due to the inelastic nature of the collision . regime of dmdm . the vertical axis is proportional to @xmath95 . _ solid lines _ : 90% cl bounds by xenon100 ( red ) , lux ( black ) and cdmslite ( purple ) , as well as the best - fit regions by cdms ii si ( blue , green ) . the large - dashed black line indicates the irreducible neutrino background @xcite . _ small - dashed magenta line _ : upper bound for @xmath96 and neutron star cooling bound @xmath97 . _ lower dotted orange line _ : upper bound for @xmath24 dominated direct detection and neutron star bound with all equal yukawa couplings . this line can be arbitrarily moved , as discussed below eq . [ e.2to2operator ] . the _ upper dotted orange line _ is for @xmath98 , in which case the vertical axis is understood to be @xmath99 . * right : * direct detection bounds on the @xmath6 regime of @xmath17 dmdm , same labeling as the left plot . the vertical axis is proportional to @xmath100 , and is understood to be @xmath101 for the @xmath44 model outlined in section [ s.indirectconstraints ] . , title="fig:",width=302 ] regime of dmdm . the vertical axis is proportional to @xmath95 . _ solid lines _ : 90% cl bounds by xenon100 ( red ) , lux ( black ) and cdmslite ( purple ) , as well as the best - fit regions by cdms ii si ( blue , green ) . the large - dashed black line indicates the irreducible neutrino background @xcite . _ small - dashed magenta line _ : upper bound for @xmath96 and neutron star cooling bound @xmath97 . _ lower dotted orange line _ : upper bound for @xmath24 dominated direct detection and neutron star bound with all equal yukawa couplings . this line can be arbitrarily moved , as discussed below eq . [ e.2to2operator ] . the _ upper dotted orange line _ is for @xmath98 , in which case the vertical axis is understood to be @xmath99 . * right : * direct detection bounds on the @xmath6 regime of @xmath17 dmdm , same labeling as the left plot . the vertical axis is proportional to @xmath100 , and is understood to be @xmath101 for the @xmath44 model outlined in section [ s.indirectconstraints ] . , title="fig:",width=302 ] fig . [ f.compare2to2to2to3 ] defines an experiment - dependent parameter map that we can use to map each collaboration s wimp bounds onto the dmdm model if @xmath24 scattering dominates . the resulting direct detection bounds are shown in fig . [ f.mappingbounds ] ( left ) . we include the irreducible neutrino background @xcite at the lux experiment , which serves as an approximate lower border of the observable dmdm parameter space . an identical procedure can be used in the @xmath25 dominant regime of dmdm . the translation map has similar qualitative features to the previous case since @xmath102 , except the faked wimp signal corresponds to somewhat higher mass . the resulting direct detection bounds are shown in fig . [ f.mappingbounds ] ( right ) . the probability for any one @xmath6 nuclear recoil event to lie above experimental detection threshold is much larger than for a @xmath3 event , due to the less severe recoil suppression . for @xmath17 , this means the former will dominate direct detection unless @xmath31 and the yukawa coupling is very small , @xmath103 . however , as discussed in section [ s.indirectconstraints ] , the neutron star cooling constraint requires at least @xmath44 . the @xmath25 process could then be arbitrarily suppressed , allowing @xmath24 direct detection with a thermal relic @xmath12 . for the @xmath24 and @xmath25 scattering regimes , direct detection probes @xmath48 and @xmath49 respectively . the neutron star cooling bound @xmath104 and the bounds on dark matter yukawa coupling @xmath26 can be combined to be shown in the direct detection planes of fig . [ f.mappingbounds ] . the assumption of a thermal relic then excludes the regions in fig . [ f.mappingbounds ] above the magenta dashed line , meaning these bounds supersede the liquid xenon experiments for @xmath105 in the @xmath24 dominant regime . there are large discoverable regions of dmdm parameter space that are not excluded . due to the nontrivial dependence of the dmdm recoil spectrum on the target- and dark - matter masses and velocity , signals at several experiments will be needed to differentiate standard wimps from our model , but dmdm offers the realistic prospect of tev - scale heavy quark discoveries pointing the way towards a sensitivity target for direct detection . _ dark mediator dark matter _ introduces the possibility that dark matter interacts with the standard model via a mediator which also carries dark charge . this `` double - dark portal '' adds the phenomenon of additional particle emission to the menu of possible interactions with nuclei , serving as an existence proof that this scattering topology can be realized . direct detection experiments are starting to probe interesting regions of parameter space compatible with a thermal relic and neutron star bounds . for observationally relevant parameters , dmdm also acts as an implementation of sidm @xcite , which can resolve various inconsistencies between many - body simulations and observations for dwarf galaxies . even more than many other dm models , dmdm discovery is aided by lowering nuclear recoil thresholds . further investigation is warranted and includes potential lhc signals , as well as possible leptophilic realizations of the model . * acknowledgements * the authors would like to gratefully acknowledge the valuable contributions of yue zhao during an earlier stage of this collaboration . we thank rouven essig , patrick fox , roni harnik and patrick meade for valuable comments on draft versions of this letter . we are very grateful to joseph bramante , rouven essig , greg gabadadze , jasper hasenkamp , matthew mccullough , olivier mattelaer , matthew reece , philip schuster , natalia toro , sean tulin , neal weiner , itay yavin and hai - bo yu for valuable discussions . d.c . and z.s . are supported in part by the national science foundation under grant phy-0969739 . y.t . is supported in part by the department of energy under grant de - fg02 - 91er40674 . the work of y.t . was also supported in part by the national science foundation under grant no . phys-1066293 and the hospitality of the aspen center for physics d. curtin and y. tsai , arxiv:1405.1034 [ hep - ph ] . p. a. r. ade _ et al . _ [ planck collaboration ] , arxiv:1303.5076 [ astro-ph.co ] . see , for example , g. jungman , m. kamionkowski and k. griest , phys . rept . * 267 * , 195 ( 1996 ) [ hep - ph/9506380 ] . g. bertone and d. merritt , mod . a * 20 * , 1021 ( 2005 ) [ astro - ph/0504422 ] . g. angloher , m. bauer , i. bavykina , a. bento , c. bucci , c. ciemniak , g. deuter and f. von feilitzsch _ et al . _ , j. c * 72 * , 1971 ( 2012 ) [ arxiv:1109.0702 [ astro-ph.co ] ] . r. agnese _ et al . _ [ cdms collaboration ] , [ arxiv:1304.4279 [ hep - ex ] ] . p. w. graham , r. harnik , s. rajendran and p. saraswat , phys . d * 82 * , 063512 ( 2010 ) [ arxiv:1004.0937 [ hep - ph ] ] . r. essig , j. kaplan , p. schuster and n. toro , [ arxiv:1004.0691 [ hep - ph ] ] . j. march - russell , j. unwin and s. m. west , jhep * 1208 * , 029 ( 2012 ) [ arxiv:1203.4854 [ hep - ph ] ] . b. feldstein , a. l. fitzpatrick and e. katz , jcap * 1001 * , 020 ( 2010 ) [ arxiv:0908.2991 [ hep - ph ] ] . y. bai and p. j. fox , jhep * 0911 * , 052 ( 2009 ) [ arxiv:0909.2900 [ hep - ph ] ] . h. davoudiasl , d. e. morrissey , k. sigurdson and s. tulin , phys . rev . d * 84 * , 096008 ( 2011 ) [ arxiv:1106.4320 [ hep - ph ] ] . j. alwall , m. herquet , f. maltoni , o. mattelaer and t. stelzer , jhep * 1106 * , 128 ( 2011 ) [ arxiv:1106.0522 [ hep - ph ] ] . s. ask , n. d. christensen , c. duhr , c. grojean , s. hoeche , k. matchev , o. mattelaer and s. mrenna _ et al . _ , arxiv:1209.0297 [ hep - ph ] . m. rocha , a. h. g. peter , j. s. bullock , m. kaplinghat , s. garrison - kimmel , j. onorbe and l. a. moustakas , mon . not . * 430 * , 81 ( 2013 ) [ arxiv:1208.3025 [ astro-ph.co ] ] . j. zavala , m. vogelsberger and m. g. walker , monthly notices of the royal astronomical society : letters * 431 * , l20 ( 2013 ) [ arxiv:1211.6426 [ astro-ph.co ] ] .
we introduce _ dark mediator dark matter _ ( dmdm ) where the dark and visible sectors are connected by at least one light mediator @xmath0 carrying the same dark charge that stabilizes dm . @xmath0 is coupled to the standard model via an operator @xmath1 , and to dark matter via a yukawa coupling @xmath2 . direct detection is realized as the @xmath3 process @xmath4 at tree - level for @xmath5 and small yukawa coupling , or alternatively as a loop - induced @xmath6 process @xmath7 . we explore the direct - detection consequences of this scenario and find that a heavy @xmath8 dmdm candidate fakes different @xmath9 standard wimps in different experiments . large portions of the dmdm parameter space are detectable above the irreducible neutrino background and not yet excluded by any bounds . interestingly , for the @xmath10 range leading to novel direct detection phenomenology , dmdm is also a form of self - interacting dark matter ( sidm ) , which resolves inconsistencies between dwarf galaxy observations and numerical simulations .
we address the modeling and simulation of elastoplastic solids in multidomain environments including also mechatronic multibody systems with nonsmooth dynamics , such as vehicles , robots and processing machinery . fast multidomain simulation is useful for concept design exploration , development of control algorithms and for interactive realtime simulators , e.g. , for operator training , human - machine interaction studies and hardware - in - the - loop testing . realizing such simulators require integrating many subsystems of different types and complexity into a single multidomain dynamics model . if the subsystems are loosely coupled , the full system simulation can be realised by means of co - simulation @xcite , but this is not the general case . for the sake of computational performance , coarse grain models are often used , with rigid and flexible bodies coupled by kinematic constraints for modeling of joints and differential algebraic equation ( dae ) models for electronics , hydraulics and powertrain dynamics @xcite . at coarse timescales the dynamics need to be treated as nonsmooth , allowing discontinuous velocities of rigid bodies undergoing impacts or frictional stick - slip transitions and instantaneous propagation of impulses through the system . this requires strong coupling through consistent mathematical formulation and coherent numerical treatment of the full system in order to achieve good stability and computational efficiency . current techniques for co - simulation does not support that . the theory and numerical methods for nonsmooth multidomain mechanics is covered in the reference @xcite . the framework in @xcite is employed for constraint regularisation and stabilisation based on a discrete variational approach with constraints introduced as the stiff limit of energy and dissipation potentials . the developed method is applicable to many different systems but one is primarily explored , namely the interaction between ground vehicles and deformable terrain @xcite . current solutions enable simulation of complex vehicle models with realtime performance but not including dynamic terrain models firmly based on solid mechanics in three dimensions . usually , empirical terrain models of bekker - wong type are used @xcite . for more accurate offline simulations with fine temporal resolution , on the other hand , many solutions exist for elastoplastic solids coupled with tire models @xcite but scarcely with more complex multibody systems and not for realtime or faster simulation . the main contribution in this paper , illustrated in figure [ fig : illustration ] , is a formulation and numerical method for simulation of elastoplastic solids as a nonsmooth multidomain multibody system on descriptor form @xcite . this enables numerical integration with large timesteps that potentially exceed current solutions and provide realtime performance or better . the regularisation and damping terms regulate the elasticity and viscous damping of deformations according to linear elasticity theory . an associative perfectly plastic drucker - prager model is employed using an elastic predictor - plastic corrector strategy to detect yielding and compute the plastic flow . in its basic form this model does not yield in hydrostatic compression in contrast to many real materials . many soils yield under hydrostatic compression by failure in the microscopic structures whereby air and fluid is released . therefore the drucker - prager model is extended to a capped version @xcite . the dynamics of other subsystems and their potentially strong coupling is treated within the same multibody dynamics framework using a variational time integrator . each timestep involve solving a block - sparse mixed linear complementarity problem mlcp , and can be efficiently integrated with tailored solvers . a meshfree method @xcite is chosen in order to handle large deformations without need for remeshing @xcite and , for future development , support fracturing and transitions to viscous flow or to granular media represented by contacting discrete elements . the displacement gradient and strain tensor is approximated by the method of moving least squares ( mls ) @xcite . matrices and vectors are represented in bold face in capital and lower case , @xmath0 and @xmath1 , respectively . latin superscripts indicate the index of a specific body @xmath2 , where @xmath3 is the total number of bodies in the system . greek subscripts indicate a specific scalar component of a vector or matrix ( may be multidimensional ) . the einstein summation convention is used where repeated indices imply summation over them , e.g. , representing matrix - vector multiplication as @xmath4 . subscripts @xmath5 or @xmath6 indicate a specific component of a vector or matrix assuming a cartesian coordinate system . as an example of these notations the position vector of a body @xmath7 is denoted @xmath8 . the relative position of particle @xmath9 to @xmath7 is @xmath10 . the @xmath11 component of this vector is @xmath12 . for rigid bodies , the position vector include also the rotation of the body . the system position vector is @xmath13 $ ] . the gradient is denoted @xmath14 and @xmath15 . the dot notation is used for the time derivative @xmath16 . the lagrangian of a constrained mechanical system is @xmath17 where @xmath18 is the kinetic energy , @xmath19 is the potential energy and @xmath20 is the rayleigh dissipation function . the system state vector @xmath21 is constrained by scleronomic holonomic constraints @xmath22 , with jacobian @xmath23 , and nonholonomic pfaffian constraints @xmath24 with corresponding lagrange multipliers @xmath25 and @xmath26 . to ensure numerical stability it is common to regularise the constraints . this may be done by treating them as the limit of strong potentials and dissipation functions @xmath27 and @xmath28 with @xmath29 @xcite . in fact , any stiff force can be transformed into a regularised constraint through a legendre transform @xcite @xmath30 this transform the euler - lagrange equations of motion into @xmath31 where @xmath32 are the explicit forces from weak smooth potentials . this constitute a system of dae of index 1 , which are easier to integrate numerically than the corresponding higher index dae in the absence of regularisation and stabilisation . a term for dissipation of motion orthogonal to the holonomic constraint surface can be added to improve the convergence . also this dissipation can be physically based by introducing it as a legendre transform of a rayleigh dissipation function @xmath34 , with damping parameter @xmath35 . this modifies eq . to @xmath36 the compliance and damping factors @xmath37 , @xmath35 and @xmath38 are not restricted to being scalar or diagonal . in what follows these are assumed to be matrices . variational integration @xcite provides a systematic approach to derive time integration schemes with good properties , e.g. , momentum preservation and symplecticity . rather than discretising the euler - lagrange equations of motion directly , the lagrangian and principle of least action is defined in time - discrete form . employing semi - implicit euler discretisation and linearising the constraint as @xmath39 leads to the following scheme @xmath40 @xmath41 where @xmath42 , @xmath43 and regularisation and stabilisation matrices @xmath44 , @xmath45 and @xmath46 . . is a linear system of @xmath47 equations . the matrix on the left hand side is block - sparse , positive definite and non - symmetric . the regularisation appearing as the diagonal perturbation matrices @xmath48 and @xmath49 are needed for handling otherwise ill - posed or ill - conditioned problems , e.g. , systems with constraint degeneracy and large mass ratios . the stabilisation terms @xmath50 on the right hand side counteract constraint violations , e.g. , sudden and large contact penetrations at impacts or small numerical constraint drift . the presented stepping scheme , referred to as spook , has been proved to be linearly stable @xcite and numerical simulations suggest a large domain of nonlinear stability . the combined effect of the regularisation and stabilisation terms is to bring elastic and viscous properties to motion orthogonal to the constraint surface @xmath51 , e.g. , for modeling elasticity in mechanical joints . the parameters @xmath37 and @xmath38 need not be chosen arbitrarily , as for the conventional approaches to constraint regularisation and stabilisation , but can be based on physics models using parameters that can be derived from first principles , found in literature or be identified by experiments . this becomes straightforward when the regularisation is introduced by potential energy as quadratic functions in @xmath52 , i.e. , @xmath53 . this has been exploited in previous work to constraint based modeling of lumped element beams @xcite , wires @xcite , meshfree fluids @xcite and granular material @xcite . in discrete time , some of the dynamics is best treated as _ nonsmooth _ @xcite , meaning that the velocity may change discontinuously to fulfil inequality constraints and complementarity conditions . introducing such conditions is an efficient way of modeling contacts , dry friction , joint limits , electric circuit switching , motor and actuator dynamics and control systems for discrete time with large step size . the alternative would be to use a fine enough temporal resolution where the dynamics appear smooth and may be modeled by daes or ordinary differential equations alone . in general , this is an intractable approach for full system simulations and interactive applications . the nonsmooth dynamics can be formally treated as _ differential variational inequalities _ @xcite . the discrete equations of motion take the form of a _ mixed linear complementarity problem _ , mlcp , when nonsmooth dynamics is included @xmath54 where @xmath55 corresponds to the left hand side of equation , @xmath56 is the negated right hand side and @xmath57 are slack variables . the terms @xmath58 and @xmath59 correspond to the upper and lower limits on the solution vector @xmath60 , respectively . the original linear system of equations is recovered by assigning @xmath61 to the upper and lower limits , i.e. , no limit . rigid body contacts are handled by the signorini - coulomb law for unilateral dry frictional contacts . this states that if the non - penetration constraint , @xmath62 , is violated then the normal contact velocity and constraint force are complementary in ensuring separation , @xmath63 , and the tangential friction force , acting to maintain zero slip , is bound to be on or inside the coulomb friction cone , @xmath64 . the latter may be linearised by approximating the cone with a polyhedral or a box . impacts are treated post facto in a separate _ impact stage _ succeeding the update of velocities and positions . at the impact stage an impulse transfer is applied enabling discontinuous velocity changes , i.e. , from @xmath65 to @xmath66 . newtons impact law , @xmath67 , is used with restitution coefficient @xmath68 in conjunction with preserving all other constraints , @xmath69 . this amounts to solving the same mlcp but with @xmath70 and @xmath71 in the right hand side of equation . a fixed timestep approach is preferred when aiming for fast full systems simulations with many nonsmooth events with inequality or complementarity conditions , e.g. , involving thousands or millions of dynamic contacts , switching in electric , hydraulic or control systems , or elastic material undergoing fracture or plastic failure . using variable timestep and exact event location become computationally intractable and may fail by occurrence of zeno points . in a multidomain simulation with strongly coupled subsystems the nonsmooth dynamics propagate instantly throughout the full system . the dynamics in the elastoplastic solid may be directly and strongly coupled with the internal dynamics of a mechatronical system . the nonsmooth multidomain dynamics approach with physics based constraint regularisation and stabilisation presented here , provides a general framework for building and efficiently solving the dynamics on mlcp form . each subsystem take the same generic form of saddle - point matrix as the full system in eq . . considering a system with two subsystems @xmath72 and @xmath73 , the full coupled system takes the following form @xmath74 \bm{h}_{a } & 0 & -\bm{g}^t_{{ab}_a}\\ 0 & \bm{h}_{b } & -\bm{g}^t_{{ab}_b}\\ \hline \bm{g}_{{ab}_a } & \bm{g}_{{ab}_b } & \bm{\sigma}_{ab } \end{bmatrix } \begin{bmatrix } \bm{z}_{a } \\ \bm{z}_b \\ \bm{\lambda}_{ab } \end{bmatrix } \label{eq : self_similar}\end{aligned}\ ] ] where @xmath75,\bm{\sigma}_{ab}$ ] are the multiplier , jacobian and regularisation of the subsystem coupling . the solid is assumed to sustain geometric large deformations . venant - kirchoff elasticity model is used in combination with the drucker - prager plasticity model @xcite . the material strain is expressed by the green - lagrange strain tensor @xmath76 where @xmath77 is the displacement field mapping a reference coordinate @xmath1 to displaced position @xmath78 . observe that the green - lagrange strain tensor transforms under large rotations and no co - rotation procedure is needed . it is convenient to use the voigt notation for representing the strain tensor on vector format , @xmath79^t$ ] , and the stress tensor @xmath80^t$ ] such that the linear constitutive law reads @xmath81 , with stiffness matrix @xmath82 where @xmath83 and @xmath84 are the first and second lam parameters which are related to the young s modulus and poisson s ratio , @xmath85 and @xmath86 , of a material as @xmath87 and @xmath88 . the corresponding strain energy density is @xmath89 plastic deformation occur when the stress of a material reaches its yield strength , @xmath90 , in which case the material undergo plastic flow if the load is increased further . the choice of yield function , @xmath91 , is based on the type of material being modeled . to represent the permanent plastic deformation , the strain tensor is decomposed into an elastic and a plastic component , @xmath92 . in ideal plasticity the plastic deformation occurs instantly according to a flow rule @xmath93 , where @xmath94 is the plastic potential and @xmath95 is the plastic multiplier . if @xmath96 the model is said to be associative and otherwise non - associative . the plastic flow last as long as it is positive , @xmath97 , incrementally reducing the stress @xmath98 , until it reaches the elastic regime , @xmath99 . this constitute a nonlinear complementarity problem known as karush - kuhn - tucker conditions @xmath100 the plastic multiplier is computed from the constitutive law , which in the plastic flow phase is @xmath101 , where the elastoplastic tangent stiffness matrix is @xmath102 predictor - corrector algorithms are conventionally used to integrate the plastic flow . in the absence of plastic hardening or softening , the plastic deformation , plastic multiplier and total stress can be computed easily using the radial return algorithm @xcite summarised in algorithm [ alg : radial_return ] . the assumed plasticity model is a capped drucker - prager model , following dolarevic @xcite , with a compaction variable @xmath103 . the yield function @xmath104 is a piecewise function consisting of the drucker - prager , tension and compression cap functions according to @xmath105 where @xmath106 is the first invariant of the stress tensor and @xmath107 is the second invariant of the deviatoric stress tensor @xmath108 . the expressions for the tension cap @xmath109 and compression cap @xmath110 are given in appendix a and the drucker - prager surface is defined as @xmath111 where @xmath112 is the internal friction angle and @xmath113 the cohesion parameter such that @xmath114 and @xmath115 . the capped yield surface is illustrated in figure [ fig : gapfunction ] . the tension cap is fix and is used to regularise the plastic flow behaviour in the corner region . the compression surface cap , on the other hand , is dynamic and the maximum hydrostatic pressure , @xmath103 , is used as main variable . the conventional drucker - prager model is a common model for the plastic deformation dynamics of soils , e.g. wet or dry sand . these materials are weak under tensile stress ( @xmath116 ) and become stronger under compression ( @xmath117 ) where it may support large shear stresses ( @xmath118 ) . the capped drucker - prager model is a generalisation that include also plastic compaction that may occur in many soils . the compaction mechanism may be failure of the individual grains whereby air or water is displaced from the soil . the compaction saturates at a maximum level , where all voids are vanished . the compaction dynamics is modeled by a variable cap on compressive side of the drucker - prager yield surface , intersecting the hydrostatic axis at @xmath119 . the compaction hardening law is chosen @xmath120 where @xmath121 is the maximum volume compaction , @xmath122 is the hardening rate and @xmath123 the initial cap position where compressive failure first occur . observe that when the plastic volume compaction @xmath124 approaches @xmath121 , the cap variable @xmath103 goes to infinity , the material does not compact further plastically and behave like the standard drucker - prager model . the elastoplastic material parameters are summarised in table [ tab : elastoplastic ] . expressions for the detailed shape of the compression cap and the derivatives of the yield functions are found in appendix a. .elastoplastic model parameters [ cols= " < , > " , ] the simulation result is compared with elasticity theory that for st . venant - kirchoff materials imply the following exact relations between applied load and uniform deformation @xmath125 where @xmath126 is the load pressure on the boundary , @xmath127 @xmath128 are the young s and bulk modulus , and @xmath129 and @xmath130 are the stretch ratios @xmath131 of the test - cube with the deformed side - length @xmath132 , initial rest length @xmath133 and cross - sectional area @xmath134 . the simulation results for @xmath135 pa and @xmath136 are presented in figure [ fig : elastic_verification ] . for large deformations ( @xmath137 % ) the simulated result agrees with the exact solution to an accuracy of @xmath138 % for uniaxial stretch and @xmath139 % for hydrostatic compression . the error decrease with size of the deformation and with finer resolution . observe that for infinitesimal deformations @xmath140 , eq . - approximates to the well - known expressions from hooke s law . increasing the stiffness or resolution further require smaller time - step for numerical stability . ) and uniaxial stretch ( @xmath141 ) with @xmath142 pa , @xmath143 and @xmath144 ( circles ) and @xmath145 ( triangles ) . the analytical solution is represented with a solid curve . , scaledwidth=45.0% ] in order to understand the nature of the error the strain field under hydrostatic compression is analysed , see figure [ fig : elastic_hydrostatic_error ] . the strain field is not uniform through the material as expected but deviate as much as @xmath146 % and most near the boundary . enforcing a uniform deformation , by initial particle displacements , produce a perfectly uniform strain and stress field . we thus conclude that the effect is caused by the way that boundary conditions are enforced . see , section [ sec : conclusions ] for a further discussion on this . through the cross - section of the elastic test cube under hydrostatic compression to @xmath147.,scaledwidth=49.0% ] the errors in the strain distribution propagate to the stress and ultimately to the plastic behaviour also . in particular the deviation from nonuniform strain will lead to a non - vanishing stress deviator in deformations where it is expected to be zero . the mean strain deviatior and its standard deviation in the case of hydrostatic compression is displayed in figure [ fig : strain_deviator ] . ^{1/2}$ ] during hydrostatic compression.,scaledwidth=49.0% ] the drucker - prager cap model is tested under hydrostatic and triaxial loading and unloading , following the tests made in @xcite with @xmath148 pa,@xmath149 , @xmath150 kpa , @xmath151 , @xmath152 , @xmath153 mm@xmath154/n . the first test examines the relation between stress and strain while undergoing plastic deformation on the compression cap , see figure [ fig : hydrostatic_plastic ] . the test produce a permanent deformation of the cube by @xmath155 compared to @xmath156 in @xcite . as can be expected this deviation is due to the deviation of uniform strain , observed in section [ sec : elastic_verification ] , that lead to non - vanishing deviatoric stress . investigating the evolution of the stress invariants , @xmath157 and @xmath158 , in the yield space , it is clear that the compression cap is not reached precisely on the hydrostatic axis and thus yield for a smaller of @xmath157 . the plastic shear is also not uniform which affect the permanent deformation . in the triaxial test , a hydrostatic pressure of @xmath159 kpa is first established . the load along one axis is then increased up to a level where the material yields and while pressure is kept fix . the material is finally unloaded . the simulated evolution of the deviatoric stress as function of strain is displayed in figure [ fig : triaxial_plastic_dp ] for a drucker - prager material and in figure [ fig : triaxial_plastic ] for a capped drucker - prager material . the simulated drucker - prager material is found to yield at a critical stress of @xmath160 kpa . this is in precise agreement with the analytical prediction from equation ( [ eq : dp ] ) with hydrostatic pressure of @xmath159 kpa . for the capped drucker - prager the yield plateau is lower and less pronounced . the difference is expected since the stress reach the compression cap surface before the drucker - prager surface . the result show a weak dependency on resolution @xmath161 and @xmath162 . the triaxial test can be used to determine the value of the plastic hardening parameter @xmath122 , whereas the maximum compaction parameter , @xmath121 , can be determined from the hydrostatic compression test . dynamic contact handling is demonstrated by dropping a beam to rest on two thin cylinders and then let if deformed by pressing a larger cylinder down on the beam . the result for an elastic and elastoplastic beam are displayed in figure [ fig : beam ] . ( green ) to @xmath163 ( red).,scaledwidth=45.0% ] in order to demonstrate multidomain capability with the method , two example systems are simulated . the first system is an articulated terrain vehicle with tracked bogies driving over a deformable terrain . images from simulation of a bogie and of a full vehicle are shown in figure [ fig : demo ] . videos from simulations are available as supplementary material at ` http://umit.cs.umu.se/elastoplastic/ ` . the vehicle weighs 4 ton and consist of roughly 200 rigid bodies including a front and rear chassis , bogie frames , wheels and track elements . these are interconnected with kinematic constraints to model chassis articulation , bogie and wheel axes , and tracks covering the wheeled bogies . the vehicle drivetrain consist of a engine , torque converter , drive shafts and differentials that distributes the engine power to front and rear part and further between left and right bogies to the wheels . the terrain is modeled as a static trimesh and a rectangular ditch with elastoplastic material . the elastoplastic material parameters are set to @xmath164 mpa , @xmath165 , @xmath166 , @xmath167 kpa , @xmath168 , @xmath169 kpa , @xmath170 mpa , @xmath171 , @xmath172 , @xmath173 kg / m@xmath174 , which represents a weak and soft forest terrain . the tracked bogie and vehicle create a rutting with permanent deformation and the rut depth can be measured . in the full vehicle simulation the solid is discretised into @xmath175 pseudo - particles . the coupled system of vehicle and terrain thus have roughly 7000 degrees of freedom and roughly the same number of constraint equations . with @xmath176 ms timestep the simulation was of the order @xmath177 slower than realtime on a conventional desktop computer and using a single core . it should be emphasised that this measure is based on prototype code with little effort on optimizing it . also the use of a direct solver for the terrain can be questioned . both the model uncertainty and spatial discretisation error are many orders in magnitude beyond the accuracy delivered by the solver . the second demonstration example is a dynamic cone penetration test where a cylindrical weight is dropped repeatedly on a cone measuring its penetration depth . this is one common way of measuring the mechanical properties of terrain . the penetration depth of the cone in simulated for two different cases . the first case is a homogeneous material and the second is material with an embedded rock , represented by a rigid body , as illustrated by figure [ fig : plastic_cone ] . the colours indicated the magnitude of displacement . the simulated cone penetration is presented in figure [ fig : plastic_cone_cm ] . the resistance is higher in the ground with an embedded rock , also before the cone actually come in contact with the rock . a meshfree elastoplastic solid model is made compatible with nonsmooth multidomain dynamics . the solid appear as a system of constrained particles in a multibody system on the same footing as articulated rigid multibodies and power transmission systems . the particles carry field variables , e.g. , the stress and strain tensor , approximated using the moving least squares method . this method provides a continuous field description throughout the solid . the dynamic interaction between the deformable solid , rigid multibodies and other geometric boundaries is modeled by unilateral contact constraints with dry friction . the full system can thus be processed using the same numerical integrator and solver framework without introducing additional coupling equations with unknown parameters and impulsive behaviour can automatically be transmitted instantly through the full and strongly coupled system . this enables fast and stable simulations of complex mechatronic systems with , or interacting with , elastoplastic materials . demonstration is made with a tracked terrain vehicle driving over deformable terrain using the capped drucker - prager model and a cone penetrometer test on terrain with and without embedded rock . the jacobian of the deformation constraint is derived and the explicit form is given in eq . ( [ eq : strainjacobiansfinal ] ) . it can be factored to a constant term that can be pre - computed and multiplied with current particle displacements . the computational bottleneck of the simulation lies in solving the block sparse mixed linear complementarity problem ( [ eq : mlcp ] ) . with dedicated hardware , parallel factorisation algorithms or using iterative solver for a more approximate solution of the terrain dynamics the performance is expected to be increased by several order in magnitude . exploring this is left for future work . the results of numerical experiments of uniform elastic deformations , presented in figure [ fig : elastic_verification ] and [ fig : elastic_hydrostatic_error ] , reveal that the solution deviate from what is expected from analytical solutions . for example , the confining pressure in hydrostatic compression of a cube discretised by @xmath178 particles is underestimated by roughly @xmath139 % when compressed up to @xmath179 % . the error decrease with reduced time - step and finer spatial discretisation . the errors are presumably due to the application of mls to the solid dynamics equations on strong form . when the problem involve traction boundary conditions , meshfree collocation methods suffer from poor accuracy and instability @xcite . the errors are not located solely to the boundaries but influence also the deformations further in the material , as seen in figure [ fig : elastic_hydrostatic_error ] , and cause the second principal deviatoric stress invariant to deviate from zero in hydrostatic pressure . this affect the plastic behaviour also . the permanent plastic deformation after loading is found to differ by the order of @xmath139 % and sometimes the material has residual stress that may cause artefacts although none have been observed . nevertheless , the developed method is applicable to many type of studies where this level of accuracy is acceptable and can not be increased further anyway because of difficulties in characterising the mechanical properties of the solid material . this deficiency can be avoided by either using a mixed strong and weak form formulation @xcite or applying correction terms of higher order on the boundary particles @xcite . future work should focus on this issue to improve the accuracy of the model . another area of improvement is to integrate the plastic yield and flow computations with the mixed complementarity problem for the constraint forces and velocity update . this will enable integration with larger time - step for strongly coupled problems than the current predictor - corrector method allow . the project was funded in part by the kempe foundation grant jck-1109 , ume university and supported by algoryx simulation . the authors are thankful fruitful discussions and ideas from prof . urban bergsten at the swedish university of agriculture , komatsu forest and olofsfors ab . this supplements section [ sec : constraint_elastoplastic ] with further details of the capped drucker - prager yield surfaces in eq . ( [ eq : dp])-([eq : caps ] ) illustrated in figure [ fig : gapfunction ] . we follow the smooth cap model of dolarevic and ibrahimbegovic @xcite with some corrections to the compression cap . the three yield functions are @xmath180 ^ 2}{r_{\textsc{\tiny c}}^{2}}+ j_2 - b\left ( \kappa \right)^2 \label{eq : c_cap}\end{aligned}\ ] ] the center , radius and intersection point of the tension cap are @xmath181 where the drucker - prager cone angle @xmath182 in the pressure - shear plane , @xmath183 is the tension cap cutoff and @xmath184^{1/2}$ ] . the compression cap intersection point is @xmath185 and @xmath186 and @xmath187 are the center and main radius of the compression cap ellipse @xcite . the stress gradient of the yield functions are @xmath188 lacoursire c. regularized , stabilized , variational methods for multibodies . in dag fritzson peter bunus and claus fhrer , editors , _ the 48th scandinavian conference on simulation and modeling ( sims 2007 ) , 30 - 31 october , 2007 , gteborg ( sr ) , sweden _ , linkping electronic conference proceedings , pages 4048 . linkping university electronic press , december 2007 . taheri s , sandu c. , taheri s. , pinto e. , gorsich d. a technical survey on terramechanics models for tire terrain interaction used in modeling and simulation of wheeled vehicles journal of terramechanics 57 , 1-22 , 2015 .
a method for simulation of elastoplastic solids in multibody systems with nonsmooth and multidomain dynamics is developed . the solid is discretised into pseudo - particles using the meshfree moving least squares method . the particles carry strain and stress tensor variables that are mapped to deformation constraints and constraint forces . the discretised solid model thus fit a unified framework for nonsmooth multidomain dynamics for realtime simulations including strong coupling of rigid multibodies with complex kinematic constraints such as articulation joints , unilateral contacts with dry friction , drivelines and hydraulics . the nonsmooth formulation allow for impulses , due to impacts for instance , to propagate instantly between the rigid multibody and the solid . plasticity is introduced through an associative perfectly plastic modified drucker - prager model . the elastic and plastic dynamics is verified for simple test systems and the capability of simulating tracked terrain vehicles driving on a deformable terrain is demonstrated .
today it is commonplace to undertake highly multiple hypothesis testing , generally in genomics and very often using tests based on student s @xmath0-statistic ; see , for example , benjamini and yekutieli ( @xcite ) , efron and tibshirani ( @xcite ) , cui and churchill ( @xcite ) , amaratunga and cabrera ( @xcite ) , page 114 , scheid and spang ( @xcite ) , shaffer ( @xcite ) , fox and dimmic ( @xcite ) , hu and willsky ( @xcite ) , qiu and yakovlev ( @xcite ) , efron ( @xcite ) , liu and hwang ( @xcite ) and van de wiel and kim ( @xcite ) . this popularity of multiple @xmath0-testing also extends to other fields ( e.g. , pawluk - koc _ et al . _ ( @xcite ) ) . the principal methods used to control the family - wise error rate and false discovery rate are founded on the assumption of independence among tests . alternative approaches are generally based either on bonferroni bounds , which are unsatisfactory for a variety of reasons ( see , e.g. , perneger ( @xcite ) ) , or on the hope that , despite ample evidence of non - independence in terms of correlation analysis , independence can be assumed in practice . the latter hope tends to be pinned either on work of benjamini and yekutieli ( @xcite ) , who argued that in some settings , the absence of independence can give conservative results , or on experience with the analysis of financial data , which suggests that in some circumstances , it might be reasonable to assume that the upper tails of the test statistics are independent , even if the joint distributions are not . upper tail independence , as it is sometimes called ( for discussion , see , e.g. , wu ( @xcite ) , falk and reiss ( @xcite ) , r. schmidt ( @xcite ) , li ( @xcite ) , r. schmidt and stadtmller ( @xcite ) , t. schmidt ( @xcite ) ) , is generally assumed to be non - asymptotic in nature . that is , tails of joint distributions are often taken to be perfectly independent beyond a certain threshold . however , this type of model is not really appropriate for the analysis of genomic data . in particular , it is difficult to determine a biological reason for , or the actual location of , a threshold . it is of greater practical interest to consider the possibility that the strength of dependence in upper tails could become successively weaker as the number of simultaneous tests , and the number of data vectors , increases . if this could be established in the context of tests based on student s @xmath0-statistic , it would lend immediate justification to the often - made assumption ( see the articles cited in the first paragraph of this paper ) that highly multiple @xmath0-statistics can be taken to be independent . the present paper will establish such a result . the mechanism for our model involves the critical points for tests becoming more extreme as the number , @xmath1 , of tests diverges ( in fact , the increase in critical points is a direct consequence of @xmath1 diverging ) so that the tests are conducted further into the tails ; furthermore , the tails of the distributions of test statistics becoming successively lighter as the number of degrees of freedom of the test statistics increases . we impose particularly weak conditions on the marginal distributions of components . in particular , the distributions need only three finite moments . with this assumption , and permitting the size , @xmath2 , of the group sample to increase a little faster than the logarithm of the number of tests , it follows from our results that the joint distributions of test statistics enjoy an asymptotic form of the upper tail independence property . this result would not be so striking if the statistics had normal distributions , but it fails for heavy - tailed distributions such as those for which not all moments are finite . of course , student s @xmath0-distribution is itself in this category , yet our results show that asymptotic independence holds in a particularly strong sense for student s @xmath0-statistic , even if it is computed from relatively heavy - tailed data . the reason this is possible is that we permit the group sample size to increase at a rate that is just sufficient to convert heavy tails to tails that are sufficiently light , to enable approximate independence at high levels . it can be seen from this property that the availability of upper - tail asymptotic independence is a bonus of working with highly multiple hypothesis testing , that is , with `` large @xmath1 and small @xmath2 '' problems . it is not available in more conventional , `` small @xmath1 and large @xmath2 '' problems , where there is a very large literature on modelling dependence in highly multiple hypothesis testing . there is a literature on comparing studentized means when the variances used for studentizing are computed from pooled data and so are common to each test statistic . however , in our experience , that approach is used less frequently , in practice , than the `` local '' standardization treated in the present paper . when using the latter method , each mean is divided by the standard deviation of the sample from which it was computed . a major motivation is that the true variances may be different in each instance . even if the variances can reasonably be assumed to be the same , it can be desirable to use the local approach since it confers greater robustness . for example , when applied to the mean alone , rather than its locally studentized form , the large - deviation properties that underpin the analysis of high - level exceedences require the data to have lighter tails . statistical literature on highly multiple hypothesis testing is outlined in helpful reviews by hochberg and tamhane ( @xcite ) , pigeot ( @xcite ) , dudoit _ et al . _ ( @xcite ) , bernhard _ et al . _ ( @xcite ) and lehmann and romano ( @xcite ) , chapter 9 . benjamini and hochberg ( @xcite ) introduced an approach , which has become very popular , to the controlling of false discovery rates ; see also simes ( @xcite ) , hommel ( @xcite ) , hochberg ( @xcite ) , sarkar and chang ( @xcite ) , sarkar ( @xcite ) , sen ( @xcite ) , hochberg and benjamini ( @xcite ) and lehmann _ et al . _ ( @xcite ) . benjamini and yekutieli ( @xcite ) specified conditions under which simultaneous , dependent hypothesis tests , conducted as though they were independent , give conservative results ; benjamini and yekutieli ( @xcite ) addressed similar issues in the context of false coverage - statement rate . sarkar ( @xcite ) extended the work of benjamini and yekutieli ( @xcite ) . efron ( @xcite ) suggested correlation corrections for large - scale simultaneous hypothesis testing . et al . _ ( @xcite ) proposed methods for controlling family - wise error rates in multiple procedures , holland and cheung ( @xcite ) discussed robustness of family - wise error rates and clarke and hall ( @xcite ) discussed robustness of testing procedures based on means . given @xmath3 , assume that for @xmath4 and @xmath5 , we observe data @xmath6 , which we use to construct @xmath0-statistics @xmath7 , where @xmath8 and @xmath9 . in practice , the statistic @xmath10 is used to test the hypothesis that the @xmath11th group has zero mean , against a one - sided alternative . when controlling the level of family - wise error rate ( fwer ) for step - down tests , we require the values of probabilities @xmath12 for @xmath13 for different levels @xmath0 and different subsets @xmath14 of @xmath15 . theorem [ thm1 ] below will enable us to compute these through approximation by the case where the @xmath10 s are all independent ; see section [ sec2.3 ] for further details . we standardize @xmath16 by dividing by @xmath2 , rather than @xmath17 , since the former is more common in nonparametric problems , but the results below are unaffected by this issue . since we studentize , there is no loss of generality in assuming that the variance of each component equals 1 . more particularly , we ask that @xmath18 where @xmath19 is a sequence of constants . the assumption that @xmath20 is made here because , in the great majority of practical applications , the hypothesis alternative to the null entails the zero level being exceeded . accordingly , the tests are one - sided , hence our preoccupation with exceedences of a level . however , minor modifications of our arguments permit the two - sided case to be treated . further , we assume that for an integer @xmath21 , @xmath22 where @xmath23 . the third moment condition in ( [ eq2.1 ] ) permits the variables @xmath6 to have relatively heavy - tailed distributions , for example , a pareto distribution with tail exponent greater than 3 . the assumption of short - range correlation in ( [ eq2.2 ] ) is , of course , an oversimplification , but it reflects the low level of correlation that is often observed in practice . for example , messer and arndt ( @xcite ) argue that correlation decays from about 0.08 , at a separation of approximately two base pairs , to about @xmath24 for a separation of ten base pairs . results reported by mansilla _ et al . _ ( @xcite ) corroborate these figures if we assume that their data are normally distributed . more generally , almirantis and provata ( @xcite ) give evidence of both short - range and long - range correlation , depending on the nature of the dna or rna under investigation . the relationship between the group size , @xmath2 , and the number of hypothesis tests , @xmath1 , is assumed to satisfy @xmath25 this allows the group size to be very much smaller than the number of tests . in the absence of more detailed assumptions about the distributions of the @xmath6 s , ( [ eq2.3 ] ) is necessary for the theorem we shall give below . to appreciate why , note that if the @xmath6 s are independent and identically distributed with an atom at zero and , in particular , if @xmath26 , then , with probability at least @xmath27 , the @xmath0-statistic @xmath28 assumes the indeterminate value @xmath29 . in such cases , we shall take @xmath30 , but in order for the theorem to have a meaningful interpretation when @xmath0 is the @xmath31-level quantile of the standard normal distribution , it is essential that the probability that @xmath32 be of smaller order than @xmath33 . therefore , we require @xmath34 for all @xmath35 and this assumption is equivalent to ( [ eq2.3 ] ) . define @xmath36 and @xmath37 . condition ( [ eq2.2 ] ) implies that @xmath38 and , of course , @xmath39 . given @xmath40 , let @xmath41 satisfy @xmath42 where @xmath19 are as in ( [ eq2.1 ] ) . if @xmath0 satisfies the first part of ( [ eq2.5 ] ) , then any function @xmath43 which satisfies , as @xmath44 , @xmath45 converges to zero as @xmath46 . in the arguments in section [ sec3 ] , we shall use this notation generically ; while @xmath43 will satisfy ( [ eq2.6 ] ) , it will alter from one appearance to another . strictly speaking , it is not essential to take @xmath1 to diverge . although that condition motivates the assumption of divergent @xmath0 and is , in turn , motivated by the contemporary high - dimensional problems that led to this work , it is not necessary for the theorem below . [ thm1 ] if @xmath47@xmath48 hold , then there exists a probability space on which are defined random variables @xmath49 and @xmath50 such that the joint distribution of @xmath51 is identical to that of @xmath52 ; the random variables @xmath50 are independent and distributed , respectively , as @xmath52 ; and with probability equal to @xmath53 , the exceedences of @xmath0 by @xmath54 occur at the same indices and take the same values as the exceedences of @xmath0 by @xmath50 . to interpret the theorem , note that we would normally expect the dependent data set @xmath52 to exhibit clusters of level exceedences , rather than the single , isolated exceedences associated with the independent sequence @xmath50 . the fact that the @xmath10 process ( or , equivalently , the @xmath55 process ) behaves like the @xmath56 process in the case of large exceedences reflects the fact that , since the marginal distribution of a @xmath0-statistic is relatively light - tailed ( if @xmath2 is sufficiently large see ( [ eq2.3 ] ) ) , exceedences of a high level are rare and so are unlikely to occur together . the case of low - level exceedences is a very different matter , of course , and so we would expect the theorem to fail if the lower bound for @xmath0 , in the first part of ( [ eq2.5 ] ) , were relaxed too far . in this section , we treat the case of the null hypothesis , where @xmath57 for each @xmath11 . this would be assumed in most applications of theorem [ thm1 ] since it represents the setting that is conventionally used for calibration . the theorem implies that , in a strong sense , exceedences down to those of the level @xmath58 are identical to the ones that would occur in the case of independent tests . now , the probability associated with an exceedence of @xmath59 is , for small @xmath60 , approximately @xmath61 . therefore , false discoveries at probability levels of approximately @xmath61 , and at lower levels , can be adequately controlled by assuming that the tests are independent , even when they are not . note that @xmath62 and that the false - discovery level controlled by the conventional family - wise error rate is only @xmath33 . next , we discuss the sorts of calculations that are enabled by theorem [ thm1 ] . let @xmath63 denote the number of indices @xmath64 $ ] for which @xmath10 lies in the interval @xmath65 $ ] , where @xmath66 and @xmath67 is determined by @xmath68 , with @xmath69 held fixed . ( we take @xmath70 . ) if @xmath52 were fixed , then the joint distribution of the @xmath71 random variables @xmath72 would be exactly multinomial with parameters @xmath1 and @xmath73 , where @xmath74)$ ] . theorem [ thm1 ] implies that for the dependent process @xmath52 , and for any @xmath75 for which @xmath76 satisfies ( [ eq2.5 ] ) , any simultaneous probability calculation based on the multinomial result , but applied to the actual @xmath77 process rather than an idealized process with independent marginals , is valid , provided that @xmath78 and the final computed probability is quantified by adding an error which is stated to be of order @xmath79 . the latter probability converges to zero , even if @xmath80 is taken as large as @xmath81 , where @xmath82 . from this point , simultaneous multinomial probability calculations based on @xmath83 , familiar from the well - understood case of independent test statistics , can be used to construct rules for controlling fwer or false discovery rate ( fdr ) ; see , for example , benjamini and hochberg ( @xcite ) . wang and hall ( @xcite ) have shown that , under the assumption of finite third moments , highly accurate approximations are available for the marginal distribution of @xmath28 . such calculations , which justify standard normal , student s @xmath0- or bootstrap approximations to the marginal distribution of @xmath28 , are already widely used in practice ( see section [ sec1 ] ) , in conjunction with the independence assumption , when controlling false discovery rates . our paper provides justification for these methods . more generally , theorem [ thm1 ] implies that if a probability statement about what the process @xmath52 does above the level @xmath0 is founded on the assumption of independence , then , no matter how complex or convoluted the statement might be , the claimed probability level is accurate to within @xmath84 . to give an example of calculations based on theorem [ thm1 ] , take @xmath85 , @xmath86 and @xmath87 , the latter denoting the upper @xmath88-level quantile of student s @xmath0-distribution with @xmath89 degrees of freedom . reflecting empirical evidence given in section [ sec2.1 ] , take @xmath90 . then , ( [ eq2.5 ] ) is in order ; the probability that at least one value of @xmath1 independent @xmath0 statistics , each on 99 degrees of freedom , exceeds @xmath87 equals @xmath91 , @xmath92 and @xmath93 for @xmath94 , @xmath95 and @xmath96 , respectively ; and ( [ eq2.6 ] ) suggests that the errors in these levels are in error by less than @xmath97 , @xmath98 and @xmath99 , respectively . most likely , the errors are much less than these since the asymptotic bound is derived only as an upper bound . if we were to make a general probability statement about exceedences of the level @xmath100 by the stochastic process of @xmath0 statistics , under the assumption of independence , then , despite the process actually being @xmath101-dependent rather than independent , we would expect to make errors no greater than these respective values . in the same general setting , relative error decreases to zero as @xmath1 and @xmath0 increase . for example , in cases where @xmath0 solves @xmath102 , with @xmath103 denoting the standard normal distribution function , we have @xmath104\to 0 $ ] as @xmath44 . theorem [ thm1 ] can be extended to other settings , in particular , to those where ( a ) a wider range of dependence , obtained by allowing @xmath101 in ( [ eq2.2 ] ) to diverge with @xmath1 , is allowed ; ( b ) the value of @xmath2 for the @xmath11th group equals @xmath105 , depending on @xmath11 , and ( [ eq2.3 ] ) is altered by requiring that @xmath106 ; ( c ) weights @xmath107 are incorporated into the construction of the @xmath0-statistics @xmath10 , by defining @xmath108 , @xmath109 and , as before , @xmath110 . provided the weights satisfy @xmath111 where @xmath112 are positive constants not depending on @xmath1 , the proof in this more general case is as in section [ sec3 ] . however , the statement of the theorem is then less elegant and less transparent , so we do not give the more general version here . incorporation of the weights @xmath107 permits the scope of the example above to be extended to hypothesis - testing problems involving linear regression.=1 to indicate the types of results that can be achieved under longer ranges of dependence , we shall discuss the case of a moving average , @xmath113 where @xmath114 is permitted to diverge to infinity at a rate not exceeding @xmath115 and the independent disturbances @xmath116 are all distributed as @xmath117 , for which @xmath118 and @xmath119 . in this setting , ( [ eq2.2 ] ) holds . we strengthen ( [ eq2.3 ] ) by asking that @xmath120 . the definition of @xmath0 implicit in ( [ eq2.5 ] ) can now be refined to @xmath121 where @xmath122 denotes a sufficiently large absolute constant . the conclusions of theorem [ thm1 ] continue to hold , with a similar proof if we replace @xmath53 by @xmath123 . the notation @xmath124 will denote constants not depending on @xmath2 or @xmath1 . let @xmath125 , @xmath126 and note that @xmath127 also , note that @xmath128 , where @xmath129 , @xmath130 , are independent and identically distributed random variables satisfying @xmath131 cf . ( [ eq2.1 ] ) . using results of wang and hall ( @xcite ) ( see also wang ( @xcite ) ) , it can be shown that , for constants @xmath132 , and whenever @xmath133 , @xmath134 we also wish to prove the following related result for pairs of exceedences . assume the conditions of theorem [ thm1 ] . there then exist @xmath135 such that for all @xmath136 with @xmath137 , and for all @xmath138 , we have @xmath139\\[-8pt ] & & \quad \leq5 \exp\bigl\{-\tfrac{1}{2 } ( 1+\alpha ) s^2 + d_5 n^{-1/2 } s^3 + 2 \sqrt n ( d_{i_1}+d_{i_2 } ) s\bigr\},\nonumber\end{aligned}\ ] ] where @xmath140 is as in @xmath141 . to establish the lemma , we write @xmath142 u_{i_21}^{(1)}&=&v_{i_21 } i(|v_{i_11}|\leq n^{1/2}/s , u_{i_21}^{(2)}=v_{i_21}-u_{i_11}^{(1 ) } .\end{aligned}\ ] ] by virtue of ( [ eq3.2a ] ) , simple calculations show that @xmath143 e\bigl\{\bigl(u_{i_11}^{(1)}\bigr)^2+\bigl(u_{i_21}^{(1)}\bigr)^2\bigr\ } & = & 2+\mathrm{o}(1 ) s/\sqrt n , \\[3pt ] e\bigl(u_{i_11}^{(1)}+u_{i_21}^{(1)}\bigr)^2 & = & e\bigl\{(v_{i_11}+ v_{i_21})-\bigl(u_{i_11}^{(2)}+u_{i_21}^{(2)}\bigr)\bigr\}^2 \\[3pt ] & \leq & 2 ( 1+\rho_{ij})+d_6 s/\sqrt n , \\[3pt ] e\bigl|u_{i_11}^{(1)}+u_{i_21}^{(1)}\bigr|^3&\leq & d_6 .\end{aligned}\ ] ] these results , and the bound @xmath144 ( valid for all real @xmath145 ) , imply that , with @xmath146 , @xmath147\nonumber \\[-7pt]\\[-7pt ] & & \quad \leq1+(\rho_{ij}-1 ) { s^2\over4n}+ d_7 { s^3\over n^{3/2 } } , \nonumber \\[2pt ] \label{eq3.5 } & & e\biggl[\exp\biggl\{\frac{1}{2 } h ( v_{i_11}+v_{i_21 } ) -\frac{1}{4 } h^2 ( v_{i_11}^2+v_{i_21}^2)\biggr\ } i(|v_{i_11}|\geq n^{1/2}/s , \mbox { or } |v_{i_21}|\geq n^{1/2}/s)\biggr]\nonumber \\[2pt ] & & \quad\leq \exp \{p(|v_{i_11}|>n^{1/2}/s ) + p(|v_{i_21}|>n^{1/2}/s)\ } \\[2pt ] & & \quad\leq d_8 \bigl(s/\sqrt n\bigr)^3 .\nonumber\end{aligned}\ ] ] results ( [ eq3.4 ] ) and ( [ eq3.5 ] ) , together with the independence of @xmath148 for each @xmath149 , imply that , for @xmath150 , with @xmath151 sufficiently small , @xmath152\nonumber \\[-8pt]\\[-8pt ] & & \quad \leq\biggl\{1+(\rho_{ij}-1 ) { s^2\over4n } + d_{10 } { s^2\over n^{3/2}}\biggr\}^{\!n } \leq\exp\biggl\{(\rho_{ij}-1 ) { s^2\over4}+d_{11 } { s^3\over\sqrt n}\biggr\ } .\nonumber\end{aligned}\ ] ] define @xmath153 . it follows from ( [ eq3.6 ] ) that whenever @xmath154 , with @xmath151 sufficiently small , we have @xmath155 & & { } \quad\ + 2 \sqrt n ( d_{i_1}+d_{i_2 } ) s \geq2 s^2 ( 1-\varepsilon ^2)\biggr\}\nonumber \\[2pt ] & \leq&\exp\biggl\{2 \sqrt n ( d_{i_1}+d_{i_2 } ) s-\frac{1}{2}s^2 ( 1-\varepsilon^2)\biggr\ } \\[2pt ] & & { } \times e\biggl[\exp\biggl\{\frac{1}{2 } h \sum_k(v_{i_1k}+v_{i_2k } ) -\frac{1}{4 } h^4 \sum_k(v_{i_1k}^2+v_{i_2k}^2)\biggr\}\nonumber \biggr ] \\[2pt ] & \leq & \exp\biggl\{-\frac{1}{2 } ( 1+\alpha ) s^2 + d_{12 } { s^3\over \sqrt n}+2 \sqrt n ( d_{i_1}+d_{i_2 } ) s\biggr\ } , \nonumber\end{aligned}\ ] ] where @xmath156 is as defined in ( [ eq2.4 ] ) . write @xmath157 and note that if @xmath158 , then @xmath159 & & \quad \subseteq\biggl\{2 h \sum_kv_{i_2k}-h^2 \sum_k v_{i_2k}^2 + 2 \sqrt n d_{i_2 } s \geq s^2 ( 1-\varepsilon ^2)\biggr\},\end{aligned}\ ] ] where @xmath160 . it can be shown that @xmath161 & & \quad = p\biggl\{\sum_k v_{i_1k}+n d_{i_1 } \geq s ( n q_{i_1}^2)^{1/2 } , \sum_k v_{i_2k}+n d_{i_2}\geq s ( n q_{i_2}^2)^{1/2}\biggr\ } \\[2pt ] & & \quad \leq\pi_{1n}+\pi_{2n}+\pi_{3n}+\pi_{4n}+\pi_{5n } , \nonumber\end{aligned}\ ] ] where @xmath162 property ( [ eq3.3 ] ) will follow from ( [ eq3.7 ] ) and ( [ eq3.8 ] ) if we prove that there exists @xmath163 such that , for @xmath164 , @xmath165 for @xmath166 . our proof of ( [ eq3.9 ] ) is based on arguments of shao ( @xcite ) ( see also the proof of proposition 4.2 of wang and hall ( @xcite ) ) and uses the following result : if @xmath167 , @xmath168 and @xmath169 , then for any @xmath170 , @xmath171 and @xmath172 , @xmath173 = 1+(\lambda ^2-\theta ) n^{-1 } x^2 + a(\lambda , \theta ) n^{-3/2 } x^3 e|x|^3 , \end{aligned}\ ] ] where @xmath174 and @xmath175 depends only on @xmath176 and @xmath177 . this result is a special case of lemma 1 of shao ( @xcite ) . also , note that @xmath178 where , noting that @xmath179 , we define @xmath180 if the random variable @xmath181 has the @xmath182 distribution and if @xmath183 , then @xmath184 and so @xmath185 for @xmath186 , with @xmath187 sufficiently small . arguments similar to those in the proof of ( [ eq3.7 ] ) yield that @xmath188 for @xmath186 with @xmath187 sufficiently small . to estimate @xmath189 , we write @xmath190 . it follows from ( [ eq3.10 ] ) with @xmath191 , @xmath192 and @xmath193 that , with @xmath194 , @xmath195 \\ & \leq & \exp\biggl\{\biggl(\frac{1}{2}-\frac{1}{6}\biggr ) s^2 -s^2 ( 1+\varepsilon ) + \frac{1}{6 } s^2 ( 1+\varepsilon ) ^2+\sqrt n d_i s+d_{15 } s^3 n^{-1/2 } \biggr\ } \\ & \leq & \exp\biggl\{-\frac{1}{2 } s^2-(5 \varepsilon s^2 /8 ) + \sqrt n d_i s+\bigl(d_{15 } s^3/\sqrt n\bigr)\biggr\ } \\ & \leq & \exp\biggl\{-\frac{1}{2 } ( 1+\alpha ) s^2+\sqrt n d_i s + \bigl(d_{15 } s^3/\sqrt n\bigr)\biggr\ } , \end{aligned}\ ] ] where we have used the fact that the function @xmath196 is increasing in @xmath197 . combining all of the above estimates , we obtain @xmath198 similarly , we may prove ( [ eq3.9 ] ) for @xmath199 . put @xmath200 . it follows from ( [ eq3.10 ] ) with @xmath191 , @xmath201 and @xmath193 that , with @xmath160 , @xmath202 \\ & \leq & \exp\bigl\{-1.5 s^2-s^2 ( 1-\varepsilon ) + 2 s^2(1-\varepsilon ) ^2+\sqrt n d_i s+\bigl(d_{16 } s^3/\sqrt n\bigr)\bigr\ } \\ & \leq & \exp\biggl\{-\frac{1}{2 } s^2 - 2 \varepsilon s^2+\sqrt n d_i s + \bigl(d_{16 } s^3/\sqrt n\bigr)\biggr\ } \\ & \leq & \exp\biggl\{-\frac{1}{2 } ( 1+\alpha ) s^2+\sqrt n d_i s + \bigl(d_{16 } s^3/\sqrt n\bigr)\biggr\ } .\end{aligned}\ ] ] similarly , we may prove ( [ eq3.9 ] ) for @xmath203 . this completes the derivation of ( [ eq3.9 ] ) and , hence , also the proof of the lemma . partition the set of positive integers into small blocks , each of length @xmath204 , where @xmath101 is as in ( [ eq2.2 ] ) , and large blocks , each of length @xmath205 , where @xmath205 is a divergent function of @xmath1 . we shall take @xmath206 where @xmath207 as @xmath1 increases . the integers in each block are consecutive , each consecutive pair of large blocks is separated by a small block and the block furthest to the left is a large block . let the small blocks be @xmath208 and the large blocks be @xmath209 indexed such that the order of the blocks is @xmath210 let @xmath211 denote the first large block and let @xmath212 be the number of indices @xmath213 for which @xmath214 . we wish to develop a bound for @xmath215 . identical bounds can be derived , uniformly in the block indices , for the versions of @xmath212 in the case of blocks @xmath216 for notational simplicity , we focus solely on @xmath217 . by hlder s inequality , @xmath218 where @xmath219 satisfy @xmath220 . define @xmath221 . in view of ( [ eq3.2 ] ) and ( [ eq3.3 ] ) , @xmath222 .\nonumber\end{aligned}\ ] ] noting that @xmath212 can be written as @xmath204 sums of @xmath223 independent and identically distributed random variables and using calculations based on the binomial distribution , it can be shown that , for the choice of @xmath205 at ( [ eq3.11 ] ) , @xmath224 is bounded as @xmath225 for each @xmath226 . hence , using ( [ eq3.12 ] ) and ( [ eq3.13 ] ) , we deduce that for each @xmath227 , @xmath228\\[-8pt ] & & { } \times \bigl[\ell \exp\bigl\{-\tfrac{1}{2 } ( \alpha + 1 ) s^2\bigr\}+\ell ^2 \exp ( -s^2 ) \bigr]^{1-\eta_2 } .\nonumber\end{aligned}\ ] ] write @xmath229 for the number of exceedences of @xmath230 that occur in the union of the small blocks @xmath231 that intersect the interval @xmath232 $ ] . there are @xmath233 such small blocks and each is of length @xmath204 , so , by ( [ eq3.2 ] ) , @xmath234 provided we choose @xmath235 to diverge to infinity in such a manner that @xmath236 it follows from ( [ eq2.3 ] ) that @xmath237 and so ( [ eq3.14 ] ) entails that @xmath238^{1-\eta_2 } .\end{aligned}\ ] ] since this is true for each @xmath239 , we have @xmath240 .\end{aligned}\ ] ] let @xmath242 denote the number of exceedences of @xmath230 which come from large blocks @xmath243 , @xmath244 , that have two or more exceedences . write @xmath245 for the sum over @xmath244 of the probability @xmath246 that @xmath214 for some @xmath247 . then ( a ) the expected number of exceedences of @xmath230 by @xmath248 equals @xmath249 and is less than or equal to @xmath250 ; ( b ) the expected number of exceedences in ( a ) is greater than or equal to @xmath251 ; and ( c ) since @xmath252 and @xmath253 satisfies ( [ eq3.2 ] ) , we have @xmath254 and an identical bound holds for @xmath255 , in particular , ( d ) @xmath256 . results ( a)(d ) imply that @xmath257 since @xmath258 , @xmath259 and bounds for @xmath215 , @xmath260 and @xmath241 are given by ( [ eq3.17 ] ) , ( [ eq3.15 ] ) and ( [ eq3.18 ] ) , it follows that ( [ eq3.19 ] ) entails , on taking @xmath205 as in ( [ eq3.11 ] ) , @xmath261 let @xmath262 denote the event that ( a ) there are no exceedences of @xmath230 in any of the small blocks that are wholly contained within @xmath232 $ ] ; ( b ) in each of the large blocks that is wholly contained within @xmath232 $ ] , there is at most one exceedence of @xmath230 ; and ( c ) there are no exceedences of @xmath230 in any block fragment that overlaps the end point @xmath1 . write @xmath263 for the complement of @xmath262 . results ( [ eq3.15 ] ) , ( [ eq3.17 ] ) and ( [ eq3.18 ] ) imply that , with @xmath205 given by ( [ eq3.11 ] ) and assuming that ( [ eq3.16 ] ) holds , @xmath264 therefore , in order for @xmath265 , it is sufficient that for some @xmath266 and all sufficiently large @xmath1 , we have @xmath267 where @xmath268 is as defined in section [ sec2 ] . this choice of @xmath230 satisfies ( [ eq3.16 ] ) and so if @xmath230 is given by ( [ eq3.22 ] ) , then @xmath269 satisfies ( [ eq3.21 ] ) . let @xmath270 , @xmath244 , be the number of times that @xmath214 for @xmath247 . then , the number , @xmath271 , say , of blocks @xmath243 for which @xmath272 is distributed as @xmath273 , where the random variables @xmath274 are independent , @xmath275 if @xmath272 and @xmath276 otherwise . as before , we define @xmath277 . conditional on @xmath271 and on the events `` @xmath278 '' and `` @xmath279 , '' where @xmath280 , the sequences @xmath281 and @xmath282 are independent . order the blocks @xmath243 for which @xmath272 , giving @xmath283 , where @xmath284 , and let @xmath285 denote a value of @xmath286 for which @xmath214 , randomly chosen among such values for which @xmath287 . write @xmath288 for the index of the value of @xmath286 that is chosen as @xmath285 . then , conditional on @xmath271 , the random variables @xmath289 are independent and identically distributed as @xmath290 , @xmath291 is a set of integers chosen independently and randomly from @xmath292 and @xmath293 is uniformly distributed among indices in @xmath294 . let @xmath295 be independent random variables having the distributions of @xmath296 , respectively , let @xmath297 denote the number of times that @xmath298 exceeds @xmath230 for @xmath247 and put @xmath299 . the numbers @xmath300 of blocks @xmath243 for which @xmath301 are distributed as @xmath302 , where the random variables @xmath303 are independent and @xmath304 if @xmath305 , @xmath306 otherwise . an argument similar to , but simpler than , that leading to ( [ eq3.20 ] ) shows that @xmath307 by enlarging the probability space if necessary , we can think of @xmath271 as denoting the number out of @xmath308 independent and random variables @xmath309 , each uniformly distributed on @xmath310 $ ] , which lie in the respective intervals @xmath311 $ ] . take @xmath300 to be the number of @xmath312 s that lie in @xmath313 $ ] . then , @xmath314 we have already constructed sequences @xmath289 , @xmath315 and @xmath291 . if @xmath316 , then , conditional on these quantities and on @xmath271 and @xmath300 , we select new values @xmath317 , @xmath318 and @xmath319 which are independent of @xmath289 , @xmath320 and @xmath291 , with @xmath321 independently distributed as @xmath290 , the values of @xmath319 independently and uniformly distributed among @xmath322 and the values of @xmath318 uniformly distributed within the blocks @xmath323 , respectively . in this instance , we take @xmath324 and @xmath325 to be identical to @xmath326 and @xmath327 , respectively . if @xmath328 , then we take @xmath329 to be the ( exceedence , block index ) pairs that remain after randomly and independently deleting @xmath330 pairs from the sequence @xmath331 . let @xmath332 denote the number of exceedences of @xmath230 by @xmath295 and let @xmath300 represent the number of large blocks @xmath243 in which there is at least one exceedence of @xmath230 by the sequence @xmath295 . then , @xmath333 . conditional on @xmath332 and @xmath300 , let @xmath334 denote independent and identically random variables , all distributed as @xmath290 , and distribute the locations @xmath335 of these exceedences independently and uniformly over the points @xmath336 , conditional on all of the variables @xmath300 , @xmath332 , @xmath324 and @xmath337 . take the values of @xmath295 that exceed @xmath230 to be the variables @xmath338 and let the locations of those exceedences be the points @xmath339 . by construction , @xmath340 are distributed as the exceedences of @xmath230 by @xmath1 independent and identically distributed random variables distributed as @xmath290 ; conjointly , @xmath339 are distributed as the locations of those exceedences and the probability that @xmath341 , @xmath342 for each @xmath343 $ ] and there are no exceedences of @xmath230 in any of the small blocks @xmath231 for any @xmath343 $ ] is bounded below by @xmath344 , where @xmath345 satisfies ( [ eq2.6 ] ) ; see also ( [ eq3.20 ] ) , ( [ eq3.21 ] ) , ( [ eq3.23 ] ) and ( [ eq3.24 ] ) . hence , provided that @xmath230 satisfies ( [ eq3.22 ] ) , we may construct a sequence @xmath295 of independent variables with the same marginal distribution as @xmath346 and such that , with probability bounded below by @xmath344 , the exceedences of @xmath248 over @xmath230 are identical to those of @xmath295 . the theorem follows from this property , ( [ eq2.3 ] ) and ( [ eq3.1 ] ) , on taking @xmath347 . this paper has benefited from helpful comments by abba krieger , to whom we are grateful . research of both authors was partially supported by an australian research council grant . hu , z. and willsky , g.r . utilization of two sample @xmath348-test statistics from redundant probe sets to evaluate different probe set algorithms in genechip studies . _ bmc bioinformatics _ * 7 * 12 ( electronic ) . pawluk - kolc , m. , zieba - palus , j. and parczewski , a. ( 2006 ) . application of false discovery rate procedure to pairwise comparisons of refractive index of glass fragments . _ forensic sci . internat . _ * 160 * 5358 .
particularly in genomics , but also in other fields , it has become commonplace to undertake highly multiple student s @xmath0-tests based on relatively small sample sizes . the literature on this topic is continually expanding , but the main approaches used to control the family - wise error rate and false discovery rate are still based on the assumption that the tests are independent . the independence condition is known to be false at the level of the joint distributions of the test statistics , but that does not necessarily mean , for the small significance levels involved in highly multiple hypothesis testing , that the assumption leads to major errors . in this paper , we give conditions under which the assumption of independence is valid . specifically , we derive a strong approximation that closely links the level exceedences of a dependent `` studentized process '' to those of a process of independent random variables . via this connection , it can be seen that in high - dimensional , low sample - size cases , provided the sample size diverges faster than the logarithm of the number of tests , the assumption of independent @xmath0-tests is often justified .
coherent population trapping was discovered in the interaction of sodium atoms with a laser field in 1976 . @xcite . due to this effect a substantial part of population , because of destructive quantum interference between different excitation pathways , is trapped in a coherent superposition of ground state sublevels dark states . with a coherent population trapping are associated dark resonances when due to this effect absorption and as a result fluorescence from atoms decreases , but intensity of the transmitted light increases when part of the atomic population is trapped in dark states . if in addition to the optical excitation an external magnetic field is applied , it can destroy coherence between ground state sublevels and return trapped population into absorbing states and , as a result , increase absorption and fluorescence , but decrease a transmitted light . a review of applications of dark resonances was published some years ago by arimondo @xcite . coherence in an atomic ground state attracted substantial attention in connection with lasing without inversion @xcite , magnetometry @xcite and laser cooling @xcite . as a result dark resonances recently are studied in detail , including open systems @xcite and systems with losses @xcite . in course of these studies a new and unexpected phenomenon was observed by authors of @xcite . in this study d@xmath0 line of @xmath1rb atoms was excited by a diode laser . radiation was tuned to the absorption from optically resolved ground state hyperfine level @xmath2 originating from atomic @xmath3s@xmath4 state . the final state of the transition was rb @xmath3p@xmath5 excited state . hyperfine components of this level was not resolved and all allowed in a dipole transition excited state hyperfine levels with quantum numbers @xmath6 were excited . in contrary to the usual dark resonance signal when in the absence of the magnetic field one can observe increased transmittance and decreased fluorescence intensity authors observed opposite effect decreased transmittance which increases with magnetic field applied and an increased fluorescence intensity which decreased when field was applied . authors of the paper @xcite write that the physical reason of this effect _ remains still unclear_. they suppose that one of the reasons for the peak in the fluorescence could be the non - coupled states on the zeeman sublevels of the excited hyperfine levels . these non - coupled states , as it is supposed in @xcite , inhibit the stimulated emission induced by the laser field . the decrease of stimulated emission leads to an increase of the fluorescence . in this letter we offer , in our opinion , very simple and strait forward explanation of the origin of these reversed dark resonances and perform experimental and numerical studies of them . in a simple qualitative explanation traditional dark resonances can be connected with a well known optical pumping phenomenon . let us assume that we excite atomic transition @xmath7 with a linearly polarized light . direction of the @xmath8 axis is chosen along the light electric field vector @xmath9 . as a result @xmath10 absorption takes place and transitions occur between ground and excited state magnetic sublevels with @xmath11 @xmath12 , where @xmath13 are magnetic quantum numbers of the ground and excited states respectively , see fig . [ fig1 ] . according to this scheme absorption does not take place from ground state magnetic sublevels with quantum number @xmath14 , because for these states there are no corresponding excited state magnetic sublevel with the same magnetic quantum number value . in the spontaneous decay dipole transitions from optically populated excited state magnetic sublevels @xmath15 to the nonabsorptive ground state sublevels @xmath14 are allowed . as a result , if relaxation in the ground state is slow , in a steady state conditions substantial part of the population will be optically pumped to the ground state sublevels with quantum numbers @xmath14 and will be trapped there . as a result traditional decrease of the absorption and fluorescence and increase of the transmittance will be observed , because the population of absorbing ground sate magnetic sublevels will be reduced . if we now apply an external magnetic field in a direction perpendicular to the @xmath8 axis , field will mix ground state sublevels effectively and will return trapped population into the states from which absorption takes place . as a result absorption and fluorescence will increase . this is a qualitative explanation of the usual dark resonance . a similar reasoning can be exploited to explain the reversed resonance observed in @xcite and in this paper . let us assume that we excite with @xmath16 radiation atomic transition @xmath17 . in this case there are no ground state sublevels not involved in the absorption that can trap atomic population . the actual relative transition rates proportional to the squared respective clebsch gordan coefficients in this system of sublevels are shown in fig . as one can see ground state magnetic sublevel @xmath18 is the most absorbing - with highest relative absorption rate . at the same time just to this sublevel intensively with high rates decay all three excited state magnetic sublevels populated by the light . one can expect that in a conditions of a steady state excitation , as a result of interplay of absorption and decay rates , population of the intensively absorbing ground state magnetic sublevel @xmath19 will be increased and , as a result , one can expect increased absorption and fluorescence from this atom and decreased transmittance of the resonant laser light . if an external magnetic field is applied perpendicularly to @xmath8 axis it will mix ground state magnetic sublevels and redistribute population between the ground state magnetic sublevels . as a result population of intensively absorbing magnetic sublevel @xmath18 will be decreased . at the same time population of less absorbing magnetic sublevels @xmath20 will be increased this means that the total absorption and fluorescence will be decreased and transmittance will be increased . this means that reversed dark resonance will be observed . to prove this qualitative consideration let us solve balance equations for the magnetic sublevel stationary population @xmath21 in the scheme shown in fig . [ fig2 ] . in a steady state conditions for the ground state magnetic sublevels we will obtain @xmath22 where @xmath23 is excited state relaxation rate , @xmath24 absorption rate , and @xmath25 ground state magnetic sublevel population in absence of the radiation . in a condition when absorption is slow @xmath26 weak absorption , we have @xmath27 if we now keep in mind absorption rates from different magnetic sublevels of the ground state , see fig . [ fig2 ] , and calculate the overall absorption from such state and compare it with the absorption from the equally populated magnetic sublevels ( when magnetic field is applied ) than we see an increase in the absorption rate by a factor @xmath28 or by approximately @xmath29 . the same calculation can be performed for the transitions @xmath30 and @xmath31 . for these schemes in a similar way we will obtain even larger increase of the absorption due to this specific optical pumping . the increase will be by a factor @xmath32 and @xmath33 respectively . this means that the described effect increases with increase of the quantum numbers of involved levels . of course presented description is only qualitative , but in our opinion gives a good idea what is happening when reversed dark resonances are observed . to have a quantitative description of the phenomenon one must solve equations for the density matrix for an open system with losses . we will not do this in present paper . instead a simple analysis will be carried out . an analysis of the probabilities of optical transitions originating from @xmath34 between excited hyperfine levels of the rb atom show that levels @xmath35 are populated in the ratio ( @xmath36 . it means that a hyperfine transition leading to the reversed dark resonances discussed above is most strongly excited . for this scheme let us calculate a signal shape using a full density matrix approach . we solved a rate equations for a density matrix , see @xcite , chapter 5 , for @xmath37 transition . a broad line approximation was used . it means that we assume that in a magnetic field all magnetic sublevels are in equally good resonance with radiation . secondly , we assumed that at a magnetic field strength used in the experiment ( resonance width is less than 100 mg ) hyperfine levels experience linear zeeman effect . no substantial hyperfine level mixing at applied field strength takes place . direct rb atom magnetic sublevel splitting in a magnetic field calculations and measurements prove that these assumption are valid , see for example @xcite . for signal simulation the following rate constants were used . excited state relaxation rate @xmath38 s@xmath39 @xcite , absorption rate @xmath40 s@xmath39 , ground state relaxation rate @xmath41 s@xmath39 ( mainly due to collisions with the walls of the cell and fly - through the excitation laser beam ) . lande factors @xmath42 , @xmath43 were calculated in a standard way from the atomic and nuclear data available in @xcite . we suppose that a magnetic field is applied along @xmath8 axis . laser light excites @xmath31 transition and is linearly polarized along @xmath44 axis . intensity of the fluorescence with the same polarization is calculated and the intensity of the transmitted beam is also calculated as a function of the magnetic field . the results are presented in a fig . they demonstrate well pronounced reverse resonances and are in a very good qualitative agreement with the measurements obtained in @xcite , see fig . 5 . there . width of these resonances can be varied by changing ground state lande factor value , ground state relaxation rate and absorption rate . at week absorption @xmath45 the dark resonance width is determined by a condition when ground state larmor frequency is equal to the ground state relaxation rate . for @xmath2 state of @xmath1rb atom at low concentration it can be as narrow as @xmath46 mg . this is a width that was actually observed in @xcite . obtained signals in some sense are the same as the ground state hanle effect measured in atoms as well as in molecules in great extent , see . for example @xcite . in case of molecules also reversed structure in ground state hanle effect was observed . in case of molecules , when optical pumping takes place in an open cycle and a total ground state population is substantially reduced , this structure can be attributed to a high order coherence created between ground state magnetic sublevels @xcite we performed measurements of these reversed resonances also in our laboratory . in our experiment we use isotopically enriched rubidium ( 99 % of @xmath1rb ) that is contained in a glass cell at room temperature to keep atomic vapor concentration low and avoid reabsorption . transition @xmath47 @xmath48 to @xmath49 @xmath50 at @xmath51 nm is excited using both temperature- and current - stabilized single - mode diode laser ( sony sld114vs ) with beam diameter @xmath52 mm . absorption signal ( transmitted light ) is monitored by a photodiode . as the laser is swept applying a ramp on a drive current , two absorption peaks with half - width of about @xmath53 mhz separated by @xmath54 ghz appear due to the @xmath1rb ground state hyperfine structure . the excited state hyperfine structure is not resolved under doppler profile . during the resonance measurements , the laser wavelength is stabilized on absorption peak originating from ground sate hyperfine level @xmath2 . helmholz coils are used to sweep the magnetic field over zero gauss region , the earth magnetic field components are compensated . signal detected in a transmitted light is averaged over 64 cycles and the result is presented in a fig . [ fig4 ] , data points and lorenz fitting , curve 1 . on the same figure a calculated signal , curve 2 , is presented . model for this calculation is the same as in case on fig . amplitude of the experimental and calculated signals are in arbitrary units and in figure are scaled in a way to make them easy to compare . for theoretical curve the parameters are chosen to have values that maximally reproduce our experimental conditions . the ground state relaxation rate @xmath55 @xmath56s@xmath39 was chosen as a reciprocal time of thermal motion of rb atoms at room temperature with most probable velocity @xmath57 mm/@xmath58 through the laser beam of radius @xmath59 mm @xcite . absorption rate was chosen to be @xmath60 @xmath56s@xmath39 . other parameters are as in fig . agreement between calculated and measured signal is remarkable . as far as our model does not account for transitions to other hyperfine levels and so does not reproduce experiment in full , we do not attempt to fit experimental points with theoretical curve , nevertheless achieved agreement fully convinces us that the proposed explanation of the reversed dark resonances is correct . to obtain quantitative description of this signal in future one must take into account in the numerical model all other hyperfine levels involved in the process . one of us ma is thankful to prof . neil shafer - ray for fruitful discussions . financial support from swedish institute visby program is greatly acknowledged .
origin of recently discovered reversed ( opposite sign ) dark resonances was explained theoretically and verified experimentally . it is shown that the reason for these resonances is a specific optical pumping of ground state level in a transition when ground state angular momentum is smaller than the excited state momentum . [ theorem]acknowledgement [ theorem]algorithm [ theorem]axiom [ theorem]claim [ theorem]conclusion [ theorem]condition [ theorem]conjecture [ theorem]corollary [ theorem]criterion [ theorem]definition [ theorem]example [ theorem]exercise [ theorem]lemma [ theorem]notation [ theorem]problem [ theorem]proposition [ theorem]remark [ theorem]solution [ theorem]summary
mesoscopic circuits with multiterminal electron wave guides have interesting and useful physical properties . figure [ fig : setup ] shows an example of a circuit where a y - branch wave guide acts as a coherent beam splitter for electrons injected from a normal or superconducting reservoir . such a circuit with normal electron reservoirs ( nyn ) has been first suggested to test statistical properties of the fermi electrons by measuring current - current correlations in the arms of the splitter.@xcite this correlation is negative for the case of fermi statistics , which was indeed observed in the experiment.@xcite however , the correlations may change the sign and become positive ( similar to the case of bose particles ) when the electrons are injected from a superconducting reservoir ( nys).@xcite this effect is related to the cooper pairing in the superconductor . another kind of interference effect , namely a phase dependent conductance , has been predicted@xcite for a circuit geometry ( syn ) where the arms of the electron beam splitter are connected to a superconductor . in this case , the electrons undergo andreev reflection from the ns interfaces@xcite , picking up the superconducting phase at the connection points . the phase difference at the connection points is created by a supercurrent flowing along the surface of superconducting electrode , e.g. due to a presence of a magnetic field ( see fig . [ fig : setup ] ) . the phase dependence of the conductance of such ns interferometers has been observed in a large number of experiments.@xcite in this paper , we investigate the properties of current - voltage characteristics of sns interferometers , i.e. sys - circuits where the arms of the y - branch splitter and the injection lead are connected to superconducting electrodes , as shown in fig . [ fig : setup ] . as is well known , the current between two superconductors at applied voltage @xmath2 ( @xmath3 is the superconducting gap ) is governed by the mechanism of multiple andreev reflections ( mar).@xcite the fingerprint of this transport mechanism is the subharmonic gap structure ( sgs ) at @xmath4.@xcite in sns interferometers one should expect the sgs to be significantly modified and become sensitive to the superconducting phase difference ( see fig . [ fig : setup ] ) . this conclusion is supported by the following qualitative argument . one may consider the interferometer as an sns junction ( @xmath5 ) which is coupled to the injection electrode @xmath6 via the y - branch wave guide . the sns junction contains bound andreev levels which are broadened due to the open connection to the injection lead . the andreev levels , the energy of which depends on the superconducting phase difference , produce resonances in the current transport between the superconducting reservoirs . these resonances will significantly modify the sgs , similar to the effect of superconducting bound states in long sns junctions.@xcite however , in the sns interferometers the resonance structures will be phase dependent.@xcite the problem has become particularly interesting due to recent experiments on a three - terminal sns interferometer by kutchinsky et al . , @xcite in which phase dependence of the conductance structures has been observed . in this experiment , the normal region of the interferometer was fabricated with diffusive two - dimensional metallic film . the effect will be more pronounced with a ballistic normal region , especially in the quantum transport regime , when the y - wave guide supports a small amount of conducting modes . in practice , such devices could be fabricated with etched or gated ballistic 2d electron gas in a multilayered semiconducting structure.@xcite quantum transport in ns interferometers has been investigated in refs . , and the resonance enhancement of the two - particle andreev current has been discussed . in sns interferometers the situation is more complex , since the interplay between the andreev bound states and mar will lead to resonances appearing in all multiparticle currents . the purpose of our study is to develop mar theory for sns interferometers , i.e. , to include the interference effect and the effect of andreev resonances in the mar calculation scheme@xcite and to analyze the phase dependent current - voltage characteristics ( cvc ) . the paper is structured as follows . in section ii we present the model of the interferometer and derive equations for the mar scattering amplitudes . in section iii , we present the results of numerical calculation of the cvc of the interferometer , and in section iv , we present analytical perturbative analysis of the resonant current structures in the cvc . we consider the sns interferometer shown in fig . [ fig : setup ] , which consists of a normal - electron y - branch quantum wire , fabricated e.g. with etched or gated 2d electron gas , which is connected to two bulk superconducting electrodes . the superconducting electrode connected to the two arms of the splitter is ring shaped , so that magnetic flux can be sent through the ring to induce a superconducting phase difference @xmath7 between the connection points @xmath8 and @xmath9 . the current is sent through the interferometer by applying voltage @xmath10 between the electrode @xmath6 and the ring . our aim is to calculate the injected current as a function of the applied voltage and the phase difference , @xmath0 . to this end we shall apply the theory of coherent multiple andreev reflections ( mar ) , which is based on the calculation of scattering states created by incoming quasiparticles from the three superconducting terminals.@xcite for the sake of simplicity , we adopt a single mode description of the wires and model the splitter by a 3@xmath113 scattering matrix @xmath12 ( see eq . ( [ s0 ] ) below ) . solving a coherent mar problem includes the following steps : first the recurrence for the scattering amplitudes is formulated by employing the boundary conditions at the scatterer and at the ns interfaces . then a solution for the recurrence can be constructed , from which the current is calculated . in the present case of a three - terminal junction standard mar technique can not be used directly . to solve this problem we will describe the andreev reflections from the two interfaces of the interferometer , seen by the injector @xmath6 , as a complex reflection from a single effective interface ; by doing this we reduce the 3-terminal problem to an equivalent 2-terminal problem , on which we can apply standard mar technique for a 2-terminal junctions developed in ref .. reflection from the interferometer ( left interface ) is then represented by a non - trivial transfer matrix that contains information about andreev states in the interferometer . to more clearly understand the effect of the phase difference on the injection current , we will neglect the length of the normal wires and assume all ns interfaces to be completely transparent . this will remove unnecessary complications due to normal - electron resonances @xcite and superconducting resonances @xcite associated with the saint - james - de gennes bound states.@xcite at the same time , the resonant property of the junction as well as the phase sensitivity of the current will persist due to the presence of andreev states in the interferometer with energy @xmath13 , @xmath14 in the mar regime , the time - dependent scattering states consist of a superposition of scattered waves with energies @xmath15 ( @xmath16 is an integer ) shifted with respect to the energy of the incoming wave @xmath17 . in each branch of the wire the wave functions of the sideband @xmath16 consists of a superposition of electron and hole plane waves moving in both directions , the corresponding amplitudes being labeled with @xmath18 according to fig . [ fig : setup2 ] . in the normal wires , the scattering state wave function has the form , @xmath19 where @xmath20 is the wave function in the @xmath21-th wire at the left side ( @xmath22 ) , @xmath23 is a coordinate along the wire , @xmath24 is the wave function at the right side and @xmath25 is the normal wave vector ( @xmath26 ) . in the superconducting electrodes , the corresponding wave functions , including six possible source terms , are : @xmath27 \nonumber \\ \psi_{l(2)}^s & = & e^{-i\sigma_z(\frac{ev}2t+\frac{\phi}4)}\bigl[(\delta_{3j } \psi_0^+ e^{i \tilde{k}^+_0 { x}}+ \delta_{4j } \psi_0 ^ - e^{-i \tilde{k}^-_0 { x}})e^{-iet } \nonumber \\ & + & \sum_{-\infty}^{\infty}(c_n \psi_n^+ e^{-i \tilde{k}^+_n { x } } + f_n \psi_n^- e^{i \tilde{k}^-_n { x}})e^{-ie_nt}\bigr ] \nonumber \\ \psi_{r}^s & = & e^{i\sigma_z\frac{ev}2t}\bigl[(\delta_{5j } \psi_0^+ e^{-i \tilde{k}^+_0 x}+ \delta_{6j } \psi_0 ^ - e^{i \tilde{k}^-_0 x})e^{-iet } \nonumber \\ & + & \sum_{-\infty}^{\infty}(k_n \psi_n^+ e^{i \tilde{k}^+_n x } + l_n \psi_n^- e^{-i \tilde{k}^-_n x})e^{-ie_nt}\bigr ] \label{ansatzs},\end{aligned}\ ] ] where @xmath28 are constant coefficients and @xmath29 are elementary solutions to the bogoliubov - de gennes equation @xcite , @xmath30 where @xmath31 and @xmath32 is the pauli matrix . the different source terms are distinguished by the index @xmath33 corresponding to injection of an electron / hole from each of the electrodes . injections from the left and from the right generate essentially different scattering states and currents , and to keep track of the side of injection we will introduce an additional side index @xmath34 corresponding to injection from the left electrodes and right electrode , respectively , @xmath35 the scattering in the y - branch splitter is described by a unitary scattering matrix : @xmath36 where the scattering matrix , @xmath37 is chosen to be symmetric , @xmath38 , with a real transmission amplitude from right to left , @xmath39 , and also to be independent of energy and thus the same for electrons and holes , @xmath40 the latter approximation is reasonable since the scale of the energy dispersion of the scattering matrix is given by the fermi energy , while we are interested in a much smaller energy interval comparable to the superconducting gap , @xmath41 . the unitary conditions for @xmath12 are expressed through the equations , @xmath42 , @xmath43 , where @xmath44 , @xmath45 and @xmath46 , and @xmath47 the transmission coefficient @xmath48 concerns the transparency of the sns junction of the interferometer and controls the position of the andreev levels , eq . ( [ al ] ) , while the transmission coefficient @xmath49 describes coupling of the sns junction to the injection lead . by means of a canonical transformation the scattering between the injection electrode and the interferometer can be separated from the scattering between the arms of the interferometer . to this end we introduce new scattering amplitudes , @xmath50 , and similarly for @xmath51 , and @xmath52 , and rewrite eq . ( [ scatt ] ) on the form , @xmath53 @xmath54 the scattering equations for holes have a similarly form . it is clear from eq . ( [ scatt2 ] ) that the transport through the interferometer , from the right electrode to the left electrodes , is independent of the ( -)-coefficients . thus , we can treat the junction as an effective two - terminal junction with transparency @xmath55 . it is convenient for the further calculations to introduce vector notations , @xmath56 and similar for the other scattering amplitudes , and to rewrite eq . ( [ scatt2 ] ) through a transfer matrix , @xmath57 where @xmath58 . the same equation also holds for the hole coefficients . it is convenient to gauge out the reflection phase @xmath59 from the transfer matrix , which can be done by a transformation of the scattering amplitudes , @xmath60 , whereby @xmath61 we now proceed with a derivation of the recurrence relation and first consider the right ns - interface , where we match the wave functions in eq . ( [ ansatzn ] ) and ( [ ansatzs ] ) , @xmath62 this is the recurrence relation for an ideal sn - interface where the andreev reflection amplitude is given by @xmath63 and the source term , @xmath64 corresponds to the injection from the right electrode , @xmath65 . similar relations hold for each of the two sn - interfaces at the left side . in terms of the vectors @xmath66 and @xmath67 introduced in eq . ( [ trans ] ) , they have the form , @xmath68 for @xmath69 . for @xmath70 there is also a source term describing injection from the left electrodes ( @xmath71 ) . now the vectors @xmath72 and @xmath73 can be eliminated by means of eq . ( [ x1 ] ) for electrons and for holes , which yields @xmath74 where @xmath75 @xmath76 . the source term , @xmath77 , for @xmath78 has the form @xmath79,\;\ ; \nu=-1 , \label{pnl}\end{aligned}\ ] ] where @xmath80 . the source term for the injection cases @xmath81 has form similar to eq . ( [ pnl ] ) with @xmath82 . the matrix @xmath83 describes complete reflection from the interferometer for energies inside the energy gap , @xmath84 , where it obeys a standard transfer matrix equation , @xmath85 . for @xmath86 this is purely andreev reflection , identical to the one described by eq . ( [ rec1 ] ) , which implies that in this case the interferometer works as a single ideal sn interface . in the general case , @xmath87 , the reflection consists of both andreev and normal reflections , the probability of andreev reflection being given by the matrix element @xmath88 . thus , at @xmath1 the probability turns to zero , and therefore the andreev transport through the interferometer is blocked . furthermore , the transfer matrix @xmath83 contains information about andreev bound states in the interferometer . assuming for a moment the sns junction disconnected from the injection lead , @xmath89 and considering a stationary version of eq . ( [ rec2 ] ) , @xmath90 , in combination with eq . ( [ t0 ] ) we get the solvability condition for these equations on the form @xmath91=0 $ ] , which gives the andreev level in eq . ( [ al ] ) . in principle the equations ( [ rec1 ] ) , ( [ rec2 ] ) and ( [ rec3 ] ) provide a complete set of recurrences for the mar amplitudes . however , following ref . , it is convenient to introduce new amplitudes , @xmath92 , which allow us to get rid of redundant mar amplitudes and to unify notation for different injection cases , @xmath93 then the recurrence relations eq . ( [ rec1 ] ) , ( [ rec2 ] ) and ( [ rec3 ] ) can be written on a compact form , @xmath94 where @xmath95 and the index @xmath96 specifies the particular form of the transfer matrices @xmath97 and @xmath98 for different side bands and for different injection directions . equations ( [ c])-([ut ] ) realize a mapping of the mar problem on the problem of wave propagation along the energy axis . the wave amplitudes for the propagation in the upward and downward directions are given by the upper , @xmath99 , and lower , @xmath100 , components of the vector @xmath101 , @xmath102 the probability current flowing along the energy axis is defined in the usual way as @xmath103 this probability current is conserved within the superconducting gap , @xmath104const for @xmath84 , due to the unitarity properties of the matrices @xmath105 and @xmath106 . violation of the unitarity for @xmath107 outside the gap indicates leakage of the probability current into the reservoirs . we notice that in real space , the probability current is carried alternatively by electrons and holes . the charge current through the interferometer is determined by the probability currents of all the side bands , @xmath108 in this equation , the spectral current @xmath109 consists of the sum over the scattering states and all the sidebands at the injection side of the junction , @xmath110 ; the factor @xmath111 is the superconducting density of states , and @xmath112 is the fermi distribution function within the superconducting reservoirs . for @xmath113 , the free solution of eqs . ( [ rec ] ) and ( [ recx ] ) can be written on the form , @xmath114 while the solution for negative @xmath16 , @xmath115 , reads @xmath116 at the injection point @xmath70 , the mar coefficients are found from eq . ( [ rec ] ) , @xmath117 this equation imposes two relations between the four unknown coefficients , @xmath118 ; another two relations are imposed by the boundary conditions at infinity in energy space @xmath119 which fix the ratios @xmath120 . to implement the boundary conditions , we introduce effective reflection coefficients , @xmath121 , which characterize the intensity of the back scattering at specific points of the mar ladder . the definition reads , @xmath122 where @xmath123 with help of the vectors @xmath124 the solution in eq . ( [ mnm ] ) can be presented in a compact form ( using the equality @xmath125 ) , @xmath126 where the brackets denote a scalar product of two - component vectors , and @xmath127 the corresponding expression for negative @xmath16 is @xmath128 the form of solution in eqs . ( [ c>0])- ( [ c<0 ] ) is particularly useful for the calculation of the sgs : the essential processes which contribute to the sgs involve transmission through the energy gap region . this transmission contains resonances which are included in the transfer matrix @xmath129 . the latter is straightforward to calculate because it does not require excursions to infinity . on the other hand , the calculation of the quantities @xmath121 includes such excursions which , however , go outside the energy gap , and the corresponding recurrences rapidly converge . to calculate the sgs , it is convenient to separate out the current associated with the scattering across the gap , which is most important for the sgs , from the current of thermal excitations which involves transitions between states below or above the gap . to this end , we first rewrite the spectral current in eq . ( [ jdef ] ) through leakage currents into the electrodes defined as , @xmath130 . by using the equality @xmath131 following from eq . ( [ recx ] ) , we get @xmath132 ( here we explicitly write the scattering state index @xmath133 , introduced in eq . ( [ ansatzs ] ) ) . we note that the conservation of the probability current inside the energy gap implies that the leakage current is zero inside the gap , @xmath134 . having made this observation , we write the total current in eq . ( [ jdef ] ) on the form , @xmath135 @xmath136 \tanh{e\over 2k_bt } . \label{isep}\ ] ] the first two terms in this equation correspond to the current of thermal excitations : they cancel each other when the temperature approaches zero due to detailed balance,@xcite @xmath137 for the partial currents defined as @xmath138 keeping the last term in eq . ( [ isep ] ) , we get for the current at low temperature , @xmath139 , the following equation , @xmath140 the current @xmath141 in eq . ( [ jnu1 ] ) can be presented after some algebra on the form which is useful for analytical study , @xmath142 where @xmath143\hat r_{0-}),\\ i_{n,\nu}(e)&= & ( \hat r_{n+},[\hat{u}_n^{\dagger-1}\sigma_z\hat{u}_n^{-1}-\sigma_z]r_{n+ } ) \label{in}\nonumber.\end{aligned}\ ] ] in this section , we present the results of numerical calculation of the current given by eq . ( [ itot ] ) , where the spectral current is given by eq . ( [ jx ] ) . in order to numerically evaluate the effective reflection amplitudes @xmath121 we use the equations , the definitions eq . ( [ rdef ] ) , using eq . ( [ mnm ] ) and eq . ( [ lim ] ) , @xmath144 which follow from the definitions in eq . ( [ rdef ] ) , and also from eq . ( [ mnm ] ) and eq . ( [ lim ] ) . for zero phase difference , the current - voltage characteristic ( cvc ) in fig . [ fig : iv0 ] shows the well known subharmonic gap structure for short junctions with current structures at @xmath4 , corresponding to thresholds for the @xmath16-particle currents@xcite . these current structures are associated with the usual gap edge singularities,@xcite and they are most pronounced if the transparency @xmath55 of the junction is small , @xmath145 . the cvcs of the interferometer for different values of the phase difference at zero temperature are shown in fig . [ fig : iv1 ] . increasing the phase difference , the structure changes and the positions of the current structures can no longer be associated with the subharmonics of the energy gap . instead , we observe enhancement of the subgap current and the appearance of large current peaks which move downwards in voltage with increasing phase difference . if the transparency is small , as in fig . [ fig : iv1 ] , the peak positions reflect the andreev spectrum of eq . ( [ al ] ) : the peaks appear at @xmath146 and @xmath147 , as shown in fig . [ fig : closeup ] . with further increase of the phase difference , the current structures decrease and completely disappear at @xmath1 . at the same time , the excess current at large voltage becomes large and negative , turning into a deficit current . the cvc is @xmath148-periodic and symmetric around @xmath149 . it is illuminating to plot the individual n - particle currents , given by eq . ( [ itotn ] ) . in fig . [ fig : iv2 ] , the first four n - particle currents are shown for @xmath150 and for the same interferometer parameters as in fig . [ fig : iv1 ] . higher order currents are too small to be visible in the same picture , and their contribution to the current may be neglected . for comparison , the inset shows the corresponding cvc when @xmath86 . it follows from the plots for @xmath87 , that the peak at low voltage originates exclusively from the four - particle current . it exists within the voltage interval @xmath151 and within a certain window of phase difference . the next peak , at larger voltage , results from a combination of two structures : the peak of the three - particle current , and the overshoot of the onset of the pair current . the onset of a single - particle current at @xmath152 is reduced in comparison with the @xmath153 case , which leads to significant reduction of the current at large voltage . the reason is spectral intensity transfer from the continuum to bound andreev states . the large deficit current at high voltage is apparently the result of this reduction as well as of the above mentioned reduction of the multiparticle currents related to the suppression of andreev reflections by the interferometer . the further analysis shows that the single- and three - particle currents ( and in general odd @xmath16-particle currents ) injected from the electrode @xmath6 and from the electrodes @xmath154 are identical . in contrast , the two- and four - particle currents possess large structures only for injection from the electrode @xmath6 while currents injected from the interferometer electrodes @xmath154 are negligible small . in the next section , we will proceed with the interpretation of these features . in the case of weak coupling of the @xmath5 junction in the interferometer to the injection lead @xmath6 , @xmath145 , it is possible to perform a perturbative analysis of the n - particle currents in order to explain the origin of the phase dependent current structures in the cvc . in this limit , the current is dominated by the contribution of the andreev resonances , which we will consider in the lowest order with respect to the small parameter @xmath145 . we will proceed with a detailed analysis of the four lowest - order n - particle currents in order to explain the major structures in the ivc . let us start the discussion with the approximation for the effective reflection amplitudes at specific levels in the mar ladder , @xmath124 , in eq . ( [ rm ] ) . in the lowest order approximation with respect to @xmath49 , these reflection amplitudes are contributed by single back scattering , @xmath155 since the denominator @xmath156 in expression eq . ( [ jx ] ) for the spectral current contains a factor @xmath157 which comes from the product of transfer matrices @xmath105 in eq . ( [ mnm ] ) , we can neglect all terms except the zero order one in eq . ( [ in ] ) , whereby @xmath158 the denominator @xmath156 in eq . ( [ jx ] ) has to be considered separately for different currents , @xmath16 , and injection cases , @xmath110 . the single - particle current involves only a single crossing of the tunnel barrier , schematically shown in fig . [ fig : mar]a . thus , it can not be resonant , and the denominator @xmath159 can be considered in the lowest - order approximation . the expression for the single particle current , derived from eqs . ( [ itotn ] ) , ( [ jx ] ) , and ( [ iappn ] ) then reads , @xmath160 where @xmath161 the form of eq . ( [ i1appr ] ) is a standard one for a tunnel current given by the tunnel hamiltonian model were @xmath162 is the density of states ( dos ) in the injection lead , while @xmath163 has the meaning of the effective phase dependent dos in the interferometer . an important property of the dos of the interferometer is the disappearance of the singularities at the gap edges for nonzero phase difference . the reason is that the resonances ( andreev states ) move down into the gap , according to eq . ( [ al ] ) ( see fig . [ fig : dos ] ) . consequently , the onset of the single - particle current at @xmath164 , which is of the order @xmath49 when @xmath86 , is reduced at finite phase difference ( see fig . [ fig : iv2 ] ) . in order to calculate the pair current we will first inspect the corresponding transfer matrix , @xmath165 . for @xmath65 ( injection from the right ) this transfer matrix has zeros in the limit @xmath89 at the energies of the andreev bound states in the interferometer , @xmath166 . the expansion of the matrix in powers of @xmath49 has the form @xmath167 + o(t^2 ) , \label{m20}\end{aligned}\ ] ] where we have used the identity , @xmath168 , derived from the unitarity condition eq . ( [ unit ] ) and @xmath76 , in the limit @xmath89 . equation ( [ m20 ] ) clearly demonstrates the resonant behavior . the first term is real and proportional to the deviation of the energy @xmath169 for the andreev reflection from the energy of the andreev state . this term determines the position of the resonance . the second , imaginary , term is proportional to @xmath49 and determines the width of the resonance . both the position and width of the resonance depend on the phase difference . at @xmath145 the spectral current is dominated by the andreev resonance , and neglecting the non - resonant current we arrive at the expression @xmath170 where @xmath171 @xmath172 . a factor of 2 appears in the current expression , in addition to the two - particle factor @xmath173 in eq . ( [ jx ] ) , since there are two andreev resonances , at @xmath174 and @xmath175 . the integrand in eq . ( [ i2appr ] ) has the form of the transmission coefficient for an asymmetric quantum mechanical double barrier structure . the difference in the tunnel rates @xmath176 is due to the different dos for the injection and exit energies , but also due to the different effective transparencies of the y - branch splitter , @xmath177 , for electrons and holes . the resonant transparency of the junction for the pair current is proportional , according to eq . ( [ i2appr ] ) , to the product @xmath178 which turns to zero at @xmath1 . thus the resonant pair current is blocked at @xmath1 , in agreement with our earlier general conclusion about the suppression of andreev reflection at the interferometer in this case . evaluation of the integral over energy in eq . ( [ i2appr ] ) under the condition , @xmath179 , yields the pair current @xmath180 the magnitude of this current is generally of order @xmath49 ; however , in the vicinity of the resonant onset , @xmath181 for deviations @xmath182 , the rate @xmath183 increases due to the enhancement of the dos in the right superconductor , @xmath184 , which yields the overshoot seen in fig . [ fig : iv2 ] . at lower voltage , @xmath185 , the pair current is non - resonant and proportional to @xmath186 , which is also the case for the pair current @xmath187 injected from the interferometer . the resonant three - particle current can be analyzed in a similar manner . we construct the corresponding transfer matrix using the pair current matrix , eq . ( [ m20 ] ) , and for @xmath65 we get @xmath188 . this equation has resonances at the same energies as the pair current , @xmath166 , but nevertheless the resonant current is small , @xmath189 , since the mar path , fig . [ fig : mar]c , includes an additional crossing of the tunnel barrier . the magnitude of the current is considerably enhanced in the vicinity of voltage @xmath190 when the second andreev reflection occurs at the gap edge of the right electrode , @xmath191 ( see fig . [ fig : mar]c ) . the singular dos at the gap edge in this electrode appears in the mar calculation as a quasibound state situated at the gap edge . thus , this mar process can be interpreted as an overlap of two resonances , one at the andreev bound state at energy @xmath175 , and another at the quasibound state at the gap edge . a similar analysis applies to the current injected from the interferometer , @xmath71 : in this case , the double resonance is formed by the andreev state at @xmath174 and the quasibound state at lower gap edge . the height of the double resonance peak is @xmath192 . a double resonance involving two andreev bound states occurs in the four - particle current injected from the right as shown in fig . [ fig : mar]d . the corresponding mar path is selected by the conditions @xmath193 and @xmath194 . the resulting peak is situated at @xmath195 and has the hight @xmath196 . in the vicinity of the peak , @xmath197 , the expression for the resonant four - particle current becomes , @xmath198 ^ 2+e_2 ^ 2{\tilde \gamma}_-^2 } , \label{i4}\ ] ] where @xmath199 the current peak under consideration is a full - scale current structure in a region of the ivc which in point contacts is dominated by non - resonant four - particle current , which is of the order @xmath200 . as any andreev resonance , the four - particle current peak strongly depends on the phase difference ; it disappears at @xmath1 . the peak exists within the voltage interval @xmath151 and within the phase difference window @xmath201 . we note that the positions of all discussed current structures are proportional to the order parameter @xmath3 and therefore temperature dependent and scale with @xmath3 . we end this section with the discussion of cvc at large voltage , @xmath202 . the current at large voltage results exclusively from the single particle and pair currents . it is straightforward to calculate the single particle current at large voltage from eq . ( [ i1appr ] ) , @xmath203 the first term in this equation is the current through the normal junction , while the second term is the single - particle contribution to the excess current , @xmath204 this excess current is large , @xmath205 , and negative , which reflects the suppression of the current onset at @xmath164 . the pair current contribution to the excess current is dominated by the resonance and is given by eq . ( [ i2a ] ) . the total excess current to leading order in @xmath49 has the form @xmath206 the excess current in the interferometer is large and negative , in sharp contrast to quantum point contacts , where the excess current is small , @xmath207 , and positive.@xcite however , it turns to zero at @xmath86 , in agreement with our observation that in this case the interferometer behaves as a quantum point contact . the excess current achieves its minimum value at @xmath1 because of blockade of the pair current . we have investigated the phase dependence of the dc current - voltage characteristics ( cvc ) in sns interferometers consisting of two superconducting reservoirs connected by a y - shaped normally conducting quantum wire . we have developed a generalization of the existing coherent mar theory , which fully incorporates the effect of interference , as well as resonance effects due to andreev bound states in the interferometer . the phase dependent current - voltage characteristic @xmath0 has been numerically investigated for single - mode wires with length smaller than the superconducting coherence length . the analytical study of the current was performed using a resonance approximation in the limit of weak coupling of the interferometer to the injection lead . we found significant enhancement of the subgap current at finite phase difference compared to the case of zero phase difference . the enhancement is accompanied by a shift of the onset of the current at @xmath208 to smaller voltages , @xmath209 , and appearance of current peaks at @xmath174 and @xmath210 . these features are produced by resonant andreev reflection of the injected particles by the s@xmath211ns@xmath212 junction of the interferometer : they are most pronounced for weak coupling of the s@xmath211ns@xmath212 junction to the injection lead . the positions of all current structures are temperature dependent and scale with @xmath3 . we further found strong suppression of the current when the phase difference approaches @xmath213 : in this region all subgap current structures disappear and , moreover , the excess current at large voltage becomes negative ( deficiency current ) . the magnitude of the deficiency current is considerably large at @xmath1 . this effect results from the suppression of the andreev reflections from the s@xmath211n and s@xmath212n interfaces due to the interference in the arms of the beam splitter ( cf . a similar effect in ns interferometers@xcite ) . we conclude our discussion with some brief comments on the connection of the studied model to more realistic experimental devices based on 2d electron gas structures . in such devices , the normal region has a typical size which exceeds the superconducting coherence length and it contains several conducting electronic modes . furthermore , the ns interfaces are not perfectly transparent . however , this effect is less important since the interface reflection coefficient in practice can be rather small ( @xmath214 ) . increasing the length of the arms of the interferometer will result in the appearance of new phase dependent current peaks due to increasing number of andreev states in the sns junction of the interferometer ( cf . ref . ) . similarly , increasing the length of the injection lead will result in the appearance of current resonances which do not depend on the phase difference , due to @xmath7-independent superconducting bound states in this lead.@xcite in interferometers with multimode wires , the resonant current peaks will be smeared ; however , one should still expect some phase dependent resonant features similar to the case of diffusive ns interferometers @xcite . the interference effect leading to the large deficient current at large applied voltage will survive in multimode junctions . the authors gratefully acknowledge support from nfr and kva ( sweden ) and nedo ( japan ) . 999 th . martin and r. landauer , phys . rev . b * 45 * , 1742 ( 1992 ) . m. bttiker , phys . b * 46 * , 12485 ( 1992 ) liu , b. odom , y. yamamoto , and s. tarucha , nature 391 , 263 ( 1998 ) . j. torrs and th . martin , europhys . j. b * 12 * , 319 ( 1999 ) . h. nakano and h. takayanagi , phys . b * 47 * , 7986 ( 1993 ) . f. w. j. hekking and yu . v. nazarov , phys . l * 71 * , 1625 ( 1993 ) . andreev , sov . jetp * 19 * , 1228 ( 1964 ) . petrashov et al . , phys . lett . * 70 * , 347 ( 1993 ) . de vegvar , et al . , * 73 * , 1416 ( 1994 ) . lambert and r. raimondi , j. phys . cond . mat . * 10 * , 901 , 1998 t.m . klapwijk , g.e . blonder and m. tinkham , physica b&c * 109 - 110 * , 1657 ( 1982 ) j.m . rowell and w.l . feldmann , phys . rev . * 172 * , 393 ( 1968 ) . e.n . bratus , v.s . shumeiko and g. wendin , phys . lett.*74 * , 2110 ( 1995 ) . ingerman , g. johansson , v. s. shumeiko , and g. wendin , phys . b * 64 * , 144504 ( 2001 ) . a phase - dependent single - particle tunnel current in 4-terminal sins interferometers with diffusive normal region was recently studied in : f.k . wilhelm and a.a . golubov , phys . rev . b * 62 * , 5353 ( 2000 ) . j. kutchinsky et al . , phys . b * 56 * , r2932 ( 1997 ) . h. takayanagi , e. toyoda , and t. akazaki , superlatt . and microstr . * 25 * , 993 ( 1999 ) . a. kadigrobov et al . , phys . b * 52 * , r8662 ( 1995 ) . p. samuelsson , v.s . shumeiko and g. wendin , phys . b * 56 * , r5763 ( 1997 ) : phys . rev . b * 62 * , 1319 ( 2000 ) chang and p.f . bagwell , phys . b * 55 * , 12678 ( 1997 ) . d. averin and a. bardas , phys . lett . * 75 * , 1831 ( 1995 ) . shumeiko , e.n . bratus , and g. wendin , low temp . phys . * 23 * , 181 ( 1997 ) . g. johansson , e.n . bratus , v. s. shumeiko , and g. wendin , superlatt . and microstr . * 25 * , 905 ( 1999 ) . g. johansson , e. bratus , v. shumeiko , and g. wendin , phys . b * 60 * , 1382 ( 1999 ) . g. de gennes and d. saint - james , phys . lett . * 4 * , 151 ( 1963 ) . p.g de gennes , _ superconductivity in metals and alloys _ , addison - wesley publishing co. , inc . , 1989 . schrieffer and j.w . wilkins , phys . * 10 * , 17 ( 1963 ) .
a theory of coherent multiple andreev reflections ( mar ) is developed for superconductor - normal metal - superconductor ( sns ) interferometers . we consider a y - shape normal electron beam splitter connecting two superconducting reservoirs , where the two connection points to the same superconductor can have different phase . the current is calculated in the quantum transport regime as a function of applied voltage and phase difference , @xmath0 . mar in interferometers incorporates two features : interference in the arms of the splitter , and interplay with andreev resonances . the latter feature yields enhancement of the subgap current and current peaks with phase - dependent positions and magnitudes . the interference effect leads to suppression of the subgap current and complete disappearance of the current peaks at @xmath1 . the excess current at large voltage decreases and changes sign with increasing phase difference .
while the rapid rotation ( v@xmath3/v@xmath4 @xmath270 - 80% of their critical velocity ; @xcite ) of classical be stars has long been speculated to be the fundamental source driving the production of their geometrically thin circumstellar disks @xcite , recent photometric surveys @xcite have suggested secondary mechanisms might contribute to the observed phenomenon . specifically , @xcite and @xcite found that the frequency of the be phenomenon seems to peak in clusters with a main sequence turn - off of b1-b2 , leading to the suggestion that the be phenomenon is enhanced with evolutionary age . several recent observational studies @xcite have also suggested that the be phenomenon may be more prevalent in low metallicity environments , based on comparisons of the fractional be populations of galactic , lmc , and smc clusters . in @xcite ( hereafter wb06 ) , we used a simple 2-color diagram ( 2-cd ) technique to identify the fractional candidate be population of numerous large magellanic cloud ( lmc ) , small magellanic cloud ( smc ) , and galactic clusters in an effort to improve the statistical database which has been used to link classical be disk formation with evolutionary age and/or metallicity . wb06 found evidence that the be phenomenon develops much earlier than previously predicted by theory @xcite , i.e before the mid - point main sequence lifetime ; furthermore , they found evidence of an additional enhancement in the fractional be content of clusters with evolutionary age . the increased statistics offered by this work , while confirming the previously suggested trend of an enhancement in the be phenomenon in low metallicity environments , lowered the average fractional be content of smc clusters from 39% @xcite to 32% . while the 2-cd has been widely used to link evolutionary age and/or metallicity with be disk formation , it is inherently unclear whether all b - type objects identified as excess h@xmath0 emitters , i.e. `` be stars '' , are truly classical be stars . it has been noted that other b - type objects , such as herbig ae / be stars , post - main - sequence b[e ] stars , and supergiants , may also exhibit h@xmath0 emission hence `` contaminate '' these claimed detections ( wb06 ) . polarimetry is a tool which has long been used to investigate the circumstellar environments of be stars @xcite . electron scattering in the innermost region of classical be circumstellar disks will polarize a small fraction of stellar photons ; furthermore , pre- and post - scattering attenuation by hydrogen atoms within the inner disk region will imprint a characteristic wavelength dependence onto this polarization signature if the inner disk is sufficiently dense @xcite . the wavelength dependence of intrinsic polarization originating from classical be stars , characterized by geometrically thin , gaseous circumstellar disks , @xcite is significantly different than that expected from the dustier circumstellar environments of herbig ae / be stars , post - main - sequence b[e ] stars , and supergiants @xcite . thus for non - pole - on disk geometries , polarimetry can be employed to investigate circumstellar environments and discriminate the evolutionary nature of sources . in this paper , we use imaging polarimetry to investigate the true nature of candidate be stars in six lmc and six smc clusters which were initially identified via photometric 2-cds . in section 2 , we outline our observations and data reduction techniques . in section 3 , we detail how we identified and removed the interstellar polarization component associated with each of our lines of sight . we develop a classification system to characterize the intrinsic polarization observed in our candidate be star sample in section 4 and also discuss the results of individual clusters . we discuss the implications of these results in section 5 and summarize our work in section 6 . the imaging polarimetry data presented in this study were obtained at the cerro tololo inter - american observatory ( ctio ) 1.5 m . the f/7.5 secondary configuration was used during our 2001 observing run , which yielded a 15 x 15 arc - minute field of view and a 0.44 arc - second pixel@xmath5 scale , while the f/13.5 secondary configuration , which provided a 8 x 8 arc - minute field of view with a 0.24 arc - second pixel@xmath5 scale , was used during our 2002 observing run . data were recorded with ctio s standard cassegrain focus ccd ( cfccd ) , a 2048 x 2048 multi - amplifier ccd . the standard cfccd configuration was modified by the addition of a rotatable half - wave plate followed by a fixed analyzer , the latter of which was placed in the upper of the 1.5 m s two filter wheels . the polymer achromatic wave - plate was manufactured by meadowlark optics and provided polarization modulation for each 90@xmath6 rotation of the wave - plate . unless otherwise noted , each of our fields of view was observed with the wave - plate rotated through 8 positions 22.5@xmath6 apart . the fixed analyzer was a double calcite block whose optical axes had been crossed to minimize astigmatism and color effects . further details about this polarimeter can be found in @xcite , @xcite , @xcite , & @xcite . we placed ctio s tek u , b , v , r , and i filters in the second filter wheel . a summary of the science clusters we observed is presented in table [ polexpose ] . nightly observations of polarization standard stars were obtained to calibrate the polarimetric efficiency of each filter and determine the zero - point of the polarization position angle in each filter , while nightly observations of unpolarized standard stars were obtained to characterize any instrumental polarization present . our unpolarized standard data indicate that the polarimeter had an instrumental polarization consistent with zero for our 2001 data . the b , v , r , and i filters of our 2002 data also had an instrumental polarization consistent with zero @xcite , while the u filter data had an instrumental polarization of @xmath20.1% , which was removed from the data . the reduction of these data began with standard image processing , including bias and flat field corrections , using standard iraf techniques . following this initial processing , aperture photometry was performed for all point sources in each of the 8 images of an observation set , using 10 apertures of size 3 - 12 pixels for the f/7.5 data and 5 - 14 pixels for the f/13.5 data . the linear polarization of each source was extracted from a least squares solution of the difference amplitudes in the 8 wave - plate positions ( @xmath7 ) , using the pccdpack software suite @xcite . both the expected photon noise errors and the actual measured errors , defined as the residuals at each wave - plate position with respect to the expected @xmath8 curve , were also calculated . with few exceptions , these two errors were consistent with one another . the polarization observed in our raw data is comprised of the superposition of an interstellar component , attributable to the dichroic absorption of starlight by partially aligned dust grains along each line of sight , and possibly an intrinsic component , attributable to the asymmetric illumination of circumstellar material and/or the illumination of an asymmetrical distribution of circumstellar material . characterizing and isolating each of these components , although often technically challenging , yields two unique diagnostics . isolating the interstellar polarization component can facilitate detailed studies of the local interstellar medium ( ism ) and provide constraints on fundamental properties such as grain size distributions and shapes , as well as diagnosing the magnitude and direction of local magnetic fields . isolating the intrinsic polarization component enables investigations of the distribution and chemistry of scatterers which comprise unresolved circumstellar environments . we stress that characterizing the polarizing agent ( i.e. gas or dust ) responsible for producing an intrinsic polarization signal is not possible simply via inspection of the wavelength dependence of the vector sum of interstellar plus intrinsic ( i.e. the total ) polarization data . fortunately , our efforts to identify and separate the interstellar and intrinsic polarization components of our data are simplified by the fact that we have observed rich stellar associations . most of the stars in these clusters should be normal main sequence stars lacking any type of asymmetrical circumstellar envelope , hence they should not exhibit an intrinsic polarization component . furthermore , we can assume that all of our stars are located at the same general distance , since they are members of a cluster population , and that the properties of the interstellar medium do not significantly vary over the angular extent of our clusters . hence each cluster in our sample provides us with a multitude of suitable `` field stars '' whose polarization we can simply average to estimate the interstellar polarization ( isp ) along each line of sight , via the commonly used field star technique @xcite . in figure [ 1818vsel ] we show the total v - band polarization of all sources in the lmc cluster ngc 1818 having polarimetric signal - to - noise ratios ( p/@xmath9 ) greater than 5.0 . the distinctive grouping of most objects in the stokes q - u diagram of this figure ( panel a ) and the narrow polarization ( panel d ) and position angle ( panel c ) histograms demonstrates that most of these objects do lack significant intrinsic polarization components and only have an isp component . after excluding objects outside of the dominant trends found in these histograms with an iterative method , we calculated the weighted average and standard deviation of these data to determine a preliminary estimate of this cluster s v - band isp , as shown in figure [ 1818serkv ] . the wavelength dependence of these estimates clearly follow the empirical serkowski law @xcite commonly used to parameterize an isp . over - plotted in figure [ 1818serkv ] is a modified serkowski law @xcite which we deemed best represented the observational data . the polarization at the midpoint wavelength location of the u , b , v , r , and i filters was then extracted from this curve and served as the final estimate of the isp . the modified serkowski parameters determined for all of our lines of sight , along with the extracted filter isp values , are summarized in table [ isptable ] . the isp values compiled in table [ isptable ] are relevant to our efforts to isolate the intrinsic polarization components associated with our target stars ; however , they do not yet yield any direct information regarding the properties of the interstellar dust grains which reside in their parent lmc and smc clusters , as these isp values are comprised of both foreground galactic isp and magellanic cloud isp contributions . following the techniques used by @xcite and @xcite , we used the interstellar polarization maps of @xcite , to identify and remove this foreground galactic isp contribution ( see table [ foreground ] ) , hence determine estimates of the isp intrinsic to our lmc and smc clusters ( see table [ ispintrin ] ) . we used a modified serkowski - law @xcite to extrapolate these maps over the wavelength range of our dataset , and assumed a nominal galactic @xmath1 value of 5500 to characterize the parameter k. the errors cited in the galactic polarization maps @xcite and those present in our total isp estimates ( table [ isptable ] ) were propagated to produce the error estimates cited in table [ ispintrin ] . most of the field stars used to derive our total isp estimates were not highly reddened objects , hence the intrinsic lmc and smc isp values listed in table [ ispintrin ] all are characterized by modest polarization amplitudes , which are often on the order of the errors of our data . in spite of this fact , trends in these data are clearly present . as expected , the magnitude of ubvri polarization towards each cluster follows a serkowski - like wavelength dependence at a wavelength independent position angle . this lack of position angle rotation indicates that our galactic interstellar polarization correction was reasonable and that we are looking at a single magnetic field orientation towards each of our clusters . all lmc clusters exhibit similar isp properties , with polarization magnitudes ranging from @xmath20.2 - 0.5% at a position angle of @xmath225 - 45 degrees ; similarly , all smc clusters also exhibit clear evidence of sharing common isp properties , with polarization magnitudes ranging from @xmath20.3 - 0.6% at a position angle of @xmath2120 - 150 degrees . the shallow curvature of the serkowski - law dependence of these data , in combination with the moderate level of uncertainty present , make identifying systematic differences in the interstellar dust grain properties of these clusters difficult . the total isp ( table [ isptable ] ) and smc isp ( table [ ispintrin ] ) of the cluster ngc 330 does however show suggestive evidence of being characterized by a short wavelength @xmath1 value of @xmath24500 ( figure [ 330serk ] ) . such an effect is commonly attributed to the presence of small dust grains , and has been previously observed in other smc sight - lines @xcite . in a future paper , we will examine the wavelength dependence of moderately and highly reddened objects in our dataset to perform a more detailed investigation of the interstellar medium properties of these lmc and smc clusters , and compliment earlier studies which parameterized the ism properties of other smc @xcite and lmc @xcite sight - lines . the total polarization vectors of our observations of the lmc cluster ngc 2100 clearly exhibit evidence of a collective , complex morphology ( see figure [ rawn2100i ] ) . while figure [ rawn2100i ] only presents i - band data , all of our other filters exhibit similar alignment patterns . recall that we expect most cluster objects should not exhibit an intrinsic polarization component ; thus we suggest that these systematic morphological changes are related to changes in the magnetic field properties within our field of view . a substantial discussion of the magnetic field properties of ngc 2100 is presented in @xcite . for the purposes of the present study , we will only discuss our efforts to parameterize and remove the isp along the line of sight to this cluster . we identified three spatial regions in our field of view , corresponding to the region around the cluster core , the region to the south of the cluster core , and the region to the north of the core , which displayed unique isp characteristics . in figures [ n2100region1isp ] , [ n2100region2isp ] , and [ n2100region3isp ] , we show the b - band polarization vector maps of what we have defined as regions 1 , 2 , and 3 respectively in the total isp along the line of sight to ngc 2100 , overlaid on dss-2 blue images . having identified these distinct regions , we then extracted total isp estimates following the technique previously described . we calculated a @xmath10 weighted average of all objects in each filter served as an initial isp estimate , determined the modified serkowski - law parameters which best represented the data , as seen in figure [ n2100region1serk ] for region 1 , and extracted final isp values ( table [ isptable2 ] ) for each region from these serkowski curves . the locations of the candidate be stars in ngc 2100 were then correlated to these 3 isp regions to determine the isp correction each should receive . we found ngc 2100:kwbbe 102 , 630 , and 712 resided in area 2 , ngc 2100:kwbbe 535 , 797 , and 1033 resided in area 3 , and the rest of the candidate be stars resided in area 1 . we subtracted the isp from the observed polarization to isolate any intrinsic polarization components present . the location of candidate be stars in our fields of view were identified by careful correlation with literature coordinates and finder charts . using several pccdpack routines , we meticulously examined the extracted polarization for each of the candidate be stars in our sample to search for undesired contamination by a ) the ordinary or extraordinary images of nearby objects and b ) cosmic ray hits or uncorrected bad pixels . the total polarization of all candidate be stars which did not suffer from these contamination issues are tabulated in table [ totpol ] , while the intrinsic polarization of these objects are tabulated in table [ inpol ] . as discussed in the introduction , electron scattering in the pure gas disks of classical be stars will polarize a small fraction of stellar photons , producing a well known wavelength dependent intrinsic polarization signature . we emphasize that as polarization is a vector quantity , identifying these signatures a ) is only possible via inspection of intrinsic polarization data ; and b ) requires either the _ simultaneous _ inspection of both the wavelength dependence of the polarization magnitude _ and _ position angle , or alternatively inspection of the behavior of the intrinsic data on a stokes q - u diagram . an observed wavelength independent intrinsic polarization magnitude and position angle across the entire ubvri wavelength regime is the expected signature of pure electron scattering , such as that expected from a gaseous classical be disk . moreover , pre- or post - scattering absorption will superimpose the wavelength dependent signature of hydrogen opacity on this signal , creating a `` saw - tooth '' polarization signature @xcite , if enough absorption events occur . specifically , this `` saw - tooth '' signature includes the presence of abrupt jumps in the magnitude of intrinsic polarization at the balmer and paschen limits , as well as a wavelength independent intrinsic polarization position angle across the entire ubvri wavelength regime . examples of these two signatures of classical be circumstellar disks are given in figure [ pbopiaqr ] , which presents multi - epoch observations of the known classical be star pi aquarii as observed by the university of wisconsin s hpol spectropolarimeter . we have contemporaneously analyzed the stokes q - u diagrams of all candidate be stars total and intrinsic polarization components to search for evidence of these signatures . we have developed a conservative 4-point classification scale to rank the likelihood that candidate be stars are truly classical be stars . results of this classification are summarized in table [ starclass ] for individual objects and in table [ polsummary ] for the net results of entire clusters . the design of our classification system emphasized the identification of objects exhibiting a `` saw - tooth '' intrinsic polarization signature and a pure electron scattering signature across the _ entire _ ubvri wavelength regime . specifically , we defined our classification system by : * * type-1 : * objects which are _ most likely _ classical be stars ; * * type-2 : * objects whose polarimetric properties are _ not inconsistent _ with those expected from classical be stars ; * * type-3 : * objects which are _ unlikely _ to be classical be stars ; and * * type-4 : * objects which are _ highly unlikely _ to be classical be stars . stars in our sample which displayed , to within 3 @xmath11 , a `` sawtooth - like '' polarization balmer jump ( bj ) along with a wavelength independent polarization position angle across the entire range of available data ( see figure [ polbal ] ) were assigned a designation of type-1 . we claim that all type-1 objects are most likely classical be stars . note that we offer a detailed discussion of other objects which could exhibit somewhat similar behavior over _ part _ of the ubvri wavelength regime in section [ bevshaebe ] . however , as we discuss in section [ nircolors ] , analysis of ancillary data demonstrates that it is highly dubious that stars classified as type-1 polarimetric sources are anything other than classical be stars . we assigned a designation of type-2 to objects whose polarimetric properties were not inconsistent with that expected from a classical be star - disk system . as previously discussed , classical be disks of sufficiently low density will not leave an imprint of hydrogen opacity in their intrinsic polarization signals ; hence , stars which exhibited , to within 3@xmath11 , a wavelength independent electron scattering ( es ) polarization signature ( see figure [ poles ] ) received a designation of type-2 . a small number of objects exhibited a nearly wavelength independent polarization magnitude along with a minor wavelength dependence in their polarization position angles ; we designated these objects as type-2 as we believe such stars most likely are exhibiting es signatures modified by a slight under- or over - correction of their isp components . we can not rule out that this minor wavelength dependence might also be produced by the additional presence of an optically thin dust disk , as present in post - main - sequence b[e ] stars @xcite . however , as we show in section [ nircolors ] , analysis of ancillary data demonstrate that most type-2 objects which we claim exhibit electron - scattering ( es ) signatures are likely to be classical be stars , and substantially less likely to be systems characterized by composite gas plus dust disks ( i.e. herbig ae / be , b[e ] stars ) . finally , we assigned all objects which appeared to be intrinsically unpolarized ( @xmath12 0.3% polarization ) , to within 3@xmath11 , a type-2 designation as this is the expected signature of pole - on or nearly pole - on classical be stars . we caution the reader that we are not 100% certain that these unpolarized objects are classical be stars as : a ) stars without gaseous disks , spuriously detected as excess h@xmath0 emitters on 2-cds , will also exhibit zero net intrinsic polarization and b ) the noise present in the observations of fainter targets may preclude us from clearly identifying them as `` contaminants '' . we assigned the designation of type-3 to type-4 to objects whose polarization , to within 3@xmath11 , appeared to be inconsistent with the aforementioned signatures expected from a classical be star - disk system ( see e.g. figure [ polunlikely ] ) . we briefly discuss some of the major types of atypical polarimetric signatures we observed . in table [ starclass ] , we assigned a designation of type 3 - 4 to several objects whose intrinsic polarization exhibited signs of a 90@xmath6 position angle reversal ( figure [ n2100intrin111 ] ) , which is a signature of a dusty bipolar nebula geometry ( see section [ bevshaebe ] ; see also @xcite ) . from an inspection of the stokes q - u diagram of such an object ( figure [ n2100intrin111 ] ) , it is clear that even if our initial isp correction was grossly miscalculated , the wavelength dependent polarization would still be inconsistent with that expected from classical be stars . while it would be interesting to further probe the circumstellar environments of such objects with follow - up investigations , for the purposes of this paper we merely remark that they are unlikely to be classical be stars . several of our candidate be stars exhibited very large ( 1 - 3.5% ) , complex intrinsic polarization signatures ( see figure [ n1948serklong ] ) . given the steep drop in their polarization magnitude at short optical wavelengths , along with subtle indications of a corresponding position angle rotation , it is possible that these objects are characterized by dusty bipolar nebulae whose polarization signature exhibits a position angle reversal at uv wavelengths , similar to hd 45677 @xcite . alternatively , the polarization of these objects could be interpreted as following a serkowski - like wavelength dependence characterized by a large @xmath1 value . as our data have already been corrected for the average isp associated with each of our clusters , this latter interpretation would require these objects to be situated in region of patchy dust , likely populated by larger grains given the long wavelength values of @xmath1 , i.e. @xmath13 7000 in figure [ n1948serklong ] @xcite . @xcite described the expected magnitude of interstellar polarization by the formulation @xmath14 ; hence , an additional reddening of 0.3 - 1.0 e@xmath15 would be required to produce the measured p@xmath16 of these objects , @xmath23% . the observed colors of at least some of these objects ( ngc 1948:kwbbe 98 , v - i = 0.67 ; keller et al . 1999 ) is likely sufficient to produce the amount of purported additional isp , although other objects ( ngc 1948:kwbbe 246 , v - i = 0.05 ; keller et al . 1999 ) clearly do not meet this criteria . while follow - up optical or infrared spectroscopy of these anomalous candidates would help to determine whether they exhibit a dusty bipolar nebula geometry or are located in a region of patchy dust , for the purpose of this paper we merely stress that it is clear that these objects are not likely to be classical be stars . we now offer discussion of the polarimetric properties of candidate be stars in individual clusters . recall that all candidates which exhibited evidence of contamination , predominantly from nearby neighbors , have been excluded from our analysis . we further note that the effective limiting magnitude of our polarimetric dataset was lower ( i.e. brighter ) than that considered in wb06 . @xcite identified 40 candidate be stars associated with the lmc cluster ngc 1818 and its surrounding field . the cluster is densely populated , and because of image overlap issues , we were only able to retrieve polarimetric information for 18 of these candidate be stars . 4 of these 18 ( 22% ) showed intrinsic polarization balmer jump ( bj ) signatures , ngc 1818:kwbbe 69 , 82 , 137 , 243 . we classified 12 of the 18 candidates ( 67% ) as type-2 objects , and found 5 of the type-2 objects ( i.e. 5 of the 18 stars , 28% , in the total population ) exhibited clear evidence of an electron scattering ( es ) polarization signature . 2 of the 18 candidates ( 11% ) , ngc 1818:kwbbe 47 and 381 , were deemed unlikely to be classical be stars . while the b , v , r , and i filter polarization of ngc 1818:kwbbe 47 were consistent with an es signature , the u filter exhibited a significant position angle rotation which was inconsistent with an es origin . although noisy , the observation of ngc 1818:kwbbe 381 also did not follow an es wavelength dependence , prompting us to assign it an `` unlikely be star '' designation . @xcite identified 27 candidate be stars associated with the cluster ngc 1948 and its nearby field . we were able to extract polarimetric information on 22 of these candidates from our data set . 6 of these 22 candidates ( 27% ) exhibited a polarization balmer jump , ngc 1948:kwbbe 71 , 75 , 102 , 153 , 172 , and 240 . we classified 12 of the 22 candidates ( 55% ) as type-2 objects , and found 6 of the type-2 objects ( i.e. 6 of the 22 , 27% , in the total population ) exhibited a definite electron scattering polarization signature , ngc 1948:kwbbe 62 , 92 , 101 , 157 , 326 , and 790 . we note that ngc 1948:kwbbe 92 does show a hint of a small polarization balmer jump ; however , we opted to describe its polarization more conservatively , i.e. as having an electron scattering signature . we found 4 of the 22 stars ( 18% ) had polarimetric properties which suggested that they were unlikely to be classical be stars . as previously discussed , 3 of these stars , ngc 1948:kwbbe 98 , 183 , and 246 exhibited intrinsic polarization signals which seemed to follow a serkowski - like wavelength dependence . ngc 1948:kwbbe 91 was also deemed unlikely to be a classical be star due to the wavelength dependent nature of its polarization position angle . @xcite identified 67 candidate be stars associated with the lmc cluster ngc 2004 and its surrounding field . we extracted polarimetric information for 43 of these candidates , and found that 9 of the 43 ( 21% ) , namely ngc 2004:kwbbe 50 , 87 , 96 , 103 , 106 , 152 , 211 , 347 , 377 , showed an intrinsic polarization balmer jump . 28 of the 43 candidates ( 65% ) were classified as type-2 objects , and note that 10 of these 28 type-2 objects ( i.e. 10 of the 43 , 23% , in the total population ) , ngc 2004:kwbbe 91 , 203 , 276 , 323 , 344 , 441 , 624 , 717 , 1175 , and 1421 exhibited an electron scattering polarization signature . we found 6 of the 43 ( 14% ) candidates in ngc 2004 were unlikely to be classical be stars based on their intrinsic polarization . of these objects , we note that the 3 filters of polarization extracted for ngc 2004:kwbbe 1315 show suggestive evidence of a polarization position angle flip , a feature which is not expected in classical be star - disk systems . wb06 identified 50 candidate be stars in the lmc cluster lh 72 , and designated 11 of these detections as tentative . we were able to obtain polarimetric information for 34 of these stars : 1 of the 34 ( 3% ) , lh 72:wbbe 5 , showed a polarization balmer jump and is most likely a bona - fide classical be star . we classified 22 of the 34 ( 65% ) stars as type-2 objects , and found 5 of these 22 type-2 objects ( i.e. 5 of the 34 , 15% , of the total population ) , lh 72:wbbe 13 , 15 , 26 , 27 , and 33 exhibited an electron scattering polarization signature . we suggest that 11 of the 34 ( 32% ) of candidates in lh 72 are unlikely to be classical be stars based on their intrinsic polarization properties . of these unlikely be stars , we found that the intrinsic polarization of lh 72:wbbe 9 followed a serkowski - like wavelength dependence , with a @xmath1 value @xmath13 6000 . we extracted polarimetric information for 7 of the 11 stars designated as possible candidate be stars by wb06 . 5 of these 7 stars were classified as type-2 objects , while 2 of the 7 appear unlikely to be classical be stars based on their intrinsic polarization signatures . wb06 identified 39 candidate be stars in the lmc cluster ngc 1858 , and we were able to extract polarimetric information for 27 of these 39 stars . 3 of the 27 ( 11% ) , ngc 1858:wbbe 6 , 9 , and 20 , exhibited polarization balmer jumps . we classified 13 of the 27 ( 48% ) as type-2 objects , and found 2 of these 13 type-2 stars ( i.e. 2 of the 27 , 7% , of the total population ) exhibited an electron scattering polarization signature . 11 of the 27 ( 41% ) candidates in ngc 1858 appear unlikely to be classical be stars based on their polarimetric signatures . as previously discussed , 2 of these unlikely be stars , ngc 1858:wbbe 3 and 12 had an intrinsic polarization which follow a serkowski - like wavelength dependence , characterized by a long @xmath1 value . 4 of the 39 photometrically identified candidate be stars in ngc 1858 were judged to be possible detections in wb06 : we were able to detect 3 of these 4 polarimetrically . we classified 2 of the 3 to be type-2 objects , and the remaining object , ngc 1858:wbbe 9 , as a type-1 object . @xcite identified 61 candidate be stars associated with the lmc cluster ngc 2100 . we were able to extract polarimetric information for 35 of the 61 candidates ( 57% ) . 8 of the 35 stars ( 23% ) , ngc 2100:kwbbe 79 , 97 , 436 , 619 , 635 , 705 , 770 , and 797 , exhibited polarization balmer jumps and hence are most likely to be classical be stars . 20 of the 35 stars ( 57% ) were classified as type-2 stars , and we found 16 of these 20 type-2 stars ( i.e. 16 of the 35 , 46% , of the total population ) exhibited an electron scattering intrinsic polarization signature . we suggest that 7 of the 35 ( 20% ) candidates in ngc 2100 are unlikely to be classical be stars based on their observed intrinsic polarization signatures . note that the intrinsic polarization of ngc 2100:kwbbe 321 seems to follow a serkowski - like wavelength dependence , while the unique intrinsic polarization signatures of ngc 2100:kwbbe 111 and 219 ( figure [ n2100intrin111 ] ) suggests these objects might be dust - disk systems . @xcite identified 48 candidate be stars in the vicinity of the smc cluster ngc 346 : we have obtained polarimetric information for 33 of these objects . 8 of the 33 objects ( 24% ) , ngc 346:wbbe 85 , 93 , 191 , 236 , 374 , 445 , 468 , and 529 , exhibited polarization balmer jumps and hence are most likely classical be stars . we classified 22 of the 33 stars ( 67% ) as type-2 stars , and found 7 of these 22 type-2 objects ( i.e. 7 of the 33 , 21% , of the total population ) exhibited a clear electron scattering intrinsic polarization signature . we suggest that 3 of the 33 candidates in ngc 346 ( 9% ) are unlikely to be classical be stars based on their polarimetric signatures . although not detected via our polarimetric survey , we note that ngc 346:kwbbe 13 and ngc 346:kwbbe 200 should not be considered to be classical be stars ; the former ( ngc 346:kwbbe 13 ) is the well - known wolf - rayet / lbv hd 5980 and the later ( ngc 346:kwbbe 200 ) has recently been shown to be the fifth known b[e ] in the smc @xcite . wb06 identified 129 candidate be stars in ngc 371 . we have obtained polarimetric information for 73 of these targets , and found 10 of these 73 stars ( 14% ) , ngc 371:wbbe 2 , 3 , 4 , 5 , 6 , 10 , 13 , 18 , 24 , and 31 exhibited polarization balmer jumps , indicating they are most likely classical be stars . we classified 49 of the 73 stars ( 67% ) as type-2 objects , and remark that 6 of these 49 type-2 stars ( i.e. 6 of the 73 , 8% , of the total population ) exhibited a clear electron scattering polarization signature . note that many of the fainter candidate be stars included in our 74 detections were not observed at high signal - to - noise levels ; thus for most of these objects we can only say that to within 3@xmath11 , their polarimetric properties are not inconsistent with those of classical be stars . we suggest that 14 of the 73 ( 19% ) candidates in ngc 371 are unlikely to be classical be stars based on their polarimetric properties . wb06 suggested that 11 of the 130 photometrically identified candidates should be viewed as `` possible detections '' ; however , we were only able to extract polarimetric information for 2 of these 11 objects , ngc 371:wbbe 64 and 87 . we classified both as type-2 stars . we were able to extract polarimetric information for 18 of the 60 candidate be stars identified photometrically by wb06 . we found 1 of the 18 ( 6% ) , bruck 60:wbbe 6 , showed a polarization balmer jump , while we classified 11 of the 18 stars ( 61% ) as type-2 objects . 8 of these 11 type-2 stars ( i.e. 8 of the 18 , 44% , of the total population ) exhibited an electron scattering intrinsic polarization signature . we suggest that 6 of the 18 ( 33% ) candidates in bruck 60 are unlikely to be classical be stars based on their polarimetric signatures . wb06 noted that 5 of the 26 photometrically identified candidate be stars in bruck 60 should be considered `` possible detections '' . we were only able to detect 1 of these 5 stars polarimetrically , bruck 60:wbbe 21 , and classified it as a type-2 object which exhibited clear evidence of an electron scattering polarization signature . 23 candidate be stars in ngc 456 were identified in the photometric survey of wb06 , and we were able to extract polarimetric information for 14 of these 23 candidates . although we did not observe any objects with polarimetric balmer jumps , we did classify 9 of the 14 stars ( 64% ) as type-2 objects . 1 of these 9 type-2 stars ( i.e. 1 of the 14 , 7% , of the total population ) , ngc 456:wbbe 1 , exhibited an electron scattering polarization signature . we suggest that 5 of the 14 ( 36% ) candidates in ngc 456 are unlikely to be classical be stars based on their observed intrinsic polarization components . wb06 suggested 1 of ngc 456 s 23 photometrically identified candidate be stars should be viewed as a `` possible detection '' : we marginally detected this object , ngc 456:wbbe 21 , and suggest that it is unlikely to be a classical be star . 30 candidate be stars were photometrically identified in ngc 458 by wb06 , and we were able to extract polarimetric information for 10 of these 30 candidates . none of the 10 detected objects exhibited polarimetric balmer jumps or clear evidence of having an electron scattering signature . we classified 6 of the 10 stars ( 60% ) , and suggest that 4 of the 10 ( 40% ) are unlikely to be classical be stars . we were able to extract polarimetric information for 1 of the 2 initially identified as `` possible detections '' by wb06 , ngc 458:wbbe19 . we classified this star as a type-2 object ; furthermore , we found it exhibited suggestive evidence of a small polarimetric balmer jump , although the data were very noisy . @xcite identified 76 candidate be stars in the vicinity of ngc 330 and we were able to extract polarimetric information for 41 of these objects . while we did not observe any of the detected candidates to have a polarimetric balmer jump , we did classify 24 of the 41 ( 59% ) candidates as type-2 stars , and note that 9 of these 24 type-2 objects ( i.e. 9 of the 41 , 22% , of the total population ) exhibited electron scattering polarimetric signatures . we suggest that 17 of the 41 ( 41% ) candidates are unlikely to be classical be stars based on their intrinsic polarimetric signatures . it is clear from table [ polsummary ] that the percentage of smc and lmc candidate be stars which exhibit polarization balmer jumps , electron scattering signatures , or were deemed unlikely to be classical be stars varies significantly from one cluster to the next . to better assess the global trends of our dataset , we extracted simple statistical mean and median values from table [ polsummary ] , as summarized in table [ polsummary2 ] . from table [ polsummary2 ] , we see that at least 25% of our dataset is populated by objects which appear unlikely to be classical be stars . the presence of such a large number of likely `` contaminants '' is interesting , as it confirms one of the initial hypotheses of this project : it is dangerous to assume , a - priori , that all objects identified as excess h@xmath0 emitters via 2-cd surveys are classical be stars . we thus suggest caution should be exercised when attempting to ascertain the role evolutionary age and/or metallicity play in the development of the be phenomenon purely via the analysis of 2-cd data . assuming these objects are truly `` contaminants '' , we recalculated the statistics of our intrinsic polarization dataset after removing these objects from consideration . the results of this exercise are listed in table [ polsum3 ] . the median prevalence of polarization balmer jumps in our smc / lmc dataset is typically between 20 - 25% ( table [ polsum3 ] ) . polarization balmer jumps appear to be significantly less prevalent in our smc clusters ( 4% ) than in our lmc clusters ( 25% ) ; however , this statistic is strongly influenced by the null detection of such signatures in three smc clusters . since 1989 , the hpol spectropolarimeter @xcite mounted on the university of wisconsin s pine bluff observatory ( pbo ) has been monitoring the optical spectropolarimetric properties of a sample of 73 known galactic be stars , and it provides a statistically significant polarimetric dataset which we will use to aid the interpretation of our results . as yet , most of the @xmath2800 hpol observations of classical be stars have not had their isp components identified and removed ; hence , we can not readily compare their intrinsic polarization characteristics to our data . however , we can determine the prevalence of polarization balmer jumps in the hpol database . data obtained prior to 1995 have been compiled in catalog form by @xcite , and observations characterized by a polarization balmer jump have already been flagged in the catalog . we have similarly analyzed all observations obtained from 1995 - 2004 to identify the presence of polarimetric balmer jumps in these data . we then compiled all hpol observations of individual classical be stars into two categories : a ) classical be stars which had * never * exhibited a polarimetric balmer jump in any pbo observation ; and b ) classical be stars which had exhibited a polarimetric balmer jump in at least one observation . the results of this classification procedure are given in table [ polsum3 ] . we found 31 of the 73 ( 42% ) observed galactic be stars exhibited a polarimetric balmer jump for at least part of the time - frame of the hpol survey . amongst early - type be stars ( o9-b5 ) in the hpol database , 22 of 45 ( 49% ) exhibited polarimetric balmer jumps . note that further restricting the sample to include only o9-b3 type be stars does not appreciably affect the observed prevalence of polarization balmer jumps , as 17 of 40 ( 43% ) exhibited such a signature . in contrast , the prevalence of polarization balmer jumps does seem to be noticeably lower for later - type ( b6-a0 ) be stars , as we found only 8 of 26 ( 31% ) exhibited evidence of a polarimetric balmer jump . the frequency of polarization balmer jumps appears to decrease with metallicity : @xmath245% of galactic be stars in the hpol database , @xmath225% of lmc be stars from the present study , and @xmath24% of smc be stars from the present study exhibit polarization balmer jumps . we suggest several possible explanations for these results . 1 . it is possible that the average disk properties of classical be stars are fundamentally different in the low metallicity environments of the smc and lmc as compared to our galaxy . the presence of intrinsic polarization in classical be stars , as well as its wavelength dependence , is a function of disk inclination angle , disk density , and the effective temperature of disk material . thus the observed lower frequency of polarization balmer jumps in our data suggest either that it might be harder to form massive disk systems in low metallicity environments or that the average disk temperature is higher in these lower metallicity environments , hence decreasing the amount of pre- and post - scattering absorption by neutral h i. 2 . our polarimetric observations provided a one time sampling of lmc and smc candidate be stars , while the hpol survey obtained a larger number of observations , albeit at quasi - random intervals , of a population of known be stars . owing to the variable nature of the be phenomenon , we do not a priori know how often any individual be star will exhibit any of the possible polarization signatures ; hence , the different sampling errors associated with these two databases might explain at least some of the observed differences in the polarization balmer jump frequencies . the statistics presented in table [ polsum3 ] were based on removing a significant number of objects deemed unlikely to be classical be stars from consideration , i.e. type-3 and type-4 objects . recall however that our definition of type-2 objects only required their polarimetric properties to be not inconsistent with those expected from classical be stars . it is possible that a population of `` contaminants '' , either a ) unpolarized objects which are nt pole - on or near pole - on classical be stars ; or b ) non - classical be stars which were not identified as such due to low data quality , still reside within our collection of type-2 objects , hence underestimate our polarization balmer jump fraction frequency . if such a scenario is true , it would further suggest that extreme caution should be exercised when assessing the nature of stellar populations purely via the use of photometric 2-cds . wb06 identified a large number of b - type objects having excess h@xmath0 emission via photometric 2-cds and suggested that these stars be classified as candidate be stars . d. gies ( 2006 , personal communication ) noted that several candidate be stars in the wb06 dataset had colors redder than that expected from classical be stars , and suggested that the frequency of observed blue ( or red ) excess h@xmath0 emitters in the 2-color diagrams presented in wb06 increased with the total number of stars of a given color range . furthermore , he postulated that if these red excess emitters were a ) a manifestation of extremely young stars with their birth disks still intact , b ) red - type giants or supergiants , or c ) false detections owing from the effects of bright background h ii emission ; measuring the prevalence of these excess emitters could serve as an estimate of the number of `` contaminants '' which affect the blue population of excess h@xmath0 emitters , i.e. the candidate be star population . we examined the 2-cd photometry of 6 clusters , common to both this survey and that of wb06 and spanning a wide range of ages ( very young , young , old ; wb06 ) and metallicities ( smc , lmc ) , and determined the frequency of excess red - type h@xmath0 emitters relative to the total number of red - type stars . we restricted our analysis such that the magnitude of ( b - v ) colors used to identify candidate be stars in each cluster ( i.e. @xmath2 -0.3 @xmath12 ( b - v ) @xmath12 0.2 in ngc 458 , wb06 ) matched the range of ( b - v ) colors used to study the frequency of red - type excess emitters ( i.e. 0.2 @xmath12 ( b - v ) @xmath12 0.7 for ngc 458 ) . furthermore , we adopted the same ( r - h@xmath0 ) cutoffs delineating excess emitters from normal stars for each cluster as used by wb06 . the results of this exercise are tabulated in column 2 of table [ photcontam ] . interestingly , with few exceptions , the frequency of potential `` contaminants '' implied by this technique is generally consistent with the minimum rate of contamination derived from our polarization results ( see column 3 of table [ photcontam ] , column 6 of table [ polsummary ] ) , i.e. the candidates we assigned a type 4 classification . clearly this photometric contamination check can not provide a diagnostic on the be classification of individual objects , unlike our polarimetry ; however , we suggest that this technique might provide a rough estimate of the role of contaminants in cases in which detailed follow - up observations have not been made or are unfeasible . one of the interesting results presented in wb06 was the identification of candidate be star in clusters of age 5 - 8 myr , in an abundance similar to the general frequency of the be phenomenon in our galaxy , @xmath217% . however , it was uncertain based on those results whether the detected excess h@xmath0 emitters a ) were true classical be star - disk systems ; b ) were b - type objects which still possessed remnant star - formation disks , i.e. herbig be stars ; or c ) merely had diffuse h@xmath0 nebulosity coincidentally associated with them . our intrinsic polarimetric dataset affords us the opportunity to further constrain the evolutionary status of candidate be stars in four clusters of age 5 - 8 myr , lh 72 , ngc 1858 , ngc 346 , and ngc 371 . as discussed in the introduction , section [ intrinpolsection ] , and references therein , it is well accepted that classical be stars are characterized by geometrically thin gas disks , and the mechanism responsible for producing their intrinsic polarization is electron scattering . herbig be stars are more complex systems in that their circumstellar environments contain both gas and dust . dominant _ polarigenic mechanism in herbig be stars is well established to be scattering by dust , as outlined in the review papers of @xcite , @xcite and @xcite and numerous references therein . while herbig be and classical be stars exhibit some similar observational traits , such as the presence of an intrinsic , often variable , polarization component @xcite , the different scattering mechanisms which characterize herbig be versus classical be stars manifest themselves in several observational manners . the magnitude of linear polarization of classical be stars is typically of order 1% or less @xcite , whereas the average ( @xmath23% ) and maximum ( 14.5% ) polarization observed in herbig ae / be stars is clearly much larger @xcite . more importantly , the different scattering mechanisms which dominate in each of these star - disk systems will produce a different wavelength dependence in these stars intrinsic polarization . the characteristic wavelength dependent intrinsic polarization signature of classical be stars has already been documented in section [ intrinpolsection ] . herbig be stars are not known to exhibit any singular `` characteristic '' intrinsic polarization behavior ( see e.g. @xcite ) ; the grain chemistry , grain size distribution , and relative size and extent of the gas and dust disks / envelopes of individual herbig be systems likely contributes to the heterogeneous mix of observed intrinsic polarization . some typical intrinsic polarization trends are observed in herbig ae / be stars however , most notably rotations ( and dramatic 90@xmath6 reversals ) in the intrinsic polarization position angle @xcite . we remark that several literature works which describe and interpret the wavelength dependence of intrinsic polarization from herbig be stars , and at times attempt to draws parallels with the intrinsic polarization behavior of classical be stars , should be viewed with caution . for example , as noted by @xcite , both @xcite and @xcite employ dubious assumptions in their efforts to constrain the interstellar polarization along the line of sight to their herbig be stars , and as such the wavelength dependence of the `` intrinsic '' polarization they report is likely inaccurate . similarly , the discussion of the wavelength dependence of intrinsic polarization of the herbig be star bd+61 154 in @xcite ( see their figure 2 ) ignores the wavelength dependence of the intrinsic polarization position angle . the @xmath240@xmath6 intrinsic polarization position angle rotation from the b to i filter in these data is inconsistent with the expected signature of a pure gas disk system ; the wavelength dependence of intrinsic polarization from this herbig be star is not similar to that observed for classical be stars . although dust scattering is the dominant polarigenic agent for herbig be stars , it is still possible to detect evidence of the gaseous inner disk component of some of these systems via polarimetry @xcite . for example , @xcite presents intrinsic spectropolarimetry of the herbig be star mwc349a covering the wavelength range @xmath25000 - 10500 ( v , r , i filters ) , which exhibits both line depolarization effects and a paschen polarization jump ( see their figure 7 ) , similar to that observed in the classical be star zeta tau @xcite . @xcite suggest electron scattering as the mechanism responsible for producing these specific features . evidence of the dusty component of mwc349a s composite disk is also evident in these data . specifically , the 30 - 40@xmath6 rotation in the intrinsic continuum polarization position angle , beginning at @xmath26200 and extending to the blue edge of the available data , @xmath25000 indicates the presence of an additional scattering mechanism which becomes more pronounced at shorter wavelengths , i.e. dust scattering . based on the available literature , we are aware of no compelling examples of herbig be stars whose intrinsic polarization magnitude _ and _ position angle fully follow the known ubvri behavior of classical be stars . while neither typical nor likely , we can not rule out the possibility that one could in principle observe a balmer jump signature in the polarization magnitude of _ some _ herbig be stars , as these systems do have composite gas plus dust disks . however , as dust scattering is the dominant polarigenic scattering agent in herbig be stars , it is dubious that the entire ubvri intrinsic polarization magnitude and position angle of such systems would exhibit no evidence of a non - electron scattering component . moreover , as astutely pointed out by our referee , it would be even more unlikely to observe one of our `` type 2 electron scattering ( es ) signatures '' in any herbig be star , unless the dust grain chemistry and size distribution would conspire to produce grey scattering . hence , the ubvri spectral range of the intrinsic polarization data presented in this paper enables us to discriminate deviations from a pure electron scattering signature over a broad spectral range . this allows us to differentiate herbig be stars analogous to mwc 349a , whose gaseous disk dominates a portion of the optical polarimetric regime , and herbig be stars whose dust disks dominate the optical polarimetric regime , from classical be stars . as noted in wb06 , many of the candidate be stars identified in clusters of age 5 - 8 myr reside nearby or even within regions characterized by diffuse nebular emission . we consider whether the observed characteristic polarimetric signatures of classical be stars , both electron scattering and balmer jump polarization features , could merely be artifacts of this background emission . such polarization signatures are only produced when the electron scattering optical depth is @xmath21.0 , i.e. , that present in the innermost disk region of classical be stars . the density of any coincident diffuse h@xmath0 nebulosity present around candidate be stars is vastly insufficient to create such a polarization signature . thus while some of the `` be star '' detections made by h@xmath0 spectroscopic surveys @xcite have been shown to be spurious detections due to the presence of coincident gas @xcite , such diffuse nebulosity can not be responsible for creating the polarimetric signals found in our dataset . further insight into the evolutionary status of candidate be stars in our very young ( 5 - 8 myr ) clusters may be elicited via inspection of their near - ir colors . it has been shown that the fundamentally different composition of classical be disks ( gas ) and herbig ae / be disks ( gas and dust ) will generally result in these objects occupying distinctly different locations in ( j - h ) versus ( h - k ) near - ir 2-color diagrams @xcite , although minor instances of overlap are observed . as such , we cross - correlated all photometrically identified candidate be stars residing in clusters studied in this paper with the 2mass @xcite survey . targets for which we were able to extract full jhk photometric data are listed in table [ ircolor ] . we plot the 2mass jhk photometry of all available sources as a function of our polarimetric classifications in figure [ ir2cd ] ( with error bars ) and figure [ ir2cdnoe ] ( without error bars ) . our type-1 sources which exhibit polarimetric balmer jumps are plotted as red circles , our type-2 sources which exhibit es polarimetric signatures are plotted as green triangles , and all other candidate be stars ( other type-2 sources , type-3 sources , type-4 sources , and polarimetric non - detections ) are plotted as blue squares . for reference , we have also plotted the near - ir colors of 101 known galactic be stars tabulated by @xcite as filled black triangles , 21 lmc `` elhc '' stars which @xcite suggest are likely to be classical be stars as yellow open triangles , 2mass and ground - based @xcite colors of 13 magellanic cloud b[e ] stars known to have dusty disks as light blue crosses , and the star `` elhc-7 '' which @xcite suggest is likely to be a lmc herbig ae / be dusty star - disk system as a pink triangle . figures [ ir2cd ] and [ ir2cdnoe ] clearly demonstrate that most of the candidate be stars which we were able to correlate with the 2mass catalog do lie in a distinctly different region of the jhk 2-color diagram than that occupied by known dustier ( herbig ae / be or b[e ] ) systems . furthermore , although the magellanic cloud candidate be stars colors exhibit a wide dispersion , likely due in part to the considerable photometric errors of these faint sources ( figure [ ir2cd ] ; table [ ircolor ] ) , we find no clear distinction amongst the near - ir colors of our candidate be stars as a function of their polarimetric classification . a small number of candidate be stars marginally overlap with the location of dustier b[e ] sources , although these sources are within 3-@xmath11 of the mean locus of classical be colors . clearly better quality near - ir photometry of these sources should be pursued to better elucidate their evolutionary status . two candidate be stars clearly exhibit near - ir colors consistent with dusty disk systems , ngc 346:kwbbe 200 and ngc 456:wbbe 7 . ngc 346:kwbbe 200 was not detected in our polarimetric survey owing to contamination from a nearby source ; however , @xcite has recently published optical spectroscopic , near - ir photometric , and ir photometric observations of the target and concluded it is a b[e ] supergiant , not a classical be star . ngc 456:wbbe 7 was observed in our polarimetric survey and assigned a classification of type-3.5 . although noisy , our intrinsic polarization data exhibited marginal evidence of a position angle rotation , which could be indicative of a dusty envelope . the star s near - ir colors seems to verify this tentative interpretation . for comparison , we also plotted the same near - ir colors for only candidate be stars residing in our very young ( 5 - 8 myr ) clusters in figure [ ir2cdvy ] . comparison of figures [ ir2cdnoe ] and [ ir2cdvy ] reveal no distinctive differences ; rather , it is quite clear that the vast majority of candidate be stars in these very young clusters have near - ir colors consistent with those expected from classical be stars and notably inconsistent with those expected from composite gas plus dust disk systems ( herbig ae / be and b[e ] stars ) . similar to figure [ ir2cdnoe ] , figure [ ir2cdvy ] reveals that there is no clear color difference amongst our various polarimetric classifications . we interpret these results as additional evidence that the many of the candidate be stars identified in clusters 5 - 8 myr old by wb06 are classical be stars , and not pre - main - sequence herbig be stars . the clearing time - scale of natal star formation material ( gas and dust ) will influence whether it is possible or likely , a - priori , to observe early - type herbig star - disk systems in clusters of age 5 - 8 myr . amongst the less massive herbig ae and t tauri stars , the median lifetime of inner optically thick accretion disks is 2 - 3 myr , and although exceptions do exist , there is little to no h- and k - band evidence of primordial disks beyond a median stellar age of 5 myr @xcite . the typical natal disk dissipation for more massive stars is less well known . from a detailed study of the young cluster ngc 6611 , @xcite suggested that the clearing of natal disk material around the more massive stars in this cluster was typically 0.1 myr or less . similarly , @xcite recently suggested that b0e star hd 53367 is a classical be star , based in part on an observed episodic loss of its disk material , a phenomenon known to characterize many classical be stars @xcite . @xcite also identified a 4 - 5 m@xmath17 pre - main - sequence binary companion which , given the pre - main - sequence evolutionary tracks of @xcite , suggests the system s age is @xmath12 0.8 myr and indicates the natal gas plus dust disk of the b0e star hd 53367 must have dissipated at least on this time - scale . although clearly not conclusive , these results do suggest that it is reasonable to expect that most of the natal disks will have been cleared from the early b - type stars in our clusters of age 5 - 8 myr . wb06 and @xcite identified a large population of b - type stars in clusters of age 5 - 8 myr which exhibited an excess of h@xmath0 emission . observational studies suggest that the natal star formation disks of early b - type stars can clear on timescales of at least 0.1 - 0.8 myr , although it is admittedly not clear if this timescale is `` typical '' for all early - type b stars . the near - ir colors of many of the excess h@xmath0 emitters in these very young clusters are consistent with those expected from gaseous classical be stars and generally inconsistent with those observed from dustier b[e ] and/or herbig ae / be disk systems . the observed wavelength dependence of intrinsic polarization of many h@xmath0 excess stars in clusters of age 5 - 8 myr are also consistent with that expected from pure gas disk systems ( classical be stars ) and exhibit no evidence of secondary contributions from dust scattering , as would be expected for composite disk systems ( herbig ae / be , b[e ] stars ) . based on these factors , we suggest that there is compelling evidence supporting the existence of classical be stars in clusters of age 5 - 8 myr . as outlined and summarized in section [ polconcl ] , data presented in this paper support the suggestion of wb06 that classical be stars do develop before the mid - point main sequence lifetime of a b star . specifically , these data support the presence of a significant number of classical be stars , having crude photometric spectral types of b0-b5 ( wb06 ) , in clusters spanning a near - continuous range of ages from log(t ) of 6.7 - 8.1 , i.e. 5 - 126 myr . these results conflict with claims made by @xcite and @xcite that the be phenomenon develops in the second - half of a b star s main sequence lifetime . @xcite do suggest that massive be stars may briefly appear near the zams in the smc and lmc , before losing their `` be status '' until the very end of the first part of their main sequence lifetimes . we note that this conclusion was essentially based on a single cluster population , which limits its robustness . wb06 identified , and the intrinsic polarization presented in this paper support , the detection of a large body of magellanic cloud be stars having crude , photometrically - assigned spectral types of b0-b5 , in clusters spanning a range of ages from 5 - 126 myr . no abrupt absence of classical be stars was observed in any of the younger - type clusters of our photometric and/or polarimetric datasets . thus , our larger , more complete investigation of be stars in magellanic cloud clusters calls into question the suggestion that the be phenomenon is primarily restricted to the second - half of the main - sequence lifetime of b - stars @xcite and contradicts the appearance - disappearance - reappearance of the be phenomenon suggested by @xcite . wb06 first reported the presence of candidate classical be stars in clusters as young as 5 myr , and noted that if such objects were truly classical be stars , they would not have spent enough time on the main sequence to spin - up to near - critical rotation velocities via the mechanism proposed by @xcite or via mass transfer in a binary system @xcite . the confirmation of the classical be status of many of these stars via the present study can be interpreted as evidence that a significant number of classical be stars emerge onto the zero - age - main - sequence at near critical rotation velocities . an alternate interpretation of our results is that these youthful classical be stars are evidence of the existence of a subset of the be phenomenon which rotate at significantly sub - critical rates , perhaps as low as 0.4 - 0.6 v@xmath4 for early - type b stars @xcite . @xcite suggested that only a subset of early - type ( o7-b2 ) be stars might be sub - critical rotators , while later - type objects ( b3-a0 ) should be all near - critical rotators . our photometric survey of these very young clusters assigned crude spectral types to these candidate be stars of b0-b5 ( wb06 ) ; however , owing to detection biases in our follow - up polarimetric observations , we have in general only been able to confirm that portions of the brightest of these candidates ( i.e. the earliest spectral sub - types , b0 to @xmath2b3 ) are bona - fide classical be stars . thus our present observations do not rule out the possibility that some of our very young bona - fide classical be stars might belong to the sub - critical rotation population predicted by @xcite . the observed emergence of the be phenomenon earlier in the main sequence lifetime than previously thought also has important implications regarding the role of magnetic fields in the formation of be disks @xcite . @xcite investigated the transport of magnetic flux tubes in 9 m@xmath18 stars ( @xmath2b2.5-b3 ) and found these structures could rise from the core to the surface within the ( at the time ) expected beginning of the be phenomenon , the mid - point main sequence lifetime . the present study has confirmed the presence of bona - fide classical be stars in clusters as young as 5 myr having crude , photometrically assigned spectral types of b0 to @xmath2b3 . although wb06 suggested the presence of an additional population of b4-b5 type candidate be stars in these clusters , the limited dynamical range probed by the present study was unable to confirm the status of these fainter candidates . although the magnetic flux transport model of @xcite had sufficient time to transport flux to the stellar surface for @xmath2b3 stars , this is unlikely to be the case for later type stars ( @xmath19b4 ) which have a much thicker radiative envelope . hence , the existence of such later - type be stars in extremely young clusters would likely require an additional mechanism to be employed to accelerate the rise times of magnetic flux ( cassinelli , 2006 personal communication ) . while challenging , follow - up spectroscopy of these young cluster populations would provide one avenue confirm the presence of these purported @xcite later - type systems . while the present study provides significant advances in identifying and understanding the biases present in earlier 2-cd studies of cluster populations , additional work is clearly required . our polarimetric survey was not sensitive to pole - on or near - pole - on systems , hence determining which of our `` type-2 '' objects are near - pole - on classical be stars and which are astrophysical objects of a fundamentally different nature is a high priority . we suggest that followup moderate resolution optical spectroscopic or infrared photometric observations would provide reasonable diagnostics to resolve this bias , and enable more quantitative determinations of the bona - fide classical be content of clusters as a function of age and/or metallicity . our data suggest that at least 25% of photometrically identified candidate be stars appear unlikely to be true classical be stars ; however , we are unable to place firm constraints on the true astrophysical nature of such objects . we suggest that followup infrared photometric and/or spectroscopic observations would be useful to further constrain the evolutionary nature of these objects . furthermore , such observations would also be useful to ascertain the true nature of candidate be stars which were too faint to be reliably probed by our polarimetric observations . our analysis of the intrinsic polarization properties of lmc and smc classical be stars suggests that the fundamental disk properties of classical be stars may depend on metallicity . modeling followup moderate resolution spectropolarimetric observations of lmc / smc classical be stars is a clear avenue one could use to further investigate these results , and we note that the advent of accurate linear spectropolarimeters on large aperture telescopes ( i.e. the focal reducer spectrograph ( fors ) at the very large telescope ( vlt ) or the robert stobie spectrograph at the south african large telescope ( salt ) ) make obtaining such observations feasible . we have used imaging polarimetric observations of six lmc and six smc clusters to investigate the evolutionary status of b - type stars identified as excess h@xmath0 emitters via optical 2-color diagram photometric techniques . we characterized the interstellar polarization components of these data in a systematic manner , allowing us to isolate the intrinsic polarization properties of our data , hence constrain the dominant polarigenic agent in each system . we found : 1 . the interstellar polarization associated with ngc 330 was characterized by @xmath1 @xmath24500 , suggesting the presence of a small dust grain population ; 2 . the isp of ngc 2100 exhibited clear evidence of a complex morphology , indicating the presence of a non - uniform magnetic field . we offer a detailed discussion of this b field and its potential origin in @xcite . the ubvri wavelength dependence of _ intrinsic _ polarization of many candidate be stars in clusters of age 5 - 8 myr exhibit either polarization balmer jumps and/or electron scattering signatures , which is the expected diagnostic of free - free scattering from a gaseous disk . no evidence of a secondary polarigenic agent , i.e. dust scattering , is observed in these system . moreover , the 2mass near - ir colors of many of these systems are most consistent with the expected colors of pure gas disks , and inconsistent with the observed colors of dustier magellanic cloud disk systems ( herbig ae / be , b[e ] ) . we conclude these data confirm the initial suggestion of @xcite that classical be stars are present in clusters of age 5 - 8 myr , contradicting claims that the be phenomenon only develops in the second half of a b star s main sequence lifetime @xcite , i.e. after 10 myr . 4 . we interpret the observed presence of classical be stars in clusters of age 5 - 8 myr as evidence that a significant population of early - type b stars must emerge onto the zams rotating at near critical velocities . we note however that these results do not exclude the possibility that we are observing the hypothesized subset of classical be stars which rotate at sub - critical velocities @xcite . 5 . comparing the polarimetric properties of our dataset to a similar survey of galactic classical be stars , we find the prevalence of polarimetric balmer jump signatures decreases with metallicity . we speculate that these results might indicate that either it is more difficult to form large disk systems in low metallicity environments , or that the average disk temperature is higher in these low metallicity environments . we find evidence that at least 25% of photometrically identified candidate be stars do not exhibit polarimetric signatures consistent with those expected from classical be stars . these data strongly suggest that caution must be exercised when attempting to correlate the onset of the be phenomenon with evolutionary age and/or metallicity based solely on statistics derived from simple 2-color diagram photometry . we thank joe cassinelli , doug gies , and jennifer hoffman for discussions which enhanced this paper . the efforts of brian babler , ken nordsieck , and the pbo observing crew for supplying the hpol data presented here are greatly appreciated . we thank the noao tac for allocating observing time for this project and jpw thanks noao for supporting his travel to ctio . portions of this work has been supported in part by nasa npp nnh06cc03b ( jpw ) , nasa grsp ngt5 - 50469 ( jpw ) , nasa ltsa nag5 - 8054 ( ksb ) , nsf ast-0307686 ( jeb ) , and fapesp 02/12880 - 0 ( ap ) grants . amm acknowledges travel support by fapesp ; he is also partially supported by cnpq . polarimetry at the university of so paulo ( usp ) is supported by fapesp . this research has made use of the simbad database operated at cds , strasbourg , france , and the nasa ads system . bjorkman , k.s . & meade , m.r . 2005 , in asp conf ser . `` astronomical polarimetry : current status and future directions '' , ed . a. adamson , c. aspin , c.j . davis , & t. fujiyoshi ( asp : san francisco ) 343 , 406 lccccc cluster & location & filter & age & date & exposure time + bruck 60 & smc & u & o & 2002 oct 19 & 1200 + & & b & & 2002 oct 21 & 720 + & & v & & 2002 oct 19 & 300 + & & r & & 2002 oct 19 & 300 + & & i & & 2002 oct 21 & 300 + ngc 330 & smc & u & y & 2001 nov 22 & 1200 + & & b & & 2001 nov 23 & 420 + & & v & & 2001 nov 25 & 300 + & & r & & 2001 nov 25 & 240 + & & i & & 2001 nov 24 & 240 + ngc346 & smc & u & vy & 2001 nov 21 & 720 + & & b & & 2001 nov 22 & 240 + & & v & & 2001 nov 21 & 180 + & & r & & 2001 nov 24 & 180 + & & i & & 2001 nov 23 & 180 + ngc 371 & smc & u & vy & 2002 oct 18 & 1200 + & & b & & 2002 oct 18 & 600 + & & v & & 2002 oct 17 & 420 + & & r & & 2002 oct 17 & 300 + & & i & & 2002 oct 18 & 420 + ngc 456 & smc & u & y & 2002 oct 24 & 1200 + & & b & & 2002 oct 26 & 720 + & & v & & 2002 oct 24 & 420 + & & r & & 2002 oct 26 & 300 + & & i & & 2002 oct 25 & 300 + ngc 458 & smc & u & o & 2002 oct 27 & 1200 + & & b & & 2002 oct 27 & 720 + & & v & & 2002 oct 26 & 300 + & & i & & 2002 oct 27 & 300 + lh 72 & lmc & u & vy & 2002 oct 26 & 1200 + & & b & & 2002 oct 25 & 720 + & & v & & 2002 oct 25 & 420 + & & r & & 2002 oct 27 & 360 + & & i & & 2002 oct 26 & 300 + ngc 1818 & lmc & u & y & 2001 nov 21 & 900 + & & b & & 2001 nov 22 & 240 + & & v & & 2001 nov 21 & 240 + & & r & & 2001 nov 22 & 180 + & & i & & 2001 nov 22 & 300 + ngc 1858 & lmc & u & vy & 2002 oct 19 & 1200 + & & b & & 2002 oct 24 & 720 + & & v & & 2002 oct 18 & 420 + & & r & & 2002 oct 21 & 300 + & & i & & 2002 oct 24 & 300 + ngc 1948 & lmc & u & y & 2001 nov 27 & 1200 + & & b & & 2001 nov 26 & 360 + & & v & & 2001 nov 25 & 180 + & & r & & 2001 nov 26 & 240 + & & i & & 2001 nov 26 & 300 + ngc 2004 & lmc & u & y & 2001 nov 25 & 900 + & & b & & 2001 nov 22 & 180 + & & v & & 2001 nov 27 & 180 + & & i & & 2001 nov 22 & 420 + ngc 2100 & lmc & u & y & 2001 nov 24 & 1200 + & & b & & 2001 nov 23 & 240 + & & v & & 2001 nov 23 & 180 + & & r & & 2001 nov 24 & 180 + & & i & & 2001 nov 23 & 180 + lcccccccccc cluster & % p@xmath20 & % p@xmath21 & % p@xmath22 & % p@xmath23 & % p@xmath24 & pa & p@xmath16 & @xmath1 & dpa & k + bruck 60 & [email protected]% & [email protected]% & [email protected]% & [email protected]% & [email protected]% & 130 & 0.55% & 5500 & 0 & 0.923 + lh 72 & [email protected]% & [email protected]% & [email protected]% & [email protected]% & [email protected]% & 28 & 0.40% & 5500 & 0 & 0.923 + ngc 330 & [email protected]% & [email protected]% & [email protected]% & [email protected]% & [email protected]% & 126 & 0.53% & 4500 & 0 & 0.737 + ngc 346 & [email protected]% & [email protected]% & [email protected]% & [email protected]% & [email protected]% & 125 & 0.37% & 5500 & 0 & 0.923 + ngc 371 & [email protected]% & [email protected]% & [email protected]% & [email protected]% & [email protected]% & 118 & 0.45% & 5500 & 0 & 0.923 + ngc 456 & [email protected]% & [email protected]% & [email protected]% & [email protected]% & 0.58% & 147 & 0.67% & 5500 & -8 & 0.923 + ngc 458 & [email protected]% & [email protected]% & [email protected]% & & [email protected]% & 120 & 0.39% & 5500 & 0 & 0.923 + ngc 1818 & [email protected]% & [email protected]% & [email protected]% & [email protected]% & [email protected]% & 39 & 0.62% & 5500 & 0 & 0.923 + ngc 1858 & [email protected]% & [email protected]% & [email protected]% & [email protected]% & [email protected]% & 45 & 0.38% & 5500 & 0 & 0.923 + ngc 1948 & [email protected]% & [email protected]% & [email protected]% & [email protected]% & [email protected]% & 35 & 0.68% & 5500 & 0 & 0.923 + ngc 2004 & [email protected]% & [email protected]% & [email protected]% & & [email protected]% & 26 & 0.38% & 5500 & 0 & 0.923 + region i & 0.37 @xmath25 0.15% & 111 & bruck 60 + region ii & [email protected]% & 123 & ngc 330 , 346 , 371 + region iii & [email protected]% & 110 & ngc 456 , 458 + region vi & [email protected]% & 30 & lh 72 + region vii & [email protected]% & 29 & ngc 1818,1858,1948 + region ix & [email protected]% & 20 & ngc 2004 + bruck 60 & [email protected]% & 132@xmath2515 & [email protected]% & 133@xmath2522 & [email protected]% & 151@xmath2523 & [email protected]% & 138@xmath2517 & [email protected]% & 144@xmath2521 + lh 72 & [email protected]% & 28@xmath2526 & [email protected]% & 27@xmath2526 & [email protected]% & 123@xmath2529 & [email protected]% & 18@xmath2590 & [email protected]% & 130@xmath2598 + ngc 330 & [email protected]% & 126@xmath2514 & [email protected]% & 126@xmath2513 & [email protected] & 129@xmath2524 & [email protected]% & 127@xmath2519 & [email protected]% & 128@xmath2528 + ngc 346 & [email protected]% & 125@xmath2524 & [email protected]% & 125@xmath2522 & [email protected]% & 130@xmath2566 & [email protected]% & 126@xmath2527 & [email protected]% & 127@xmath2546 + ngc 371 & [email protected]% & 118@xmath2517 & [email protected]% & 117@xmath2517 & [email protected]% & 111@xmath2532 & [email protected]% & 116@xmath2519 & [email protected]% & 114@xmath2527 + ngc 456 & [email protected]% & 148@xmath2515 & [email protected]% & 149@xmath2515 & [email protected]% & 155@xmath2515 & [email protected]% & 151@xmath2516 & 0.56% & 153.2 + ngc 458 & [email protected]% & 121@xmath2537 & [email protected]% & 121@xmath2528 & [email protected]% & 129@xmath2539 & [email protected]% & 125@xmath2532 & & + ngc 1818 & [email protected]% & 40@xmath2512 & [email protected]% & 40@xmath2512 & [email protected]% & 49@xmath2517 & [email protected]% & 43@xmath2513 & [email protected]% & 45@xmath2516 + ngc 1858 & [email protected]% & 47@xmath2525 & [email protected]% & 48@xmath2519 & [email protected]% & 74@xmath2530 & [email protected]% & 55@xmath2530 & [email protected]% & 63@xmath2532 + ngc 1948 & [email protected]% & 35@xmath2512 & [email protected]% & 36@xmath2510 & [email protected]% & 40@xmath2517 & [email protected]% & 37@xmath2512 & [email protected]% & 38@xmath2515 + ngc 2004 & [email protected]% & 27@xmath2518 & [email protected]% & 28@xmath2518 & [email protected]% & 75@xmath2572 & [email protected]% & 44@xmath2567 & & + lcccccccccc cluster & % p@xmath20 & % p@xmath21 & % p@xmath22 & % p@xmath23 & % p@xmath24 & pa & p@xmath16 & @xmath1 & dpa & k + ngc 2100 area 1 & 1.32% & 1.41% & 1.44% & 1.37% & 1.20% & 139 & 1.45% & 5000 & 0 & 0.83 + ngc 2100 area 2 & 1.26% & 1.35% & 1.37% & 1.31% & 1.14% & 100 & 1.38% & 5000 & 0 & 0.83 + ngc 2100 area 3 & 1.64% & 1.69% & 1.65% & 1.54% & 1.32% & 78 & 1.70% & 4500 & 0 & 0.737 + lcccccc candidate be star & filter & % p & pa & % q & % u & % err + bruck 60:wbbe 1 & u & 1.75 & 107.3 & -1.44 & -0.99 & 0.61 + & b & & & & & + & v & 1.33 & 113.7 & -0.90 & -0.98 & 0.30 + & r & 1.33 & 117.6 & -0.76 & -1.09 & 0.35 + & i & & & & & + lcccccc candidate be star & filter & % p & pa & % q & % u & % err + bruck 60:wbbe 1 & u & 1.46 & 100.9 & -1.36 & -0.54 & 0.61 + & b & & & & & + & v & 0.91 & 104.4 & -0.80 & -0.44 & 0.30 + & r & 0.88 & 110.0 & -0.67 & -0.56 & 0.35 + & i & & & & & + lccccc cluster & detection rate & # pol . bj & # es & not inconsistent & unlikely + & & ( type-1 ) & ( type-2 ) & ( type-2 ) & ( type-3,4 ) + bruck 60 & 18/26 ( 69% ) & 1/18 ( 6% ) & 8/18 ( 44% ) & 11/18 ( 61% ) & 6/18 ( 33% ) + ngc 330 & 41/76 ( 54% ) & 0/41 ( 0% ) & 9/41 ( 22% ) & 24/41 ( 59% ) & 17/41 ( 41% ) + ngc 346 & 33/48 ( 69% ) & 8/33 ( 24% ) & 7/33 ( 21% ) & 22/33 ( 67% ) & 3/33 ( 9% ) + ngc 371 & 73/129 ( 57% ) & 10/73 ( 14% ) & 6/73 ( 8% ) & 49/73 ( 67% ) & 14/73 ( 19% ) + ngc 456 & 14/23 ( 61% ) & 0/14 ( 0% ) & 1/14 ( 7% ) & 9/14 ( 64% ) & 5/14 ( 36% ) + ngc 458 & 10/30 ( 33% ) & 0/10 ( 0% ) & 0/10 ( 0% ) & 6/10 ( 60% ) & 4/10 ( 40% ) + lh 72 & 34/50 ( 67% ) & 1/34 ( 3% ) & 5/34 ( 15% ) & 22/34 ( 65% ) & 11/34 ( 32% ) + ngc 1818 & 18/40 ( 45% ) & 4/18 ( 22% ) & 5/18 ( 28% ) & 12/18 ( 67% ) & 2/18 ( 11% ) + ngc 1858 & 27/39 ( 69% ) & 3/27 ( 11% ) & 2/27 ( 7% ) & 13/27 ( 48% ) & 11/27 ( 41% ) + ngc 1948 & 22/27 ( 81% ) & 6/22 ( 27% ) & 6/22 ( 27% ) & 12/22 ( 55% ) & 4/22 ( 18% ) + ngc 2004 & 43/67 ( 64% ) & 9/43 ( 21% ) & 10/43 ( 23% ) & 28/43 ( 65% ) & 6/43 ( 14% ) + ngc 2100 & 35/61 ( 57% ) & 8/35 ( 23% ) & 16/35 ( 46% ) & 20/35 ( 57% ) & 7/35 ( 20% ) + lcccccc cluster & mean bj & median bj & mean es & median es & mean unlikely & median unlikely + very young & 13% & 13% & 13% & 12% & 25% & 26% + young & 13% & 21% & 22% & 23% & 26% & 20% + smc & 7% & 3% & 17% & 15% & 30% & 35% + lmc & 18% & 22% & 24% & 25% & 23% & 19% + all smc+lmc & 13% & 13% & 21% & 22% & 26% & 26% + lcccc cluster & mean bj & median bj & mean es & median es + very young & 17% & 18% & 17% & 18% + young & 16% & 24% & 28% & 31% + smc & 7% & 4% & 25% & 17% + lmc & 22% & 25% & 31% & 29% + all lmc+smc & 16% & 18% & 28% & 25% + milky way galaxy & 42 - 49% & & & + lccccc star & classification & ( j - h ) & ( h - k ) & ( j - h)@xmath26 & ( h - k)@xmath26 + bruck60:wbbe 5 & type2 - 4,other & 0.09 & 0.04 & 0.14 & 0.24 + bruck60:wbbe 20 & type2 - 4,other & 0.29 & 0.00 & 0.13 & 0.20 + lh72:wbbe 5 & type1:bal - j & 0.07 & 0.21 & 0.10 & 0.15 + lh72:wbbe 15 & type2:es & 0.54 & 0.24 & 0.14 & 0.17 + , greater than 5 are plotted on a stokes q - u diagram ( panel a ) , as a polarization map as a function of pixel location ( panel b ) , and in position angle ( panel c ) and polarization ( panel d ) histograms . as many cluster stars in panels a , c , and d exhibit the same general polarimetric properties , we suggest that most of these objects lack significant intrinsic polarization components and instead only exhibit evidence of interstellar polarization . [ 1818vsel ] ] , with that expected from a classical be star . such objects would receive a designation of type-3 or type-4 , depending on the severity of their deviation from any of the expected polarimetric signatures . these objects ( especially those which receive a type-4 designation ) provide strong evidence that the 2-cd technique can mis - identify classical be star - disk systems . [ polunlikely ] ] '' ( light blue cross ) correspond to 2-mass and ground - based @xcite photometry of known magellanic cloud b[e ] supergiants , and `` mc haebe '' ( pink triangle ) corresponds to the star elhc-7 , which @xcite assert is a herbig ae / be star which resides in the lmc . as discussed in section [ nircolors ] , most of the candidate be stars which we were able to correlate with the 2mass catalog exhibit colors consistent with those expected from classical be stars , and inconsistent with those expected from composite gas plus dust disk systems ( i.e. herbig ae / be or b[e ] stars ) . [ ir2cd ] ] , corresponding to candidate be stars residing in `` very young '' clusters of age 5 - 8 myr , is presented . as discussed in section [ nircolors ] , the data presented in this figure and figure [ ir2cdnoe ] exhibit similar properties . regardless of polarimetric classification type , all candidate be stars in clusters of age 5 - 8 myr appear consistent with the expected colors of classical be stars and inconsistent with the observed colors of composite gas plus dust systems ( i.e. herbig ae / be , b[e ] stars ) . we interpret these data as additional evidence that many of the photometrically identified candidate be stars in clusters 5 - 8 myr old by wb06 are classical be stars and not pre - main - sequence herbig be stars . [ ir2cdvy ] ]
we present the first detailed imaging polarization observations of six smc and six lmc clusters , known to have large populations of b - type stars which exhibit excess h@xmath0 emission , to constrain the evolutionary status of these stars and hence better establish links between the onset of disk formation in classical be stars and cluster age and/or metallicity . we parameterize the interstellar polarization ( isp ) along the lines of sight to these twelve clusters , thereby providing a diagnostic of the fundamental properties of the dust which characterizes their localized interstellar medium . we determine that the isp associated with the smc cluster ngc 330 is characterized by a modified serkowski law with @xmath1 @xmath24500 , indicating the presence of smaller than average dust grains . furthermore , the morphology of the isp associated with the lmc cluster ngc 2100 suggests that its interstellar environment is characterized by a complex magnetic field . removing this interstellar polarization component from our data isolates the presence of any intrinsic polarization ; the wavelength dependence of this intrinsic polarization provides a diagnostic of the dominant and any secondary polarigenic agents present , enabling us to discriminate pure gas disk systems , i.e. classical be stars , from composite gas plus dust disk systems , i.e. herbig ae / be or b[e ] stars . our _ intrinsic _ polarization results , along with available near - ir color information , strongly supports the suggestion of wisniewski et al . that classical be stars are present in clusters of age 5 - 8 myr , and contradict assertions that the be phenomenon only develops in the second half of a b star s main sequence lifetime , i.e. no earlier than 10 myr . our data imply that a significant number of b - type stars must emerge onto the zero - age - main - sequence rotating at near - critical rotation rates , although we can not rule out the possibility that these data instead reveal the presence of a sub - group of the be phenomenon characterized by sub - critically rotating objects . comparing the polarimetric properties of our dataset to a similar survey of galactic classical be stars , we find that the prevalence of polarimetric balmer jump signatures decreases with metallicity . we speculate that these results might indicate that either it is more difficult to form large disk systems in low metallicity environments , or that the average disk temperature is higher in these low metallicity environments . we have characterized the polarimetric signatures of all candidate be stars in our data sample and find @xmath225% are unlikely to arise from true classical be star - disk systems . this detection of such a substantial number `` contaminants '' suggests one should proceed with caution when attempting to determine the role of evolutionary age and/or metallicity in the be phenomenon purely via 2-cd results .
since the famous heisenberg s paper @xcite had been published , uncertainty relations are the subject of long researches @xcite . a new interest was stimulated by recent advances in quantum information processing . there are two well - known approaches to formulating the uncertainty principle . the first was initiated by robertson @xcite who showed that a product of the standard deviations of two observables is bounded from below . here we may run across some disputable topics such as the number - phase case @xcite . both the well - defined hermitian operator of phase and corresponding number - phase uncertainty relation have been fit within the pegg - barnett formalism @xcite . the second approach is generally characterized by posing the uncertainty principle via information - theoretic terms especially via entropies @xcite . although the first relation of such a kind was derived by hirschman @xcite , a general statement of the problem is examined in the papers @xcite . mutually unbiased bases @xcite , the time - energy case @xcite and tomographic processes @xcite have been considered within an entropic approach as well . the most of known uncertainty relations deals with observables or , more generally , with povm measurements . nevertheless , there exist relations for unitary transformations @xcite and non - hermitian annihilation operator @xcite . both the measurement and unitary evolution are rather simplest types of a state change in quantum theory @xcite . the formalism of quantum operations , or super - operators , is now a standard tool for treating quantum processes . in the present work , we address a question how to formulate the uncertainty principle for super - operators . it turns out that one of possible ways is naturally provided with the notion of extremal unraveling of a super - operator . for the shannon entropy , this notion was examined in @xcite . as an entropic measure , we will use the tsallis entropy , which has found use in various physical problems ( see references in @xcite ) , and the rnyi entropy . the paper is organized as follows . in section [ sec2 ] , the properties of the tsallis entropies and super - operators are recalled . for given input state and super - operator , we find the extremal unraveling that minimizes all the tsallis entropies simultaneously . relations between the ensemble entropy and the entropy of a generated density matrix are revisited . in section [ sec3 ] , we derive the uncertainty relation for two generalized resolutions of the identity in terms of tsallis entropies . the previous result on the rnyi entropies is refined . the obtained entropic relations are directly used for the extremal unravelings of two super - operator . section [ sec4 ] concludes the paper with a summary of results . first , we briefly recall the definitions of used entropic measures ( for a discussion and further references , see @xcite ) . for real @xmath1 and @xmath2 , we define the non - extensive @xmath3-entropy of probability distribution @xmath4 by @xcite @xmath5 this can be rewritten as @xmath6 , where @xmath7 is the @xmath3-logarithmic function , defined for @xmath8 , @xmath2 and @xmath9 . the quantity ( [ tsaent ] ) will be referred to as tsallis @xmath3-entropy , though it was previously discussed by havrda and charvt @xcite . in the limit @xmath10 , @xmath11 and the quantity ( [ tsaent ] ) recovers the shannon entropy . we will also use the rnyi @xmath3-entropy defined for @xmath2 as @xcite @xmath12 the rnyi @xmath3-entropy also coincides with the shannon entropy in the limit @xmath10 . there are many various forms of extrapolation between different entropies @xcite . we will only need in the equality @xmath13 which immediately follows from the definitions ( [ tsaent ] ) and ( [ renent ] ) . in the following , some notation of linear algebra will be used . let @xmath14 be finite - dimensional hilbert space . for given two vectors @xmath15 , their inner product is denoted by @xmath16 . for two operators @xmath17 and @xmath18 on @xmath14 , we define the hilbert schmidt inner product by @xcite @xmath19 this inner product naturally induces so - called frobenius norm @xmath20 . the frobenius norm can be re - expressed as @xmath21 , where the singular values @xmath22 are eigenvalues of @xmath23 . the largest singular value of @xmath17 gives the spectral norm @xmath24 of this operator @xcite . a quantum measurement is described by a generalized resolution of the identity ( or by a positive operator - valued measure ) . this is a set @xmath25 of positive semidefinite operators obeying the completeness relation @xmath26 , where @xmath27 is the identity @xcite . consider a linear map @xmath28 that takes linear operators on @xmath14 to linear operators on @xmath29 , and also satisfies the conditions of trace preservation and complete positivity . following @xcite , this map will be referred to as super - operator . each super - operator has a kraus representation , namely @xcite @xmath30 where the kraus operators @xmath31 map the input space @xmath14 to the output space @xmath29 and obey @xmath32 ( the preservation of the trace ) . representations of such a kind are never unique @xcite . for given super - operator @xmath28 , there are many sets @xmath33 that enjoy ( [ opsum ] ) . in the paper @xcite , each concrete set @xmath33 resulting in ( [ opsum ] ) is called an unraveling of the super - operator @xmath28 . this terminology is due to carmichael @xcite who introduced this word for a representation of the master equation . it is well - known that two kraus representations of the same super - operator are related as @xmath34 where the matrix @xmath35 $ ] is unitary @xcite . [ we assume in ( [ eqvun ] ) that unravelings @xmath36 and @xmath37 have the same cardinality by adding zero operators , if needed . ] for given density operator @xmath38 and super - operator unraveling @xmath39 , we introduce the matrix @xmath40\equiv [ [ { { \rm{tr}}}({\mathsf{a}}_i^{\dagger}{\mathsf{a}}_j\rho ) ] ] \ . \label{pimdef}\ ] ] the diagonal element @xmath41 is clearly positive and gives the @xmath42th effect probability . then the entropies @xmath43 and @xmath44 are merely defined by ( [ tsaent ] ) and ( [ renent ] ) respectively . by definition , the matrix @xmath45 is hermitian . suppose that the two sets @xmath33 and @xmath46 fulfill ( [ eqvun ] ) . using the properties of the inner product , we have @xmath47 or @xmath48 as the matrix relation . that is , if the sets @xmath36 and @xmath37 are both unravelings of the same super - operator then the matrices @xmath45 and @xmath49 are unitarily similar . due to hermiticity , all the matrices of a kind @xmath45 assigned to one and the same super - operator are unitarily similar to a unique ( up to permutations ) diagonal matrix @xmath50 , where the @xmath51 s and perhaps zeros are the eigenvalues of each of these matrices . so any @xmath45 is positive semidefinite . for given unraveling @xmath33 , we obtain the concrete matrix @xmath45 and diagonalize it through a unitary transformation @xmath52 . let us define a specific unraveling @xmath53 related to given @xmath36 as @xmath54 where the unitary matrix @xmath55 diagonalizes @xmath45 . it turns out that the unraveling ( [ excalc ] ) enjoys the _ extremality _ property with respect to all the tsallis @xmath3-entropies for @xmath56 and the rnyi @xmath3-entropies for @xmath57 . [ extr ] for given density operator @xmath38 and super - operator @xmath28 , each unraveling @xmath36 of the @xmath28 satisfies @xmath58 where the extremal unraveling @xmath53 is defined by the formula ( [ excalc ] ) . * proof * we firstly note that @xmath59 with the probabilities @xmath60 of effects . in view of @xmath61 , the probabilities @xmath41 of different effects of the unraveling @xmath36 is related to the @xmath60 s by @xmath62 where @xmath63 . the matrix @xmath64 $ ] is unistochastic , whence @xmath65 for all @xmath66 and @xmath67 for all @xmath42 . we now use these relations and an obvious fact that for @xmath1 the function @xmath68 is concave in the range @xmath69 $ ] . according to jensen s inequality ( see , e.g. , the book @xcite ) , there holds @xmath70 this completes the proof for ( [ tsaex ] ) . further , we note that for @xmath71 the function @xmath72 is increasing in the range @xmath73 . combining the equality ( [ rentsa ] ) with ( [ tsaex ] ) then gives ( [ renex ] ) . @xmath74 for the shannon entropy , a problem of minimal unraveling was considered in @xcite and later in @xcite . diagonalizing the matrix @xmath45 is actually equivalent to the extreme condition that has been derived by the method of lagrange s multipliers in @xcite . latter reasons local in spirit are quite complemented by the above proof based on the concavity . our treatment allows to find the extremal unraveling easily from ( [ excalc ] ) . thus , for prescribed state @xmath38 all the tsallis entropies of order @xmath56 and the rnyi entropies of order @xmath57 are minimized by one and the same unraveling of given super - operator . however , this unraveling does not minimize other rnyi entropies in general . an unraveling extremal with respect to rnyi s entropy of order @xmath75 may also be dependent on @xmath3 . a search for such an unraveling seems to be difficult since rnyi s entropy of such an order is not purely convex nor purely concave . the relations ( [ tsaex ] ) and ( [ renex ] ) can be put in the context of a state ensemble having a prescribed density operator . in line with ( [ tsaent ] ) and ( [ renent ] ) , we introduce the quantum tsallis and rnyi entropies of a density matrix @xmath38 by @xmath76 in the limit @xmath10 both the expressions coincides with the von neumann entropy @xmath77 . let @xmath78 be an ensemble of normalized pure states , @xmath79 and @xmath80 , that leads to the density operator @xmath38 , namely @xmath81 in general , the states @xmath82 are not mutually orthogonal . the right - hand side of ( [ ensres ] ) poses the spectral decomposition of @xmath38 , so that the vectors @xmath83 form an orthonormal basis in @xmath14 . the ensemble classification theorem says that @xcite @xmath84 for some unitary matrix @xmath85 $ ] . it follows from ( [ unfr ] ) and @xmath86 that @xmath87 , where @xmath88 are elements of a unistochastic matrix . changing notation in the formula ( [ jens ] ) appropriately , we finally get @xmath89 in the limit @xmath10 , both the inequalities are reduced to @xmath90 . the last is a known relation between the von neumann entropy of the generated density operator and the shannon entropy of an ensemble @xcite . for the tsallis entropies this result can be proceeded to ensembles of mixed states . let an ensemble @xmath91 of normalized density operators , @xmath92 and @xmath80 , give rise to a density operator @xmath93 where the spectral decomposition @xmath94 . applying the first inequality of ( [ concpur ] ) to the right - hand side of ( [ ensmix ] ) , we have @xmath95 here we used the identity @xmath96 . adding a consequence of concavity of the function @xmath68 , we finally write @xmath97 the inequality on the left can be shown as follows . for any concave function @xmath98 , the functional @xmath99 is also concave on hermitian @xmath17 ( for details , see section iii in @xcite ) . we further note that @xmath100 . in the limit @xmath10 , the inequalities ( [ concmix ] ) are reduced to the well - known bounds on the von neumann entropy of a state mixture ( for instance , see @xcite ) . it seems that such a treatment fails in the case of rnyi s entropies . indeed , the tsallis @xmath3-entropy does enjoy so - called strong additivity of degree @xmath3 , whereas the rnyi @xmath3-entropy does not @xcite . to obtain entropic relations , we will use a version of riesz s theorem ( see theorem 297 in the book @xcite ) . let @xmath101 be @xmath102-tuple of complex numbers @xmath103 and let @xmath104 $ ] . define @xmath105 to be maximum of @xmath106 , i.e. @xmath107 . the fixed matrix @xmath108 describes a linear transformation @xmath109 . that is , to each @xmath110 we assign @xmath111-tuple @xmath112 with elements @xmath113 for @xmath114 , the @xmath115 norm of vector @xmath110 is @xmath116 . for @xmath117 , we use a like function @xmath118 of probability distribution @xmath119 , though it is not a norm for @xmath120 . riesz s theorem is formulated as follows . [ riesz ] suppose the matrix @xmath108 satisfies @xmath121 for all @xmath101 , @xmath122 and @xmath123 ; then @xmath124 we will now obtain an improved version of the statement emerged in the paper @xcite . to each generalized resolution @xmath125 and given mixed state @xmath38 , we assign the probabilistic vector @xmath126 with elements @xmath127 . another resolution @xmath128 is assigned by the vector @xmath129 with elements @xmath130 . [ resol ] for any two resolutions @xmath125 and @xmath128 of the identity and given density operator @xmath38 , there holds @xmath131 where @xmath132 , @xmath133 and the function @xmath134 is defined by @xmath135 * proof * we first consider the case when both the resolutions @xmath125 and @xmath128 are orthogonal . for those values of labels @xmath42 and @xmath66 that satisfy @xmath136 and @xmath137 , we put ( generally non - hermitian ) operators @xmath138 it is clear that @xmath139 and @xmath140 . since the resolutions @xmath25 and @xmath141 are orthogonal , we further have @xmath142 and @xmath143 . the matrix elements of transformation @xmath108 are then defined by @xmath144 we now rewrite ( [ lintrans ] ) as @xmath145 , where @xmath146 by definition . due to @xmath147 and @xmath148 , where @xmath149 for all @xmath42 , we have @xmath150 and the precondition of lemma [ riesz ] herewith . so we apply ( [ suppos ] ) to the special values @xmath151 and @xmath152 . by the completeness relation , @xmath153 whence @xmath154 , i.e. the values @xmath155 are related to @xmath156 via the transformation @xmath108 too . since the operators @xmath157 and @xmath158 are projective , @xmath159 and @xmath160 , whence @xmath161 and @xmath162 with @xmath163 , @xmath164 . the statement of lemma [ riesz ] then results in the inequality @xmath165 , under the conditions @xmath132 and @xmath133 . noting that the maximum of modulus of ( [ tijdefm ] ) is actually @xmath134 , we resolve the case when the resolutions @xmath166 and @xmath167 are both orthogonal . to generalize ( [ eqresol ] ) to the case of arbitrary two resolutions , we will use the method proposed in @xcite and further developed in @xcite . in the extended space @xmath168 , the resolutions @xmath166 and @xmath167 are realized as new resolutions @xmath169 and @xmath170 respectively , and the @xmath169 is now orthogonal by naimark s extension ( see , e.g. , sect . 5.1 in @xcite ) . it can be made in such a way that for any density matrix @xmath38 on @xmath14 , @xmath171 , @xmath172 , and @xmath173 ( the @xmath174 is built from @xmath38 by adding zero rows and columns ) . hence we have the same values of entropies @xcite and @xmath175 . by a similar extension of @xmath170 , the question is quite reduced to the above case of two orthogonal resolutions . @xmath74 using simple algebra ( see the proof of proposition 3 in @xcite ) , the inequality ( [ eqresol ] ) can be rewritten as @xmath176 for a pure state , this relation in terms the rnyi entropies coincides with the one deduced in the previous work @xcite . with respect to the spectral decomposition @xmath177 , we define @xmath178 where the probabilities @xmath179 and @xmath180 . the above inequality with @xmath181 instead of @xmath134 was obtained in @xcite . for an impure state , we have @xmath182 and a stronger bound in ( [ renuim ] ) . for canonically conjugate variables , the uncertainty relations in terms of rnyi s entropies were given in @xcite . let us proceed to a relation with the tsallis entropies . the sum @xmath183 can not be arbitrarily small because of the constraint ( [ eqresol ] ) . we rewrite this sum as @xmath184 in terms of the variables @xmath185 , @xmath186 and the function @xmath187 . assuming @xmath188 , we obviously have @xmath189 and @xmath190 . adding ( [ eqresol ] ) in the form @xmath191 , where @xmath192 , we have arrived at a task of minimizing @xmath187 under the above conditions . it is important here that @xmath193 and , therefore , @xmath194 . indeed , the cauchy - schwarz inequality for the hilbert - schmidt inner product directly shows that the modulus of ( [ tijdefm ] ) does not exceed one . the task is solved in appendix [ app1 ] , and the sum @xmath183 can not be less than @xmath195 where we used @xmath196 due to @xmath132 . by @xmath197 , we claim the following . [ tsasol ] for any two resolutions @xmath125 and @xmath128 of the identity and given density operator @xmath38 , there holds @xmath198 where @xmath132 and @xmath199 . the inequality ( [ eqtsasol ] ) gives the uncertainty relation in terms of the tsallis entropies for two generalized measurements . in spirit and origin , it is like to the relation via the rnyi entropies ( [ renuim ] ) . note that relations in terms of the tsallis entropies have been considered for particular cases of the position and momentum @xcite and the spin-1/2 components @xcite . it is of some interest to get a state - independent version of ( [ eqtsasol ] ) . since the function @xmath200 is increasing for @xmath201 , @xmath202 and @xcite @xmath203 we have the state - independent bound @xmath204 note that @xmath205 is provided by @xmath206 . here the inequality on the left is a cauchy - schwarz inequality for ordinary matrix products and the spectral norm ( see , e.g. , the result ( 4.50 ) in @xcite ) and the inequality on the right follows from the completeness relation . the inequality @xmath205 is saturated if and only if for some two elements @xmath207 , @xmath208 and nonzero @xmath209 there hold @xmath210 and @xmath211 simultaneously . in other words , @xmath212 implies that there exist those operators @xmath213 and @xmath214 that act as commuting projectors in a nonempty subspace @xmath215 . otherwise , the entropic bound in the relation ( [ eqtsind ] ) is nontrivial . so , we have extended the relation , conjectured in @xcite and later proved in @xcite , to the tsallis entropies and general quantum measurements . let us consider two concrete examples of specific interest . [ fdisc ] the first example is a pair of complementary observables in a @xmath0-level system ( for the rnyi formulation , see @xcite ) . let complex amplitudes @xmath216 and @xmath217 be connected by the discrete fourier transform @xmath218 and the corresponding probability distributions @xmath219 and @xmath220 . the transformation ( [ discf ] ) is the canonical example that leads to complementary observables @xcite . it follows from @xmath221 and ( [ suppos ] ) that @xmath222 where @xmath122 and @xmath123 . squaring , we get @xmath223 , @xmath224 under the conditions on @xmath3 and @xmath225 from lemma [ resol ] . in its symmetric form , the uncertainty relation in terms of the tsallis entropies is written as @xmath226 , where @xmath132 and @xmath199 . [ angl ] the angle @xmath227 and the angular momentum @xmath228 can similarly be treated . taking one and the same size @xmath229 for all the angular bins ( i.e. the ratio @xmath230 is a strictly positive integer ) , we introduce probabilities @xmath231 where the coefficients @xmath217 s are related to the expansion @xmath232 , with respect to the eigenstates of the @xmath228 . using theorem 192 of the book @xcite for integral means and assuming @xmath233 , we have @xmath234 and the inversed inequality with @xmath3 instead of @xmath225 . summing these inequalities with respect to @xmath235 and then raising them to the powers @xmath236 and @xmath237 respectively , we finally write @xmath238 where the norm @xmath239 and @xmath240 , @xmath241 . combining the relations ( [ cor192 ] ) with the young - hausdorff inequalities ( see , e.g. , section 8.17 in @xcite ) , which are viewed in our notation as @xmath242 we get @xmath243 , @xmath244 . so , there holds @xmath245 , where @xmath132 and @xmath199 . when @xmath246 , this new inequality in terms of tsallis entropies coincides with the relation in terms of rnyi s entropies deduced in @xcite . finally , we consider the case of extremal unravelings within the general formulation ( [ eqtsasol ] ) . for the extremal unravelings @xmath53 and @xmath247 of the super - operators @xmath248 and @xmath249 , we obtain the entropic uncertainty relation @xmath250 where the @xmath251 is put by ( [ gfdef ] ) with @xmath252 , @xmath253 in terms of kraus operators @xmath254 , @xmath255 . we can also write the uncertainty relation for unravelings extremal with respect to the rnyi entropies . using ( [ renuim ] ) , for @xmath75 we obtain the relation @xmath256 where @xmath257 denotes the unraveling of @xmath248 extremal for the rnyi entropy of order @xmath75 . this unraveling differs from the one given by ( [ excalc ] ) and also depends on the parameter @xmath3 in general . a search of explicit analytic expression for @xmath257 seems to be complicated , because concavity ( convexity ) things can not be used here . nevertheless , the entropic relation in terms of the rnyi entropies holds , as a mathematical inequality at least . we have considered the unraveling ( i.e. the concrete set of kraus operators ) of a super - operator that is extremal with respect to all the tsallis entropies of a positive order and the rnyi entropies of an order @xmath258 . this general result is formally posed in theorem [ extr ] . if one of unravelings is given explicitly , then this minimizing unraveling is easily calculated by diagonalizing a certain hermitian matrix . the known relation between the shannon entropy of an ensemble of pure states and the von neumann entropy of the risen density operator is extended to both the tsallis and rnyi entropies of the mentioned orders . the case of tsallis entropy allows further extension to a mixture of density operators ( see the bounds ( [ concmix ] ) ) . due to riesz s theorem , there exists an inequality between certain functions of the probability distributions generated by two resolutions of the identity ( see lemma [ resol ] ) . this inequality gives an origin for the uncertainty relations , given by ( [ renuim ] ) for the rnyi entropies and by theorem [ tsasol ] for the tsallis entropies . the general formulation ( [ eqtsasol ] ) is a new result in the topic . it has been illustrated within the two interesting cases ( see examples [ fdisc ] and [ angl ] ) . in the formulae ( [ enunun ] ) and ( [ renunn ] ) , both the entropic uncertainty relations are naturally recast for the extremal unravelings of two given trace - preserving super - operators . to obtain a lower bound on the sum of tsalis entropies , we find the minimal value of the function ( [ sumhh ] ) in the domain @xmath259 such that @xmath260 , @xmath261 , @xmath191 . when @xmath262 , the curve @xmath263 cuts off the down right corner of the rectangle @xmath264 and herewith the point @xmath265 in which @xmath266 ( see fig . 1 ) . in the interior of @xmath259 , we have @xmath267 due to @xmath188 . so the minimum is reached on the boundary of the domain @xmath259 . using ( [ partder ] ) , the task is merely reduced to minimizing @xmath187 on segment of the curve @xmath268 between the point @xmath269 , where @xmath270 , and the point @xmath271 . substituting @xmath272 in ( [ sumhh ] ) and differentiating with respect to @xmath273 , we obtain @xmath274 where we used @xmath275 because of @xmath132 . when @xmath276 , the quantity ( [ deriv ] ) is strictly positive for @xmath277 and the minimal value is @xmath278 too . rnyi a 1961 _ on measures of entropy and information _ , proceedings of the 4th berkeley symposium on mathematical statistics and probability , pp . 547 - 561 ( university of california press : berkeley los angeles )
a way to pose the entropic uncertainty principle for trace - preserving super - operators is presented . it is based on the notion of extremal unraveling of a super - operator . for given input state , different effects of each unraveling result in some probability distribution at the output . as it is shown , all tsallis entropies of positive order as well as some of rnyi s entropies of this distribution are minimized by the same unraveling of a super - operator . entropic relations between a state ensemble and the generated density matrix are revisited in terms of both the adopted measures . using riesz s theorem , we obtain two uncertainty relations for any pair of generalized resolutions of the identity in terms of the rnyi and tsallis entropies . the inequality with rnyi s entropies is an improvement of the previous one , whereas the inequality with tsallis entropies is a new relation of a general form . the latter formulation is explicitly shown for a pair of complementary observables in a @xmath0-level system and for the angle and the angular momentum . the derived general relations are immediately applied to extremal unravelings of two super - operators .
recently , @xcite ( hereafter a06 ) measured a tight correlation between the bondi accretion rates @xmath3 and jet powers @xmath4 of the nuclei of nine nearby , x - ray luminous elliptical galaxies using _ chandra _ x - ray observations . it is imperative to understand how this correlation is established and what constraints on the properties of the central black holes can be derived from it . this is the goal of the present work . the origin of jets is likely to be related to the spin of the central black hole ( e.g. , @xcite ) . with this in mind , we employ two physical models of jet production which relate the spin @xmath5 ( @xmath6 ) and accretion rate @xmath7 onto the black hole to the observed jet power in order to understand the empirical correlation of a06 : the blandford - znajek ( bz ) model ( e.g. , @xcite ) , in which magnetic fields threading the hole extract its rotational energy and drive the jet ; and a `` hybrid model '' @xcite which combines the blandford - payne and bz mechanisms , in which the fields tap energy from the accretion flow and the spinning hole . we assume that in the nuclei of the observed elliptical galaxies the accretion flow around the black hole is advection - dominated ( advection - dominated accretion flows , hereafter adaf , e.g. , @xcite , see also contribution of f. yuan in this volume ) . adafs are expected to be associated with the production of radio jets ( e.g. , @xcite ) . we take into account general relativistic effects not fully appreciated before in such models , in particular the dependence of the radius of the marginally stable orbit ( @xmath8 ) on the black hole spin , and frame - dragging effects that intensify the magnetic field strength . figure 1 ( left ) shows the spin dependence of @xmath4 we obtain from the modelling . figure 1 ( right ) illustrates the accretion rates and spins required to reproduce the correlation between @xmath3 and @xmath4 for @xmath9 . for typical values of the viscosity parameter @xmath0 , we obtain that the tight correlation of a06 implies the narrow range of spins @xmath10 and accretion rates @xmath11 . these results are discussed in more depth in @xcite . allen , s. w. , et al . 2006 , mnras , 372 , 21 blandford , r. d. 2002 , in lighthouses of the universe , ed . m. gilfanov , r. sunyaev , & e. churazov ( berlin : springer ) , 381 meier , d. l. 2001 , apj , 548 , l9 nemmen , r. s. , et al . 2006 , apj , 643 , 652 nemmen , r. s. , bower , r. g. , babul , a. , & storchi - bergmann , t. 2007 , mnras , submitted ( astro - ph/0612354 )
recently , allen et al . measured a tight correlation between the bondi accretion rates and jet powers of the nuclei of nearby x - ray luminous elliptical galaxies . we employ two models of jet powering to understand the above correlation and derive constraints on the spin and accretion rate of the central black holes . the first is the blandford - znajek model , in which the spin energy of the hole is extracted as jet power ; the second model is an hybrid version of the blandford - payne and blandford - znajek processes , in which the outflow is generated in the inner parts of the accretion disk . we assume advection - dominated accretion flows ( adaf ) and account for general relativistic effects . our modelling implies that for typical values of the disk viscosity parameter @xmath0 the tight correlation implies the narrow range of spins @xmath1 and accretion rates @xmath2 . our results provide support for the `` spin paradigm '' scenario and suggest that the central black holes in the cores of clusters of galaxies must be rapidly rotating in order to drive radio jets powerful enough to quench the cooling flows .
the homunculus nebula around @xmath0 carinae is composed of a striking pair of lumpy bipolar lobes divided by a ragged equatorial skirt ( morse et al . 1998 ; duschl et al . 1995 ; smith et al . 2000 ; smith & gehrz 2000 ) . it is primarily a reflection nebula at uv , visual , and near - ir wavelengths , since it is highly polarized ( thackeray 1956b , 1961 ; visvanathan 1967 ; meaburn et al . 1993 ; schulte - ladbeck et al.1999 ) , and because stellar wind emission lines are seen reflected by the nebula ( allen & hillier 1991 ; davidson et al . 1995 ; smith et al . 2003a ; stahl et al . 2005 ) . while spectacular in appearance , scattered - light images do not give a clear picture of the three dimensional ( 3d ) shape , tilt angle , or detailed structure of the nebula because the optical depth is near unity and the inclination angle hides parts of the nebula . additionally , the underlying structure giving rise to the clumpy , mottled appearance of the polar lobes in _ hubble space telescope _ ( _ hst _ ) and infrared ( ir ) images defies easy explanation . different datasets imply different mass distributions in dark lanes , holes , and bright cells or clumps ( e.g. , morse et al . 1998 ; smith et al . 1998 ) , evoking entertaining comparisons to structures seen in convection cells and vegetables . because of its prominence and a wealth of high - quality observational data , the homunculus is an important prototype of bipolar nebulae , being the focus of numerous theoretical studies ( frank et al . 1995 , 1998 ; owocki & gayley 1997 ; dwarkadas & balick 1998 ; langer et al.1999 ; dwarkadas & owocki 2002 ; owocki 2003 ; soker 2001 , 2004 ; matt & balick 2004 ; gonzalez et al . 2004a , 2004b ) . yet , the formation of its clumpy bipolar lobes remains an enduring mystery . even more perplexing is its ragged equatorial disk , which makes the homunculus unique among bipolar nebulae ( davidson et al . 1997 , 2001 ; zethson et al . 1999 ; smith & gehrz 1998 ; smith et al . 1998 , 2002 ) . detailed study of this nebula provides one of our most valuable ways to learn about the great eruption of @xmath0 carinae in the mid-19th century , since proper motions indicate that the nebula was ejected in that event ( gaviola 1950 ; ringuelet 1958 ; currie et al . 1996 ; smith & gehrz 1998 ; morse et al . an important observational goal is to measure the exact shape and mass distribution in the nebula , because it bears the direct imprint of the physics of the great eruption as a function of latitude . this , in turn , can constrain the roles of rotation , binarity , and the hydrodynamics of interacting winds in the formation of bipolar structure . thermal - ir observations have yielded important clues about the structure of the homunculus , because the nebula is optically thin at these wavelengths ( they have also provided our best estimates of its total mass of more than 10 m@xmath5 and ir luminosity of 4.3@xmath1110@xmath12 l@xmath5 ; smith et al . mid - ir maps by hackwell , gehrz , & grasdalen ( 1986 ) first revealed that the bipolar nebula consisted of a pair of hollow , limb - brightened osculating spheroids . this basic structure was later supported by a much higher resolution monochromatic 12 @xmath13 image ( smith et al . 1995 ) , multiwavelength mid - ir images ( smith et al . 1998 ; polomski et al.1999 ) , mid - ir imaging polarimetry ( aitken et al . 1995 ) , and spectroscopy ( hillier & allen 1992 ; meaburn et al . 1993 ) . while the homunculus had roughly the same size in thermal emission from 8 to 13 @xmath13 , smith et al . ( 1998 ) noted that it was considerably larger in scattered light images at @xmath14 . this mystery was resolved with higher resolution thermal - ir imaging using larger telescopes . smith et al . ( 2003b ) used images from the magellan observatory to demonstrate that the nebula was also larger at 1825 @xmath13 than at 813 @xmath13 . in that paper , we proposed a _ double - shell structure _ for the polar lobes , wherein a warm ( @xmath10200 k ) inner shell dominates the 813 @xmath13 images , and a cooler ( @xmath10140 k ) outer skin dominates at longer wavelengths . the double - shell structure will be an important recurring theme in the present investigation . this model also indicated that most of the @xmath1510 m@xmath5 in the homunculus was associated with the cool outer skin of the polar lobes , settling an earlier debate about the origin of the far - ir emission from @xmath0 car ( davidson & smith 2000 ; morris et al . thermal - ir maps show complex dust structure in the homunculosity core as well ( smith et al . 1995 , 2002 ; chesneau et al.2005 ) , which will be discussed in detail at the end of this paper . since the homunculus is essentially a ballistic expanding hubble flow , measured doppler shifts in the ejecta provide another viable method to trace the 3d shape . several authors have utilized this technique with spectroscopy of visual - wavelength emission lines ( thackeray 1951 , 1956a , 1956b , 1961 ; meaburn et al . 1987 , 1993 ; hillier & allen 1992 ; allen & hillier 1991 , 1993 ; currie et al . 1996 ; meaburn 1999 ) , as well as imaging polarimetry ( schulte - ladbeck et al . 1999 ) . however , none of these various studies gave a definitive measure of the shape , largely due to the low spatial resolution of ground - based spectroscopy , plus the very bright scattered continuum light and high optical depth of the dusty nebula at visual wavelengths ( see hillier 1997 for a review of proposed model shapes ) . this situation improved significantly when high spatial resolution optical spectra from the space telescope imaging spectrograph ( stis ) became available . davidson et al . ( 2001 ) studied the kinematic structure of narrow [ ni ii ] and [ fe ii ] emission lines in stis spectra to give the first clear picture of the shape of the polar lobes and their inclination angle , plus the first genuine detection of the thin equatorial disk . however , our major result in that study was still rather unsatisfying : our model shape that best fit the kinematic data ( `` model 1 '' of davidson et al . 2001 ) yielded a projected image of the polar lobes that was _ far too small and skinny _ to explain images of the homunculus . we shall see below that the solution to this mystery is that the [ ni ii ] and [ fe ii ] lines studied at visual wavelengths actually trace a smaller inner shell of the polar lobes , and not the dominant outer layer of the homunculus seen in images . the discovery of molecular gas in the homunculus ( smith 2002a ; smith & davidson 2001 ) provided a powerful new tool to study its structure via near - ir emission from h@xmath1 @xmath16=10 s(1 ) @xmath221218 . this emission line is bright and narrow , is only seen in intrinsic emission from the polar lobes ( no confusing reflected emission ) , and is at a wavelength of @xmath102 @xmath13 where we can begin to see through to the back of the homunculus . hence , spectra of h@xmath1 revealed emission from the far side of the homunculus lobes for the first time ( smith 2002a ) . smith ( 2002a ) showed that the h@xmath1 emission most likely originates in an outer shell , separate from the inner [ fe ii ] shell . thus , h@xmath1 probably traces gas associated with the same outer cool dust shell seen in thermal - ir images ( smith et al . it is therefore our most powerful kinematic diagnostic of the majority of the mass in the homunculus , and traces the component of the nebula that dominates images . however , earlier [ fe ii ] and h@xmath1 spectra had inadequate spectral resolution to clearly discern the relative thickness of the two shells or their detailed structure . the present study aims to rectify this situation using long - slit spectra of h@xmath1 and [ fe ii ] obtained with much higher dispersion . high - resolution ( @xmath1760,000 ; @xmath105 km s@xmath18 ) near - ir spectra of @xmath0 car were obtained on 2005 march 1 and 2 using the phoenix spectrograph ( hinkle et al . 2003 ) on the gemini south telescope . phoenix has a 1024@xmath11256 insb detector with a pixel scale of 0@[email protected] km s@xmath18 at a wavelength of @xmath101.6 @xmath20 m , and 1.26 km s@xmath18 at @xmath102 @xmath13 . sky conditions were photometric , and the seeing was typically 0@xmath1930@xmath195 ( better at the longer wavelength ) . removal of airglow lines was accomplished by subtracting an observation of an off - source position 35@xmath21 southeast of the star . the 0@xmath1925-wide long - slit aperture was oriented at p.a.=310@xmath22 along the polar axis of the homunculus . to sample the kinematics across the nebula , the slit was positioned on the bright central star , plus offsets of 1@xmath21 and 2@xmath21 in either direction perpendicular to the slit axis as shown in figure 1 . the observations were very similar to those described in earlier papers presenting phoenix data ( smith 2004 , 2005 ) . hr 3685 and hr 4621 were observed with phoenix on the same nights with the same grating settings in order to correct for telluric absorption and for flux calibration . telluric lines were used for wavelength calibration , using the telluric spectrum available from noao . velocities were calculated adopting a vacuum rest wavelength of 16439.981 for the [ fe ii ] ( @xmath23 ) line , and 21218.356 for h@xmath1 @xmath16=10 s(1 ) . these velocities were corrected to a heliocentric reference frame ; heliocentric velocities will be quoted here . uncertainty in the resulting velocities is @xmath91 km s@xmath18 , dominated by scatter in the dispersion solution for telluric lines . at all five slit placements , long - slit spectra were obtained for the two emission lines : [ fe ii ] @xmath216435 and h@xmath1 @xmath221218 . these were suspected to be the most unambiguous tracers of the structure in the polar lobes of the homunculus at near - ir wavelengths that can penetrate the dusty nebula ( smith 2002a ) . the phoenix slit is shorter than the @xmath1018 extent of the homunculus , so each offset position in figure 1 actually consisted of exposures at three steps along the slit that were shifted and combined later to yield a greater effective slit length . at each pointing , two or three exposures , each consisting of 2@xmath1160-second integrations ( to avoid severe saturation ) were obtained . depending on the overlap at various positions along the slit , the total effective integration time after adding all exposures for a given slit position varied from 240 to 1080 seconds . earlier experience with phoenix ( smith 2004 , 2005 ) revealed that the maximal extent of the blue and redshifted h@xmath1 velocities in the homunculus reached slightly beyond the usable bandwith of a single phoenix exposure . therefore , for h@xmath1 emission , observations were obtained at two different grating settings separated by about 400 km s@xmath18 ; one optimized for including the structure of the blushifted polar lobe and one for the redshifted lobe . the exposures at the redshifted setting were then scaled to have exactly the same dispersion per pixel , and combined with the blueshifted data . ( since the extreme redshifted velocities are only seen in the northwest polar lobe , the southeast pointings along the slit were ignored . this explains the data gap in the lower right corner of each panel on the right side of fig . 2 . ) figure 2 shows the resulting long - slit data for [ fe ii ] and h@xmath1 , respectively , at the five different slit positions , where a linear fit to the bright reflected continuum light in the homunculus has been subtracted out to enhance the contrast of the line emission ( see smith 2002a , 2004 , 2005 for more details ) . the vertical dashed lines in figure 2 mark the systemic velocity of @xmath0 car at 8.1 km s@xmath18 , measured from earlier phoenix spectra of h@xmath1 in the homunculus ( smith 2004 ) . three distinct kinematic components can be seen in figures 2 and 3 : 1 ) a thin outer shell of h@xmath1 at @xmath9600 km s@xmath18 and @xmath98 , 2 ) a thicker inner skin of [ fe ii ] at @xmath9500 km s@xmath18 and @xmath98 , and 3 ) [ fe ii ] emission at @xmath9250 km s@xmath18 and @xmath92 from the smaller bipolar nebula known as the `` little homunculus '' , which was the subject of a previous paper ( smith 2005 ) . emission in figs . 2 and 3 . this is because it is younger and thus follows a different hubble law than the larger homunculus ( smith 2005 ; ishibashi et al . 2003 ) . ] it is clear from figure 3 that h@xmath1 and [ fe ii ] emission in the polar lobes do not share the same spatial distribution . in fact , they are strongly anticorrelated . the observed kinematic structure gives a dramatic and independent confirmation of the double - shell structure proposed intially from thermal - ir images ( smith et al . 2003b ) , consisting of a warm inner dust shell , plus a cooler and denser outer shell . here , the [ fe ii ] emission clearly traces the warmer inner shell , while the h@xmath1 emission corresponds to the more massive thin outer skin . [ fe ii ] emission in figures 2 and 3 coincides with the warmer inner dust shell of the polar lobes , and therefore traces only about 10% of the total mass in the homunculus ( smith et al . for this reason , the main focus of the present paper will be on the h@xmath1 emission . however , a few comments are in order , as the kinematic structure of the inner [ fe ii ] shell is closely related to the more massive cool h@xmath1 skin . at all latitudes , [ fe ii ] emission from the lobes has a smaller radius than the h@xmath1 emission and clearly represents a separate component of the nebula . the infrared [ fe ii ] emission shown here is emitted by the same gas that gives rise to optical [ ni ii ] and [ fe ii ] lines . this explains why davidson et al.(2001 ) found that their shape for the homunculus derived from spectra was too small and skinny compared to images . the inner [ fe ii ] layer and the double - shell structure may also hold clues for understanding polarization studies of the homunculus . while dust in the outer h@xmath1 skin traces the main scattering layer seen in visual images , dust associated with the inner [ fe ii ] zone may contribute significantly to the observed polarization . if the outer layer contains predominantly large grains that behave like black - bodies at thermal - ir wavelengths , and the inner [ fe ii ] layer contains smaller grains with higher temperatures ( as observed ) that are seen through holes in the clumpy h@xmath1 skin , it may go a long way toward explaining the apparent contradictions in grain properties inferred independently from thermal - ir emission and visual to near - ir polarization ( walsh & ageorges 2000 ; schulte - ladbeck et al . 1999 ; smith et al . 2003b ) . the [ fe ii ] layer is thicker and more irregular than the h@xmath1 shell , partially filling the interior of the lobes , especially evident in figure 3 . its origin and relationship to the more massive h@xmath1 shell are not clear . on one hand , the observed double - shell structure resembles a classic two - shock structure of a wind plowing into an ambient medium . in this interpretation , the [ fe ii ] shell might represent emission from cooling stellar wind material that has passed through a reverse shock . the actual shock velocity could be quite low ( less than 100 km s@xmath18 ) , since the homunculus expansion speed at high latitudes is not much different than the post - eruption stellar wind speed ( smith et al . on the other hand , the [ fe ii ] ( and warm dust ) layer may simply be a photodissociation region , where the radiation field from the central star is still too strong to allow h@xmath1 to exist . dust in this region attenuates uv light , eventually allowing h@xmath1 to survive at larger radii ( see , e.g. , ferland et al . an almost identical double - shell structure of h@xmath1 and [ fe ii ] also exists in the famous planetary nebula m 2 - 9 ( hora & latter 1994 ; smith , balick , & gehrz 2005 ) . at all five slit positions , the [ fe ii ] emission from the little homunculus is much brighter in the blueshifted ( se ) lobe , while the [ fe ii ] from the larger homunculus is always brighter in the redshifted ( nw ) lobe . previously , this had been interpreted as attenuation of light from the nw lobe of the little homunculus as it passed through dust in the equator ( smith 2005 ) , while the brightness asymmetry in the larger lobes was caused by the thick polar caps and thinner side walls of the nebula ( davidson et al . 2001 ; hillier & allen 1992 ) . together , these two explanations have a hint of inconsistency . a more plausible conjecture might be that the redshifted little homunculus lobe has a lower density and is therefore optically thinner in the balmer continuum , allowing more uv to leak through to illuminate the larger polar lobe . a higher density in the se polar lobe of the little homunculus would explain why it is brighter in the near - ir , and may help explain the variable extended radio continuum and predominantly blueshifted radio recombination lines ( duncan & white 2003 ; duncan et al . in any case , the spatial extent of the little homunculus ( smith 2005 ) matches the extent of variable radio continuum emission ( duncan & white 2003 ) , so it would appear that the radio outbursts of @xmath0 car during the 5.5 yr cycle are due the little homunculus absorbing all the lyman continuum radiation that escapes from a variable central source . some faint [ fe ii ] emission from the equatorial skirt is present as well in panels b and c of figures 2 and 3 , as had been seen previously in [ fe ii ] , [ ni ii ] , [ ca ii ] , and he i @xmath210830 ( zethson et al . 1999 ; davidson et al . 2001 ; smith 2002a ; hartman et al . the reader should be cautioned that the curved [ fe ii ] emission feature extending from the center to the lower right in each panel of figure 2 is not intrinsic emission from gas in the homunculus ; it is a narrow emission line from the weigelt knots near the star that is reflected and redshifted by expanding dust in the se polar lobe ( see smith 2002a ; davidson et al . 2001 ) . while bright [ fe ii ] emission is seen in all nebulae around luminous blue variables ( smith 2002b ) , @xmath0 car is the only one where h@xmath1 emission has been detected . this unique emission feature in the homunculus turns out to be extremely valuable in deducing the shape and other physical properties of the homunculus , as detailed in the rest of this paper . this is because the h@xmath1 emission traces the cool and dense outer skin of the homunculus , which thermal - ir observations indicate contains the majority of the nebula s mass and kinetic energy ( smith et al . 2003b ) . seen at the high spectral resolution provided by the phoenix spectrograph on gemini south , the outer h@xmath1 skin is remarkably thin and uniform . the thickness of the h@xmath1 shell ( @xmath3r ) is typically 300600 au ( 0.51@xmath1110@xmath24 cm ) , or roughly 23% of the polar radius ( r ) . _ this gives us an important clue to the short duration of the most important mass - loss during the great eruption_. in previous studies , i have commented that the thickness of the lobes compared to their largest radius ( @xmath3r / r ) is about 1 / 8 , which is the same as the duration of the bright phase of the great eruption compared to the time elapsed since then : @xmath3t / t = 20yr/160yr ( smith et al . 1998 , 2003b ; smith 2002a ) . however , now we see that this estimate corresponds to both the inner and outer portions of the double shell structure , with only a small fraction of the mass ( @xmath410% ) found in the thicker , warmer inner shell . the very thin outer shell , where 90% of the mass resides , would seem to imply that @xmath3t / t is only a few per cent , so that the duration of the dominant mass ejection phase was @xmath45 yr . this is supported by the small age dispersion derived from high - precision proper motion measurements with _ hst _ ( e.g. , morse et al . this , in turn , would imply a _ huge _ instantaneous mass - loss rate during the peak of the eruption of about 25 m@xmath5 yr@xmath18 . this may be higher than the limit for a super - eddington continuum - driven wind ( owocki et al . 2004 ; shaviv 2000 ) . such a short duration for the eruption and the corresponding high ratio of mechanical to radiative energy may point instead to a mechanism more like a hydrodynamic explosion originating deep within the star ( e.g. , arnett et al . 2005 ; young 2005 ) . however , the very thin h@xmath1 skin also requires a very small velocity dispersion at each latitude during the original ejection , which is not a property that one usually associates with an explosion . a caveat , of course , is that even though the homunculus is seen as a hubble flow today , we can not be certain that @xmath3r / r exactly traces @xmath3t / t , since thermal collapse of the shell or `` pile - up '' in the ejecta may have occurred during or shortly after the eruption . shell . ] no observations currently exist with which one can reliably rule out either possibility , so both wind and explosion models are worth exploring theoretically . in any case , the outer h@xmath1 skin is remarkably thin at the present epoch thinner than had been recognized in previous near - ir spectroscopy or thermal - ir imaging ( smith 2002a ; smith et al . this narrow outermost layer gives rise to a sharp absorption feature at 513 km s@xmath18 where it crosses our line of sight to the central star , seen in h@xmath1 as well as several low - ionization metals like ti ii and fe i in uv spectra of @xmath0 car ( gull et al.2005 ; nielsen et al . instead of tracing just one sightline , long - slit near - ir spectra have the advantage of showing emission patterns from the full homunculus . comparing the uv absorption profiles of gull et al . ( 2005 ) to the spectra in figure 2 , it is clear that the 513 km s@xmath18 component is from the outer h@xmath1 skin , the 146 km s@xmath18 component arises in the `` little homunculus '' ( as noted already by smith 2002a and smith 2005 ) , and the several intermediate velocity components seen in low ionization metals correspond to the thick lumpy [ fe ii ] emitting layer seen in the near - ir ( note that the 513 and 146 km s@xmath18 absorption components are indentified by small white dots on the spectrum with the slit passing through the star in figs . 2 and 3 ) . in the discussion in following sections , it will ease matters if it is fair to assume that the h@xmath1 layer can be approximated as a thin skin of uniform density and constant thickness . the h@xmath1 skin obviously has some clumping and some minor brightness variations , but overall the shell has a remarkably constant thickness and brightness certainly within @xmath950% . one apparent exception to the large - scale symmetry is in the portions of the walls of the polar lobes that run tangent to our line of sight , projected near the position of the star . these portions appear much fainter in figure 2 than the corresponding parts of the lobes at the same latitudes ( 040 ) which are nearly in the plane of the sky and are seen perpendicular to the dispersion direction . these differences give the impression of strong point symmetry in the brightness distribution at low latitudes in the homunculus . however , it is not certain that this effect is real . the problem is that the fainter portions are narrow features stretching along the dispersion direction , and coincide spatially with very bright reflected continuum emission that has been subtracted . it is quite possible that some h@xmath1 emission was subtracted - out when attempting to fit the continuum at these positions . also , these regions may suffer considerable internal extinction as we look through the limb of the outer dust shell . on the other hand , if these differences mark a real deficit of material , then the possible departures from a uniform - density model discussed here are in the sense that there will be additional missing material at low latitudes in the side walls of the polar lobes , as compared to the assumption of a uniform skin . this would only reinforce the general conclusion that most of the mass is at high latitudes ( see 5 ) . since near - ir h@xmath1 emission traces the outer dust shell with most of the mass in the homunculus , and the primary dust scattering layer that dominates continuum images of the reflection nebula ( e.g. , davidson & ruiz 1975 ) , it is our most valuable diagnostic of the true shape of the homunculus . shell ( davidson et al . 2001 ; smith 2002a ) . however , it is seen in absorption , tracing only the near side of each lobe . ] the slit position passing through the central star is best for this purpose , as it is oriented along the symmetry axis of the nebula at p.a.@xmath25310 . the age of the homunculus is @xmath10160 yr based on proper motion measurements and on the most dramatic change in brightness during the great eruption ( currie et al . 1996 ; smith & gehrz 1998 ; morse et al . 2001 ; frew 2004 ) , and the entire outflow is essentially a hubble flow . therefore , the observed doppler shift in km s@xmath18 ( @xmath26 ) translates directly to distance from the plane of the sky as @xmath27=33.5@xmath28 , measured in au . similarly , for a given distance from earth , the pixel scale in arcseconds can be converted to a physical distance in the plane of the sky in order to create a model shape with the correct physical proportion . assuming axial symmetry then yields an independent estimate of the distance to @xmath0 car , and the orientation of the homunculosity ( see also davidson et al . 2001 ) . figure 4 shows the measured h@xmath1 emission shape for a cross section through the symmetry axis of the homunculus ( with the slit centered on the star ) . the relative scale of the spatial and velocity axes was adjusted to yield a model shape with the greatest degree of axial symmetry , and then rotated so that the polar axis is vertical in figure 4 ( figs . 3f h were created the same way ) . the distance and inclination angle adopted to make figure 4 are @xmath29=2350@xmath950 pc and @xmath30=41@xmath310@xmath90@xmath315 , respectively . the inclination angle @xmath30 given here corresponds to the tilt of the polar axis out of the line of sight . it is the same quantity as the inclination defined for binary system orbits , so that the polar axis is tilted from the plane of the sky by 49 . the uncertainty is dominated by real deviations from axial symmetry in the homunculus , which are of order 23% of the radius . this distance estimate is larger than previous estimates using the same technique ( smith 2002a ; davidson et al.2001 , allen & hillier 1993 ) , but the various studies agree to within the uncertainty . the distance derived here , however , is the first to use high - resolution data for the outer h@xmath1 shell that is actually seen in images , instead of the interior [ fe ii ] and [ ni ii ] emitting layer . the best fitting idealized symmetric model for the shape of the polar lobes is plotted over the data in figure 4 . this curve provides a direct measure of the size ( radial separation from the star in au ) of the homunculus as a function of latitude , which is listed in table 1 . this is a definitive measure of the true 3d shape of the nebula ; it supersedes previous efforts largely because doppler shifts of the narrow molecular hydrogen line constitute such a powerful and unique tracer of the outer massive shell of the homunculus . earlier studies using optical [ fe ii ] and [ ni ii ] were not measuring the outer shell at all , but instead measured the inner shell , which has an entirely different shape . the word `` definitive '' in this context is not hyperbole , because one ca nt do much better with observational data uncertanties in the derived model shape are dominated by the inherent corrugations in polar lobe surfaces , which are , in turn , comparable to the thickness of the shell . hopefully , the model shape in figure 4 and table 1 will be used to constrain future theoretical investigations of the formation of bipolar nebulae in general , and the homunculus in particular . of course , the homunculus lobes are not _ perfectly _ symmetric . departures from axial symmetry are most apparent in the near side of the se lobe , which appears somewhat scrunched or `` snub - nosed '' overall , and has a few large dents at some slit positions , especially at the corners and right at the polar axis . these features have been mentioned before as they are apparent in images ( e.g. , morse et al.1998 ) . this is only obvious in the se lobe by contrast , the nw lobe shape obeys axial symmetry to a higher degree , showing only a few minor corrugations . overall , the h@xmath1 shell appears smoother in the side walls than in the polar caps . the polar regions of the lobes show several clumps or cells in the h@xmath1 spectra , as well as in high - resolution _ hst _ images ( morse et al . these correspond to minor wiggles or corrugations in the shape of the nebula but these deviations from a smooth shape in the radial direction are smaller than sizes of cells in images . in other words , these are not three - dimensional `` clumps '' in the normal sense , but rather , fractal divisions of the outer shell s surface . they are more like irregular plates in the thin shell , which bring to mind analogies with rapid cooling and thermal instabilities or convection cells . these analogies may be appropriate during the optically thick , rapidly cooling phase after the material was ejected . the bright h@xmath1 features , presumably tracing the densest material , correspond to bright features in optical images , rather than dark lanes . the maximum radial expansion speed does not occur at the pole , but at latitudes around 6570 ( fig . this gives rise to a hint of a trapezoidal shape to the lobes of the homunculus . this trapezoidal shape is more obvious in the polar lobes of the lh ; this coincidence suggests a common physical mechanism at work . one possibility is that this results from extra ram pressure inside the corners of the lobes , arising from the deflection of the post - eruption wind that the inner side - walls of the polar lobes at an oblique angle . such effects are present in simulations of bipolar wind - blown bubbles ( e.g. , cunningham , frank , & hartmann 2005 ; frank et al . this extra ram pressure would not be related to the opening angle of the shock cone in a binary system , since during the eruption it is likely that the shock cone was collapsed onto the secondary star . the homunculus bears the imprint of the great eruption s mass loss as a function of latitude . therefore , the shape and structure of the h@xmath1 emission in high - resolution spectra provide important insight to the physics of the outburst , to the extent that h@xmath1 is a good tracer of the mass . figure 5 shows the latitude dependence of velocity , mass loss , and mechanical energy in the homunculus , based on h@xmath1 emission . the velocity distribution follows directly from the shape of the nebula , since it is a hubble flow . however , calculating these distributions of mass and kinetic energy requires a few assumptions . as a first approximation , the thin outer h@xmath1 shell of the homunculus is taken to be of uniform density with a constant width . the apparent width in phoenix spectra is very nearly constant , with variations typically less than a factor of 2 . note that the thicker appearance of the h@xmath1 shell at slit offset positions ne2 and sw2 ( see fig . 1 ) is somewhat misleading , since these slit positions slice through the lobes at an oblique angle to their surfaces . if this assumption applies , then the mass in various parts of the shell is simply proportional to its geometric surface area at each latitude . while the nebula is obviously not perfectly uniform , we will see below that the important characteristics of the latitudinal distribution in the nebula are not compromised by these minor effects . to calculate various quantities as functions of latitude , the outer shell was approximated as a series of adjacent loops or tori centered on the polar axis of the nebula ( see figure 6 ) . the mass for each representative latitudinal loop is given by 2@xmath32 ( @xmath33)(2@xmath34r@xmath35 ) , where r@xmath35=r@xmath35(@xmath36 ) is the horizontal radius of each torus measured from the polar axis ( parallel to the equator ) , and 2@xmath37 is the thickness of the torus , roughly equal to the overall shell thickness of @xmath10600 au ( the extra factor of 2 accounts for the mass in both polar lobes ) . following the assumption of constant shell thickness and uniform density , however , the fraction of the total mass at each latitude @xmath38(@xmath36)=m(@xmath36)/@xmath39m(@xmath36 ) is simply proportional to r@xmath35 . this quantity is plotted in figure 5b , and avoids the uncertainty in the absolute physical density , shell thickness ( as long as it is constant ) , and total mass . the fraction of the total kinetic energy at each latitude @xmath40(@xmath36)=m(@xmath36)@xmath16(@xmath36)@xmath41/@xmath39[m(@xmath36)@xmath16(@xmath36)@xmath41 ] in figure 5c is calculated for each torus from its value of @xmath38(@xmath36 ) and its radial expansion speed . these rings were regularly spaced along the surface of the nebula ( fig . 6 ) , but not spaced regularly in latitude . therefore , before plotting in figure 5 , the values for @xmath38(@xmath36 ) and @xmath40(@xmath36 ) were rebinned in latitude intervals of 5 . figure 5b indicates that most mass lost during the great eruption was aimed to high latitudes . only 25% of the total homunculus mass was ejected at @xmath43 , with the remaining 75% ejected between 45 and the pole . the latitudinal distribution of the ejecta s kinetic energy is even more extreme , with almost all the mechanical energy ( @xmath1094% ) escaping between 45 and the pole . the efficiency of imparting mechanical energy to the ejecta seems to peak at latitudes around 60 - 65 , while the peak mass loss occurs around 5060 . for comparison , the latitudinal distribution of mass and kinetic energy for a uniform spherical shell is shown by the dotted curves in figure 5 , while figure 7 shows the mass and kinetic energy per solid angle ( so that a spherical shell is a horizontal line ) . the differences between a spherical shell and the homunculus are striking . this latitudinal behavior strongly refutes the idea that a bipolar nebula like the homunculus was caused by an otherwise spherical explosion or wind that is simply pinched by a pre - existing circumstellar torus , because in that case we would expect to see more mass at low latitudes ( see also dwarkadas & balick 1998 and soker 2001 ) . likewise , the mass loss during the great eruption could not have been redirected toward the poles by deflection from its companion star , because the amount of kinetic energy in the polar ejecta is greater than the binding energy of the current putative binary system , and more than three orders of magnitude greater than the available energy in the companion s wind during the eruption . instead , the concentration of mass at high latitudes suggests that the original explosion itself directed the mass and momentum toward the poles . an alternative way to express same result is that perhaps much of the mass that would have been ejected at 040 failed to reach escape velocities and fell back onto the star . in any case , having the majority of the mass a significant fraction of the total mass of the star lost at polar latitudes will effect the star s angular momentum evolution . a polar explosion will tend to take away less than its share of angular momentum , leaving the post - outburst star with higher angular momentum per unit mass . this is related to its presently - observed bipolar stellar wind ( smith et al . additionally , after ejecting its own massive lbv nebula , there is evidence that ag carinae is also left as a rapid rotator ( groh et al . ultimately , without a circumstellar torus or deflection by a companion , the bipolar shape must come from rapid rotation of the primary star ; in that view the role of a companion star is relegated to helping to spin up the primary star through tidal friction during repeated periastron passes ( e.g. , smith et al.2003a ) . if the mass loss is driven by radiation , then gravity darkening on a rapidly rotating star will tend to drive most of the mass flux toward high latitudes ( e.g. , owocki et al . 1996 , 1998 ; owocki & gayley 1997 ; owocki 2003 ) , providing a possible explanation for the polar mass concentration . however , it remains to be seen if even a continuum - driven wind can fuel the required mass loss ( owocki et al . 2004 ) . the latitudinal mass distribution has some interesting connections with other famous bipolar nebulae , especially the circumstellar nebula around sn1987a . both @xmath0 car and sn1987a have nebulae with equatorial and bipolar mass loss , but the two polar rings around sn1987a are particularly vexing . in the homunculus of @xmath0 car , we see that most of the mass outflow was concentrated toward intermediate latitudes around 6570 . one can imagine that if the mechanism that produced this mass concentration were pushed to some extreme limit , it might produce a pair of rings as it flowed out from the star . while energy arguments rule out bipolar shaping by a companion star in an eccentric 5.5 yr orbit that survives as the putative binary system seen in @xmath0 car today ( see above ) , a more exotic model where the 1843 great eruption involved the merger of a tight binary system is admittedly more difficult to reject . in fact , a recent theoretical study of a merger scenario for the progenitor of sn1987a found that most of the mass loss is directed toward mid latitudes , with a significant deficit of mass loss near the equator ( morris & podsiadlowski 2005 ) . on the other hand , if one wishes to explain @xmath0 car s great eruption as a merger of two massive stars , then it is a mystery why the similar eruption of p cygni in 1600 a.d . ( see humphreys et al . 1999 ) ejected a spherical shell ( smith & hartigan 2006 ) and apparently did not leave that star as a rapid rotator . finally , an important caveat is needed . the notorious equatorial skirt of @xmath0 carinae to be discussed in 7 is not detected in near - ir h@xmath1 emission , presumably because molecules simply can not survive or never formed there . this equatorial spray of ejecta seen in images is quite extensive and reaches large radii from the star , comparable to some parts of the polar lobes . thus , one should expect that the true velocity , mass , and energy distributions in figure 5 would be modified with a spike at @xmath36=0 . the magnitude of this equatorial spike in the mass distribution is not known . judging by the general absence of these equatorial features in thermal - ir images ( smith et al . 2003b ) , however , the skirt likely contains less than @xmath100.5 m@xmath5 . a crude estimate of the expected mass density in the h@xmath1 shell can be made from the total mass of the homunculus and the volume occupied by the thin h@xmath1 skin . the h@xmath1 skin corresponds to the cool 140 k dust component that occupies the outer shell seen in thermal - ir images , which smith et al . ( 2003b ) estimate to contain at least 11 m@xmath5 . this 11 m@xmath5 contains about 90% of the total mass in the homunculus , but its actual value depends on the highly uncertain gas : dust mass ratio , assumed to be 100 . as noted by smith et al . ( 2003b ) , this is most likely a lower limit to the mass , since the ejecta of @xmath0 car are depleted of important grain constituents like c and o , and much of the fe is evidently in the gas phase as re - enforced by the spectra presented in this paper . therefore , the density derived below will also be a lower limit . the total volume of the model h@xmath1 skin in figure 6 is v@xmath44=4.6@xmath1110@xmath45 ( r/300 au)@xmath41 @xmath46 . thus , the equivalent pure atomic hydrogen density in the h@xmath1 skin is given by @xmath47 where @xmath48 is the proton mass and 2@xmath37=600 au is the assumed thickness of the h@xmath1 skin . two points need to be stressed : 1 ) @xmath49@xmath153@xmath1110@xmath12 @xmath8 is a lower limit because the total mass estimate is a lower limit ( smith et al . 2003b ) and because parts of the shell are thinner than adopted here , and 2 ) this is an _ average _ density , so densities above 10@xmath50 @xmath8 may be present in clumps even if the total mass estimate is accurate . the assumption of constant average density throughout the h@xmath1 skin may be optimistic , of course , but given the very thin side walls of the homunculus lobes , the latitudinal mass trend in figure 7 could only be erased ( flattened ) if the density were of order 10@xmath51 @xmath8 or more . corresponding column densities would be @xmath5210@xmath53 @xmath54 , which would contradict the fact that the side walls of the polar lobes are more optically thin than the caps ( e.g. , davidson et al . 2001 ; allen & hillier 1993 ) . such high column densities would also contradict the much lower value of @xmath521.5@xmath1110@xmath55 @xmath54 from scattered hard x - rays in the homunculus ( corcoran et al . 2004 ) . similar values of @xmath5610@xmath50 @xmath8 have been derived by gull et al . ( 2005 ) from uv absorption along our sightline to the star in the lowest ionization species at a radial velocity of 513 km s@xmath18 . this rough agreement with the density value quoted here gives independent confirmation of the very high mass estimate for the homunculus of more than 10 m@xmath5 ( smith et al . in fact , the volume of the h@xmath1 shell derived from phoenix spectra and the higher density derived independently by gull et al . would favor an even higher mass for the homunculus of perhaps @xmath1020 m@xmath5 . while the shape , orientation , and density structure of the polar lobes seem understandable in the framework of the double - shell structure discussed here , the morphology of the messy equatorial skirt is still perplexing . however , the new phoenix spectra from gemini south shed some light on this mystery . one important clue is that the material in the extended equatorial skirt shows no detectable infrared h@xmath1 emission , despite residing at large radii where it should be shielded by dense dusty structures in the core region . the features in the equatorial skirt , while prominent in reflected light at visual wavelengths ( morse et al.1998 ; duschl et al . 1995 ) , have weak emission in the thermal - ir ( smith et al . these facts suggest that the equatorial disk has much lower density than the walls of the lobes , and so dust and molecule formation was less efficient there in the decades after the great eruption . in fact , some equatorial features seem to be optical illusions ( davidson et al . 2001 ) . thus , the equatorial skirt contains only a small fraction of the total mass in the homunculus ; probably @xmath40.5 m@xmath5 , as noted above . the spatial and kinematic structure in high - resolution spectra of h@xmath1 give the first clear picture of the shape and structure of the homunculus at low latitudes near equator where the two polar lobes meet . this region is hidden in images and optical spectra due to obscuration from the foreground lobe , and the [ fe ii ] emission there is very confusing due to the little homunculus ( ishibashi et al . 2003 ; smith 2005 ) and structures near the star like the so - called `` weigelt knots '' and other clumps ( weigelt et al . 1995 ; davidson et al . 1997 ; smith et al . 2004 ; smith 2005 ; chesneau et al . thermal emission from dust traces an extremely complicated distribution of interconnected clumps , arcs , and diffuse emission seen most clearly in high resolution images at 28 @xmath13 ( chesneau et al . 2005 ; smith et al . 2002 , 2003b ; smith & gehrz 2000 ; rigaut & gehring 1995 ) . this is the warmest dust ( t@xmath15250 k ) in the core of the homunculus , long thought to constitute some sort of a warm torus ( hyland et al . 1979 ; mitchell et al . 1983 ; hackwell et al . 1986 ; smith et al . 1995 , 1998 ; polomski et al . 1999 ) . two recent papers with the best ir images to date suggested two very different interpretations of these complex structures : 1 ) smith et al . ( 2002 ) proposed that they mark a distorted and disrupted torus or ring where the two polar lobes of the homunculus meet at the equator . in this view , the complex structures are a result of lower - density structures getting swept back by the post eruption stellar wind , and higher density clumps resisting that expansion . 2 ) chesneau et al . ( 2005 ) proposed the existence of a new structure called the `` butterfly nebula '' , which was a dusty bipolar nebula different from the little homunculus , and which had the dust reaching back toward the star along the polar axis . the new h@xmath1 spectra distinguish clearly between these two hypotheses . figure 8 demonstrates the correspondence between the structures seen in a high - resolution map of the warm dust column density near the star from smith et al . ( 2003b ) and the spatial and kinematic structure of h@xmath1 emission in spectra . where the thin h@xmath1 walls of the homunculus meet at the equator , they do not reach all the way in to the star , but instead leave a gap in the apparent distribution of h@xmath1 along the slit . this gap is slightly different at each slit position , and is noted in each spectral image . each of these gaps is superposed on top of the ir continuum image , where we see that indeed , the spatial limits of h@xmath1 emission correspond nicely to the empty interior of the so - called `` butterfly nebula '' or torus . the agreement between h@xmath1 in the walls of the homunculus and the dust structures prove that this really is an equatorial torus or ring where the polar lobes meet at the equator , rather than the separate nebular structure proposed by chesneau et al . the toroidal interpretation also gives a more natural explanation for the deficit of ir emission within its boundaries ( chesneau et al . no kinematic evidence for the putative butterfly structure is seen in spectra , and the dust structures are not coincident with the little homunculus either ( fig.8 ) . the question then is : if these structures represent parts of an equatorial torus or ring , why does it exhibit such prominent azimuthal asymmetry ( clumps , arcs , etc . ) , while the rest of the h@xmath1-emitting portions of the homunculus , by comparison , show such well - ordered axisymmetric structure ? the strange clumps and arcs in this equatorial dust structure and the lack of similar features in the polar lobes can be understood as follows ( smith et al . 2002 ) : in the 160 yr since the homunculus was ejected in the great eruption , it has been followed by the post - eruption stellar wind . at the poles , the speed for the bulk of the post eruption wind is about 600 km s@xmath18 ( traced by the deepest part of the h@xmath57 p cygni absorption ; smith et al . 2003a ) , which is almost identical to the expansion speed of the homunculus ( fig 5a ) . with such a small @xmath58 , we expect little or no hydrodynamic influence from the wind . the equator presents a very different story . although the present - day stellar wind is bipolar , it appears to maintain relatively high wind speeds of @xmath10400 km s@xmath18 at the equator ( smith et al . 2003a ) , whereas the pinched waist of the homunculus seen in h@xmath1 has very slow velocities of only a few dozen km s@xmath18 . this large value of @xmath58 right at the equator should lead to rayleigh - taylor or nonlinear thin - shell instabilities ( vishniac 1994 ) in the torus where the polar lobes meet , much like any other dense ejecta shell followed by a faster wind ( e.g. , garcia - segura et al . the homunculus is inherently clumpy at all latitudes , perhaps even more so at the equator ( ud doula & owocki 2002 ) . since the ram pressure of the incident wind depends on ( @xmath58)@xmath41 , the post - eruption wind is most influential at the equator . any clumps in this ring will resist the ram pressure of the wind more than the less dense regions between them . the inter - clump sections of the torus would be swept back , leading to structures akin to dust pillars in h ii regions . in other words , the values in the first row of table 1 are azimuthally dependent . this is essentially a more extreme version of the rayleigh - taylor instabilities that are thought to have led to the early development of hot spots in the famous equatorial ring around sn1987a ( e.g. , michael et al . 1998 ) . in the inter - clump regions of the torus , where the wind pushes outward , there may be an interesting link to the unusual structures seen in @xmath0 car s equatorial skirt . ( the equatorial skirt is the more extended disk with peculiar radial streamers or spokes that seem to point back to the star ; duschl et al . 1995 ; morse et al . 1998 . ) in the inter - clump regions of the torus that are pushed to larger radii , light may more easily leak through to illuminate outer regions of the skirt ( i.e. searchlights like the radial features in the egg nebula ; sahai et al . 1998 ) . in some extreme cases , the stellar wind or ejecta from the great eruption or the 1890 eruption may even have broken through the homunculus . some material in the skirt is indeed seen to have kinematics consistent with ejection during the 1890 eruption ( davidson et al . 2001 ; smith 2002a ) , and some more extended equatorial features like the nn jet and the s condensation appear to have originated in the great eruption ( morse et al . these searchlights or ejecta breaking through the homunculus in the equator provide a natural explanation for the quasi - radial spokes or streamers seen in the equatorial skirt . these features in the skirt may be largely an illumination effect , while the true distribution of matter in the equatorial disk may be considerably more azimuthally symmetric than it appears . new high - resolution , long - slit spectra of h@xmath1 @xmath221218 and [ fe ii ] @xmath216435 from the phoenix spectrograph on gemini south provide new clues to the detailed structure of the homunculus . the main conclusions of this study are summarized as follows : \1 . the relative spatio - kinematic structure of infrared h@xmath1 and [ fe ii ] emission gives an independent confirmation of the double - shell structure for the polar lobes of the homunculus proposed on the basis of thermal - ir images ( smith et al . the inner [ fe ii ] layer corresponds to the warmer ( 200 k ) inner dust layer that traces only about 10% of the mass of the homunculus , while the outer h@xmath1 skin traces the coolest dust component ( 140 k ) that contains the remaining 90% of the nebular mass . the high - resolution spectra of h@xmath1 show for the first time how thin and surprisingly smooth the outer molecular skin of the homunculus is . if the h@xmath1 skin s @xmath3r / r traces @xmath3t / t of the mass ejection , then the most important mass - loss phase during the eruption is constrained to be about 5 yr or less . the corresponding mass - loss rate is so high that the mechanical luminosity would have greatly exceeded the radiative luminosity at that time , implying that the great eruption was more of an explosion than a wind . \4 . since h@xmath1 emission traces most of the mass in the homunculus as well as the dominant dust scattering layer seen as a reflection nebula in images , h@xmath1 spectra give the first definitive picture of the 3d shape of the homunculus , which is given in table 1 . earlier estimates of the shape and orientation traced the inner [ fe ii]-emitting layer , which has a different size and structure . deviations from a smooth symmetric shape are minor generally comparable to the thickness of the lobes . the redshifted nw lobe appears more symmetric than the approaching se lobe , which shows more prominent dents and corrugations . the h@xmath1 distribution yields our first estimate of the latitudinal dependence of the velocity , mass loss , and kinetic energy of the great eruption . about 75% of the mass and more than 90% of the mechanical energy was released at latitudes between 45 and the pole . this latitudinal mass distribution rules - out a model where the bipolar shape of the nebula arises from an otherwise spherical explosion being constricted by a pre - existing circumstellar torus . the kinetic energy of the polar ejecta rule out a model where the ejecta were deflected toward the poles by a companion star , because the putative companion could not have supplied the required energy . this means that the ejecta were directed toward the poles by the explosion itself . a merger event is not as easy to rule out . . the spatio - kinematic structure of h@xmath1 emission at the pinched waist of the nebula helps explain the unusual and complex structures seen in high - resolution ir images as an equatorial torus disrupted by the post - eruption wind . the radial structures in the equatorial skirt seen in visual - wavelength images are the result of either beams of light or ejecta that have broken through the inner dust torus . i thank verne smith and sybil adams for assistance during the gemini south observing run , and keivan stassun for diligently obtaining numerous dark frames with phoenix earlier in the same observing nights . i also thank kris davidson and stan owocki for comments on the manuscript and numerous discussions . i was supported by nasa through grant hf-01166.01a from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc . , under nasa contract nas5 - 26555 . 0.0 & 2100 & 62 & 2100 & 0.0 + 2.0 & 2349 & 70 & 2347 & 82 + 4.0 & 2997 & 89 & 2990 & 209 + 8.6 & 3709 & 111 & 3668 & 555 + 13.9 & 4365 & 130 & 4237 & 1049 + 18.7 & 4961 & 148 & 4698 & 1594 + 21.9 & 5608 & 167 & 5202 & 2097 + 25.3 & 6276 & 187 & 5675 & 2681 + 28.6 & 6858 & 205 & 6021 & 3284 + 31.6 & 7422 & 222 & 6323 & 3887 + 34.2 & 7999 & 239 & 6613 & 4501 + 36.5 & 8654 & 258 & 6957 & 5149 + 38.7 & 9259 & 276 & 7225 & 5790 + 40.7 & 9935 & 297 & 7530 & 6481 + 42.1 & 10545 & 315 & 7824 & 7071 + 43.9 & 11200 & 334 & 8075 & 7761 + 45.5 & 11880 & 355 & 8327 & 8474 + 46.8 & 12487 & 373 & 8551 & 9100 + 48.3 & 13156 & 393 & 8756 & 9820 + 49.4 & 13823 & 413 & 8990 & 10500 + 50.5 & 14472 & 432 & 9210 & 11164 + 51.6 & 15044 & 449 & 9336 & 11798 + 52.6 & 15622 & 466 & 9481 & 12417 + 53.8 & 16275 & 486 & 9624 & 13126 + 54.7 & 16886 & 504 & 9753 & 13786 + 55.8 & 17503 & 523 & 9837 & 14478 + 56.9 & 18097 & 540 & 9869 & 15169 + 58.1 & 18716 & 559 & 9883 & 15895 + 59.1 & 19255 & 575 & 9878 & 16528 + 60.2 & 19779 & 591 & 9819 & 17170 + 61.4 & 20374 & 608 & 9739 & 17896 + 62.9 & 20898 & 624 & 9526 & 18601 + 64.3 & 21323 & 637 & 9249 & 19213 + 65.9 & 21641 & 646 & 8850 & 19749 + 67.7 & 21906 & 654 & 8327 & 20262 + 69.5 & 22014 & 657 & 7710 & 20621 + 71.3 & 22012 & 657 & 7060 & 20850 + 73.1 & 22001 & 657 & 6400 & 21050 + 74.9 & 21967 & 656 & 5740 & 21204 + 76.6 & 21932 & 655 & 5100 & 21332 + 78.3 & 21892 & 654 & 4450 & 21435 + 79.9 & 21873 & 653 & 3800 & 21540 + 81.8 & 21833 & 652 & 3100 & 21613 + 83.8 & 21774 & 650 & 2370 & 21645 + 85.6 & 21730 & 649 & 1660 & 21667 + 87.3 & 21711 & 648 & 1005 & 21688 + 89.1 & 21690 & 648 & 335 & 21688 +
high resolution long - slit spectra obtained with the phoenix spectrograph on gemini south provide our most accurate probe of the three dimensional structure of the homunculus nebula around @xmath0 carinae . the new near - infrared spectra dramatically confirm the double - shell structure inferred previously from thermal dust emission , resolving the nebula into a very thin outer shell seen in h@xmath1 @xmath221218 , and a warmer , thicker inner layer seen in [ fe ii ] @xmath216435 . the remarkably thin and uniform h@xmath1 skin has @xmath3r / r of only a few per cent at the poles , hinting that the most important mass loss during the 19th century eruption may have had a very short duration of @xmath45 yr . h@xmath1 emission traces the majority of the more than 10 m@xmath5 of material in the nebula , and has an average density of order @xmath610@xmath7 @xmath8 . this emission , in turn , yields our first definitive picture of the exact shape of the nebula , plus a distance of 2350@xmath950 pc and an inclination angle of @xmath1041 ( the polar axis is tilted 49 from the plane of the sky ) . the distribution of the h@xmath1 emission provides the first measure of the latitude dependence of the speed , mass loss , and kinetic energy associated with @xmath0 car s 19th century explosion . almost 75% of the total mass and more than 90% of the kinetic energy in the ejecta were released at high latitudes between 45 and the polar axis . this rules out a model for the bipolar shape wherein an otherwise spherical explosion was pinched at the waist by a circumstellar torus . also , the ejecta could not have been deflected toward polar trajectories by a companion star , since the kinetic energy of the polar ejecta is greater than the binding energy of the putative binary system . instead , most of the mass appears to have been directed poleward by the explosion itself or the star failed to launch material from low latitudes , which would have important consequences for the angular momentum evolution of the star . in any case , comparing h@xmath1 and [ fe ii ] emission resolves some puzzles about structure noted in previous studies . emission also provides our first reliable picture of the critical innermost waist of the homunculus , yielding clues to the observed morphology of the core and the more extended equatorial debris .
[ sec1 ] the literature on functional data analysis has seen a spectacular growth in the last twenty years , showing promise in applications ranging from genetics [ @xcite ] to proteomics [ @xcite ] , epidemiology [ @xcite ] and oceanography [ @xcite ] . because functional data are inherently complex , functional clustering is useful as an exploratory tool in characterizing variability among subjects ; the resulting clusters can be used as a predictive tool or simply as a hypothesis - generating mechanism that can help guide further research . some examples of functional clustering methods include @xcite , who use b - spline fitting coupled with @xmath0-means clustering ; @xcite , who apply @xmath0-means clustering via the principal points of random functions ; @xcite , who develop methods for sparsely sampled functional data that employ spline representations ; @xcite , where the robust @xmath0-means method for functional clustering is developed ; @xcite , who use a fourier representation for the functions along with @xmath0-means clustering ; @xcite , where a bayesian hierarchical clustering approach that relies on spline representations is proposed ; @xcite , who build a hierarchical bayesian model that employs a bayesian nonparametric mixture model on the coefficients of the wavelet representations ; and @xcite , where a @xmath0-centers functional clustering approach is developed that relies on the karhunen love representation of the underlying stochastic process generating the curves and accounts for both the means and the modes of variation differentials between clusters . all of the functional clustering methods described above have been designed for situations where a single curve is observed for each subject or experimental condition . extensions to nested designs where multiple curves are collected per subject typically assume that coefficients describing subject - specific curves arise from a common parametric distribution , and clustering procedures are then applied to the parameters of this underlying distribution . the result is a procedure that generates clusters of subjects based on their average response curve , which is not appropriate in applications in which subjects vary not only in the average but also in the variability of the replicate curves . for example , in studies of trajectories in reproductive hormones that collect data from repeated menstrual cycles , the average trajectory may provide an inadequate summary of a woman s reproductive functioning . some women have regular cycles with little variability across cycles in the hormone trajectories , while other women vary substantially across cycles , with a subset of the cycles having very different trajectory shapes . in fact , one indication of impending menopause and a decrease in fecundity is an increase in variability across the cycles . hence , in forming clusters and characterizing variability among women and cycles in hormone trajectories , it is important to be flexible in characterizing both the mean curve and the distribution about the mean . this situation is not unique to hormone data , and similar issues arise in analyzing repeated medical images as well as other applications . this paper discusses hierarchical bayes models for clustering nested functional data in the context of the early pregnancy study ( eps ) [ @xcite ] , where progesterone profiles were collected for both conceptive and nonconceptive women from multiple menstrual cycles . our models use spline bases along with mixture priors to create sparse but flexible representations of the hormone profiles , and can be applied directly to other basis systems such as wavelets . we start by introducing a hierarchical random effects model on the spline coefficients which , along with a generalization of the dirichlet process mixture ( dpm ) prior [ @xcite ] , allows for mean - response - curve clustering of women , in the spirit of @xcite . then , we extend the model to generate distribution - based clusters using a nested dirichlet process ( ndp ) [ @xcite ] . the resulting model simultaneously clusters both curves and subjects , allowing us to identify outlier curves within each group of women , as well as outlying women whose distribution of profiles differs from the rest . to the best of our knowledge , there is no classical alternative for this type of distribution - based multilevel clustering . in order to provide some insight into the challenges associated with functional clustering in the context of the eps , consider the hormonal profiles depicted in figure [ fihorpro ] . frames ( a ) to ( c ) depict the hormone profiles for 3 women , while frame ( d ) shows the mean profile corresponding to each one of them , obtained by simply averaging all available observations at a given day within the cycle . when looking at the mean profiles in ( d ) , women 43 and 36 seem to have very similar hormonal responses , which are different from those of woman 3 . however , when the individual profiles are considered , it is clear that most of the cycles of woman 43 look like those of woman 3 and that the big difference in the means is driven by the single abnormal cycle . this result , although not surprising if we consider that means are notoriously nonrobust , suggests that simple approaches that average profiles over individuals ( or , equivalently , use gaussian distributions to describe subject - specific variability ) might not be the most appropriate for this type of data . the use of bayesian nonparametric mixture models for clustering has a long history [ @xcite ] and presents a number of practical advantages over other model - based clustering techniques . nonparametric mixtures induce a probability distribution on the space of partitions of the data , therefore , we do not need to specify in advance the number of clusters in the sample . once updated using the data , this distribution on partitions allows us to assess variability , and hence characterize uncertainty in the clustering structure ( including that associated with the estimation of the curves ) , providing a more complete picture than classical methods . in this paper , we work with a generalized dirichlet process ( gdp ) first introduced by @xcite and study some of its properties as a clustering tool . in particular , we show that the gdp generates a richer prior on data partitions than those induced by popular models such as the dirichlet process [ @xcite ] or the two - parameter poisson dirichlet process [ @xcite ] , as it allows for an asymptotically bounded number of clusters in addition to logarithmic and power - law rates of growth . the paper is organized as follows : section [ semixfunc ] reviews the basics of nonparametric regression and functional clustering , while section [ sebnpmixfunc ] explores the design of nonparametric mixture models for functional clustering . building on these brief reviews , section [ sefunclustnested ] describes two bayesian approaches to functional clustering in nested designs , while section [ secompu ] describes markov chain monte carlo algorithms for this problem . an illustration focused on the eps is presented in section [ seillus ] . finally , section [ seconclusion ] presents a brief discussion and future research directions . to introduce our notation , consider first a simple functional clustering problem where multiple noisy observations are collected from functions @xmath1 . more specifically , for subjects @xmath2 and within - subject design points @xmath3 , observations consist of ordered pairs @xmath4 , where @xmath5 for example , in the eps , @xmath6 corresponds to the level of progesterone in the blood of woman @xmath7 collected at day @xmath8 of the menstrual cycle , and @xmath9 denotes a smooth trajectory in progesterone for woman @xmath7 ( initially supposing a single menstrual cycle of data from each woman ) , and clusters in @xmath10 could provide insight into the variability in progesterone curves across women , while potentially allowing us to identify abnormal or outlying curves . if all curves are observed at the same covariate levels ( i.e. , @xmath11 and @xmath12 for every @xmath7 ) , a natural approach to functional clustering is to apply standard clustering methods to the data vectors , @xmath13 . for example , in the spirit of @xcite , one could apply hierarchical or @xmath0-means clustering to the first few principal components [ @xcite ] . alternatively , with uneven spacings or missing observations , nonparametric regression techniques such as kernel regression [ e.g. , see @xcite ] could be used to interpolate the value of the curves to a common grid , and then traditional clustering techniques could be applied . from a model - based perspective , one could instead suppose that @xmath14 is drawn from a mixture of @xmath0 multivariate gaussian distributions , with each gaussian corresponding to a different cluster [ @xcite ] . the number of clusters could then be selected using the bic criteria [ @xcite ] or a nonparametric bayes approach could be used to bypass the need for this selection , while allowing the number of clusters represented in a sample of @xmath15 individuals to increase stochastically with sample size [ @xcite ] . a more general approach that allows us to deal with unevenly spaced and missing observations is to fit a nonparametric model to each curve and then project all the curves onto a common space . for example , we can represent the unknown function @xmath9 as a linear combination of prespecified basis functions @xmath16 , that is , we can write @xmath17 where @xmath18 are basis coefficients specific to subject @xmath7 , with variability in these coefficients controlling variability in the curves @xmath10 . a common approach to functional clustering is to induce clustering of the curves through clustering of the basis coefficients [ @xcite ] . then the methods discussed above for clustering of the data vectors @xmath19 in the balanced design case can essentially be applied directly to the basis coefficients @xmath20 . although the methods apply directly to other choices of basis functions , our focus will be on splines , which have been previously used in the context of hormone profiles [ @xcite ] ; given a set of knots @xmath21 , the @xmath0th member of the basis system is defined as @xmath22 where @xmath23 . given the knot locations , inferences on @xmath24 and @xmath25 can be carried out using standard linear regression tools , however , selecting the number and location of the nodes @xmath26 can be a challenging task . a simple solution is to use a large number of equally spaced knots , together with a penalty term on the coefficients to prevent overfitting . from a bayesian perspective , this penalty term can be interpreted as a prior on the spline coefficients ; for example , the maximum likelihood estimator ( mle ) obtained under an @xmath27 penalty on the spline coefficients is equivalent to the maximum a posteriori estimates for a bayesian model under a normal prior , while the mle under an @xmath28 penalty is equivalent to the maximum a posterior estimate under independent double - exponential priors on the spline coefficients . instead of the more traditional gaussian and double exponential priors , in this paper we focus on zero - inflated priors , in the spirit of @xcite . priors of this type enforce sparsity by zeroing out some of the spline coefficients and , by allowing us to select a subset of the knots , provides adaptive smoothing . in their simpler form , zero - inflated priors assume that the coefficients are independent from each other and that @xmath29 where @xmath30 denotes the degenerate distribution putting all its mass at @xmath31 , @xmath32 controls the overdispersion of the coefficients with respect to the observations and @xmath33 is the prior probability that the coefficient @xmath34 is different from zero . in order to incorporate a priori dependence across coefficients , we can reformulate the hierarchical model by introducing bernoulli random variables @xmath35 such that @xmath36 where @xmath37 and @xmath38 are , respectively , the vectors of responses and covariates associated with subject @xmath7 , @xmath39 is the matrix of basis functions also associated with subject @xmath7 with entries @xmath40_{tk } = b_k(x_{it})$ ] , @xmath41 and @xmath42 equals 1 independently with probability @xmath33 , and @xmath43 is a @xmath44 covariance matrix . note that if @xmath43 is a diagonal matrix and @xmath45_{kk } = \omega_k$ ] , we recover the independent priors in ( [ eqzeroinfpr ] ) . for the single curve case , choices for @xmath43 based on the regression matrix @xmath46 are discussed in @xcite , @xcite and @xcite . although the preceding two - stage approach is simple to implement using off - the - shelf software , it ignores the uncertainty associated with the estimation of the basis coefficients while clustering the curves . in the spirit of @xcite , an alternative that deals with this issue is to employ a mixture model of the form @xmath47 where @xmath48 is the vector of coefficients associated with the @xmath0th cluster , @xmath49 is the diagonal selection matrix for the @xmath0th cluster , @xmath50 is the observational variance associated with observations collected in the @xmath0th cluster , @xmath51 can be interpreted as the proportion of curves associated with cluster @xmath0 , and @xmath52 is the maximum number of clusters in the sample . from a frequentist perspective , estimation of this model can be performed using expectation maximization ( em ) algorithms , while selection of the number of mixture components can be carried out using the bic . alternatively , bayesian inference can be performed for this model using markov chain monte carlo ( mcmc ) algorithms once appropriate priors for the vector @xmath53 and the cluster - specific parameters @xmath54 have been chosen , opening the door to simple procedures for the estimation of the number of clusters in the sample . note that the model in ( [ eqsimplmodel ] ) can be rewritten as a hierarchical model by introducing latent variables @xmath55 so that @xmath56\\[-8pt ] g(\cdot ) & = & \sum_{k=1}^{k } w_k \delta_{({\bolds{\theta}}^{*}_k , \sigma ^{*2}_k , { \bolds{\lambda}}_k^ { * } ) } ( \cdot).\nonumber\end{aligned}\ ] ] therefore , specifying a joint prior on @xmath57 and @xmath58 is equivalent to specifying a prior on the discrete distribution @xmath59 generating the latent variables @xmath60 . in this section we discuss strategies to specify a flexible prior distribution on this mixing distribution in the context of functional clustering . in particular , we concentrate on nonparametric specifications for @xmath59 through the class of stick - breaking distributions . a stick - breaking prior [ @xcite ] with baseline measure @xmath61 and precision parameters @xmath62 and @xmath63 is defined as @xmath64 where the atoms @xmath65 are independent and identically distributed samples from @xmath61 and the weights @xmath66 are constructed as @xmath67 , with @xmath68 another independent and identically distributed sequence of random variables such that @xmath69 for @xmath70 and @xmath71 . for example , taking @xmath72 , @xmath73 and @xmath74 for @xmath75 and @xmath76 yields the two - parameter poisson dirichlet process [ @xcite ] , denoted @xmath77 , with the choice @xmath78 resulting in the dirichlet process [ @xcite ] , denoted @xmath79 . in mixture models such as ( [ eqsimplmodelh ] ) , @xmath61 acts as the common prior for the cluster - specific parameters @xmath80 , while the sequences @xmath81 and @xmath82 control the a priori expected number and size of the clusters . the main advantage of nonparametric mixture models such as the poisson dirichlet process as a clustering tool is that they allow for automatic inferences on the number of components in the mixture . indeed , these models induce a prior probability on all possible partitions of the set of observations , which is updated based on the information contained in the data . however , poisson dirichlet processes have two properties that might be unappealing in our eps application ; first , they place a relatively large probability on partitions that include many small clusters and , second , they imply that the number of clusters will tend to grow logarithmically ( if @xmath78 ) or as a power law ( if @xmath83 ) as more observations are included in the data set . however , priors that favor of increasing numbers of clusters without bound as the number of subjects increase have some disadvantages in terms of interpretability and sparsity in characterizing high - dimensional data . for example , in applying dp mixture models for clustering of the progesterone curves in eps , @xcite obtained approximately 32 different clusters , with half of these clusters singletons . many of the clusters appeared similar , and it may be that this large number of clusters was partly an artifact of the dp prior . @xcite proposed a local partition process prior to reduce dimensionality in characterizing the curves , but this method does not produce easily interpretable functional clusters . hence , it is appealing to use a more flexible global clustering prior that allows the number of clusters to instead converge to a finite constant . with this motivation , we focus on the generalized dirichlet process ( gdp ) introduced by @xcite , denoted @xmath84 . the gdp corresponds to a stick - breaking prior with @xmath72 , @xmath85 and @xmath86 for all @xmath0 . when compared against the poisson dirichlet process , the gdp has quite distinct properties . [ thnumclust ] let @xmath87 be the number of distinct observations in a sample of size @xmath88 from a distribution @xmath59 , where @xmath89 . the expected number of clusters @xmath90 is given by @xmath91 the proof can be seen in appendix [ approofexpnumc ] . note that for @xmath92 , this expression simplifies to @xmath93 , a well - known result for the dirichlet process [ @xcite ] . letting @xmath94 denote the change in the number of clusters in adding the @xmath88th individual to a sample with @xmath95 subjects , stirling s approximation can be used to show that @xmath96 where @xmath97 . hence , @xmath98 as and new clusters become increasingly rare as the sample size increases . note that for @xmath99 , the number of clusters will grow slowly but without bound as @xmath88 increases , with @xmath100 . the rate of growth in this case is proportional to @xmath101 , which is similar to what is obtained by using the poisson dirichlet prior [ @xcite ] . however , when @xmath102 the expected number of clusters instead converges to a finite constant , which is a remarkable difference compared with the dirichlet and poisson dirichlet process . as mentioned above , there may be a number of practical advantages to bounding the number of clusters . in addition , a finite bound on the number clusters seems to be more realistic in many applications , including the original species sampling applications that motivated much of the early development in this area [ @xcite ] . in order to gain further insight into the clustering structure induced by the @xmath84 , we present in figure [ ficluststruc ] the relationship between the size of the largest cluster and the mean number of clusters in the partition ( left panel ) , and the mean cluster size and the number of clusters ( right panel ) for a sample of size @xmath103 . the numbers presented in this figure are based on the results from 20,000 simulations from the stick - breaking prior . each continuous line corresponds to a combination of shape parameters such that @xmath104 is constant , while the dashed line in the plots corresponds to the combinations available under a dirichlet process . the plots demonstrate that the additional parameter in the gdp allows us to simultaneously control the number of clusters and the relative size of the clusters , increasing the flexibility of the model as a clustering procedure . for a sample of size @xmath103 . panel shows the relationship between the size of the largest cluster and the mean number of clusters for different gdps , where each curve shares a common @xmath105 . panel shows the relationship between the average cluster size and the mean number of clusters . the dashed lines correspond to the combinations available under a standard dirichlet process . ] the previous discussion focused on the impact of the prior distribution for the mixture weights on the clustering structure . another important issue in the specification of the model is the selection of the baseline measure @xmath61 . note that in the functional clustering setting @xmath106 and , therefore , a computationally convenient choice that is in line with our previous discussion on basis selection and zero - inflated priors is to write @xmath107 a prior of this form allows differential adaptive smoothing for each cluster in the data ; the level of smoothness is controlled by @xmath33 ( the prior probability of inclusion for each of the spline coefficients ) and , therefore , it is convenient to assign to it a hyperprior such as @xmath108 . consider now the case where multiple curves are collected for each subject in the study . in this case , the observations consist of ordered pairs @xmath109 where @xmath110 where @xmath111 is the @xmath112th functional replicate for subject @xmath7 , with @xmath113 , @xmath114 and @xmath115 . for example , in the eps , @xmath111 is the measurement error - corrected smooth trajectory in the progesterone metabolite pdg over the @xmath112th menstrual cycle from woman @xmath7 , with @xmath116 indexing the sample number and @xmath117 denoting the day within the @xmath118 menstrual cycle relative to a marker of ovulation day . a natural extension of ( [ eqsimplmodelh ] ) to nested designs arises by modeling the expected evolution of progesterone in time for cycle @xmath112 of woman @xmath7 as @xmath119 and using a hierarchical model for the set of curve - specific parameters @xmath120 in order to borrow information across subjects and/or replicates . in the following subsections , we introduce two alternative nonparametric hierarchical priors that avoid parametric assumptions on the distribution of the basis coefficients , while inducing hierarchical functional clustering . as a first approach , we consider a gaussian mixture model , which characterizes the basis coefficients for functional replicate @xmath112 from subject @xmath7 as conditionally independent draws from a gaussian distribution with subject - specific mean and variance , in the spirit of @xcite : @xmath121\\[-8pt ] g_i & = & \mathsf{n}\bigl ( { \bolds{\lambda}}_i { \bolds{\theta}}_i , \sigma^2_{i } { \bolds{\sigma}}\bigr),\nonumber\end{aligned}\ ] ] where @xmath122 are as described in expression ( [ eqsimplmodelh ] ) and @xmath123 is a @xmath124 covariance matrix . in this model , the average curve for subject @xmath7 is obtained as @xmath125 , with @xmath126 providing a mechanism for subject - specific basis selection , so that the curves from subject @xmath7 only depend on the basis functions corresponding to nonzero diagonal elements of @xmath126 . the variability in the replicate curves for the same subject is controlled by @xmath127 , with the subject - specific multiplier allowing subjects to vary in the degree of variability across the replicates . the need to allow such variability is well justified in the hormone curve application . in order to borrow information across women , we need a hyperprior for the woman specific parameters @xmath128 . since we are interested in clustering subjects , a natural approach is to specify this hyperprior nonparametrically through a generalized dirichlet process centered around the baseline measure in ( [ eqbaselme ] ) , just as we did for the single curve case . this yields @xmath129 with @xmath61 given in ( [ eqbaselme ] ) . since the distribution @xmath59 is almost surely discrete , the model identifies clusters of women with similar average curves . this is clearer if we integrate the curve - specific coefficients @xmath130 and the unknown distribution @xmath59 out of the model to obtain @xmath131\\[-8pt]\nonumber & & \qquad \sim \sum_{k=1}^{k } w_k \biggl\ { \prod_{j=1}^{n_i } \mathsf{n } \bigl ( { \mathbf{b } } ( { \mathbf{x}}_{ij } ) { \bolds{\lambda}}_k^ { * } { \bolds{\theta}}_k^ { * } , \sigma^{*2}_{k } ( \mathbf{i } + { \bolds{\sigma } } ) \bigr ) \biggr\}.\end{aligned}\ ] ] by incorporating the distribution of the selection matrices @xmath132 in the random distribution @xmath59 , this model allows for a different smoothing pattern for each cluster of curves . this is an important difference with a straight generalization of the model in @xcite , who instead treat the selection matrix as a hyperparameter in the baseline measure @xmath61 and therefore induce a common smoothing pattern across all clusters . the model is completed by assigning priors for the hyperparameters . for the random effect variances we take inverse - wishart priors : @xmath133 in the spirit of the unit information priors [ @xcite ] , the hyperparameters for these priors can be chosen so that @xmath134 and @xmath135 are proportional to @xmath136 finally , the concentration parameters @xmath137 and @xmath138 are given gamma priors @xmath139 and @xmath140 and the probability of inclusion @xmath33 is assigned a beta prior , @xmath141 . because the subject - specific distributions @xmath142 were assumed to be gaussian and the nonparametric prior was placed on their means , the model in the previous section clusters subjects based on their average profile . however , as we discussed in section [ seintro ] , clustering based on the mean profiles might be misleading in studies such as the eps in which there are important differences among subjects in not only the mean curve but also the distribution about the mean . in hormone curve applications , it is useful to identify clusters of trajectories over the menstrual cycle to study variability in the curves and identify outlying cycles that may have reproductive dysfunction . it is also useful to cluster women based not simply on the average curve but on the distribution of curves . with this motivation in mind , we generalize our hierarchical nonparametric specification to construct a model with additional parameters that enables clustering based on more than proximity between the mean curves . this generalization further enables clustering within subjects as well as subjects . to motivate our nonparametric construction , consider first the simpler case in which there are only two types of curves in each cluster of women ( say , normal and abnormal ) , so that it is natural to model the subject - specific distribution as a two - component mixture where @xmath143\\[-8pt]\nonumber & & \qquad \sim \varpi_i \mathsf{n}\bigl ( { \mathbf{b}}({\mathbf{x}}_{ij } ) { { \bolds{\lambda}}}_{1i } { { \bolds{\theta}}}_{1i } , { \sigma}^2_{1i } \mathbf{i } \bigr ) + ( 1-\varpi_i ) \mathsf{n}\bigl({\mathbf{b } } ( { \mathbf{x}}_{ij } ) { { \bolds{\lambda}}}_{2i } { { \bolds{\theta}}}_{2i } , { \sigma}^2_{2i } \mathbf{i } \bigr),\end{aligned}\ ] ] @xmath144 can be interpreted as the proportion of curves from subject @xmath7 that are in group 1 ( say , normal ) , and @xmath145 are the parameters that describe curves from a normal cycle , @xmath146 is a diagonal variable selection matrix for the normal cycles that zeroes out unnecessary coefficients to avoid overfitting , @xmath147 are the parameters describing the curves from an abnormal cycle , and @xmath148 is the variable selection matrix for the abnormal cycles . note that in this case we have not one but two variance parameters for each individual , which provide additional flexibility by allowing each cluster of curves to present a different level of observational noise . this feature is desirable in the eps because , for a given woman , observational noise in abnormal cycles tends to be larger than in normal cycles . under this formulation , the subject - specific distribution is described by the vector of parameters @xmath149 , and clustering subjects could be accomplished by clustering these vectors . we can accomplish this by using another mixture model that mimics ( [ eqsimplmodel ] ) and ( [ eqjointliksinglecurve ] ) , so that @xmath150 to formulate our bayesian nonparametric model for clustering , we start by rewriting ( [ eqtwocomponent ] ) as a general mixture model where @xmath151 and @xmath152 is a discrete distribution which is assigned a nonparametric prior . note that this is analogous to the formulation in ( [ eqmeanclustre ] ) , but by replacing the gaussian distribution with a random distribution with a nonparametric prior we are modeling the within - subject variability by clustering curves into groups with homogeneous shape . now , we need to define a prior over the collection @xmath153 that induces clustering among the distributions . for example , we could use a discrete distribution _ whose atoms are in turn random distributions _ , for example , @xmath154 where @xmath155 , @xmath156 and @xmath157 independently . this implies that @xmath158 with @xmath159 and @xmath160 and @xmath61 as in ( [ eqbaselme ] ) . therefore , if we were to replace the collection @xmath161 with random discrete distributions with only two atoms and we were to integrate over the random distributions @xmath153 , this model would be equivalent to ( [ eqtwocomponent ] ) with @xmath72 . this model on the collection @xmath153 is a generalization of the nested dirichlet process introduced in @xcite and , as with other models based on nested nonparametric processes , interesting special cases can be obtained by considering the limit of the precision parameters . for example , letting @xmath162 while keeping @xmath163 fixed induces a model where all menstrual cycles within a woman are assumed to have the same profile , and subjects are clustered according to their mean cycle . such a model is equivalent to the one obtained by taking @xmath164 in ( [ eqmeanclustre ] ) . on the other hand , by letting @xmath165 while keeping @xmath166 constant , we obtain a model where all subjects are treated as different and menstrual cycles are clustered within each women . in this case , information is borrowed across the menstrual cycles of each women , but not across women . intuitively , we can think of the model we just described as first clustering curves within a subject via a model - based version of @xmath0-means applied to the subject - specific basis coefficients . then , two subjects having similar distributions of curve clusters will be clustered together . because we use a joint hierarchical model , these stages are done simultaneously . again , the model is completed by specifying prior distributions on the remaining parameters . as before , we let @xmath167 , @xmath168 and @xmath108 , providing a conditionally conjugate specification amenable for simple computational implementation . finally , for the precision priors of the gdps we set @xmath169 as is commonplace in bayesian inference , we resort to markov chain monte carlo ( mcmc ) algorithms [ @xcite ] for computation in our functional clustering models . given an initial guess for all unknown parameters in the model , the algorithms proceed by sequentially sampling blocks of parameters from their full conditional distributions . in particular , we design our algorithms using truncated versions of the gdp and the nested gdp , where a large but finite number of atoms are used to approximate the nonparametric mixture distributions . well - known results on the convergence of truncations as the number of atoms grows that were originally presented in @xcite and @xcite can be directly extended to this problem ( see appendix [ aptruncgdp ] ) . furthermore , because most components of the model are conditionally conjugate , most of the full conditional distributions can be directly sampled using gibbs steps . full details of the algorithm can be seen in the online supplementary materials [ @xcite ] . convergence of the mcmc algorithms was assessed using the multi - chain method described in @xcite , which was applied to monitor the ( unnormalized ) posterior distribution , as well as the number of occupied clusters on each level of the model and the frobenius norm of the matrices @xmath43 and @xmath123 . progesterone plays a crucial role in controlling different aspects of reproductive function in women , from fertilization to early development and implantation . therefore , understanding the variability of hormonal profiles across the menstrual cycle and across subjects is important in understanding mechanisms of infertility and early pregnancy loss , as well as for approaches for identifying abnormal menstrual cycles and women for diagnostic purposes . our data , extracted from the early pregnancy study [ @xcite ] , consists of log daily creatinine - corrected concentrations of pregnanediol-3-glucuronide ( pdg ) for 60 women along multiple menstrual cycles , measured in micrograms per milligram of creatinine ( g@xmath170ml cr ) . we focus on day intervals extending from 10 days before ovulation to 2 days after ovulation . according to the results in @xcite , this interval should include the fertile window of the menstrual cycle during which noncontracepting intercourse has a nonneglible probability of resulting in a conception . also , for this illustration we considered only nonconceptive cycles and women with at least four cycles in each record . therefore , the number of curves per woman varies between 4 and 9 . we analyzed the eps data using both the mean - based clustering model described in section [ semeanclustering ] and the distribution - based clustering model of section [ sedistclustering ] . these models where fitted using the algorithms from section [ secompu ] . in the mean - based clustering algorithm , the gdp was truncated so that @xmath171 , while in the distribution - based algorithm the nested gdp was truncated so that @xmath171 and @xmath172 . although these numbers might seem large given the sample sizes involved , a large number of empty components are helpful in improving the mixing of the algorithms . in both cases , we used piecewise linear splines ( @xmath173 ) and @xmath174 knots , corresponding to each of the days considered in the study . all inferences presented in this section are based on 100,000 samples obtained after a burn - in period of 10,000 iterations . prior distributions in the mean - based clustering algorithm were set as follows . for the concentration parameters , we used proper priors @xmath175 and @xmath176 , and for the observational variance , we set @xmath177 , so that @xmath178 . note that setting @xmath179 a priori is natural because ( as we discussed in section [ sebnpmixfunc ] ) the asymptotic behavior of the expected number of clusters is different if @xmath102 or @xmath180 . on the other hand , the prior mean for @xmath181 is based on information from previous studies . to allow uncertainty in the probability of basis selection within the base measure , we let @xmath182 , implying that we expect about one - third of the spline basis functions to be used in any given cluster . priors for the distribution - based clustering algorithm were chosen in a similar way , with @xmath183 , @xmath184 , @xmath185 and @xmath186 , while for the baseline measure we picked a prior with the inclusion probabilities @xmath187 and the prior on the group specific variances as given by @xmath188 . , @xmath138 , @xmath166 , @xmath189 , @xmath163 and @xmath190 . @xmath191 ( subject - level clustering ) , @xmath192 ( curve - level clustering ) . ] to better understand the effect of our prior on @xmath137 , @xmath138 , @xmath166 , @xmath189 , @xmath163 and @xmath190 , we show in figure [ fipriorcomp ] the prior expected number of clusters implied by our choices . in particular , figure [ fipriorcomp](a ) shows that , a priori , we expect between 1 and 8 clusters of subjects with high probability , with the most likely prior value being 3 clusters ( note that this applies to both the mean - based and the distribution - based clustering methods ) . on the other hand , figure [ fipriorcomp](b ) shows that , for a subject for which 7 cycles have been observed , we expect between 1 and 3 clusters with high probability . on the other hand , to assess the effect of the prior choices on the results , we conducted a small sensitivity analysis . in particular , the priors for @xmath137 , @xmath138 , @xmath166 , @xmath163 , @xmath189 and @xmath190 were replaced with exponential distributions with mean 2 . the induced priors on the number of clusters have similar modes to our original specifications but have higher variability , placing substantially more mass on larger number of clusters . although the posterior distribution for these parameters was somewhat affected , we did not see any substantial change in our posterior inferences on the clustering structure or the curve shapes ( which are the main focus of our analysis ) . similarly , we explored the effect of a @xmath193 prior on @xmath33 and an @xmath194 prior on @xmath181 without any significant change in the estimates of the hormonal profiles . we start our analysis by comparing the clustering structure generated by the mean - based and distribution - based models considered in section [ sefunclustnested ] . for this purpose , we show in figures [ fipairhormeanclust ] and [ fipairhordistclust ] heatmaps of the average pairwise clustering probability matrix under these two models . entry @xmath195 of the matrix contains the posterior probability that observations @xmath7 and @xmath112 are assigned to the same cluster . the black squares in the plots correspond to point estimates of the clustering structure obtained through the method described in @xcite . in our case , the point estimate is obtained by minimizing a loss function that assigns equal weights to all pairwise misclassification errors . therefore , the resulting plots provide information about the optimal clustering structure for the data as well as the uncertainty associated with it . figures [ fipairhormeanclust ] and [ fipairhordistclust ] show that the structure of the clusters generated by both models are similar . for example , cluster 1 from mean - based clustering ( counting from the bottom left corner of figure [ fipairhormeanclust ] ) contains similar subjects as cluster 1 from distribution - based clustering ( counting from the bottom left corner of figure [ fipairhordistclust ] ) . the same is true for cluster 2 from both algorithms , and for cluster 5 from the mean - based clustering model and cluster 4 from the distribution - based model . there are also similarities in the structure of the outlier clusters ( corresponding to the small clusters at the top and right of both heat maps ) . both models assign subject 25 to a singleton cluster , while grouping subjects 21 and 53 into a pair , with subjects 40 and 51 in a separate pair . there are also some important differences between the two methods . mean - based clustering groups subjects 36 and 43 together and assigns subject 3 to a different cluster ( figure [ fihorpro ] shows differences in the mean curve of subject 3 ) , while distribution - based cluster groups subjects 3 and 43 together and assigns 36 to a different group . the behavior illustrates the robustness of distribution - based clustering , because except for a single outlier cycle for subject 43 , the curves for this subject are much more similar to those of subject 3 than those of 36 . mean - based clustering also treats subjects 11 and 56 as singletons , while distribution - based clustering assigns them to the large cluster 4 ( counting from the bottom left ) . visually the raw profiles for subjects 11 and 56 are not very different from other profiles in cluster 4 [ see panels ( d ) , ( e ) and ( f ) of figure [ fihorraw2 ] ] , so this behavior seems reasonable . . white corresponds to zero probability , while red corresponds to 1 . the squares correspond to a point estimate of the cluster structure in the data . ] . white corresponds to zero probability , while red corresponds to 1 . the squares correspond to a point estimate of the cluster structure in the data . ] figure [ firechor ] shows reconstructed profiles under the distribution - based clustering model for some representative women in each of the main four groups . most profiles are flat before ovulation , when hormone levels start to increase . also , in most clusters the profiles tend to be relatively consistent for any single woman . however , we can see some outliers , typically corresponding to elevated post - ovulation levels and/or early increases in the hormone levels . cluster 3 corresponds to women with very low hormonal levels , even after ovulation . this group has few outliers , and those present are characterized by a slightly larger increase in progesterone after ovulation , which is still under 1 g@xmath170ml cr . group 2 shows much more diversity in the hormonal profiles , as well as a slightly higher baseline level in progesterone level and an earlier rise in progesterone than group 3 . group 1 tends to show few outliers , and otherwise differs from the previous ones in a higher baseline level and an early and very fast increase in progesterone . finally , group 4 presents `` normal '' cycles with the highest baseline level of progesterone ( 1 g@xmath170ml cr ) and the fastest increase in progesterone after ovulation , along with `` abnormal '' cycles with even higher baseline levels and very extreme levels of progesterone after ovulation ( close to 5 g@xmath170ml cr ) . for comparison we also applied a simple functional clustering approach based on kernel smoothing and hierarchical clustering . we first fitted a gaussian kernel smoother [ @xcite ] to each of the curves , generated fitted values over a common grid ( in this example we used 6 equally - spaced knots ) , then computed the average predicted values for each subject at each point of the grid , and finally applied complete - linkage hierarchical clustering ( as implemented in the ` r ` package ` mclust ` ) , with the number of clusters selected using bic . this approach identified 7 clusters ; figure [ fidendogram ] shows the associated dendogram , where colors are used to represent the clusters . some of these clusters are similar to the ones identified by the mean - based clustering model . for example , the small cluster of 5 subjects ( 3 , 10 , 37 , 45 and 48 ) , corresponding to the fourth cluster from the bottom left in figure [ fipriorcomp ] , is perfectly identified by hierarchical clustering . similarly , subjects 21 , 25 , 40 , 51 and 53 , which are all identified as outliers by the mean - based clustering model , are allocated to a single cluster by the hierarchical clustering approach . however , the majority of the clustering pattern generated by this method is quite different from the one obtained with mean - based or the distribution - based clustering . furthermore , the number and structure of clusters depends heavily on the number and location of grid points used to interpolate the functions . for example , when the interpolation grid contains 5 knots we obtain 8 clusters , while 18 clusters are obtained when 11 knots are used for interpolation . part of this difference could be due to the fact that , because most of the missing values tend to concentrate at the beginning of the cycle , many of the fitted functional values correspond to extrapolations rather than interpolations when a large number of knots are used . we also note that posterior estimates of the precision parameters on the gdp suggest that a logarithmic rate of growth for the expected number of clusters might be reasonable for this data . for the mean - based clustering , the posterior mean for @xmath137 was @xmath196 and the 95% posterior symmetric credible interval was @xmath197 , while the posterior mean for @xmath138 was @xmath198 with 95% credible interval @xmath199 . for the distribution - based clustering , the corresponding estimates are @xmath200 @xmath201 and @xmath202 @xmath203 for @xmath166 and @xmath189 , and @xmath204 @xmath205 and @xmath206 @xmath207 for @xmath163 and @xmath190 . finally , we investigated the relationship between the clusters we identified using our distribution - based clustering algorithm and a series of covariates available for each subject , including age at the beginning of the study ( age ) , age of first menses ( age_mense ) , average length of menses ( mense_length ) , race ( race ) , body mass index ( bmi ) , whether the subject had been taking contraceptives before the start of the study ( ev_pill ) and whether the subject self reports as a consumer of marijuana during the period of the study ( marijuana_use ) . to investigate this relationship , we used a classification tree , as implemented in the ` r ` package ` rpart ` . to simplify interpretation , we merged all the small outlier clusters into a single group ( which we label cluster 5 ) . the resulting classification tree ( obtained using a cp value of 0.03 for pruning ) can be seen in figure [ ficlasstree ] . note that race and marijuana consumption do not appear to play any role in predicting the shape of the hormonal profiles . on the other hand , bmi , along with age , menses length and menses age , seems to have some predictive capability . for example , a bmi higher than 22 and an age of menses under 12 years old seem to predict progesterone profiles like the ones observed in cluster 3 . we have presented two approaches to functional clustering in nested designs . these approaches look into different features of the nested samples and are therefore applicable in different circumstances . our mean - based clustering approach is easier to interpret and provides an excellent alternative when within - subject samples are homogenous . however , when within - subject curves are heterogeneous , mean - based clustering can lead to biased results . therefore , in studies such as the eps , distribution - based models such as the one described here provide a viable alternative that acknowledges the heterogeneity in the function replicates from a subject . one interesting insight that can be gathered from the results of the eps data is that , for small numbers of functional replicates per subject and rare outliers , the effect of the distribution - based clustering is to perform clustering based on the modal rather than the mean profile . that is , the distribution - based clustering model is able to automatically discount the abnormal curves , leading to more appropriate clustering patterns if the effect of outliers needs to be removed . naturally , this perceived advantage of the distribution - based clustering method implicitly assumes that abnormal curves should be discounted . although that assumption is justified in our application , users should be aware of it when applying our method to other data sets . let @xmath208 be a sequence of independent and identically distributed samples from a random distribution @xmath59 , which follows a @xmath84 distribution . also , let @xmath209 be 1 if @xmath210 is different from every @xmath211 , and zero otherwise . clearly , @xmath212 is the number of distinct values among the first @xmath88 samples form a @xmath84 . @xcite shows that @xmath213 which completes the proof . [ thbounds ] assume that samples of @xmath88 observations have been collected for each of @xmath214 distributions and are contained in vector @xmath215 . also , let @xmath216 be , respectively , the prior distribution of the model parameters under the nested gdp model and its corresponding truncation after integrating out the random distributions , and @xmath217 and @xmath218 be the prior predictive distribution of the observations derived from these priors . then @xmath219 where @xmath220^{j } \biggr),\hspace*{25pt}\mbox{if } l = \infty , k < \infty , \vspace*{8pt}\cr \displaystyle4 \biggl ( 1 - \biggl [ 1 - \biggl ( \frac{b_2}{a_2+b_2 } \biggr)^{l-1 } \biggr]^{nj } \biggr),\qquad\mbox{if } l < \infty , k=\infty , \vspace*{8pt}\cr \displaystyle4 \biggl ( 1 - \biggl [ 1 - \biggl ( \frac{b_1}{a_1+b_1 } \biggr)^{k-1 } \biggr]^{j } \biggl [ 1 - \biggl ( \frac{b_2}{a_2+b_2 } \biggr)^{l-1 } \biggr]^{nj } \biggr ) , \vspace*{5pt}\cr \hspace*{170pt}\mbox{if } l < \infty , k < \infty.}\end{aligned}\ ] ] the proof is a direct extension of results in @xcite ( @xcite ) and @xcite and it is omitted for reasons of space . this result is particularly important since it justifies the use of computational algorithms based on finite mixtures and allows us to choose adequate truncation levels . we would like to thank nils hjort for his helpful comments and allen wilcox for providing access to the eps data .
we discuss functional clustering procedures for nested designs , where multiple curves are collected for each subject in the study . we start by considering the application of standard functional clustering tools to this problem , which leads to groupings based on the average profile for each subject . after discussing some of the shortcomings of this approach , we present a mixture model based on a generalization of the nested dirichlet process that clusters subjects based on the distribution of their curves . by using mixtures of generalized dirichlet processes , the model induces a much more flexible prior on the partition structure than other popular model - based clustering methods , allowing for different rates of introduction of new clusters as the number of observations increases . the methods are illustrated using hormone profiles from multiple menstrual cycles collected for women in the early pregnancy study .
it is well - known that the julia set @xmath3 of a transcendental entire function @xmath4 can be the whole complex plane . ( basic definitions and notations are reviewed in section [ sec_prel ] . ) as far as we know , there is no function with this property for which the topological dynamics has been completely understood . the results in this paper provide such a description for a wide class of examples , including maps such as @xmath8 . ( a description of the _ combinatorial _ dynamics of the latter map was previously given by schleicher in @xcite . ) in particular , we give an answer to the question of bergweiler ( personal communication ) whether the escaping set @xmath9 of a cosine map @xmath10 with strictly preperiodic critical values is connected : this is not the case ; see corollary [ cor_bergweiler ] below . ( on the other hand , there are entire maps whose julia set equals @xmath11 but for which the escaping set is connected ; see @xcite . ) in fact , our results are considerably more general . a transcendental entire map @xmath4 is called _ subhyperbolic _ if the intersection of the fatou set @xmath1 and the postsingular set @xmath2 is compact and the intersection of the julia set @xmath3 and @xmath2 is finite . a subhyperbolic map is called _ hyperbolic _ if @xmath12 . we are interested in the following class of subhyperbolic functions which includes all hyperbolic maps . a subhyperbolic transcendental entire map @xmath4 is called _ strongly subhyperbolic _ if @xmath3 contains no asymptotic values of @xmath4 and the local degree of @xmath4 at the points in @xmath3 is bounded by some finite constant . note that the map @xmath13 mentioned previously is strongly subhyperbolic , since @xmath14 is finite , all critical points are simple and there are no asymptotic values . our main theorem describes the julia set of any strongly subhyperbolic entire function as a quotient of the julia set of a ( particularly simple ) hyperbolic function in the same parameter space . [ thm_maintheorem ] let @xmath4 be strongly subhyperbolic , and let @xmath15 be such that @xmath16 is hyperbolic with connected fatou set . then there exists a continuous surjection @xmath17 , such that @xmath18 for all @xmath19 . moreover , @xmath20 restricts to a homeomorphism between the escaping sets @xmath7 and @xmath21 . the hypothesis will be automatically satisfied whenever @xmath22 is sufficiently small . also , any two maps @xmath6 and @xmath23 as in the theorem are quasiconformally conjugate on a neighbourhood of their julia sets @xcite , so it is sufficient to prove the theorem for any such map . as in @xcite , we say that a hyperbolic function @xmath6 with connected fatou set is of _ disjoint type_. for simple families , such as @xmath24 , the dynamics of disjoint type functions is well - understood . hence theorem [ thm_maintheorem ] extends this understanding to all strongly subhyperbolic functions in these families . ( in appendix a , we present a detailed description of the topological dynamics of @xmath25 . ) more generally , suppose that @xmath6 is of disjoint type and has _ finite order _ , i.e. , @xmath26 as @xmath27 , or , more generally , can be written as a finite composition of finite - order maps with bounded singular sets . then it is known that @xmath28 is a _ cantor bouquet _ , i.e. , homeomorphic to a _ straight brush _ in the sense of @xcite . [ cor1 ] let @xmath29 be a strongly subhyperbolic map , where every @xmath30 is an entire map with bounded set of singular values and with finite order of growth . then @xmath3 is a _ pinched cantor bouquet _ ; that is , the quotient of a cantor bouquet by a closed equivalence relation defined on its endpoints . the escaping set of a disjoint type entire function is always disconnected , so we also have the following corollary , settling bergweiler s question for all strongly subhyperbolic maps . [ cor_bergweiler ] the escaping set of a strongly subhyperbolic transcendental entire function is disconnected . let us comment on the assumption of strong subhyperbolicity . theorem [ thm_maintheorem ] is not true for all subhyperbolic maps : it is known that @xmath31 ( disjoint type ) and @xmath32 ( subhyperbolic ) are not topologically conjugate on their escaping sets ( * ? ? ? * proposition 2.1 ) . ( in fact , for exponential maps the escaping set consists of curves to infinity , called _ dynamic rays _ @xcite . for @xmath33 , all these rays have a landing point in @xmath11 , while for @xmath34 there are uncountably many dynamic rays , each of which accumulates everywhere upon itself . ) rempe asked whether two cosine maps with strictly preperiodic critical values can be conjugate on their escaping sets ( * ? ? ? * question 12.1 ) . theorem [ thm_maintheorem ] together with the mentioned result in @xcite on disjoint type maps gives an affirmative answer to this question . for cosine maps @xmath35 that are subhyperbolic , theorem [ thm_maintheorem ] implies that every point in the julia set is either on a dynamic ray or the landing point of a dynamic ray . this has already been shown by schleicher @xcite when both critical values of @xmath36 are assumed to be preperiodic . nonetheless , his results do not explain the topological embedding of the escaping set of such a map in the complex plane . furthermore , theorem [ thm_maintheorem ] formulated for subhyperbolic cosine maps can be proved in a concise and fairly elementary way , which is why we have included the modifications of our proof for this special case in section [ subs_cosine ] . for hyperbolic maps , theorem [ thm_maintheorem ] is due to rempe , and our proof is in the spirit of the ideas presented in @xcite . however , the attempt to transfer the construction in the hyperbolic case to the setting of strongly subhyperbolic maps fails due to the existence of singular values in the julia sets . this obstruction is overcome by studying julia sets as subsets of _ hyperbolic riemann orbifolds_. these can be thought of as images of the unit disk under branched coverings , for which the set of critical values is discrete and `` tame '' . this yields a description of a riemann orbifold as a riemann surface together with a discrete set of _ ramified points _ , each of which has finite ramification value . the use of orbifolds in dynamics goes back to thurston and has been used with great success by douady and hubbard in their work on subhyperbolic rational maps . the following result is not only crucial for the proof of theorem [ thm_maintheorem ] but also interesting in its own right , since it provides us with a global estimate of the hyperbolic metric on certain hyperbolic riemann orbifolds . [ thm_main2 ] let @xmath37 and let @xmath38 be an infinite sequence of points satisfying @xmath39 . let @xmath40 be the orbifold with @xmath11 as the underlying surface and whose ramified points are the points @xmath38 with ramification value @xmath41 . then the density @xmath42 of the hyperbolic metric on @xmath40 satisfies @xmath43 the orbifold for which the set of ramified points is given by @xmath44 shows that our estimate is best possible ( see proof of proposition [ prop_cos_est ] ) . if we replace the ramified points by punctures , i.e. , if we consider the hyperbolic domain @xmath45 instead of the orbifold @xmath40 , the same bound for the asymptotic behaviour of the density map near @xmath46 can be obtained using standard estimates of the hyperbolic metric in the twice - punctured plane ( * ? ? ? * lemma 2.1 ) . it will become clear in section [ sec associate ] why strongly subhyperbolic maps are exactly those maps which can be approached with orbifold theory . this still leaves open the question what can be said about topological dynamics of maps that are subhyperbolic but not strongly subhyperbolic . we believe that if @xmath3 contains an asymptotic value of @xmath4 then , similar to the case of exponential maps , there is no conjugacy between @xmath4 and any ( suitable ) disjoint type map @xmath6 on their escaping sets ; this is work in progress . however , we have no indication of what to expect for maps whose julia sets contain no asymptotic values but sequences of points with unbounded local degree . it would be very interesting to explore this problem , in particular since there are prominent examples of such maps like _ poincar functions _ corresponding to certain hyperbolic polynomials ; an elaboration of such an example is given in appendix b. in section [ sec associate ] we develop the concept of orbifolds _ dynamically associated _ to a strongly subhyperbolic map @xmath4 ; the main consequence is that we obtain a hyperbolic orbifold @xmath47 such that @xmath4 is expanding with respect to the corresponding hyperbolic metric . later , in section [ sec uniform ] , we prove that the expansion of @xmath4 is actually uniform . the key for this will be ( the proof of ) theorem [ thm_main2 ] . finally , section [ sec semiconjugacy ] addresses the construction of the semiconjugacy itself . my special thanks go to my supervisor , lasse rempe for introducing me to the research problem and for his continuous help and support . furthermore , i would like to thank adam epstein and mary rees for interesting and helpful discussions . if not stated differently , we will assume throughout this article that the considered maps are transcendental entire . we denote the complex plane by @xmath11 , the riemann sphere by @xmath48 , and the punctured plane by @xmath49 . we write @xmath50 for the unit disk and @xmath51 for the upper half - plane . the euclidean disk centred at @xmath52 with radius @xmath53 is denoted by @xmath54 . if not stated differently , the boundary @xmath55 and the closure @xmath56 of a set @xmath57 is always understood to be taken relative to the complex plane . the euclidean distance between two sets @xmath58 will be denoted by @xmath59 . for a sequence @xmath60 of natural numbers we write @xmath61 for their least common multiple . the _ fatou set _ of a map @xmath4 is the set of all points in @xmath11 that have a neighbourhood in which the iterates @xmath62 form a normal family in the sense of montel . its complement @xmath63 is called the _ julia set _ of @xmath4 . recall that the _ escaping set _ of @xmath4 is given by @xmath64 we say that a point @xmath65 is a _ periodic point _ of @xmath4 if there exists an integer @xmath66 such that @xmath67 . the smallest @xmath68 with this property is called the _ period _ of @xmath69 . a periodic point of period one is called a _ fixed point_. we call a point @xmath69 _ preperiodic _ under @xmath4 if some image @xmath70 , @xmath71 , of @xmath69 is periodic . note that every periodic point is also preperiodic . to avoid confusion , we say that a point @xmath69 is _ strictly preperiodic _ if it is preperiodic but not periodic . let @xmath69 be a periodic point of @xmath4 of period @xmath68 . we call @xmath72 the _ multiplier _ of @xmath69 . a periodic point @xmath69 is called _ attracting _ if @xmath73 , _ indifferent _ if @xmath74 and _ repelling _ if @xmath75 . since the multiplier of an indifferent periodic point is of the from @xmath76 with @xmath77 , we can distinguish between _ rationally _ and _ irrationally indifferent _ points , according to whether @xmath78 is rational or not . the set of all points whose orbits converge to an attracting periodic cycle is called the _ attracting basin _ of this cycle . we denote the set of all critical points of @xmath4 by @xmath79 , the set of all critical values by @xmath80 and the set of all ( finite ) asymptotic values by @xmath81 . the set of _ singular values _ of @xmath4 , denoted by @xmath82 , is the smallest closed set such that @xmath83 is a covering map . it is well - known that @xmath84 . finally , we denote the _ postsingular set _ of @xmath4 by @xmath85 . for more background on holomorphic dynamics we refer the reader to @xcite . an _ orbifold _ is a space which is locally modelled on the quotient of an open set in @xmath86 by the linear action of a finite group . for a general introduction see ( * ? ? ? * @xmath87 13 ) . throughout this article we will need only orbifolds modelled on riemann surfaces , and for a more detailed introduction to this topic see e.g. @xcite . [ def r o ] a _ riemann orbifold _ is a pair @xmath88 , where @xmath89 is a riemann surface and @xmath90 is a map called the _ ramification map _ , such that @xmath91 is discrete . a point @xmath92 with @xmath93 is called a _ ramified _ or _ marked point_. the _ signature _ of an orbifold is the list of values that the ramification map @xmath94 assumes at the ramified points , where a value is repeated as often as it occurs as @xmath95 for some ramified @xmath92 . a traditional riemann surface can be regarded as a riemann orbifold with ramification map @xmath96 . throughout this article , whenever we use the expression orbifold , we will always mean a riemann orbifold . recall that for a holomorphic map @xmath97 between riemann surfaces , the _ local degree _ @xmath98 of @xmath4 at a point @xmath99 is the unique integer @xmath100 , such that @xmath101 and @xmath102 . thus @xmath103 is a _ critical _ or _ branch point _ if and only if @xmath104 . the map @xmath4 is called a _ branched covering map _ if every point in @xmath89 has a connected neighbourhood @xmath105 such that @xmath4 maps any component of @xmath106 onto @xmath105 as a proper map . recall that a map @xmath107 is called _ proper _ if the preimage @xmath108 of any compact set @xmath109 is a compact subset of @xmath110 . let @xmath111 and @xmath112 be riemann orbifolds . a _ holomorphic map _ @xmath113 is a holomorphic map @xmath97 between the underlying riemann surfaces such that , for each @xmath114 , @xmath115 if @xmath97 is a branched covering map with @xmath116 for all @xmath117 , then @xmath113 is an _ orbifold covering map_. if additionally the surface @xmath118 is simply - connected , then we call @xmath119 a _ universal covering orbifold _ of @xmath40 . in the standard terminology , where an orbifold is defined via atlases and group actions , the definition of a holomorphic map @xmath4 between two orbifolds is equivalent to a `` local lifting property '' ; if @xmath4 is a covering then every such local lift can be chosen to be an embedding . for more details , see ( * ? ? ? * a2 ) . note that if @xmath120 is a covering then this is not necessarily true for the map @xmath121 between the underlying surfaces . recall that by the uniformization theorem for riemann surfaces , every riemann surface has a universal cover that is conformally equivalent to either @xmath122 , @xmath11 or @xmath50 . the following theorem tells us that the same is true for almost all riemann orbifolds . * theorem a2 ) [ uniform ] let @xmath123 be a riemann orbifold . then @xmath40 has no universal covering orbifold if and only if @xmath40 is isomorphic to @xmath122 with signature @xmath124 or @xmath125 , where @xmath126 . in all other cases the universal cover is unique up to conformal isomorphism over the surface @xmath89 and hence given by either @xmath122 , @xmath11 or @xmath50 . in analogy to riemann surfaces , we will call an orbifold @xmath40 _ elliptic , parabolic _ or _ hyperbolic _ if it is covered by @xmath127 or @xmath50 , respectively . let @xmath128 be orbifolds that have a universal cover . then a map @xmath120 is a covering if and only if it lifts to a conformal isomorphism between the universal covering spaces ( * ? ? ? * lemma e.2 ) . for a connected orbifold @xmath123 the _ euler characteristic _ @xmath129 is given by the equation @xmath130 where @xmath131 denotes the euler characteristic of the surface @xmath89 . note that ramified point cause a reduction of @xmath129 . as for riemann surfaces , a riemann orbifold with negative euler characteristic is always hyperbolic . this also implies that , roughly speaking , most orbifolds are hyperbolic . for the precise list of spherical and parabolic orbifolds , see the details of ( * ? ? ? * theorem a2 ) . let @xmath132 be the uniformized universal covering surface of @xmath40 ( i.e. , @xmath133 ) and denote by @xmath134 its unique complete conformal metric of constant curvature @xmath135 , @xmath136 or @xmath137 , respectively . by pushing forward this metric by a universal covering map we obtain a riemannian metric on @xmath40 that can be written as @xmath138 ( in terms of a local uniformizing parameter @xmath139 ) , and @xmath140 is nonzero and smooth except at the ramified points of @xmath40 . we call this metric the _ orbifold metric _ of @xmath40 . observe that at a ramified point , say @xmath141 , with ramification value @xmath142 , the density has a singularity of the type @xmath143 more precisely , if we choose a local branched covering near @xmath136 , e.g. , @xmath144 , then the induced metric @xmath145 is smooth and nonsingular throughout some neighbourhood of @xmath136 in the @xmath69-plane . note that @xmath138 is again a complete metric with constant curvature @xmath135 , @xmath136 or @xmath137 , respectively , everywhere except at the marked points ( which are singularities of the curvature ) . in this article we are mainly interested in hyperbolic riemann orbifolds . the well - known pick theorem for hyperbolic surfaces generalizes to hyperbolic orbifolds as well and will be of great use for us . * proposition 17.4 ) [ pick ] a holomorphic map between two hyperbolic orbifolds can never increase distances as measured in the hyperbolic orbifold metric . distances are strictly decreased , unless the map is a covering map ; in this case it is a local isometry . in particular , if @xmath119 and @xmath40 are two orbifolds such that @xmath146 is holomorphic , then @xmath147 for all @xmath148 . recall that our main result is the existence of a semiconjugacy between a strongly subhyperbolic and a disjoint type map on their julia sets . we will now recall the definitions and present briefly the relevant dynamical properties of those types of functions . [ def_hyp ] a transcendental entire function @xmath6 is called _ hyperbolic _ if @xmath149 is a compact subset of @xmath150 . it follows from classical results in holomorphic dynamics that the fatou set of a hyperbolic ( transcendental entire ) map is non - empty and consists of attracting basins that correspond to finitely many attracting cycles . a hyperbolic _ rational _ map is classically defined as a function which _ expands _ a conformal riemannian metric defined on a neighbourhood of its julia set . for entire transcendental maps one can give a similar description : it is easy to derive from definition [ def_hyp ] that if @xmath6 is hyperbolic , then there exists a bounded neighbourhood @xmath151 of @xmath149 such that @xmath152 . note that by montel s theorem , @xmath153 is a normal family , hence @xmath154 is a neighbourhood of @xmath28 . furthermore , the map @xmath155 is then a covering map which _ uniformly expands _ the hyperbolic metric on the domain @xmath154 ( * ? ? ? * lemma 5.1 ) . note that not every hyperbolic rational map satisfies definition [ def_hyp ] : by ( * ? ? ? * theorem 19.1 ) , a rational map is hyperbolic if and only if @xmath156 or , equivalently , every critical point converges to an attracting periodic orbit . in certain cases ( including all nonconstant polynomials ) , the point at @xmath46 can be such an attractor as well . [ disjoint type ] a hyperbolic transcendental entire map @xmath6 is said to be of _ disjoint type _ if @xmath150 is connected . since every component of the fatou set of a hyperbolic transcendental entire map is simply - connected ( * ? ? ? * proposition 3 , theorem 1 ) , it follows that the fatou set of a disjoint type map is connected and simply - connected . [ prop_dt ] the following statements are equivalent : 1 . the map @xmath6 is of disjoint type . 2 . @xmath6 has a unique attracting fixed point and @xmath149 is a compact subset of its immediate basin of attraction . 3 . there exists a bounded jordan domain @xmath157 such that @xmath152 . [ prop_disjoint_type ] let @xmath6 be of disjoint type . in particular , @xmath6 is hyperbolic . hence @xmath149 is a compact subset of @xmath150 . by definition , @xmath150 is connected , hence @xmath149 is a compact subset of a completely invariant component of @xmath150 , which can only be the immediate attracting basin of an attracting fixed point of @xmath6 , showing that @xmath158 implies @xmath159 . we will just give a sketch of @xmath159 implies @xmath160 , for more details see e.g. proof of proposition 3.7 in @xcite . let @xmath103 be the unique attracting fixed point of @xmath6 with immediate attracting basin @xmath161 . note that @xmath149 must contain @xmath103 . furthermore , @xmath149 has positive distance to @xmath162 , hence there is a simply - connected domain @xmath163 compactly contained in @xmath164 that contains a neighbourhood of @xmath149 . let @xmath105 be a linearising neighbourhood of @xmath103 and let @xmath68 be the smallest integer such that @xmath165 . taking the union of @xmath166 and the compact sets @xmath167 we obtain a compact connected set @xmath168 that is mapped into its interior . @xmath168 is not necessarily simply - connected ; by the open mapping theorem , we can fill the holes and obtain a full set @xmath169 that is mapped into its interior . now we choose @xmath151 to be a jordan domain such that @xmath170 . to see that @xmath160 implies @xmath158 , let us choose a domain @xmath157 such that @xmath152 . by montel s theorem , @xmath151 is contained in a component of @xmath150 , so , in particular , @xmath6 is hyperbolic and @xmath150 is the union of attracting basins . since every immediate attracting basin contains at least one singular value ( * ? ? ? * theorem 7 ) , @xmath6 has a unique attracting cycle , which is a fixed point contained in @xmath151 . hence every point @xmath171 is eventually mapped into @xmath151 , showing that @xmath172 . on the other hand , @xmath173 and @xmath150 is connected . let @xmath6 be a transcendental entire map with bounded singular set @xmath174 , let @xmath175 be a bounded jordan domain containing @xmath174 and let @xmath176 . then @xmath177 is a covering map and each component @xmath178 of @xmath179 is a simply connected unbounded jordan domain ; we call every such domain @xmath178 a _ tract _ of @xmath6 . note that if @xmath6 is of disjoint type , then by proposition [ prop_disjoint_type ] we can choose @xmath151 such that @xmath180 is disjoint from @xmath181 . using such a domain one can easily prove the following ( well - known ) property of the escaping set of a disjoint type map . [ thm_esc_set_dis_type ] let @xmath6 be of a disjoint type . then @xmath7 is disconnected . let @xmath182 be a jordan domain such that @xmath180 is disjoint from @xmath181 and let @xmath183 be a simple curve that connects @xmath184 to @xmath46 . ( such a curve always exists since the boundary of the open set @xmath185 in @xmath122 is locally connected . ) the preimages of @xmath186 split every tract of @xmath6 ( w.r.t . @xmath151 ) in simply - connected unbounded domains on which @xmath6 restricts as a conformal isomorphism ; we call every such domain a fundamental domain . since @xmath187 , it follows that @xmath7 is contained in the union of the fundamental domains . by ( * ? ? ? * theorem ) , there exists a point @xmath188 . hence every fundamental domain must intersect @xmath7 since it contains a preimage of @xmath139 . let @xmath189 be an arbitrary but fixed fundamental domain and let @xmath105 denote the union of all fundamental domains other than @xmath189 . then @xmath105 and @xmath189 are two disjoint nonempty open sets whose union covers @xmath7 . thus @xmath7 is disconnected . [ dfn_subh ] a transcendental entire function @xmath4 is called _ subhyperbolic _ if * @xmath190 is finite , * @xmath191 is compact . it follows again from classical results that if @xmath4 is subhyperbolic , then @xmath1 is either empty or consists of attracting basins , and by condition @xmath192 there can be only finitely many attracting cycles . by condition @xmath193 , every singular value of @xmath4 in @xmath3 is preperiodic and every critical point in @xmath3 is strictly preperiodic . it also follows that @xmath4 has no cremer points , i.e. , irrationally indifferent periodic points that lie in the julia set . this means that each singular value in @xmath3 eventually lands on some repelling orbit ( for a proof of the above statements see e.g. ( * ? ? ? * proposition 2.5 ) ) . a rational map is said to be subhyperbolic if it is _ expanding _ with respect to an orbifold metric . this characterization is equivalent to saying that every critical orbit is finite or converges to an attracting periodic orbit . note that the postsingular set of a subhyperbolic rational function does not have to be bounded . however , for an _ arbitrary _ subhyperbolic transcendental map it is not possible to define an orbifold metric on a neighbourhood of the julia which is expanded by the map . as we will see later , strongly subhyperbolic maps are exactly those maps for which we can construct the required orbifold metric . for completeness of this section , let us recall the definition of a strongly subhyperbolic map . [ dfn_tame ] a subhyperbolic transcendental entire map @xmath4 is called _ strongly subhyperbolic _ , if @xmath194 and there is a constant @xmath195 such that @xmath196 holds for all @xmath197 . let @xmath4 be a strongly subhyperbolic map . the first step towards the proof of theorem [ thm_maintheorem ] is to find hyperbolic orbifolds @xmath198 such that @xmath199 is expanding with respect to the hyperbolic metric on @xmath47 . we will make use of the following simple observation . [ expansion ] let @xmath123 and @xmath111 be hyperbolic orbifolds with metrics @xmath200 and @xmath201 , respectively . let @xmath113 be a covering map and assume that the inclusion @xmath146 is holomorphic but not a covering . then @xmath202 wherever this is defined . by theorem [ pick ] @xmath4 is a local isometry , hence @xmath203 since the inclusion is only holomorphic , it is a strict contraction , and hence @xmath204 . so altogether , @xmath205 implying that @xmath206 . in order to find hyperbolic orbifolds @xmath47 and @xmath207 such that the assumptions of proposition [ expansion ] are satisfied , we will roughly follow the approach of douady and hubbard for subhyperbolic rational maps @xcite . we need to be able to compute sufficiently good estimates of the orbifold metrics on @xmath47 and @xmath207 . our requirements are formalized in the following proposition . [ prop_or ] let @xmath4 be a strongly subhyperbolic function . then there exist orbifolds @xmath208 and @xmath209 with the following properties : * the set @xmath210 of ramified points of @xmath47 is a finite set that contains @xmath211 . furthermore , there exists a point @xmath212 such that @xmath213 for some @xmath214 . * @xmath215 while @xmath216 . * @xmath47 is a hyperbolic orbifold containing a punctured neighbourhood of @xmath46 . * @xmath199 is a covering map . * the inclusion @xmath217 is holomorphic but not a covering map . furthermore , if @xmath218 , then @xmath219 . we say that the pair @xmath220 of riemann orbifolds is _ dynamically associated to @xmath4 _ , if @xmath207 and @xmath47 satisfy @xmath221-@xmath222 . we start with the construction of @xmath223 . if @xmath224 , then define @xmath225 . otherwise , the fatou set of @xmath4 consists of attracting basins only and , as in the proof of proposition [ prop_dt ] , we can find a bounded neighbourhood @xmath105 of the set @xmath226 such that @xmath227 , @xmath228 is connected and @xmath229 . we choose a set @xmath105 with this property and define @xmath230 note that @xmath223 is connected and that @xmath3 is entirely contained in @xmath223 . now assume that there is a point @xmath231 , such that for every point @xmath232 with @xmath233 there exists @xmath214 with @xmath234 . then we define the ramification value of a point @xmath235 to be @xmath236 if there is no point @xmath237 with such a property , then pick a repelling fixed point @xmath238 of @xmath4 . observe that such a point exists , since every map with a bounded set of singular values has infinitely many fixed points @xcite , and since @xmath4 is subhyperbolic , only finitely many of them can be non - repelling . since @xmath239 , it also belongs to @xmath223 and we define the ramification value of every point @xmath240 to be the value defined by equation ( [ ram ] ) , and assign @xmath237 the ramification value @xmath241 . observe that in both cases , there is a point @xmath242 such that @xmath243 is a multiple of @xmath41 . let @xmath208 . since @xmath244 if and only if @xmath69 belongs to @xmath245 , the set of ramified points of @xmath47 is finite . furthermore , no critical point @xmath246 belongs to a periodic cycle and since we have assumed that the local degree at all points in @xmath3 is globally bounded by some constant @xmath247 , the ramification value @xmath248 is necessarily a finite number for each @xmath235 . hence @xmath208 is a riemann orbifold and statement @xmath221 follows by construction . note that part @xmath249 is an immediate consequence of the definition of @xmath223 . next we will prove that @xmath47 is hyperbolic . observe that it is sufficient to restrict to the case when @xmath250 , since every orbifold that is holomorphically included in a hyperbolic orbifold has to be hyperbolic as well . so we consider the orbifold @xmath47 with @xmath251 . we will give a proof by contradiction , so let us assume that @xmath47 is not hyperbolic . since @xmath252 , it follows that @xmath47 must be parabolic . by assumption @xmath253 , so it follows from ( * ? ? ? * theorem a4 ) that @xmath47 must be isomorphic to @xmath11 with signature either @xmath254 or @xmath255 . this implies that @xmath224 , hence all singular values of @xmath4 belong to @xmath3 . assume first that @xmath47 is isomorphic to @xmath11 with signature @xmath254 . then @xmath4 has only one singular value , say at @xmath256 . it follows from a covering space argument that @xmath257 , where @xmath258 and @xmath259 . but then @xmath136 is an asymptotic value ( and since it is omitted , it can not be a fixed point either ) , yielding a contradiction . so @xmath47 must be isomorphic to @xmath11 with signature @xmath255 . by the previous argument , @xmath4 has exactly two singular values @xmath260 and @xmath261 which are necessarily critical values , and any of their preimages is either a critical point of local degree two or a regular point . signature @xmath255 also implies that @xmath262 , meaning that both critical values are either fixed points of @xmath4 or they form a two - cycle . since @xmath4 is subhyperbolic but not hyperbolic , @xmath260 and @xmath261 are both repelling . in particular , @xmath263 . let @xmath264 be the orbifold which has exactly the regular preimages of @xmath260 and @xmath261 as ramified points , assigning them the ramification value two . clearly , @xmath265 . then @xmath266 is a covering map and since @xmath47 is parabolic , so is @xmath264 , which means that @xmath260 and @xmath261 are the only ramified points in @xmath264 . hence @xmath267 . by conformal conjugacy we can assume that @xmath268 and @xmath269 . then the map @xmath270 is a universal covering map . since @xmath271 is a covering map , it lifts to a conformal @xmath11-isomorphism @xmath272 , @xmath273 , yielding the relation @xmath274 by periodicity and symmetry of the cosine map , @xmath275 and @xmath276 . but this means that @xmath4 or @xmath277 is a chebyshev polynomial , contradicting the fact that @xmath4 is transcendental . hence @xmath47 is hyperbolic . by construction , @xmath47 is the complement of a , possibly empty , compact set , hence @xmath278 follows . define @xmath279 and @xmath280 by equation ( [ ram ] ) , @xmath281 is a positive integer for every @xmath282 and by the identity theorem , the set of points @xmath283 with @xmath284 is discrete . hence @xmath209 is a riemann orbifold . since @xmath285 , the map @xmath286 is a branched covering . furthermore , @xmath287 hence @xmath288 is an orbifold covering map , proving statement @xmath289 . we will show now that the inclusion @xmath290 is holomorphic but not a covering . first note that @xmath291 by construction of @xmath223 . moreover , if @xmath218 , then @xmath292 is a relatively compact subset of @xmath223 ( see equation ( [ eqn_s_f ] ) ) . recall that by @xmath221 , there is a point @xmath212 such that @xmath243 is a multiple of @xmath41 . the fact that @xmath293 implies that @xmath237 has infinitely many preimages @xmath294 under @xmath4 , and for every such preimage point we have @xmath295 . moreover , @xmath296 holds for all but finitely many of the preimages of @xmath237 , since by @xmath221 , @xmath47 has only finitely many ramified points . on the other hand , @xmath297 , which means that @xmath298 , hence the inclusion is not a covering map . let @xmath235 . observe that the definition of @xmath299 ( see equation ( [ ram ] ) ) together with the fact that for any point @xmath300 the local degree of an iterate @xmath301 of @xmath4 is given by @xmath302 implies that @xmath303 divides @xmath304 . since @xmath199 is a covering , @xmath305 . hence @xmath248 divides @xmath281 and this proves that the inclusion @xmath290 is a holomorphic map . let @xmath220 be dynamically associated to @xmath4 . note that @xmath207 is usually not connected . however , if @xmath4 has finitely many tracts over @xmath46 , e.g. if @xmath4 has finite order , then the number of components of @xmath207 is finite . observe also that it follows from proposition [ prop_or]@xmath222 that the set @xmath210 of ramified points of @xmath47 satisfies @xmath306 . finally we would like to mention that we have proved more than the existence of a hyperbolic orbifold @xmath47 satisfying the remaining assumptions of proposition [ prop_or ] . in fact , we have also shown the following . _ let @xmath4 be a transcendental entire function and let @xmath119 and @xmath40 be any two orbifolds such that @xmath120 is a covering map . then @xmath40 ( and hence @xmath119 ) is hyperbolic . _ [ cor_(f ) ] let @xmath4 and @xmath207 be as in proposition [ prop_or ] . then there is a constant @xmath37 and an infinite sequence of points @xmath38 for which @xmath307 is a multiple of @xmath41 , such that @xmath308 holds for all @xmath309 . let @xmath310 be a point such that @xmath243 is a multiple of @xmath41 and let @xmath311 be a jordan curve in @xmath11 such that the bounded component of @xmath312 contains @xmath82 but not @xmath237 . the components of the preimage of the unbounded component of @xmath312 are tracts of @xmath4 and every such tract contains infinitely many preimages of @xmath237 . let @xmath38 denote the preimages of @xmath237 lying in one such ( arbitrary but fixed ) tract . using standard estimates on the hyperbolic metric in a simply - connected domain ( * ? ? ? * corollary a.8 ) , one easily deduces that there is a constant @xmath37 such that @xmath313 holds for infinitely many @xmath309 . ( for details , see e.g. ( * ? ? ? * proof of lemma 5.1 ) or ( * ? ? ? * proof of proposition 3.4 ) . ) however , since all but finitely many @xmath38 are regular points of @xmath4 , it follows that @xmath314 and this is , by @xmath221 , a multiple of @xmath41 . for a pair @xmath220 of orbifolds dynamically associated to @xmath4 , we denote by @xmath315 and @xmath316 the densities of the hyperbolic metrics of @xmath207 and @xmath47 , respectively . we conclude this section with the following simple observation which justifies our restriction to strongly subhyperbolic maps . let @xmath4 be a subhyperbolic map for which there is a pair of dynamically associated orbifolds . then @xmath4 is strongly subhyperbolic . if @xmath317 is an asymptotic value of @xmath4 , then for any compact set @xmath318 containing @xmath317 , there exists a component of @xmath108 which is not compact . hence there is no domain @xmath319 such that @xmath320 is a proper map . hence , if an asymptotic value of @xmath4 belongs to @xmath3 , then there is no domain @xmath321 such that @xmath322 is a ( branched ) covering map . assume now that @xmath4 has a critical value @xmath323 , such that for every @xmath324 there exists a point @xmath325 with @xmath326 and @xmath327 . if there was a pair @xmath220 of dynamically associated orbifolds , then proposition [ prop_or]@xmath289 would imply that @xmath139 is a puncture of @xmath47 , contradicting the fact that @xmath328 . let @xmath4 be a strongly subhyperbolic map and let @xmath220 be dynamically associated to @xmath4 . by proposition [ expansion ] , @xmath329 wherever defined , so in particular for all @xmath330 . if @xmath331 is a point such that @xmath332 , then it follows by proposition [ prop_or]@xmath289,@xmath222 that @xmath333 divides @xmath334 . in this case we define @xmath335 and so if @xmath336 , then @xmath337 is a finite number , while otherwise @xmath338 . our goal is to show that the expansion of @xmath4 is uniform . [ thm_uniform_expansion ] let @xmath4 be strongly subhyperbolic and let @xmath220 be dynamically associated to @xmath4 . then there is a constant @xmath339 such that @xmath340 for all @xmath330 . the remainder of section [ sec uniform ] is devoted to the proof of theorem [ thm_uniform_expansion ] . we want to show the following continuity statement : let @xmath40 be a riemann orbifold , let @xmath237 be a regular and @xmath341 a ramified point of @xmath40 . then the value of the density map @xmath42 of the orbifold metric at @xmath237 depends continuously on @xmath341 , i.e. , if we perturb the point @xmath341 slightly , then the density of the corresponding orbifold metric at the point @xmath237 will also undergo only a small change . this statement is surely not new but we were not able to locate a reference . hence we include a proof for completeness . we will restrict to orbifolds whose underlying surface is a subset of the sphere ; the proof in the general case , where the underlying surface of @xmath40 is an arbitrary riemann surface , can be derived using exactly the same arguments , with the additional step of taking charts . so let @xmath342 and let @xmath123 . denote by @xmath343 the set of ramified points of @xmath40 . let @xmath344 be an arbitrary but fixed integer . for a point @xmath345 we define a new orbifold @xmath346 , where @xmath347 furthermore , we assume that every such orbifold has a universal cover ( hence we exclude the case when @xmath89 is a sphere and the set @xmath343 is either empty or consists of only one point with ramification value @xmath348 ) . note that any two such orbifolds @xmath349 and @xmath350 have the same signature and hence the same uniformized universal cover . let us denote the density of the orbifold metric on @xmath349 by @xmath351 . for a point @xmath352 we define the map @xmath353,\quad q\mapsto\rho_q(p).\end{aligned}\ ] ] [ thm_continuity ] let @xmath352 be arbitrary but fixed . then the map @xmath354 is continuous at every point in @xmath355 . let @xmath356 be an arbitrary but fixed point . we want to show that @xmath354 is continuous at @xmath357 . pick a sufficiently small jordan domain @xmath358 such that @xmath359 . let @xmath317 be a point in @xmath360 . by conformal conjugacy , we can assume that @xmath361 . for two points @xmath362 let us denote by @xmath363 the distance between @xmath364 and @xmath365 measured in the hyperbolic metric of @xmath151 . for a point @xmath366 consider the unique riemann map @xmath367 which maps @xmath368 and @xmath369 , where @xmath370 . let @xmath371 , @xmath372 . then @xmath373 is a @xmath374-quasiconformal self - map of @xmath51 . define @xmath375 it is easy to see that @xmath376 extends continuously to the complement of @xmath151 as the identity map and that the extended map , which we will also denote by @xmath376 , is a @xmath374-quasiconformal map ( see e.g. ( * ? ? ? * lemma 5.2.3 ) ) . observe that @xmath377 as @xmath378 . let @xmath132 be the uniformized universal covering surface of @xmath379 and @xmath349 and let @xmath380 and @xmath381 be universal covering maps , both normalized such that @xmath382 and @xmath383 . considered as a map between orbifolds , @xmath384 is a homeomorphism and hence can be lifted to a homeomorphism on @xmath132 . from now on we will assume that @xmath385 since the other two cases follow by the same strategy , using even simpler calculations . there is a unique lift @xmath386 of @xmath376 such that @xmath387 . let @xmath388 denote the covering group of @xmath50 over @xmath349 and assume that there exist two distinct lifts @xmath389 , @xmath390 of @xmath376 that fix @xmath136 . there exists a mapping @xmath391 such that @xmath392 holds for all @xmath393 . it follows from our assumption that @xmath394 , hence @xmath395 , where @xmath396 denotes the stabilizer of @xmath136 in @xmath388 . but @xmath397 and @xmath136 is a non - ramified point of @xmath349 , which means that @xmath398 is trivial . hence @xmath399 and @xmath400 . we have the following commutative diagram : & ^_q & + ^_q^ * & & _ _ q + _ q^ * & _ _ q & _ q since @xmath376 is a @xmath374-quasiconformal map and @xmath401 and @xmath402 are holomorphic , the map @xmath386 is also @xmath374-quasiconformal . moreover , @xmath389 is conformal when restricted to the set @xmath403 , so in particular in a sufficiently small neighbourhood of any point in the set @xmath404 . furthermore , the lifts @xmath389 converge to @xmath405 ( locally uniformly ) as @xmath378 and , due to the chosen normalization , @xmath406 . moreover , when restricted to @xmath407 , the maps converge in the @xmath408-norm , hence @xmath409 when @xmath378 . by the above diagram we can write @xmath410 for every @xmath393 . recall that if @xmath411 and @xmath412 , then the value of the density function @xmath351 at @xmath139 is given by @xmath413 and this does not depend on the choice of the preimage of @xmath139 . similarly , if @xmath414 , then @xmath415 . hence , @xmath416 observe first that @xmath417 , since @xmath418 . let us fix a point @xmath419 . then @xmath420 let @xmath421 be the unique point such that @xmath422 . we obtain @xmath423 hence @xmath424 since @xmath425 in the @xmath408-norm in a neighbourhood of @xmath426 when @xmath378 , the expression @xmath427 tends to @xmath428 and hence @xmath429 . we are now able to prove theorem [ thm_main2 ] , which is the key - statement for the proof of theorem [ thm_uniform_expansion ] . [ thm_orb_main ] let @xmath37 and let @xmath38 , @xmath430 , be an infinite sequence of points satisfying @xmath313 . let @xmath431 , where @xmath432 then the density @xmath42 of the hyperbolic metric on @xmath40 satisfies @xmath43 first observe that , by affine conjugacy , we can assume that @xmath136 is one of the ramified points of @xmath40 . let @xmath433 be an arbitrary but fixed point in @xmath40 . depending on @xmath69 , we choose @xmath434 , where @xmath435 satisfies @xmath436 and is minimal with this property , i.e. , if @xmath437 then @xmath438 . it follows immediately that @xmath439 as next we set @xmath440 , where @xmath441 is minimal with the property @xmath442 . we then obtain @xmath443 for any three pairwise distinct points @xmath444 denote by @xmath445 the orbifold defined by @xmath446 note that every such orbifold is hyperbolic , since its euler characteristic equals @xmath447 . we denote by @xmath448 the density of the hyperbolic metric on @xmath449 . observe first that @xmath40 is holomorphically embedded in @xmath450 , and it follows from theorem [ pick ] that @xmath451 holds for all @xmath452 . let @xmath453 and @xmath454 . then the map @xmath455 is obviously a conformal isomorphism and hence a local isometry . altogether , we obtain @xmath456 let @xmath457 . then equations ( [ eqn_b ] ) and ( [ eqn_c ] ) yield @xmath458 i.e. , @xmath459 and @xmath460 belong to compact annuli disjoint from @xmath461 ( see figure [ figure_annuli ] ) . by theorem [ thm_continuity ] the map @xmath462 is a composition of two continuous maps and hence itself continuous . furthermore , it attains its minimum ( and maximum ) on the compact set @xmath463 . hence , there exist constants @xmath464 depending only on @xmath168 such that @xmath465 by setting @xmath466 in equation ( [ eqn_abc ] ) , we finally get @xmath467 and the assertion of the theorem follows . and @xmath468 belong to the compact annuli @xmath469 and @xmath470 . ] theorem [ thm_orb_main ] and corollary [ cor_(f ) ] immediately imply the following . [ estimate_rhotilde ] let @xmath4 be strongly subhyperbolic and let @xmath220 be dynamically associated to @xmath4 . then @xmath471 where @xmath472 denotes the density of the hyperbolic metric on @xmath207 . using theorems [ thm_orb_main ] and [ thm_continuity ] , we can finally deduce that @xmath199 is a uniform expansion with respect to the hyperbolic metric of @xmath47 . let @xmath315 and @xmath316 denote the densities of the hyperbolic metrics on @xmath207 and @xmath47 , respectively . since @xmath288 is a covering map , our claim is equivalent to the statement that there is a constant @xmath339 such that @xmath473 if @xmath474 , recall that by proposition [ prop_or]@xmath222 , @xmath47 is modelled on a hyperbolic domain @xmath223 with @xmath475 , implying that @xmath47 and @xmath207 have no common boundary points in @xmath11 . the same is true if the underlying surface is @xmath11 , which means that @xmath46 is the only common boundary point of @xmath47 and @xmath207 . hence it only remains to check that for some @xmath476 , @xmath477 let @xmath478 be the complement of a closed euclidean disk centred at @xmath136 such that @xmath479 for all @xmath480 , and denote by @xmath481 the density of the hyperbolic metric on @xmath132 . then there is a right half - plane @xmath482 such that the map @xmath483 is a covering . hence the asymptotic behaviour of @xmath481 is given by @xmath484 by theorem [ pick ] @xmath485 , and so @xmath486 it now follows from corollary [ estimate_rhotilde ] that @xmath487 and hence @xmath488 we already mentioned in the introduction that if we replace the ramified points by punctures , i.e. , if we consider the hyperbolic domain @xmath45 instead of the orbifold @xmath40 , the same bound for the asymptotic behaviour of the density map near @xmath46 can be obtained using standard estimates of the hyperbolic metric in the twice - punctured plane ( * ? ? ? * lemma 2.1 ) . as will become clear in the proof of proposition [ prop_cos_est ] , the orbifold for which the set of ramified points is given by @xmath44 shows that our estimate is best possible . we say that @xmath36 is a _ cosine map _ , if it can be written as @xmath489 for some @xmath490 . every such function @xmath36 has exactly two critical values , namely @xmath491 and @xmath492 . furthermore , if @xmath65 is a preimage of a critical value , then @xmath69 is a critical point satisfying @xmath493 , which implies that @xmath260 and @xmath261 are totally ramified . it is easy to check that @xmath36 has no asymptotic values , hence the critical values @xmath260 and @xmath261 are the only singular values . in particular , every subhyperbolic cosine map is automatically strongly subhyperbolic . in @xcite , schleicher studied landing properties of those cosine maps for which the critical values are strictly preperiodic . note that the julia set of every such function equals @xmath11 . he proved that for such a map , every point in @xmath11 is either on a dynamic ray or the landing point of a dynamic ray . this result will also follow from theorem [ thm_maintheorem ] . moreover , our proof in the case of strongly subhyperbolic cosine maps is considerably more concise and elementary than the proof of the general statement and the proof ( of the weaker statement ) given in @xcite . the reason is that for strongly subhyperbolic cosine maps , we can compute _ explicitly _ the required estimates of the metrics of certain dynamically associated orbifolds . let us start with a simple observation . let @xmath494 be strongly subhyperbolic but not hyperbolic . then there exists a point @xmath495 . since @xmath496 is not hyperbolic , it follows that at least one critical value of @xmath496 , say @xmath260 , belongs to @xmath497 . now assume that the claim is wrong , i.e. , @xmath498 . since @xmath211 is forward invariant , this can only occur if @xmath499 or @xmath260 and @xmath261 form a cycle . however , since @xmath260 is totally ramified , it would then be a superattracting periodic point of @xmath496 , contradicting the assumption that @xmath500 . for simplicity , let us assume that @xmath501 ; the case when @xmath502 can be treated in a very similar way ( and is even easier ) . let @xmath503 and @xmath504 , where @xmath505 it is straightforward to check that @xmath506 is a pair of orbifolds dynamically associated to @xmath496 . ( in fact , this is how we constructed dynamically associated orbifolds in the proof of proposition [ prop_or ] . ) in particular , @xmath507 for all @xmath65 . let us fix a point @xmath495 . then @xmath237 has only regular preimages @xmath294 , for which necessarily @xmath508 . since @xmath496 is @xmath509-periodic , the orbifold @xmath510 is holomorphically embedded in the orbifold @xmath511 defined by @xmath512 in particular , if @xmath513 and @xmath514 denote the densities of the hyperbolic metrics on @xmath510 and@xmath515 , respectively , then @xmath516 holds for all @xmath65 . for @xmath514 we can give the following explicit lower bound . [ prop_cos_est ] the density function @xmath517 satisfies @xmath518 for all @xmath65 , with @xmath519 . for all points in the punctured halfplane @xmath520 , the density @xmath521 of the hyperbolic metric of @xmath522 can be bounded from below by @xmath523 where @xmath524 and @xmath525 @xcite . let @xmath526 , where @xmath527 and @xmath528 for all @xmath529 . we easily see that the map @xmath530 is a covering map and hence a local isometry . a simple calculation yields @xmath531 where @xmath532 is the density of the hyperbolic metric of @xmath533 and @xmath534 denotes the principle branch of the squareroot . since the map @xmath535 is a covering map , it follows that @xmath536 for every @xmath537 . let us simplify the above expression . we note , by expanding with @xmath538 , that @xmath539 and one can easily see that the obtained expression is bounded from above by @xmath540 . let us now consider @xmath541 . since @xmath542 , it is enough to restrict to the case when @xmath543 . here we get @xmath544 and hence @xmath545 together , these estimates yield the proof . let @xmath546 and denote by @xmath547 the @xmath11-orbifold with signature @xmath548 , with @xmath317 , @xmath549 and @xmath52 being the ramified points . then there exists a ( unique ) mbius map @xmath550 mapping @xmath136 , @xmath135 and @xmath137 to @xmath317 , @xmath549 and @xmath52 , respectively . moreover , the map @xmath551 provides a covering map from @xmath515 to @xmath547 . this observation enables us to estimate the hyperbolic metric of an arbitrary @xmath11-orbifold with signature @xmath548 using simple calculations , and hence provides an alternative way of proving theorem [ thm_orb_main ] , which though it is less elegant uses only elementary observations . recall that our goal is to construct a continuous and surjective map @xmath552 , where @xmath6 is any map of disjoint type that belongs to the family @xmath553 such that @xmath554 holds for all @xmath19 . recall that by ( * ? ? ? * theorem 5.2 ) , any two such maps @xmath6 and @xmath555 are conjugate on their julia sets , hence it is enough to prove the statement for one such map . we start with the construction of such a map @xmath6 . let us fix a pair of orbifolds @xmath220 dynamically associated to @xmath4 with underlying surfaces @xmath292 and @xmath223 , respectively . note that , by proposition [ prop_or]@xmath278 , @xmath223 can be written as @xmath556 , where @xmath132 is a , possibly empty , compact set . observe that for every @xmath557 , @xmath558 . let @xmath559 be sufficiently large , such that @xmath560 . since @xmath4 is entire , it maps bounded sets to bounded sets , hence there exists @xmath561 such that @xmath562 let us fix a constant @xmath561 with this property and define @xmath563 . it then follows that if @xmath564 and @xmath69 is a point with @xmath565 , then @xmath566 and hence @xmath567 . this means , @xmath568 and , in particular , it follows from proposition [ prop_dt ] that @xmath6 is of disjoint type . define @xmath569 note that @xmath570 holds for all @xmath571 , such as @xmath572 , since @xmath6 is of disjoint type . furthermore , @xmath28 is the set of those points that are never mapped into @xmath573 , hence @xmath28 equals the limit of the domains @xmath574 . and the isomorphism @xmath575 . ] we want to construct a sequence of conformal isomorphisms @xmath576 for @xmath577 and with @xmath578 , such that @xmath579 we will proceed inductively . since @xmath578 , the map @xmath580 is given by the formula @xmath581 . for a point @xmath582 let @xmath583 be the straight line segment connecting @xmath69 and @xmath584 ( we can actually choose @xmath585 to be any rectifiable curve which connects @xmath69 and @xmath584 within the domain @xmath586 ) . to define @xmath575 at a point @xmath587 , we consider the line segment @xmath588 . by definition of @xmath589 , @xmath590 since @xmath591 , there is a preimage component @xmath592 of @xmath593 , such that one endpoint of @xmath592 is @xmath584 . we define @xmath594 to be the other endpoint of @xmath592 ( see figure [ fig_semi ] ) . continuing inductively , we define the curve @xmath595 to be the pullback of @xmath596 under @xmath4 with one endpoint at @xmath597 , and we define @xmath598 to be the other endpoint @xmath599 . we want to give some properties of the maps @xmath600 . since @xmath4 and @xmath6 are holomorphic and in particular continuous , each map @xmath600 is continuous as well . by induction , it also follows that each map @xmath600 is injective and surjective . hence each map @xmath600 is a conformal isomorphism , mapping a component of @xmath601 onto a component of @xmath602 . [ thm_main_repeat ] the maps @xmath603 converge uniformly with respect to the hyperbolic orbifold metric @xmath604 on @xmath47 to a continuous surjective function @xmath605 so that @xmath606 . moreover , @xmath607 and @xmath608 is a homeomorphism . with respect to the hyperbolic metric on @xmath47 , we denote by @xmath609 the distance between two points @xmath610 , and by @xmath611 the length of a rectifiable curve @xmath612 . let @xmath613 . since @xmath572 , both @xmath600 and @xmath614 are defined in a neighbourhood of @xmath69 and it follows from our construction that @xmath615 since for every point @xmath616 , @xmath617 we obtain an upper bound for @xmath618 by computing its length with respect to the hyperbolic metric in @xmath619 , which is given by @xmath620 . hence @xmath621 recall that by lemma [ expansion ] , there is a constant @xmath339 , such that @xmath622 holds for all @xmath330 . since @xmath623 is obtained as a pullback of @xmath624 under the map @xmath625 , it follows from equation ( [ phi_j ] ) that @xmath626 this means that the maps @xmath603 from a cauchy sequence , and since the orbifold metric is complete , there is a continuous limit function @xmath627 note that @xmath20 necessarily satisfies @xmath628 as well as @xmath629 for all @xmath324 and all @xmath19 . we want to derive some properties of the limit function @xmath20 . by equation ( [ eqn_escape ] ) , @xmath630 so together with equation ( [ eqn_fe ] ) this implies that @xmath631 . in particular , it follows that @xmath632 , since @xmath633 and @xmath634 @xcite . now let @xmath635 . then @xmath636 for all sufficiently large @xmath637 , so we can consider the sequence @xmath638 . let @xmath69 be an accumulation point of the points @xmath639 . note that by relation ( [ eqn_phi_infty ] ) , @xmath640 . let @xmath641 be a subsequence of @xmath639 that converges to @xmath69 . then , @xmath642 showing that @xmath643 is surjective . before we show that @xmath608 is injective , let us recall that @xmath644 is a countable collection of simply - connected domains , so - called tracts , on which @xmath6 acts as a covering map . now let @xmath645 be a curve connecting @xmath646 with @xmath46 . then every component of @xmath647 is a simply - connected unbounded subdomain of @xmath648 , called a _ fundamental domain _ , and the restriction of @xmath6 to any such domain @xmath496 is a conformal map . now let @xmath649 be two points such that @xmath650 . by definition , @xmath651 and it follows from the inductive definition of the maps @xmath600 that for every sufficiently large @xmath637 , there exists a fundamental domain @xmath652 such that @xmath653 . on the other hand , it follows from equation ( [ eqn_fe ] ) that @xmath654 holds for all @xmath655 . furthermore , equation ( [ eqn_escape ] ) implies that @xmath656 by standard expansion estimates ( see e.g. ( * ? ? ? * lemma 2.7 ) ) , the distance between @xmath657 and @xmath658 must be unbounded , unless the points @xmath69 and @xmath461 are equal , implying that @xmath20 is injective . observe that by equation ( [ eqn_phi_infty ] ) , @xmath20 can be extended ( sequentially ) continuously to a map @xmath659 with @xmath660 . the set @xmath661 is compact since it is the continuous image of a compact set . furthermore , @xmath662 is necessarily closed . so @xmath663 and as @xmath664 is closed , it follows that @xmath665 , hence @xmath20 is surjective . since the restriction of the map @xmath20 in theorem [ thm_main_repeat ] to the escaping set of the disjoint type map is a homeomorphism , we obtain the following result as an immediate consequence of theorem [ thm_main_repeat ] and theorem [ thm_esc_set_dis_type ] . the escaping set of a strongly subhyperbolic map is not connected . dierk schleicher kindly pointed out that the escaping set of the cosine map @xmath666 mentioned in the introduction is obviously disconnected : the imaginary axis consists of points with bounded orbits and it disconnects the escaping set ( for details on this special function , see appendix a ) . to state our next corollary , we need to introduce the notion of a dynamic ray . [ defn_ray ] a _ ray tail _ of a transcendental entire map @xmath4 is an injective curve @xmath667 ( where @xmath668 ) such that for each @xmath324 , @xmath669 and such that , as @xmath670 , @xmath671 uniformly in @xmath78 . a _ dynamic ray _ of @xmath4 is then a maximal injective curve @xmath672 such that @xmath673 is a ray tail for every @xmath668 . in terms of dynamic rays , our main result implies the following topological description of the julia set of certain strongly subhyperbolic maps . [ cor_pcb ] let @xmath29 be a strongly subhyperbolic map , where every @xmath30 has finite order and a bounded set of singular values . then @xmath3 is a pinched cantor bouquet , consisting of dynamic rays of @xmath4 and their endpoints . in particular , all dynamic rays of @xmath4 land and every point in @xmath3 is either on a dynamic ray or the landing point of a dynamic ray of @xmath4 . recall that by a pinched cantor bouquet we mean a quotient of a cantor bouquet by a closed equivalence relation on its endpoints . note that corollary [ cor_pcb ] implies corollary [ cor1 ] from the introduction . let @xmath6 and @xmath674 be maps as in theorem [ thm_main_repeat ] . by ( * theorem 4.7 , theorem 5.10 ) , @xmath28 is an `` absorbing brush '' ( this implies that every connected component @xmath132 of @xmath28 is a closed arc to infinity , and all points of @xmath132 except possibly the finite endpoint escape ) . in fact , @xmath28 is homeomorphic to a straight brush in the sense of @xcite ( see remark in @xcite ) , and hence @xmath28 is a cantor bouquet . by ( * ? ? ? * theorem 4.7 ) , @xmath28 consists of dynamic rays of @xmath6 and their endpoints . since @xmath20 is surjective , @xmath3 is a pinched cantor bouquet , consisting of dynamic rays of @xmath4 and their endpoints . the statement that the absorbing brush in ( * ? ? ? * theorem 4.7 ) is homeomorphic to a straight brush in the sense of @xcite can be deduced , for instance , using a topological characterization like the one given in ( * ? ? ? * theorem 3.11 ) . however , we will not state the details here since it would require a lot vocabulary from point - set topology which would be of no further use in this article . if @xmath4 is a function as in corollary [ cor_pcb ] , then one can encode the `` pinching '' of the cantor bouquet ( which equals @xmath3 ) combinatorially using _ itineraries_. we will elaborate this explicitly in the case of the map @xmath13 in the following section . however , such a concept can be developed in the more general setting of corollary [ cor_pcb ] ; this is contained in the author s thesis . for more information on itineraries of exponential and cosine maps , see e.g. @xcite . this section is dedicated to the description of the topological dynamics of the function @xmath675 we will define a `` simple '' model consisting of a topological space @xmath676 and a map @xmath677 such that if @xmath6 is any map of disjoint type in the family @xmath678 then * @xmath28 is homeomorphic to @xmath676 , and * @xmath679 is conjugate to @xmath680 . we will transfer the ideas from @xcite , where such a model was constructed for exponential maps whose singular value belongs to some attracting basin . the adoption of @xcite to the maps we are interested in is particularly simple since in left and right half - planes , sufficiently far away from the imaginary axis , any map @xmath681 with @xmath682 is essentially the same ( i.e. , up to a constant factor ) as @xmath683 and @xmath684 , respectively . for this reason , we will skip the details and refer , for further consideration , to @xcite as well as the extensive work on dynamics of cosine maps by rottenfuer and schleicher @xcite . once we have constructed such a model for a disjoint type map @xmath685 , theorem [ thm_main_repeat ] tells us that there is a semiconjugacy between @xmath6 and @xmath4 on their julia sets , and hence also between the model map @xmath686 and @xmath4 . the combinatorial dynamics of @xmath4 on @xmath3 was already established in @xcite and we will summarize here the required results . let us consider the family @xmath687 with @xmath682 ( hence @xmath688 ) . the critical values of @xmath681 are @xmath689 . every map @xmath690 is a homeomorphism with @xmath691 and @xmath692 . furthermore , @xmath693 $ ] . both critical values of @xmath4 are mapped by @xmath4 to the repelling fixed point @xmath136 . since @xmath4 ( as well as every other @xmath681 ) has no asymptotic values , the postsingular set of @xmath4 equals @xmath694 . hence @xmath4 is postsingularly finite and strongly subhyperbolic . furthermore , @xmath695 . for @xmath682 chosen sufficiently small , the origin is an attracting fixed point and the subinterval @xmath696 $ ] of the imaginary axis is mapped into itself and thus belongs to the immediate basin of attraction of @xmath136 . hence by choosing @xmath22 sufficiently small , we obtain a map @xmath681 of disjoint type ( see proposition [ prop_disjoint_type ] ) . from now on , we will fix @xmath697 such that the corresponding map @xmath698 is of disjoint type . note that for every @xmath699 , the horizontal line @xmath700 is mapped by @xmath6 ( or any other @xmath681 with @xmath701 ) to @xmath702 $ ] , hence every point @xmath19 is contained in one the horizontal half - strips @xmath703 note that the restriction of @xmath6 ( or any other @xmath681 with @xmath701 ) to any of the half - strips is a conformal isomorphism onto its image which is the left or right half - plane . let @xmath704 be the space of infinite sequences of elements in @xmath705 , where @xmath706 and @xmath707 are two disjoint copies of @xmath708 . by the previous argument , we can assign to a point @xmath19 a unique sequence @xmath709 defined by @xmath710 . we will call such a sequence the _ external address _ of @xmath69 . for every @xmath711 we define @xmath712 . furthermore , since @xmath28 consists of ( asymptotically horizontal ) dynamic rays and their endpoints ( * ? ? ? * theorem 4.1 ) , our model @xmath676 should be a subset of the space @xmath713 note that the relation @xmath714 defines an order on @xmath715 . thus @xmath716 is equipped with the product topology of the topology on @xmath715 ( induced by the order relation ) and the standard topology on @xmath717 . let @xmath718 be a point in @xmath716 . we should think of the first entry @xmath719 in @xmath720 as the imaginary part of the point ( or its height corresponding to our horizontal strips ) , together with the information whether it is lying left or right from the imaginary axis . the second entry @xmath78 should be thought of as the absolute value of the real part of the point . hence it is helpful to think of a point @xmath721 in its `` complexified '' version @xmath722 . let us denote by @xmath723 the projection onto the second coordinate . we can now define our model map to be @xmath724 where @xmath725 denotes the one - sided shift map and @xmath726 denotes the standard model map for exponential growth . recall that the maps we consider behave like the exponential in each of the halfplanes . the essential characteristic of our model map now is that as for exponential maps , the size of the image @xmath727 of a point @xmath718 is roughly the exponential of its real part . more precisely , @xmath728 whenever @xmath729 . hence we define the model sets @xmath676 and @xmath730 to be @xmath731 by ( * ? ? ? * observation 3.1 ) , @xmath676 is homeomorphic to a straight brush . in particular , for every external address @xmath720 there exists a unique @xmath732 $ ] such that @xmath733 . we denote by @xmath734 the set of _ endpoints _ of @xmath676 . by iterating forward under the model map @xmath686 and backwards under @xmath6 , we obtain a sequence of maps that converges to a homeomorphism @xmath735 such that @xmath736 for all @xmath737 . the key argument for such a limit to exist is again uniform hyperbolic contraction of the map @xmath6 and the fact that the mapping behaviour of the model map reflects that of @xmath6 . ( for a precise statement see ( * ? ? ? * section 3 ) or ( * ? ? ? * proposition 3.3 ) . ) a proof of the above statement is essentially the same as in the case of exponential maps in ( * ? ? ? * theorem 9.1 ) , which is why we skip the details here . a proof can also be derived by essentially the same estimates as given in the proof of theorem [ thm_main_repeat ] . by theorem [ thm_main_repeat ] there is a surjective map @xmath738 such that @xmath739 holds for all @xmath19 . moreover , @xmath20 restricts to a homeomorphism between @xmath7 and @xmath21 . as already mentioned , every point @xmath740 escapes within the strips @xmath741 with @xmath742 , since the forward orbit of any point in the boundary of the strips has a bounded orbit . recall from the proof of theorem [ thm_main_repeat ] that by choosing the inverse branches of the maps @xmath743 appropriately , the conjugacy @xmath20 relates the escaping points of @xmath6 and @xmath4 with respect to the combinatorics in terms of their external addresses . from corollary [ cor_pcb ] , we obtain that @xmath686 projects to a function @xmath744 on @xmath745 , where @xmath746 is an equivalence relation on the set @xmath747 of endpoints of @xmath676 , such that @xmath748 is conjugate to @xmath749 . the equivalence relation @xmath746 tells us which dynamic rays are being `` pinched '' . we will now describe @xmath746 explicitly using the results from @xcite . , showing its structure of a pinched cantor bouquet . this picture was kindly provided by arnaud chritat . ] for every @xmath699 we set @xmath750 one can easily see that the restrictions @xmath751)$ ] and @xmath752)$ ] are conformal isomorphisms . we will call a sequence @xmath753 an _ itinerary_. if @xmath311 is a dynamic ray such that for every @xmath754 there exists a domain @xmath755 with @xmath756 then we assign to @xmath311 the ( well - defined ) itinerary @xmath757 . since a dynamic ray of @xmath4 is either contained in some half - strip @xmath758 or is completely contained in the boundary of such a domain , it follows that an itinerary can not be assigned to a ray @xmath311 if and only if there is @xmath759 such that @xmath760 equals @xmath761 or @xmath762 , or equivalently , if @xmath763 or @xmath764 for all @xmath571 , where @xmath765 is the external address of @xmath311 . this means that to every external address @xmath720 in @xmath766 we can assign a unique itinerary @xmath767 . let us first comment on those external addresses that belong to @xmath768 the mapping behaviour of the map @xmath4 is fairly simple and allows us to describe completely all tuples and quadruples of external addresses in @xmath769 for which the respective dynamic rays land together . for instance , for all addresses @xmath770 that belong to either the left of right quadruple @xmath771 where @xmath772 and @xmath759 , we define @xmath773 . we will not list the remaining combinations since there are not so many of them and each one is easy to determine using elementary computations . the remaining task is to determine those external addresses in @xmath774 such that the corresponding dynamic rays land together . let @xmath311 be a dynamic ray with external address @xmath775 and let @xmath139 be its landing point . then either @xmath6 is the only dynamic ray that lands at @xmath139 or there is exactly one more such dynamic ray @xcite ; the latter case occurs if and only if @xmath139 is eventually mapped into @xmath776 $ ] ( and remains there without ever being mapped to @xmath136 ) . let @xmath777 be the external addresses of two dynamic rays landing at the same point @xmath139 . it follows that @xmath778 and @xmath779 must be of the form @xmath780 or with @xmath137 instead of @xmath136 . on the other hand , it follows from ( * ? ? ? * lemma 5 ) and elementary computations that two dynamic rays with itineraries as in equation ( [ eqn_itin ] ) do indeed land together : such dynamic rays have a forward image that lands in the interval @xmath781 $ ] and its landing point is never mapped to @xmath136 . so let @xmath782 . it follows that @xmath783 if and only if @xmath778 and @xmath779 are of the form given by equation ( [ eqn_itin ] ) ( or with @xmath137 instead of @xmath136 ) . one can certainly relate the model @xmath784 directly to @xmath3 . the reason to incorporate a disjoint type map is simply to show what to do when the considered strongly subhyperbolic map @xmath4 can be embedded in a family where the topological dynamics of disjoint type maps is well understood . we want to give an explicit example of a map @xmath785 that is subhyperbolic , has no asymptotic values but such that the local degree at points in @xmath786 is unbounded . let @xmath237 be a complex polynomial of degree @xmath787 and let @xmath103 be a repelling fixed point of @xmath237 with multiplier @xmath788 . by poincar s theorem @xcite , there exists an entire map @xmath785 , which is called a _ poincar function of @xmath237 at @xmath103 _ , such that the functional equation @xmath789 is satisfied for all @xmath65 . now let @xmath790 and let @xmath791 denote a poincar function of @xmath237 at the point @xmath792 . the unique finite critical point of @xmath237 is @xmath136 . since @xmath793 and @xmath794 , the cycle @xmath795 is superattracting and , in particular , @xmath796 . note that @xmath237 has no other attracting or parabolic cycles in @xmath11 , since every such cycle attracts at least one critical point of @xmath237 ( * ? ? ? * theorem 8.6 ) . observe also that @xmath237 has no exceptional values ( points with a finite backward orbit ) in @xmath11 ; it is now not hard to check that @xmath797 ( see e.g. ( * ? ? ? * theorem 2.10 ) ) . the multiplier of the repelling fixed point @xmath103 is given by @xmath798 . by @xcite , the order of @xmath791 is given by the formula @xmath799 and it follows then from the denjoy - carleman - ahlfors theorem ( * ? ? ? * xi , @xmath874 , p.313 ) that @xmath791 has at most one finite asymptotic value . let us assume that @xmath800 and let @xmath139 be the unique asymptotic value of @xmath791 . by ( * ? ? ? * theorem 1 ) , @xmath139 is an attracting periodic point of @xmath237 , hence either @xmath801 or @xmath802 . let @xmath803 be an asymptotic path for @xmath139 , i.e. , @xmath804 and @xmath805 . then @xmath806 hence @xmath807 is an asymptotic path of @xmath791 leading to the asymptotic value @xmath808 . but this contradicts the fact that @xmath809 , and hence @xmath810 . since @xmath136 is a critical value of @xmath791 , there exists a point @xmath461 such that @xmath811 let @xmath812 . using the functional equation ( [ eqn_poincare ] ) , it is not difficult to see that for every @xmath324 , @xmath813 hence for every @xmath324 , we have @xmath814 . this also implies that at least one of the critical values must have a regular preimage ; otherwise , both critical values would have ramification index strictly larger than @xmath815 and this is not possible by the fact that the sum of the ramification indices of an entire map never exceeds one ( * ? ? ? * chapter x , @xmath873 , 235 ) . we can assume w.l.o.g . that @xmath136 has a regular preimage , say @xmath317 , i.e. , @xmath816 and @xmath817 . let @xmath549 be a preimage of @xmath137 , chosen sufficiently large such that @xmath818 . now consider the map @xmath819 note that @xmath820 , since @xmath785 and @xmath791 differ only by postcomposition with a conformal map . it follows also immediately that @xmath821 , @xmath822 and @xmath823 , hence @xmath785 is postsingularly finite and in particular subhyperbolic . moreover , since @xmath824 , the critical value @xmath317 is a repelling fixed point of @xmath785 and hence belongs to @xmath786 , implying that @xmath785 is not hyperbolic . finally note that the points @xmath325 are mapped to @xmath317 under @xmath785 satisfying @xmath825 , so @xmath785 is not strongly subhyperbolic . 99 , ` the geometry of julia sets ' , _ trans . _ 338 ( 1993 ) , no . 2 , 897918 . , ` the poincar metric of plane domains ' , _ j. london math . ( 2 ) _ 18 , no . 3 ( 1978 ) , 475483 . , ` iteration of meromorphic functions ' , _ bull . _ 29 , no . 2 ( 1993 ) , 151188 , arxiv : math.ds/9310226 . , ` etude dynamique des plynomes complexes ' , _ publications mathmatique dorsay , 84 , universit de paris - sud _ ( 1984 ) . , ` singularities of schrder maps and unhyperbolicity of rational functions ' , _ comput . methods funct . theory _ 8 , no . 1 ( 2008 ) , 285302 , arxiv : math.ds/0704.3309 . , ` on the iteration of entire functions ' , _ dynamical systems and ergodic theory , proc . banach center publ . warsaw _ 23 ( 1989 ) , 339345 . , ` dynamical properties of some classes of entire functions ' , _ ann . fourier , grenoble _ 42 , no . 4 ( 1992 ) , 9891020 . , ` the real fatou conjecture ' , _ annals of mathematics studies , princeton university press _ ( 1998 ) . , ` on the fixpoints , multipliers and value distribution of certain classes of meromorphic functions ' , _ ann . _ 23 ( 1998 ) , 135150 . , ` complex dynamics and renormalization ' , _ annals of mathematics studies , princeton university press 135 , 3rd edition _ ( 1994 ) . , ` a landing theorem for dynamic rays of geometrically finite entire functions ' , _ preprint _ ( 2008 ) , arxiv : math.ds/0707.4135 . , ` dynamics in one complex variable ' , _ annals of mathematics studies , princeton university press , princeton , nj , vol . 160 , edition 3 _ ( 2006 ) . , ` eindeutige analytische funktionen ' , _ springer - verlag / berlin , gttingen , heidelberg , 2nd edition _ ( 1953 ) . , ` sur une classe nouvelle de transcendantes uniformes ' , _ j. math . pures appliques iv ser . _ 6 ( 1890 ) , 316365 . , ` topological dynamics of exponential maps on their escaping sets ' , _ ergodic theory dynamical systems _ 26 , no . 6 ( 2006 ) , 19391975 , arxiv : math.ds/0309107 . , ` rigidity of escaping dynamics for transcendental functions ' , to appear in _ acta math . _ , arxiv : math.ds/0605058 . , ` the escaping set of the exponential ' , _ preprint _ ( 2009 ) , arxiv : math.ds/0812.1768 . , ` escaping points of entire functions of small growth ' , _ math . z. _ 261 , no . 3 ( 2009 ) , 557570 , arxiv : math.ds/0801.3605 . , ` escaping points of the cosine family ' , _ transcendental dynamics and complex analysis , london mathematical society lecture notes series , cambridge university press _ 348 , ( p. j. rippon and g. m. stallard , eds . ) ( 2008 ) . , ` dynamic rays of bounded type entire functions ' , _ preprint _ ( 2005 ) , arxiv : math.ds/0704.3213 . , ` periodic points and dynamic rays of exponential maps ' , _ ann . _ 28 ( 2003 ) , 327354 . , ` the dynamical fine structure of iterated cosine maps and a dimension paradox ' , _ duke math . journal _ 136 , no . 2 ( 2007 ) , 343356 , arxiv : math.ds/0406255 . , ` hausdorff dimension , its properties , and its surprises ' , _ american mathematical monthly _ 114 ( 2007 ) , 509528 , arxiv : math.ds/0505099 . , ` geometry and topology of three - manifolds ' , _ princeton lecture notes _ ( 1979 ) . , ` on the combinatorics and dynamics of iterated rational maps ' , _ preprint _ ( 1984 ) . , ` fonctions analytiques ' , _ presses universitaires de france _ ( 1954 ) .
let @xmath0 be a transcendental entire map that is _ subhyperbolic _ , i.e. , the intersection of the fatou set @xmath1 and the postsingular set @xmath2 is compact and the intersection of the julia set @xmath3 and @xmath2 is finite . assume that no asymptotic value of @xmath4 belongs to @xmath3 and that the local degree of @xmath4 at all points in @xmath3 is bounded by some finite constant . we prove that there is a hyperbolic map @xmath5 with connected fatou set such that @xmath4 and @xmath6 are semiconjugate on their julia sets . furthermore , we show that this semiconjugacy is a conjugacy when restricted to the escaping set @xmath7 of @xmath6 . in the case where @xmath4 can be written as a finite composition of maps of finite order , our theorem , together with recent results on julia sets of hyperbolic maps , implies that @xmath3 is a _ pinched cantor bouquet _ , consisting of dynamic rays and their endpoints . our result also seems to give the first complete description of topological dynamics of an entire transcendental map whose julia set is the whole complex plane .
the global dynamics of networks of @xmath2 coupled oscillators with different types of coupling has been studied in @xcite . in their formalism it is shown that the symmetry group of the network can be considered a subgroup of @xmath3 , as long as the oscillators taken individually have no internal symmetries . these ideas have been investigated in the above cited paper , for general @xmath4 and @xmath5 cases . however , it is a routine considering @xmath6 so the case @xmath0 is sometimes not explicitly taken into account . another reason for carrying out our study is the fact that among the dihedral groups , @xmath0 is the only abelian group , which makes its action and therefore its analysis slightly different . in this paper we are concerned with two properties of networks with @xmath0 symmetry : hopf bifurcation and low coupling case leading to heteroclinic cycles . firstly we use the methodology developed by ashwin and stork @xcite to construct a network of differential systems with @xmath0 symmetry . our case is a particular situation of a network of @xmath2 coupled oscillators with symmetry to be a subgroup of @xmath7 that analyzes different types of coupling between the oscillators , as shown in @xcite . while this approach has been successfully used for the subgroups @xmath5 and @xmath8 our approach is interesting not only because of the particularities already mentioned of @xmath9 but also because it offers the possibility of analyzing the weak coupling limit , where dynamics of the network is governed only by the phases of the oscillators . + a large variety of subgroups of @xmath3 can be easily generated by the method described by ashwin and stork @xcite based on graph theory . the automorphism group of the colored graph can be defined to be its symmetry group . therefore , it is possible , via cayley graphs , to design oscillatory networks with the prescribed symmetry of a subgroup of @xmath10 while the hopf bifurcation is in fact a simple adaptation of the theory developed by golubitsky , stewart and schaeffer in @xcite,@xcite and @xcite , the low coupling case is much more interesting since it allows the existence of heteroclinic cycles in systems with @xmath0 symmetry . in this case it is possible to reduce the asymptotic dynamics to a flow on an four - dimensional torus @xmath11 by assuming a weak coupling we average the whole network and introduce an extra @xmath12 symmetry . we determine the two - dimensional invariant subspaces on this torus and show that a heteroclinic cycle consisting of four one - dimensional routes among the four zero - dimensional equilibria can appear . we then apply the general stability theory for heteroclinic cycles developed by krupa and melbourne in @xcite , to analyze the stability of the heteroclinic cycles . our aim is to construct an oscillatory system with the @xmath0 symmetry . to achieve this goal we need first to represent the group by a cayley diagram . a cayley diagram is a graph ( that is , a set of nodes and the arrows between them ) to represent a group . vertex or nodes of the graph are the group elements and the arrows show how the generators act on the elements of the group . at any vertex there are arrows pointing towards it , others pointing away . proceeding as in @xcite or @xcite , let @xmath13 be the generating set of @xmath14 this implies that * @xmath15 generates @xmath9 * @xmath15 is finite and * @xmath16 since @xmath0 is finite we can forget assumptions @xmath17 and @xmath18 then the cayley graph of @xmath0 , is a directed colored graph whose arrows indicate the action the group elements have on each other and the vertices are the group elements . as shown in @xcite the generating set @xmath15 is a set of `` colors '' of the directed edges and the group elements @xmath19 and @xmath20 are connected through an edge from @xmath19 to @xmath20 of color @xmath21 if and only if @xmath22 group . solid arrows represent left - multiplication with @xmath23 dot - and - dashed arrows left multiplication with @xmath24 the two generators of this group.,width=293,height=291 ] ( 0,0 ) ( -90,235 ) ( 83,235 ) ( -90,55 ) ( 83,55 ) the cayley graphs for @xmath0 is shown in figure . following the approach by stork et al . @xcite , we may identify the vertices as cells with a certain dynamics and the edges as couplings between the cells . in this way we can construct an ode system which has the symmetry of @xmath0 . the action of the group @xmath0 on the cells can be written as @xmath25 where the two generators and their commuting product @xmath26 act on @xmath27 as @xmath28 if we assign coupling between cells related by the permutations in , we can build the following pairwise system in with the @xmath0 symmetry . @xmath29 where @xmath30 and @xmath31 . as shown by ashwin and stork @xcite we can think of @xmath32 as being generic functions that assure that the isotropy of this vector field under the action of @xmath33 is generically @xmath0 . the group @xmath0 has order @xmath34 , the generators being @xmath35 since all irreducible representations of the @xmath0 group are one dimensional , we restrict our study to the actions of the generators on @xmath36 the orthogonal representation of @xmath0 @xmath37 on the complex vector space @xmath38 is irreducible if and only if the only @xmath39invariant subspaces of @xmath38 are the trivial ones . in fact it can be shown that @xmath40 the group element @xmath41 acts on the point @xmath42 by @xmath43 the orbit of a point @xmath44 under the action of @xmath0 is @xmath45 therefore the orbits are * the origin , @xmath46 * points on the @xmath47-axis , @xmath48 with @xmath49 * points on the @xmath50-axis , @xmath51 with @xmath52 * points off the axes @xmath53 with @xmath54 the isotropy subgroup of a point @xmath44 under the action of @xmath0 is @xmath55 @xmath0 has four isotropy subgroups : * @xmath0 corresponding to the origin * @xmath56 corresponding to @xmath57 with @xmath58 * @xmath59 corresponding to @xmath60 with @xmath61 * @xmath62 corresponding to @xmath63 with @xmath64 in this section we basically recall the theory of golubitsky and stewart @xcite and @xcite on one parameter hopf bifurcations with symmetry . the group @xmath0 only has four @xmath66dimensional real irreducible representations and so they are absolutely irreducible . one of them is the trivial representation and the other three are non - isomorphic each with a nontrivial @xmath67 action . therefore , the only possible way to have a hopf bifurcation is for the linearization at the equilibrium to be from a @xmath68simple subspace of the form @xmath69 where @xmath70 is an absolutely irreducible representation which has either a trivial or a @xmath67 action . generically , as a parameter is varied , the critical eigenspace with purely imaginary eigenvalues is two - dimensional of the form @xmath69 described above . the periodic solution bifurcating is unique and has the same symmetry has the @xmath67 action on v , or has no symmetry . in fact , the @xmath0 symmetric coupled system does not support any periodic solution . one can easily show that the isotypic decomposition at a trivial steady - state is made up of the trivial representation @xmath71 and the complement decomposes into the remaining three non - isomorphic irreducible representations . therefore , the @xmath69 subspace can not exist in the tangent space at the trivial equilibrium . each cell in the network needs to be two - dimensional to hope to have a hopf bifurcation with @xmath0 symmetry in such a @xmath72cell system . for the @xmath39equivariant hopf bifurcation in @xmath73 , there is a unique branch of periodic solutions consisting of rotating waves with @xmath67 spatial symmetry . first , unicity of a branch of periodic solutions is based on the fact that @xmath0 is abelian , and we recall the abelian hopf theorem in @xcite . therefore , the periodic solution predicted by the equivariant hopf theorem @xcite with the symmetry of the unique isotropy subgroup with a two - dimensional fixed point subspace is the only one . we begin with recalling results on the normal form for the @xmath39equivariant bifurcation from @xcite on the @xmath74 group , which is isomorphic to @xmath14 this means that we can adapt the theory of @xmath74 to our case , withe merely a re - interpretation of the branches . we say that the bifurcation problem @xmath75 commutes with @xmath0 if @xmath76 the proof is almost identical to the one for @xmath74 in @xcite but we prefer to give it adapted to the action of the group generators of @xmath14 we write @xmath75 as @xmath81 commutativity with equation implies @xmath82 now @xmath83 transforms @xmath84 into @xmath85 , i.e. @xmath86 and if the action of @xmath87 is @xmath88 then we get that @xmath89 is odd in @xmath47 while @xmath90 is even in @xmath91 when @xmath83 acts as identity and @xmath87 acts as @xmath92 then @xmath89 is even in @xmath50 and @xmath90 is odd in @xmath93 now the rest of the proof is exactly the same as in @xcite . + + in the following we will follow again the route and results in @xcite to obtain the jacobian matrix of equation @xmath94 what is more , we can reformulate or adapt to our case lemma @xmath95 ( page @xmath96 in @xcite ) and add the proof again corresponding to our case , which in the cited reference has been left as an exercise for the reader . * if @xmath99 is a trivial or a pure mode solution , then @xmath100 in is diagonal and its eigenvalues which are real have the signs listed as follows : * * trivial solution : @xmath101 * * @xmath102mode solution : @xmath103 * * @xmath104mode solution : @xmath105 * if @xmath99 is a mixed mode solution , then * * @xmath106 at @xmath107 * * @xmath108 at @xmath109 accordingly , if @xmath114 equation reduces to @xmath115 in the first case . moreover , since @xmath116 for the @xmath102mode solution , then @xmath117 and therefore the jacobian matrix becomes @xmath118 similarly , since @xmath119 for the @xmath104mode solution , then @xmath120 and therefore the jacobian matrix becomes @xmath121 under the conditions in lemma , the stability of the heteroclinic cycle is guaranteed by the fact that it is restricted to the dynamics of system within the fixed point spaces . we use the criteria of krupa and melbourne @xcite to study the stability of the heteroclinic cycle . we now recall these stability criteria , which are based on four hypotheses . + the first hypothesis guarantees that the heteroclinic cycle is robust . * ( @xmath126 ) there is an isotropy subgroup @xmath127 with the fixed - point subspace @xmath128 such that @xmath129 and @xmath130 is a sink in @xmath131 + corresponding to each isotropy subgroup @xmath127 in ( @xmath126 ) is the isotypic decomposition @xmath132 of @xmath133 into isotypic components . let @xmath134 and @xmath135 denote the normalizer of @xmath127 in @xmath136 + * ( @xmath137 ) the eigenspaces corresponding to @xmath138 and @xmath139 lie in the same @xmath140isotypic component ; + * ( @xmath141 ) @xmath142 + * ( @xmath143 ) all transverse eigenvalues of @xmath144 with positive real part lie in the same @xmath140isotypic component . [ krupa , melbourne ] let @xmath147 be a finite group acting on @xmath133 and @xmath148 be a @xmath68equivariant vector field . suppose that @xmath149 is a heteroclinic cycle for @xmath150 satisfying hypotheses @xmath151 then generically the stability of @xmath149 is described by precisely one of the following possibilities . * asymptotically stable ( @xmath152 and @xmath153 for each @xmath154 ) , * unstable but essentially asymptotically stable ( @xmath152 and @xmath155 for each @xmath154 and @xmath156 for some @xmath154 ) , * almost completely unstable ( @xmath157 or @xmath158 for some @xmath154 ) , * completely unstable if @xmath159 the stability is expressed by @xmath162 where @xmath163 in equation , @xmath164 is the expanding eigenvector at the @xmath165th point of the cycle , @xmath166 is the contracting eigenvector and @xmath167 is the tangential eigenvector of the linearization . for the heteroclinic cycle we have @xmath168 so from equation we obtain @xmath169(a-2\epsilon ) ^2(b-2\epsilon)}{\left[b+2(\epsilon+2q)\right](a+2\epsilon ) ^2(b+2\epsilon)}}. \end{array}\ ] ] then the proof follows by applying theorem @xmath170 in @xcite . 10 , _ oscillator networks with the symmetry of the unit quaternion group _ , in dynamics , bifurcation and symmetry , series c : mathematical and physical sciences , vol . 437 , p. chossat ed . , pp . 35 - 48 . , _ permissible symmetries of coupled cell networks _ , math . , 116 , ( 1994 ) , pp . 2736 . , _ the dynamics of n identical oscillators with symmetric coupling _ , j. nonlin . , 2 , ( 1992 ) , pp . _ singularities and groups in bifurcation theory i _ , applied mathematical sciences 51 , springer - verlag , ( 1985 ) . , _ the abelian hopf @xmath172 theorem _ , siam j. appl . dynam . sys . 9 , ( 2010 ) , pp . 283291 . , _ hopf bifurcation with dihedral group symmetry : coupled nonlinear oscillators . in : multiparameter bifurcation theory _ , m. golubitsky , j. guckenheimer , eds . , contemporary mathematics 56 , ams ( 1986 ) , pp . 131173 . , _ nonasymptotically stable attractors in @xmath173 mode interactions . normal forms and homoclinic chaos _ , ( w. langford and w. nagata eds . ) fields institute communications 4 , amer . soc . , providence , ri , ( 1995 ) , pp .
in this paper we analyze a generic dynamical system with @xmath0 constructed via a cayley graph . we study the hopf bifurcation and find conditions for obtaining a unique branch of periodic solutions . our main result comes from analyzing the system under weak coupling , where we identify the conditions for heteroclinic cycle between four equilibria in the two - dimensional fixed point subspace of some of the isotropy subgroups of @xmath1 we also analyze the stability of the heteroclinic cycle . equivariant dynamical system , cayley graph , hopf bifurcation , heteroclinic cycle . 37c80 , 37g40 , 34c15 , 34d06 , 34c15
this talk briefly summarizes a growing body of work@xcite done with tom cohen on unstable baryon resonances in the @xmath0 qcd expansion . more in - depth synopses appear in ref . . the @xmath0 expansion about the large @xmath1 qcd limit has been extensively used to study baryons treated as stable states . at @xmath2 the lowest - lying baryons fill a completely symmetric spin - flavor su(6 ) representation , the * 56*. although the lightest @xmath3 baryons have @xmath1 valence quarks , the su(6 ) representation remains symmetric , as one shows@xcite by imposing order - by - order ( in @xmath0 ) unitarity in @xmath4 scattering ( _ consistency conditions _ ) . low - spin baryons in this multiplet ( analogues to @xmath5 , @xmath6 ) are split in mass at @xmath7,@xcite meaning that such states are stable for large @xmath1 . such analyses use a hamiltonian @xmath8 with asymptotic baryon eigenstates , which is expanded in su(6)-breaking operators accompanied by powers of @xmath0 . baryon resonances , on the other hand , generically lie above the ground - state multiplet by an @xmath9 mass gap ; since mesons like @xmath10 have @xmath9 masses , such states are generically unstable against strong decay . while treating excited multiplets like the arbitrary-@xmath1 analogue of the su(6 ) * 70 * using a hamiltonian approach has produced many interesting phenomenological results , such a treatment for resonances is not entirely reasonable since it entails grafting _ ad hoc _ sources of @xmath11 pairs ( the production / decay mechanism ) onto @xmath8 , violating unitarity . moreover , resonances in such an @xmath8 are not manifested as excited states of stable baryons . in fact , a resonance @xmath12 is properly represented by a complex - valued pole in a scattering amplitude at position @xmath13 . such an approach can be developed using tools originally appearing in the context of chiral soliton models , for such models naturally accommodate baryon resonances as excitations resulting from scattering of mesons off ground - state baryons . they are consistent with the large @xmath1 limit because the solitons are heavy , semiclassical objects compared to the mesons . in contrast to quark models , chiral soliton models tend to fall short in providing detailed spectroscopy and decay parameters for baryon resonances , particularly at higher energies . nevertheless , one may use the scattering or soliton _ picture _ ( no specific model ) to study baryon resonances as well as the full scattering amplitudes in which they appear , and relate it for @xmath1 large to the hamiltonian or quark _ picture_. the relevant physical observables in real data are scattering amplitudes , in which resonances appear as complex poles . the role of large @xmath1 is to treat the ground - state baryon , off which mesons scatter , as a heavy field manifesting the consistency conditions . these conditions turn out to impose@xcite t - channel constraints : amplitudes with @xmath14 first appear at order ( both relative and absolute ) @xmath15 , leading to relations among the amplitudes holding at various orders in @xmath0 . the t - channel expressions may be reexpressed in the s channel using group - theoretical crossing relations , in which the @xmath16 rule becomes@xcite the statement that the underlying dynamical degrees of freedom are _ reduced amplitudes _ @xmath17 labeled by @xmath18 , the eigenvalue of @xmath19 . the @xmath4 and @xmath20 partial waves of orbital angular momentum @xmath21 are @xmath22 respectively , which are special cases of an expression@xcite that holds for ground - state baryons of arbitrary spin (= isospin ) and mesons of arbitrary spin and isospin . expressions like ( [ mpeqn1 ] ) are useful because full observable partial waves outnumber reduced amplitudes @xmath17 , giving rise to a number of linear relations among the measured amplitudes that hold at leading [ @xmath9 ] order . a pole in one partial wave implies a pole in one of the @xmath17 , which in turn contributes to several others partial waves . a single pole therefore generates a multiplet of baryon resonances degenerate in mass and width.@xcite moreover , the existence of an s - channel pole in an intermediate state can not depend upon the states used to form it ( except through conservation of quantum numbers ) , and so the resonant poles are labeled uniquely by @xmath18 . for example , eqs . ( [ mpeqn1 ] ) applied to @xmath24 , @xmath25 states ( @xmath26 ) gives @xmath27 and @xmath28 , while in the @xmath29 , @xmath30 ( @xmath31 ) channel , @xmath32 ( parity being determined by @xmath21 ) . a pole in @xmath33 ( @xmath34 ) implies a degenerate pole in @xmath35 , because the latter contains a @xmath36 amplitude . clearly , degenerate towers of resonances develop with @xmath37 , _ etc . _ on the other hand , one type of @xmath26 pole ( @xmath36 ) couples predominantly to @xmath10 and the other ( @xmath38 ) to @xmath39 . in comparison , the quark - picture large @xmath1 analogue of the excited multiplet @xmath40 has 2 @xmath26 and 2 @xmath31 states , just as for @xmath41 . one can show that the numerous nonstrange states in this multiplet have only 3 distinct mass eigenvalues out to @xmath9,@xcite which we label @xmath42 . comparing the quantum numbers and multiplicities of such states with the results of eqs . ( [ mpeqn1 ] ) , one immediately finds@xcite that exactly the required resonant poles are obtained if each @xmath18 amplitude , @xmath43 , contains precisely one pole , located at the value @xmath44 . this quark - picture multiplet of excited baryons is found to be _ compatible _ with , _ i.e. _ , consist of a complete set of , multiplets classified by @xmath18 . one can prove@xcite compatibility for all nonstrange baryon multiplets in the su(6)@xmath45o(3 ) shell picture , as well as for the whole @xmath40 multiplet@xcite if @xmath46 poles are added . it is important to note that compatibility does _ not _ imply su(6 ) is an exact symmetry at large @xmath1 for resonances as it is for ground states ; rather , it says that su(6)@xmath45o(3 ) multiplets are complete but _ reducible _ at large @xmath1 . in the above example , @xmath42 are split at @xmath9 and therefore label distinct multiplets . indeed , large @xmath1 by itself does not mandate the existence of _ any _ baryon resonances , but it does say that if even one exists , it must be a member of a well - defined multiplet . it is possible to obtain testable predictions for the decay channels , even using just the leading order in @xmath0 . for example , we noted that the @xmath47 @xmath26 resonance couples only to @xmath39(@xmath10 ) . indeed , the @xmath48 resonance decays to @xmath20 with a 3055@xmath49 branching ratio ( br ) despite lying barely above that threshold , while the @xmath50 decays to @xmath20 with only a 310@xmath49 br despite having much more comparable @xmath4 and @xmath20 phase spaces . this pattern clearly suggests@xcite that the @xmath10-phobic @xmath48 should be identified with @xmath51 and the @xmath39-phobic @xmath50 with @xmath52 , the first fully field theory - based explanation for these observations . similarly , the @xmath53 lies only 5 mev above the @xmath54 threshold but has a 1025@xmath49 br to this channel , and thus appears to lie in a @xmath55 su(3 ) octet with the @xmath48.@xcite large @xmath1 also provides@xcite su(3 ) selection rules : octet resonances preferentially produce @xmath10 and @xmath39 in their decays , while singlet @xmath56 s produce @xmath57 s . configuration mixing between states with 1 , 2 , _ etc . _ excited quarks , which requires an _ ad hoc _ treatment in the quark picture , naturally occurs in the scattering picture . if one finds two resonances of comparable mass and identical quantum numbers , and one has a natural broad @xmath9 width , can the other have a narrow @xmath7 width ? unless a conservation law forbids the mixing , both are expected to have @xmath58 . indeed , one can exhibit@xcite @xmath9 operators that allow such mixing . the mandate that resonances occur in multiplets applies not only to conventional but exotic baryons as well . for example , a state with the quantum numbers of the controversial @xmath59 ( @xmath60 , @xmath61 ) would necessarily have partners@xcite with @xmath62 , @xmath63 and @xmath64 , @xmath65 , degenerate in mass and width up to @xmath7 corrections . indeed , such a @xmath66 in an su(3 ) @xmath67 would have degenerate * 27 * and * 35 * partners as well,@xcite an enumeration matching that found in a quark picture calculation.@xcite first - order @xmath0 corrections in amplitude relations like eqs . ( [ mpeqn1 ] ) are handled@xcite ( at least in the nonstrange case ) simply by including @xmath68 amplitudes in the analysis and identifying linear relations surviving among the enlarged set of physical amplitudes . when confronted with data , relations obtained in this way are typically a factor @xmath2 better satisfied than those that only hold to @xmath9 . processes other than meson - baryon scattering can also be treated using the scattering approach , given only the @xmath69 and @xmath70 quantum numbers of the field coupling to the baryon along with the corresponding @xmath1 power suppression of each coupling . for example , each multipole amplitude in pion photoproduction can be expressed in this way,@xcite once the photon is divided into its @xmath71 and @xmath72 pieces . amplitude relations thus derived , including @xmath0 corrections , can be compared to data . the comparisons are favorable except in the resonant region , where the effects of finite @xmath0 splittings between resonances degenerate in the @xmath73 limit obscure agreement . in fact , the helicity amplitudes on corresponding resonances derived from such seemingly disparate amplitudes turn out to agree remarkably well with @xmath0 expectations.@xcite the most interesting unsolved problems using the scattering method involve a proper treatment of @xmath0 effects . first , the @xmath74 and its incorporation of @xmath0-suppressed amplitudes must be generalized to strange resonances . second , many spurious states arise when @xmath3 ; explaining how they decouple when @xmath1 is tuned to 3 will open a broad new front for phenomenological results . finally , a marriage of this approach with chiral symmetry ( straightforward because of the connection to soliton models ) will provide a very powerful additional set of physical constraints . the author s research was supported by the n.s.f . through grant phy-0140362 . lebed , hep - ph/0406236 , in _ continuous advances in qcd 2004 _ , ed . by t. gherghetta , world scientific , singapore ( 2004 ) ; hep - ph/0501021 , in _ large @xmath1 qcd 2004 _ , ed . by j.l . goity _ et al . _ , world scientific , singapore ( 2005 ) ; t.d . cohen , _ nucl . phys . _ a * 755 * , 40 ( 2005 ) .
i give a brief overview of the model - independent analysis of unstable baryon resonances in the @xmath0 qcd expansion . this approach produces numerous surprising semi - quantitative phenomenological predictions that are supported by available observations .
we discuss specific aspects of the core of a vortex line in layered high t@xmath1 superconductors . the physics of these vortices is governed by two distinct length scales , the london penetration depth in the planes , @xmath2 , and the coherence length in the planes , @xmath3 . the penetration depth is the electromagnetic length scale of a vortex . the physics on this length scale is well described by a combination of macroscopic electromagnetism , london s theory for supercurrents along the layers , and interlayer josephson coupling . this description breaks down in the core of the vortex , _ i.e. _ at distances of order @xmath4 from the center of the vortex . thus , physical properties of the core carry information on the microscopic physics of high t@xmath1 superconductivity . the small coherence length of high t@xmath1 superconductors makes the vortex core a good potential sensor for microscopic mechanisms of superconductivity . our discussion of the vortex core in high t@xmath1 superconductors is based on the fermi - liquid model of superconductivity . the physical properties of the vortex core predicted by this model are spectacular , unique , and could serve as fingerprints of the traditional pairing theory of superconductivity . the vortex core of traditional high-@xmath5 superconductors is well described by the bardeen - stephen model@xcite which represents the core by a region of normal electrons . the bardeen - stephen model is justified as long as the mean free path , @xmath6 , is much shorter than the core size , so that the motion of an electron gets randomized before it leaves the core . this condition is not fulfilled in high t@xmath1 superconductors which are generally clean superconductors with @xmath7 . the core of a vortex in a clean superconductor was first studied in the classic papers of caroli , matricon , and de gennes.@xcite these authors calculated the spectrum of quasiparticle states in the core , and showed that electrons and holes form bound states at energies below the bulk energy gap . further early studies of the excitation spectrum in the core can be found in refs . . more recent theoretical work was stimulated by the direct observation of core states in @xmath8 by scanning tunnelling spectroscopy ( sts).@xcite the recent report of sts in ybco @xcite provides new information on the excitation spectrum of vortices in the high @xmath9 cuprates . consequently , theoretical efforts focused on the tunnelling density of states of bound states in isolated vortices and vortex lattices.@xcite these calculations show that the bound states in the core have a different nature compared with the usual quantum mechanical bound states in a potential well . the core states are coherent superpositions of particle states and hole states and are formed by repeated andreev scattering from the pair potential ( order parameter ) in the core . andreev scattering is a process of `` retroreflection '' of excitations : spatial variations of the amplitude or the phase of the order parameter induce branch conversion of electron - like excitations into hole - like excitations , and vice versa . bound states occur at energies at which the phases of multiply reflected electron - like and hole - like states interfere constructively . the charge current carried by an incoming electron and an outgoing andreev reflected hole is identical because the reversal of the velocity in an andreev reflection process is compensated by the reversal of the charge due to electron - hole conversion . consequently , andreev bound states can transport a charge current , unlike bound states in a potential well . charge conservation requires that the current carried by the bound states inside the core is transported outside the core by bulk supercurrents . this leads to an interplay between supercurrents flowing past the core and the bound states in the core . hence , the physics of the ` normal core ' in clean superconductors is basically the physics of the bound states in contact and intimate exchange with the superconducting environment outside the core . consider a stack of `` pancake '' vortices forming an isolated vortex line whose axis is oriented perpendicular to the layers . we investigate the current distribution in the core of a pancake vortex , and show how this distribution changes if the vortex is exposed to a bulk supercurrent , or the circulation is changed from @xmath10 to @xmath11 . we calculate the _ spectral current density _ , which carries the information on the contribution of the states in a given energy interval to the total current density . a supercurrent in homogeneous superconductors is distributed over all continuum states . these states exhibit doppler shifts of their energies , @xmath12 , in the presence of a phase gradient in the order parameter ( or superfluid momentum , @xmath13 ) . the total current is obtained by adding the contributions of states with positive shifts from quasiparticles co - moving with the flow and the contributions with negative shifts from quasiparticles that are counter - moving relative to the flow field . we find that the currents in the core have a very different spectral distribution from bulk supercurrents . the continuum states ( scattering states ) show smeared out doppler shifts , and contribute very little to the total current . the dominant contributions to the circulating currents around the vortex center , as well as the currents through the core , come from andreev bound states . hence , the physics of vortex cores in clean superconductors ( @xmath14 ) is very different from the physics of the vortex core in a dirty superconductor ( @xmath15 ) , which is well described by a continuum of normal electronic states . the calculations presented in this paper concentrate on stationary properties of the vortex core of clean layered superconductors . we expect more spectacular effects in the dynamic properties . the bound states react sensitively to the environment outside of the core . this leads to a coupling of the collective degrees of freedom in the london range of the vortex and the bound states in the core , which will produce a rich spectrum of largely unexplored dynamical phenomena . below we present analytical and numerical calculations for the states in the vortex core . we use two versions of a quasiclassical formulation of the bcs theory of superconductivity : a ) _ andreev s theory_@xcite which represents the quasiclassical limit of bogolyubov s equations,@xcite and b ) the _ quasiclassical theory _ of eilenberger @xcite , larkin , and ovchinnikov@xcite which represents the quasiclassical limit of gorkov s green s function theory . andreev s theory and the quasiclassical theory are essentially equivalent for clean superconductors , and in this limit the choice of approach is largely a matter of taste . however , the quasiclassical theory has a wider range of application . it is the generalization of landau s fermi liquid theory to the superconducting state , and is capable of describing a broader range of superconducting materials and phenomena , such as dirty superconductors or superconductors with short inelastic lifetimes ( strong - coupling superconductors).@xcite section ii contains analytical results for the bound states and the spectral current density for a pancake vortex with a superimposed bulk supercurrent . these results are obtained from andreev s hamiltonian@xcite by the methods described in ref . . in section iii we discuss the numerical results , which are obtained using the quasiclassical theory of fermi - liquid superconductivity . we solve the quasiclassical transport equations to obtain self - consistently the pair amplitude ( order parameter ) for pancake vortices . given the pair amplitude we calculate the excitation spectrum in the core of the vortex , and deduce from it the spectral current density . the numerical calculations are done for layered superconductors with s - wave pairing . the analytical and numerical results confirm and complement each other , and they establish the important role of the bound states for the currents in the core region of a vortex . in this section , we investigate the spectrum of current carrying states of a two - dimensional pancake vortex in equilibrium at temperature @xmath16 , in the presence of an externally imposed supercurrent . we ignore the spin degree of freedom of a quasiparticle excitation;@xcite in this case it is sufficient to work in the two - dimensional space of particle - hole degrees of freedom . operators in this space are @xmath17 matrices , and we use the notation @xmath18 , @xmath19 , @xmath20 , for the three pauli matrices in particle - hole space ( nambu space ) , and @xmath21 for the unit matrix . the hamiltonian for the quasiparticle excitations , @xcite the bogolyubov hamiltonian , then reads : @xmath22 where @xmath23 and @xmath24 are the momentum and position operators appropriate to a particle moving in two dimensions , @xmath25 is the fermi momentum and @xmath26 is the electromagnetic vector potential . the order parameter , @xmath27 is an off - diagonal matrix and is generally represented by a linear combination of @xmath18 and @xmath19 . in the absence of an externally imposed supercurrent , we write the order parameter of the vortex as @xmath28 where @xmath29 is the magnitude of the order - parameter of a bulk superconductor at temperature @xmath16 , @xmath30 is the normalized profile of the vortex , which is a monotonically increasing function of @xmath31 obeying @xmath32 , and @xmath33 is the angular coordinate of @xmath34 with @xmath35 , @xmath36 the main assumption we make in this section is that the order - parameter in the presence of a superflow @xmath37 has the form @xmath38 we assume throughout this section that @xmath39 is small compared to the bulk critical current , @xmath40 . the principal physical quantities with which we shall concern ourselves are the _ spectral current density _ , and the total equilibrium current density which is related to @xmath41 by @xmath42 we shall also make reference to the local density of states , @xmath43 . the quantities @xmath44 and @xmath45 may be expressed in terms of the one - particle greens function , @xmath46 or , equivalently , the `` spectral function '' @xmath47 . using the spectral function , we find that in dirac notation , @xmath48 @xmath49 where the subscript @xmath50 denotes the upper left element of the @xmath17 matrices , thereby selecting out the _ particle sector _ of the spectral function and the factor @xmath51 takes into account both spin projections of the quasiparticles . most calculations of the properties of superconductors with inhomogeneous order parameters are simpler in the quasiclassical limit , where one takes advantage of the separation in the scales of the wavelength of quasiparticles near the fermi energy and the characteristic scale for spatial variations of the pair potential , i.e. @xmath52 . the quasiclassical limit of the bogolyubov hamiltonian ( [ bogoliubov ] ) is the andreev hamiltonian in which the kinetic energy in ( [ bogoliubov ] ) is replaced by an operator that is _ linear _ in the gradient.@xcite let us define the normal - state density of states at the fermi level , @xmath53 , and introduce the directions , @xmath54 and @xmath55 , which are , respectively , parallel and perpendicular to trajectories of a quasiparticle wavepacket in the quasiclassical description , _ i.e. _ @xmath56 . the coordinates along these directions are defined by @xmath57 . in addition we work in the limit @xmath58 , in which case the vector potential is approximately constant in the vicinity of the vortex core and can be neglected . the order parameter in ( [ delta ] ) can be written as @xmath59 .\ ] ] by performing a gauge transformation that removes the factor of @xmath60 in eq . ( [ op ] ) we obtain the spectral current density and the local density of states in terms of the andreev hamiltonian for an isolated vortex , @xmath61 @xmath62)|\zeta \rangle _ { 1,1 } \,,\ ] ] @xmath63)|\zeta \rangle _ { 1,1}\,,\ ] ] where @xmath64 is an eigenvector of the `` one - dimensional '' trajectory coordinate operator , @xmath65 : @xmath66 . the operators @xmath67 and @xmath68 appearing in @xmath69 are canonically conjugate : @xmath70=i\hbar$ ] . the quasiclassical interpretation given to ( [ andreev ] ) is as follows : quantum - mechanical evolution in particle - hole space takes place along classical trajectories parallel to @xmath71 having a fixed value of @xmath72 . thus , @xmath73 is identified as a c - number impact parameter.@xcite let us write the current density at temperature @xmath16 as @xmath74 where @xmath75 is a high energy cutoff that serves to make manipulations of @xmath76 well defined ; large positive energies are automatically cut off by the fermi function , @xmath77 . where no ambiguity arises , we shall take @xmath78 . defining @xmath79 , and using ( [ andreev ] ) we have @xmath80 next , we split up the energy integrals into the following terms @xmath81 @xmath82 @xmath83 \,,\ ] ] @xmath84 the three contributions to the current have different interpretations . 1 . since @xmath85 is large , @xmath86 may be replaced by its high energy , normal - state limit , @xmath87 and @xmath88 this term coincides with the @xmath89 current of a uniform superconductor . the term @xmath90 contributes to `` backflow '' , since it always yields a current with a component in the @xmath91 direction.@xcite the term @xmath90 contains the current carried by the bound states and also a correction to the @xmath89 current due to the thermal breaking of pairs . to appreciate these points we look at this term in two limits , assuming @xmath92 . ( i ) for @xmath89 we have @xmath93 the small size of @xmath94 ensures that the energy integral only selects states in the gap , thus @xmath95 only obtains contributions from the bound states . ( ii ) for @xmath96 and assuming @xmath97 to be that of a uniform system , @xmath98 we write @xmath99@xmath100@xmath101 and obtain @xmath102= -ev_f^2n_f{\bf p}_sy(\beta \delta_{0 } ) \,,\ ] ] where @xmath103 is the yosida function which gives a quantitative measure of the thermal breaking of cooper pairs . the term @xmath104 is independent of @xmath105 and is simply the current density of a vortex in the absence of an externally imposed supercurrent . consider the spectral properties at the center of the vortex . for @xmath106 @xmath107 with @xmath108 accounting for the @xmath109 phase change across the vortex core . this special case of the andreev hamiltonian is identical in form to the continuum hamiltonian used to describe _ trans_-polyacetylene containing a single topological soliton.@xcite it is known that this hamiltonian always has a non - degenerate bound state at zero energy.@xcite whether or not it has other bound states depends on the form of the profile , @xmath110 . for the single quantum vortex and trajectories through the center there are no other bound states . the eigenfunction for the zero - energy bound state , @xmath111 , is found by solving @xmath112 \psi_0(\zeta ) = 0.$ ] the normalized solution is @xmath113 @xmath114 where @xmath115 is a profile dependent quantity with the dimensions of length , @xmath116 . analytical estimates of the bound states at distances far from the vortex are given in appendix [ far ] . for energies @xmath117 only the bound state of @xmath118 will contribute to the spectral current density ( and the local density of states ) , @xmath119_{1,1}\ ] ] @xmath120 \hat{{\bf p}}_s\ , , \quad there is a simple relation between @xmath121 and @xmath122 when @xmath123 in ( [ jo ] ) , the delta function in the integrand of @xmath121 effectively replaces @xmath71 by @xmath124 . taking this factor outside the integral leaves an integral identical to that of the local density of states . consequently , @xmath125 note that the contribution of negative energy ( bound ) states to the total current density lies in the @xmath126 direction , i.e. _ opposite _ to the externally imposed supercurrent . at zero temperature the total current density originates from the bound states having energies @xmath127 , @xmath128 the current density of an isolated vortex with @xmath129 vanishes at the center of the vortex , _ i.e. _ @xmath130 . we can combine the result in ( [ j_bound ] ) with @xmath131 given in ( [ j1 ] ) to obtain the total current density at @xmath89 : @xmath132 thus , for sufficiently small * p*@xmath133 the bound - state contribution dominates ( [ center ] ) and @xmath134 will point in the @xmath135 direction . for equilibrium conditions the divergence of the current density vanishes . from ( [ j1 ] ) , @xmath131 has a vanishing divergence , and for an undisturbed vortex we have @xmath136 . thus , @xmath137 . at @xmath89 , @xmath138 in eq . ( [ andreevspectrum ] ) of appendix [ simplification ] we show that @xmath139 \,,\ ] ] yielding latexmath:[\[\label{sso}s({\bf r})=4\pi ev_f\delta_{0}n_f\int_0^{2\pi } \frac { d\varphi _ k}{2\pi } \int_0^{-{\bf q\cdot}\hat{\bf k}}d\epsilon \text{\/}\frac { f(r)}r\mbox{tr}\left [ ( \zeta \hat\tau_{2}-\eta \hat\tau_{1})\langle \zeta @xmath92 , only the bound state , @xmath141 , contributes to the @xmath142 integral . thus , @xmath143 where @xmath144 is the bound - state energy for an impact parameter @xmath73 . at small distances from the center of the vortex ( @xmath145 ) @xmath146 , with @xmath147 , and the lowest energy bound state is then @xmath148 @xmath149 substituting @xmath150 and @xmath151 into ( [ s ] ) and setting @xmath152 yields @xmath153 which is non - zero , indicating that the ansatz ( [ delta ] ) is not physically correct for any value of @xmath154 . this failure to satisfy the conservation law is due to the lack of self - consistency of the order parameter in eq . ( [ delta ] ) in the presence of the flow field . in the absence of pinning , the vortex will move in response to a flow field , even one of arbitrarily small strength . the results in section ii implicitly assume a _ pinned _ vortex . thus , there will be distortion of the vortex away from its cylindrically symmetric equilibrium form ( [ delta ] ) . in the following section we show that the self - consistently determined vortex order parameter , which includes the deformation by the flow field , restores the conservation law . a versatile and efficient method for calculating local spectral properties of superconductors is the quasiclassical theory of superconductivity.@xcite this theory is well adapted for a numerical approach to microscopic problems in superconductivity , such as the calculation of the structure and the excitation spectrum of vortex cores for superconductors with isotropic , anisotropic or unconventional order parameters . the quasiclassical theory is the only theoretical formulation which can handle equally well clean and dirty superconductors , as well as more complicated geometries than that of an isolated vortex with cylindrical symmetry or a perfect vortex lattice . it can be interpreted as the generalization of landau s theory of normal fermi liquids to the superconducting state . the quasiclassical theory shares with landau s theory the semiclassical description of the orbital degrees of freedom of quasiparticle excitations . on the other hand , the internal degrees of freedom , _ i.e. _ the spin and the particle - hole degrees of freedom , are described by quantum mechanics . quantum mechanical coherence of particle and hole excitations is the basis of all superconducting effects such as persistent supercurrents , flux quantization , josephson effects and andreev reflection . here we use the quasiclassical theory for our investigations of the vortex core . numerical work on vortices in superconductors using the quasiclassical theory started with a series of publications by kramer , pesch , and watts - tobin . @xcite more recent work includes pinning of vortices at small defects,@xcite vortices in superfluid @xmath155he and other systems with unconventional pairing,@xcite the excitation spectrum of bound quasiparticles,@xcite and the spectrum of moving pancake vortices.@xcite we use in this article the notation of refs . . the central objects of the quasiclassical theory of superconductors in equilibrium are the quasiclassical propagators @xmath156 , which are @xmath17 matrices in the particle - hole index , @xmath157 the variables @xmath158 and @xmath159 stand for the energy of an excitation and its momentum ( on the fermi surface ) . the momentum variable reduces to @xmath160 for an isotropic fermi surface in two dimensions ( see section ii ) . general symmetries lead to the following fundamental relation between @xmath161 and @xmath162 , @xmath163 we use , as described in section ii , the notation @xmath18 , @xmath19 , @xmath20 , for the three pauli matrices in particle - hole space , and @xmath21 for the unit matrix . the off - diagonal terms @xmath164 in eq . ( [ matrx ] ) are the pair amplitudes . they vanish in the normal state , and measure the amount of particle - hole mixing in the superconducting state . the diagonal elements of the propagators determine the density of states , @xmath165 and the equilibrium current density . the most detailed information on the current distribution is obtained from the _ spectral current density _ , @xmath166 where @xmath167 is the dimensionless density of states for co - moving ( @xmath168 ) and counter - moving ( @xmath169 ) excitations along the trajectory line defined by @xmath159 , and @xmath170 is the fermi velocity at the point @xmath159 on the fermi surface . this spectral density measures the contributions of quasiparticle states with energy @xmath158 and momentum near the fermi surface point @xmath159 to the current density at position @xmath171 . the full current density is obtained by weighting the spectrally resolved current density by the occupation probability of the quasiparticle states , then integrating over fermi momenta and energies . for equilibrium states , @xmath172 where @xmath173 is the fermi distribution function . the symbol @xmath174 denotes a weighted integral over the fermi surface . the weight at @xmath159 is @xmath175 , and the integral is normalized , @xmath176 . the spectral current density is particularly well suited for our study of the importance of andreev bound states for the current flow in a vortex core . these bound states appear as delta functions in the spectral current density at energies below the bulk energy gap . the spectral weight of the delta function , combined with the occupation of the bound state , determine its contribution to the total current density . we calculate @xmath177 from eilenberger s transport equation@xcite @xmath178 + i\hbar{\bf v}_f\cdot\mbox{\boldmath $ \nabla$}\hat g^r({\bf p}_f,{\bf r};\epsilon)=0\ , , \ ] ] supplemented by the condition of analyticity in the upper half of the complex @xmath158-plane , and the normalization condition @xmath179 for a fixed fermi momentum @xmath159 this is a first order differential equation along a straight - line classical trajectory in the direction of the fermi velocity @xmath170 . the propagator @xmath180 at a chosen point of interest , @xmath181 , is determined by the solution of ( [ eilen ] ) along the trajectory through @xmath171 in the direction @xmath170 . complete information on the local physical properties at point @xmath171 , such as the current density , is obtained by sampling all trajectories through @xmath171 . the propagator @xmath182 is intimately related to the @xmath183 density matrix of the particle - hole degrees of freedom of a quasiparticle moving along the classical trajectory specified by @xmath159 , @xmath171 . thus , @xmath177 describes the state of the internal degrees of freedom of the excitation . the internal state , _ i.e. _ the amount of particle - hole mixing , may change along the trajectory as a consequence of the off - diagonal pair potential , @xmath184 , which acts as a driving term that ` rotates ' the particle - hole pseudo spin . the pair potential couples particle and hole excitations , and is the origin of particle - hole coherence . it depends on the real space position , @xmath171 , and , for anisotropic superconductors , on the fermi surface position , @xmath159 , @xmath185 the pair potential must be calculated self - consistently from the gap equation , @xmath186 where @xmath187 is the pairing interaction , which determines the orbital symmetry of the pair potential , its magnitude and @xmath9 . our procedure for numerical calculation of the currents in the core of 2d pancake vortices is the following . we first solve self - consistently the gap equation and eilenberger s equation at matsubara energies . this allows us to determine the pair potential and the supercurrent density . we then insert the pair potential into eilenberger s differential equation at real energies , and obtain from its solution the excitation spectrum : the density of states and the spectral current density . the differential equations are solved by standard 4@xmath188 order runge - kutta and multiple shooting methods , and self - consistency is achieved iteratively by using alternatively a relaxation method and the mbius - eschrig algorithm.@xcite we consider three examples of pancake vortices : isolated , i ) singly - quantized and ii ) doubly - quantized @xmath189-wave vortices , and iii ) a pinned @xmath189-wave vortex in the presence of a uniform transport supercurrent . we choose a temperature of @xmath190 , unless otherwise noted , and assume @xmath191 . in this limit the vector potential is essentially constant in the core region , and can be neglected . figure 1 shows the amplitude of the order parameter of a singly - quantized @xmath189-wave vortex . the amplitude is isotropic and vanishes linearly in the core . the variation of the amplitude and phase along two trajectories are also shown in fig . 1 . for trajectory @xmath192 passing through the center of the vortex , the phase changes discontinuously and the amplitude vanishes linearly at the vortex center . for trajectory @xmath193 , with impact parameter @xmath194 , there is only a small change in the amplitude of @xmath195 . for singly - quantized vortices the phase of the order parameter is the more important factor determining the spectrum of bound states . figure 2 shows the angle - resolved local density of states for the two trajectories shown in fig . 1 . for trajectory ( a ) through the center of the vortex , the spectrum shows a zero - energy bound state separated from the continuum that begins at the bulk gap . the bound state results from constructive interference of particle- and hole - like quasiparticles that undergo andreev reflections from the vortex order parameter . this bound state corresponds to the zero angular momentum bound state found by caroli , de gennes and matricon.@xcite a zero - energy bound state is always present for trajectories in which the order parameter is real ( up to a constant phase factor ) and has different signs when going to @xmath196 along the trajectory.@xcite bound states with non - zero energies are found for trajectories with a nonzero impact parameter measured from the vortex center . these bound states correspond to the spectrum of bound states with non - zero angular momenta obtained by caroli et al.@xcite figure 2b shows the spectrum for a trajectory with an impact parameter of @xmath197 and @xmath198 measured at the point of closest approach to the vortex center . the bound state is shifted down in energy to @xmath199 , and the continuum states are shifted and inhomogeneously broadened by the doppler energy , @xmath200 . the spectrum near the onset at point @xmath201 in fig . 2b has low weight and corresponds to the continuum edge at @xmath202 far from the impact point , while the peak in the spectrum at point @xmath51 corresponds to the maximum doppler shift , @xmath203 at the impact point @xmath204 . note the development of the bcs coherence peak as the density of states is sampled further from the vortex center . 2 = = the density of states of an @xmath189-wave vortices has been investigated by several authors.@xcite our emphasis is on the importance of the andreev bound states for the current distribution in the vortex core . we show in fig . 2c the spectral current density for the trajectory with @xmath205 and @xmath206 . the net current carried by the states at the point @xmath207 on the fermi surface is obtained by weighting this spectrum by the equilibrium distribution and integrating over all energies . thus , for @xmath208 only the negative energy states contribute . it is clear from fig . 2c that the current in the vicinity of the vortex core is carried almost entirely by the bound states with @xmath209 . the continuum states give almost no net contribution to the current in the core . figure 2d shows the spectral current density of the set of bound states with trajectories @xmath210 as a function of the impact parameter @xmath73 for @xmath211 . the peak at @xmath212 corresponds to the trajectory with impact parameter @xmath213 . the bound state energy decreases with increasing distance from the core . for small @xmath73 we obtain , @xmath214 , in reasonable agreement with the analytic estimate in eq . ( [ eig ] ) . as indicated in fig . 2d the contribution of the bound state to the current density decreases as the impact parameter increases . however , even at a relatively large distance , @xmath215 , the bound state still contributes significantly to the circulating current density of the vortex . the evolution of the bound state energy for small impact parameters can be written in terms of the angular momentum of an excitation about the vortex center , @xmath216 ; _ i.e. _ @xmath217 , where @xmath218 . this spectrum was originally obtained by caroli et al.@xcite by solving the bogolyubov equations . in the bogolyubov or gorkov formulation the spectrum is discrete : @xmath219 with @xmath220 and @xmath221 defining the level spacing of the low - lying bound states in the core . the lowest energy bound state in the core has a zero - point energy of @xmath222 which is outside the resolution of the quasiclassical or the andreev theory . the discrete spectrum of the bogolyubov theory corresponds to the continuous andreev spectrum in the limit where the level spacing is small compared to all other relevant energy scales , _ i.e. _ @xmath223 , etc . this is generally an excellent approximation in conventional type ii superconductors . for the high @xmath9 cuprates the discrete level structure is expected to play a more important role , particularly in the transverse response of vortices in the ultra - clean limit , _ i.e. _ @xmath224 , where @xmath225 is the mean scattering time.@xcite = it is interesting to compare the single - quantum vortex with the axially symmetric , @xmath11 vortex , @xmath226 . the double quantum vortex has higher energy than a pair of isolated single - quantum vortices ; however , once created the double - quantum vortex is metastable against dissociation into singly - quantized vortices . the amplitude of the order parameter for the double - quantum vortex decreases as @xmath227 for @xmath228 as shown in fig . = in contrast to the @xmath10 vortex there is no sign change of the order parameter for trajectories passing through the center of the vortex core . this difference has a profound effect on the spectrum of andreev bound states in the core . 3b shows the excitation spectrum of the doubly - quantized vortex at the center of a trajectory passing through the center of the vortex core . a symmetric spectrum of two bound states at @xmath229 are separated from the continuum . figure 3 also shows the current density of the doubly - quantized vortex . the remarkable feature is the _ reversal _ of the current direction in the core , i.e. for @xmath230 ( see fig . 3c and 3d ) . this current anomaly is associated with the appearance of a counter - moving andreev bound state below the fermi level ( @xmath231 ) . the evolution of the spectral current density is shown in fig . the trajectories are parallel to @xmath232 and the spectral current density is shown as a function of the impact parameter . at distances greater than @xmath233 two bound states lie below zero energy , and both states are co - moving with the circulating phase gradient , @xmath105 . as the vortex core is approached the co - moving bound state nearest the fermi level moves to higher energy , and a counter - moving bound state above the fermi level ( not shown ) moves to lower energy . these two states cross the fermi energy ( @xmath231 ) at approximately @xmath234 , leading to a reversal of the integrated current density inside the core . the cumulative current density for each trajectory is shown as the thick solid line in each panel of fig . 4 . = finally , consider the current and excitation spectrum of a @xmath10 vortex in the presence of a uniform supercurrent @xmath235 . in the absence of pinning the vortex will move in the direction ( @xmath236 ) in order to reduce the kinetic energy . thus , in order to investigate the excitation spectrum in the presence of a transport current we must pin the vortex to the lattice . our model for the pinning center is a normal metal inclusion where the pairing interaction ( or the local @xmath9 ) vanishes . 5a shows the order parameter of a pinned vortex for an @xmath189-wave superconductor and a pinning center with a radius of @xmath237 . in the presence of a transport current the amplitude of the order parameter deforms ; it is suppressed on the high current side of the vortex as shown in fig . 5a . for relatively small transport currents , e.g. @xmath238 , the center for the phase winding lies within the normal inclusion . however , as shown in fig . 5b , the vortex current no longer vanishes at the center of the vortex core ; there is substantial current _ through _ the vortex core region , including the normal inclusion . the current density inside the normal inclusion is carried by the andreev bound states and is a consequence of the proximity effect . the bound state spectrum at the center of the vortex is shown for a trajectory parallel to the transport current and passing through the vortex center . the negative energy bound state carries the transport current inside the normal inclusion . 5d shows the spectral current density measured at the center of the normal inclusion for the trajectories with @xmath239 . note that the bound state dominates the current and that this current is opposite to the applied transport current . this result was also obtained in section [ analyt ] , without taking into account the distortion of the vortex core order parameter . this led to a violation of charge conservation in the core . our numerical calculation shows that the main features of the analytic model for the bound state spectrum and the self - consistent determination of the order parameter for the pinned vortex in the presence of a transport current guarantees that charge is conserved . we have discussed the current carried by the excitations of s - wave vortices in clean layered superconductors . the _ spectral current density _ was introduced in order to identify the excitations that determine the transport and circulating currents of a vortex . the bound states of the vortex carry most of the current in the vicinity of the core , including transport currents that flow through the core of a pinned vortex . far from the vortex core currents are carried primarily by the bound - pair continuum that forms the condensate . for currents flowing through a pinned vortex , current conservation is maintained by `` spectral transfer '' of the current carried by the andreev bound states to the continuum states outside the core . a novel example of the evolution of the spectral current density is provided by the double quantum vortex which shows the connection between the spectrum of bound states and the symmetry or topology of the order parameter . at low temperatures ( @xmath240 ) the double quantum vortex exhibits a ` current reversal ' relative to the asymptotic direction of the circulation . the counter - circulating current in the core is due to a counter - moving bound state that appears below the fermi level and dominates the current for distances of order @xmath241 . at high temperatures , @xmath242 , this counter - moving bound state is thermally depopulated with the result that the current reversal in the core disappears in the ginzburg - landau limit . in summary , we find that the andreev bound states dominate the current of vortices on the scale of a few coherence lengths . the nonequilibrium properties of vortices on this scale are expected to be dominated by the spectral evolution and dynamics of these bound states . this work was initiated when the authors were participants at a workshop at the institute for scientific interchange , villa gualino , torino . the research of dw was supported in part by the engineering and physical sciences research council , while that of dr and jas was supported in part by nsf grant dmr 91 - 20000 through the science and technology center for superconductivity , the max planck gesellschaft and the alexander von humboldt stiftung . we also thank dr . m. j. graf for his comments on the manuscript . in this appendix we derive a form for a matrix element used in section [ div ] . the matrix element in question is @xmath243|\zeta \rangle _ { 1,1}$ ] where we write @xmath244 we have @xmath245 = i\left ( \hat\tau_{3 } ( \hat{h}_a-\hat{\delta } ) \delta ( \epsilon -\hat{h}_a)-\delta ( \epsilon -\hat{h}_a)(\hat{h}_a-\hat{\delta } ) \hat\tau_{3}\right)\ ] ] @xmath246-i\{\hat\tau_{3}% \hat{\delta } , \delta ( \epsilon -\hat{h}_a)\}\,.\ ] ] then @xmath247|\zeta \rangle _ { 1,1 } = i \mbox{tr}\left [ \frac 12(1+\hat\tau_{3})\langle \zeta |([\hat\tau_{3}\epsilon , \delta ( \epsilon -\hat{h}_a)]- \{\hat\tau_{3 } \hat{\delta } , \delta ( \epsilon -\hat{h}_a)\})|\zeta \rangle \right]\ ] ] @xmath248 \,.\ ] ] substituting the explicit form for @xmath249 yields the relation @xmath250|\zeta \rangle _ { 1,1}=\delta _ { 0}\frac{f(r)}r\mbox{tr}\left [ ( \zeta \hat\tau_{2}-\eta \hat\tau_{1})\langle \zeta |\delta ( \epsilon -\hat{h}% _ a)|\zeta \rangle \right ] .\ ] ] the parameter @xmath252 appearing in the above equation has the semiclassical interpretation as a c - number impact parameter . in this appendix we present approximations to the above equation for large values of the impact parameter , @xmath252 . for @xmath253 we are justified to replace @xmath254 by its asymptotic value of unity . furthermore , making the somewhat crude approximation @xmath255 yields for the bound state @xmath256 this indicates that at large impact parameters , the bound states of the andreev equation are found very close to the threshold of the continuum . when required , more refined estimates may be obtained by writing @xmath257\ ] ] and then performing a unitary transformation @xmath258 , @xmath259 , with @xmath260 $ ] to remove the phase from the order parameter . the transformed andreev equation is @xmath261 \widetilde{\psi } ( \zeta ) = e\widetilde{\psi } ( \zeta ) \,,\ ] ] this is a one dimensional dirac equation with a weak scalar potential , which has weakly bound states with energies near @xmath262 . a `` non - relativistic '' treatment is appropriate in this case and we approximate the dirac equation by a schrdinger equation . for example , for @xmath263 we write @xmath264 with @xmath265 scalars . straightforward manipulations indicate that @xmath266 approximately obeys the schrdinger equation @xmath267 \psi _ l(\zeta ) = ( e-\delta _ { 0})\psi _ l(\zeta ) .\ ] ] all the machinery of schrdinger theory may be used on this equation to estimate e.g. the bound states . we can put a lower limit on the bound state energy . this may be obtained from the fact that the eigenvalues of the schrdinger operator are @xmath268 , the minimum of the potential . thus , @xmath269 $ ] , i.e. @xmath270 note that @xmath272 is proportional to an @xmath273 integral of @xmath274 \langle \zeta |\delta ( \epsilon-\hat{h}_a)|\zeta \rangle _ { 1,1}$ ] . since @xmath275 and@xmath276 \leq 0 $ ] it immediately follows that @xmath277 for a recent review see n. kopnin , _ theory of flux flow hall effect in clean type ii superconductors _ , to be published in `` quasiclassical methods in the theory of superconductivity and superfluidity '' , eds . d. rainer and j. a. sauls , springer - verlag , 1996 .
we calculate the spectrum of quasiparticle excitations in the core of isolated pancake vortices in clean layered superconductors . we show that both the circular current around the vortex center as well as any transport current through the vortex core is carried by localized states bound to the core by andreev scattering . hence the physical properties of the core are governed in clean high-@xmath0 superconductors ( e.g. the cuprate superconductors ) by the andreev bound states , and not by normal electrons as it is the case for traditional ( dirty ) high-@xmath0 superconductors .
the advanced camera for surveys ( acs ; ford et al . 1998 ) virgo cluster survey is a project based on observations with the acs on the _ hubble space telescope _ ( ) , aimed at the exploration of the properties of 100 early type galaxies in the virgo cluster , the study of their globular cluster systems , and the reconstruction of virgo s three dimensional structure . to measure the galaxy distances needed to realize these goals , we have used measurements of surface brightness fluctuations ( sbf ) . the sbf method was first proposed by tonry & schneider ( 1988 ) and it is based on the fact that the variance of normalized poissonian fluctuations of the galaxy stellar population depends on galaxy distance ( for reviews see jacoby et al . 1992 and blakeslee et al . 1999 ) . at present , sbf measurements have been used to determine galaxy distances out to @xmath11 7000 km s@xmath12 , using data from ground based telescopes and ( ajhar 1997 , 2001 ; tonry 1997 , 2001 ; jensen 1999 , 2003 ; blakeslee 2001 , 2002 ; mei 2001 , 2003 ; liu 2001 , 2002 ; mieske et al . 2003 ; and references in mei et al . 2005 , hereafter paper iv ) . for the virgo cluster in particular , sbf measurements have been published by several groups using data from both ground based telescopes ( tonry et al . 1990 , 2000 , 2001 ; pahre & mould 1994 ; jensen et al . 1998 ; jerjen et al . 2004 ) and ( ajhar et al . 1997 ; neilsen & tsvetanov 2000 ; jensen et al . 2003 ) . west & blakeslee ( 2000 ) have used published ground - based sbf data in a first examination of the three - dimensional structure of this cluster , particularly the cluster s `` principal axis '' and its relations to the surrounding large - scale structure . our new acs dataset ( for details , see ct et al . 2004 ; paper i ) constitutes the largest , most complete , and most homogeneous sbf sample of virgo galaxies yet assembled . since the absolute magnitude of the sbf varies as a function of the stellar population age and metallicity , sbf distance measurements in a given observational bandpass must be carefully calibrated in terms of stellar population observables , typically the galaxy color . using an extensive ground - based @xmath13-band sbf survey , tonry et al . ( 1997 , 2001 ) have established that it is possible to calibrate the sbf method using the integrated galaxy colors . in particular , they find that for bright ellipticals : 1 ) the absolute sbf magnitude @xmath14 is a linear function of the galaxy @xmath15 color over the range @xmath16 with an internal scatter @xmath17 mag ; 2 ) the zero - point of this relation is universal over this color range ; 3 ) the slope of the relation is consistent with theoretical predictions from recent stellar population models . for observational bands other than the @xmath13-band , the relation between sbf absolute magnitude and galaxy colors has to be calibrated in that particular band . ajhar et al . ( 1997 ) and jensen et al . ( 2003 ) have calibrated sbf absolute magnitudes as a linear function of galaxy color for use , respectively , with the wide field and planetary camera-2 ( wfpc-2 ) and nicmos on the . this paper provides the first calibration for sbf using the f850lp filter with the acs , i.e , calibration of the absolute f850lp sbf magnitude @xmath3 , as a function of dereddened color in the f475w and f850lp filters . observations and sbf reductions have already been described in paper i , jordn et al . ( 2004 ; paper ii ) and paper iv . in 2 the main steps of the sbf reduction are briefly summarized . in 3 the calibration method is discussed and the results are presented and analyzed . in 4 the theoretical predictions from bruzual & charlot ( 2003 ) stellar population models are discussed . we summarize and conclude in 5 . the calibration presented here will be used in future papers to derive sbf distances for the galaxies in our sample and to explore the three - dimensional structure of the virgo cluster . our observations and data reduction procedures have been discussed in paper i and paper ii , respectively . in paper iv , we have addressed the specific issues related to sbf measurements with the acs . since the acs camera presents significant geometrical distortions , the correction of these distortions involves image resampling by an interpolation kernel . for our sbf measurements we use images geometrically corrected with the _ drizzle _ software ( fruchter & hook 2002 ) using the lancsoz3 interpolation kernel . we have shown in paper iv that , when using this kernel , the distortion correction does not add significant error to the sbf distance measurements . the main steps of sbf measurements are also described in paper iv . we summarize here the main points . we measure sbf in the f850lp bandpass of the acs wide field camera . the process involves fitting a smooth model to each galaxy in both the f850lp and the f475w band ( papers ii and iv ) and subtracting the models from the original images to produce `` residual images '' . the residual images in each band are then used to construct catalogs of external sources , i.e. globular clusters and background galaxies . from these catalogs we create a source mask to remove the contamination from the sources to the surface brightness variance . all sources brighter than a completeness magnitude @xmath18 , that corresponds to a completeness of @xmath19 at a given radius , are masked from the image . the power spectrum of the residuals is then calculated in multiple concentric annuli centered on the galaxy center , by multiplying successive annular mask functions with the source mask ( i.e. , the total mask function ) . the image power spectrum is the sum of two components : a flat , white noise power spectrum and the combined power spectrum of the sbf and undetected external sources , both convolved by the psf ( point spread function ) in the spatial domain . in the fourier domain , this sum can be written as : @xmath20 where the `` expectation power spectrum '' @xmath21 is the convolution of the power spectra of the normalized psf and of the total mask function of the annular region being analyzed . in paper iv we have shown that this power spectrum is not significantly modified by pixel correlation induced by our interpolation kernel ( see paper iv for details ) . armed with a clearer understanding of the systematic errors arising from the data interpolation , we have measured unbiased values of @xmath22 and @xmath23 for multiple annuli in each galaxy as described in paper iv . as the minimum innermost radius , we have taken @xmath24 . however when dust is present in the galaxy center , the dust regions are masked ( see paper i ) , and the minimum radius without dust contamination has been taken as the minimum annular radius . for the maximum outermost annular radius , we have chosen the radius where the galaxy brightness falls to half that of the sky ( the range of maximum outermost radii being between @xmath25 and @xmath26 ) . the next step is then to extract the sbf signal from @xmath22 by subtracting off the residual variance @xmath27 from faint sources including undetected globular clusters and distant background galaxies . because of the exquisite resolution and depth of our acs images , these correction are very small ( see the following section ) . the contribution to the power spectrum from these external sources is indistinguishable from the sbf contribution , since both are spatially convolved with the psf . following tonry ( 1990 ) and blakeslee & tonry ( 1995 ) , we can write @xmath28 where @xmath29 @xmath30 is the contribution to the fluctuations due to globular clusters , and @xmath31 is the contribution due to background galaxies . the mean galaxy brightness in the denominator is chosen for compatibility with the @xmath22 definition above . to estimate the @xmath27 contribution , sources were extracted from the same catalog used for masking external sources from the residual images ( see above and paper iv ) . source counts were fitted to a composite globular cluster and external source luminosity function using a maximum likelihood approach . for the former , we adopted a gaussian luminosity function , where @xmath32 is a function of both radius @xmath33 ( distance from the galaxy center ) and magnitude @xmath34 : @xmath35 for the virgo cluster in the f850lp filter , we expect @xmath36 23 mag and @xmath37 @xcite . we use these values as our initial estimates but iterate on @xmath38 to improve the fitted luminosity functions . for the background galaxies , we assumed a power - law luminosity function : @xmath39 with @xmath40 . from the acs faint galaxy counts of benitez et al . ( 2004 ) @xmath41 in the f814w filter . at present , a value for @xmath42 in f850lp is unavailable in the literature . however , the f814w and f850lp bandpasses are quite similar so we expect no strong differences in @xmath42 between these filters . a fit of faint galaxy counts in our acs blank fields gives us a slope of @xmath43 ( peng et al . 2005 ) . if we vary @xmath42 in the range 0.33 to 0.37 , our fits produce differences of only a few thousandths of a magnitude in the final sbf magnitudes . in fitting the total source counts ( @xmath44 and @xmath45 ) , we kept as fixed parameters @xmath46 and @xmath42 , and iterated on @xmath38 , as a function of the galaxy distance . sources identified as bright foreground stars were removed from the images but not included in the fits . from the estimated @xmath44 and @xmath45 per pixel , @xmath47 was calculated as the sum of @xcite @xmath48}\\ \nonumber & \,\times\ , & \hbox{erfc}[\frac{m_{cut}-m^{gc}_{peak}+0.8\sigma^2\ln(10)}{\sqrt{2}\sigma}]\end{aligned}\ ] ] and @xmath49 where @xmath50 is the photometric zero - point on the ab system ( 24.862 mag for the f850lp ; sirianni et al . 2005 ) plus reddening corrections ( schlegel et al . 1998 ) as given in paper ii . an example of the external source luminosity function is shown in fig [ gc_example ] for ngc 4472 ( vcc 1226 ) . finally , the @xmath51-band sbf magnitudes are given by @xmath52 the uncertainty on @xmath53 in each galaxy has been calculated by propagating the measurement errors in eq . [ eq : mag ] for the different annuli . this means that we have added in quadrature the errors due to the uncertainty in the fit of @xmath22 , the uncertainty in the galaxy flux , and the uncertainty in @xmath47 ( @xmath11 25% @xmath47 , calculated as the variance in @xmath47 when different initial fit parameter are chosen ) . we neglect the @xmath81% systematic uncertainty in @xmath54 ( sirianni et al . 2005 ) for the calibration , since it is common to all galaxies . the total uncertainty on the average @xmath53 for each galaxy has been calculated as the quadrature sum of the errors for the different annuli which we are averaging , divided by the number of annuli . finally , since @xmath55 , where @xmath56 is the galaxy distance , two more sources of uncertainty must be added in quadrature to the measurement error in @xmath53 for the purpose of this calibration . first , we include an expected `` cosmic scatter '' of 0.05 mag in @xmath3 ( see tonry et al . 1997 ) , representing the intrinsic dispersion in @xmath3 for galaxies at the same observed color . second , an uncertainty of 0.11 mag is introduced to account for the actual distance dispersion from depth effects in the virgo cluster . this value has been derived from the rms scatter of @xmath57 in position on the sky of the 100 galaxies in our sample . we seek to calibrate @xmath3 as a function of the galaxy color ( f475w@xmath58f850lp ) , hereafter referred to , since the acs @xmath59 approximates the sdss @xmath60 and f850lp the sdss @xmath51 . the galaxy colors are measured in the same annuli , and with the same masks , as those used for measuring the sbf . the sky values in the two filters for our final calibration sample are the same as in paper ii . galaxies for which the sky estimates were especially difficult ( due to strong gradients from bright neighboring galaxies ) have been excluded for the purposes of this calibration ( see below ) . the errors on the colors have been calculated by summing in quadrature the error in galaxy flux and the error due to sky subtraction . when the colors have been averaged on more annuli , the errors in each one have been added in quadrature , squared and divided by number of annuli . an additional uncertainty of 0.01 mag in color is added in quadrature as an estimate of the error due to imperfect flatfielding ( see , e.g. , sirianni 2005 ) . the colors have been dereddened as explained in paper ii , with the reddening for each galaxy position from schlegel et al . the mean reddening for the sample is @xmath61 mag , with a standard deviation of 0.008 mag . the adopted extinction corrections are appropriate for elliptical galaxies , @xmath62 @xmath63 and @xmath64 @xmath63 . hereafter , we use @xmath65 to refer to the reddening - corrected galaxy color . for the calibration of the @xmath13-band sbf survey , tonry et al . ( 1997 ) divided their sample into galaxy groups and clusters , and assumed that , since the galaxies within an individual group or cluster are at nearly the same distance , sbf magnitudes will simultaneously fit the relation : @xmath66\ ] ] where @xmath67 is the group mean value at the chosen fiducial color @xmath68 mag , that was fitted for each group to establish the universality of the sbf calibration . to obtain the absolute sbf magnitude @xmath69 , tonry et al . ( 1997 ) calibrated the zero - point in eq . [ tonry ] with cepheid distances to galaxies in their groups . thus , they obtained the following calibration , valid in the color range @xmath70 : @xmath71.\ ] ] the ajhar et al . ( 1997 ) /wfpc2 sample spans a range in galaxy @xmath65 color between 1.15 and 1.3 mag . their calibration for the wfpc2 f814w bandpass is @xmath72.\ ] ] the jensen et al . ( 2003 ) hst / nicmos sample in the f160w ( @xmath73 m ) filter was calibrated over the range @xmath74 as : @xmath75.\ ] ] these three samples were heavily weighted towards bright , red , early type galaxies . the ajhar and jensen samples ( due the small field sizes ) were mainly confined to the central regions of bright galaxies , which tend to be quite red . blakeslee ( 2001 ) concluded that this fact explained the steep slope found by ajhar , since the stellar population models predict a steepening of the sbf color relation for very red populations . calibrations of bluer samples of dwarf early - type galaxies have been proposed by jerjen et al . ( 1998 , 2000 , 2003 , 2004 ) and mieske et al . ( 2003 ) using ground based observations of @xmath76 dwarfs coupled with the predicted color - dependence from stellar population models . however , until now , there has not been a sufficient sample of high - quality sbf measurements in dwarf ellipticals to permit a dependable _ empirical _ calibration . our acs vcs sample is the most extensive and homogeneous early type galaxy sample available for sbf measurements . the galaxies span a range in color of @xmath77 mag , corresponding to a range in @xmath65 between 0.94 and 1.3 mag ( based on the color transformations from sirianni et al . we adopt here the same technique that tonry et al . ( 1997 ) have used for galaxy groups , fitting the relation : @xmath78\ ] ] for the virgo cluster . among the 100 galaxies in our full sample , 14 have been omitted from our calibration sample because they are edge on disk or barred galaxies which are especially challenging for sbf measurements , or because of difficult sky subtraction . these galaxies include : vcc 571 , vcc 654 , vcc 685 , vcc 1025 , vcc 1125 , vcc 1192 , vcc 1199 , vcc 1327 , vcc 1535 , vcc 1627 , vcc 1720 , vcc 1857 , vcc 2048 , and vcc 2095 . one galaxy in the sample , vcc 1499 , has @xmath79 , and is too blue to be reliably included in a calibration . there are also several outliers that do not appear to lie on the average sbf magnitude versus color relation for the cluster . [ allsample ] shows the sbf and color measurements for the remaining 85 galaxies . to calibrate our sbf measurements in the f850lp filter as a function of the average galaxy @xmath0 color , we calculate average sbf measurements for each galaxy and fit the relation in eq . [ eq : cal ] to average galaxy colors , as tonry et al . ( 1997 ) have done for each group in their sample . all galaxies included in the fit have an uncertainty on @xmath53 less than 0.15 mag ( not considering the uncertainty due to the ` cosmic ' scatter on the sbf calibration ; see above ) . errors on average galaxy colors are typically @xmath80.01 mag . both uncertainties on magnitude and color are included in the @xmath80 calculations below . to exclude the outliers from the final calibration , we first make a linear fit to eq . [ eq : cal ] to the overall sample with a robust least absolute deviation technique ( press et al . we obtain @xmath81 mag and @xmath82 . we identify four outlier galaxies from iterative 3-@xmath46 clipping . the four 3@xmath46 outliers are : vcc 21 , vcc 538 , vcc 575 , and vcc 731 . of these , vcc 21 contains numerous bright young star clusters , indicative of recent star formation , while vcc 731 ( ngc 4365 ) is known from the ground - based sbf survey to be a virgo member which lies on the far side of the cluster ; we suspect that the other two may be similar cases . after excluding these four outliers and the very blue vcc 1499 , our final sample consists of 82 galaxies . in fig . [ slope ] our final galaxy sample and an overall fit ( the continuous straight line ) are shown . both uncertainties on magnitude and color have been included in the fit @xmath80 estimation . we obtain : @xmath83 \nonumber\ ] ] in the overall color range @xmath84 . this color range includes both bright and faint galaxies . various lines of evidence have suggested that the slope of the @xmath13-band sbf color relation is steeper for redder galaxies : for instance , the steep empirical slope found by ajhar et al . ( 1997 ) for the red centers of bright ellipticals , which was also reproduced by the reddest composite stellar population models of blakeslee ( 2001 ) . conversely , the few dwarf galaxies in the tonry ( 1997 ) sample indicated a significantly shallower slope at the blue end , and the sbf measurements for galactic globular clusters ( ajhar & tonry 1994 ) showed no dependence on color . most recent models also show a flattening of the relation for bluer populations . our sample , more clearly than any previous study , suggests that changes in color range affect the fitted slope , so we have studied the stability of the above relation as a function of the lower color limit , @xmath85 . when we repeat the fit in different ranges @xmath86 , we obtain the results shown in fig [ col_cut ] . on the top of the figure the fitted slope @xmath87 is shown as a function of @xmath85 . on the bottom is shown the corresponding value of @xmath80 . red galaxies tend to have slopes that are steeper ( @xmath11 2 ) than those obtained by fitting to the overall sample . this suggests that our sample would be better represented by two distinct samples of red and blue galaxies , each one having its own calibration slope . we fit eq . [ eq : cal ] in two complementary color ranges : @xmath88 ( blue ) and @xmath89 ( red ) , with @xmath90 ( as suggested from fig . [ col_cut ] ) . in fig . [ col_div ] the fitted slopes ( top ) and the respective @xmath80 ( bottom ) are shown as a function of @xmath91 . dot - dashed lines correspond to the fit in the color range @xmath92 and dashed lines to the fit in the color range @xmath88 . the continuous line is the total @xmath80 from both fits . the minimum total @xmath80 is obtained at @xmath93 , where the red slope is @xmath94 and the blue slope is @xmath95 . the two slopes cross at @xmath96 , which we take to be the color at which the two regimes separate . the @xmath80 for the total fit is minimized for both slope and zero - point , and the errors in both magnitude and color have been taken into account . as in tonry ( 1997 ) , we find the sbf in s0 and ellipticals to follow the same trend as a function of @xmath0 . interestingly , as figs . 2 and 5 show , the dividing color of @xmath97 is close to that which separates the two most basic galaxy types in our survey : giant and dwarf early - type galaxies . in fact , in our final sample , of 37 galaxies with color @xmath98 , only one is classified as dwarf . for @xmath99 , 10 galaxies are dwarfs and 8 are giants . for @xmath100 , of 26 galaxies , 20 are dwarfs . physically then , the two different calibration slopes correspond to different galaxy types ( giants corresponding to the red end , and dwarfs to the blue end ) . jerjen et al ( 2000 , 2004 ) adopted a theoretical calibration to dwarf elliptical galaxies , based on a two branch prediction in the r band sbf versus the @xmath101 galaxy color from worthey et al . ( 1994 ) theoretical stellar population models using the padova isochrones . while the scatter becomes larger for the bluest dwarfs , we see no evidence for distinct branches in our large data set ; nor do the bruzual & charlot ( 2003 ) models predict such branches in @xmath102 vs @xmath0 . our sample is sufficiently complete at the blue end that we are able to provide , for the first time in sbf studies , an empirical calibration of sbf in early type dwarf galaxies . splitting the galaxy sample at @xmath103 , we derive the following calibration for @xmath104 : @xmath105 \nonumber\ ] ] in the range @xmath106 , and @xmath107 \nonumber\ ] ] in the range @xmath108 . the reduced @xmath80 values for the red and blue subsamples are 0.99 and 0.8 , respectively . the value of the total reduced @xmath80 is calculated as the appropriately weighted sum of the @xmath80 values from the two independent fits . when normalized by the total number of objects used in the fits , its value is 0.9 . the reduced @xmath80 for the blue subsample is lower than that for the red sample . this might be due to the difficulty in accounting for the intrinsic distance dispersion within virgo , or to overly conservative error estimates for the dwarfs . one of the outliers rejected above , vcc 21 , is no longer an outlier with respect to the final , shallower , blue - end calibration , and we have added it to the final calibration above , by iteration . we show in fig . [ scatter ] the average and standard deviation of @xmath53 in @xmath0 color bins ( top ) and the color histogram ( bottom ) of our final calibration sample ( 82 galaxies , plus vcc 21 ) . from the scatter in @xmath53 , we are aware that our calibration has a large scatter for colors @xmath109 . we will limit our absolute magnitude calibration to the color range @xmath110 . we observe a similar behavior of @xmath111 as a function of the @xmath0 color for the multiple annuli within individual galaxies . however , the precise value of the slope is less robust and more sensitive to errors in the sky . empirical calibration of the zero - point of the sbf relation requires computing the sbf absolute magnitude @xmath112 from independent distance measurements . the ideal situation would be to have cepheid distances for multiple galaxies in our sample , however there are no galaxies in our sample with cepheid distances . we thus calibrate our zero - points to the virgo cluster distance modulus found in the @xmath13-band sbf survey of tonry et al . ( 2001 ) , but revised in accordance with the final set of key project cepheid distances from freedman ( 2001 ) . tonry et al . ( 2001 ; table 4 ) reported @xmath113 mag as the virgo sbf distance modulus , where the sbf zero - point was calibrated from sbf measurements for individual early - type spiral galaxies with cepheid distances given by ferrarese et al . ( 2000 ) . however , using the final set of key project distances from freedman et al . ( 2001 ) , who adopted the udalski et al . ( 1995 ) period - luminosity relation and added a metallicity correction , the tonry et al . ( 2001 ) distance modulus for virgo has to be corrected with a shift of @xmath580.06 mag , and becomes @xmath114 mag ( see discussions by ajhar et al . 2001 ; blakeslee et al . 2002 ) . here , the errorbar reflects the internal error on the mean virgo sbf distance ; the uncertainties in tying the ground - based sbf distances to the cepheid distance scale , and in the zero - point of the cepheid distance scale itself , contribute another @xmath80.15 mag systematic error . if the metallicity corrections to the cepheid distances tabulated by freedman are not applied , then the tonry et al . ( 2001 ) virgo distance modulus would shift instead by @xmath115 mag to @xmath116 mag ( see jensen et al . 2003 ) . for our sbf calibration , we will adopt @xmath114 mag as the virgo distance modulus . thus , scaling the zero - points from eq.[eq : zero ] , we obtain @xmath117 \nonumber\ ] ] in the range @xmath5 , and @xmath118 \nonumber\ ] ] in the range @xmath119 . we now compare our results to predictions from the bruzual & charlot ( 2003 ) stellar population models , which we have calculated from the models and stellar luminosity functions kindly provided by s. charlot . in fig . [ cha ] we show our data as compared to bruzual & charlot ( 2003 ) model predictions for different stellar populations . the lines connect models of the same metallicity : the red line is for solar metallicity , the green for 40% solar , the blue for 20% solar , the light blue for 2% solar , and the orange for 2.5 times solar metallicity . the dots on the lines represent , respectively from left to right , ages of 1 , 3 , 5 , 8 , 12 and 15 gyr . the bruzual & charlot models predict @xmath3 . to compare these predictions with the data , we add to the theoretical @xmath3 a distance modulus such that the predicted average @xmath53 for solar and 40% solar metallicity at @xmath120 mag is equal to the average observed @xmath53 at the same @xmath0 for our calibration sample . this procedure yields a theoretically predicted sbf virgo cluster distance modulus of @xmath121 mag , which is entirely independent of the cepheid distance scale , but we caution that it depends on the color range we have chosen for shifting the models into agreement with the data . this theoretically calibrated sbf virgo cluster distance modulus differs by @xmath110.15 mag with respect to the revised tonry et al . ( 2001 ) virgo distance modulus of 31.09 mag adopted above . however , the two are consistent within the uncertainties from the zero - point of the cepheid distance scale , the anchoring of the ground - based sbf measurements to the cepheids , and the various uncertainties in the stellar population models ( see blakeslee et al . 2001 for a discussion ) . it is also worth noting that if we had adopted the virgo distance modulus of @xmath116 obtained by calibrating sbf to the freedman ( 2001 ) cepheid distances that use the new period luminosity calibration but omit the metallicity correction , then the difference with respect to the model - predicted distance modulus would be only @xmath122 mag . our sample of early - type galaxies spans a wide range in metallicity and , presumably , age . like the observations , the theoretical predictions also show a steeper slope at the red end , and a shallower at the blue end . the slopes predicted from linear fits to the overall model predictions are @xmath123 for @xmath124 mag , @xmath125 for @xmath126 mag , and @xmath127 for @xmath128 mag ( in very good agreement with our empirical slope over the full range ) . if we fit only metallicities less than or equal to solar , at the red end the slope will be @xmath129 , compared to @xmath130 in the blue . the agreement between theoretical predictions and our empirical calibration is somewhat better at the red end . this is likely due to better accuracy in the modeling of highly evolved populations . the sbf predictions for young to intermediate - age populations is less certain due mainly to the uncertainties in the modeling of the asymptotic giant branch contributions . furthermore , models of old stellar populations are better constrained by observations of galactic globular clusters . for these reasons , we prefer our empirical blue - end slope , although additional sbf measurements of large homogeneous samples of dwarf galaxies ( e.g. , in the fornax cluster ) would be helpful in refining the sbf calibration for blue early - type galaxies . considering the uncertainty in the virgo distance modulus , we conclude that our @xmath2 sbf data for the giant early - type galaxies are well reproduced by the models for stellar populations ages greater than 3 gyr and metallicities in the range between 40% and 250% solar . the data for the dwarf galaxies are consistent with either much younger ( less than 3 gyr ) solar metallicity populations or intermediate - to - old stellar populations with metallicities in the range 20%40% solar . as part of the acs virgo cluster survey , we have observed 100 early type galaxies in the virgo cluster . we have measured sbf magnitudes in these galaxies with the aim of determining their distances and the three - dimensional structure of virgo . we have presented the first calibration of the sbf method using the acs instrument , giving the f850w sbf magnitude as a function of galaxy @xmath0 color . obvious outliers , most likely due to a real dispersion in distance for galaxies in the virgo cluster region , have been omitted from this calibration , leaving 82 galaxies with @xmath0 colors ranging between 0.9 and 1.6 mag . over this color range , the blue ( mainly dwarf ) and red ( mainly giant ) galaxies follow different linear relations , with the slope being shallower at the blue end . for the first time in sbf studies , our sample allows us to empirically calibrate sbf for dwarf early type galaxies . by comparing with stellar population models from bruzual & charlot ( 2003 ) , we find that the slope of the model prediction for @xmath53 with @xmath65 color is highly consistent with the observed relation for the red galaxy subsample . there is also general consistency at the blue end , though the stellar population models with young ages and low - metallicities predict a somewhat shallower slope than observed . this underscores the importance of empirical calibrations , as well as more realistic composite populations in any effort to model the behavior of the sbf method theoretically , especially for dwarf galaxies , for which the scatter in stellar population properties is larger . a forthcoming paper will present the distances to the individual galaxies derived with this calibration , and examine the issue of virgo s three dimensional structure and its implications for galaxy and globular cluster properties . we thank stphane charlot for providing theoretical stellar population models in the acs filters and gerhard meurer for useful discussions . support for program go-9401 was provided through a grant from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc . , under nasa contract nas5 - 26555 . acs was developed under nasa contract nas 5 - 32865 . s.m . and j.p.b . acknowledge additional support from nasa grant nag5 - 7697 to the acs team . p.c . acknowledges support provided by nasa ltsa grant nag5 - 11714 . m.j.w . acknowledges support through nsf grant ast-0205960 . d.m . acknowledges support provided by nsf grants ast-0071099 , ast-0206031 , ast-0420920 and ast-0437519 , by nasa grant nng04gj48 g , and by grant hst - ar-09519.01-a from stsci . m.m . acknowledges support from the sherman m. fairchild foundation . this research has made use of 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 .
as part of the advanced camera for survey ( acs ) virgo cluster survey , we have measured surface brightness fluctuations ( sbf ) in a sample of 100 early - type virgo galaxies . distances derived from these measurements are needed to explore the three - dimensional structure of the virgo cluster , study the intrinsic parameters of globular clusters associated with the program galaxies , and compare with the galaxy distances derived from globular cluster luminosity functions . our sbf measurements have been performed in the f850lp bandpass of the wide field channel of the acs on the hubble space telescope . these are the first measurements of this kind , and we present the first sbf calibration for this bandpass . the measured fluctuations depend on galaxy stellar population variations , which we quantify by galaxy color @xmath0 , where @xmath1 and @xmath2 are the galaxy magnitudes , respectively , in the f475w and f850lp acs filters . we derive the following calibration for the absolute sbf magnitude @xmath3 : @xmath4 \nonumber\ ] ] in the range @xmath5 , and @xmath6 \nonumber\ ] ] in the range @xmath7 . the quoted zero - point uncertainty here includes all sources of internal error ; there is an additional systematic uncertainty of @xmath80.15 mag , due to the uncertainty in the distance scale calibration . physically , the two different color regimes correspond to different galaxy types : giant ellipticals and s0s at the red end , and early - type dwarfs at the blue end . for the first time in sbf studies , we are able to provide a firm empirical calibration of sbf in early type dwarf galaxies . our results agree with stellar population model predictions from bruzual & charlot ( 2003 ) in the range @xmath9 , while our empirical slope is somewhat steeper than the theoretical prediction in the range @xmath10 .
gale duality is a classical notion in convex geometry . since its appearance in @xcite it allowed to prove many strong and nontrivial results for convex polytopes and configurations of points on a sphere ( the survey of this field can be found in @xcite ) . in this paper we describe a surprisingly simple connection between gale diagrams and combinatorial alexander duality . for any set of points on a sphere @xmath4 we associate a covering of @xmath4 by hemispheres . from the theory of gale duality follows that the nerves of such coverings are exactly those complexes , which are alexander dual to boundaries of simplicial polytopes , or , more generally , to nerve - complexes of polytopes ( see claim [ claimgalealex ] for the precise statement ) . on one hand , this gives a combinatorial characterization of complexes dual to boundaries of simplicial polytopes . on the other hand , geometrical considerations , involving coverings by hemispheres allowed to prove particular statements about convex polytopes . in section [ secsimpconstr ] we review and define basic constructions , used in the work . these include alexander duality for simplicial complexes ; nerve - complexes of nonsimplicial polytopes and the construction of a _ constellation complex _ for a configuration of points on a sphere . in section [ secconstel ] are listed the most important topological and combinatorial properties of constellation complexes . in section [ secgale ] the gale duality is applied to show that constellation complexes are alexander dual to nerve - complexes of polytopes . alexander duality allows to simplify and treat topologically many well known results . in section [ secbetti ] we provide basic definitions from commutative algebra . arguments , similar to those used by eagon and reiner in @xcite are applied to constellation complexes . we use hochster formula to show that the stanley reisner ideal of a constellation complex @xmath5 has a linear resolution . this means that all generators of modules in the minimal resolution are concentrated in prescribed degrees . alexander duality leads to the following result : if @xmath6 is a gale diagram of a simplicial polytope @xmath1 , then bigraded betti numbers of the constellation complex @xmath5 coincide with the @xmath3-vector of @xmath1 ( proposition [ propbetticonstel ] ) . this correspondence can be naturally generalized to polytopes which are not simplicial ( proposition [ propbetticonstelgen ] ) . on the other hand , one can calculate bigraded betti numbers of a polytope @xmath1 by studying the combinatorial topology of @xmath5 . originally , this research was motivated by the study of buchstaber invariant ( definition [ definbuchnumber ] ) . this invariant of simplicial complexes and polytopes appeared naturally in toric topology , and in 2002 v.m.buchstaber posed a problem : to describe this number combinatorially . since then several approaches to this problem had been developed . i.izmestiev @xcite found a connection between buchstaber invariants and a chromatic number . this connection was further developed in work @xcite . in the paper @xcite real buchstaber invariant of skeleta of simplices was determined by integer linear programming . n.yu.erokhovets in his thesis @xcite , and other works @xcite built the theory of buchstaber invariants , constructed many examples and estimations , and found equivalent definitions for these numbers . we refer the reader to his surveys @xcite or @xcite to find out more about particular results and open problems in this field . recent result @xcite and gale duality allowed to prove the following conjecture , made in @xcite . if @xmath1 is a polytope , then @xmath7 if and only if @xmath1 is a pyramid ( theorem [ thmpyramids ] ) . result of @xcite can also be applied to construct polytopes with @xmath2 from their gale diagrams . this consideration had led to interesting combinatorial consequences ( theorem [ propfanocircle ] and statement [ propfanocircle ] ) . we hope that gale diagrams will allow to answer the question posed by nickolai erokhovets in @xcite : whether buchstaber invariant is determined by bigraded betti numbers of a simplicial complex ? probably , gale diagrams will lead to the solutions of other open problems concerning buchstaber invariants of polytopes . the author is grateful to professor v.m.buchstaber for his suggestion to consider the invariant @xmath8 on the class of nonsimple polytopes and for his interest to this work . also i wish to thank the participants of the student geometry and topology seminar in moscow state university . a few talks made at this seminar turned out to be very useful for the understanding of topics mentioned in this paper . let @xmath9 be a simplicial complex on a set of vertices @xmath10=\{1,\ldots , m\}$ ] . in the following the complex and its geometrical realization are denoted by the same letter for the sake of simplicity . by @xmath11}$ ] or @xmath12 we denote the simplex on a set @xmath10 $ ] . suppose , @xmath13}$ ] . in this case the _ dual complex _ @xmath14 ( or @xmath15 ) is a complex on a set @xmath10 $ ] defined by @xmath16\mid [ m]\setminus i\notin k\}.\ ] ] obviously , double dual @xmath17 coincides with @xmath9 . in the literature ( e.g. @xcite ) this duality is also called combinatorial alexander duality , since both @xmath9 and @xmath14 can be embedded in barycentric subdivision @xmath18})'\cong s^{m-2}$ ] as alexander dual subcomplexes ( see @xcite ) . as a consequence , @xmath19 where @xmath20 is a field or @xmath21 . for a simplicial complex @xmath9 on a set @xmath10 $ ] the following notions and notation will be used in the paper : * if @xmath22 $ ] , then @xmath23 is a _ full subcomplex _ on a set @xmath24 . its simplices are those simplices of @xmath9 which are subsets of @xmath24 . * if @xmath25 is a simplex , then its _ link _ is a complex on a set @xmath10\setminus i$ ] defined by @xmath26 \setminus i\mid i\sqcup j\in k\}$ ] . * if @xmath27 $ ] , but @xmath28 , then @xmath29 is called the _ ghost vertex _ of @xmath9 . ghost vertices do not affect the geometry of simplicial complex but they make combinatorial reasoning simpler . in particular , by definition , links usually have many ghost vertices . * a set @xmath30 $ ] is called a _ minimal nonsimplex _ ( it is also called a missing face in the literature ) if @xmath31 , but any proper subset of @xmath32 is a simplex . the set of all minimal nonsimplices of @xmath9 will be denoted by @xmath33 . * @xmath34 denotes @xmath35-dimensional skeleton of @xmath9 . * @xmath36 , where @xmath37 is a complex , which do not have nonempty simplices . one can show that @xmath38\setminus i } \mbox { and } ( k_j)^{\wedge } = \operatorname{link}_{{{\widehat{k}}}}([m]\setminus j),\ ] ] when @xmath25 and @xmath39 . note , that these conditions imply @xmath10\setminus i \notin { { \widehat{k}}}$ ] and @xmath10\setminus j\in { { \widehat{k}}}$ ] , which makes all objects well - defined . now consider a convex polytope @xmath40 , given as a convex hull of its vertices @xmath41 . suppose @xmath42 . for such a polytope construct an abstract simplicial complex @xmath43 on a set @xmath10 $ ] . its simplices are those subsets @xmath44 $ ] for which corresponding vertices @xmath45 belong to a common facet of @xmath1 . obviously , if @xmath1 is simplicial , then @xmath43 coincides with the boundary @xmath46 . in general , when @xmath1 is not simplicial , @xmath43 has more complicated structure it is not a simplicial sphere and it can be non - pure ( see fig . [ pictprismnerve ] ) . such complexes were called _ nerve - complexes _ and their properties were described in @xcite . example .. let @xmath1 be a triangular prism . the complex @xmath43 and its maximal simplices are illustrated in fig . [ pictprismnerve ] for a triangular prism . ] another way to define nerve - complexes ( which explains their name ) is the following . consider a polytope @xmath47 with facets @xmath48 . define a simplicial complex @xmath49 on a set @xmath10 $ ] as a nerve of the cover @xmath50 . in other words , @xmath51 iff facets @xmath52 intersect . then , @xmath53 , where @xmath54 is a polar dual polytope to @xmath1 . in @xcite we used @xmath49-construction rather than @xmath43-construction for toric topology reasons , but the outcome of these constructions is the same . in particular , it is proved that @xmath43 defines the face lattice of a polytope @xmath1 uniquely . so far there is no loss of combinatorial information when @xmath43 is considered instead @xmath1 . let @xmath4 be a unit sphere in euclidian space : @xmath55 . the notation @xmath56 is reserved for geometrical object ; the letter @xmath57 is used for sphere as a topological space or a homotopy type . for a point @xmath58 define a subset @xmath59 . if @xmath60 , then @xmath61 is empty . if @xmath62 , the set @xmath61 is the open hemisphere , corresponding to @xmath63 . let @xmath64 be a collection of points , @xmath65 ( repetitions are allowed , so @xmath6 is a multiset ) . such configurations also appear as _ spherical codes _ in the literature in connection with geometrical optimization problems ( e.g. @xcite ) . consider the covering @xmath66 . the nerve @xmath5 of this covering will be called the constellation complex of a configuration @xmath6 . it means that @xmath5 is a simplicial complex on a set @xmath10 $ ] , and @xmath67 iff @xmath68 . remark .. the condition @xmath69 means that there exist @xmath70 such that @xmath71 . so far this is equivalent to @xmath72 . therefore , the set of indices @xmath73 forms a simplex iff corresponding points @xmath74 lie in a common open hemisphere . this explains the terminology : it seems reasonable to call a set of stars on a celestial sphere a _ constellation _ if they can be observed from some point on earth at the same time . similar considerations and comparison also appeared in @xcite . remark .. if @xmath75 , the vertex @xmath29 is the ghost vertex of @xmath5 . example .. a few examples of constellation complexes for points on @xmath76 are represented in fig.[pict2sphere ] . in the last image we took @xmath77 points , which contain @xmath78 in their convex hull . in this case any three hemispheres intersect , but not four : @xmath79 . in the following we suppose that @xmath80 , so every point on a sphere is covered by some hemisphere . in this case @xmath5 is homotopy equivalent to @xmath81 , since the nonempty intersections in the covering are contractible ( they are given as intersections of open cones with a sphere ) . in particular , this implies @xmath82 . for a subset of labels @xmath83 we denote the ( multi)set of points @xmath84 by @xmath85 . [ claimcaratheodory ] let @xmath86 . subset @xmath32 is a nonsimplex of @xmath5 if and only if the points @xmath85 contain @xmath78 in their convex hull . the cardinality of any minimal nonsimplex does not exceed @xmath87 . the first statement follows from standard separation arguments in convex geometry . the second one is caratheodory s theorem ( see , e.g. ( * ? ? ? * sec.2.3 ) ) . example .. let @xmath88 be a configuration of @xmath89 points on a circle @xmath90 placed in vertices of a regular pentagon ( fig . [ pictcircle5 ] ) . the colored arcs on the left image show the open hemispheres corresponding to points . in this example @xmath91 is a mbius band any 3 consecutive points form a simplex . claim [ claimcaratheodory ] is illustrated by the list of minimal nonsimplices of @xmath91 . for 5 points on a circle . ] example .. similarly , for @xmath92 points on @xmath90 placed in vertices of a regular hexagon we have fig . [ pictcircle6 ] . for 6 points on a circle . ] the configuration @xmath6 is called _ nondegenerate _ if for each @xmath24 such that @xmath93 there holds @xmath94 . equivalently , @xmath6 is nondegenerate if all minimal nonsimplices of @xmath5 have cardinality exactly @xmath87 . the configuration @xmath88 on fig.[pictcircle5 ] is nondegenerate . configuration @xmath95 ( fig . [ pictcircle6 ] ) is degenerate . let @xmath9 be a complex on a set @xmath96 and @xmath97 its subcomplex . consider a new simplicial complex @xmath98 on a set @xmath99 the result of attaching a cone with apex @xmath100 to @xmath9 along @xmath101 . the operation @xmath102 will be called a _ construction step _ in the case when @xmath101 is a contractible space . any complex @xmath9 can be decomposed as @xmath103\setminus w}\cup_{\operatorname{link}w}\operatorname{cone}\operatorname{link}w$ ] for any vertex @xmath100 . the possibility to make @xmath9 from @xmath103\setminus w}$ ] by construction step means that @xmath104 is contractible . [ propconstelproperties ] let @xmath105 be a nondegenerate configuration , such that @xmath106 , and @xmath5 its constellation complex . then 1 . for each subset @xmath30 $ ] the full subcomplex @xmath107 is either a simplex or homotopy equivalent to a sphere @xmath81 . if @xmath107 is homotopy equivalent to a sphere , then so is @xmath108 for @xmath109 . @xmath5 can be obtained from @xmath110 by a sequence of construction steps . note that a full subcomplex @xmath107 coincides with @xmath111 the constellation complex of the smaller set . to prove ( 1 ) consider two possibilities : @xmath112 or @xmath113 . in the first case we actually have @xmath114 because of nondegeneracy condition . therefore , open hemispheres of @xmath85 cover @xmath4 and @xmath111 is homotopy equivalent to @xmath115 by the nerve theorem . in the second case , when @xmath113 , the set @xmath85 is covered by an open hemisphere , therefore @xmath116 , so @xmath117 . to prove ( 2 ) we proceed as follows . at first , find @xmath118 $ ] , such that @xmath119 and @xmath120 . it exists by caratheodory theorem and gives a minimal nonsimplex @xmath121 . therefore , @xmath122 . now let @xmath123 . the complex @xmath124 is obtained from @xmath111 by attaching a cone along @xmath125 . this link is given by all @xmath126 such that @xmath127 . this means @xmath125 coincides with the nerve of the covering of @xmath128 by hemispheres corresponding to @xmath129 . this nerve is contractible , since @xmath128 is contractible . therefore , @xmath124 is obtained from @xmath111 by construction step . applying this operation several times allows to build @xmath5 from @xmath130 by a sequence of construction steps . remark .. the proof shows that there is , actually , a variety of ways to build @xmath5 from @xmath110 . once the initial nonsimplex @xmath24 is installed other vertices can be added in any order . remark .. the last part of the proof works well in a more general situation . consider a covering @xmath131 of a sphere @xmath4 by `` hats '' of the form @xmath132 for real @xmath133 between @xmath78 and @xmath134 . if some smaller set @xmath135 of hats also covers @xmath4 , then the nerve of @xmath131 is built from the nerve of @xmath136 by a sequence of construction steps . an interesting question is how complicated could be the starting nerve @xmath136 , say , for fixed @xmath133 . remark .. things become more complicated for degenerate configurations . for example , consider a configuration @xmath137 from fig.[pictdegens2 ] . points @xmath138 are north and south pole respectively . points @xmath139 lie on equatorial circle and @xmath140 . in this case @xmath141 . therefore , @xmath142 . so in degenerate case may appear subcomplexes which are not homotopy equivalent to @xmath81 . ] two objects , defined in section [ secsimpconstr ] nerve - complexes and constellation complexes are strongly related . roughly speaking , they are alexander dual to each other . let @xmath143 be such a configuration of points on @xmath4 that any point @xmath62 lies in at least two open hemispheres @xmath144 . so the covering @xmath66 wraps the sphere at least twice . in this case configuration @xmath6 will be called _ good_. [ claimgalealex ] 1 . a complex @xmath9 on @xmath145 vertices is isomorphic to @xmath43 for @xmath146-dimensional polytope @xmath1 if and only if its alexander dual @xmath14 is a constellation complex @xmath5 for a good configuration @xmath147 . such @xmath6 can be constructed as affine gale diagram @xmath148 of a polytope @xmath1 . 2 . a complex @xmath9 on @xmath145 vertices is a boundary of a simplicial @xmath146-dimensional polytope if and only if its alexander dual @xmath14 is a constellation complex @xmath5 for a good nondegenerate configuration @xmath6 on a sphere @xmath149 . this follows directly from the properties of affine gale diagrams . if @xmath150 the set of points , then one can construct its gale diagram , which is a configuration of points : @xmath151 . we refer to ( * ? ? ? * sec.5.4 ) for the definition of this construction and its properties . let @xmath1 be a polytope with the set of vertices @xmath152 , so @xmath153 and suppose @xmath154 . we assert that @xmath155 . indeed , let @xmath156 . , @xmath10\setminus i\notin k(p)$ ] . equiv . , @xmath157\setminus i)$ ] is not a subset of a proper face in @xmath1 . , @xmath32 does not contain a coface of @xmath1 ( coface = complement to the set of vertices of a face ) . , by gale duality , @xmath32 does not contain a subset @xmath24 such that @xmath158 . , @xmath159 . , @xmath116 . the fact that resulting configuration @xmath6 should be good and the second statement of the claim follow from the properties of gale diagrams , listed in @xcite . example .. [ ex6prism ] by comparing figures [ pictprismnerve ] and [ pictcircle6 ] one can see that @xmath43 from the first picture is alexander dual to @xmath160 from the second . indeed , maximal simplices of a complex are the complements to minimal nonsimplices of its dual . equivalent way of saying this : @xmath95 is a combinatorial gale diagram for a triangular prism . example .. the mbius band @xmath91 from fig.[pictcircle5 ] is alexander dual to the boundary of a pentagon . configuration @xmath88 is a gale diagram of a pentagon . + + the following proposition is well known in the theory of gale duality . as before , @xmath161 the set of vertices of @xmath153 and @xmath162 is a gale diagram of @xmath163 denoted by @xmath164 or @xmath148 for short . [ proppyrzero ] 1 . if @xmath165 , then @xmath166 . @xmath1 is a pyramid with apex @xmath167 if and only if @xmath75 . more generally , @xmath1 decomposes as a join @xmath168 , if and only if @xmath148 decomposes as direct sum of @xmath169 and @xmath170 . example .. the gale diagram of a square is the multiset @xmath171 . the gale diagram of a pyramid over a square is @xmath172 . + + the following proposition is also well known ( original paper of gale @xcite or ( * ? ? ? * cor.5 ) ) . we provide a simple reformulation in terms of alexander duality . recall , that a polytope @xmath1 is called @xmath173__-neighborly _ _ if any @xmath173 of its vertices belong to a common proper face of @xmath1 . let @xmath1 be a polytope with @xmath145 vertices and @xmath6 its gale diagram . then @xmath1 is @xmath173-neighborly if and only if @xmath174 . the condition for @xmath1 to be @xmath173-neighborly is equivalent to @xmath11}^{(k-1)}\subseteq k(p)$ ] . applying alexander duality to this inclusion gives @xmath175}^{(k-1)}\right)^{\wedge}$ ] which can be rewritten as @xmath176}^{(m - k-2)}$ ] . the last statement is equivalent to @xmath174 . remark .. here we do not assume @xmath1 is simplicial . the statement holds in general . remark .. configurations @xmath6 of points on a sphere , for which @xmath174 were called _ positive @xmath177-spanning sets _ in @xcite . the notion reflects the fact that after deleting any subset @xmath178 with @xmath179 the remaining set @xmath180 contains @xmath78 in the interior of convex hull . this means vectors @xmath180 span @xmath181 with positive coefficients . + + recall that a simplicial polytope @xmath1 is called flag if any set of pairwise connected vertices is a face . in other words , @xmath1 is flag if @xmath182 for any minimal nonsimplex @xmath183 . [ lemmaflag ] if @xmath6 is the gale diagram of a simplicial polytope @xmath1 with @xmath145 vertices , then @xmath1 is flag if and only if @xmath184 for any maximal simplex @xmath116 . this follows from the fact that any maximal simplex of @xmath185 is the complement to a minimal nonsimplex of @xmath46 . suppose @xmath186 is a nondegenerate good configuration of points and @xmath187 . then there is a maximal simplex of @xmath5 which has less than @xmath188 points . suppose the contrary . then by lemma [ lemmaflag ] @xmath6 is a gale diagram for a flag simplicial polytope @xmath1 with @xmath145 vertices and @xmath189 . it is known that any flag polytope @xmath1 has at least @xmath190 vertices ( see e.g. ( * ? ? ? * lemma 2.1.14 ) ) the contradiction . another important example of using gale diagrams is the iterated simplicial wedge construction . example .. in the work @xcite an iterated wedge operation was defined . if @xmath1 is a simple @xmath146-polytope with @xmath145 facets , then a new polytope @xmath191 is defined , which has @xmath192 facets and dimension @xmath193 . the corresponding operation for simplicial complexes @xmath194 can be described combinatorially in several different ways ( see @xcite ) . the gale diagram @xmath195 can be constructed from @xmath196 by assigning multiplicities @xmath197 to points @xmath198 . example .. fig.[pictpentdoubling ] shows how gale diagram changes when a simplicial wedge construction is applied to the boundary of a simplicial polytope . when vertex @xmath134 is `` wedged '' the corresponding vertex in gale diagram doubles . let @xmath20 be a ground field and @xmath199 = { \bbbk}[v_1,\ldots , v_m]$ ] the ring of polynomials in @xmath145 variables . the ring @xmath200 $ ] is graded by @xmath201 . the field @xmath20 is given the @xmath199$]-module structure by the epimorphism @xmath199\to { \bbbk}$ ] , @xmath202 . let @xmath9 be a simplicial complex on @xmath145 vertices . the _ stanley reisner algebra _ $ ] is defined as a quotient algebra @xmath199/i_{sr}$ ] , where the stanley reisner ideal @xmath204 is generated by square - free monomials @xmath205 corresponding to nonsimplices @xmath206 . an algebra @xmath203 $ ] is graded and it has a natural @xmath199$]-module structure . let @xmath207 $ ] be a free resolution of the module @xmath203 $ ] by graded @xmath199$]-modules @xmath208 . we have @xmath209 . the tor - module of a complex @xmath9 therefore has a natural double grading : @xmath210}({\bbbk}[k],{\bbbk } ) = \bigoplus\limits_{i , j\in { \mathbb{z}}_{\geqslant } } \operatorname{tor}^{-i,2j}_{{\bbbk}[m]}({\bbbk}[k],{\bbbk}).\ ] ] the bigraded betti numbers of a complex @xmath9 are defined as the dimensions of the graded components of the tor - module : @xmath211}({\bbbk}[k],{\bbbk}).\ ] ] these numbers depend on a field @xmath20 but we will omit @xmath20 to simplify notation . the number @xmath212 is the rank of a module @xmath208 in a minimal resolution . note that @xmath213 and @xmath214 for @xmath215 . numbers @xmath216 represent a lot of combinatorial , topological and algebraical information about simplicial complex ( see e.g. @xcite ) . by hochster formula @xcite , ( * ? ? ? * th.3.2.8 ) , bigraded betti numbers can be expressed in terms of ordinary betti numbers of full subcomplexes in @xmath9 : @xmath217 , |j|=j}\dim { \tilde{h}}^{j - i-1}(k_j;{\bbbk}),\ ] ] the result of [ secconstel ] can be stated in terms of betti numbers . [ propresolution ] let @xmath105 be as in statement [ propconstelproperties ] . then @xmath218 for @xmath219 and @xmath220 . in other words , each module @xmath208 of the minimal resolution for @xmath221 $ ] is generated in degree @xmath222 for @xmath219 . this directly follows from statement [ propconstelproperties ] and formula . here is another equivalent statement : the stanley reisner ideal @xmath223 has a linear resolution as a graded module over @xmath199 $ ] . it means that all maps in the minimal resolution ( which are @xmath199$]-linear maps ) are linear in variables @xmath224 ( see @xcite for the details ) . using claim [ claimgalealex ] we can interpret statement [ propconstelproperties ] in terms of alexander duality . let @xmath9 be the boundary of a simplicial @xmath146-polytope @xmath1 with @xmath145 vertices , thus a simplicial @xmath225-sphere . then @xmath226 is a simplicial ( and even polytopal ) @xmath227-sphere with @xmath228 vertices . by @xmath229 should be a homological sphere of homological dimension @xmath230 . on the other hand , by , @xmath229 is the same as @xmath231\setminus i}$ ] . by claim [ claimgalealex ] , @xmath14 is the constellation complex for the nondegenerate configuration on a sphere @xmath149 . these considerations provide another explanation why full subcomplexes of a nondegenerate constellation complex should be homology @xmath232-spheres . the proposition [ propresolution ] can also be explained in terms of eagon reiner theorem @xcite , which states that @xmath9 is cohen macaulay if and only if stanley reisner ideal of its dual @xmath14 has a linear resolution . for a simplicial polytope @xmath1 the boundary @xmath46 is a simplicial sphere , thus a cohen macaulay complex . therefore @xmath233 has a linear resolution . moreover , there is a natural correspondence between simplices of @xmath234 and full subcomplexes of @xmath14 homotopy equivalent to @xmath235 . this correspondence sends @xmath25 to the full subcomplex @xmath231\setminus i}\simeq s^{m - d-2}$ ] . thus , full subcomplexes , which are homotopy equivalent to a sphere represent the face lattice of @xmath9 . [ propbetticonstel ] let @xmath1 be a simplicial @xmath146-polytope with @xmath145 vertices , and @xmath236 its gale diagram . then for @xmath219 we have @xmath237 , where @xmath238 is the number of @xmath239-dimensional simplices of @xmath46 . we prove more general statement . let @xmath9 be a simplicial @xmath225-sphere . then @xmath240 the number of @xmath239-dimensional simplices of @xmath9 . the proposition then follows by claim [ claimgalealex ] . for @xmath241 we have : @xmath242\setminus j))^{\wedge})=\\ = \sum\limits_{|j|=j}\dim{\tilde{h}}_{m-(m - j)-(j - i-1)-3}(\operatorname{link}_k([m]\setminus j))=\sum\limits_{|j|=j}\dim { \tilde{h}}_{i-2}(\operatorname{link}_k([m]\setminus j))=\\ = \sum\limits_{|j|=j}\dim { \tilde{h}}_{i-2}(s^{d-1-m+j}).\end{gathered}\ ] ] thus @xmath243 if @xmath244 , that is @xmath245 . on the other hand , @xmath246 equals the number of simplices in @xmath9 with @xmath247 vertices . each such simplex contributes @xmath134 in the last sum of . to extend this result to general polytopes we use some basic facts from the theory of nerve - complexes developed in @xcite . let @xmath1 be a @xmath146-dimensional polytope with @xmath145 vertices ( possibly not simplicial ) . consider the numbers @xmath248 the number of @xmath249-dimensional proper faces of @xmath1 with @xmath35 vertices . in addition , set @xmath250 this corresponds to the `` empty face '' of a polytope . obviously , in the case when @xmath1 is simplicial we have @xmath251 and @xmath252 if @xmath253 . generally , numbers @xmath254 provide much more detailed information on a polytope , than the ordinary @xmath3-vector . in the work @xcite numbers @xmath254 appear as the coefficients of the so called 2-dimensional @xmath255-polynomial of a polytope ( it was defined for the dual polytope , so the definition in that work is slightly different ) . [ propbetticonstelgen ] let @xmath1 be a @xmath146-polytope with @xmath145 vertices , and @xmath256 its gale diagram . then for @xmath219 we have @xmath257 . the calculation from proposition [ propbetticonstel ] gives @xmath258\setminus j)).\ ] ] in the case @xmath259 it gives @xmath260\setminus j)).\ ] ] for each face @xmath261 denote by @xmath262 the set of its vertices . in ( * lemma 4.7 ) we proved that @xmath263 for any face @xmath261 and for all other simplices @xmath264 the complex @xmath265 is contractible . therefore , each face @xmath255 with @xmath266 and @xmath267 contributes @xmath134 to the sum in formula iff @xmath268 and @xmath269 . therefore , @xmath270 . and @xmath160 ] example .. bigraded betti numbers of @xmath91 ( fig.[pictcircle5 ] ) and @xmath160 ( fig . [ pictcircle6 ] ) are depicted on fig . [ pictbetti56 ] . one can see , that bigraded betti numbers for @xmath88 are concentrated only in one dimension for each @xmath29 . this illustrates proposition [ propbetticonstel ] , since @xmath88 is nondegenerate . the numbers in the left table represent the @xmath3-vector of a pentagon , since @xmath88 is its gale diagram . the numbers from the right table represent @xmath254 for a triangular prism ( fig.[pictprismnerve ] ) , since @xmath95 is its gale diagram , as was discussed in example [ ex6prism ] . indeed , a triangular prism has @xmath272 @xmath272-dimensional triangular faces , @xmath273 @xmath272-dimensional quadratic faces , @xmath274 edges , @xmath92 vertices and @xmath134 empty face . we can use duality between subcomplexes and links in another direction . bigraded betti numbers of @xmath46 are described by homology of links in @xmath5 , where @xmath6 is the gale diagram of @xmath1 . [ propbettiofpolytope ] let @xmath1 be a simplicial polytope and @xmath6 its gale diagram . then for @xmath219 there holds @xmath275 the proof is similar to proposition [ propbetticonstel ] . this calculation can be used to count bigraded betti numbers for polytopes with small @xmath276 , where @xmath145 is the number of vertices and @xmath146 the dimension . -gon and the links of its simplices . ] example .. let @xmath1 be a polytope with @xmath277 . its gale diagram is up to combinatorial equivalence a set @xmath278 of points placed in vertices of a regular @xmath279-gon ( * ? ? ? * sect.6.3 ) . for simplicity assume that @xmath280 and any vertex of @xmath6 appears with multiplicity @xmath134 , so @xmath281 . the case @xmath282 is shown on fig.[pictcircle5 ] and @xmath283 on fig.[pict9gon ] . any @xmath284 consecutive points of the configuration @xmath278 lie in a common halfcircle , thus form a simplex of @xmath5 . the links of simplices @xmath285 have nontrivial reduced homology only in the following cases : 1 . @xmath32 is the maximal simplex , that is a set of @xmath284 consecutive points of @xmath6 . @xmath286 up to ghost vertices and @xmath287 . @xmath32 is the set of @xmath173 consecutive points of @xmath6 . in this case @xmath288 up to ghost vertices and @xmath289 . @xmath290 , @xmath291 and @xmath292 . one can simply prove that for other choices of @xmath32 the complex @xmath293 is contractible ( see fig . [ pict9gon ] ) . by proposition [ propbettiofpolytope ] for the corresponding simplicial polytope @xmath1 we have the expressions : @xmath294 , @xmath295 , @xmath296 and @xmath297 for all other @xmath298 with @xmath241 . the same computation can be made if nontrivial multiplicities are assigned to vertices of the gale diagram @xmath278 , but the answer is more sophisticated . the answer coincides with the result of @xcite , where bigraded betti numbers of simplicial spheres with few vertices were calculated using different approach . let @xmath9 be a simplicial complex on @xmath145 vertices and @xmath299 be the unit disk and unit circle in @xmath300 . for any simplex @xmath25 define the subset @xmath301 , @xmath302 , if @xmath303 . then the _ moment - angle complex _ of @xmath9 is a topological space @xmath304 there is a canonical coordinatewise action of a torus @xmath310 on @xmath311 , which can be restricted to the action on a subset @xmath312 . similarly , we have an action @xmath313 . in both cases the action is not free for @xmath314 . [ definbuchnumber ] _ buchstaber invariant _ @xmath315 of the complex @xmath9 is the maximal rank of torus subgroups @xmath316 which act freely on @xmath317 . real buchstaber invariant @xmath318 is the maximal rank of subgroups @xmath319 which act freely on @xmath305 . we briefly sketch here the main consideration , which allows to study buchstaber invariants . the action @xmath312 have stabilizer subgroups of the form @xmath320 the coordinate subtori , corresponding to simplices @xmath25 . the subgroup @xmath321 acts freely on @xmath317 iff it intersects stabilizers trivially . by dimensional reasons , @xmath322 . therefore , @xmath323 . similarly @xmath324 . if @xmath13}$ ] , then @xmath325 and @xmath326 , since diagonal subgroups act freely . by taking real part of a moment - angle complex one can also prove that @xmath327 . so , we have : @xmath328 the `` if '' part is simple . if @xmath1 is a pyramid with @xmath145 facets , then @xmath54 is a pyramid with @xmath145 vertices . all of these vertices except apex lie in a same facet in the base . these vertices by definition form a simplex @xmath338 , such that @xmath339 . therefore , @xmath340 and the desired relation @xmath7 follows from . now the `` only if '' part . suppose that @xmath1 is not a pyramid . then @xmath54 is not a pyramid and its gale diagram @xmath196 does not contain points at @xmath341 by proposition [ proppyrzero ] . by claim [ claimgalealex ] the complex @xmath342 is a constellation complex for the configuration @xmath343 . it does not have ghost vertices . let @xmath344 be such a vector , that @xmath345 for each point @xmath346 . then every point from @xmath196 lies either in @xmath347 or @xmath348 ( the open hemispheres introduced in [ secsimpconstr ] ) . let @xmath349 and @xmath350 be the sets of points of gale diagram sitting in @xmath347 and @xmath348 respectively . these sets are by definition the simplices of constellation complex @xmath351 , and @xmath352 $ ] . since @xmath353 , we have @xmath354\setminus i_{+}\notin k(p^*)$ ] and @xmath355\setminus i_{-}\notin k(p^*)$ ] by definition of alexander dual complex . also @xmath356 . @xmath349 and @xmath350 are nonsimplices of @xmath357 so they contain minimal nonsimplices @xmath358 and @xmath359 , @xmath360 , which , obviously , satisfy @xmath361 . now we use recent result of n.yu.erokhovets @xcite which is the following . let @xmath9 be a simplicial complex . if @xmath9 has three minimal nonsimplices or two minimal nonsimplices , which do not intersect , then @xmath362 . applying this result to @xmath357 finishes the proof . if we restrict to the class of simple polytopes , the theorem [ thmpyramids ] is known @xcite : for simple polytope @xmath1 the condition @xmath7 is equivalent to @xmath165 . obviously , a simplex is the only simple polytope , which is a pyramid . [ remmaptononsimplices ] equivalently , there exist a mapping @xmath369 from @xmath370 to the set of all nonsimplices ( not necessary minimal ) , satisfying @xmath367 for any minimal linear dependence @xmath368 . indeed , if there is such a map , we can choose a minimal nonsimplex inside each @xmath371 and these subsets satisfy the same nonintersecting condition . [ properokhreform ] let @xmath1 be a @xmath146-dimensional polytope with @xmath145 facets . then @xmath2 if and only if there exist a map @xmath375 such that @xmath376 for any odd minimal linear dependence @xmath368 in @xmath377 . recall that @xmath378 . by remark [ remmaptononsimplices ] @xmath379 is equivalent to the existence of a map @xmath369 from @xmath380 to nonsimplices of @xmath357 satisfying certain conditions . the complement to any nonsimplex of @xmath357 is a simplex of @xmath342 , therefore we have a map @xmath381 , @xmath382\setminus\xi(a)$ ] . map @xmath383 satisfies the condition @xmath384\ ] ] for any minimal linear dependence @xmath368 . for each binary vector @xmath385 consider a simplex @xmath386 . by claim [ claimgalealex ] @xmath387 , so the points of @xmath196 with labels from @xmath388 lie in an open hemisphere . so , there exist a vector @xmath389 , such that the hemisphere @xmath390 contains @xmath391 . set @xmath392 . [ corconstrbuchgiven ] consider an arbitrary map @xmath396 . let @xmath397 be a spherical configuration such that @xmath398 for any odd minimal linear dependence @xmath368 in @xmath377 . suppose @xmath399 covers @xmath400 at least twice , so @xmath401 for some @xmath402-dimensional polytope @xmath1 with @xmath145 vertices . then @xmath403 . this construction has an unexpected purely combinatorial consequence . let us call the coloring @xmath404 proper , if any odd minimal affine dependence @xmath405 is not single - colored , that is @xmath406 . \(1 ) the statement , obviously , holds for @xmath408 , so in the following let @xmath409 . suppose the contrary . let @xmath410 $ ] be a proper coloring . consider a set of points @xmath411 , representing the vertices of a regular simplex , inscribed in @xmath412 . we have @xmath413 , and @xmath414 for @xmath415 . take antipodal points @xmath416 for @xmath417 $ ] . then consider a map @xmath422 , defined by @xmath423 . this map associates to binary vector @xmath385 one of the points @xmath167 on a sphere according to the coloring . since @xmath424 is a proper coloring , for any minimal affine dependence @xmath405 the set @xmath425 contains at least two points @xmath167 and @xmath426 , @xmath415 . by ( 1 ) , the union of hemispheres @xmath427 covers the set @xmath428 . now take each point @xmath429 with multiplicity at least @xmath272 to get a configuration @xmath6 of @xmath145 points on @xmath412 . configuration @xmath6 is good and nondegenerate according to ( 2 ) ( recall , that good means that corresponding hemispheres cover @xmath412 at least twice ) . therefore , by claim [ claimgalealex ] there exist a simple polytope @xmath1 of dimension @xmath430 with @xmath145 facets , such that @xmath431 . then , by corollary [ corconstrbuchgiven ] @xmath2 . on the other hand , by estimation we have @xmath432 . this gives a contradiction . \(2 ) let @xmath433 be the basis of a space @xmath434 . consider the crosspolytope@xmath435 . it is simplicial and has @xmath436 vertices , so its gale diagram @xmath148 lies on a sphere @xmath437 . one can show , that @xmath438 the vertices of a regular simplex , inscribed in @xmath437 , each taken two times . the dual polytope @xmath54 is a cube . one can easily derive from ( * ? ? ? item 4 ) ( or from the fact that @xmath54 is delzant ) that @xmath439 . then by statement [ properokhreform ] there exist a map from @xmath440 , such that @xmath441 contain @xmath442 for each odd minimal linear dependence @xmath368 . to each @xmath385 assign a color @xmath443 $ ] if @xmath444 does not contain @xmath419 . then to each @xmath445 at least one color is assigned , since @xmath446 do not lie in a common hemisphere . if several colors are assigned to binary vector @xmath445 , choose any of them . we claim that coloring @xmath447 $ ] obtained by this procedure is proper . suppose the contrary : @xmath448 for some minimal affine dependence @xmath405 . in this case all hemispheres @xmath449 do not contain @xmath419 which contradicts the construction . recall , that fano plane is a finite projective geometry @xmath450 . its points are nonzero binary vectors in @xmath451 and lines are triples @xmath452 , @xmath453 that is exactly minimal linear dependencies . we would like to mention that the technic used to prove theorem [ thmcolorzt ] looks very similar to the proof of kneser conjecture , found by brny @xcite . though a direct connection of these subjects is not clarified yet . + + here is another unexpected fact , derived from the theory of buchstaber invariant . suppose the contrary . then , for each fano line @xmath457 holds either @xmath458 ( or the same for other pair of points ) ; or @xmath459 . in both cases @xmath460 , where @xmath461 is a finite set ( see fig . [ pictcircle3 ] ) , @xmath462 . now , let @xmath6 be an arbitrary good nondegenerate configuration of points on a circle @xmath90 . then @xmath463 for a simple polytope @xmath1 . by slightly rotating @xmath6 we can assume that @xmath6 does not intersect the finite set @xmath464 . then @xmath6 is covered by @xmath465 for any fano line @xmath466 . then , by corollary [ corconstrbuchgiven ] @xmath467 . this consideration shows that @xmath467 for any simple polytope @xmath1 with @xmath145 facets of dimension @xmath468 . this would contradict the following result of @xcite : for each @xmath469 there exist a simple polytope @xmath1 with @xmath145 facets , such that @xmath470 and @xmath471 . a. ayzenberg , _ relation between the buchstaber invariant and generalized chromatic numbers _ , far - eastern math . j.,11:2(2011),113 - 139 . ( anton ayzenberg , _ the problem of buchstaber number and its combinatorial aspects _ , arxiv:1003.0637 [ math.co ] ) m.hochster , _ cohen - macaulay rings , combinatorics , and simplicial complexes _ , in ring theory , ii ( proc . second conf.,univ . oklahoma , norman , okla . , 1975 ) , lecture notes in pure and appl . math . , v. 26 , pp.171223 , dekker , new york , 1977 .
in the paper we treat gale diagrams in a combinatorial way . the interpretation allows to describe simplicial complexes which are alexander dual to boundaries of simplicial polytopes and , more generally , to nerve - complexes of general polytopes . this technique and recent results of n.yu.erokhovets are combined to prove the following : buchstaber invariant @xmath0 of a convex polytope equals 1 if and only if @xmath1 is a pyramid . in general , we describe a procedure to construct polytopes with @xmath2 . the construction has purely combinatorial consequences . we also apply gale duality to the study of bigraded betti numbers and @xmath3-vectors of polytopes .
heat conductivity is a central problem in thermodynamics . it is well known that the classical description in irreversible thermodynamics , essentially fourier s law , has unattractive features . in particular , it predicts an instantaneous propagation of thermal signals . this is in contradiction to the expected hyperbolic nature of physical laws . in fact , the associated non - causality would be completely unacceptable within a relativistic theory . resolving this issue has been a main motivating factor behind the development of extended irreversible thermodynamics @xcite , a model which introduces additional dynamical fields in order to retain hyperbolicity and causality . this model has proved useful in various application areas , ranging from superfluid systems to heat conduction in solids . in this paper we consider the problem of heat conductivity within the flux - conservative multi - fluid framework developed by @xcite ( with the corrections discussed by @xcite ) . this model builds on the non - dissipative variational model developed by @xcite , and represents the natural non - relativistic counterpart to carter s convective variational hydrodynamics in general relativity @xcite , see @xcite for a recent review . these models have so far primarily been used to investigate the dynamics of compact stars ( see @xcite for details ) . yet , one would expect the general framework to be relevant . it is therefore of some interest to consider applications in other problem areas . this serves several purposes . first of all , it is often the case that the elegant geometric view of a relativistic analysis ( where time and space are treated `` equally '' ) simplifies the description of a complex system . the close link between the variational multi - fluid model and the relativistic counterpart may lead to insights that would be hard to reach otherwise . secondly , it is interesting to learn from the requirements of models for a range of different systems . ultimately , this will improve our understanding of the general multi - fluid framework , and the role of the various parameters in the model . in this paper we focus on the simplest `` conducting '' system , with a single species of particle together with a massless entropy component . if the entropy drifts relative to the particles , the system is heat conducting . throughout the discussion we will assume that the entropy can be treated as a ( massless ) fluid . in essence , this means that the `` phonon mean free path '' is not too large . we aim to demonstrate that our flux - conservative formulation for this system captures key aspects of extended thermodynamics . we illustrate this by writing down a model for heat conductivity in a rigid solid and comparing to various results in the literature . this exercise makes it clear that our model incorporates a finite propagation speed for heat . we also learn that the associated relaxation timescale is directly linked to the entrainment between particles and entropy . this entropy entrainment also played a central role in our recent discussion of finite temperature superfluids @xcite . in fact , the model we consider here is formally equivalent to our recent model for superfluid helium , the key element being that the massless entropy component is allowed to flow relative to the particles in the system . however , the discussion here differs in that we do not impose the irrotationality constraint associated with a superfluid system . instead , we focus on the heat conductivity . we wish to understand to what extent this , conceptually rather elegant , model captures the complex physics associated with heat flow . as an interesting by - product , the present discussion suggests how the helium model could be extended ( using a three - fluid model ) to account for the thermal conductivity associated with the interaction between rotons and phonons that dominates at higher temperatures @xcite . this is an interesting problem since the development of a causal model for this phenomenon is still outstanding . our analysis also hints at the key ingredients of a causal relativistic model for heat conductivity . the main lesson is that the entropy entrainment must be retained in order to avoid pathological behaviour . we consider a simple system with two dynamical degrees of freedom . we distinguish between the mass carrying `` atoms '' and the massless `` entropy '' . the former will be identified by constituent index @xmath0 , it has mass @xmath1 , number density @xmath2 and flows with a velocity @xmath3 , while the latter is represented by @xmath4 , with number density @xmath5 ( the entropy per unit volume ) and a flow given by @xmath6 . the canonical momentum ( density ) for each fluid is determined by the variational analysis of @xcite . this leads to _ i^= mn v_i^- 2 w_i^ , for the atoms and ^_i = 2 w_i^ , [ smom]for the entropy . this latter relation encodes the inertia of heat . we see that the heat s inertia vanishes if the entrainment between the two components , quantified by @xmath7 , vanishes . in these expressions we have used the relative velocity @xmath8 . we recall that the variational analysis takes as its starting point an energy functional @xmath9 , representing the equation of state . the coefficient @xmath7 then follows from = . e w_^2 |_n , s . as discussed by @xcite the momentum equations for the dissipative two - component system can be written f_i^= _ _ j ( v_^j _ i^+ d^j _ i ) + n _ i ( _ - 1 2 m v_^2 ) + _ j^_i v_^j , [ eulern]and f_i^= _ t _ i^+ _ j ( v_^j _ i^+ d^j _ i ) + s _ i t + _ j^_i v_^j . [ eulers]here the terms @xmath10 and @xmath11 represent the dissipation and @xmath12 and @xmath13 are the respective forces acting on the matter and entropy . the chemical potentials are determined by _ = . e n |_s , w_^2 , and _ = . e s |_n , w_^2 t , [ tdef]where we have identified the temperature , @xmath14 , in the usual way ( note that we use units such that boltzmann s constant is equal to unity , @xmath15 ) . since we have no particle creation or destruction , mass conservation leads to _ n + _ j ( n v_^j ) = 0 . [ conn]at the same time the entropy can increase , so we have _ t s + _ j ( s v_^j ) = _ , [ cons]where the second law of thermodynamics requires @xmath16 . finally , the dissipative terms are constrained by the fact that we consider a closed system . as discussed by @xcite , this means that we must have f_i^= - f_i^ , and d_ij^= d_ij - d_ij^ . here , @xmath17 is the total `` dissipation '' . it is distinguished by the fact that it is symmetric in its indices whereas @xmath18 and @xmath19 do not have to be . following the steps taken in @xcite , i.e. constructing the dissipative terms from the relevant thermodynamic fluxes and taking account of the onsager symmetry , we arrive at and @xmath20 , respectively . indices , which range from 1 to 3 , can be raised and lowered with the ( flat space ) metric @xmath21 , i.e. , @xmath22 . derivatives are expressed in terms of the covariant derivative @xmath23 which is consistent with the metric in the sense that @xmath24 . this formulation has great advantage when one wants to discuss the geometric nature of the different dissipation coefficients . we also use the volume form @xmath25 which is completely antisymmetric , and which has only one independent component ( equal to @xmath26 in the present case ) . ] - f_i^= 2 ^ w^_i + 2 ^ w^_i , [ f1]- d_ij^= ^ _ ijk w_^k + g_ij ( ^ _ + ^ _ ) + 2 ^ ^_ij + 2 ^^_ij + ^ _ ijk w_^k , [ f2]and - d_ij = g_ij ( ^ _ + _ ) + 2 ^ ^_ij + 2 ^_ij . [ f3]here we have defined the expansion _ = _ j v_^j , the trace - free shear ^_ij = 1 2 ( _ i v^_j + _ j v^_i - 2 3 g_ij _ ) , and the `` vorticity '' w_^i = 1 4 ^ijk ( _ j v^_k - _ k v^_j ) , of the entropy flow . the quantities @xmath27 , @xmath28 and @xmath29 are constructed from @xmath30 analogously . so far , the development has been formal . the model was developed by combining the non - dissipative equations of motion from the variational analysis @xcite with the general form for the dissipative terms assuming a quadratic deviation from thermodynamic equilibrium . we have accounted for the onsager symmetry between the dissipative terms , and imposed the conservation of total angular momentum . this reduces the problem to one with nine , as yet unspecified , dissipation coefficients ; @xmath31 , @xmath32 , @xmath33 , @xmath34 , @xmath35 , @xmath36 , @xmath37 , @xmath38 and @xmath39 . it should be noted that a number of , essentially , analogous formulations for multi - component systems exist in the literature , see for example @xcite . however , most of these are not arrived at by variational arguments , and hence they do not distinguish the fluxes from the canonically conjugate momenta . this means that the entrainment effect tends to be ignored . in fact , the novelty of our approach relates to the entrainment between particles and entropy . from an intuitive point of view , this encodes the inertia of heat and allows us to assign an effective mass to the entropy component . it is well known that these concepts are central to causal models of heat conductivity , and we will soon see why this is so . it is natural to begin the discussion of heat conductivity by sketching the classic approach to the problem . this provides a useful contrast to the more general model that we will develop . the following argument is more or less identical to that provided by @xcite . first we rewrite as _ t s + _ i [ s v_^i + s w_^i ] = _ . in general , @xmath40 if the system is out of equilibrium . next we introduce the heat flux vector q^i = st w_^i , [ qdef]such that _ t s + _ i = _ . keeping only the linear friction term associated with @xmath41 , and using eq . ( 2.9 ) from @xcite , i.e. t _ = - f_i^w_^i - d^j _ i _ j v_^i - d^j _ i _ j w_^i , [ tgs ] we arrive at _ = - 1 t f_i^w_^i = 2 t ^ w_^2 = 2^ ( q s t ) ^2 0 . the second law of thermodynamics thus requires that @xmath42 . in the following we will often omit the subscripts on the resistivity coefficient @xmath31 . this makes the equations somewhat clearer and we do not believe that it should cause any confusion . finally , if we consider the case of vanishing entrainment ( cf . @xcite ) then it follows from that f_i^= s _ i t = - f_i^= 2 w_i^ = - 2 st q_i , or q_i = - 1 2 s^2 t _ i t . comparing to fourier s law we identify the thermal conductivity as = 1 2 s^2 t . this completes the traditional description of the heat conductivity problem . however , for reasons that we have already discussed , this analysis is not entirely satisfactory . fortunately , the multi - fluid approach advocated here allows for a more general view . this alternative description has a richer phenomenology and recovers a number of features of extended irreversible thermodynamics @xcite . illustrating this is one of the main purposes of our discussion the analysis in the previous section focussed on the entropy conservation law . the spirit of the discussion was very much that of classic irreversible thermodynamics . we will now depart from this view , and ask what we can learn by taking the two - fluid model , e.g. the entropy inertia , at face value . the massless entropy flow and the associated entrainment then play central roles . let us consider the entropy momentum equation ( [ eulers ] ) a bit closer . we can rewrite this equation as @xmath43 for simplicity , we first include only the resistive contribution to the dissipation , i.e. , we set d^_ij = 0 , f^_i = 2 w_i^ , as before . then we have @xmath44 where we recall the definition of the entropy momentum , . it turns out to be instructive to view this equation in the context of extended irreversible thermodynamics @xcite . to facilitate the comparison , let us focus on the simplified case of heat conduction in a rigid solid . in that case we have @xmath45 and our momentum equation simplifies to _ t _ i^- _ j ( 1 2 ^j_^_i ) - _ j^_i ( 1 2 ^j _ ) + s _ i t = _ i^ , moreover , as before it makes sense to introduce the heat flux @xmath46 as w_^i = 1 st q^i = - 1 2 ^_i . [ wdef]in the following we will often choose to keep the relative velocity as the main variable ( we also omit the constituent indices , using @xmath47 for clarity ) . this simplifies many of the equations as it suppresses factors of @xmath48 . expressing all results in terms of the heat flux vector is trivial given . it follows that - _ t ( 2 w_i ) - _ j ( 2 w^j w_i ) - _ i w^2 + s _ i t + 2 w_i = 0 . [ cat1]this is one of our main results . consider the limit of vanishing entrainment . letting @xmath49 we are left with w_i = - s 2 _ i t , or q_i = - s^2 t 2 _ i t = - _ i t , where we have identified the thermal conductivity , @xmath50 , as in the previous section . in the limit of vanishing entrainment we recover fourier s law , as expected . if we instead linearise the equation ( with respect to thermal equilibrium , @xmath51 ) we find that - _ t ( 2 st q_i ) + s _ i t + 2 st q_i = 0 . [ linear]for constant parameters should be `` constant '' . we only make the assumption in order to facilitate a direct comparison with the cattaneo equation . the discussion of @xcite makes it quite clear that the parameter needs to be temperature dependent in order for this kind of model to be able to reproduce data from second sound experiments . ] , this can be written _ + q_i = - _ i t , [ cattaneo]where we have introduced the thermal relaxation time , @xmath52 , according to = - . ( [ cattaneo ] ) is known as the cattaneo equation @xcite . it resolves the `` paradox '' associated with instantaneous propagation of heat predicted in classic irreversible thermodynamics , see @xcite for a brief discussion . it also leads to the presence of a second sound in solids , an effect that has been observed in laboratory experiments on dielectric crystals , see for example the discussion by @xcite . as described by @xcite , the cattaneo equation provided key stimulus for the development of extended irreversible thermodynamics . the above exercise shows that the variational multi - fluid formalism contains a key element of extended thermodynamics . moreover , we have a natural interpretation of the entropy entrainment , @xmath7 . in terms of the relaxation time @xmath52 we have = - = - s^2 t 2 ( ) . that is , we should expect to have @xmath53 . is it surprising that the cattaneo equation can be deduced from the variational two - fluid model ? not really . in retrospect the result is , more or less , obvious . this is easy to see if we compare the nature of the variational energy functional @xmath9 @xcite to the generalised entropy used in extended irreversible thermodynamics @xcite . in the variational case , we have @xmath54 and in the case of a solid ( taking @xmath2 to be constant ) it follows that de = e s ds + e w^2 dw^2 = t ds + dw^2 = t ds + d ( q st)^2 . [ gibbs1]meanwhile , one of the tenets of extended irreversible thermodynamics is the generalised entropy s(u , q)=s_0(u)- 1 2 ( t ) q^2 , where @xmath55 is the internal energy . as discussed by @xcite , this assumption is common to a number of alternative approaches to irreversible thermodynamics . the entropy satisfies the generalised gibbs identity ds = 1 t du - dq^2 . [ gibbs2]it is now apparent that , once we identify @xmath56 , and contain similar information . the key point is that the energy / entropy of the system depends on the heat flux . this is natural since the heat flux corresponds to the flow of energy relative to the matter current . the variational approach provides a slightly different perspective on the problem , but the predicted dynamics should be equivalent . before moving on , it is worth discussing the role of the non - linear terms that were discarded in . the variational analysis naturally leads to the presence of quadratic terms in the heat flux . due to their origin , i.e. the entropy momentum equation , these terms ( essentially ) take the same form as the non - linear terms in the standard euler equation . in other words , they do not represent the most general non - linearities that one might envisage ( in the dissipative problem ) . in the interest of clarity , we have chosen not to compare the non - linear terms in our model to various attempts at constructing non - linear heat - conducting models , see for example @xcite and @xcite . such a comparison would obviously be interesting given that non - linearities are relevant for the development of both shocks and turbulence . however , our main initial aim is to establish the viability of the multi - fluids approach to the heat problem . for this purpose , the evidence provided by the linear comparison should be adequate . the model discussed in the previous section is the simplest in a hierarchy of possible heat conductivity models . by relaxing the assumptions that led to we can easily obtain more complicated models . such models are interesting , because may not provide a faithful representation of all relevant phenomena . it is obviously important to compare a more general model to analogous efforts in the literature . this comparison provides further insight into the usefulness of the multi - fluid framework and the interpretation of the various coefficients . so far we have neglected the viscous stresses in the entropy momentum equation . let us now relax this assumption , i.e. , account also for @xmath57 . in order to simplify the analysis somewhat , we will still assume that @xmath58 . that is , we ignore the coupling to vorticity . we now need d^_ij = d_ij - d^_ij = - g_ij [ ( - ^ ) _ + ( ^- ^ ) _ ] - 2 [ ( - ^ ) ^_ij + ( ^- ^ ) ^_ij ] . rewriting this expression in terms of the particle velocity @xmath3 and the relative flow @xmath30 we have d^_ij = - g_ij ( - ^ ) _ - 2 ( - ^ ) ^_ij - g_ij | _ - 2 | ^_ij , [ ds]where we have introduced | = 2 ^- ^ - , | = 2 ^- ^ - . again focussing on the case of a rigid solid , the first two terms in vanish and we also have _ = - _ i w^i , and ^_ij = 1 2 , adding the relevant terms to ( [ cat1 ] ) , taking @xmath59 and @xmath60 to be constant , for simplicity , we have @xmath61 we can simplify things further by i ) assuming that all parameters are constant , and ii ) linearising in @xmath46 ( the caveats regarding these assumptions remain as before ) . this leads to the equation - 2 st _ t q_i + s _ i t + 2 st q_i + thermal relaxation time and the thermal conductivity , _ t q_i + _ i t + q_i + l_^2 ^2 q_i + ( l_^2 + 1 3 l_^2 ) _ i _ j q^j = 0 . [ phoneq]we have also introduced the two lengthscales l_^2 = | 2 , and l_^2 = | 2 . these can be taken to represent the mean - free path associated with the dissipative terms . this result can be directly compared to the `` phonon hydrodynamics '' model developed by @xcite ( see @xcite , and @xcite for alternative descriptions ) . their model is the most celebrated attempt to account for non - local heat conduction effects . it accounts for interaction of phonons with each other and the lattice . resistive terms are represented by @xmath52 while momentum conserving interactions are associated with @xmath62 and @xmath63 . our model , , completely reproduces the @xcite result , once we set l_^2 = 1 3 l_^2 . this leads to _ t q_i + _ i t + q_i + l_^2 ( ^2 q_i + 2 _ i _ j q^j ) = 0 . the usefulness of this result is due to the fact that it can be used both in the collision dominated and the ballistic phonon regime . in the former , the resistivity dominates , the nonlocal terms can be neglected and heat propagates as waves . in the opposite regime , the momentum conserving interactions are dominant and we can neglect the thermal relaxation . in this regime , heat propagates by diffusion . the transition between these two extremes has recently been discussed by @xcite . interestingly , the non - local heat conduction model may be useful in the description of nano - size systems . if a system has characteristic size @xmath64 , and @xmath65 , then one would not necessarily expect a fluid model to apply . nevertheless , @xcite have argued that the expected behaviour of the thermal conductivity as the size of the system decreases ( as discussed by @xcite one would expect the `` effective '' conductivity to scale as @xmath66 ) can be reproduced from provided that the appropriate slip condition for @xmath46 is applied at the boundaries . a key part of this analysis is the close analogy between and the navier - stokes equation . in the latter case it is well - known that an applied non - slip condition at a surface leads to the formation of a viscous boundary layer that dominates the dissipation of the bulk flow . it appears that the heat problem is quite similar in the ballistic phonon regime , although the required slip condition is different . this is an interesting problem that requires more detailed study . at this point we have demonstrated that the multi - fluid formalism , with one fluid representing the massless entropy , reproduces a number of non - trivial results for heat conductivity in rigid solids . however , this is a simplified problem since one of the degrees of freedom in the system was `` frozen '' . in order to complete the model , we will now relax this assumption and allow @xmath67 . this leads to a system of equations with interesting applications . in particular , one could imagine modelling systems with spatial transitions to superfluidity ( as in a neutron star core ) . in one regime , the system would be dominated by resistivity , while the thermal relaxation timescale determines the dynamics elsewhere . a unified model for this problem could prove very useful indeed . let us return to the general two - fluid model and focus on the matter degree of freedom . as in the case of superfluid helium @xcite , it is natural to work with the total momentum equation . by combining and we have @xmath68 we can rewrite this using _ i^+ _ i^= v^_i , _ l^_i v_^l + _ l^_i v_^l = n _ i ( 1 2 m v_^2 ) - 2 w_l^ _ i w_^l , and v_^l _ i^+ v_^l _ i^= mnv_^l v_i^- 2w_^l w_i^ . we also use the fact that the pressure follows from @xcite _ i p = n_i _ + s _ i t - _ i w_^2 . combining these results we arrive at _ t ( v_i^ ) + _ l ( v_^l v^_i ) - _ l ( 2w_^l w_i^ ) + _ i p = -_j d^j _ i . using our definition for the heat flux , , we see that if we linearise in @xmath69 ( or equivalently , @xmath70 ) then the problem simplifies considerably . if we also use the continuity equation ( [ conn ] ) , we arrive at an equation that resembles the standard navier - stokes result @xcite ; ( _ t + v_^l _ l ) v^_i + _ i p = -_j d^j _ i . [ euler]the right - hand side is , however , different . keeping @xmath58 ( as before ) we find that -_j d^j _ i = - ( + 1 3 ) _ i ( _ j w_^j ) - ^2 w^_i + ( + 1 3 ) _ i ( _ j v_^j ) + ^2 v^_i , where we have defined = ^- , = ^- . for simplicity , we have also assumed that the equilibrium configuration is uniform . this leads to the final result @xmath71 the first two terms on the right - hand side are familiar from the navier - stokes equation @xcite , but the last two terms are new . they represent the dissipative coupling between the total momentum and the heat flux . an interesting question concerns whether there are situations where these terms have decisive impact on the dynamics . are there , for example , situations where the last term is similar in magnitude to the second term ? to complete the model , we need the momentum equation for the heat flux . starting from , linearising in the heat flux and using the results from the previous section , we find that @xmath72 \\ + { 1 \over 2 \alpha } \left[\left ( \bar{\zeta } + { 1 \over 3 } \bar{\eta } \right ) \nabla_i ( \nabla_j q^j ) + \bar{\eta } \nabla^2 q_i \right ] \ . \label{mom2}\end{gathered}\ ] ] comparing to the corresponding equation for a rigid solid , we recognize several terms . some dissipative terms are , however , new . before we conclude our discussion , it is worth making the following observation . in our formulation of the problem we have assumed that the entropy conservation law is used explicitly . an alternative , more common , strategy is to use the energy equation . the two approaches are , in principle , equivalent . nevertheless , it is useful to complement our analysis with a brief consideration of the energy equation . from @xcite we have ( for an isolated system ) the energy equation _ t + _ i q^i = 0 . after some work , using the various definitions , we find that q^i = ( 1 2 v_^2 - 2 w_^2 + n + st ) v_^i - ( st + 2v_^j w^_j ) w_^i . here , we can use the fundamental relation p + e = n+ st , to get q^i = ( 1 2 v_^2 - 2 w_^2 + p + e ) v_^i - ( st + 2v_^j w^_j ) w_^i . we also have = 1 2 v_^2 - 2w_^2 + e . combining these results , and linearising in @xmath30 , we immediately arrive at the standard energy equation ( c.f . t ( 12 v_^2 + e ) + _ i = _ i ( st w_^i ) = - _ i q^i . this confirms , at least at the linear level , the definition of the heat flux . we now have a `` complete '' model for a heat conducting fluid . it combines the equations of motion ( [ mom1 ] ) and ( [ mom2 ] ) with the two continuity equations ( [ conn ] ) and ( [ cons ] ) . this model should be relevant for dynamics on timescales such that the thermal relaxation can not be ignored . in a typical system , this would correspond to the extreme high - frequency regime . however , as we already know from the discussion of @xcite the model also applies to superfluid condensates at finite temperatures . in essence , we have a unified framework for modelling the transition to superfluidity . finally , there may be situations where the model applies even if the thermal relaxation can be safely ignored . this would be the case when the `` phonon '' mean free path exceeds the size of the system . as discussed in the previous section , the dissipative terms in then play the leading role and the crucial boundary effects may be incorporated by imposing suitable surface conditions on the heat flux . this possibility has so far been discussed only for nano - systems , but it is worth noting that it may be relevant also for large scale systems . in fact , the ballistic phonon regime should apply to cold superfluid condensates in neutron stars . to what extent the present analysis can be applied to that problem is an interesting question for the future . we have discussed heat conductivity from the point of view of the variational multi - fluid model developed by @xcite and @xcite . we have shown that a two - fluid model that distinguishes between a massive fluid component and a massless entropy can reproduce a number of key results from extended irreversible thermodynamics . in particular , we have demonstrated that the entropy entrainment , that played a central role in our recent discussion of superfluid helium @xcite , is intimately linked to the thermal relaxation time that is required to make heat propagation in solids causal . we have also considered non - local terms that arise naturally in the dissipative multi - fluid model , and related them to models of phonon hydrodynamics . this discussion may provide useful insight into the modelling of both nano - systems and superfluids at low temperatures where the phonon mean - free path is large compared to the size of the system . finally , we formulated a `` complete '' heat conducting two - fluid model and identified a number of `` new '' dissipative terms . future work needs to establish whether there are physical situations where these terms play a decisive role . what is the importance of this work ? first of all , we believe that the discussion provides strong support for the main assumptions of our model : that one can treat the entropy as an additional fluid , endowed with the inertial and dynamical properties generally associated with fluids . the entrainment between particles and the massless entropy plays a key role in this approach . the connection between this entropy entrainment and the thermal relaxation time provides an immediate interpretation , and illustrates the importance , of the main parameter of the model . the simple fact that the mathematical framework can be used to model both heat conduction and finite temperature superfluids @xcite is , in our view , clear evidence of the elegance and promise of the variational multi - fluid approach . moreover , since our model has its origin in a fully relativistic variational analysis , see @xcite for a review , the present discussion suggests a promising strategy for developing a causal relativistic model for heat conductivity . this is known to be a challenging problem where a number of issues remain to be resolved . andersson , n. , & comer , g. l. , 2006 , a flux - conservative formalism for convective and dissipative multi - fluid systems , with application to newtonian superfluid neutron stars , class . quantum grav . * 23 * 5505 carter , b. , 1989 , `` covariant theory of conductivity in ideal fluid or solid media '' , in a. anile and m. choquet - bruhat , eds . , _ relativistic fluid dynamics ( noto , 1987 ) _ , pp . 164 , heidelberg : springer - verlag
fluid dynamics ; heat conductivity we discuss heat conductivity from the point of view of a variational multi - fluid model , treating entropy as a dynamical entity . we demonstrate that a two - fluid model with a massive fluid component and a massless entropy can reproduce a number of key results from extended irreversible thermodynamics . in particular , we show that the entropy entrainment is intimately linked to the thermal relaxation time that is required to make heat propagation in solids causal . we also discuss non - local terms that arise naturally in a dissipative multi - fluid model , and relate these terms to those of phonon hydrodynamics . finally , we formulate a complete heat conducting two - component model and discuss briefly the new dissipative terms that arise .
minimal supersymmetric standard model ( mssm ) is one of the most promising candidates of the physics beyond the standard model ( sm ) that may solve the hierarchy problem in the higgs sector @xcite . the model predicts a set of superpartners ( sparticles ) whose charges are exactly the same to that of the sm partners but the spins are different by one half . these sparticles will be directly searched at the large hadron collider ( lhc ) at cern , which is a @xmath3 collider with 14 tev center of mass energy . the lhc is scheduled to start in 2007 . one of the interesting features of the mssm model is the conserved r parity , and all sparticles are assigned to have odd r parity . because of the conservation law , the lightest supersymmetric particle ( lsp ) is stable , and is assumed as a good candidate for the dark matter . at collider experiments , sparticles are produced in pairs , and a sparticle decay should produce at least one lsp . susy events at collider experiments therefore have a distinctive missing transverse momentum . if supersymmetry ( susy ) is an exact symmetry , a particle and its superpartner have a same mass . therefore susy must be broken spontaneously . the susy is considered to be broken in `` hidden sector '' , and the susy breaking will be mediated to our sector by mediation fields . the mechanism is already under severe constraints . for example , the mass of the sfermions with same charge must be universal so that they do not cause dangerous fcnc , or sfermion masses must be so heavy that correction from the susy sector is suppressed . this suggests that yet unknown symmetry / dynamics exists in the sector . the origin of the supersymmetry breaking and mediation mechanism may be understood indirectly by measuring masses of the susy particles . therefore both the discovery and mass determination of the susy particle have been studied intensively in recent years @xcite . if susy is broken at some high energy scale and the boundary conditions are universal at the scale , strongly interacting ( si ) susy particles are heavier than electroweakly interacting ( ewi ) susy particles . for example , gaugino masses follow the relation @xmath4 in the minimal supergravity model , therefore the gluino is six times heavier than the bino - like particles . at the lhc , si susy particles will be copiously produced in @xmath3 collisions , and subsequently decay into lighter ewi susy particles . the gluino and squark decays are associated with jets with high transverse momentum ( @xmath5 ) . the transverse momentum is the order of the gluino and squark masses . moreover , because the lsp is significantly lighter than the gluino , the lsp from the gluino decay also has a high @xmath5 . they would give a large missing transverse momentum to the susy events . in addition , decays of the ewi sparticles may produce high @xmath5 leptons . events from the standard model ( sm ) processes do not have such high @xmath5 particles . motivated by these observations , following cuts are often applied to reduce the sm background events to the susy signal events @xcite ; * an event is required to have at least one jet with @xmath6 gev and three jets with @xmath7 gev within @xmath8 , * the effective mass of the event must satisfy @xmath9 gev , where the effective mass is defined using the transverse missing energy ( @xmath2 ) and the transverse momentum of four leading jets as : @xmath10 if the event has hard isolated leptons , the effective mass may be defined as follows : @xmath11 where sum of the lepton @xmath5 can be taken over the leptons with @xmath12 gev and @xmath13 gev . * the missing transverse energy must satisfy the relation : @xmath14 * the transverse sphericity @xmath15 must be greater than 0.2 , where @xmath15 is defined as @xmath16 , with @xmath17 and @xmath18 being the eigenvalues of the @xmath19 sphericity tensor @xmath20 formed by summing over the transverse momentum of all calorimeter cells . to reduce the background further , hard , isolated lepton(s ) may be required . these cuts are good enough to reduce the sm backgrounds from @xmath21+@xmath22-jets and @xmath23+@xmath22-jets productions down to a manageable level , although the production cross section of the sm processes may be @xmath24 higher than signal cross sections . while the susy production section reduces very quickly as sparticle masses increase beyond 1 tev , but the signature peaks at higher @xmath1 where backgrounds can be ignored . previous studies show that the squark and gluino with mass around 2.5 tev can be found at the lhc in the minimal super gravity model ( msugra ) . in msugra , the sm background after the cuts can be neglected safely . then , the distribution of accepted events are also useful to determine the mass scale of susy particles . for example , the peak of the @xmath1 distribution is sensitive to the squark and gluino masses . for the events with same flavor opposite sign dileptons , the invariant mass distributions , @xmath25 , @xmath26 , and @xmath27 , are useful to reconstruct the susy particle masses @xmath28 , @xmath29 , @xmath30 and @xmath31 . recently it is pointed out that a string inspired model based on the flux compactification ( kklt models ) @xcite predicts a mass relation different from that of the msugra @xcite . this is called mixed modulus anomaly mediation ( mmam ) model . this model has a volume modulas @xmath32 and a compensator field of minimum supergravity model @xmath33 as a messenger of the susy breaking . the susy mass spectrum depends on the ratio of the two susy breaking parameters @xmath34 and @xmath35 . the unification scale of the sparticle masses depends on the ratio . it is interesting that the unification scale of the soft susy parameters can be much lower than the gut scale in this model . they may unify even at the weak scale for a special choice of the model parameters . when sparticle masses are degenerated at the weak scale , we expect a reduced probability to have high @xmath5 jets , and smaller @xmath1 and @xmath2 for given squark and gluino masses . this means that the standard susy cuts reduce the signal events as well , and susy discovery is more affected by the sm background . quantitative understanding of the sm background may be required in this case . existing background estimates have large theoretical uncertainty coming from the scale of the strong coupling . recently , several groups has emphasized importance to include the matrix element correction @xcite in the previous parton shower estimate of the background , which significantly change the background distribution in the signal region . when overlap between the signal and background distribution is large , as it is expected in the degenerated susy spectrum , the uncertainty of the background must be taken seriously , and we also need to reconsider the cuts to reduce the background . the purpose of this paper is to illustrate the phenomenology of the degenerated supersymmetry . we take the mmam model as an example . by changing the ratio @xmath36 , the mass spectrum changes smoothly from msugra - like one to anomaly - mediation - like one . in between , there are regions of parameters where squark , slepton , and gaugino masses are significantly degenerated compared with those expected in msugra . the model therefore provides one dimensional parametrization from the `` transverse signature '' to its non - transverse limit . the investigation of our analysis may easily be extended to other susy / non - susy scenarios with @xmath2 signature , such as the universal extra dimension model or little higgs model with @xmath32 parity . this paper is organized as follows . in section 2 we describe the mmam model and its mass spectrum with an emphasis on the region of parameter space where sparticle masses are degenerated . in section 3 , we describe our monte carlo simulations . in section 4 , we study how susy event distribution depends on sparticle mass degeneracy and compare it with the background . we find that the signal @xmath1 distribution is quite similar to that of the background if @xmath37 . however we find a certain relation among @xmath2 and @xmath1 that is universal for the model with heavy colored particle that decays into the stable neutral particle lhc will miss the susy signature if @xmath38 . section 5 is devoted to discussions and conclusions . in this section , we briefly describe the mmam model following the notation in @xcite . in this model , all shape modulus and dilaton will be fixed by non - zero flux on a cy manifold in the type iib string theory . the low energy @xmath39 lagrangian of this model is given by unfixed volume modulus @xmath32 , compensator field @xmath33 , and gauge and matter fields @xmath40 and @xmath41 as @xmath42 . \end{aligned}\ ] ] here @xmath43 , @xmath44 is the 4d metric in the super conformal frame . @xmath45 and @xmath46 are the volume modulas and the chiral compensator superfield of @xmath47 sugra , respectively . @xmath48 is the superpotential for @xmath32 and matter . @xmath32 dependent function @xmath49 and @xmath50 may be expressed as @xmath51 where @xmath52 is the modular weights and @xmath53 for matter fields located on d7 ( d3 ) branes , and @xmath54 for matter living at brane intersections @xcite . in kklt model , @xmath55 , where the last term of @xmath56 expresses the non - perturbative effect such as the gaugino condensation in d7 brane , which fix the volume modulus , @xmath57 is the contribution of the flux . in addition to the @xmath39 supersymmetric action , there are contributions from anti - d3 branes which break supersymmetry and uplift the potential from ads vacuum to ( nearly minkowski ) de sitter vacuum . the term is expressed by a spurion operator depending on @xmath32 and @xmath33 , and minimum of the potential will be obtained by solving the effective @xmath39 action and the lifting potential . the resulting theory is parametrized by @xmath58 and @xmath59 . the susy breaking terms are obtained by expanding the action by @xmath34 and @xmath35 . here we define the soft terms as @xmath60 where @xmath61 is a canonically normalized yukawa coupling @xmath62 they are explicitly written as functions of @xmath63 and @xmath64 as follows ; @xmath65 where @xmath66 parameter is such that the off - diagonal element of @xmath67 mass matrix is @xmath68 . ] and @xmath69 with @xmath70 , namely for matter in the fundamental representation @xmath71 , @xmath72 , @xmath73 . we only include the effect of the top yukawa coupling @xmath74 , therefore @xmath75 , @xmath76 , @xmath77 and @xmath78 otherwise . the scale dependence of @xmath79 is expressed as @xmath80 where @xmath81 and @xmath82 here , @xmath83 , @xmath84 , @xmath85 , @xmath86 , and @xmath87 , @xmath88 , @xmath89 , @xmath90 . finally , @xmath91 in eq . ( [ eq : soft ] ) , the highest term in @xmath92 is the contribution of modulas @xmath32 , while the terms independent from @xmath92 is that of pure anomaly mediation . in @xcite , it is pointed out that the mass spectrum in the mmam model shows a special feature if @xmath93 , where @xmath94 this can be seen by investigating the low energy mass parameters as functions of @xmath95 . for example , gaugino masses at the scale @xmath96 may be expressed as @xmath97\ ] ] in one loop level . when @xmath98 , the @xmath99 term in the equation becomes 0 at @xmath100 . namely , if @xmath101 does not depend on the gauge group , the gaugino masses unify at the weak scale , rather than at the gut scale . the mass spectrum in the matter sector depends on @xmath52 . if yukawa couplings @xmath61 are non - vanishing only for the combination @xmath102 satisfying @xmath103 , or if the effect of the yukawa couplings in the rge can be ignored , the scalar masses and trilinear couplings also unify at the same scale of the gaugino mass unification . this relation is satisfied for the choice @xmath104 and @xmath105 and we call this choice of the boundary condition as model a. even if this condition is not satisfied , squark and slepton masses tend to be close with each other at @xmath106 when gaugino masses unify at low energy scale , because the gaugino loop corrections to the sfermion masses are roughly equal . in table [ table : alphar ] , we show the relation between @xmath107 and @xmath95 for a fixed gluino mass at the gut scale of 450 gev . this corresponds to @xmath108 tev . we can see that the unification of gaugino or squark and slepton masses occurs at @xmath109 . @xmath107 @xmath110 & 0.45 & 0.1 0.30 & 0.50 & 10 0.61 & 0.56 & 20 0.92 & 0.63 & 30 1.25 & 0.73 & 40 1.58 & 0.86 & 50 1.92 & 1.06 & 60 for the following discussion , we calculate the low energy mass spectrum using isajet @xcite version 7.72 which solves the boundary condition given in section 2.1 . the sparticle mass spectrum for the model a is listed in table [ table : massa ] . as @xmath95 increases , gaugino masses , and slepton and squark masses get closer , and the model shows the mass pattern different from that of msugra . the numerical values roughly agree with those in @xcite . @xmath111 & @xmath96 & @xmath112 & @xmath113 & @xmath114 & @xmath115&@xmath95 0.1 & 1055 & 184 & 350 & 700 & 957 & 435 & 354 & 450 & @xmath110 30&1045 & 436 & 536 & 607 & 913 & 531 & 476 & 631 & 0.92 40&1038 & 573 & 653 & 545 & 879 & 578 & 541 & 729 & 1.24 45&1034 & 657 & 717 & 499 & 852 & 604 & 576 & 790 & 1.41 55&1020 & 882 & 892 & 339 & 765 & 671 & 675 & 951 & 1.74 as can be seen in table [ table : massa ] , the higgsino mass parameter @xmath96 decreases as @xmath95 increases . the @xmath96 is determined by solving the minimization condition of one loop higgs effective potential . in minimal supergravity model ( @xmath116 limit ) , the higgs soft mass square at weak scale is driven to large negative value by stop / top loop , while the correction from gaugino loops is small because @xmath117 . the @xmath96 parameter is chosen to compensate the negative value to get correct gauge symmetry breaking , therefore @xmath96 can be as large as the stop mass in the minimal supergravity model . when @xmath95 is large , @xmath118 at the gut scale . the higgs soft mass parameters get large positive contribution from the gaugino masses . top / stop loop effect is largely compensated by the gaugino corrections at the weak scale . as a result , @xmath119 for @xmath120 , and @xmath121 for larger value of @xmath107 . although @xmath122 and @xmath111 get closer to @xmath123 for @xmath124 , the mass splitting among susy particles increases again . we have seen in the model a that decrease of the @xmath96 parameter limits the mass degeneracy . in our study , we are interested in the model point where the mass splitting among the susy particles are the smallest . the @xmath96 parameter would be largest when the gut scale value of the higgs soft susy breaking mass is the smallest . on the other hand , sfermion masses at gut scale must be large to avoid the @xmath125 or @xmath126 lsp . we therefore also consider the model b , where @xmath127 and @xmath128 as another example of the mmam model . the mass spectrums of the models a and b for @xmath129 gev and @xmath130 are compared in table [ table : massab ] . the mass difference between squark ( gluino ) and lsp hits minimum when @xmath119 at @xmath131 . @xmath132 for the model a and 55 for the model b , where @xmath133 and @xmath134 , respectively . as expected , the model set b has more degenerated mass spectrum at @xmath135 , because the @xmath96 parameter is larger for the same gaugino masses . @xmath136 & @xmath137 & @xmath136 & @xmath138 0 & 995 ( 1055 ) 182 & 961 & 1041 ( 1061 ) 189 & 1007 10 & 986 ( 1053 ) 246 & 924 & 1043 ( 1061 ) 248 & 98420 & 973 ( 1049 ) 326 & 793 & 1044 ( 1060 ) 330&94030 & 951 ( 1045 ) 426 & 726 & 1045 ( 1060 ) 434 & 865 40 & 916 ( 1038 ) 507 & 635 & 1044 ( 1059 ) 569 & 733 50 & 854 ( 1027 ) 426 & 641 & 1038 ( 1057 ) 713 & 548 55 & 803 ( 1021 ) 335 & 663 & 1033 ( 1056 ) 721 & 529 60 & no ewsb & & 1023 ( 1055 ) 700 & 543 in the following sections , we discuss the discovery potential in the mass degenerated susy models . phenomenologically important parameter for the hadron collider is the typical mass scale of the event . this can be expressed by the energy of the jet from @xmath139 decay in the @xmath140 rest frame , which is expressed as @xmath141 because sparticles are pair produced , @xmath138 is the typical order of the effective transverse mass @xmath1 of the susy production process . we list the value of @xmath138 for the model points in table [ table : massab ] . @xmath142 can be reduced by a factor of @xmath143 to @xmath144 for these models . in the following sections , we find that discovery of susy particles will be non - trivial in this region . it is worth noting that gaugino and higgsino are highly mixed when @xmath145 is minimum . in this case , relatively large nucleon - lsp scattering cross section and smaller dark matter density are expected , possibly consistent with cosmological constraints . another phenomenologically important aspect in terms of collider physics is the branching ratios of the gluino and squark into leptons . the leptons would be produced from the neutralino or chargino decays arising from the squark decays . for the model a , the decay @xmath146 @xmath147 channels are open unless @xmath148 . the chargino and neutralino may decay into ( s)leptons . especially , if the decay channel @xmath149 is open , the large branching ratio into the golden mode @xmath150 is expected . for example , @xmath151 80% and @xmath15214% for the model a , @xmath153 gev , @xmath130 , and @xmath154 . for the model a , @xmath155 if @xmath156 , namely , the decay channel is open especially in the degenerated region . for the model b , the squark and slepton masses are so heavy that the decay into slepton is always closed . here we describe our event simulation method used in the next section . as explained earlier , susy mass spectrum is calculated by isajet @xcite which is interfaced to the herwig @xcite event generator using isawig @xcite . herwig generates hard processes , takes care of initial and final state radiations , and fragments partons into hadrons . to estimate event distributions to be measured at lhc detectors , we smear particle energies , identify isolated leptons , and reconstruct jets . we independently developed a fast detector simulation program , which takes the following steps ; 1 . * finding isolated leptons : * if a lepton ( @xmath157 or @xmath96 ) with @xmath158 gev and @xmath159 is found in an event record , we take a cone with a size @xmath160 around the lepton . if the sum of @xmath161 of the particles in the cone ( except the lepton at the center of the cone ) is less than 10 gev , we regard it as an isolated lepton candidate . after the jet reconstruction described below , isolated lepton candidates with @xmath162 are accepted as isolated leptons . * reconstruction of jets*. we adopt a simple jet finding algorithm pycell , a subroutine in the pythia package , with minor modification on its treatment of leptons and energy smearing . namely , 1 . we remove isolated lepton candidates for the jet reconstruction . particle energies are measured by the electromagnetic and hadronic calorimeters . we assume that the number of calorimeter cells in @xmath163 direction is 50 within @xmath164 and 50 in @xmath165 direction , respectively . transverse energy deposit @xmath161 in a cell is summed for electrons , photons , and hadrons as @xmath166 which correspond to the hits in the electromagnetic and hadronic calorimeter cells , respectively . transverse energy deposit in each cell is smeared by gaussian energy resolutions @xmath167 } $ ] and @xmath168}$ ] . after the energy smearing , we regard sum of the smeared @xmath169 and @xmath170 as the measured energy deposit in the cell . we take the highest energy cell as the initiator and sum the energy deposits in the cell within @xmath171 . we accept the cluster as a jet if @xmath172 gev . we repeat this procedure after removing the cells which have been already used for the jet reconstruction . the reconstruction process ( 1 and 2 ) is similar to the fast simulation codes used by the atlas and cms groups . note that we do not smear energy of isolated leptons and assume they are identified correctly , because the energy resolution and particle identification are excellent for leptons in both of the lhc detectors . we smear photon and non - isolated electrons energies with better resolution than that for hadrons , however we do not assume fine grained electromagnetic calorimeter in our simulation for simplicity . our simulation does not take care of other detector effects , such as misidentification of leptons and hadrons , non - uniformity of detector responses ( cracks ) , non - gaussian smearing in the energy measurements . the simulation , however , reproduces the signal distributions of past simulation studies reasonably well . to check the validity of our independently developed simulation program , we compare published distributions of susy signals using the atlfast simulator @xcite with our simulation results . figure [ fig : msps1a ] show distributions of the same flavor and opposite sign dilepton invariant mass ( @xmath25 ) and the jet - lepton - antilepton invariant mass ( @xmath173 ) at sps1a @xcite , where @xmath174 is one of the two highest @xmath5 jets with @xmath175 . the distributions are important to reconstruct the cascade decay @xmath176 . contribution of background events with accidental two leptons can be subtracted using the distributions of @xmath177 events . the plots are produced by applying the same cuts as used in @xcite ; * @xmath178 gev . * at least 4 jets with @xmath179 gev , @xmath180 gev , and @xmath181 gev . * two and only two leptons with @xmath182 gev and @xmath183 gev . * @xmath184 * @xmath185 . we calculate the sphericity and @xmath2 , based on the smeared energy deposits to the cells in @xmath186 in addition to the momentums of isolated lepton candidates . we find that accepted number of events is consistent with the previous simulation studies . in figure [ fig : msps1a ] we show the @xmath25 and @xmath27 distribution for @xmath187 susy events ( corresponding to @xmath188 1.8 fb@xmath189 ) . the edge and the end point appear at correct positions . the @xmath25 distribution has a sharper edge compared with that in @xcite , because we do not smear lepton energies in our simulation . we find that 445 events remain after the cuts and the background subtraction . in the report @xcite roughly @xmath190 events remain after the cuts for @xmath191 100 fb@xmath189 . the acceptances of the two simulations therefore agree within 1@xmath192 . ( left ) and @xmath27(right ) distribution at the point sps1a by our fast simulation program . we produce @xmath187 susy events for the plot , corresponding to @xmath193 fb@xmath189 . after the cuts and background subtractions , 445 dilepton events remain . , title="fig:",width=226 ] ( left ) and @xmath27(right ) distribution at the point sps1a by our fast simulation program . we produce @xmath187 susy events for the plot , corresponding to @xmath193 fb@xmath189 . after the cuts and background subtractions , 445 dilepton events remain . , title="fig:",width=226 ] in this sub - section , we study the @xmath1 distributions at the susy model points with degenerated mass spectrum . we first show the @xmath5 distribution of the first jet in the left panel of figure [ fig : ptmeffb ] . here we take the model b , @xmath153 gev , @xmath130 and vary degeneracy by changing @xmath107 from 0 to 55 . as @xmath107 increases , sparticle mass difference gets smaller , resulting in softer @xmath5 distribution . the peak positions are positively correlated with @xmath145 which ranges from @xmath194 gev to @xmath195 gev when we change @xmath107 from @xmath196 to @xmath197 . from the figure , we can read that the acceptance of susy events depends on @xmath107 . for example , events with @xmath198 gev is only 16% for the point with @xmath116 but it is 35% for the point with @xmath199 . this means that about one third of the events are rejected by the requirement on the @xmath5 of the first jet in the previous section for the latter point . distributions of the highest @xmath5 jet for model b , @xmath153 gev and @xmath130 . 50000 events are generated for each points(@xmath200 from right to the left ) . right ) @xmath1 distributions for @xmath201 for model b with @xmath153 gev @xmath202 ( solid histograms from right to left ) and @xmath203 gev @xmath204(dashed histograms from right to left ) respectively . , title="fig:",width=226 ] distributions of the highest @xmath5 jet for model b , @xmath153 gev and @xmath130 . 50000 events are generated for each points(@xmath200 from right to the left ) . right ) @xmath1 distributions for @xmath201 for model b with @xmath153 gev @xmath202 ( solid histograms from right to left ) and @xmath203 gev @xmath204(dashed histograms from right to left ) respectively . , title="fig:",width=226 ] the @xmath1 distribution shows a similar behavior . in figure [ fig : ptmeffb ] ( right ) , we show the @xmath1 distributions for the model b , with @xmath153 gev ( solid histograms ) and @xmath203 gev ( dashed histograms ) , respectively . we apply the cut ( a ) and also require one hard isolated lepton with @xmath12 gev and @xmath205 . here we compare the distributions with @xmath206 ( msugra like ) and @xmath207 to @xmath197 ( degenerated ) . it should be noted that , while the mass spectrums are considerably different , the power low of the distributions beyond their peaks are roughly the same . these high @xmath5 events are originated from the collisions with its center of mass energy @xmath208 much higher than the squark and gluino masses . therefore the power low of the distributions only depends on the dominant luminosity function . the peak position of the @xmath1 distribution has a direct correlation with the produced sparticle mass . in @xcite it is found that the correlation between the susy mass scale defined as @xmath209 where @xmath210 is the lsp mass and @xmath211 . the peak position of @xmath1 is linear for msugra and gauge mediation model . we also find that the linear relation holds for the signal distribution . we do not provide the fit results here , because the existence of the standard model background is very important for this model , as will be discussed in the following subsections . signals . ] we now turn into the relation between @xmath2 and @xmath1 for degenerated points . in figure [ fig : etmissmeffb ] we show the distribution for the model b with @xmath212 gev , @xmath130 and @xmath213 , respectively . the @xmath214- and @xmath74- axes are @xmath1 and @xmath2 , respectively , in figure [ fig : etmissmeffb ] . we see that the @xmath2 takes a significant part of @xmath1 for degenerated points @xmath215 , and @xmath216 if @xmath1 is higher , while , for @xmath206 the @xmath2 for a fixed @xmath1 is broader in the plot . t + this distribution can be understood as follows . suppose we have events with two uncorrelated jets with energy @xmath217 . the dominant part of the cross section is squark / gluino pair production near the threshold for our case , so if they decay directly into two jets and two lsps , the event kinematics are indeed of this type up to the boost to the beam direction . we now calculate missing energy @xmath218 and effective energy @xmath219 of the events . the momentum of the two jets and the missing momentum can be expressed as follows , @xmath220 where we take the direction of the @xmath221 momentum as @xmath222-axis and events are in @xmath214-@xmath222 plane . then the missing energy @xmath218 and effective mass @xmath219 are defined as @xmath223 where @xmath219 ranges from @xmath138 to @xmath224 . the relation between @xmath2 and @xmath1 is somewhat similar to that of @xmath218 and @xmath219 as can be seen in figure [ fig : etmissmeffb ] . for each plot @xmath2 tends to be small when @xmath225 , where @xmath226 tev for @xmath213 , respectively . it increases linearly as a function of @xmath1 up to @xmath227 , and the number of events reduces quickly beyond @xmath228 . the relation indeed should be true when two particle are pair produced and each of them decays into a visible particle and an invisible particle at hadron collider . on the other hand , for the dominant background coming from @xmath23 + @xmath22-jets or @xmath21 + @xmath22-jets , the @xmath229 should be smaller than those of the signal . this is because the sm events have many other jet activities which are not associated with the missing momentum productions . the background comes from highly boosted events , where @xmath2 is typically a fraction of the missing energy originated from high energy @xmath230 , @xmath231 , and @xmath232 . the canonical cut to reduce the sm background is @xmath233 . if the background distribution in @xmath1 and @xmath2 plane is significantly different from the signal distribution , one can improve the s / n by applying an improved cut in the plane rather than the simple cuts @xmath9 gev and @xmath234 described in section 1 . because there are no background distribution study on the relation between @xmath1 and @xmath2 , we do not attempt to improve the cuts in following sections and use the cut given in section 1 . we now show the signal @xmath1 distributions for the degenerated model . we start our discussion from the @xmath235 distributions of the susy events without high @xmath5 leptons . this channel has been studied in @xcite with an emphasis on the model - independent reconstruction of the susy scale , as discussed in the previous subsection . distributions for non - lepton sample . the thick solid histogram is the background estimated in @xcite . solid and dashed histograms are signal distributions for @xmath153 gev @xmath206 ( msugra ) , and @xmath154 ( degenerate ) in the left panel , and for @xmath203 gev @xmath236 ( msugra ) and @xmath237 ( degenerate ) in the right panel , respectively . all plots are normalized for @xmath238 fb@xmath189 . , title="fig:",width=226 ] distributions for non - lepton sample . the thick solid histogram is the background estimated in @xcite . solid and dashed histograms are signal distributions for @xmath153 gev @xmath206 ( msugra ) , and @xmath154 ( degenerate ) in the left panel , and for @xmath203 gev @xmath236 ( msugra ) and @xmath237 ( degenerate ) in the right panel , respectively . all plots are normalized for @xmath238 fb@xmath189 . , title="fig:",width=226 ] in figure [ fig : meffa ] , the signal distributions for the model a , with @xmath153 gev and 650 gev are plotted for @xmath239 and @xmath240 . the signal distributions ( solid and dashed histograms ) are parallel to the background distribution ( bars in the plots ) beyond the peak of @xmath1 distribution . the difference appears only in the overall normalizations . as already discussed , this is because the signal distribution is determined mainly by the luminosity function , once @xmath241 . the background shown in the plots has a large uncertainty . the estimation has been made by the lowest order monte carlo generator isajet . the distribution is subjet to the scale uncertainty which is typically @xmath242(30% ) . we also require at least 4 high @xmath5 jets , while the number of final state parton involved in the tree level diagram is much less than 4 . the number of additional jets is estimated by parton shower ( ps ) approximation . the ps approximation is good to estimate distributions of jets collinear to the leading jets . recently , several groups have emphasized the importance of the matrix element ( me ) correction to the sm background processes . the matrix elements of the diagram involving @xmath243 + @xmath22-partons are calculated by the generator devoted for the process with multi jets @xcite . experimental groups are also working to take into account the corrections , for example , in the atlas group @xcite , the backgrounds from @xmath244+@xmath22-jets are calculated up to @xmath245 respectively using alpgen @xcite . after the inclusion of the me corrections , the sm background is increased by a factor of 2@xmath2464 for @xmath247 tev . even in the highest bin in the plot , the expected background is above 1 . this is caused by the increased high @xmath5 jets in the sm events . the improved background estimation is still preliminary and in the level of the lowest order , and subject of coupling scale uncertainty . we find that the signal rate is the same order of the background rate for the degenerated parameters . if the uncertainty of the overall normalization of the background is 100% , no bound would be obtained from no - lepton channel by looking into @xmath235 distribution only . however , as we noted previously , the structure of the signal distribution may be visible over the background distribution in the @xmath235 and @xmath2 plane . the significant fraction of susy production gives rise to events with isolated leptons ( @xmath248 gev and @xmath13 ) . we show the signal distribution for the model a in figure [ fig : meffla ] . here , we do not show the background distributions , because they are not given in previous literatures . in @xcite , the distribution of the sm background is given , which is roughly 1/20 of the background of the zero lepton channel with a similar shape . distributions of one lepton events . the solid and dashed histograms correspond to model a , @xmath239 ( msugra - like ) and @xmath240 ( degenerated mass spectrum ) respectively . @xmath249 , 750 , 850 gev from left to right . all plots are normalized for @xmath250fb@xmath189 . , title="fig:",width=170 ] distributions of one lepton events . the solid and dashed histograms correspond to model a , @xmath239 ( msugra - like ) and @xmath240 ( degenerated mass spectrum ) respectively . @xmath249 , 750 , 850 gev from left to right . all plots are normalized for @xmath250fb@xmath189 . , title="fig:",width=170 ] distributions of one lepton events . the solid and dashed histograms correspond to model a , @xmath239 ( msugra - like ) and @xmath240 ( degenerated mass spectrum ) respectively . @xmath249 , 750 , 850 gev from left to right . all plots are normalized for @xmath250fb@xmath189 . , title="fig:",width=170 ] distributions for one lepton signal events for model b , @xmath153 gev ( left panel ) , @xmath203 gev ( right panel ) and @xmath130 . th solid histograms for @xmath206 ( msugra like ) and dashed histograms for @xmath199 and @xmath216 respectively corresponding to degenerated mass spectrum . all plots are normalized for @xmath251fb@xmath189 . , title="fig:",width=226 ] distributions for one lepton signal events for model b , @xmath153 gev ( left panel ) , @xmath203 gev ( right panel ) and @xmath130 . th solid histograms for @xmath206 ( msugra like ) and dashed histograms for @xmath199 and @xmath216 respectively corresponding to degenerated mass spectrum . all plots are normalized for @xmath251fb@xmath189 . , title="fig:",width=226 ] and @xmath27 distribution of signal for a degenerated susy point . @xmath252 susy events are generated for model a , @xmath123(gut)@xmath253 gev , @xmath130 and @xmath154 . , title="fig:",width=226 ] and @xmath27 distribution of signal for a degenerated susy point . @xmath252 susy events are generated for model a , @xmath123(gut)@xmath253 gev , @xmath130 and @xmath154 . , title="fig:",width=226 ] the histograms in figure [ fig : meffla ] show signal distributions for @xmath203 gev , @xmath254 gev , @xmath255 gev from left to right , and @xmath239 for solid line and @xmath256 for dashed lines , respectively . as can be seen in table [ table : massab ] , the minimum mass difference is smaller for the model b than that for the model a , because the higgsino mass @xmath96 is higher . because of the degeneracy , the signal is more suppressed as can be seen in figure [ fig : mefflb ] where we take @xmath153 gev ( left ) and @xmath257 gev ( right ) , @xmath199 and @xmath216 ( dashed histograms ) , respectively . overall normalization of the signal above @xmath258 tev reduces by a factor of 1/10 for for model b @xmath203 gev ( figure [ fig : mefflb ] , right panel ) compared to the model a ( figure [ fig : meffla ] , left panel ) for most degenerate points . we conclude that one lepton signal is not accessible at the most degenerated points by looking into @xmath259 distribution alone , assuming that sm background in the one lepton channel in @xcite , and a factor 2@xmath246 3 uncertainty , even though the susy scale is relatively small for this point , @xmath260 tev the mass reach at the lhc will be revisited again in section 4.5 . for the model a , there are significant two lepton events in the @xmath261 region because @xmath262 is lighter than the @xmath263-like neutralino . we show the @xmath264 and @xmath27 distribution for the model a , with @xmath129 gev and @xmath154 in figure [ fig : mlla ] . here we require that one of the two isolated leptons has @xmath12 gev and the other has @xmath265 gev . we also apply the standard cuts described in section i. the relevant mass spectrum for the most prominent endpoints are @xmath266 gev , @xmath267 gev , @xmath268 gev , @xmath269 gev , and @xmath270 gev . the predicted @xmath25 edge and @xmath27 endpoint for the decay chain @xmath271 is @xmath272 gev and @xmath273 gev , respectively , consistent with figure [ fig : mlla ] . the edges and endpoints of other cascade decays are also visible in the plots . to select the jet from @xmath140 decays , we have taken the one of two most highest @xmath5 jets with smaller @xmath27 . if mass difference between @xmath274 and @xmath28 is smaller , we may found the smaller probability to find correct jets that arise from @xmath275 decay . we do not observe such effect in figure [ fig : mlla ] . the number of accepted events @xmath276 ( where @xmath277 denotes number of isolated leptons with @xmath12 gev and @xmath278 ) for this point are @xmath279 , @xmath280 , @xmath281 for @xmath252 generated events ( corresponding to @xmath246 20 fb@xmath189 ) . number of two lepton events below the dominant edge is 2933 after the background subtraction . because @xmath1 distribution of the one lepton channel is not very prominent over the background , susy particles may be discovered in the two lepton channel for the degenerate region . note that for the most degenerated susy spectrum , we have @xmath282 so that all neutralinos are highly mixed . in such situation , dominant source of @xmath283 is @xmath284 , because @xmath285 has significant decay branching ratios into charginos . it is worth to point out that the relative weight of the decays @xmath286 may be studied by looking into the charge asymmetry in @xmath26 distributions , @xmath287 @xmath288 @xcite . the charge asymmetry comes from the polarization of @xmath283 arising from @xmath140 , however as discussed in @xcite , the effect tends to be suppressed if sparticle masses are degenerated . for the model b , the decay of @xmath289 into sleptons are closed for any value of @xmath107 . for @xmath153 gev and @xmath215 , we have @xmath290 , @xmath291 , @xmath292 for @xmath252 generated events . while @xmath293 and @xmath294 is about same as that of the model a , @xmath295 is reduced by a factor of 4 . in addition no significant events with opposite sign same flavor leptons are seen after the subtraction of @xmath296 events . in msugra , the signal @xmath1 distribution is clearly harder than the background . especially , one can see the increase of the signal towards its peak , where the background is negligible . if the total @xmath1 distribution shows a bump or a clear change in the power behavior , one may claim discovery of susy particles without precise understanding of the background distribution . on the other hand , when susy mass spectrum is degenerated , the peak of the signal distribution shifts to a lower position where the background may be @xmath24610 times higher . the signal distribution shows a similar power low behavior for @xmath297 as can be seen in figure [ fig : meffa ] . the bump structure may not be detected and precise understanding of the background would be required . searching edges in @xmath25 distribution from the decay @xmath298 @xmath299 may become more important , however , as we have discussed already , the signal rate is model dependent . we do not consider the possibility to improve s / n ratio by selecting signal region @xmath300 vs @xmath2 in this subsection . table [ table : meffnum ] shows the number of signal events after the cut for @xmath1 intervals for @xmath252 generated event ( upper row ) and number of events for 10 fb@xmath189 ( lower row ) for both model a and model b. we take a moderate value of @xmath203 gev for the table , which corresponds to @xmath301 tev . @xmath1(tev)&0.8 - 1.2&1.2 - 1.6&1.6 - 2.0&2.0 - 2.4&2.4 - 3.2 & 3.2 - 4.0 & @xmath192&@xmath138(tev ) model a & & & & & & & & @xmath206&448&905&1267&999&568 & 69&0.252pb&1.35 & 22.5 & 45.6&63.9&50.3 & 28.8 & 3.4 & & 10&752&1201&1378&896 & 483&59&0.245pb&1.29 & 36.8 & 58.8 & 67.4 & 43.8 & 23.9 & 2.9 & & 20 & 813 & 1421 & 1319 & 676 & 293&17 & 0.248pb&1.20&40.3&70.5 & 65.4&33.7 & 14.5 & 0.8 & & 30 & 1246&1691&1018&336 & 95 & 9 & 0.267pb & 1.04 & 66.6&90.4 & 54.4 & 18.0&5.1 & 0.5 & & 40 & 2446&2337&758&212&63&9 & 0.314pb&0.87 & 152.3 & 145.5 & 47.2 & 13.2 & 4.0 & 0.6 & & 50 & 882 & 960 & 529 & 198 & 65 & 4 & 0.457pb & 0.90 & 80.7&87.8 & 48.4 & 18.1 & 5.9&0.4 & & 55 & 173 & 249 & 175 & 71&26 & 5 & 0.666pb & 0.94 & 23.0 & 33.2 & 23.2 & 9.5 & 3.5 & 0.7 & & model b & & & & & & & & @xmath206 & 450 & 1108 & 1405&1192 & 784 & 96 & 0.206pb&1.42 & 18.6 & 45.7 & 57.9 & 49.2 & 32.3 & 4.0 & & 10&649 & 1267 & 1585 & 1145 & 654 & 67&0.187pb & 1.38 & 24.3 & 47.4 & 59.3&42.8 & 24.5 & 2.5 & & 20 & 836 & 1528 & 1645 & 962 & 432 & 42 & 0.174pb&1.31 & 29.1 & 53.2 & 57.2 & 33.5 & 15.0 & 1.5 & & 30 & 943&1838&1362&561 & 217 & 26&0.165pb&1.2 & 31.2 & 60.7 & 45.0 & 18.5 & 7.2&0.9 & & 40 & 1711 & 1591 & 711 & 184 & 86 & 4 & 0.158pb & 0.99&54.0 & 50.2 & 22.4 & 5.8 & 2.7 & 0.1 & & 50 & 2428 & 803 & 215 & 67 & 25 & 5 & 0.154pb & 0.73 & 74.9 & 24.8 & 6.6 & 2.1 & 0.8 & 0.2 & & in the msugra limit @xmath302 , the background rate is negligible in the signal region . the number of events in @xmath303 tev is greater than 10 for @xmath304 , where the number of background events would be around 1 according to @xcite . however , the signal rate is reduced by more than a factor of @xmath305 for @xmath306 compared to that at @xmath239 . another important question is if we have a flatter power spectrum in a certain @xmath1 region . the number of background events is reduced steeply @xmath1 bins in figure [ fig : meffa ] , while we have nearly flat signal distribution between @xmath307-@xmath308 tev for msugra - like ( @xmath239 ) points . the ratio between the signal in the bins @xmath309 and @xmath310 for @xmath311 for the model a and for @xmath312 for the model b , therefore the power law for the signal is similar to that of background in figure [ fig : meffa ] . we can describe the discovery potential in terms of the mass degeneracy rather than @xmath107 . in the table [ table : meffnum ] , @xmath138 ranges from @xmath313 tev . ( @xmath314 tev ) . once @xmath138 reduces below 1 tev , 70% of @xmath315 , the separation between signal and background are not good in @xmath1 distributions . good understanding of the background distribution and the detector effect is required to exclude the model in early stage of lhc runs @xmath316fb@xmath189 . a factor 2 decrease of the signal is found for the model a between @xmath317 and @xmath199 . in this model , squark mass decreases as @xmath107 increases . while the production cross section increases as @xmath318 gets lighter , the masses of heavier neutralino and chargino increase . they are heavier than @xmath140 at @xmath199 , so that they would not be produced from @xmath140 decay at all . the lighter neutralinos and charginos are still lighter than @xmath140 but they are degenerated , so that only soft leptons are produced from the decay . the two lepton signal is also suppressed for this model point . this is an example that the rate of signal events with lepton is model dependent . the tendency of the signal reduction is same for the higher mass spectrum . the luminosity needed to discover the susy signal is larger at the degenerated susy points . for example , the production cross section is @xmath319 fb for the model a with @xmath320 gev and @xmath116 , and 15.1 events/10 fb@xmath189 is expected in the region @xmath308 tev @xmath321 tev , which may be enough to claim the discovery . on the other hand , for @xmath154 only 3.8 events/10 fb@xmath189 is expected , excluding the possibility of discovery with low luminosity . in this paper , we study the discovery potential of minimal supersymmetric models with degenerated mass spectrum at the lhc . such parameter regions have not been systematically studied in the past . susy studies have been mainly performed in the models where gluino and squark are significantly heavier than the lsp , i.e. in the msugra model , gauge mediation model , and so on . we do not study the general mssm model here , but take mixed modulas anomaly mediation ( mmam ) models , which are parametrized by @xmath322 , @xmath323 , and @xmath324 . the parameter @xmath107 is the ratio of the @xmath325 terms of the volume modulas @xmath32 in kklt model and the compensator field @xmath33 in @xmath39 msugra . this parameter @xmath107 gives us one parameter description from the msugra - like mass spectrum , where @xmath326 , to ( moderately ) degenerated mass spectrum . the models we have studied include the mssm points where @xmath327 . when sparticle mass spectrum is degenerated , energy of a particle from squark and gluino decays is small relative to the mass scale . this is reflected in the distributions of the first @xmath5 jet , @xmath2 and @xmath1 . they tend to peak at smaller values than msugra prediction for the same gluino and squark masses . the background from the sm processes increases rapidly for such low @xmath5 and low @xmath1 region , therefore , the susy signal has to compete with large background . indeed , the signal and background @xmath1 distributions are quite similar for degenerated parameters we have studied . we find the @xmath328 ratio can be far below 1 for our most degenerated parameter points . nevertheless , we found the signal distribution of degenerated susy mass spectrum show special universal pattern in @xmath1 and @xmath2 plane . this may help to determine the appropriate signal region and discriminate signal from background better . the discovery potential at the degenerated points therefore depends on our knowledge on the background distribution . the main background source to the susy productions is multijet final states involving @xmath230 , @xmath329 and @xmath232 . there are significant improvements in the lowest order estimation of the background recently ; matrix element corrections of the backgrounds are now included . theoretical uncertainty comes from matching between parton shower approximation and matrix corrections , and higher order qcd corrections ( scale uncertainty ) . experimentally backgrounds are partly estimated by the experimental data itself . for example , overall normalization of dominant @xmath21 background can be determined by looking into the events with low @xmath2 , where contamination of the susy events are small . for the degenerated mass spectrum , however , such calibration may also be affected . note that the background estimation in the susy signal region require extrapolation using monte carlo , while there is a contamination in the region for the calibration in non - trivial manner for degenerated susy mass spectrum . we look into @xmath2 distribution and find that the number of events with @xmath330 gev is increased by factor of @xmath331 for our most degenerated point ( the model b with @xmath199 , @xmath153 gev ) from msugra limit ( @xmath206 ) . even with the uncertainty discussed above , the discovery of supersymmetry in msugra models would not be a problem up to @xmath246 2 tev region . however , as we have discussed in this paper , this is not the case even for moderately degenerated susy parameters in mixed modulas anomaly mediation model . we stress the importance to include the background and its uncertainty to perform susy discovery and parameter studies for degenerated or general mssm mass spectrum , while we find that they tend to be ignored in recent studies @xcite . we would like to thank dr . tasuo kobayashi , dr . shoji asai and dr . junichi kanzaki for discussion . we also thank dr . giacomo polesello for comments to the drafts . this work is supported in part by grant in grant - in - aid for science research , ministry of education , science and culture , japan , 16081207 , 18340060 for mmn and 15340076 , 16081208 for kk . 99 s. p. martin , `` a supersymmetry primer , '' arxiv : hep - ph/9709356 . atlas collaboration , _ atlas detector and physics performance technical design report _ , cern / lhcc 99 - 14/15 ( 1999 ) . i. hinchliffe , f. e. paige , m. d. shapiro , j. soderqvist and w. yao , phys . d * 55 * , 5520 ( 1997 ) [ arxiv : hep - ph/9610544 ] . + h. bachacou , i. hinchliffe and f. e. paige , phys . d * 62 * , 015009 ( 2000 ) [ arxiv : hep - ph/9907518 ] . arxiv : hep - ph/0410364 . section 5.1.2 " a detailed analysis of the measurement of susy masses with the atlas detector at the lhc s. kachru , r. kallosh , a. linde and s. p. trivedi , phys . d * 68 * , 046005 ( 2003 ) [ arxiv : hep - th/0301240 ] . k. choi , k. s. jeong and k. i. okumura , jhep * 0509 * , 039 ( 2005 ) [ arxiv : hep - ph/0504037 ] . a. falkowski , o. lebedev and y. mambrini , jhep * 0511 * , 034 ( 2005 ) [ arxiv : hep - ph/0507110 ] . m. endo , m. yamaguchi and k. yoshioka , phys . d * 72 * , 015004 ( 2005 ) [ arxiv : hep - ph/0504036 ] . t. plehn , d. rainwater and p. skands , arxiv : hep - ph/0510144 , s. asai , talk in 4th tev4lhc , oct 20 -22 , 2005 at fermilab . m. l. mangano , m. moretti , f. piccinini , r. pittau and a. d. polosa , jhep * 0307 * , 001 ( 2003 ) [ arxiv : hep - ph/0206293 ] . see also http://mlm.home.cern.ch/mlm , [ merging matrix element and shower ] g. corcella _ et . al . _ _ _ , _ jhep _ * 01 * ( 2001 ) 010 , [ arxiv : hep - ph/0011363 ] . richardn / herwig / isawig/ http://www.physics.ucdavis.edu/ conway / research / software / pgs / pgs4-general.htm e. richter - was , d. froidevaux and l. poggioli , `` atlfast 2.0:a fast simulation package for atlas '' , tech . rep.atl-phys-98-131(1998 ) b. c. allanach _ et al . _ , in _ proc . of the aps / dpf / dpb summer study on the future of particle physics ( snowmass 2001 ) _ ed . n. graf , eur . j. c * 25 * ( 2002 ) 113 [ econf * c010630 * ( 2001 ) p125 ] [ arxiv : hep - ph/0202233 ] . a. j. barr , phys . b * 596 * , 205 ( 2004 ) [ arxiv : hep - ph/0405052 ] . t. goto , k. kawagoe and m. m. nojiri , at the cern large hadron collider , phys . d * 70 * , 075016 ( 2004 ) [ erratum - ibid . d * 71 * , 059902 ( 2005 ) ] [ arxiv : hep - ph/0406317 ] . j. m. smillie and b. r. webber , jhep * 0510 * , 069 ( 2005 ) [ arxiv : hep - ph/0507170 ] . h. baer , e. k. park , x. tata and t. t. wang , arxiv : hep - ph/0604253 . n. arkani - hamed , g. l. kane , j. thaler and l. t. wang , arxiv : hep - ph/0512190 .
discovery of supersymmetric ( susy ) particles at the large hadron collider ( lhc ) has been studied for the models where squarks and gluino are much heavier than the lightest supersymetric particle ( lsp ) . in this paper , we investigate the susy discovery in the models with degenerated mass spectrum up to @xmath0 . such mass spectrum is predicted in certain parameter region of the mixed modulas anomaly mediation ( mmam ) model . we find that the effective transverse mass of the signal for the degenerated parameters shows the distribution similar to that of the background . experimental sensitivity of the susy particles at the lhc therefore depends on the uncertainty of the background in this class of model . we also find that susy signal shows an interesting correlation between @xmath1 and @xmath2 which may be used to determine the signal region properly to enhance the s / n ratio even if the sparticle masses are rather degenerated . the structure is universal for the models with new heavy colored particles decaying into visible particles and a stable neutral particle , dark matter .
we know from observational constraints that the hydrogen in the intergalactic medium ( igm ) was reionized by redshift @xmath5 ( schneider , schmidt , & gunn 1991 ) and that the was reionized at @xmath2 ( reimers et al . 1997 ) . however , the exact nature of the ionizing sources is still uncertain . the observed drop in the space density of bright quasars at @xmath6 ( pei 1995 ) suggests that early stellar populations played a role in the reionization of hydrogen , but it is not known whether hot stars produced photons at rates sufficient to ionize the universe before this epoch . our understanding of reionization is closely connected to our knowledge of the extragalactic radiation field . models of reionization assume an extragalactic spectrum composed of the distinct intrinsic spectra of active galactic nuclei ( agn ) and star - forming galaxies . this composite spectrum determines the reionization epoch , the / ionization ratio , and metal - line absorption ratios at @xmath7 . models by giroux & shull ( 1997 ) of the observed / ratio at @xmath2 ( songaila & cowie 1996 ) and subsequent work by haardt & madau ( 1996 ) and fardal et al . ( 1998 ) on the gunn - peterson effect demonstrated that the observations are consistent with an extragalactic spectrum produced by a mixture of qsos and hot stars . these conclusions depend , however , on the assumed shape of the ionizing spectrum of stars . while the details of these spectra vary ( sutherland & shull 1999 ; leitherer et al . 1999 ) , one common element is the assumption that hot stars contribute few ionizing photons . some models of reionization assume a phenomenological prescription for star formation , in which a single parameter describes the conversion efficiency of mass to stars and then to ionizing photons ( gnedin & ostriker 1997 ) . others use existing model grids of stellar structure and atmospheres designed for application to metal - poor environments ( haiman & loeb 1997 ) . the first method ignores the details of star formation , ionizing photon production , and radiation escape from the immediate regions of star formation . the second method has the drawback of applying theoretical calculations of metal - poor stars to the very different regime of zero metallicity . the existing grids of stellar evolution tracks extend to @xmath8 ( schaerer et al . 1996 and references therein ) . existing model atmospheres extend to @xmath9 but are limited in the range of stellar parameters ( kurucz 1992 ) . these models are meant for application to low - metallicity starbursts ( leitherer et al . 1999 ) and metal - poor galactic halo populations . as shown by existing models of metal - free stars ( ezer 1972 ; ezer & cameron 1971 ; el eid et al . 1983 ) , however , stars with @xmath10 are quite different from their @xmath11 counterparts . thus , when we consider the effects of stellar populations on reionization , we must use true metal - free models to predict the ionizing photon production of the first generation of stars . in this _ letter _ we adopt the common term `` population iii '' or `` pop iii '' for metal - free stars , which are understood to have formed from primeval gas . although extremely metal - poor populations ( @xmath12 ) may fit an observer s definition of pop iii , we apply that label to metal - free stars only . in 2 we present structure and atmosphere models of metal - free stars , and in 3 we predict ionizing photon yields from these model stars . in 4 we evaluate the effects of these models on the epoch of reionization , and in 5 we comment on further cosmological implications of metal - free stellar populations . the models presented here are static stellar structure models calculated using a fitting - method technique that incorporates opal radiative opacities ( rogers & iglesias 1992 ) and analytic expressions for energy generation . these models were used to predict the effective temperature ( @xmath13 ) , luminosity ( @xmath14 ) , and surface gravity ( @xmath15 ) of stars with mass 2 90 @xmath16 ( at 5 10 @xmath16 intervals ) . there is currently no full set of evolutionary tracks for metal - free stars , and our models do not incorporate evolution . however , the existing evolutionary tracks for metal - free stars ( castellani , chieffi , & tornambe 1983 ; chieffi & tornambe 1984 ) show that , like their metal - enriched counterparts , these stars become systematically cooler , larger , and more luminous over their h - burning lifetimes , which are similar in duration . therefore , we assume that metal - free tracks differ in their first - order characteristics only in their starting point on the hertzsprung - russell ( hr ) diagram . if so , the `` gain '' in ionizing photons at @xmath11 is maintained throughout the main sequence ( ms ) lifetime of the star , when most of its ionizing radiation is released . existing evolutionary tracks of metal - free stars show that these stars may build up a small fraction of c nuclei ( @xmath17 ) via the triple-@xmath18 process before they join the h - burning main sequence ( el eid et al . 1983 ; castellani et al . 1983 ) . following this result , we assume that stars with @xmath19 @xmath16 are enriched to @xmath20 via triple-@xmath18 burning prior to the onset of ms h - burning . we use these pre - enriched models in all our analysis . we will explore the detailed effects of pre - ms self - enrichment in a later paper , once a comprehensive grid of tracks is available . figure [ fig1 ] shows an hr diagram for zero - age main sequence models with pop i and pop iii metallicities . the most striking feature of the metal - free models is the high temperature they maintain at their photospheres . these stars derive their nuclear energy from a combination of inefficient proton - proton burning and cno burning with the small fraction of c built up in the pre - ms phase ( castellani et al . 1983 ; el eid et al . 1983 ) . as a result of lower energy generation rates in the convective core , they maintain core temperatures in excess of 10@xmath21 k to support the mass against gravitational collapse . together with reduced radiative opacity in their envelopes , the high core temperatures of pop iii stars make these stars hotter and smaller than their metal - enriched counterparts . using homology relations , and assuming constant ( electron scattering ) opacity and a cno burning rate @xmath22 , we find that massive stars have @xmath23 and @xmath24 , in good agreement with our numerical models . with @xmath15 and @xmath13 in hand , we need a model atmosphere to derive the spectral luminosity distribution @xmath25 . a model atmosphere is necessary because a simple blackbody curve for each @xmath13 will not accurately reproduce the spectrum near the ionization edges of , , and . for each stellar model and @xmath26 @xmath13 pair , we calculated a static , non - lte model atmosphere that was used to predict the spectral luminosity distribution for the star . we used the atmosphere code tlusty ( hubeny & lanz 1995 ) to produce the model atmospheres and its included package synspec to produce the continuum spectra . figure [ fig2 ] illustrates the change in the spectral distribution of a 15 @xmath16 star at @xmath8 ( @xmath27 36,000 k ) and @xmath28 ( @xmath27 63,000 k ) . the key between the two spectra if the high @xmath13 of the @xmath11 model . a pop ii star of the same @xmath13 would exhibit a continuum shape similar to the pop iii model but attenuated by metal - line absorption . however , a pop ii star at 15 @xmath16 is unlikely to reach @xmath27 63,000 k during its ms lifetime . the model atmospheres were used to predict the rates of ionizing photon production , @xmath29 , in photons s@xmath30 , for all the modeled stars : @xmath31 where @xmath32 is the radius of the star , @xmath33 is the spectral flux distribution in erg @xmath34 s@xmath30 hz@xmath30 , and the indices @xmath35 = 0 , 1 and 2 correspond to integration over the , , and ionizing continua , @xmath36 ( 13.60 , 24.58 , and 54.40 ev , respectively ) . the @xmath29 for pop iii stars appear in figure [ fig3 ] . individual metal - free stars emit far more of their energy in photons with @xmath37 ev than do their pop i and pop ii counterparts . this increase produces a 50% gain in the total ionizing photon production at high mass ( 40 70 @xmath16 ) and gains by factors of @xmath38 at moderate mass ( 10 30 @xmath16 ) . above 70 @xmath16 , a larger fraction of the star s energy is released above 1 ryd , but the average photon energy increases such that the overall gain in @xmath39 is modest . a striking feature of the ionizing photon production is the high fraction of the photons emitted with energy sufficient to ionize and . these fractions are expressed by the ratios @xmath40 and @xmath41 , displayed in figure [ fig3 ] . high - mass pop iii stars emit 60 70% of their ionizing photons in the continuum and up to 12% of these photons in the continuum . for comparison , @xmath42 0.2 0.4 for pop i o3 o5 stars with @xmath13 in the range 45,000 51,000 k ( vacca , garmany , & shull 1996 ; schaerer & de koter 1997 ) . the overall gain in integrated ionizing photon production is best evaluated with synthetic spectra of stellar clusters ( sutherland & shull 1999 ; leitherer et al . 1999 ) , which compare the rates @xmath29 per unit mass of stellar material . synthetic spectra of pop iii and pop ii zero - age clusters with a standard initial mass function ( salpeter imf with @xmath43 ) appear in figure [ fig4 ] . the gain in @xmath39 ( s@xmath30 ) is near 50% for this imf . however , @xmath44 increases from 52.4 to 52.9 , and @xmath45 increases from 46.1 to 51.7 . this dramatic increase in the capability of stars to ionize and are not predicted by populations that approximate pop iii with existing metal - poor models . metal - free stars emit 50% more ionizing radiation per unit mass of stellar material than metal - enriched stars . given the uncertain efficiency of star formation out of primeval gas ( abel et al . 1998 ) , our conclusion that pop iii stars are more efficient ionizing sources ( per unit mass ) does not necessarily mean that they are more capable than other stellar populations of reionizing the universe . for this reason , we calculate a star formation rate ( sfr ) for metal - free stars necessary to reionize the universe and test this quantity against observations . madau , haardt , & rees ( 1999 ) imposed the condition that the universe is reionized when the number of ionizing photons emitted in one recombination time equals the mean number of hydrogen atoms . accounting for clumping of the igm and assuming a typical rate for ionizing photon production per unit mass , they found that the critical sfr required to reionize by @xmath1 is 0.013 @xmath46 @xmath16 yr@xmath30 mpc@xmath47 , where @xmath48 is the fraction of ionizing photons that escape into the igm ( see tumlinson et al . 1999 for a review ) . based on the per - mass photon emission rates of the pop iii cluster in figure [ fig4 ] , we estimate the critical rate of metal - free star formation to be @xmath49 @xmath16 yr@xmath30 mpc@xmath47 . this requirement would be comparable to the inferred ( highly uncertain ) sfr at @xmath1 if @xmath48 = 0.40 ( madau , mozetti , & dickenson 1998 ) . in addition to enhanced ionizing capability , pop iii stars convert less of their initial mass into metals . preliminary evaluation of the metal yields of pop iii stars ( woosley & weaver 1995 ) indicates that for every solar mass of stars formed , 0.007 @xmath16 in metals are released . assuming that this value holds throughout the pop iii epoch , then the critical sfr for h reionization by pop iii stars implies a metal enrichment rate of @xmath50 @xmath16 yr@xmath30 mpc@xmath47 . over the 500 myr between @xmath51 and @xmath1 , pop iii stars would enrich the universe to a mean metallicity @xmath52 ( @xmath53 km s@xmath30 mpc@xmath30 , @xmath54 ) , 20% of the minimum value @xmath55 observed at @xmath2 and @xmath562% of the average metallicity in damped ly@xmath18 systems at that redshift ( pettini 1999 ) . our new models of metal - free stars raise the possibility that stars are also responsible for reionization . various studies ( reimers et al . 1997 ; heap et al . 1999 ; hogan , anderson , & rugers 1997 ) constrain the epoch of reionization to @xmath57 with observations of patchy absorption . conventional wisdom ( fardal et al . 1998 ; madau et al . 1999 ) states that only the hard spectra of agn could produce enough photons with @xmath58 54.4 ev to reionize , based in part on the result @xmath59/@xmath60 for low - metallicity stellar spectra to which wolf - rayet stars contribute only minimally ( leitherer & heckman 1995 ) . however , for the pop iii cluster in figure [ fig4 ] , @xmath59/@xmath61 . thus , for @xmath62 ( @xmath63 ) and assuming full ionization of h and he , the region excited by this cluster will have 50% the radius of its strmgren sphere . this large region may imply that reionization differs in the denser regions , compared to the low - density igm where recombinations are not important . pop iii stars can be distinguished from metal - enriched stars by their theoretical radii and effective temperatures . unfortunately , these features are not observable directly . however , these characteristics modify the spectrum of the stars in ways are that are potentially observable . we discuss these in brief below , but defer detailed studies of metal - free stars to future work . if the ionizing spectra of clusters follow a power law , @xmath64 , we can use @xmath18 to characterize the effects of pop iii stars on the intergalactic radiation field . the pop iii cluster in figure [ fig4 ] has a hard ( and inverted ) spectral index @xmath65 in the range 1 4 ryd and @xmath66 for @xmath0 ryd . for comparison , the pop ii cluster in figure [ fig4 ] has @xmath67 and negligible flux above 4 ryd . thus , metal - free stars could contribute harder radiation to the extragalactic spectrum than qsos with intrinsic @xmath68 between 0.9 2.6 ryd ( zheng et al . 1997 ) . at @xmath5 this hard euv spectrum is accessible to ground- and space - based instruments in the optical and near - uv . the dramatic increase in ionizing photons from metal - free stars implies that they will have distinctive effects on their neighborhoods . simple models of pop iii regions show that pop iii stars excite sizable regions . the @xmath691640 and @xmath694686 recombination lines of might be detected for targets with @xmath70 in the 1 - 5 @xmath71 m range with ngst . similarly , the radio recombination lines of he ii may provide a unique signature of these stars . however , recombination emission observed in the spectra of metal - poor extragalactic regions has often been attributed to wolf - rayet stars , x - ray binaries , or shocks . ( garnett et al . 1991 ; de mello et al . 1998 ; izotov et al . these sources are likely to be less important at high @xmath72 , but they may complicate the identification of metal - free stellar populations . the photodissociation of h@xmath73 by fuv radiation ( 912 1126 ) from the first luminous sources in the universe is suggested to have inhibited subsequent star formation near these sources by destroying their only coolant ( haiman , rees , & loeb 1997 ; ciardi , ferrara , & abel 1999 ) . photons with @xmath74 = 11.2 13.6 ev can propagate freely into neutral , dust - free gas and dissociate h@xmath73 by exciting permitted transitions to the @xmath75 state , 10 15% of which decay into the continuum . this `` negative feedback '' is presumed to precede the ionization front by a distance dependent on the spectrum of the ionizing source . pop iii clusters , with enhanced ionizing photon production and suppressed fuv flux , may dissociate h@xmath73 with their ionization fronts . the hard spectra of pop iii stars may also affect the igm ionization ratios through changes to the extragalactic spectrum . fardal et al . ( 1998 ) define the ratio @xmath76 to express the relative column densities of and . this ratio is sensitive to the shape of the extragalactic spectrum and can be used to predict the optical depths @xmath77 and @xmath78 in the igm . an increase in the ionizing flux favors and decreases @xmath79 . fardal et al . ( 1998 ) estimate that @xmath80 is necessary to explain @xmath81 1 5 observed at @xmath2 ( davidsen , kriss , & zheng 1996 ) . the @xmath11 stellar spectra give @xmath82 , inconsistent with the observed opacity if these stars are still forming at that epoch . thus , @xmath11 stars may not dominate the intergalactic radiation field at @xmath83 . in summary , we outline a general picture of the era of pop iii stars based on our models and consistent with the observations discussed above . we assume that the first stars formed at @xmath84 , consistent with simulations of large - scale structure . if the total cosmic star formation rate exceeded the critical rate calculated in 4 , then these first stars may have reionized hydrogen and helium in the universe . upon their deaths , they enriched the universe to an average metallicity 20% of that observed at @xmath2 . population iii then faded , leaving their metal - enriched progeny to provide the rest of the metals and the softer radiation seen at @xmath2 . abel , t. , anninos , p. , norman , m. l. , & zhang , y. 1998 , , 508 , 518 castellani , v. , chieffi , a. , & tornambe , a. 1983 , , 272 , 249 chieffi , a. , & tornambe , a. 1984 , , 287 , 745 ciardi , b. , ferrara , a. , & abel , t. 1999 , , in press , astro - ph/9911137 davidsen , a. f. , kriss , g. a. , & zheng , w. 1996 , nature , 380 , 47 de mello , d. f. , schaerer , d. , heldmann , j. , & leitherer , c. 1998 , , 507 , 199 el eid , m. f. , fricke , k. j. , & ober , w. w. 1983 , a&a , 119 , 54 ezer , d. 1961 , , 133 , 159 ezer , d. , & cameron , a. g. w. 1971 , ap&ss , 14 , 399 fardal , m. a. , giroux , m. l. , & shull , j. m. 1998 , , 115 , 2206 garnett , d. r. , kennicutt , r. c. , jr . , chu , y .- h . , & skillman , e. d. 1991 , , 373 , 458 giroux , m. l. , & shull , j. m. 1997 , , 113 , 1505 gnedin , n. y. , & ostriker , j. p. 1997 , , 486 , 581 haardt , f. , & madau , p. 1996 , , 461 , 20 haiman , z. , & loeb , a. 1997 , , 483 , 21 haiman , z. , rees , m. j. , & loeb , a. 1997 , , 476 , 458 heap , s. r. , et al . 1999 , , in press , astro - ph/9812429 hogan , c. j. , anderson , s. f. , & rugers , m. h. 1997 , , 113 , 1495 hubeny , i. , & lanz , t. 1995 , , 439 , 875 izotov , y. , et al . 1997 , , 476 , 698 kurucz , r. 1992 , in stellar populations of galaxies , ed . b. barbuy & a. renzini ( dordrecht : kluwer ) leitherer et al . 1999 , , 123 , 3 leitherer , c. , & heckman , t. m. 1995 , , 96 , 9 madau , p. , haardt , f. , & rees , m. j. 1999 , , 514 , 648 madau , p. , mozetti , l. , & dickinson , m. 1998 , , 498 , 106 pei , y.c . 1995 , , 438 , 623 pettini , m. 1999 , in chemical evolution from zero to high redshift , ed . j. walsh , & m. rosa ( berlin : springer ) , astro - ph/9902173 reimers , d. , kohler , s. , wisotzki , l. , groote , d. , rodriguez - pascual , p. , & wamsteker , w. 1997 , a&a , 327 , 890 rogers , f. j. , & iglesias , c. a. 1992 , , 79 , 507 schaerer , d. , de koter , a. , schmutz , w. , & maeder , a. 1996 , a&a , 312 , 475 schaerer , d. , de koter , a. 1997 , a&a , 322 , 598 schneider , d. , schmidt , m. , & gunn , j. e. 1991 , , 101 , 2004 songaila , a. , & cowie , l. l. 1996 , aj , 112 , 335 sutherland , r. s. , & shull , j. m. 1999 , in preparation tumlinson , j. , giroux , m. l. , shull , j. m. , & stocke , j. t. 1999 , , 118 , in press , astro - ph/9907174 vacca , w. d. , garmany , c. d. , & shull , j. m. 1996 , , 460 , 914 woosley , s. e. , & weaver , t. a. 1995 , , 101 , 181 zheng , w. , kriss , g. a. , telfer , r.c . , grimes , j. p. , & davidsen , a. f. 1997 , , 475 , 469
we present stellar structure and atmosphere models of metal - free stars and examine them from a cosmological point of view . metal - free stars exhibit high effective temperatures and small sizes relative to metal - enriched stars of equal mass . these unique physical characteristics enhance the ionizing photon production by metal - free stars , particularly in the ( @xmath0 ryd ) continuum . the star formation rate of metal - free stars necessary to reionize the hydrogen in the universe by @xmath1 is consistent with the inferred star formation rate at that epoch . however , the hard stellar spectra are inconsistent with the observations of opacity in the igm at @xmath2 , indicating that the period of metal - free star formation ended before that epoch . we examine the effects of these stars on the ionization balance of the igm , the radiative feedback of the first luminous objects , and the extragalactic radiation field . we comment on the prospects for detecting metal - free stellar populations with the @xmath3 and @xmath4 recombination lines of .
a _ software design pattern _ is not a finished design , it is a description or template that can be instanciated in order to be used in many different situations . in this paper , we propose _ inference system patterns _ that can be instanciated with monads or comonads in order to be used for proving properties of different effects . in order to formalize computational effects one can choose between types and effects systems @xcite , monads @xcite and their associated lawvere theories @xcite , comonads @xcite , or decorated logics @xcite . starting with moggi s seminal paper @xcite and its application to haskell @xcite , various papers deal with the effects arising from a monad , for instance @xcite . each of these approaches rely on some classification of the syntactic expressions according to their interaction with effects . in this paper we use decorated logics which , by extending this classification to equations , provide a proof system adapted to each effect . this paper presents equational - based logics for proving first order properties of programming languages involving effects . we propose two dual patterns , consisting in a language with an inference system , for building such a logic . the first pattern provides inference rules which can be interpreted in the cokleisli category of a comonad and the kleisli category of the associated monad . in a dual way , the second pattern provides inference rules which can be interpreted in the kleisli category of a monad and the cokleisli category of the associated comonad . the logics combine a three - levels effect system for terms consisting of pure terms and two other kinds of effects called observers / constructors and modifiers , and a two - levels system for strong and weak equations . each pattern provides generic rules for dealing with any comonad ( respectively monad ) , and it can be extended with specific rules for each effect . the paper presents two use cases : a language with state and a language with exceptions . for the language with state we use a comonadic semantics and we prove that the equational theory obtained is hilbert - post complete , which provides a new proof for a result in @xcite . for the language with exceptions we extend the standard monadic semantics in order to catch exceptions ; this relies on the duality between states and exceptions from @xcite . we do not claim that each effect arises either from a comonad or from a monad , but this paper only deals with such effects . intuitively , an effect which observes features may arise from a comonad , while an effect which constructs features may arise from a monad @xcite . however , some interesting features in the comonad pattern stem from the well - known fact that each comonad determines a monad on its cokleisli category , and dually for the monad pattern . more precisely , on the monads side , let @xmath0 be a monad on a category @xmath1 and let @xmath2 be the kleisli category of @xmath0 on @xmath1 . then @xmath3 can be seen as the endofunctor of a comonad @xmath4 on @xmath2 , so that we may consider the cokleisli category @xmath5 of @xmath4 on @xmath2 . the canonical functors from @xmath1 to @xmath2 and from @xmath2 to @xmath5 give rise to a hierarchy of terms : pure terms in @xmath1 , constructors in @xmath2 , modifiers in @xmath5 . this corresponds to the three translations of a typed lambda calculus into a monadic language @xcite . on the comonads side , we get a dual hierarchy : pure terms in @xmath1 , observers in @xmath2 , modifiers in @xmath5 . we instanciate these patterns with two fundamental examples of effects : state and exceptions . following @xcite , we consider that the states effect arise from the comonad @xmath6 ( where @xmath7 is the set of states ) , thus a decorated logic for states is built by extending the pattern for comonads . the comonad itself provides a decoration for the lookup operation , which observes the state , while the monad on its cokleisli category provides a decoration for the update operation . following @xcite , we consider that the exceptions effect arise from the monad @xmath8 ( where @xmath9 is the set of exceptions ) , thus a decorated logic for exceptions is built by extending the pattern for monads . the monad itself provides a decoration for the raising operation , which constructs an exception , while the comonad on its kleisli category provides a decoration for the handling operation . in fact the decorated logic for exceptions is not exactly dual to the decorated logic for states if we assume that the intended interpretation takes place in a distributive category , like the category of sets , which is not codistributive . other effects would lead to other additional rules , but we have chosen to focus on two effects which are well known from various points of view . our goal is to enligthen the contributions of each approach : the annotation system from the types and effects systems @xcite , the major role of monads for some effects @xcite , and the dual role of comonads @xcite , as well as the flexibility of decorated logics @xcite . moreover , proofs in decorated logics can be checked with the coq proof assistant ; a library for states is available there : http://coqeffects.forge.imag.fr . in this paper we focus on finite products and coproducts ; from a programming point of view this means that we are considering languages with @xmath10-ary operations and with case distinction , but without loops or higher - order functions . in a language with effects there is a well - known issue with @xmath10-ary operations : their interpretation may depend on the order of evaluation of their arguments . in this paper we are looking for languages with case distinction and with _ sequential products _ , which allows to force the order of evaluation of the arguments , whenever this is required . it is well known that ( co)monads fit very well with composition but require additional assumptions for being fully compatible with products and coproducts . this corresponds to the fact that in the patterns from section [ sec : patterns ] , which are valid for any ( co)monad , the rules for products and coproducts hold only under some decoration constraints . however , such assumptions are satisfied for several ( co)monads . this is in particular the case for the state comonad and the exceptions monad . in section [ sec : patterns ] we describe the patterns for a comonad and for a monad . the first pattern is instanciated with the comonad for state in section [ sec : states ] , and we prove the hilbert - post completeness of the decorated theory for state . in section [ sec : exceptions ] we instanciate the second pattern to the monad for exceptions . in this section we define a grammar and an inference system for two logics @xmath11 and @xmath12 , then we define an interpretation of these logics in a category with a comonad and a monad , respectively . the logics @xmath11 and @xmath12 are called _ decorated _ logics because their grammar and inference rules are essentially the grammar and inference rules for a `` usual '' logic , namely the equational logic with conditionals ( denoted @xmath13 ) , together with _ decorations _ for the terms and for the equations . the decorations for the terms are similar to the _ annotations _ of the types and effects systems @xcite . decorated logics are introduced in @xcite in an abstract categorical framework , which will not be explicitly used in this paper . the grammar of the equational logic with conditionals is reminded in figure [ fig : logeq - grammar ] . each term has a source type and a target type . as usual in categorical presentations of equational logic , a term has precisely one source type , which can be a product type or the unit type . each equation relates two parallel terms , i.e. , two terms with the same source and the same target . this grammar will be extended with decorations in order to get the grammar of the logics @xmath11 and @xmath12 . @xmath14}\mid { \mathit{in}}_{t , t,1}\mid{\mathit{in}}_{t , t,2}\mid{[\;]}_t\mid \\ \textrm{equations : } & e::= f\equiv f \\ \hline \end{array}\ ] ] the rules in figure [ fig : pattern - rules ] are _ patterns _ , in the following sense : when the boxes in the rules are removed , we get usual rules for the logic @xmath13 , which may be interpreted in any bicartesian category . when the boxes are replaced by decorations , we get a logic which , according to the choice of decorations , may be interpreted in a bicartesian category with a comonad or a monad . there may be other ways to decorate the rules for @xmath13 , but this is beyond the scope of this paper . @xmath15[c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}f } \qquad { \textrm{(sym ) } } \quad \dfrac{f^\box { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}g^\box}{g { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}f } \qquad { \textrm{(trans ) } } \quad \dfrac{f^\box { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}g^\box { \;\;}g^\box { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}h^\box}{f { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}h } \\ { \textrm{(repl ) } } \quad \dfrac{f_1^\box{{\,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}f_2^\box\colon a\to b { \;\;}g^\box\colon b\to c } { g\circ f_1 { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}g\circ f_2 } \qquad { \textrm{(subs ) } } \quad \dfrac{f^\box\colon a\to b { \;\;}g_1^\box{{\,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}g_2^\box\colon b\to c } { g_1 \circ f { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}g_2\circ f } \\ \hline \mbox{categorical rules } \\ { \textrm{(id ) } } \quad \dfrac{a}{{\mathit{id}}_a^\box\colon a\to a } \qquad { \textrm{(comp ) } } \quad \dfrac{f^\box\colon a\to b \quad g^\box\colon b\to c } { ( g\circ f)^\box \colon a\to c } \\ { \textrm{(id - source ) } } \quad \dfrac{f^\box\colon a\to b}{f\circ { \mathit{id}}_a { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}f } \qquad { \textrm{(id - target ) } } \quad \dfrac{f^\box\colon a\to b}{{\mathit{id}}_b\circ f { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}f } \\ { \textrm{(assoc ) } } \quad \dfrac{f^\box\colon a\to b { \;\;}g^\box\colon b\to c { \;\;}h^\box\colon c\to d } { h\circ ( g\circ f ) { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}(h\circ g)\circ f } \\ \hline \mbox{product rules } \\ { \textrm{(prod ) } } \quad \dfrac{b_1 \quad b_2 } { b_1{\!\times\!}b_2 \quad { \mathit{pr}}_1^\box\colon b_1{\!\times\!}b_2 \to b_1 \quad { \mathit{pr}}_2^\box\colon b_1{\!\times\!}b_2 \to b_2 } \\ { \textrm{(pair ) } } \quad \dfrac { f_1^\box\colon a \to b_1 \quad f_2^\box\colon a \to b_2 } { { \langle f_1,f_2 \rangle}^\box\colon a\to b_1{\!\times\!}b_2 } \\ { \textrm{(pair - eq ) } } \quad \dfrac{f_1^\box\colon a\to b_1 { \;\;}f_2^\box\colon a\to b_2 } { { \mathit{pr}}_1\circ{\langle f_1,f_2 \rangle } { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}f_1 \quad { \mathit{pr}}_2\circ{\langle f_1,f_2 \rangle } { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}f_2 } \\ { \textrm{(pair - u ) } } \quad \dfrac{f_1^\box{\!\colon\!}a\!{\!\to\!}\ ! b_1 { \;\;}f_2^\box{\!\colon\!}a\!{\!\to\!}\ ! b_2 { \;\;}g^\box{\!\colon\!}a\!{\!\to\!}\ ! b_1{\!\times\!}b_2 { \;\;}{\mathit{pr}}_1\circ g{{\,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}f_1 { \;\;}{\mathit{pr}}_2\circ g{{\,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}f_2 } { g { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}{\langle f_1,f_2 \rangle } } \\ { \textrm{(unit ) } } \quad \dfrac{}{{\mathbb{1 } } } \qquad { \textrm{(final ) } } \quad \dfrac{a}{{\langle \ ; \rangle}_a^\box\colon a\to { \mathbb{1 } } } \qquad { \textrm{(final - u ) } } \quad \dfrac{f^\box\colon a\to { \mathbb{1}}}{f { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}{\langle \ ; \rangle}_a } \\ \hline \mbox{coproduct rules } \\ { \textrm{(coprod ) } } \quad \dfrac{a_1 \quad a_2 } { a_1{\!+\!}a_2 \quad { \mathit{in}}_1^\box\colon a_1\to a_1{\!+\!}a_2 \quad { \mathit{in}}_2^\box\colon a_2\to a_1{\!+\!}a_2 } \\ { \textrm{(copair ) } } \quad \dfrac { f_1^\box\colon a_1 \to b \quad f_2^\box\colon a_2 \to b } { { [ f_1|f_2 ] } ^\box\colon a_1{\!+\!}a_2 \to b } \\ { \textrm{(copair - eq ) } } \quad \dfrac { f_1^\box\colon a_1 \to b { \;\;}f_2^\box\colon a_2 \to b } { { [ f_1|f_2 ] } \circ { \mathit{in}}_1 { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}f_1 \quad { [ f_1|f_2 ] } \circ { \mathit{in}}_2 { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}f_2 } \\ { \textrm{(copair - u ) } } \quad \dfrac{g^\box{\!\colon\!}a_1\!{\!+\!}\ ! a_2 { \!\to\!}b { \;\;}f_1^\box{\!\colon\!}a_1 { \!\to\!}b { \;\;}f_2^\box{\!\colon\!}a_2 { \!\to\!}b { \;\;}g\circ { \mathit{in}}_1 { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}f_1 { \;\;}g\circ { \mathit{in}}_2 { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}f_2 } { g { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\ , } } { [ f_1|f_2 ] } } \\ { \textrm{(empty ) } } \quad \dfrac{}{{\mathbb{0 } } } \qquad { \textrm{(initial ) } } \quad \dfrac{b}{{[\;]}_b^\box\colon { \mathbb{0}}\to b } \qquad { \textrm{(initial - u ) } } \quad \dfrac{g^\box\colon { \mathbb{0}}\to b}{g { { \,\,\makebox[0pt][c]{{$\box$}}\makebox[0pt][c]{\raisebox{2pt}{\tiny\ensuremath{\equiv}}}\,\,}}{[\;]}_b } \\ \hline \end{array}\ ] ] in the logic @xmath11 for comonads , each term has a decoration which is denoted as a superscript @xmath16 , @xmath17 or @xmath18 : a term is _ pure _ when its decoration is @xmath16 , it is an _ accessor _ ( or an _ observer _ ) when its decoration is @xmath17 and a _ modifier _ when its decoration is @xmath18 . each equation has a decoration which is denoted by replacing the symbol @xmath19 either by @xmath20 or by @xmath21 : an equation with @xmath20 is called _ strong _ , with @xmath21 it is called _ the inference rules of @xmath11 are obtained by introducing some _ conversion _ rules and by decorating the rules in figure [ fig : pattern - rules ] . when writing terms , if a decoration does not matter or if it is clear from the context , it may be omitted . * the conversion rules are : @xmath22 the conversions for terms are _ upcasting _ conversions . + we will always use them in a _ safe _ way , by interpreting them as injections . this allows to avoid any specific notation for these conversions ; an accessor @xmath23 may be converted to a modifier which is denoted @xmath24 : both have the same name although they are distinct terms ; similarly , a pure term @xmath25 may be converted to @xmath26 or to @xmath27 . an equation between terms with distinct decorations does not imply any downcasting of its members ; for instance , if @xmath28 then it does not follow that @xmath29 is downcasted to @xmath30 . the conversions for equations mean that strong and weak equations coincide on pure terms and accessors and that each strong equation between modifiers can be seen as a weak one . * all rules of @xmath13 are decorated with @xmath16 for terms and @xmath20 for equations : the pure terms with the strong equations form a sublogic of @xmath11 which is isomorphic to @xmath13 . thus we get @xmath31 , @xmath32 , @xmath33 , @xmath34 , @xmath35}^{{(0)}}$ ] . * the congruence rules for equations take all decorations for terms and for equations , with one notable exception : the replacement rule for weak equations holds only when the replaced term is pure : @xmath36 * the categorical rules hold for all decorations and the decoration of a composed terms is the maximum of the decorations of its components . * the product rules hold only when the given terms are pure or accessors and the decoration of a pair is the maximum of the decorations of its components . thus , @xmath10-ary operations can be used only when their arguments are accessors . * the coproduct rules hold only when the given terms are pure and a copair is always pure , which is the maximum of the decorations of its components . thus , case distinction can be done only for pure terms . in order to give a meaning to the logic @xmath11 , let us consider a bicartesian category @xmath37 with a comonad @xmath38 satisfying the epi requirement , i.e. , @xmath39 is an epimorphism for each object @xmath40 ( the dual assumption is discussed in @xcite ) . then we get a model @xmath41 of the decorated logic @xmath11 as follows . * the types are interpreted as the objects of @xmath37 . * the terms are interpreted as morphisms of @xmath37 : a pure term @xmath42 as a morphism @xmath43 in @xmath37 ; an accessor @xmath44 as a morphism @xmath45 in @xmath37 ; and a modifier @xmath46 as a morphism @xmath47 in @xmath37 . * the conversion from pure terms to accessors is interpreted by mapping @xmath48 to @xmath49 . the epi requirement implies that this conversion is safe . * the conversion from accessors to modifiers is interpreted by mapping @xmath50 to @xmath51 . it is easy to check that this conversion is safe . * when a term @xmath29 has several decorations ( because it is pure or accessor , and thus can be upcasted ) we will denote by @xmath29 any one of its interpretations : a pure term @xmath42 may be interpreted as @xmath43 and as @xmath52 and as @xmath47 , and an accessor @xmath44 as @xmath52 and as @xmath47 . the choice will be clear from the context , and when several choices are possible they will give the same result , up to conversions . for this reason , we will describe the interpretation of the rules only for the largest possible decorations . * the identity @xmath53 is interpreted as @xmath54 in @xmath37 ; * the composition of two modifiers @xmath55 and @xmath56 is interpreted as @xmath57 in @xmath37 . * an equation between modifiers @xmath58 is interpreted by an equality @xmath59 in @xmath37 . * a weak equation between modifiers @xmath60 is interpreted by an equality @xmath61 in @xmath37 . * the unit type is interpreted as the final object of @xmath37 and the term @xmath62 as the unique morphism from @xmath40 to @xmath63 in @xmath37 . * the product @xmath64 with its projections is interpreted as the binary product in @xmath37 and the pair of @xmath65 and @xmath66 as the pair @xmath67 in @xmath37 . * the empty type is interpreted as the initial object of @xmath37 and the term @xmath35}_a^{{(0)}}\colon { \mathbb{0}}\to a$ ] as the unique morphism from @xmath68 to @xmath40 in @xmath37 . * the coproduct @xmath69 with its coprojections is interpreted as the binary coproduct in @xmath37 and the copair of @xmath70 and @xmath71 as the copair @xmath72 } \colon a_1+a_2\to { t}b$ ] in @xmath37 . the dual of the decorated logic @xmath11 for a comonad is the decorated logic @xmath12 for a monad . thus , the grammar of @xmath12 is the same as the grammar of @xmath11 , but a term with decoration @xmath17 is now called a _ constructor_. the rules for @xmath12 are nearly the same as the corresponding rules for @xmath11 , except that for weak equations the replacement rule always holds while the substitution rule holds only when the substituted term is pure : @xmath73 in the rules for pairs and copairs , the decorations are permuted . the logic @xmath12 can be interpreted dually to @xmath11 . let @xmath37 be a bicartesian category and @xmath0 a monad on @xmath37 satisfying the mono requirement , which means that @xmath74 is a monomorphism for each object @xmath40 . then we get a model @xmath75 of the decorated logic @xmath12 , where a constructor @xmath44 is interpreted as a morphism @xmath76 in @xmath37 and a weak equation @xmath60 is interpreted as an equality @xmath77 in @xmath37 . let us consider a distributive category @xmath37 with epimorphic projections and with a distinguished object @xmath7 called the _ object of states_. we consider the comonad @xmath38 with endofunctor @xmath78 , with counit @xmath79 made of the projections @xmath80 , and with comultiplication @xmath81 which `` duplicates '' the states , in the sense that @xmath82 where @xmath83 is the projection . we call this comonad the _ comonad of state_. it is sometimes called the _ product comonad _ , and it is different from the _ costate comonad _ or _ store comonad _ with endofuntor @xmath84 @xcite . the category @xmath37 with the comonad of states provides a model of the logic @xmath11 . we can extend @xmath11 into a logic @xmath85 dedicated to the state comonad . first , because of the specific choice of the comonad @xmath78 , we can add new decorations to the rule patterns for pairs in @xmath11 , involving modifiers : there is a _ left pair _ @xmath86 of an accessor @xmath87 and a modifier @xmath88 , satisfying the first three rules in figure [ fig : state - prod ] . there are also three rules ( omitted ) , symmetric to these ones , for the _ right pair _ @xmath89 of a modifier @xmath90 and an accessor @xmath91 . the interpretation of the left pair @xmath92 is the pair @xmath93 of @xmath94 and @xmath95 . moreover , the rule expresses the fact that , when @xmath78 , two modifiers coincide as soon as they return the same result and modify the state in the same way . @xmath96_l}^{{(2)}}\colon a \to b_1\times b_2 } \\ { \textrm{(l - pair - eq ) } } & \dfrac { f_1^{{(1)}}\colon a\to b_1 \quad f_2^{{(2)}}\colon a \to b_2 } { { \mathit{pr}}_1^{{(0)}}\circ { \langle f_1,f_2 \rangle_l}^{{(2)}}{\sim}f_1^{{(1)}}\quad { \mathit{pr}}_2^{{(0)}}\circ { \langle f_1,f_2 \rangle_l}^{{(2)}}{\cong}f_2^{{(2 ) } } } \\ { \textrm{(l - pair - u ) } } & \dfrac{g^{{(2)}}{\!\colon\!}a { \!\to\!}b_1{\!\times\!}b_2 { \;\;}f_1^{{(1)}}{\!\colon\!}a{\!\to\!}b_1 \quad f_2^{{(2)}}{\!\colon\!}a { \!\to\!}b_2 { \;\;}{\mathit{pr}}_1^{{(0)}}\circ g { \sim}f_1 { \;\;}{\mathit{pr}}_2^{{(0)}}\circ g { \cong}f_2 } { g^{{(2)}}{\cong}{\langle f_1,f_2 \rangle_l}^{{(2 ) } } } \\ \hline { \textrm{(effect ) } } & \dfrac{f , g\colon a \to b \quad f{\sim}g \quad { \langle \ ; \rangle}_a\circ f { \cong}{\langle \ ; \rangle}_a\circ g } { f{\cong}g } \\ \hline \end{array}\ ] ] for each set @xmath97 of _ locations _ ( or identifiers ) , additional grammar and rules for the logic @xmath85 are given in figure [ fig : state - lookup ] . we extend the grammar of @xmath11 with a type @xmath98 , an accessor @xmath99 and a modifier @xmath100 for each location @xmath101 , and we also extend its rules . the rule asserts that two functions without result coincide as soon as they coincide when observed at each location . together with the rule it implies that two functions coincide as soon as they return the same value and coincide on each location . @xmath102 for each family of objects @xmath103 in @xmath37 such that @xmath104 we build a model @xmath105 of @xmath85 , which extends the model the model @xmath41 of @xmath11 with functions for looking up and updating the locations . the types @xmath98 are interpreted as the objects @xmath98 and the accessors @xmath99 as the projections from @xmath7 to @xmath98 . then the interpretation of each modifier @xmath106 is the function from @xmath107 to @xmath7 defined as the tuple of the functions @xmath108 where @xmath109 is the projection from @xmath107 to @xmath98 and @xmath110 is made of the projection from @xmath107 to @xmath7 followed by @xmath111 when @xmath112 . the logic we get , and its model , are essentially the same as in @xcite : thus , the pattern for a comonad in section [ sec : patterns ] can be seen as a generalization to arbitrary comonads of the approach in @xcite . since we have assumed that the category @xmath37 is _ distributive _ we get new decorations for the rule patterns for coproducts : the copair of two modifiers now exists , the corresponding decorated rules are given in figure [ fig : state - coprod ] . the interpretation of the modifier @xmath72}$ ] , when both @xmath113 and @xmath114 are modifiers , is the composition of @xmath72}\colon ( a_1\times s)+(a_2\times s)\to b\times s$ ] with the inverse of the canonical morphism @xmath115 : this inverse exists because @xmath37 is distributive . @xmath116}^{{(2)}}\colon a_1\!+\!a_2 \to b } \\ { \textrm{(copair - eq ) } } & \dfrac { f_1^{{(2)}}\colon a_1 \to b { \;\;}f_2^{{(2)}}\colon a_2 \to b } { { [ f_1|f_2 ] } \circ { \mathit{in}}_1 { \cong}f_1 \quad { [ f_1|f_2 ] } \circ { \mathit{in}}_2 { \cong}f_2 } \\ { \textrm{(copair - u ) } } & \dfrac{f_i^{{(2)}}{\!\colon\!}a_i { \!\to\!}b { \;\;}g^{{(2)}}{\!\colon\!}a_1\!{\!+\!}\!a_2 { \!\to\!}b { \;\;}g\circ { \mathit{in}}_i { \cong}f_i } { g^{{(2)}}{\cong } { [ f_1|f_2 ] } ^{{(2 ) } } } \\ \hline \end{array}\ ] ] to conclude with states , let us look at the constructions for conditionals and binary operations in the language for states . the rules in figure [ fig : state - coprod ] provide conditionals . there is no binary product of modifiers , but there is a left product of a constructor and a modifier and a right product of a modifier and a constructor . it follows that the _ left and right sequential products _ of two modifiers @xmath113 and @xmath114 can be defined , as in @xcite , by composing , e.g. , the left product of an identity and @xmath113 with the right product of @xmath114 and an identity . a major feature of this approach is that , for states , sequential products are defined without any new ingredient : no kind of strength , in contrast with the approach using the strong monad of states @xmath117 @xcite , no `` external '' decoration for equations , in contrast with @xcite . this property is due to the introduction of the intermediate notion of _ accessors _ between pure terms ( or _ values _ ) and modifiers ( or _ computations _ ) . now we use the decorated logic @xmath85 for proving that the decorated theory for states is hilbert - post complete . this result is proved in ( * ? ? ? * prop.2.40 ) in the framework of lawvere theories . here we give a proof in the decorated logic for states . this proof has been checked in coq . the logic we use is the fragment @xmath118 of @xmath85 which involves neither products nor coproducts nor the empty type ( but which involves the unit type ) . the _ theory of state _ , denoted @xmath119 , is the family of equations which may be derived from the axioms of @xmath118 using the rules of @xmath118 . more generally , a _ theory _ @xmath120 with respect to @xmath118 is a family of equations between terms of @xmath118 which is saturated with respect to the rules of @xmath118 . a theory @xmath121 is an _ extension _ of a theory @xmath120 if it contains all the equations of @xmath120 . two families of equations are called _ equivalent _ if each one can be derived from the other with the rules of @xmath118 . as in ( * prop.2.40 ) , for the sake of simplicity it is assumed that there is a single location @xmath101 , and we write @xmath122 , @xmath123 and @xmath124 instead of @xmath98 , @xmath125 and @xmath126 . then there is a single axiom @xmath127 . in addition , it is assumed that all types are _ inhabited _ , in the sense that for each type @xmath101 there exists a closed pure term with type @xmath101 . [ theo : equations ] every equation between terms of @xmath118 is equivalent to four equations between pure terms . the proof is obtained by merging the two parts of proposition [ prop : equations ] , which is proved in appendix [ app - complete ] . [ prop : equations ] 1 . [ prop : equations - acc ] every equation between accessors is equivalent to two equations between pure terms . [ prop : equations - modi ] every equation between modifiers is equivalent to two equations between accessors . roughly speaking , a theory ( with respect to some logic ) is said _ syntactically complete _ if no unprovable axiom can be added to the theory without introducing an inconsistency . more precisely , a theory with respect to the equational logic is _ hilbert - post complete _ if it is consistent and has no consistent proper extension ( * ? ? ? * definition 2.8 . ) . since we use a decorated version of the equational logic , we have to define a decorated version of hilbert - post completeness . [ defi : complete ] with respect to the logic @xmath118 , a theory @xmath120 is _ consistent _ if there is an equation which is not in @xmath120 . an extension @xmath121 of a theory @xmath120 is a _ pure extension _ if it is generated by @xmath120 and by equations between pure terms . it is a _ proper extension _ if it is not a pure extension . a theory @xmath120 is _ hilbert - post complete _ if it is consistent and has no consistent proper extension . the proof of theorem [ theo : complete ] relies on theorem [ theo : equations ] . we do not have to assume that the interpretation of the type @xmath122 is a countable set . we have assumed that @xmath97 is a singleton , but we conjecture that our result can be generalized to any set of locations , without any finiteness condition . [ theo : complete ] the theory for state is hilbert - post complete . the theory @xmath119 is consistent : it can not be proved that @xmath128 . let us consider an extension @xmath120 of @xmath119 and let @xmath129 be the theory generated by @xmath119 and by the equations between pure terms in @xmath120 . thus , @xmath129 is a pure extension of @xmath119 and @xmath120 is an extension of @xmath129 . let us consider an arbitrary equation @xmath130 in @xmath120 , according to theorem [ theo : equations ] we get a family @xmath9 of equations between pure terms which is equivalent to the given equation @xmath130 . since @xmath130 is in @xmath120 and @xmath120 is saturated , the equations in @xmath9 are also in @xmath120 , hence they are in @xmath129 . since @xmath9 is in @xmath129 and @xmath129 is saturated , the equation @xmath130 is also in @xmath129 . this proves that @xmath131 , so that the theory @xmath119 has no proper extension . let us consider a bicartesian category @xmath37 with monomorphic coprojections and with a distinguished object @xmath9 called the _ object of exceptions_. we do not assume that @xmath37 is distributive ( it would not help ) nor codistributive , because usually this is not the case . the _ monad of exceptions _ on @xmath37 is the monad @xmath0 with endofunctor @xmath132 , its unit @xmath133 is made of the coprojections @xmath134 , and its multiplication @xmath135 `` merges '' the exceptions , in the sense that @xmath136}\colon ( a+e)+e\to a+e$ ] where @xmath137 is the coprojection . it satisfies the mono requirement because the coprojections are monomorphisms . thus , the category @xmath37 with the monad of exceptions provides a model of the logic @xmath12 . the name of the decorations is adapted to the monad of exceptions : a constructor is called a _ propagator _ : it may raise an exception but can not recover from an exception , so that it has to propagate all exceptions ; a modifier is called a _ catcher_. for this specific monad @xmath132 , it is possible to extend the logic @xmath12 as @xmath138 , called the _ logic for exceptions _ , so that @xmath37 with @xmath132 can be extended as a model @xmath139 of @xmath138 . first , dually to the left and right pairs for states in figure [ fig : state - prod ] , we get new decorations to the rule patterns for copairs in @xmath12 , involving modifiers , as in figure [ fig : exc - coprod ] for the left copairs ( the rules for the right copairs are omitted ) . the interpretation of the left copair @xmath72_l}^{{(2 ) } } : a_1+a_2\to b$ ] is the copair @xmath72}:a_1+a_2+e\to b+e$ ] of @xmath140 and @xmath141 in @xmath37 . for instance , the coproduct of @xmath142 , with coprojections @xmath143 and @xmath35}_a^{{(0)}}:{\mathbb{0}}\to a$ ] , gives rise to the left copair @xmath72_l}^{{(2)}}:a \to b$ ] of any constructor @xmath144 with any modifier @xmath145 , which is characterized up to strong equations by @xmath72_l}{\sim}f_1 $ ] and @xmath72_l}{\cong}f_2 $ ] . this will be used in the construction of the @xmath146 expressions . moreover , the rule expresses the fact that , when @xmath132 , two modifiers coincide as soon as they coincide on ordinary values and on exceptions . @xmath147_l}^{{(2)}}\colon a_1\!+\!a_2 \to b } \\ { \textrm{(l - copair - eq ) } } & \dfrac { f_1^{{(1)}}\colon a_1 \to b { \;\;}f_2^{{(2)}}\colon a_2 \to b } { { [ f_1|f_2 ] _ l}^{{(2)}}\circ { \mathit{in}}_1^{{(0)}}{\sim}f_1^{{(1)}}\quad { [ f_1|f_2 ] _ l}^{{(2)}}\circ { \mathit{in}}_2^{{(0)}}{\cong}f_2^{{(2 ) } } } \\ { \textrm{(l - copair - u ) } } & \dfrac{g^{{(2)}}{\!\colon\!}a_1\!+\!a_2 { \!\to\!}b { \;\;}f_1^{{(1)}}{\!\colon\!}a_1 { \!\to\!}b { \;\;}f_2^{{(2)}}{\!\colon\!}a_2 { \!\to\!}b { \;\;}g\circ { \mathit{in}}_1 { \sim}f_1 { \;\;}g\circ { \mathit{in}}_2 { \cong}f_2 } { g^{{(2)}}{\cong } { [ f_1|f_2 ] _ l}^{{(2 ) } } } \\ \hline { \textrm{(effect ) } } & \dfrac{f , g\colon a \to b \quad f{\sim}g \quad f\circ { [ \;]}_a { \cong}g\circ { [ \;]}_a } { f{\cong}g } \\ \hline \end{array}\ ] ] for each set @xmath148 of _ exception names _ , additional grammar and rules for the logic @xmath138 are given in figure [ fig : exc - tag ] . we extend the grammar of @xmath12 with a type @xmath149 , a propagator @xmath150 and a catcher @xmath151 for each exception name @xmath152 , and we also extend its rules . the logic @xmath138 obtained performs the _ core _ operations on exceptions : the _ tagging _ operations encapsulate an ordinary value into an exception , and the _ untagging _ operations recover the ordinary value which has been encapsulated in an exception . this may be generalized by assuming a hierarchy of exception names @xcite . the rule asserts that two functions without argument coincide as soon as they coincide on each exception . together with the rule it implies that two functions coincide as soon as they coincide on their argument and on each exception . @xmath153}_{v_t}\circ{\mathtt{tag}}_r } \\ { \textrm{(local - global ) } } & \dfrac { f , g\colon { \mathbb{0}}\to b \quad \mbox{for all } t\in { \mathit{exn}}\ ; f\circ { \mathtt{tag}}_t { \sim}g\circ { \mathtt{tag}}_t } { f{\cong}g } \\ \hline \end{array}\ ] ] for each family of objects @xmath154 in @xmath37 such that @xmath155 we build a model @xmath139 of @xmath138 , which extends the model the model @xmath156 of @xmath12 with functions for tagging and untagging the exceptions . the types @xmath149 are interpreted as the objects @xmath149 and the propagators @xmath150 as the coprojections from @xmath149 to @xmath9 . then the interpretation of each catcher @xmath151 is the function @xmath157 defined as the cotuple ( or case distinction ) of the functions @xmath158 where @xmath159 is the coprojection of @xmath149 in @xmath160 and @xmath161 is made of @xmath162 followed by the coprojection of @xmath9 in @xmath160 when @xmath163 . this can be illustrated , in an informal way , as follows : @xmath164 encloses its argument @xmath165 in a box with name @xmath152 , while @xmath166 opens every box with name @xmath152 to recover its argument and returns every box with name @xmath167 without opening it : @xmath168^{{\mathtt{tag}}_t } & & * + [ f-]{a } \ar@{}[r]_(.2){t } & & * + [ f-]{a } \ar@{}[r]_(.2){t } \ar[rr]^{{\mathtt{untag}}_t } & & a \\ & & & & & * + [ f-]{a } \ar@{}[r]_(.2){r } \ar[rr]^{{\mathtt{untag}}_t } & & * + [ f-]{a } \ar@{}[r]_(.2){r } & \\ } \ ] ] since we did not assume that the category @xmath37 is codistributive we can not get products of modifiers in a way dual to the coproducts of modifiers for states . however these rules have not been used for proving the hilbert - post completeness of the theory for state . thus by duality from theorem [ theo : complete ] we get `` for free '' a result about the core language for exceptions . [ coro : complete ] the core theory for exceptions is hilbert - post complete . we have obtained a logic @xmath138 for exceptions , with the core operations for tagging and untagging . this logic provides a direct access to catchers ( the untagging functions ) , which is not provided by the usual mechanism of exceptions in programming languages . in fact the core operations remain _ private _ , while there is a _ programmer s _ language , which is _ public _ , with no direct access to the catchers . the programmer s language for exceptions provides the operations for _ raising _ and _ handling _ exceptions , which are defined in terms of the core operations . this language has no catcher : the only way to catch an exception is by using a @xmath146 expression , which itself propagates exceptions . thus , all terms of the programmer s language are propagators . this language does not include the private tagging and untagging operations , but the public @xmath169 and @xmath146 constructions , which are defined in terms of @xmath170 and @xmath171 . for the sake of simplicity we assume that only one type of exception is handled in a @xmath146 expression , the general case is treated in @xcite . the main ingredients for building the programmer s language from the core language are the coproducts @xmath172 and a new conversion rule for terms . the _ downcast _ conversion of a catcher to a propagator could have been defined in section [ sec : patterns ] for the logic @xmath11 , and dually for the logic @xmath12 ; the rule is : @xmath173 this downcasting conversion from catchers to propagators is interpreted by mapping @xmath174 to @xmath175 . it is related to weak equations : @xmath176 , and @xmath177 if and only if @xmath178 . but the downcasting conversion is _ unsafe _ : several catchers may be downcasted to the same propagator . this powerful operation turns an effectful term to an effect - free one ; since it is not required for states nor for the core language for exceptions , we did not introduce it earlier . [ defi : exc ] for each type @xmath179 and each exception name @xmath152 , the propagator @xmath180 is : @xmath181}_b^{{(0)}}\circ { \mathtt{tag}}_t^{{(1)}}\colon v_t\to b\ ] ] for each each propagator @xmath182 , each exception name @xmath152 and each propagator @xmath183 , the propagator @xmath184 is defined in three steps , involving two catchers @xmath185 and @xmath186 , as follows : @xmath187}_b^{{(0)}}\ ; ] } ^{{(1)}}\circ{\mathtt{untag}}_t^{{(2)}}\colon { \mathbb{0}}\to b \\ { \mathtt{try}}(f){\mathtt{catch}}(t{\rightarrow}g)^{{(2)}}= { [ \ ; { \mathit{id}}_b \;|\ ; { \mathtt{catch}}(t{\rightarrow}g ) \ ; ] _ l}^{{(2)}}\circ f^{{(1)}}\colon a\to b \\ { \mathtt{try}}(f){\mathtt{catch}}(t{\rightarrow}g)^{{(1)}}= { { \downarrow}}({\mathtt{try}}(f){\mathtt{catch}}(t{\rightarrow}g ) ) \colon a\to b \\ \end{array}\ ] ] this means that raising an exception with name @xmath152 consists in tagging the given ordinary value ( in @xmath149 ) as an exception and coerce it to any given type @xmath179 . for handling an exception , the intermediate expressions @xmath188 and @xmath189 are private catchers and the expression @xmath190 is a public propagator : the downcast operator prevents it from catching exceptions with name @xmath152 which might have been raised before the @xmath190 expression is considered . the definition of @xmath190 corresponds to the java mechanims for exceptions @xcite . the definition of @xmath190 corresponds to the following control flow , where ` exc ? ` means `` _ _ is this value an exception ? _ _ '' , an _ abrupt _ termination returns an uncaught exception and a _ normal _ termination returns an ordinary value ; this corresponds , for instance , to the java mechanims for exceptions @xcite . @xmath191 & & \\ & { \txt{exc?}}\ar[ld]_{y}\ar[rd]^{n } & & \\ { \boxed{\mathit{abrupt } } } & & f^{{(1)}}\ar[d ] & \\ & & { \txt{exc?}}\ar[ld]_{y}\ar[rd]^{n } & \\ & { \mathtt{untag}}_t^{{(2)}}\ar[d ] & & { \boxed{\mathit{normal}}}\\ & { \txt{exc?}}\ar[ld]_{y}\ar[rd]^{n } & & \\ { \boxed{\mathit{abrupt } } } & & g^{{(1)}}\ar[d ] & \\ & & \txt{{\boxed{\mathit{normal}}}\mbox { or } { \boxed{\mathit{abrupt } } } } \\ } \ ] ] to conclude with exceptions , let us look at the constructions for case distinction and binary operations in the programmer s language for exceptions , which means , copairs and pairs of _ constructors_. the general rules of the logic @xmath12 include coproducts of constructors ( figure [ fig : pattern - rules ] ) , which provide case distinction for all terms in the programmer s language for exceptions . but the general rules for a monad do not include binary products involving a constructor , hence they can not be used for dealing with binary operations in the programmer s language for exceptions when at least an argument is not pure . indeed , if @xmath65 is pure and @xmath192 does raise an exception , it is in general impossible to find @xmath193 such that @xmath194 and @xmath195 . however , there are several ways to formalize the fact of first evaluating @xmath113 then @xmath114 : for instance by using a strong monad @xcite , or a sequential product @xcite , or productors @xcite . the sequential product approach can be used in our framework ; it requires the introduction of a third kind of `` equations '' , in addition to the strong and weak equations , which corresponds to the usual order between partial functions : details are provided in @xcite . we have presented two patterns giving sound inference systems for effects arising from a monad or a comonad . we also gave detailed examples of applications of these patterns to the state and the exceptions effects . the obtained decorated proof system for states has been implemented in coq , so that the given proofs can be automatically verified . we plan to adapt this logic to local states ( with allocation ) in order to provide a decorated proof of the completeness theorem in @xcite . from this implementation , we plan to extract the generic part corresponding to the comonad pattern , dualize it and extend it to handle the programmer s language for exceptions . then a major issue is scalability : how can we combine effects ? within the framework of this paper , it may seem difficult to guess how several effects arising from either monads or comonads can be combined . however , as mentioned in the introduction , this paper deals with two patterns for instanciating the more general framework of decorated logics @xcite . decorated logics are based on spans in a relevant category of logics , so that the combination of effects can be based on the well - known composition of spans . [ [ acknowledgment . ] ] acknowledgment . + + + + + + + + + + + + + + + we are grateful to samuel mimram for enlightning discussions . 00 csar domnguez , dominique duval . diagrammatic logic applied to a parameterization process . mathematical structures in computer science 20 , p. 639 - 654 ( 2010 ) . jean - guillaume dumas , dominique duval , laurent fousse , jean - claude reynaud . decorated proofs for computational effects : states . accat 2012 . electronic proceedings in theoretical computer science 93 , p. 45 - 59 ( 2012 ) . jean - guillaume dumas , dominique duval , laurent fousse , jean - claude reynaud . a duality between exceptions and states . mathematical structures in computer science 22 , p. 719 - 722 ( 2012 ) . jean - guillaume dumas , dominique duval , jean - claude reynaud . cartesian effect categories are freyd - categories . journal of symbolic computation 46 , p. 272 - 293 ( 2011 ) . jean - guillaume dumas , dominique duval , jean - claude reynaud . a decorated proof system for exceptions . arxiv:1310.2338 ( 2013 ) . jean - guillaume dumas , dominique duval , burak ekici , damien pous . formal verification in coq of program properties involving the global state effect . arxiv:1310.0794 ( 2013 ) . jeremy gibbons , michael johnson . relating algebraic and coalgebraic descriptions of lenses bx 2012 . eceasst 49 ( 2012 ) . james gosling , bill joy , guy steele , gilad bracha . the java language specification , third edition . addison - wesley longman ( 2005 ) . bart jacobs . a formalisation of java s exception mechanism . esop 2001 . lncs , vol . 2028 , p. 284 - 301 springer ( 2001 ) . bart jacobs and jan rutten . an introduction to ( co)algebras and ( co)induction . in : d. sangiorgi and j. rutten ( eds ) , advanced topics in bisimulation and coinduction , p.38 - 99 , 2011 . paul blain levy . monads and adjunctions for global exceptions . mfps 2006 . electronic notes in theoretical computer science 158 , p. 261 - 287 ( 2006 ) . john m. lucassen , david k. gifford . polymorphic effect systems . popl 1988 . acm press , p. 47 - 57 . paul - andr mellis . segal condition meets computational effects . lics 2010 . p. 150 - 159 , ieee computer society ( 2010 ) . eugenio moggi . notions of computation and monads . information and computation 93(1 ) , p. 55 - 92 ( 1991 ) . eugenio moggi and sonia fagorzi . a monadic multi - stage metalanguage . fossacs 2003 , lncs , vol . 2620 , p. 358 - 374 , springer ( 2003 ) . matija pretnar . the logic and handling of algebraic effects . university of edinburgh 2010 . gordon d. plotkin , john power . notions of computation determine monads . fossacs 2002 . lncs , vol . 2620 , p. 342 - 356 , springer ( 2002 ) . gordon d. plotkin , matija pretnar . handlers of algebraic effects . esop 2009 . lncs , vol . 5502 , p. 80 - 94 , mpringer ( 2009 ) . lutz schrder , till mossakowski . generic exception handling and the java monad . amast 2004 . lncs , vol . 3116 , p. 443 - 459 , springer ( 2004 ) . sam staton . completeness for algebraic theories of local state . fossacs 2010 . lncs , vol . 6014 , p. 48 - 63 , springer ( 2010 ) . ross tate . the sequential semantics of producer effect systems . popl 2013 . acm press , p. 15 - 26 ( 2013 ) . tarmo uustalu , varmo vene . comonadic notions of computation . cmcs 2008 . entcs 203 , p. 263 - 284 ( 2008 ) . philip wadler . the essence of functional programming . popl 1992 . acm press , p. 1 - 14 the logic used in this appendix is the fragment @xmath118 of the decorated logic for states @xmath85 which involves neither products nor coproducts nor the empty type , but which involves the unit type . for the sake of simplicity it is assumed that there is a single location @xmath101 , and we write @xmath122 , @xmath196 and @xmath197 instead of @xmath98 , @xmath198 and @xmath199 . then there is a single axiom @xmath200 . in section [ sec : states ] , the proof of hilbert - post completeness in theorem [ theo : complete ] relies on proposition [ prop : equations ] , which is restated here as proposition [ prop : equations - app ] . the aim of this appendix is to prove proposition [ prop : equations - app ] . 1 . [ lemm : eqns - rule - unit ] @xmath201 2 . [ lemm : eqns - rule - val ] @xmath202 3 . [ lemm : eqns - strong ] + @xmath203 4 . [ lem : pure_mid_id ] @xmath204 5 . [ lemm : onepureacc ] @xmath205 6 . [ lemm : lkppureequal ] @xmath206 7 . [ lemm : lkpuseless ] @xmath207 1 . consequence of the observational rule with only one location . 2 . consequence of [ lemm : eqns - rule - unit ] applied to @xmath208 : indeed , from the axiom @xmath209 we get @xmath210 . 3 . from axiom @xmath209 by substitution we get @xmath211 ; thus , point [ lemm : eqns - rule - unit ] implies @xmath212 . 4 . from @xmath213 , as @xmath214 is pure , by the weak replacement we have @xmath215 . then , weak substitution with @xmath165 yields @xmath216 . 5 . we know that @xmath217 . + it follows that @xmath218 . let @xmath219 . composing with @xmath197 we get @xmath220 . using the axiom @xmath209 and the replacement rule for @xmath21 , which can be used here because both @xmath221 and @xmath214 are pure , we get @xmath222 . since weak and strong equations coincide on pure terms we get @xmath223 . let @xmath224 . by point [ lemm : onepureacc ] above we get @xmath225 , thus @xmath226 . then by point [ lemm : lkppureequal ] above we get @xmath227 . now , let us prove proposition [ prop : canonical - form ] , which says that , up to strong equations , it can be assumed that there is at most one occurrence of @xmath196 in any accessor and at most one occurrence of @xmath197 in any modifier . 1 . for each accessor @xmath228 , if @xmath165 is not pure then there is a pure term @xmath229 such that @xmath230 2 . for each modifier @xmath231 , if @xmath29 is not an accessor then there is an accessor @xmath232 and a pure term @xmath233 such that @xmath234 1 . if @xmath228 is not pure then it contains at least one occurrence of @xmath235 . thus , it can be written in a unique way as @xmath236 for some pure term @xmath229 and some accessor @xmath237 . since @xmath237 is such that @xmath238 , the result follows . if @xmath231 is not an accessor then it contains at least one occurrence of @xmath239 . thus , it can be written in a unique way as @xmath240 for some accessor @xmath241 and some modifier @xmath242 . from point [ prop : canonical - form - acc ] , we also have that @xmath243 for some pure term @xmath229 so that @xmath244 . * if @xmath113 is an accessor , the result follows with @xmath245 . * otherwise , @xmath90 contains at least one occurrence of @xmath239 . thus , it can be written in a unique way as @xmath246 for some accessor @xmath247 and some modifier @xmath248 . according to point [ prop : canonical - form - acc ] applied to the accessor @xmath249 , either @xmath249 is pure or @xmath250 for some pure term @xmath251 * * if @xmath250 then @xmath252 . the axiom @xmath253 and the replacement and substitution rules for @xmath21 ( since @xmath254 is pure ) yield @xmath255 . then it follows from point [ lemm : eqns - rule - val ] in lemma [ lemm : eqns ] that @xmath256 , and since @xmath257 we get @xmath258 . the result follows by induction on the number of occurrences of @xmath197 in @xmath29 : indeed , there is one less occurrence of @xmath197 in @xmath259 than in @xmath260 . * * if @xmath249 is pure then @xmath261 from point [ lemm : onepureacc ] in lemma [ lemm : eqns ] . thus the previous proof applies by replacing @xmath249 with @xmath262 . 1 . * if @xmath267 is pure , since @xmath268 ( because @xmath269 is an accessor ) we get @xmath270 , thus the result is obtained with @xmath271 . * otherwise , we have just proved that @xmath272 with @xmath273 , then @xmath274 and @xmath275 . if @xmath276 is an accessor , since @xmath277 we get @xmath278 , thus the result is obtained with @xmath279 . * otherwise , we have just proved that @xmath280 with @xmath281 , then @xmath282 and @xmath283 . we can now prove proposition [ prop : equations - app ] on which the hilbert - post completeness theorem relies . this proof has been checked with the coq proof assistant using the system for states of @xcite . the coq library with the inference system is available there : http://coqeffects.forge.imag.fr . the single proof of the following proposition ( roughly 16 pages in coq ) is directly available there : http://coqeffects.forge.imag.fr/hpcompletecoq.v . 1 . we prove that for any accessors @xmath290 there are three cases : 1 . either they are both pure and @xmath291 is the required equation between pure terms . 2 . either they are both accessors and it can be derived from @xmath292 that @xmath293 for some pure terms @xmath294 . 3 . or one of them is pure and the other one is an accessor and it can be derived from @xmath292 that @xmath293 and @xmath295 for some pure terms @xmath294 and @xmath296 . + we prove , moreover , that the converse also hold . 1 . as already mentioned , if @xmath297 and @xmath298 are both pure and @xmath291 is the required equation between pure terms . 2 . if neither @xmath297 nor @xmath298 is pure , then according to proposition [ prop : canonical - form ] @xmath299 and @xmath300 for some pure terms @xmath301 . * starting from the equation @xmath302 we thus get @xmath303 . then , using the assumption , for any function @xmath289 , we have that @xmath304 . now @xmath305 . this , together with the axiom @xmath306 and the replacement rule for @xmath21 ( which can be used here because both @xmath254 and @xmath307 are pure ) yield @xmath308 . as the latter are both pure terms we also have @xmath309 . * conversely , if @xmath310 then @xmath311 , which means that @xmath312 . the only remaining case is w.l.o.g . if @xmath297 is pure and @xmath298 is not . * then @xmath313 from proposition [ prop : canonical - form ] as previously and @xmath314 satisfies @xmath315 for any assumed @xmath289 . indeed from @xmath316 we get @xmath317 but , on the one hand , @xmath318 so that point [ lemm : onepureacc ] in lemma [ lemm : eqns ] gives @xmath319 with @xmath320 . on the other hand , @xmath321 so that @xmath322 . thus equation ( [ eq : fullpureacc ] ) rewrites as @xmath323 and point [ lemm : lkpuseless ] in lemma [ lemm : eqns ] yields @xmath324 thus now we also have @xmath325 . from the original equation @xmath326 we finally get @xmath327 * conversely , we start from @xmath328 and @xmath329 satisfying both equations ( [ eq : pureaccfirst ] ) and ( [ eq : pureaccsecond ] ) . then , we define @xmath330 which satisfies @xmath331 thanks to equation ( [ eq : pureaccfirst ] ) . the latter is also @xmath332 which is thus @xmath333 thanks to equation ( [ eq : pureaccsecond ] ) . the rule for states means that two modifiers coincide as soon as they return the same result and modify the state in the same way . this means that @xmath334 if and only if @xmath335 and @xmath336 . thanks to corollary [ coro : downcast ] the equation @xmath335 is equivalent to an equation between accessors . it remains to prove that the equation @xmath336 is also equivalent to an equation between accessors . + for @xmath337 , since @xmath338 , proposition [ prop : canonical - form ] says that @xmath339 for some accessor @xmath340 . thus , @xmath336 if and only if @xmath341 . let us check that this equation is equivalent to @xmath291 . + clearly if @xmath342 then @xmath341 . conversely , if @xmath343 then @xmath344 and since @xmath253 we get @xmath345 , which is the same as @xmath291 because @xmath297 and @xmath298 are accessors . + thus , @xmath336 if and only if @xmath291 , as required .
this paper presents equational - based logics for proving first order properties of programming languages involving effects . we propose two dual inference system patterns that can be instanciated with monads or comonads in order to be used for proving properties of different effects . the first pattern provides inference rules which can be interpreted in the kleisli category of a monad and the cokleisli category of the associated comonad . in a dual way , the second pattern provides inference rules which can be interpreted in the cokleisli category of a comonad and the kleisli category of the associated monad . the logics combine a 3-tier effect system for terms consisting of pure terms and two other kinds of effects called constructors / observers and modifiers , and a 2-tier system for up - to - effects and strong equations . each pattern provides generic rules for dealing with any monad ( respectively comonad ) , and it can be extended with specific rules for each effect . the paper presents two use cases : a language with exceptions ( using the standard monadic semantics ) , and a language with state ( using the less standard comonadic semantics ) . finally , we prove that the obtained inference system for states is hilbert - post complete .
a sketch of the nb / gaas / algaas structure is shown in fig . the semiconductor portion consists of a 1 @xmath0 m thick @xmath1-gaas(001 ) buffer layer si - doped nominally at @xmath2 @xmath3 grown by molecular beam epitaxy ( mbe ) on a @xmath1-gaas(001 ) substrate , followed by a 4 nm thick al@xmath4ga@xmath5as barrier . this was followed by the growth of a 12 nm thick gaas(001 ) epilayer si - doped at @xmath2 @xmath3 and by 14 nm of gaas doped by a sequence of six si @xmath6-doped layers spaced by 2 nm . a 1-@xmath0m - thick amorphous as cap layer was deposited in the mbe growth chamber to protect the surface during transfer in air to an uhv sputter - deposition / surface analysis system . after thermal desorption of the as protective cap layer , a 100-nm - thick nb film was then deposited _ in situ _ by dc - magnetron sputtering . the thickness of the gaas epilayer sandwiched between the superconductor and the algaas barrier was selected in order to have an experimentally - accessible single quasi - bound state below the superconductive gap , and the si @xmath6-doped layers at the nb / gaas interface were employed to achieve the required transmissivity . more detail on the contact fabrication procedure is reported elsewhere @xcite where the reference sn junctions are also described . the latter consisted of a nb/@xmath6-doped - gaas junction _ without _ the algaas insulating barrier . a qualitative sketch of the energy - band diagram of our snin structure is depicted in fig . 100@xmath7160@xmath0m@xmath8 junctions were defined by standard photolithographic techniques and reactive ion etching . 4-wire measurements were performed with two leads on the junction under study and the other two connected to the sample back contact . the measured differential conductances @xmath9 bias ( @xmath10 ) for the resonant structure ( snin structure , panel ( a ) ) and for the reference junction ( sn structure , panel ( b ) ) at @xmath11 k are shown in fig . comparison of the two characteristics clearly shows the presence of a marked dgsj subgap conductance peak in the snin , resonant device . the resonance is superimposed to the typical andreev - dominated subgap conductance . the symmetry of conductance and the zero - bias conductance peak ( zbcp ) peculiar to reflectionless tunneling ( rt ) @xcite further demonstrate the effectiveness of the fabrication protocol . quantitative determination of the resonant transport properties in such a system is not trivial as it can be inferred by inspecting fig . 1(b ) and considering the diffusive nature of the normal regions . dgsj - enhancement , however , is an intrinsically ballistic phenomenon so that its essential features can be captured with relative ease . one study particularly relevant for our system was performed in ref.@xcite . in the context of ballistic transport a one - dimensional snin structure was studied as a function of the n interlayer thickness @xmath12 and it was demonstrated that resonances can occur in the subgap conductance spectrum for suitable geometric conditions . the insulating barrier was simulated by a @xmath6-like potential . customarily the barrier strength is described by the dimensionless coefficient @xmath13 , where @xmath14 is the electron fermi velocity @xcite and @xmath15 the barrier amplitude , and in the following we shall make use of it to characterize our system . the key - results of the analysis are : ( _ a _ ) the number of resonances increases for larger thickness @xmath12 of the metallic interlayer and is virtually independent of @xmath16 ; ( _ b _ ) the energy - width of such resonances decreases for increasing @xmath16 . this model is rather idealized , but is a useful tool to grasp the essential features of our system such as number and position of resonances @xcite . in order to apply it we must first determine parameters such as barrier strength , electron density and mean free path . also , any quantitative comparison with experiment requires us to estimate the sample series resistance , which influences the experimental energy position of the resonance peak . the above parameters can be obtained from an analysis of the rt - driven zbcp and from hall measurements . we performed hall measurements at 1.5 k and obtained carrier density @xmath17@xmath3 and mobility @xmath18@xmath19/v s. these data allow us to estimate the thermal coherence length @xmath20m@xmath21 , where @xmath22cm@xmath8/s is the diffusion constant , and the electron mean free path @xmath24 nm . @xmath25 compares favorably with the geometrical constraints of the structure ( @xmath26 nm @xmath27 ) and further supports our ballistic analysis . the zbcp can be described following the work by beenakker and co - workers @xcite . their model system consists of a normal disordered region of length @xmath28 and width @xmath29 separated by a finite - transparency barrier from a superconductor . under appropriate conditions a zbcp is predicted to occur , whose width @xmath30 ( i.e. , the voltage at which the zbcp is suppressed ) was estimated to be of the order of the thouless voltage , @xmath31 . similarly , upon application of a magnetic field , the zbcp is suppressed for fields of the order of @xmath32 . in real systems , at finite temperatures , @xmath28 and @xmath29 are to be replaced by @xmath33 , if it is smaller @xcite . at 0.34 m and in our junctions we calculate @xmath35@xmath0v . comparison with the experimental value @xmath36@xmath0v in fig . 2(c ) ( see solid line ) , allowed us to estimate the series resistance contribution to the measured conductivity . this rather large effect stems mainly from the algaas barrier and the sample back - contact resistance . we calculated the conductance for the nominal n - thickness value ( @xmath37 nm ) and @xmath38 , as appropriate for the algaas barrier @xcite . we emphasize , however , that the essential features such as the number and energy position of the dgsj resonances are virtually independent of z and are controlled instead by the value of @xmath12 . our calculations yield a single conductance peak at energies corresponding to about @xmath39 , where @xmath40 is the superconductor energy gap . by including the above - determined series resistance contribution , the resonance peak is positioned at about @xmath41mv ( see fig . 2(c ) , dash - dotted line , @xmath11 k ) . the corresponding experimental value is about 3 mv ( solid line in fig . 2(c ) ) , but the observed energy difference is well within the uncertainty resulting from the determination of the series resistance . the results of our model calculations strongly support our interpretation of the experimental structure in terms of dgsj resonant transport . ordinary resonant tunneling in the normal double - barrier potential schematically shown in fig . 1(b ) can not explain the observed subgap structure . this is indicated by the symmetry in the experimental data for positive and negative bias and is further proven by the temperature and magnetic field dependence of the differential conductance . figure 3 shows a set of @xmath10s measured in the 0.34 - 1.55 k range for the zbcp , ( a ) , and for the resonance peak , ( b ) . both effects show a strong dependence on temperature and at @xmath42 k are totally suppressed . this temperature value is within the range where rt suppression is expected @xcite . at higher temperatures , the conductance in the resonance region resembles that of the reference sn junction of fig . notably , the zbcp and the resonance peak disappear at the same temperature , hence indicating the coherent nature of the observed effect . further confirmation of the nature of the resonance peak can be gained observing its dependence from the magnetic field . figure 4 shows @xmath10 at @xmath43k for several values of the magnetic field applied in the plane of the junction for the zbcp , ( a ) , and for the resonance peak , ( b ) . the measurements confirm the known sensitivity of zbcp to the magnetic field @xcite , and clearly indicate that the dgsj resonance is easily suppressed for critical fields of the order of 100 mt . in the perpendicular field configuration at @xmath44k ( see fig . 4(c),(d ) ) the resonance and the zbcp displayed a similar dependence . such behavior is fully consistent with dgsj - related origin but is not compatible with a normal resonant tunneling description of the data @xcite . in summary , we have experimentally observed dgsj resonant states in nb / gaas / algaas hybrid heterostructures . transport was studied as a function of temperature and magnetic field and was successfully described within the ballistic model of riedel and bagwell@xcite . the present results suggest that the nb / gaas / algaas system is a good candidate for the implementation of complex mesoscopic structures that can take advantage of the mature algaas nanofabrication technology . such structures may represent ideal prototype systems for the study of coherent transport and the implementation of novel hybrid devices . this work was sponsored by infm under the pais project eiss . we confined our calculation to a one - dimensional case in light of the results by s. chaudhuri and p. f. bagwell , phys . b * 51 * , 16936 ( 1995 ) . these authors showed the insensitivity to dimensionality of the essential properties of transport resonances . resonances are determined by inspecting the @xmath10 behavior .
resonant transport in a hybrid semiconductor - superconductor microstructure grown by mbe on gaas is presented . this structure experimentally realizes the prototype system originally proposed by de gennes and saint - james in 1963 in _ all_-metal structures . a low temperature single peak superimposed to the characteristic andreev - dominated subgap conductance represents the mark of such resonant behavior . random matrix theory of quantum transport was employed in order to analyze the observed magnetotransport properties and ballistic effects were included by directly solving the bogoliubov - de gennes equations . in this contribution we report the experimental observation of de gennes - saint - james ( dgsj)-type resonances @xcite in a microstructure consisting of a nb / gaas / algaas / gaas hybrid microjunction . this microstructure realizes the model system introduced by de gennes and saint - james in 1963 , consisting of a superconductor - normal metal - insulator - normal metal - type ( snin ) layer sequence . multiple reflections off the superconductive gap ( i.e. , andreev reflections @xcite ) and off the insulating barrier ( i.e. , normal reflections ) can give rise to quasi - bound states @xcite that manifest themselves as conductance resonances . this result was made possible by the exploitation of semiconductor epitaxial growth to tailor electronic states according to the original dgsj model system . we analyzed the data within the random matrix theory of quantum transport and included ballistic effects by directly solving the bogoliubov - de gennes equations in a model potential profile @xcite . our description of the system is confirmed by the observed temperature and magnetic - field dependence .
an unequivocal prediction of cold dark matter models ( cdm ) is that the mass of the milky way , in the form of its dark matter halo , builds up hierarchically , by accretion of lower - mass halos . when these sub - systems ( often refereed to as subhalos ) escape the tidal disruption in the milky way halo they survive in the form of satellite galaxies until today . numerical simulations confirmed the theory predicting 500 satellites within 500 kpc from the milky way center ( kauffman et al . 1993 ; klypin et al . 1999 ; moore et al . the modest populations of observed dwarf galaxies orbiting the milky way and andromeda , however , seems to conflict with this prediction . indeed , in the current cdm paradigm of structure formation , dark matter satellites outnumber the known spheroidal by a factor of 10 to 100 . this discrepancy between the expected and observed numbers of dwarf galaxies has become widely known as the _ missing dwarf problem_. the newly discovered population of ultra - faint dwarfs around the milky way and m31 found in the sloan digital sky survey ( willman et al . 2005a , b ; belokurov et al . 2006 , zucker et al . 2006 ) increases of a factor of two in the number of known satellites . it is unclear , however , whether these new discovered low surface brightness satellite galaxies can reconcile the outnumber of subhalos predicted in cosmological simulations and solve the problem ( simon & geha 2007 ) . so far a number of mechanisms have been proposed to the substructure problem . cosmological solutions include modifying the power spectrum at small scales ( kamionkowski & liddle 2000 ; zentner & bullock 2003 ) and changing the nature of the dark matter particles , such as assuming a warm dark matter particle ( e.g. colin et al . 2000 ; avila - reese et al . 2001 ) or invoking a decay from a nonrelativistic particle ( strigari et al . alternative solutions typically appeal to feedback effects associated with stellar evolution or including heating from uv radiation in order to suppress the formation of dwarf galaxies by preventing low mass dark matter halos from acquiring enough gas to form stars ( e.g. , bullock et al . 2000 ; somerville 2002 ; benson et al . 2002 ) . an additional scenario has been proposed by kravtsov , gnedin & klypin ( 2004 ) suggesting that the dwarf spheroidal we observe today were once much more massive objects that have been reduced to their present mass by tidal stripping processes . although these models can reconcile the shallow faint end of the luminosity function of galaxies with the low- mass end of the halo mass function , they did not explain so far the peculiarities of the dwarf spheroidal of the local group . for instance , it is known that dsphs of the local group tend to cluster tightly around the giant spirals . proximity to a giant central galaxy prevent them from accreting material and from further star formation and allow for tidal interactions to convert them from dwarf irregulars into spheroids . however , the presence of isolated dsphs found in the outskirts of the local group like tucana or cetus ( grebel , gallagher & harbeck 2003 ) seem to imply that dsphs might have been formed away from giant spirals and been accreted already as spheroids , in disagreement with the tidal stripping scenario , which is the common picture of dsph formation models ( see e.g. mayer et al . 2007 and reference therein ) . clues to these questions may be gathered from the current observations of the metallicities of a large sample of stars in four nearby dwarf spheroidal galaxies : sculptor , sextans , fornax , and carina . recent work from helmi et al . ( 2006 ) shows that all the four systems lack significantly of stars with low metallicity with the dsph metallicity distribution significantly different from that of the galactic halo . in this context sales et al . ( 2007 ) pointed out that satellite galaxies might be accreted into the main halo as multiple systems and more recently li & helmi ( 2007 ) showed that subhalos are often accreted in small groups . lake & donghia ( 2008 ) have collected data for nearby dwarf associations from tully et al . ( 2006 ) including the list of candidates of dwarf galaxies associated with the magellanic plane group : lmc , smc , sagitarius , ursa minor , draco , sextans and leoii ( lynden bell 1976 , fusi pecci et al . 1995 ; kroupa 2005 ) . the authors showed that seven out of ten dwarfs within @xmath1kpc from the milky way center might well be part of a group of dwarfs recently accreted into the milky way showing that there are natural mechanisms that lead to less suppression of satellite galaxies in dwarf galaxy groups and pointing to a resolution for the missing dwarf problem in the milky way . in this paper , we present a simulation of a loose group containing a milky way sized halo showing that the magellanic group likely entered the milky way at a redshift 1 and breaks up outside the virial radius . in lake & donghia ( 2008 ) we showed that this picture can naturally explain why galaxy satellites like draco or ursa minor are luminous and why other dwarfs are expected to be dark . here we focus on the peculiarities of the satellite galaxies of the local group and show that they are consistent with the tidal breakup of an infalling group of dwarf galaxies dominated from the lmc and smc systems . our analysis is based on a high - resolution cosmological simulation of a loose group containing a milky way sized halo . the target halo of the loose group has a virial mass @xmath27x@xmath3m@xmath4 with @xmath2 6 millions particles within the virial radius . the group was identified in a cosmological simulation of box 90 mpc ( or 67.5 @xmath5 mpc ) ( comoving ) on a side of 300@xmath6 particles with cosmological parameters chosen to match the 3-yr wilkinson microwave anisotropy probe ( wmap3 ) constraints ( spergel et al . these are characterized by the present - day matter density parameter , @xmath7 ; a cosmological constant contribution , @xmath8=0.762 ; and a hubble parameter @xmath9 ( @xmath10 km s@xmath11 mpc@xmath11 ) . the mass perturbation spectrum has a spectral index , @xmath12 , and is normalized by the linear rms fluctuation on @xmath13 mpc radius spheres , @xmath14 . the candidate halo was selected according to the following criteria : _ ( i ) _ environment location along a filament at least 5 @xmath5 mpc far from any rich galaxy group or galaxy cluster at z=0 . _ ( ii ) _ the the peak velocity of the milky way halo inside the group is 206 km s@xmath11 the group candidate containing the milky way halo was selected and resimulated to higher resolution using grafic2 ( bertschinger 2001 ) . the simulation was run with the tree - code pkdgrav ( stadel 2001 ) . gravitational interactions between pairs of particles are softened with a fixed comoving softening length of 0.8 @xmath5 kpc . we use skid ( stadel 2001 ) to identify subhalos in the high resolution region . skid finds subhalos within main halos by locating high density regions within the main halo and identifying the bound group of particles associated with each overdensity using a friends - of - friends ( fof ) method . particles which are not bound to their group are removed from the group by an unbinding procedure . we list for our catalog all subhalos with more than 20 particles . the output of skid is a list of subhalos with their structural properties . we focus on the following satellite properties : _ ( i ) _ the subhalo mass m@xmath15 ; _ ( ii ) _ the peak of the subhalo circular velocity profile @xmath16 ; _ ( iii ) _ the location of the subhalo center , identified with the most bound particle , which is the particle with the minimum gravitational potential energy . skid was run on all snapshot up to z=5 , allowing us to track the evolution of individual subhalos . rrrr 1 & 62 & 1.2x10@xmath17 & 134 + 1 & 53 & 6.5x10@xmath18 & 189 + 1 & 40 & 4x10@xmath18 & 720 + 1 & 35 & 2.5x10@xmath18 & 20 + 2 & 29 & 2x10@xmath18 & 69 - 189 + 1 & 27 & 1x10@xmath18 & 1072 + 1 & 18 & 4.5x10@xmath19 & 341 + 2 & 16 & 4x10@xmath19 & 55 - 770 + 1 & 14 & 2.9x10@xmath19 & 277 + 3 & 13 & 3x10@xmath19 & 153 - 247 - 352 + 3 & 12 & 1.5x10@xmath19 & 58 - 66 - 131 + 1 & 11 & 1.3x10@xmath19 & 85 + 5 & 10 & 1.3x10@xmath19 & 39 - 45 - 344 - 404 - 734 + we have identified all the surviving subhalos within the virial radius of the milky way halo at present time . within the virial radius of the simulated milky way , there are a total of 70 satellites with circular velocities greater than 10@xmath20 at present time . there are five magellanic clouds - sized subhalos inside the milky way halo with peak circular velocities between 50 and 62 km s@xmath11 . we traced each representative subhalos backwards in time and we searched for the group it belongs to at preceding snapshot . we focused on two target subhalos with circular velocity of 62 and 53 @xmath20 respectively and found out that these two subhalos were both satellites of an infalling group of substructures . hereafter we refer to this infalling group as the `` magellanic group '' . figure 1 shows the magellanic group ( at the bottom the picture ) infalling into the milky way halo at z=1.12 . the group infalls late and it breaks up outside the virial radius of the final milky way halo and spreads its surviving satellites over such a wide range of radii . the infalling group of satellites is loose and resembles association of dwarf galaxies discovered from tully et al . ( 2006 ) within 1 and 3 mpc from the local group . surprisingly the group becomes unbound through tidal disruption before entering within the milky way halo . the basic properties of the members of the magellanic group at z=0 are presented in table 1 . the group has roughly only 23 members surviving as satellites which contribute to the total budget of subhalos of the milky way at present time . table 1 lists the total number of subhalos found with similar peak of the circular velocity , @xmath21 and the mass associated to the members . figure 2 shows the spatial distribution of all the satellites within the virial radius of the milky way ( blue filled circles ) as compared to the subhalos of the disrupted magellanic group at z=0 ( magenta stars ) . despite the late infall , this particular group appears very well mixed , however almost half of the surviving subhalos of the group ( 10 subhalos ) are at present time located outside the virial radius of the final milky way . a few of them are in the outskirts of the milky way . these subhalos may intuitively reproduce the special cases like tucana or cetus which seem to be located in the outskirts of the local group ( grebel , gallagher & harbeck 2003 ) . in particular one of the satellite of the destroyed group is almost 1 mpc away from the milky way center ( see table 1 ) . out of 23 surviving subhalos of the lmc group the remaining subhalos and material of the group is destroyed and contributes to the diffuse milky way halo . we traced backwards in time the other representative magellanic systems and found that they belong to infalling multiple systems . figure 3 shows the peak circular velocity cumulative distribution of the satellites contributed by the simulated infalling magellanic group measured at z=0 compared to the observed magellanic plane group ( filled squared symbols)(see also figure 2 in lake & donghia 2008 ) . the magellanic plane group includes : lmc , smc , sagittarius , ursa minor , draco , sextans and leo ii ( lynden bell 1976 ; fusi pecci et al . 1995 ; kroupa 2005 ) . the figure displays also the circular velocity cumulative distribution of the observed satellites in the milky way including the newest dwarfs with a minimum @xmath22 and a correction for incomplete sky coverage ( simon & geha 2007 ) . we note that while the simulated magellanic group contributes less than 10% of the milky way halo mass , it is responsible for nearly 20% of the subhalos within the milky way virial radius at present time . having only 10% of the milky way mass any accreted gas inside the group would cool down quicly preventing its dwarf members from being stripped by ram pressure . as a remarkable consequence the dsphs might have been accreted into the milky way already as spheroids . figure 4 shows the two target subhalos of 62 and 53 @xmath20 representative in our sample of the magellanic clouds orbiting within the milky way halo . elongated tidal streams are formed well outside the virial radius . these two effects , the survivor satellites and long tidal debris outside the virial radius reinforce recent findings suggesting that the traditional virial radius and virial mass of the dark halos might be underestimated of a factor of two or three when a bulk infall is occurring ( e.g. see figure 3 in donghia & navarro 2007 ; cuesta 2007 ) . this is also consistent with results from n - body simulations suggesting that many satellite galaxies are bound on extreme orbits with apocentric distances of three virial radii ( e.g. moore et al . 2004 ; diemand et al . 2007 ; sales et al . 2007 ) . we performed a simulation of a milky way halo in the @xmath23cdm models focusing on the properties of the satellite population at the present day . our simulation probes that in the cosmological context of the hierarchical universe the magellanic clouds might have been accreted into the milky way from an infalling group and that the lmc and smc became unbound only recently . this results is consistent with the recent findings from kallivayalil et al . ( 2006 ) , besla et al . ( 2007 ) who argue that the lmc fell into the milky way halo only 2 gyrs ago and is moving at nearly the escape velocity at its radius approaching its orbital pericenter for the first time . one might argue that the time of the accretion of magellanic clouds is @xmath24 in this specific simulation , much earlier than the observational results but groups with 10 % of the milky way mass may infall any time our main results may be summarized as follows . _ ( i ) _ the tidal break up of the magellanic group occurs outside the current virial radius of the milky way and spreads its surviving satellites over 2 or 3 virial radii . half of its satellites survive outside the current virial radius of the milky way in the form of satellites , whereas the remaining material contributes to the diffuse milky way halo . _ ( ii ) _ at z=0 the disrupted magellanic group contributes less than 10% to the milky way mass but 20% of the brightest dwarf galaxies of the milky way have been part of this group . _ ( iii ) _ the circular velocity cumulative distribution of the satellites of the simulated magellanic group matches the function of the observed magellanic plane satellites the in the local group , suggesting that the gas physics in dwarf groups plays a key role in keeping dwarf galaxies luminous and might be the key to solve the missing dwarf problem as already discussed in lake & donghia ( 2008 ) . _ ( iv ) _ elongated tidal streams are formed as remnants of the breakup of the magellanic group well outside the traditional virial radius estimate , reinforcing the idea that the traditional estimates of the virial masses around galaxies might be underestimated of a factor of 2 when an infall is happening . a remarkable consequence of the last result is that our models predict a great amount of low mass satellites to be discovered well outside the virial radius of the milky way . it predicts groups of dwarf galaxies to be in the neighborhood of the local group . some associations of dwarf galaxies within 5 mpc from the milky way have been already discovered ( tully et al . 2006 ) but more associations are expected in the @xmath23cdm models and might be a challenge for future observations . in particular these association of dwarfs have the properties expected of bound systems with 1 - 10x10@xmath25 m@xmath4 , but they have too little gas and too few stars with a consequent mass - to - light ratio of 100 - 1000 m@xmath4/l@xmath4 . such a high mass - to - light ratio breaks the relationship between the m@xmath4/l@xmath4 and the mass of bound systems and leads to a sharp increase of m@xmath4/l@xmath4 for systems below the mass of 6x10@xmath25 m@xmath4 . a similar break and sharp increase of the m@xmath4/l@xmath4 at low mass systems has been pointed out from van den bosch et al . ( 2003 ; 2005 ) . the authors found a paucity of light at low mass systems comparing the shallow faint - end of the observed luminosity functions of galaxies with the corresponding halo mass function . it has not escaped our notice that the scenario we proved explains a lot of properties and peculiarities of the milky way substructures . it suggests that dwarf galaxies have been form in a different group environment away from the giant spirals and been accreted already as spheroids . this would naturally explain the mismatch in metallicity between the stellar halo of the milky way and the dwarf spheroidals that many have suspected dissolved to build it . the simulations were run on zbox2 at university of zurich . the author thanks george lake for useful discussions , joachim stadel for providing the tree - code pkdgrav and doug potter and peter engelmaier for technical support . she is grateful for the hospitality at the harvard - smithsonian center for astrophysics . in particular she thanks lars hernquist , gurtina besla , nitya kallivayalil , beth willman and all the theory group for many useful discussions during the theory meetings which have improved this work . this work has been supported by a eu marie curie fellowship under contract meif-041569 . avila - reese , v. , colin , p. , valenzuela , o. , donghia , e. , firmani , c. 2001 , apj , 559 , 516 belokurov , v. et al . 2007 , apj , 654 , 897 benson , a.j . , frenk , c.s . , lacey , c.g . , baugh , c.m . , cole , s. 2002 , mnras , 333 , 177 bertschinger , e. 2001 , apjs , 137 , 1 besla , g. , kallivayalil , n. , hernquist , l. , robertson , b. , cox , t. j. , van der marel , r.p . , alcock , c. 2007 , apj , 668 , 949 bullock , j.s . , kravtsov , a.v . , weinberg , d.h . 2000 , apj , 539 , 517 colin , p. , avila - reese , v. , valenzuela , o. 2000 , apj , 542 , 622 cuesta , a. j. , prada , f. , klypin , a. , moles , m. 2007 , arxiv0710.5520 diemand , j. , kuhlen , m. , madau , p. 2007 , apj , 667 , 859 donghia , e. , navarro , j.f . 2007 , mnras , 380 , l58 fusi pecci , f. ; bellazzini , m. ; cacciari , c. ; ferraro , f. r. 1995 , aj , 110 , 1664 grebel , e. k. , gallagher , j.s . , harbeck , d. 2003 , aj , 125 , 1926 helmi , a. 2006 , , 651 , l121 . kallivayalil , n. , van der marel , r.p . , alcock , c. 2006 , apj , 652 , 1213 kamionkowski , m. , liddle , a.r . 2000 , phrvl , 84 , 4525 kauffmann , g. , white , s. d. m. , guiderdoni , b. 1993 , mnras , 264 , 201 klypin , a. , kravtsov , a.v . , valenzuela , o. , prada , f. 1999 , apj , 522 , 82 kravtsov , a.v . , gnedin , o.y . , klypin , a.a 2004 , apj , 609 , 482 kroupa , p. , theis , c. , boily , c. m. 2005 , a&a , 431 , 517 lake , g. , donghia , e. 2008 , apj , submitted li , y & helmi , a. 2007 , mnras , submitted , arxiv0711.2429 lynden - bell , d. 1976 , mnras , 174 , 695 mayer , l. , kazantzidis , s. , mastropietro , c. , wadsley , j. 2007 , nature , 445 , 738 moore , b. , ghigna , s. , governato , f. , lake , g. , quinn , t. , stadel , j. , tozzi , p. 1999 , apj , 524 , 19 moore , b. , diemand , j. , stadel , j. 2004 , iau colloquium 195:outskirts of galaxy clusters : intense life in the suburbs , p.513 , arxiv:0406615 sales , l.v . , navarro , j.f . , abadi , m.g . , steinmetz , m. 2007 , mnras , 379 , 1475 simon , j.d . , geha , m. 2007 , apj , 670 , 313 somerville , r.s . 2002 , apj , 572 , l23 spergel , d.n . 2007 , apjs , 170 , 377 stadel , j.g . 2001 , thesis ( phd ) , university of washington , source dai - b 62/08 , p.3657 , 141 pages strigari , l.e . , bullock , j.s . , kaplinghat , m. , diemand , j. , kuhlen , m. , madau , p. 2007 , apj , 669 , 676 tully , r.b . , rizzi , l. , dolphin , a. e. , karachentsev , i. d. , karachentseva , v. e. , makarov , d. i. , makarova , l. , sakai , s. , shaya , e.j . 2006 , aj , 132 , 729 van den bosch , f. c. , mo , h. j. , yang , x. 2003 , mnras , 345 , 923 van den bosch , f.c . , yang , x. , mo , h. j. , norberg , p. 2005 , mnras , 356 , 1233 white , s. d. m. , @xmath26 reese , m. j. 1978 , , 183 , 341 willman , b. et al . 2005a , apj , 626 , l85 willman , b. et al . 2005b , aj , 129 , 2692 zentner , a.r . , bullock , j.s . 2003 , apj , 598 , 49 zucker , d. b. et . al . 2006 , mnras apj , 650 , l41
we use a numerical simulation of a loose group containing a milky way halo to probe that in the hierarchical universe the magellanic clouds and some dsphs have been accreted into the milky way halo from a late infalling group of dwarfs . our simulations show that the tidal breakup of the magellanic group occurs before it enters the milky way halo . only half of the satellites contributed from the group are predicted to be inside the milky way virial radius . half of its subhalos survive outside the current virial radius in the form of satellites , whereas the remaining material contributes to the diffuse milky way halo . at @xmath0 the disrupted group contributes less than 10% to the milky way halo mass but 20% of the brightest dwarf galaxies of the milky way have been part of this group . this scenario points out that some dsphs might have been formed away from giant spirals and been accreted already as spheroids , by a late infall group in contrast with the classical picture of tidal stripping of dsph formation models . this would naturally explain several peculiarities of the local dsph : why draco and the other luminous dsphs exist compared to other ultra - faint satellite galaxies , the location of tucana and cetus in the outskirts of the local group and the mismatch in metallicity between the stellar halo of the milky way and the dwarf galaxies that many have suspected dissolved to build it .
let @xmath5 be an arbitrary finite set . in the case we deal with boolean clones we have @xmath6 . a function @xmath7 is called a total function on @xmath5 . a function @xmath8 with @xmath9 is called partial function on @xmath5 and we denote the domain by @xmath10 . the set @xmath11 is the set of all total functions on @xmath5 , and @xmath12 is the set of all partial functions on @xmath5 . the function @xmath13 defined by @xmath14 is called the @xmath15-ary _ projection _ onto the @xmath16-th coordinate . for each @xmath17 the function @xmath18 is defined as @xmath19 for all @xmath20 . let @xmath21 be @xmath15-ary and let @xmath22 be @xmath23-ary . the _ composition _ @xmath24 is an @xmath23-ary partial function defined by @xmath25 and @xmath26 @xmath27 is called a partial clone if it is composition closed and contains the projections . if additionally @xmath28 then @xmath0 is a total clone . let @xmath29 . then @xmath30 is a restriction ( or subfunction ) of @xmath31 if @xmath32 and @xmath33 for all @xmath34 , short @xmath35 . let @xmath36 . then the set @xmath37 is defined by @xmath38 if @xmath39 then @xmath40 is called _ strong _ , or _ restiction closed_. that means , that @xmath40 contains every restriction of every of its functions , i.e. , @xmath41 for every @xmath21 and @xmath42 with @xmath35 . let @xmath43 $ ] be the set of all @xmath44-ary relations on @xmath5 for some @xmath45 , i.e. , @xmath43 : = \ { x \mid x \subseteq a^h \}$ ] . furthermore , let @xmath46 $ ] . let @xmath47 $ ] , and @xmath8 with @xmath9 an @xmath15-ary partial function . then @xmath30 preserves @xmath48 iff @xmath49 for any @xmath50 matrix @xmath51 whose rows belong to the domain of @xmath30 , i.e. @xmath52 for all @xmath16 , and whose columns belong to @xmath48 . let @xmath53 be the set of all partial functions preserving every relation @xmath54 . let @xmath55 the set of all total functions preserving every relation @xmath54 . there at least three different types of intervals which we consider here . let @xmath0 be a total clone of @xmath11 . then we can define the three intervals @xmath2 , @xmath4 , and @xmath56 by @xmath57 clearly , @xmath58 holds . the following total boolean clones are needed in this paper , and every other total boolean clone can be written as the intersection of some of these . @xmath59\right)\end{aligned}\ ] ] = [ circle , fill = black , scale=0.5 ] = [ scale=2,draw = none , fill = none , label distance=-4 ] ( p2 ) [ label = above:@xmath60 at ( 7.0 , 10.0 ) ; ( t0 ) [ label = above left:@xmath61 at ( 7.0 - 1.5,10.0-.5 ) ; ( t1 ) [ label = above right:@xmath62 at ( 7.0 + 1.0,10.0-.5 ) ; ( t ) at ( 7.0-.5,10.0 - 1.0 ) ; ( m ) [ label = above left:@xmath63 at ( 6.76 , 9.38 ) ; ( t0 m ) at ( 7.23 - 1.5 , 9.51-.5 ) ; ( t1 m ) at ( 6.76 + 1.0 , 9.38-.5 ) ; ( tm ) at ( 7.23-.5 , 9.51 - 1.0 ) ; ( l ) [ label = above right:@xmath64 at ( 7.0 , 5.2 ) ; ( t0l ) at ( 7.0 - 1.5 , 5.2-.5 ) ; ( t1l ) at ( 7.0 + 1.0 , 5.2-.5 ) ; ( tl ) at ( 7.0-.5 , 5.2 - 1.0 ) ; ( av ) at ( 7.0 , 3.0 ) ; ( t0av ) at ( 7.0 - 1.5 , 3.0-.5 ) ; ( t1av ) at ( 7.0 + 1.0 , 3.0-.5 ) ; ( tav ) at ( 7.0-.5 , 3.0 - 1.0 ) ; ( p21 ) [ label = right:@xmath65 at ( 7.0,3.8 ) ; ( s ) [ label = right:@xmath66 at ( 6.85,7.0 ) ; ( sl ) at ( 6.75,4.7 ) ; ( sp21 ) at ( 6.65,2.5 ) ; ( st ) at ( 6.5,6.5 ) ; ( sm ) at ( 6.15,6.0 ) ; /in p2/t0 , p2/t1 , t0/t , t1/t , m / t0 m , m / t1 m , t0m / tm , t1m / tm , p2/m , t0/t0 m , t1/t1 m , t / tm , l / t0l , l / t1l , t0l / tl , t1l / tl , p2/l , t0/t0l , t1/t1l , t / st , st / tl , av / t0av , av / t1av , t0av / tav , t1av / tav , l / p21 , p21/av , t0l / t0av , t1l / t1av , tl / tav , p2/s , s / st , s / sl , l / sl , sl / tl , sl / sp21 , p21/sp21 , sp21/tav , st / sm , sm / tav ( ) ( ) ; ( t02 ) [ label = left:@xmath67 at ( .5 , 8.5 ) ; ( t02 t ) at ( .5 + 1.0 , 8.5-.7 ) ; ( t02 m ) at ( .5 + 2.0 , 8.5-.3 ) ; ( t02tm ) at ( .5 + 3.0 , 8.5 - 1.0 ) ; ( t03 ) [ label = left:@xmath68 at ( .5 , 7.5 ) ; ( t03 t ) at ( .5 + 1.0 , 7.5-.7 ) ; ( t03 m ) at ( .5 + 2.0 , 7.5-.3 ) ; ( t03tm ) at ( .5 + 3.0 , 7.5 - 1.0 ) ; ( t0h ) [ draw = none , fill = none , scale=0.1 ] at ( .5 , 7.0 ) ; ( t0ht ) [ draw = none , fill = none , scale=0.1 ] at ( .5 + 1.0 , 7.0-.7 ) ; ( t0hm ) [ draw = none , fill = none , scale=0.1 ] at ( .5 + 2.0 , 7.0-.3 ) ; ( t0htm ) [ draw = none , fill = none , scale=0.1 ] at ( .5 + 3.0 , 7.0 - 1.0 ) ; ( t0e ) [ label = left:@xmath69 at ( .5 , 5.5 ) ; ( t0et ) at ( .5 + 1.0 , 5.5-.7 ) ; ( t0em ) at ( .5 + 2.0 , 5.5-.3 ) ; ( t0etm ) at ( .5 + 3.0 , 5.5 - 1.0 ) ; \(a ) [ label = below:@xmath70 at ( .5 + 3.5,3.7 ) ; ( at1 ) at ( .5 + 4.5,3.2 ) ; ( at0 ) at ( .5 + 2.0,3.2 ) ; ( at ) at ( .5 + 3.0,2.7 ) ; ( t12 ) [ label = right:@xmath71 at ( 13.0 , 8.5 ) ; ( t12 t ) at ( 13.0 - 1.0 , 8.5-.7 ) ; ( t12 m ) at ( 13.0 - 2.0 , 8.5-.3 ) ; ( t12tm ) at ( 13.0 - 3.0 , 8.5 - 1.0 ) ; ( t13 ) [ label = right:@xmath72 at ( 13.0 , 7.5 ) ; ( t13 t ) at ( 13.0 - 1.0 , 7.5-.7 ) ; ( t13 m ) at ( 13.0 - 2.0 , 7.5-.3 ) ; ( t13tm ) at ( 13.0 - 3.0 , 7.5 - 1.0 ) ; ( t1h ) [ draw = none , fill = none , scale=0.1 ] at ( 13.0 , 7.0 ) ; ( t1ht ) [ draw = none , fill = none , scale=0.1 ] at ( 13.0 - 1.0 , 7.0-.7 ) ; ( t1hm ) [ draw = none , fill = none , scale=0.1 ] at ( 13.0 - 2.0 , 7.0-.3 ) ; ( t1htm ) [ draw = none , fill = none , scale=0.1 ] at ( 13.0 - 3.0 , 7.0 - 1.0 ) ; ( t1e ) [ label = right:@xmath73 at ( 13.0 , 5.5 ) ; ( t1et ) at ( 13.0 - 1.0 , 5.5-.7 ) ; ( t1em ) at ( 13.0 - 2.0 , 5.5-.3 ) ; ( t1etm ) at ( 13.0 - 3.0 , 5.5 - 1.0 ) ; \(v ) [ label = below:@xmath74 at ( 13.0 - 3.5,3.7 ) ; ( vt0 ) at ( 13.0 - 4.5,3.2 ) ; ( vt1 ) at ( 13.0 - 2.0,3.2 ) ; ( vt ) at ( 13.0 - 3.0,2.7 ) ; /in t02/t02 t , t02/t02 m , t02t / t02tm , t02m / t02tm , t03/t03 t , t03/t03 m , t03t / t03tm , t03m / t03tm , t0e / t0et , t0e / t0em , t0et / t0etm , t0em / t0etm , t12/t12 t , t12/t12 m , t12t / t12tm , t12m / t12tm , t13/t13 t , t13/t13 m , t13t / t13tm , t13m / t13tm , t1e / t1et , t1e / t1em , t1et / t1etm , t1em / t1etm , t02/t03 , t02t / t03 t , t02m / t03 m , t02tm / t03tm , t03/t0h , t03t / t0ht , t03m / t0hm , t03tm / t0htm , t12/t13 , t12t / t13 t , t12m / t13 m , t12tm / t13tm , t13/t1h , t13t / t1ht , t13m / t1hm , t13tm / t1htm , t0em / at0 , t0etm / at , t1em / vt1 , t1etm / vt , a / at0 , a / at1 , at0/at , at1/at , v / vt0 , v / vt1 , vt0/vt , vt1/vt , a / av , at0/t0av , at1/t1av , at / tav , v / av , vt0/t0av , vt1/t1av , vt / tav , t02tm / sm , t12tm / sm , t0/t02 , t0m / t02 m , t / t02 t , tm / t02tm , t1/t12 , t1m / t12 m , t / t12 t , tm / t12tm ( ) ( ) ; ( m ) edge [ out=215 , in=90 ] ( a ) ; ( t1 m ) edge [ out=215 , in=90 ] ( at1 ) ; ( m ) edge [ out=325 , in=90 ] ( v ) ; ( t0 m ) edge [ out=325 , in=90 ] ( vt0 ) ; /in t0h / t0e , t0ht / t0et , t0hm / t0em , t0htm / t0etm , t1h / t1e , t1ht / t1et , t1hm / t1em , t1htm / t1etm ( ) ( ) ; plot coordinates ( @xmath75 ) ( @xmath76 ) ( @xmath77 ) ( @xmath78 ) ( @xmath79 ) ; plot coordinates ( @xmath80 ) ( @xmath81 ) ( @xmath82 ) ( @xmath83 ) ( @xmath84 ) ; plot coordinates ( @xmath85 ) ( @xmath86 ) ( @xmath87 ) ( @xmath88 ) ( @xmath89 ) ; at ( @xmath90 ) * section [ section : introduction ] * ; at ( @xmath91 ) * section [ section : t02 ] * ; at ( @xmath92 ) * section [ section : t02 ] * ; at ( @xmath93 ) * sec . [ section : lambda ] * ; at ( @xmath94 ) * section [ section : l ] * ; in @xcite the finite intervals @xmath2 have been determined and in @xcite the finite intervals of the form @xmath56 . these results can be assembled into the following theorem . the column with the sizes for @xmath4 can be easily deduced from the sizes @xmath56 and post s lattice . the finite intervals @xmath95 and @xmath96 are displayed in appendix [ section : finiteintervalsdrawings ] . [ theorem : finiteintervals ] let @xmath0 be a total boolean clone with @xmath97 then @xmath2 and @xmath56 are finite sets . furthermore it holds that @xmath98 in @xcite it was shown that the intervals @xmath2 for subclones @xmath99 with @xmath100 have the size of the continuum . then in @xcite the remaining intervals were determined to be infinite . the authors of @xcite then finished the determination of the intervals of the form @xmath2 to yield the following theorem . let @xmath0 be a total boolean clone such that @xmath99 and @xmath101 . then the set @xmath2 has the cardinality of the continuum . as stated in @xcite this yields a dichotomy on the size of the intervals @xmath2 for boolean clones @xmath0 . let @xmath0 be a total boolean clone . then @xmath2 is either finite or has the cardinality of the continuum . furthermore , @xmath2 is finite if and only if @xmath102 or @xmath103 . the aim of this paper is to show that this result can be strengthend in the sense that only strong partial clones are considered . that means @xmath2 will be replaced by @xmath4 in the statement of the last theorem . since @xmath58 for every boolean clone @xmath0 , we already have that @xmath4 is finite if @xmath102 or @xmath103 . thus we will show that the interval @xmath4 has the cardinality of the continuum for all other boolean clones @xmath0 . although we focus on the case of boolean clones , there have been some investigations into the general case with @xmath104 , for example @xcite and @xcite . some of these results will be extended with the help of lemma [ lemma : diffconstants ] . for some natural numbers @xmath105 with @xmath106 we define the sets @xmath107 : = \{n , n+1,\dots , m\}$ ] , and @xmath108 : = [ 1,n]$ ] . tuples will be written with boldface small letters , and with the exception of @xmath109 a small boldface letter signifies a tuple . for a tuple @xmath110 we define the set of its entries by @xmath111 : = \{x_1,\dots , x_n\}$ ] , and let @xmath112|$ ] . for @xmath113 $ ] we let @xmath114 . for @xmath115^l$ ] with @xmath116 we define @xmath117 . we will often use the two special tuples @xmath118 and @xmath119 ; the length of these tuples can be deduced from the context . the statement of theorem [ theorem : definability ] proven by romov in @xcite gives a nice characterization of the constructability of relations in the co - clone of a strong partial clone . this enables us to prove the theorems [ theorem : t02:nonconstructible ] and [ theorem : lambda : nonconstructible ] just with relational methods . the relation @xmath120 $ ] is called _ irredundant _ iff it fulfills the following two conditions : 1 . @xmath121 has no duplicate rows , i.e. , for all @xmath122 with @xmath123 , there is a tuple @xmath124 with @xmath125 ; 2 . @xmath121 has no fictitious coordinates , i.e. , there is no @xmath126 , such that @xmath124 implies @xmath127 for all @xmath128 . for a relation @xmath129 $ ] we define @xmath130 . [ theorem : definability ] let @xmath131 and @xmath132 $ ] be relations . furthermore let @xmath121 be irredundant . then @xmath133 iff there are some @xmath134^{\operatorname{arity}\sigma}$ ] for all @xmath135 such that @xmath136 and @xmath137 = \bigcup_{\sigma \in \sigma } \bigcup_{\ii \in \gamma_\sigma } [ \ii].\ ] ] the aim of this section is the introduction of classes of partial functions ( or shorter : partial classes ) similar to the ideas presented by harnau in @xcite for total clones . this concepts will prove fruitful in the extension of theorem 8 @xcite as shown in lemma [ lemma : strongdownwards ] . since we do not need the full power of the galois connection presented by harnau we will only prove statements about partial classes relevant to this paper . for the definition of a partial class we need to define the following maltsev - operations @xmath138 , @xmath139 , @xmath140 , @xmath141 , and @xmath142 . let @xmath143 $ ] and @xmath144 $ ] . then we define @xmath145 let @xmath36 . then @xmath40 is called a _ partial class _ if it closed under the operations @xmath142 , @xmath138 , @xmath139 , @xmath141 , and @xmath140 . [ lemma : intersectingclasses ] let @xmath146 be two partial classes . then @xmath147 is also a partial class . the partial classes containing the projections are exactly the partial clones . if @xmath148 , then we define the set @xmath149 by @xmath150 similar to the work done by harnau in @xcite we introduce relation pairs to characterize strong partial classes . for each @xmath45 let @xmath151 $ ] be the set of all pairs @xmath152 with @xmath153 , and @xmath154 $ ] . let @xmath155 $ ] for some @xmath45 , and @xmath143 $ ] for some @xmath156 . then @xmath30 _ preserves _ the relation pair @xmath157 , if for all matrices @xmath158 with columns in @xmath121 , and lines in @xmath159 the tuple @xmath160 belongs to @xmath161 . we write @xmath162 , or @xmath163 . if @xmath164 then the preservation of the relation pair @xmath157 coincides with the preservation of the relation @xmath121 , i.e. , @xmath165 . if @xmath36 , and @xmath166 , then we define @xmath167 [ lemma : starcomposing ] let @xmath162 and @xmath168 with @xmath169 . then @xmath170 . let @xmath171 ( \rho,\rho')$ ] and @xmath172 ( \sigma , \sigma')$ ] with @xmath173 $ ] . let @xmath158 be an @xmath174-matrix with columns @xmath175 , and rows @xmath176 . let @xmath177 for each @xmath178 $ ] . then @xmath179 by the definition of @xmath142 , and @xmath180 . thus @xmath181 . from this @xmath182 and thus @xmath170 . [ lemma : cpolqisclass ] let @xmath183 . then @xmath184 is a non - empty strong partial class of @xmath12 . let @xmath185 . we first show that @xmath186 is a partial class let @xmath187 . it is easy to see that @xmath188 . from lemma [ lemma : starcomposing ] with @xmath189 and @xmath190 follows @xmath170 . thus @xmath186 is a partial class of @xmath12 . we now want to show that @xmath186 is strong . let @xmath162 and @xmath191 , and assume to the contrary that @xmath192 . then there is a matrix @xmath158 with columns @xmath193 and rows @xmath194 , such that @xmath195 . since @xmath196 and @xmath197 . thus @xmath198 contradicting the assumption . thus @xmath186 is strong . it is non - empty since the partial function @xmath199 with empty domain perserves any relation pair @xmath200 . let @xmath166 . then @xmath201 is a non - empty strong partial class of @xmath12 . by lemma [ lemma : cpolqisclass ] we have that @xmath184 is a strong partial class for all @xmath202 . then by lemma [ lemma : intersectingclasses ] and the definition of @xmath201 , we see that @xmath201 is a partial class . furthermore , the intersection of two strong sets is also strong . it is non - empty since @xmath203 . it is possible to show , that for every non - empty strong partial class @xmath36 , there is some @xmath166 with @xmath204 . since this and other further properties of the operators @xmath205 and @xmath206 are not needed in this paper , they will not be proven here . [ lemma : emptyconsequentnototalfunction ] let @xmath207 with @xmath208 . then @xmath209 . let @xmath210 $ ] , @xmath120 $ ] , and @xmath211 . let @xmath158 be the matrix formed by @xmath15-fold repetition of the column @xmath212 . let the rows of @xmath158 be called @xmath213 . clearly , @xmath214 for all @xmath215 $ ] since @xmath30 is a total function . but @xmath216 , and thus @xmath217 . let @xmath207 , @xmath21 and @xmath218 . then @xmath219 . let @xmath120 $ ] , @xmath143 $ ] and @xmath172 ( \rho , \emptyset)$ ] . if @xmath220 , then @xmath221 . thus @xmath219 . let @xmath208 . assume to the contrary , that @xmath222 . then there is a matrix @xmath158 with columns @xmath175 , and rows @xmath223 . we can now look at the matrix @xmath224 formed by the first @xmath23 columns , and with rows @xmath225 . then @xmath226 implies @xmath227 for all @xmath215 $ ] . but since @xmath228 we get @xmath229 in contradiction to the assumption . let @xmath36 and @xmath207 . then @xmath230 . the following corollary follows from lemma [ lemma : starcomposing ] . . then @xmath231 . the last two corollaries can now be combined into the final statement of this subsection . [ corollary : emptysetisveryhungry ] let @xmath207 , @xmath232 and @xmath233 . then @xmath234 and @xmath235 . in the proof that the interval @xmath236 are of continuum cardinality for some total clone @xmath237 , we try to make as few constructions as possible . this can be achieved if we find some clone @xmath0 with @xmath238 , construct a set @xmath239 of continuum cardinality , and then find restrictions of the partial clones in @xmath240 , such that these restricted partial clones lie in @xmath236 , and @xmath240 does not collapse . for this purpose we prove a stronger version of theorem 8 @xcite as follows . [ lemma : strongdownwards ] let @xmath0 and @xmath237 be clones of @xmath11 with @xmath238 , @xmath241 a strong partial class of @xmath12 , and @xmath239 , such that the following conditions hold 1 . [ enum : strongdownwards : totalempty ] @xmath242 , 2 . [ enum : strongdownwards : qihungry ] @xmath243 , and @xmath244 , 3 . [ enum : strongdownwards : different ] @xmath245 for all @xmath246 with @xmath247 . then @xmath248 for each @xmath249 we define @xmath250 by @xmath251 we let @xmath252 , and show that @xmath253 . by [ enum : strongdownwards : different ] we have that @xmath254 . let @xmath249 be arbitrary . by [ enum : strongdownwards : totalempty ] we have that @xmath255 thus we only have to show that @xmath250 is a strong partial clone . since @xmath256 , @xmath40 , and @xmath241 are strong partial classes , we see that @xmath257 , and that @xmath250 is closed with respect to @xmath138 , @xmath139 , @xmath141 and @xmath140 . furthermore , @xmath250 contains the projections , since @xmath258 , and @xmath237 is a clone . it remains to show that @xmath250 is closed with respect to @xmath142 . let @xmath259 . we want to show that @xmath260 . since @xmath261 , @xmath262 and @xmath40 is a partial clone , we have @xmath263 . there are several cases : * @xmath264 . then @xmath265 , since @xmath256 is a strong partial clone . * @xmath266 . then @xmath267 , since @xmath268 is a strong partial class . * @xmath269 , and @xmath270 ; or @xmath271 , and @xmath272 . by [ enum : strongdownwards : qihungry ] we have @xmath273 . thus @xmath274 thus @xmath250 is a strong partial clone with @xmath275 . this implies @xmath276 . therefore @xmath253 , and consequently @xmath277 . one example of the strong partial class @xmath241 needed in the preceding lemma is the partial class @xmath278 of all partial functions not defined on @xmath279 . this was implicitly used for example in @xcite and @xcite . each of the sets @xmath240 defined in this paper will be indexed by the subsets of a countable infinite set @xmath280 . as such the set @xmath240 has the same cardinality as the powerset of @xmath281 , which has the cardinality of the continuum , and therefore @xmath240 is of continuum cardinality . first we use lemma [ lemma : strongdownwards ] in a general setting , involving two clones @xmath0 and @xmath237 in @xmath11 with @xmath238 and @xmath282 for some @xmath17 . for a partial function @xmath143 $ ] and some @xmath17 we define the @xmath283-ary partial function @xmath284 by @xmath285 [ lemma : constants ] let @xmath28 be a clone with @xmath286 , and @xmath287 . then @xmath288 if and only if @xmath289 . assume @xmath288 . then @xmath290 , and thus @xmath288 . now assume that @xmath289 . additionally , we have @xmath291 . thus @xmath292 . [ lemma : noconstant ] let @xmath293 be a clone with @xmath294 . then there is some @xmath295 with @xmath296 . assume to the contrary , that @xmath297 for all @xmath295 . then @xmath298 for all @xmath295 , and thus @xmath299 . contradiction . [ lemma : diffconstants ] let @xmath300 be clones with @xmath282 and @xmath238 . then @xmath301 . by lemma [ lemma : noconstant ] there is some relation @xmath121 with @xmath296 and @xmath302 . let @xmath303 , and @xmath304 . we want to use lemma [ lemma : strongdownwards ] . since @xmath305 we have condition [ enum : strongdownwards : totalempty ] , and by corollary [ corollary : emptysetisveryhungry ] we have condition [ enum : strongdownwards : qihungry ] . now we want to show condition [ enum : strongdownwards : different ] . now let @xmath306 with @xmath247 ; w.l.o.g . there is some @xmath307 . by lemma [ lemma : constants ] we have @xmath308 . we just need to show that @xmath309 . assume to the contrary that @xmath310 . let @xmath311 be @xmath15-ary , and @xmath121 be @xmath44-ary . then there is a matrix @xmath158 such that * its row @xmath312 , and * its columns @xmath313 . by the definition of @xmath311 and choice of @xmath121 we see that @xmath314 . this is a contradiction . thus @xmath309 , and consequently @xmath245 . therefore all conditions of lemma [ lemma : strongdownwards ] are fulfilled , and we get @xmath301 . this lemma can be applied to the main results of theorems 10 and 19 in @xcite . let @xmath315 be the set of all @xmath44-universal relations ( @xmath316 ) , and let @xmath317 be the set of all prime affine relations on @xmath5 . then for each @xmath318 the following properties hold * @xmath319 is a maximal clone of @xmath11 , * @xmath298 for all @xmath17 , * @xmath320 has the cardinality of the continuum . with lemma [ lemma : diffconstants ] we obtain the following statement . let @xmath293 a clone with @xmath302 for some @xmath318 , and @xmath294 for some @xmath17 . then @xmath236 has the cardinality of the continuum . in this section we use the results from @xcite to show that the interval @xmath236 has continuum cardinality for all clones @xmath322 . we need to define some functions first as given in @xcite . let @xmath323 , @xmath324 and @xmath325 . define the @xmath326-ary partial function @xmath327 by @xmath328 we define @xmath329 by @xmath330 and @xmath331 for all @xmath332 . set @xmath333 for all @xmath334 . [ lemma : xisindependent ] let @xmath334 . then @xmath335 $ ] . as a consequence we get the following theorem . [ theorem : lcontinuum ] the interval @xmath336 has the cardinality of the continuum . let @xmath337 $ ] for every @xmath338 . by lemma [ lemma : xisindependent ] we see that @xmath339 if @xmath340 , and thus the set @xmath341 has the cardinality of the continuum . furthermore , @xmath342 . since @xmath321 is a maximal clone and @xmath343 , we conclude that @xmath336 has the cardinality of the continuum . [ lemma : lt0lt1lscontinuum ] let @xmath322 be a clone with @xmath344 with @xmath345 . then @xmath236 has the cardinality of the continuum . we have @xmath346 , and @xmath347 , @xmath348 , @xmath349 . thus lemma [ lemma : diffconstants ] is applicable with @xmath350 , and by [ theorem : lcontinuum ] follows that @xmath236 has the cardinality of the continuum . the only two subclones of @xmath321 not covered yet are @xmath351 $ ] and @xmath352 $ ] . let @xmath353 , @xmath354 and @xmath355 be three 4-ary relations defined as @xmath356 although the fact that @xmath357 holds , is not used directly , the similar structure makes the proof of lemma [ lemma : xipreserve ] a bit easier . as shown by blochina in @xcite ( see also section 10.2 @xcite ) the relations @xmath353 , @xmath354 and @xmath355 characterize the clones @xmath358 , @xmath359 , and @xmath321 , respectively . that means the following equalities hold : @xmath360 [ lemma : xipreserve ] let @xmath334 . then @xmath361 and @xmath362 . let @xmath363 . assume to the contrary , that @xmath364 does not preserve @xmath121 . then there is a matrix @xmath158 such that * its rows @xmath365 , * its columns @xmath366 , and * @xmath367 . we will show that @xmath368 for some @xmath369 $ ] . let @xmath370 \mid \xx_k \neq \1 \}$ ] . clearly , @xmath371 iff @xmath372 . since @xmath373 , we have that @xmath374 and thus there are @xmath375 $ ] with @xmath376 and @xmath377 . this implies @xmath378 . by the construction of @xmath364 each row @xmath379 for @xmath372 has at most @xmath329-many @xmath380 s . but @xmath364 has an arity of @xmath381 thus there is some column @xmath382 with @xmath383 for all @xmath372 . furthermore , @xmath384 for all @xmath385 \setminus k$ ] . thus @xmath386 . but this contradicts @xmath387 and @xmath388 . therefore @xmath389 . [ lemma : c01continuum ] @xmath390 has the cardinality of the continuum . let @xmath337 $ ] for every @xmath338 . then the set @xmath341 has the cardinality of the continuum . let @xmath391 . * @xmath392 . * as @xmath393 we have @xmath394 , and @xmath395 . * by lemma [ lemma : xipreserve ] and the definition of @xmath240 , we get @xmath245 for all @xmath246 with @xmath247 . then we apply lemma [ lemma : strongdownwards ] , and yield the result . by setting @xmath396 in the previous proof we obtain the proof for the following lemma . [ lemma : omega1continuum ] @xmath397 has the cardinality of the continuum . now we can conclude from theorem [ theorem : lcontinuum ] , and lemmas [ lemma : lt0lt1lscontinuum ] , [ lemma : c01continuum ] , [ lemma : omega1continuum ] , that the following theorem holds . [ theorem : subl ] let @xmath322 . then @xmath236 has the cardinality of the continuum . in this section we first give an alternative proof for the fact that @xmath399 has the cardinality of the continuum . the relations used are similar to the ones given in @xcite , but the proof here only uses relations . the proof given in this section uses some ideas from the proof in @xcite , but changes the basic building blocks of the relations used . furthermore , while the former proof depended on working with functions , this proof here only deals with relations . let @xmath400 . then we remember that @xmath401 . let @xmath402 and @xmath403 be two @xmath15-ary relations defined by @xmath404 } \rho_{0,2}(x_i , x_{i+1 \bmod{n } } ) , \\ { r^{0,2}}_{k , n}(x_1,\dots , x_n ) & : = \bigwedge_{\substack{i , j \in [ n ] \\ i \neq j } } \rho_{0,2}(x_i , x_j ) . \\ \intertext{furthermore , let } { r^{0,2}}_n & : = { r^{0,2}}_{c , n } \times { r^{0,2}}_{k , n}.\end{aligned}\ ] ] the names @xmath0 and @xmath405 in the indices of the relations are in correspondance with the circular graph @xmath406 and the complete graph @xmath407 on @xmath15 vertices . the relations @xmath402 have the same definition as @xmath408 in @xcite . the idea behind replacing @xmath409 with @xmath403 stemmed from the fact that with graphs the following holds : * let @xmath410 be two odd numbers . then there is no graph homomorphism from @xmath406 into @xmath411 . * for @xmath410 there is no graph homomorphism from @xmath412 into @xmath407 . * for @xmath413 there is no graph homomorphism from @xmath407 into @xmath411 . the relation @xmath414 represents in this model the disjoint union @xmath415 of @xmath406 and @xmath407 . let @xmath416 denote the fact , that there is some graph homomorphism from @xmath417 to @xmath418 . we consider the possible homomorphisms from @xmath419 to @xmath420 . then we see * for @xmath410 that @xmath421 , since @xmath422 and @xmath423 ; and * for @xmath424 that any homomorphism from @xmath425 to @xmath426 actually maps into @xmath427 , since @xmath428 and @xmath423 . but for the construction of @xmath429 this would mean that the first @xmath15 coordinates are not essential , a contradiction . for @xmath430 the relations @xmath431 and @xmath432 look like this : @xmath433 . then @xmath435 . by construction we have that @xmath436 , @xmath437 , and @xmath438 . since @xmath439 we obtain @xmath440 , and @xmath441 . from this follows @xmath442 , and consequently @xmath443 . let @xmath444 . let @xmath445 and @xmath446 for the rest of this section . we want to show that @xmath447 holds . we assume to the contrary , that is false . this means that by theorem [ theorem : definability ] we can write @xmath448 for some auxiliary relations @xmath449 for all @xmath450 . furthermore , we can assume that no condition is superfluous . [ lemma : seperateck ] let @xmath450 with @xmath451 , and @xmath452 . then @xmath453 } \subseteq [ n]$ ] or @xmath453 } \subseteq [ n+1,2n]$ ] . + similarly , @xmath454 } \subseteq [ n]$ ] or @xmath454 } \subseteq [ n+1,2n]$ ] . we only consider the first statement ; the second one follows similarly . assume the statement is not true . then there is some @xmath455 $ ] such that @xmath456 $ ] and @xmath457 $ ] . by construction of @xmath414 ( or , more specifically @xmath402 ) this means , that @xmath458 holds , i.e. , @xmath459 and @xmath460 can not both be @xmath380 at the same time . but by construction of @xmath414 we have @xmath461{\ensuremath{i_j}}\vphantom{\ensuremath{i_j}}}}{1}},0,\dots,0,{\underset{\substack{\uparrow \\ \makebox[0pt]{\ensuremath{i_{j+1 \bmod{m}}}}\vphantom{\ensuremath{i_{j+1 \bmod{m}}}}}}{1}},0,\dots,0 ) \in { r^{0,2}}_n.\ ] ] this is a contradiction , and thus @xmath453 } \subseteq [ n]$ ] or @xmath453 } \subseteq [ n+1,2n]$ ] . [ lemma : r02mlessn ] let @xmath450 and @xmath451 . then @xmath462 . assume to the contrary , that @xmath450 , @xmath451 , and @xmath463 . then there is some @xmath452 . by lemma [ lemma : seperateck ] we have @xmath454 } \subseteq [ n]$ ] or @xmath454 } \subseteq [ n+1,2n]$ ] . thus @xmath464 for some @xmath465 $ ] with @xmath466 . by construction of @xmath414 ( or , more specifically @xmath403 ) this means , that @xmath467 holds , i.e. , @xmath468 . but by construction of @xmath414 we have @xmath461{\ensuremath{i_j}}\vphantom{\ensuremath{i_j}}}}{1}},0,\dots,0 ) \in { r^{0,2}}_n.\ ] ] this is a contradiction , and thus @xmath462 . since @xmath414 is not a trivial relation , there is at least one @xmath462 with non - empty @xmath449 . thus we can assume that @xmath469 . [ lemma : r02intokn ] let @xmath450 , @xmath462 , and @xmath470 . then @xmath453 } \subseteq [ n+1,2n]$ ] , and @xmath454 } \subseteq [ n+1,2n]$ ] . we only consider the first statement ; the second one follows similarly . assume to the contrary that @xmath453 } \not\subseteq [ n+1,2n]$ ] holds . by lemma [ lemma : seperateck ] we have @xmath453 } \subseteq [ n]$ ] . if @xmath471}| \leq 2 $ ] then there is some @xmath455 $ ] with @xmath472 , implying @xmath468 . but by construction of @xmath414 we have @xmath461{\ensuremath{i_j}}\vphantom{\ensuremath{i_j}}}}{1}},0,\dots,0 ) \in { r^{0,2}}_n.\ ] ] thus @xmath471}| \geq 3 $ ] . since @xmath402 has a circular structure , and @xmath473 , we have some @xmath474 $ ] with @xmath475 and @xmath476 . but @xmath477 $ ] and by construction of @xmath414 ( or , more specifically @xmath402 ) this means , that @xmath478 holds , i.e. , @xmath459 and @xmath479 can not both be @xmath380 at the same time . but by construction of @xmath414 we have @xmath461{\ensuremath{i_j}}\vphantom{\ensuremath{i_j}}}}{1}},0,\dots,0,{\underset{\substack{\uparrow \\ \makebox[0pt]{\ensuremath{i_{j'}}}\vphantom{\ensuremath{i_{j'}}}}}{1}},0,\dots,0 ) \in { r^{0,2}}_n.\ ] ] this is a contradiction , and thus @xmath453 } \not\subseteq [ n]$ ] . this shows that in the right hand side of the variables @xmath480 are inessential . but this contradicts the fact , that these variables are essential in @xmath414 . thus follows : [ theorem : t02:nonconstructible ] @xmath481 . let @xmath482 be non - empty sets . then @xmath483 now we will look at the intervals @xmath236 for all subclones @xmath484 . we use the fact that @xmath485 , and let @xmath486 . in this way the conditions [ enum : strongdownwards : totalempty ] and [ enum : strongdownwards : qihungry ] of lemma [ lemma : strongdownwards ] are fulfilled due to corollary [ corollary : emptysetisveryhungry ] . the only condition left to show is [ enum : strongdownwards : different ] for the set @xmath240 defined by @xmath487 let @xmath445 and @xmath488 . then @xmath489 . we need to show that there is some @xmath490 by theorem [ theorem : t02:nonconstructible ] we have that there is some @xmath491-ary partial function @xmath492 if @xmath493 , then @xmath494 and thus we are done . we now assume that @xmath495 . since @xmath496 there is some matrix @xmath158 with columns @xmath497 and rows @xmath498 such that * @xmath499 , and * @xmath500 . let @xmath501 be defined by @xmath502 and @xmath503 for all @xmath504 $ ] . thus we see that @xmath505 if @xmath506 , then @xmath507 and thus we are done . thus there is some @xmath508 $ ] , such that @xmath509 . we define the @xmath510-ary partial function @xmath31 by @xmath511 \setminus \{j\ } \ } , \\ g(1,\yy_j ) & : = f'(\yy_j ) , \\ g(0,\yy_i ) & : = f'(\yy_i ) \text { for all } i \in [ 2n ] \setminus \{j\}. \end{aligned}\ ] ] as @xmath512 we see that @xmath513 . because @xmath514{\ensuremath{j}}\vphantom{\ensuremath{j}}}}{1}},0,\dots,0 ) \in { r^{0,2}}_n$ ] we have @xmath515 but all points are defined . therefore @xmath516 holds , and this implies @xmath517 if @xmath518 , then @xmath519 and thus we are done . otherwise , repeating the steps from @xmath520 to @xmath31 yields finally a desired @xmath521 . [ corollary : subt02:itdifferent ] let @xmath246 with @xmath247 . then @xmath245 . [ theorem : subt02 ] let @xmath484 be a clone on @xmath522 . then @xmath236 has the cardinality of the continuum . by corollary [ corollary : subt02:itdifferent ] and the properties of @xmath241 mentioned at the beginning of this subsection all conditions of lemma [ lemma : strongdownwards ] hold , and therefore @xmath277 . from the results of the previous sections we see that the clones @xmath523 , @xmath524 , @xmath525 , and @xmath526 are the only clones for which we need to determine the size of @xmath4 . we will show in this section that @xmath527 and @xmath528 have both the cardinality of the continuum . by symmetry of post s lattice this implies the same statement for @xmath529 and @xmath530 . creignou , kolaitis and zanuttini have given in @xcite the set of relations defining the smallest element in the interval @xmath4 for each boolean clone @xmath0 . they call these the plain basis . since the least element in @xmath4 is @xmath531 for each total clone @xmath0 , we can conclude from @xcite that @xmath532 where @xmath533 . equivalently , @xmath534 . the clone @xmath523 is denoted by @xmath535 in @xcite , and the plain basis can be found in the entry @xmath536 of table 2 . any @xmath15-ary relation @xmath121 in the partial co - clone of @xmath537 can be constructed from a selection of @xmath538 , i.e. , there are ( possibly empty ) @xmath539-ary auxiliary relations @xmath540 on @xmath108 $ ] for each @xmath541 such that @xmath542 since @xmath538 is totally symmetric on the last @xmath543 coordinates , and @xmath544 , the tuples @xmath545 can be represented by pairs of the form @xmath546 \})$ ] . this notation makes the symmetry of the relation more obvious , and exposes the special element more visibly . for such pairs @xmath547 with @xmath548 $ ] and @xmath549 $ ] we can define the @xmath15-ary relation @xmath550 by @xmath551 we note that @xmath552 whenever @xmath553 , due to the tautology @xmath554 in the definition of @xmath550 . let @xmath555 , j \subseteq [ n ] , i \notin j \}$ ] . then we define the relation @xmath556 $ ] by @xmath557 then equation holds if and only if there is some suitable @xmath558 with @xmath559 let @xmath548 $ ] , @xmath560 $ ] . then @xmath561 . follows from the definition . in this subsection we will define relations @xmath562 for @xmath563 , which will be independent and be used to show that @xmath527 has the cardinality of the continuum . the relations @xmath562 will be called _ monsters _ , as they `` kill '' this problem . let @xmath563 . we define @xmath564 , j \subseteq [ m+1 ] , i \notin j \}$ ] and @xmath565 $ ] by @xmath566 ) \ } \cup { } \\ & \phantom{{}:= { } } \ { ( i,\{1,j_1,j_2\ } ) \mid i , j_1,j_2 \in [ 2,m+1 ] , |\{i , j_1,j_2\}| = 3 \ } , \\ { r^{\lambda}}_m & : = \lambda^{m+1}_{\gamma_m}. \intertext{a more visual represention of $ \gamma_4 $ and $ { r^{\lambda}}_4 $ is given in table~\ref{table : gamma4}.\newline furthermore , we define the ternary relation $ { r^{\lambda}}_\lambda$ by } { r^{\lambda}}_\lambda & : = \{(0,0,0),(0,0,1),(0,1,0),(1,1,1 ) \}.\end{aligned}\ ] ] @xmath567 as shown by blochina in @xcite ( see also section 10.2 @xcite ) the relation @xmath568 characterizes the clone @xmath523 , i.e. , @xmath569 now we give some properties of the relations @xmath562 . let @xmath563 . then @xmath570 . we have the following connections : @xmath571 thus @xmath568 is constructible from @xmath562 as @xmath572 thus @xmath570 by theorem [ theorem : definability ] . . then @xmath574 , i.e. , @xmath575 . since @xmath562 can be constructed from @xmath576 and @xmath577 we have @xmath578 , and thus @xmath579 . on the other hand we have @xmath570 and thus @xmath580 . [ lemma : rlambda : properties ] let @xmath563 . then the following properties hold . 1 . [ enum : rlambda : properties:1inr ] @xmath581 . 2 . [ enum : rlambda : properties:11011notinr ] @xmath582{\ensuremath{i}}\vphantom{\ensuremath{i}}}}{0}},1,\dots,1 ) \notin { r^{\lambda}}_m$ ] for all @xmath583 $ ] . [ enum : rlambda : properties:0xinr ] @xmath584 . [ enum : rlambda : properties:00100100inr ] @xmath585{\ensuremath{i}}\vphantom{\ensuremath{i}}}}{1}},0,\dots,0,{\underset{\substack{\uparrow \\ \makebox[0pt]{\ensuremath{j}}\vphantom{\ensuremath{j}}}}{1}},0,\dots,0 ) \in { r^{\lambda}}_m$ ] for all @xmath586 $ ] with @xmath587 . * since @xmath588 for any @xmath16 and @xmath589 , we have @xmath590 . * if @xmath591 , then @xmath592 ) } \supseteq { r^{\lambda}}_m$ ] . otherwise , if @xmath593 then @xmath594 . thus @xmath595 . * by the definition of @xmath596 we see that @xmath597 if @xmath598 . thus @xmath599 ) } \\ & = \ { 0 \ } \times ( { { \mathbf{2}}}^m \setminus \{(1,\dots,1)\ } ) . \end{aligned}\ ] ] * let @xmath600 . since the set @xmath589 contains at least three elements for every condition @xmath596 in the construction of @xmath562 , there is some @xmath601 with @xmath602 . thus @xmath603 , and consequently @xmath604 . similar to the case of @xmath398 we want to show that there are continuum many strong partial clones with total part equal to @xmath523 . let @xmath605 . let @xmath445 and @xmath446 for the rest of this section . we want to show that @xmath606 holds . we assume to the contrary , that is false . this means that by theorem [ theorem : definability ] we can write @xmath607 with some auxiliary relations @xmath449 for all @xmath450 . furthermore , we can assume that no condition is superfluous . [ lemma : gammamnoidents ] let @xmath563 , @xmath452 , and distinct @xmath608 $ ] . then @xmath609 . assume to the contrary that there are distinct @xmath608 $ ] with @xmath464 . there are a few cases distinguished by the size of the set @xmath610 $ ] . for each @xmath611 $ ] let @xmath612 \mid i_y = i_x \}$ ] . * @xmath613 . since @xmath614 the condition @xmath615 is superfluous in contradiction to the assumption for . * @xmath616 . we have three subcases . * * @xmath617 . then set @xmath618{\ensuremath{i_2}}\vphantom{\ensuremath{i_2}}}}{1}},0,\dots,0 ) \in { r^{\lambda}}_n$ ] . but we have @xmath619 , i.e. , this case can not appear in the construction of @xmath620 . * * @xmath621 . for each constraint @xmath622 in the construction of @xmath562 we have some @xmath623 . thus @xmath624 \setminus t_1 $ ] , i.e. , these coordinates get identified . therefore this constraint is superfluous . + since this holds for every such constraint the complete condition @xmath625 is superfluous . * * @xmath626 . let @xmath627 with @xmath628 and @xmath629 . let @xmath630 \setminus t_1)$ ] , and define @xmath618{\ensuremath{i_1}}\vphantom{\ensuremath{i_1}}}}{1}},0,\dots,0 ) \in { r^{\lambda}}_n$ ] . + from @xmath631 and @xmath632 follows that @xmath633 . this contradicts @xmath634 and therefore this case can not happen . * @xmath635 . + since there are distinct @xmath608 $ ] with @xmath609 there is some @xmath611 $ ] with @xmath636 . + let @xmath637 . since @xmath635 we have @xmath638 $ ] , and thus the proof for @xmath639 and @xmath626 works if we replace @xmath640 by @xmath641 . [ corollary : gammamemptyforbigm ] let @xmath642 . then @xmath643 . [ lemma : gammamnooneforsmallm ] let @xmath462 and @xmath452 . then @xmath644 $ ] . assume to the contrary , that @xmath645 $ ] . by lemma [ lemma : gammamnoidents ] there are no identifications , i.e. , @xmath609 for all distinct @xmath608 $ ] . there are two cases * @xmath646 . we may assume w.l.o.g . that @xmath647 for all @xmath611 $ ] . + we define @xmath648 . then @xmath649 but @xmath634 . thus this contradicts . * @xmath650 for some @xmath651 $ ] . w.l.o.g . let @xmath652 . + let @xmath653 with @xmath654 and define @xmath655{\ensuremath{u_1}}\vphantom{\ensuremath{u_1}}}}{1}},0,\dots,0,{\underset{\substack{\uparrow \\ \makebox[0pt]{\ensuremath{u_2}}\vphantom{\ensuremath{u_2}}}}{1}},0,\dots,0,{\underset{\substack{\uparrow \\ \makebox[0pt]{\ensuremath{u_3}}\vphantom{\ensuremath{u_3}}}}{1}},0,\dots,0).\ ] ] then @xmath656 since @xmath657 . but @xmath634 . thus this contradicts . [ theorem : lambda : nonconstructible ] let @xmath445 and @xmath446 . then @xmath658 . from corollary [ corollary : gammamemptyforbigm ] and lemma [ lemma : gammamnooneforsmallm ] follows that @xmath644 $ ] for all @xmath452 and @xmath450 . thus in the right hand side of the variable @xmath659 is inessential . but this contradicts the fact , that this variables is essential in @xmath620 . therefore is not true , and by theorem [ theorem : definability ] follows the statement of this theorem . let @xmath482 non - empty sets . then @xmath660 from this follows that @xmath240 has continuum cardinality and with @xmath661 we obtain the following statement . [ theorem : lambdacontinuum ] the interval @xmath527 has the cardinality of the continuum . [ theorem : lambdat1continuum ] the interval @xmath528 has the cardinality of the continuum . we have @xmath662 . thus lemma [ lemma : diffconstants ] is applicable , and by [ theorem : lambdacontinuum ] follows that @xmath528 has the cardinality of the continuum . combining theorems [ theorem : subt02 ] , [ theorem : lambdacontinuum ] , [ theorem : lambdat1continuum ] , [ theorem : subl ] , and [ theorem : finiteintervals ] we obtain the dichotomy result for intervals of strong partial clones . let @xmath0 be a total boolean clone . then @xmath4 is either finite or has the cardinality of the continuum . furthermore , @xmath4 is finite if and only if @xmath102 or @xmath103 . does the dichotomy between finite intervals and intervals of continuum cardinality also hold if we consider the clones on some set @xmath5 with @xmath663 ? or , do there exists some @xmath5 and some total clone @xmath0 in @xmath11 such that the interval @xmath4 is countably infinite ? another question in this direction is concerning the two different intervals @xmath2 and @xmath4 for some total clone @xmath0 in @xmath11 . clearly , @xmath664 holds . in the boolean case for each total clone @xmath0 either both intervals are finite , or both intervals have the cardinality of the continuum . but is this also the case on every @xmath5 with @xmath663 ? for some subclones of @xmath321 , and ( in principle ) all subclones of @xmath523 and @xmath525 , respectively , we have shown a strong relation between the intervals . let @xmath665 and @xmath237 a total boolean clone with @xmath238 and @xmath666 . then there is some partial class @xmath241 , such that @xmath667 and @xmath668 hold for all @xmath306 . this means that there is some order - preserving embedding of the interval @xmath4 into @xmath236 . the author would be interested , if such an embedding is possible for all pairs boolean clones @xmath0 and @xmath237 with @xmath238 ? since in this paper the structure of the lattice was used explicitely , for example for the subclones of @xmath321 , a more difficult question arises : if the embedding is possible , can this be proven in general without directly using the description of all clones ? what about this statement for some @xmath5 with @xmath663 ? the partial classes introduced in section [ section : partialclasses ] are an equivalent of the classes considered by harnau in @xcite . in there he presents the galois connection and also describes the closure operator for the relation pairs . the difference on the relational side between clones and strong partial clones is the omission of the projection operator . does this also work when switching from classes to strong partial classes ? 10 .. and .. . . , 6(4):5879 , 1994 . . .. . . , 16 , 1970 . m. couceiro , l. haddad , k. schlzel , and t. waldhauser . a solution to a problem of d. lau : complete classification of intervals in the lattice of partial boolean clones . in _ multiple - valued logic ( ismvl ) , 2013 ieee 43rd international symposium on _ , pages 123128 , 2013 . nadia creignou , phokion kolaitis , and bruno zanuttini . structure identification of boolean relations and plain bases for co - clones . , 74(7):11031115 , november 2008 . j. fugere and l. haddad . on partial clones containing all idempotent partial operations . in _ multiple - valued logic , 1998 . proceedings . 1998 28th ieee international symposium on _ , pages 369373 , 1998 . l. haddad and d. lau . . , 48(1):257280 , 2007 . l. haddad , d. lau , and i.g . , 11(4):399421 , 2006 . l. haddad and g.e . simons . . in _ multiple - valued logic , 33rd ieee international symposium on _ , pages 315320 , 2003 . w. harnau . . , ( 28):517 , 1985 . w. harnau . . , ( 31):1120 , 1987 . w. harnau . . , ( 32):1524 , 1987 . , 2006 . d. lau and k. schlzel . . , pages 189194 , 2010 . .. . . , 2:111 , 1981 . . b. strauch . . , 1995 . b. strauch . . , 1996 . b. strauch . b. strauch . , 43(4):510524 , 1997 . in the figures [ figure : mt ] and [ figure : st ] we present the two finite intervals @xmath95 and @xmath96 , respectively . these were given in @xcite , but the drawings have been improved to show the structure in a better way . the following short - hand notation is used for some of these partial boolean clones . all unlabeled points can be written as the intersection of some of these .
the following result has been shown recently in the form of a dichotomy : for every total clone @xmath0 on @xmath1 , the set @xmath2 of all partial clones on @xmath3 whose total component is @xmath0 , is either finite or of continuum cardinality . in this paper we show that the dichotomy holds , even if only strong partial clones are considered , i.e. , partial clones which are closed under taking subfunctions : for every total clone @xmath0 on @xmath3 , the set @xmath4 of all strong partial clones on @xmath3 whose total component is @xmath0 , is either finite or of continuum cardinality .
one of the most important predictions of lattice qcd is the existence of a transition from ordinary hadronic matter , where quarks and gluons are confined , to a quark gluon plasma at high temperatures . current estimates are that this transition takes place at @xmath0 mev , and is most likely a crossover at zero chemical potential . at higher chemical potential a first - order phase transition is predicted , ending in a tricritical point at @xmath1 mev . these predictions are currently being put to the test at rhic and other heavy - ion colliders . however , interpreting the results of these experiments and comparing them to lattice qcd predictions is far from straightforward . one important reason for this is that the process to a large extent takes place out of thermal equilibrium , and equilibrium field theory methods such as lattice monte carlo are therefore not sufficient . one major , still unsolved puzzle of heavy - ion physics is whether the system ever actually reaches thermal equilibrium , and if so , what the equilibration time is . many aspects of the collision can be successfully described ( eg , using hydrodynamics @xcite ) by assuming a very short equilibration time of 1 @xmath2 , but it is far from understood how this would come about . clearly , a proper understanding of the thermalisation process is essential if we are to have a coherent and reliable description of the heavy - ion collision . at late stages of the collision process , as the system expands , the particles eventually decouple . this typically occurs in two phases : first , inelastic collisions cease , causing the ratios of different particle species to be fixed ( chemical freeze - out ) ; later , also the mean free path for elastic collisions becomes too large and the momentum distribution of the particles deviates from that of thermal equilibrium ( thermal or kinetic freeze - out ) . it is at this point that the final particle yields of the collision are fixed . thus , non - equilibrium dynamics is needed to understand also this aspect of the process . another field where non - equilibrium field theory is needed , is early - universe physics . examples of non - equilibrium processes which require a field - theoretical approach include ( p)reheating , electroweak and qcd phase transitions in the early universe , and baryogenesis . field theories out of equilibrium is a notoriously difficult problem to study nonperturbatively . a large number of approaches have been employed , including hartree and large - n approximations @xcite , dyson schwinger - related approaches based on the 2pi effective action @xcite , and kinetic theory @xcite . all of these have their strengths as well as drawbacks and limitations . within its area of applicability , the classical approximation has the advantage of being fully nonperturbative , easy and relatively inexpensive to implement numerically , and straightforwardly applicable to gauge theories . the major drawback is obviously that quantum effects are not taken into account . classical statistical physics is quantum statistical physics in the limit of large occupation numbers . it follows that the classical approximation can be used when the occupation numbers of the relevant or dominant modes of the system are large . in the context of heavy - ion collisions it is arguable that the multiplicity of soft gluons in the initial ( pre - equilibrium ) stages is very high , and the classical approximation may therefore be valid . the classical approximation has also been applied to the chiral dynamics during freeze - out . in the linear sigma model and related models , the effective potential of the chiral @xmath3 field changes from symmetric to mexican - hat type as the temperature drops , giving rise to an instability where the low - momentum modes increase exponentially . in this scenario , the resulting high occupation numbers justify the use of the classical approximation . a particular hazard with simulating classical dynamics on a lattice is connected with high - momentum lattice artefacts . these will in general interact with the soft modes which carry the interesting physics and for which the classical approximation is in principle valid . if there is sufficient strength in the hard modes , they may equilibrate classically with the soft modes on a much shorter timescale than that of the interesting physics @xcite . in that case , not only the hard modes , but the entire system will be dominated by classical lattice artefacts . to avoid this problem , it is important that the high - momentum modes should be , and remain , strongly suppressed . at early stages , this can be ensured by choosing appropriate initial conditions . the initial conditions are a crucial part of the simulation . they should reflect the salient features of the system at the outset . this is also the only place where information about the quantum nature of the real world enters into the simulation . a quantum system may be represented as an ensemble of classical configurations initially distributed according to quantum statistics @xcite . one example of this may be to choose the 2-point correlators of the fields and their canonical momenta to obey the bose einstein distribution for free fields at some temperature @xmath4 , after subtracting the quantum vacuum fluctuations . for scalar fields @xmath5 with momentum fields @xmath6 and mass @xmath7 one would then have @xmath8 where @xmath9 . such an initialisation also provides an exponential cutoff for the hard modes , which will help in avoiding the dangerous lattice artefacts . following this , each configuration evolves independently according to the classical hamiltonian equations of motion , and time - dependent correlators are computed as averages over the initial conditions . given sufficient time , the system will eventually thermalise classically , resulting in classical equipartition @xmath10 and giving rise to rayleigh jeans type divergences . the hope is that this will happen on much longer time scales than those under consideration in the simulation . there is no unique definition of local particle numbers and energies for interacting fields out of equilibrium . still , the system may exhibit effective particle - like behaviour , which may be used to characterise the approach to thermal equilibrium or to an equilibrium - like distribution . given a definition of local particle numbers , these can also be used to give an effective description of the system in terms of kinetic theory . the effective particle numbers may be extracted from the two - point field correlators , which in the free - field case ( where the particle description is appropriate and well - defined ) contain all the information there is about the system . for example , for a homogeneous , free scalar field we have @xmath11 where @xmath12 is the occupation number for the mode with momentum @xmath13 and @xmath14 is the associated energy . for interacting fields , this may in turn be used as a _ definition _ of the instantaneous particle numbers and energies @xmath15 and @xmath16 @xcite : @xmath17 in the classical approximation the @xmath18 is left out . in a non - abelian gauge theory , the correlation functions will in general be gauge dependent , so the distribution functions will contain ambiguities due to the gauge choice . this ambiguity may be removed by constructing gauge invariant correlators using parallel transporters ; however , this introduces path dependence . in particular , in a lattice regularisation there is in general no one preferred path between two points . although the distribution functions are not unique , all physical observables extracted from them , such as masses , temperatures and chemical potentials , should not depend on the definition and in particular on the gauge . as long as these quantities are not well - defined on the other hand ( such as when the system is very far from equilibrium and the quasiparticle picture does not apply ) , one may expect `` masses '' and `` temperatures '' to be definition - dependent . thus , studying the gauge dependence ( or path dependence ) of distribution functions may serve the double purpose of monitoring the approach to equilibrium and verifying the validity of the approach used . one natural choice of gauge is the coulomb gauge , which is a smooth gauge . in a system with spontaneously broken gauge symmetry ( e.g. , a higgs system ) , the unitary gauge , where the ( effective ) higgs field has only one non - zero , real component , is another natural choice . other gauges , such as maximal abelian gauge , axial gauges or random gauge , may also be considered . in the coulomb gauge , the gauge potential @xmath19 ( but not its conjugate momentum @xmath20 ) is purely transverse , and it can be shown that the transverse free correlators behave analogously to the scalar case @xcite . thus the particle numbers and energies can be defined as @xmath21 here @xmath22 are the transverse @xmath23- and @xmath24-correlators respectively , constructed from the two - point functions @xmath25 according to @xmath26 + + in figure [ fig : n - gauge ] the gauge dependence of effective particle numbers is illustrated in the su(2)higgs model @xcite . in this case , the system was prepared in such a way that all the energy initially was in the higgs field , while the gauge potential was initialised to zero . since the angular modes of the higgs fields are absorbed into the gauge fields in the unitary gauge , the inital occupation numbers are very different . however , already after @xmath27 the two distributions appear almost identical . however , while the particle numbers in the coulomb gauge change very little from here on , in the unitary gauge they continue to fluctuate and it is only from @xmath28 on that one with some confidence can claim the numbers are gauge independent . this agrees roughly with the point where the dispersion relation in the unitary gauge begins to show stable , particle - like behaviour in this particular model . in a non - homogeneous system , and in general in a kinetic - theory description , it is appropriate to think in terms of _ local _ particle numbers @xmath29 . these may be related to the wigner functions constructed from gauge invariant two - point functions @xcite , or more generally to the two - point functions fourier transformed on a region @xmath30 centred on @xmath31 . in the case of a scalar theory we may have ( suppressing the common @xmath32-coordinate for brevity ) @xmath33 here @xmath34 is the volume of the region @xmath30 . this coarse - graining creates an intrinsic unsharpness in the momentum @xmath13 and position @xmath31 of the quasiparticles , given by the size ( and shape ) of the region @xmath35 . if the system under consideration is homogeneous , we may improve statistics by performing an average over all space . this can be shown to be equivalent to local averaging in momentum space , with a weight function @xmath36 depending on the size and shape of @xmath35 : @xmath37 where @xmath38 denotes the correlation function evaluated on the total volume @xmath39 , and the sum is over the discrete momenta available on this volume . in practice , it is simpler to work backwards , choosing a simple form of momentum - space averaging which may correspond to rather complicated spatial regions . for instance , binning in the absolute value of the momentum , @xmath40 corresponds to spherical shells with thickness approximately @xmath41 in position space . at the earliest stages of heavy - ion collisions , the gluon density is expected to be so high that the classical approximation can be justified . the same approximation also justifies ignoring the back - reaction of the quarks , since their number density will be much lower ; leaving us with classical yang mills equations of motion , which may be solved numerically on a lattice . the lattice equations of motion in the temporal gauge ( @xmath42 ) read @xmath43\ ] ] where @xmath44\ ] ] is the canonical momentum to @xmath45 . here , @xmath46 denotes the backward lattice derivative , while @xmath47 is the backward covariant lattice derivative . the equations of motion for @xmath48 constitute the gauss constraint , @xmath49 which must be satisfied by the initial conditions but is conserved by the equations of motion . the initial gluon fields should be related to the gluon distributions of the two colliding nuclei : in principle they should just be the superposition of two lorentz - boosted nuclear gluon distributions . simulations have been carried out over a number of years by krasnitz , nara and venugopalan @xcite ( see also @xcite ) using the `` colour glass condensate '' model of the nuclear wave function to provide the initial conditions . in these studies , the numerical work has been simplified by considering only the mid - rapidity region where the physics is assumed to be boost - invariant . this reduces the system to effectively 2 + 1 dimensions . with these assumptions , the authors have been able to provide an estimate of the initial energy density and gluon distribution which may be used as input into hydrodynamic or kinetic calculations . an alternative approach would be to determine the nuclear gluon field from e.g. a bag model , give this a lorentz boost , and perform a 2 + 1 + 1-dimensional simulation with the longitudinal lattice spacing @xmath50 , where @xmath51 is the lattice spacing in the transverse ( @xmath52 ) direction . work is underway to implement this . the classical approximation may be applied to a range of problems in non - equilibrium field theory where occupation numbers are high , such as the earliest stages of heavy - ion collisions . it has the advantage of being non - perturbative and computationally relatively inexpensive . effective particle numbers may be defined out of equilibrium in a self - consistent manner , and their gauge dependence ( or that of derived quantities such as masses and temperatures ) can be used as a check on the validity of the quasi - particle picture . this work was supported by fom / nwo . i am thankful to jan smit and anders tranberg for numerous fruitful discussions . p. f. kolb and u. heinz , nucl - th/0305084 . f. cooper _ et al . _ , phys . rev . * d50 * , 2848 ( 1994 ) [ hep - ph/9405352 ] . d. boyanovsky , h. j. de vega and r. holman , phys . rev . * d51 * , 734 ( 1995 ) [ hep - ph/9401308 ] . j. berges and j. cox , phys b517 * , 369 ( 2001 ) [ hep - ph/0006160 ] . j. berges and j. serreau , hep - ph/0302210 . r. baier , a. h. mueller , d. schiff and d. t. son , phys . lett . * b502 * , 51 ( 2001 ) [ hep - ph/0009237 ] . g. d. moore , jhep * 11 * , 021 ( 2001 ) [ hep - ph/0109206 ] . m. sall , j. smit and j. c. vink , phys . rev . * d64 * , 025016 ( 2001 ) [ hep - ph/0012346 ] . g. aarts and j. smit , phys . rev . * d61 * , 025002 ( 2000 ) [ hep - ph/9906538 ] . skullerud , j. smit and a. tranberg , jhep * 08 * , 045 ( 2003 ) [ hep - ph/0307094 ] . a. krasnitz and r. venugopalan , phys . * 84 * , 4309 ( 2000 ) [ hep - ph/9909203 ] . a. krasnitz , y. nara and r. venugopalan , nucl . phys . * a717 * , 268 ( 2003 ) [ hep - ph/0209269 ] . t. lappi , phys . rev . * c67 * , 054903 ( 2003 ) [ hep - ph/0303076 ] .
the classical approximation may be applied to a number of problems in non - equilibrium field theory . the principles and limits of classical real - time lattice simulations are presented , with particular emphasis on the definition of particle numbers and energies and on applications to the earliest stages of heavy - ion collisions . mathmargin = 0pt
the united nations department of economic and social affairs estimates that the population of the world exceeded six billion people for the first time on october 12 , 1999 . there is no doubt that the world of human society has become quite large in recent times . nonetheless , people routinely claim that , global statistics notwithstanding , it s still a small world . and in a certain sense they are right . despite the enormous number of people on the planet , the structure of social networks the map of who knows whom is such that we are all very closely connected to one another ( kochen 1989 , watts 1999 ) . one of the first quantitative studies of the structure of social networks was performed in the late 1960s by stanley milgram , then at harvard university ( milgram 1967 ) . he performed a simple experiment as follows . he took a number of letters addressed to a stockbroker acquaintance of his in boston , massachusetts , and distributed them to a random selection of people in nebraska . ( evidently , he considered nebraska to be about as far as you could get from boston , in social terms , without falling off the end of the world . ) his instructions were that the letters were to be sent to their addressee ( the stockbroker ) by passing them from person to person , and that , in addition , they could be passed only to someone whom the passer new on a first - name basis . since it was not likely that the initial recipients of the letters were on a first - name basis with a boston stockbroker , their best strategy was to pass their letter to someone whom they felt was nearer to the stockbroker in some social sense : perhaps someone they knew in the financial industry , or a friend in massachusetts . a reasonable number of milgram s letters did eventually reach their destination , and milgram found that it had only taken an average of six steps for a letter to get from nebraska to boston . he concluded , with a somewhat cavalier disregard for experimental niceties , that six was therefore the average number of acquaintances separating the pairs of people involved , and conjectured that a similar separation might characterize the relationship of any two people in the entire world . this situation has been labeled `` six degrees of separation '' ( guare 1990 ) , a phrase which has since passed into popular folklore . given the form of milgram s experiment , one could be forgiven for supposing that the figure six is probably not a very accurate one . the experiment certainly contained many possible sources of error . however , the general result that two randomly chosen human beings can be connected by only a short chain of intermediate acquaintances has been subsequently verified , and is now widely accepted ( korte and milgram 1970 ) . in the jargon of the field this result is referred to as the * small - world effect*. the small - world effect applies to networks other than networks of friends . brett tjaden s parlor game `` the six degrees of kevin bacon '' connects any pair of film actors via a chain of at most eight co - stars ( tjaden and wasson 1997 ) . tom remes has done the same for baseball players who have played on the same team ( remes 1997 ) . with tongue very firmly in cheek , the _ new york times _ played a similar game with the the names of those who had tangled with monica lewinsky ( kirby and sahre 1998 ) . all of this however , seems somewhat frivolous . why should a serious scientist care about the structure of social networks ? the reason is that such networks are crucially important for communications . most human communication where the word is used in its broadest sense takes place directly between individuals . the spread of news , rumors , jokes , and fashions all take place by contact between individuals . and a rumor can spread from coast to coast far faster over a social network in which the average degree of separation is six , than it can over one in which the average degree is a hundred , or a million . more importantly still , the spread of disease also occurs by person - to - person contact , and the structure of networks of such contacts has a huge impact on the nature of epidemics . in a highly connected network , this year s flu or the hiv virus can spread far faster than in a network where the paths between individuals are relatively long . in this paper we outline some recent developments in the theory of social networks , particularly in the characterization and modeling of networks , and in the modeling of the spread of information or disease . the simplest explanation for the small - world effect uses the idea of a random graph . suppose there is some number @xmath0 of people in the world , and on average they each have @xmath1 acquaintances . this means that there are @xmath2 connections between people in the entire population . the number @xmath1 is called the * coordination number * of the network . we can make a very simple model of a social network by taking @xmath0 dots ( `` nodes '' or `` vertices '' ) and drawing @xmath2 lines ( `` edges '' ) between randomly chosen pairs to represent these connections . such a network is called a * random graph * ( bollobs 1985 ) . random graphs have been studied extensively in the mathematics community , particularly by erds and rnyi ( 1959 ) . it is easy to see that a random graph shows the small - world effect . if a person a on such a graph has @xmath1 neighbors , and each of a s neighbors also has @xmath1 neighbors , then a has about @xmath3 second neighbors . extending this argument a also has @xmath4 third neighbors , @xmath5 fourth neighbors and so on . most people have between a hundred and a thousand acquaintances , so @xmath5 is already between about @xmath6 and @xmath7 , which is comparable with the population of the world . in general the number @xmath8 of degrees of separation which we need to consider in order to reach all @xmath0 people in the network ( also called the * diameter * of the graph ) is given by setting @xmath9 , which implies that @xmath10 . this logarithmic increase in the number of degrees of separation with the size of the network is typical of the small - world effect . since @xmath11 increases only slowly with @xmath0 , it allows the number of degrees to be quite small even in very large systems . as an example of this type of behavior , albert ( 1999 ) studied the properties of the network of `` hyperlinks '' between documents on the world wide web . they estimated that , despite the fact there were @xmath12 documents on the web at the time the study was carried out , the average distance between documents was only about 19 . there is a significant problem with the random graph as a model of social networks however . the problem is that people s circles of acquaintance tend to overlap to a great extent . your friend s friends are likely also to be your friends , or to put it another way , two of your friends are likely also to be friends with one another . this means that in a real social network it is not true to say that person a has @xmath3 second neighbors , since many of those friends of friends are also themselves friends of person a. this property is called * clustering * of networks . .the number of nodes @xmath0 , average degree of separation @xmath13 , and clustering coefficient @xmath14 , for three real - world networks . the last column is the value which @xmath14 would take in a random graph with the same size and coordination number . [ wstable ] [ cols="<,^,^,^,^",options="header " , ] a random graph does not show clustering . in a random graph the probability that two of person a s friends will be friends of one another is no greater than the probability that two randomly chosen people will be . on the other hand , clustering _ has _ been shown to exist in a number of real - world networks . one can define a * clustering coefficient * @xmath14 , which is the average fraction of pairs of neighbors of a node which are also neighbors of each other . in a fully connected network , in which everyone knows everyone else , @xmath15 ; in a random graph @xmath16 , which is very small for a large network . in real - world networks it has been found that , while @xmath14 is significantly less than 1 , it is much greater than @xmath17 . in table , we show some values of @xmath14 calculated by watts and strogatz ( 1998 ) for three different networks : the network of collaborations between movie actors discussed previously , the neural network of the worm _ c. elegans , _ and the western power grid of the united states . we also give the value @xmath18 which the clustering coefficient would have on random graphs of the same size and coordination number , and in each case the measured value is significantly higher than for the random graph , indicating that indeed the graph is clustered . in the same table we also show the average distance @xmath13 between pairs of nodes in each of these networks . this is not the same as the diameter @xmath8 of the network discussed above , which is the _ maximum _ distance between nodes , but it also scales at most logarithmically with number of nodes on random graphs . this is easy to see , since the average distance is strictly less than or equal to the maximum distance , and so @xmath13 can not increase any faster than @xmath8 . as the table shows , the value of @xmath13 in each of the networks considered is small , indicating that the small - world effect is at work . ( the precise definition of `` small - world effect '' is still a matter of debate , but in the present case a reasonable definition would be that @xmath13 should be comparable with the value it would have on the random graph , which for the systems discussed here it is . ) so , if random graphs do not match well the properties of real - world networks , is there an alternative model which does ? such a model has been suggested by duncan watts and steven strogatz . it is described in the next section . in order to model the real - world networks described in the last section , we need to find a way of generating graphs which have both the clustering and small - world properties . as we have argued , random graphs show the small - world effect , possessing average vertex - to - vertex distances which increase only logarithmically with the total number @xmath0 of vertices , but they do not show clustering the property that two neighbors of a vertex will often also be neighbors of one another . the opposite of a random graph , in some sense , is a completely ordered lattice , the simplest example of which is a one - dimensional lattice a set of vertices arranged in a straight line . if we take such a lattice and connect each vertex to the @xmath1 vertices closest to it , as in fig . a , then it is easy to see that most of the immediate neighbors of any site are also neighbors of one another , i.e. , it shows the clustering property . normally , we apply periodic boundary conditions to the lattice , so that it wraps around on itself in a ring ( fig . b ) , although this is just for convenience and not strictly necessary . for such a lattice we can calculate the clustering coefficient @xmath14 exactly . as long as @xmath19 , which it will be for almost all graphs , we find that @xmath20 which tends to @xmath21 in the limit of large @xmath1 . we can also build networks out of higher - dimensional lattices , such as square or cubic lattices , and these also show the clustering property . the value of the clustering coefficient in general dimension @xmath22 is @xmath23 which also tends to @xmath21 for @xmath24 . low - dimensional regular lattices however do not show the small - world effect of typical vertex vertex distances which increase only slowly with system size . it is straightforward to show that for a regular lattice in @xmath22 dimensions which has the shape of a square or ( hyper)cube of side @xmath25 , and therefore has @xmath26 vertices , the average vertex vertex distance increases as @xmath25 , or equivalently as @xmath27 . for small values of @xmath22 this does not give us small - world behavior . in one dimension for example , it means that the average distance increases linearly with system size . if we allow the dimension @xmath22 of the lattice to become large , then @xmath27 becomes a slowly increasing function of @xmath0 , and so the lattice does show the small - world effect . could this be the explanation for what we see in real networks ? perhaps real networks are roughly regular lattices of very high dimension . this explanation is in fact not unreasonable , although it has not been widely discussed . it works quite well , provided the mean coordination number @xmath1 of the vertices is much higher than twice the dimension @xmath22 of the lattice . ( if we allow @xmath1 to approach @xmath28 , then the clustering coefficient , eq . , tends to zero , implying that the lattice loses its clustering properties . ) watts and strogatz ( 1998 ) however have proposed an alternative model for the small world , which perhaps fits better with our everyday intuitions about the nature of social networks . their suggestion was to build a model which is , in essence , a low - dimensional regular lattice say a one - dimensional lattice but which has some degree of randomness in it , like a random graph , to produce the small - world effect . they suggested a specific scheme for doing this as follows . we take the one - dimensional lattice of fig . b , and we go through each of the links on the lattice in turn and , with some probability @xmath29 , we randomly `` rewire '' that link , meaning that we move one of its ends to a new position chosen at random from the rest of the lattice . for small @xmath29 this produces a graph with is still mostly regular but has a few connections which stretch long distances across the lattice as in fig . c. the coordination number of the lattice is still @xmath1 on average as it was before , although the number of neighbors of any particular vertex can be greater or smaller than @xmath1 . in social terms , we can justify this model by saying that , while most people are friends with their immediate neighbors neighbors on the same street , people that they work with , people that their friends introduce them to some people are also friends with one or two people who are a long way away , in some social sense people in other countries , people from other walks of life , acquaintances from previous eras of their lives , and so forth . these long - distance acquaintances are represented by the long - range links in the model of watts and strogatz . clearly the values of the clustering coefficient @xmath14 for the watts strogatz model with small values of @xmath29 will be close to those for the perfectly ordered lattice given above , which tend to @xmath21 for fixed small @xmath22 and large @xmath1 . watts and strogatz also showed by numerical simulation that the average vertex vertex distance @xmath13 is comparable with that for a true random graph , even for quite small values of @xmath29 . for example , for a random graph with @xmath30 and @xmath31 , they found that the average distance was about @xmath32 between two vertices chosen at random . for their rewiring model , the average distance was only slightly greater , at @xmath33 , when the rewiring probability @xmath34 , compared with @xmath35 for the graph with no rewired links at all . and even for @xmath36 , they found @xmath37 , a little over twice the value for the random graph . thus the model appears to show both the clustering and small - world properties simultaneously . this result has since been confirmed by further simulation as well as analytic work on small - world models , which is described in the next section . most of the recent work on models of the small world has been performed using a variation of the watts strogatz model suggested by newman and watts ( 1999a ) . in this version of the model , instead of rewiring links between sites as in fig . c , extra links , often called * shortcuts * , are added between pairs of sites chosen at random , but no links are removed from the underlying lattice . this model is somewhat easier to analyze than the original watts strogatz model , because it is not possible for any region of the graph to become disconnected from the rest , whereas this can happen in the original model . mathematically a disjoint section of the graph can be represented by saying that the distance from any vertex in that section to a vertex somewhere on the rest of the graph is infinite . however , this means that , when averaged over all possible realizations of the graph , the average vertex vertex distance @xmath13 in the model is also infinite for any finite value of @xmath29 . ( a similar problem in the theory of random graphs is commonly dealt with by averaging the reciprocal of vertex vertex distance , rather than the distance itself , but this approach does not seem to have been tried for the watts strogatz model . ) in fact , it is possible to show that the series expansion of @xmath38 in powers of @xmath29 about @xmath39 is well - behaved up to order @xmath40 , but that the expansion coefficients are infinite for all higher orders . for the version of the model where no links are ever removed , the expansion coefficients take the same values up to order @xmath40 , but are finite for all higher orders as well . generically , both versions of the model are referred to as * small - world models * , or sometimes * small - world graphs*. many results have been derived for small - world models , and many of their other properties have been explored numerically . here we give only a brief summary of the most important results . barthlmy and amaral ( 1999 ) conjectured that the average vertex vertex distance @xmath13 obeys the scaling form @xmath41 , where @xmath42 is a universal scaling function of its argument @xmath43 and @xmath44 is a characteristic length - scale for the model which is assumed to diverge in the limit of small @xmath29 according to @xmath45 . on the basis of numerical results , barthlmy and amaral further conjectured that @xmath46 . barrat ( 1999 ) disproved this second conjecture using a simple physical argument which showed that @xmath47 can not be less than 1 , and suggested on the basis of more numerical results that in fact it was exactly 1 . newman and watts ( 1999b ) showed that the small - world model has only one non - trivial length - scale other than the lattice spacing , which we can equate with the variable @xmath44 above , and which is given by @xmath48 for the one - dimensional model , or @xmath49 in the general case . thus @xmath47 must indeed be 1 for @xmath50 , or @xmath51 for general @xmath22 and , since there are no other length - scales present , @xmath13 must be of the form @xmath52 where @xmath53 is another universal scaling function . ( the initial factor of @xmath54 before the scaling function is arbitrary . it is chosen thus to give @xmath55 a simple limit for small values of its argument see eq . . ) this scaling form is equivalent to that of barthlmy and amaral by the substitution @xmath56 if @xmath57 . it has been extensively confirmed by numerical simulation ( newman and watts 1999a , de menezes 2000 ) and by series expansions ( newman and watts 1999b ) ( see fig . ) . the divergence of @xmath44 as @xmath58 gives something akin to a critical point in this limit . ( de menezes ( 2000 ) have argued that , for technical reasons , we should refer to this point as a `` first order critical point '' ( fisher and berker 1982 ) . ) this allowed newman and watts ( 1999a ) to apply a real - space renormalization group transformation to the model in the vicinity of this point and prove that the scaling form above is exactly obeyed in the limit of small @xmath29 and large @xmath25 . . tells us that although the average vertex vertex distance on a small - world graph appears at first glance to be a function of three parameters@xmath29 , @xmath1 , and @xmath25it is in fact entirely determined by a single scalar function of a single scalar variable . if we know the form of this one function , then we know everything . actually , this statement is strictly only true if @xmath59 , when it is safe to ignore the other length - scale in the problem , the lattice parameter of the underlying lattice . thus , the scaling form is expected to hold only when @xmath29 is small , i.e. , in the regime where the majority of a person s contacts are local and only a small fraction long - range . ( the fourth parameter @xmath22 also enters the equation , but is not on an equal footing with the others , since the functional form of @xmath55 changes with @xmath22 , and thus eq . does not tell us how @xmath13 varies with dimension . ) both the scaling function @xmath53 and the scaling variable @xmath60 have simple physical interpretations . the variable @xmath43 is two times the average number of shortcuts on the graph for the given value of @xmath29 , and @xmath53 is the average fraction by which the vertex vertex distance on the graph is reduced for the given value of @xmath43 . from the results shown in fig . , we can see that it takes about @xmath61 shortcuts to reduce the average vertex - vertex distance by a factor of two , and 56 to reduce it by a factor of ten . in the limit of large @xmath29 the small - world model becomes a random graph or nearly so . hence , we expect that the value of @xmath13 should scale logarithmically with system size @xmath25 when @xmath29 is large , and also , as the scaling form shows , when @xmath25 is large . on the other hand , when @xmath29 or @xmath25 is small we expect @xmath13 to scale linearly with @xmath25 . this implies that @xmath53 has the limiting forms @xmath62 in theory there should be a leading constant in front of the large-@xmath43 form here , but , as discussed shortly , it turns out that this constant is equal to unity . the cross - over between the small- and large-@xmath43 regimes must happen in the vicinity of @xmath63 , since @xmath44 is the only length - scale available to dictate this point . neither the actual distribution of path lengths in the small - world model nor the average path length @xmath13 has been calculated exactly yet ; exact analytical calculations have proven very difficult for the model . some exact results have been given by kulkarni ( 2000 ) who show , for example , that the value of @xmath13 is simply related to the mean @xmath64 and mean square @xmath65 of the shortest distance @xmath66 between two points on diametrically opposite sides of the graph , according to @xmath67 unfortunately , calculating the shortest distance between opposite points is just as difficult as calculating @xmath13 directly , either analytically or numerically . newman ( 2000 ) have calculated the form of the scaling function @xmath53 for @xmath50 small - world graphs using a mean - field - like approximation , which is exact for small or large values of @xmath43 , but not in the regime where @xmath68 . their result is @xmath69 this form is also plotted on fig . ( dotted line ) . since this is exact for large @xmath43 , it can be expanded about @xmath70 to show that the leading constant in the large-@xmath43 form of @xmath53 , eq . , is @xmath71 as stated above . newman also solved for the complete distribution of lengths between vertices in the model within their mean - field approximation . this distribution can be used to give a simple model of the spread of a disease in a small world . if a disease starts with a single person somewhere in the world , and spreads first to all the neighbors of that person , and then to all second neighbors , and so on , then the number of people @xmath72 who have the disease after @xmath73 time - steps is simply the number of people who are separated from the initial carrier by a distance of @xmath73 or less . newman and watts ( 1999b ) previously gave an approximate differential equation for @xmath74 on an infinite small - world graph , which they solved for the one - dimensional case ; moukarzel ( 1999 ) later solved it for the case of general @xmath22 . the mean - field treatment generalizes the solution for @xmath50 to finite lattice sizes . ( a similar mean - field result has been given for a slightly different disease - spreading model by kleczkowski and grenfell ( 1999 ) . ) the resulting form for @xmath74 is shown in the inset of fig . , and clearly has the right general sigmoidal shape for the spread of an epidemic . in fact , this form of @xmath72 is typical also of the standard logistic growth models of disease spread , which are mostly based on random graphs ( sattenspiel and simon 1988 , kretschmar and morris 1996 ) . in the next section we consider some ( slightly ) more sophisticated models of disease spreading on small - world graphs . a variety of authors have looked at dynamical systems defined on small - world graphs built using either the watts strogatz rewiring method or the alternative method described in section . we briefly describe a number of these studies in this section . watts and strogatz ( 1998 , watts 1999 ) looked at cellular automata , simple games , and networks of coupled oscillators on small - world networks . for example , they found that it was much easier for a cellular automaton to perform the task known as density classification ( das 1994 ) on a small - world graph than on a regular lattice ; they found that in an iterated multi - player game of prisoner s dilemma , cooperation arose less frequently on a small - world graph than on a regular lattice ; and they found that the small - world topology helped oscillator networks to synchronize much more easily than in the regular lattice case . monasson ( 1999 ) investigated the eigenspectrum of the laplacian operator on small - world graphs using a transfer matrix method . this spectrum tells us for example what the normal modes would be of a system of masses and springs built with the topology of a small - world graph . or , perhaps more usefully , it can tell us how diffusive dynamics would occur on a small world graph ; any initial state of a diffusive field can be decomposed into eigenvectors which each decay independently and exponentially with a decay constant related to the corresponding eigenvalue . diffusive motion might provide a simple model for the spread of information of some kind in a social network . barrat and weigt ( 2000 ) have given a solution of the ferromagnetic ising model on a @xmath50 small - world network using a replica method . since the ising model has a lower critical dimension of two , we would expect it not to show a phase transition when @xmath39 and the graph is truly one - dimensional . on the other hand , as soon as @xmath29 is greater than zero , the effective dimension of the graph becomes greater than one , and increases with system size ( newman and watts 1999b ) . thus for any finite @xmath29 we would expect to see a phase transition at some finite temperature in the large system limit . barrat and weigt confirmed both analytically and numerically that indeed this is the case . the ising model is of course a highly idealized model , and its solution in this context is , to a large extent , just an interesting exercise . however , the similar problem of a potts antiferromagnet on a general graph has real practical applications , e.g. , in the solution of scheduling problems . although this problem has not been solved on the small - world graph , walsh ( 1999 ) has found results which indicate that it may be interesting from a computational complexity point of view ; finding a ground state for a potts antiferromagnet on a small - world graph may be significantly harder than finding one on either a regular lattice or a random graph . newman and watts ( 1999b ) looked at the problem of disease spread on small - world graphs . as a first step away from the very simple models of disease described in the last section , they considered a disease to which only a certain fraction @xmath75 of the population is susceptible ; the disease spreads neighbor to neighbor on a small - world graph , except that it only affects , and can be transmitted by , the susceptible individuals . in such a model , the disease can only spread within the connected cluster of susceptible individuals in which it first starts , which is small if @xmath75 is small , but becomes larger , and eventually infinite , as @xmath75 increases . the point at which it becomes infinite the point at which an epidemic takes place is precisely the percolation point for site percolation with probability @xmath75 on the small - world graph . newman and watts gave an approximate calculation of this epidemic point , which compares reasonably favorably with their numerical simulations . moore and newman ( 2000a , 2000b ) later gave an exact solution . lago - fernndez ( 2000 ) investigated the behavior of a neural network of hodgkin huxley neurons on a variety of graphs , including regular lattices , random graphs , and small - world graphs . they found that the presence of a high degree of clustering in the network allowed the network to establish coherent oscillation , while short average vertex vertex distances allowed the network to produce fast responses to changes in external stimuli . the small - world graph , which simultaneously possesses both of these properties , was the only graph they investigated which showed both coherence and fast response . kulkarni ( 1999 ) studied numerically the behavior of the bak sneppen model of species coevolution ( bak and sneppen 1993 ) on small - world graphs . this is a model which mimics the evolutionary effects of interactions between large numbers of species . the behavior of the model is known to depend on the topology of the lattice on which it is situated , and kulkarni and co - workers suggested that the topology of the small - world graph might be closer to that of interactions in real ecosystems than the low - dimensional regular lattices on which the bak sneppen model is usually studied . the principal result of the simulations was that on a small - world graph the amount of evolutionary activity taking place at any given vertex varies with the coordination number of the vertex , with the most connected nodes showing the greatest activity and the least connected ones showing the smallest . although most of the work reviewed in this article is based on the watts strogatz small - world model , a number of other models of social networks have been proposed . in section we mentioned the simple random - graph model and in section we discussed a model based on a regular lattice of high dimension . in this section we describe briefly three others which have been suggested . one alternative to the view put forward by watts and strogatz is that the small - world phenomenon arises not because there are a few `` long - range '' connections in the otherwise short - range structure of a social network , but because there are a few nodes in the network which have unusually high coordination numbers ( kasturirangan 1999 ) or which are linked to a widely distributed set of neighbors . perhaps the `` six degrees of separation '' effect is due to a few people who are particularly well connected . ( gladwell ( 1998 ) has written a lengthy and amusing article arguing that a septuagenarian salon proprietor in chicago named lois weisberg is an example of precisely such a person . ) a simple model of this kind of network is depicted in fig . , in which we start again with a one - dimensional lattice , but instead of adding extra links between pairs of sites , we add a number of extra vertices in the middle which are connected to a large number of sites on the main lattice , chosen at random . ( lois weisberg would be one of these extra sites . ) this model is similar to the watts strogatz model in that the addition of the extra sites effectively introduces shortcuts between randomly chosen positions on the lattice , so it should not be surprising to learn that this model does display the small - world effect . in fact , even in the case where only one extra site is added , the model shows the small - world effect if that site is sufficiently highly connected . this case has been solved exactly by dorogovtsev and mendes ( 1999 ) . another alternative model of the small world has been suggested by albert ( 1999 ) who , in their studies of the world wide web discussed in section , concluded that the web is dominated by a small number of very highly connected sites , as described above , but they also found that the distribution of the coordination numbers of sites ( the number of `` hyperlinks '' pointing to or from a site ) was a power - law , rather than being bimodal as it is in the previous model . they produced a model network of this kind as follows . starting with a normal random graph with average coordination number @xmath1 and the desired number @xmath0 of vertices , they selected a vertex at random and added a link between it and another randomly chosen site if that addition would bring the overall distribution of coordination numbers closer to the required power law . otherwise the vertex remains as it is . if this process is repeated for a sufficiently long time , a network is generated with the correct coordination numbers , but which is in other respects a random graph . in particular , it does not show the clustering property of which such a fuss has been made in the case of the watts strogatz model . albert found that their model matched the measured properties of the world wide web quite closely , although related work by adamic ( 1999 ) indicates that clustering is present in the web , so that the model is unrealistic in this respect . it is worth noting that networks identical to those of albert can be generated in a manner much more efficient than the monte carlo scheme described above by simply generating @xmath0 vertices with a power law distribution of lines emerging from them ( using , for instance , the transformation method ( newman and barkema 1999 ) ) , and then joining pairs of lines together at random until none are left . if one were interested in investigating such networks numerically , this would probably be the best way to generate them . a third suggestion has been put forward by kleinberg ( 1999 ) , who argues that a model such as that of watts and strogatz , in which shortcuts connect vertices arbitrarily far apart with uniform probability , is a poor representation of at least some real - world situations . ( kasturirangan ( 1999 ) has made a similar point . ) kleinberg notes that in the real world , people are surprisingly good at finding short paths between pairs of individuals ( milgram s letter experiment , and the kevin bacon game are good examples ) given only local information about the structure of the network . conversely , he has shown that no algorithm exists which is capable of finding such paths on networks of the watts strogatz type , again given only local information . thus there must be some additional properties of real - world networks which make it possible to find short paths with ease . to investigate this question further , kleinberg has proposed a generalization of the watts strogatz model in which the typical distance traversed by the shortcuts can be tuned . kleinberg s model is based on a two - dimensional square lattice ( although it could be generalized to other dimensions @xmath22 in a straightforward fashion ) and has shortcuts added between pairs of vertices @xmath76 with probability which falls off as a power law @xmath77 of the distance between them . ( in this work , @xmath78 is the `` manhattan distance '' @xmath79 , where @xmath80 and @xmath81 are the lattice coordinates of the vertices @xmath82 and @xmath83 . this makes good sense , since this is also the distance in terms of links on the underlying lattice that separates those two points before the shortcuts are added . however , one could in principle generate networks using a different definition of distance , such as the euclidean distance @xmath84 , for example . ) it is then shown that for the particular value @xmath85 of the exponent of the power law ( or @xmath86 for underlying lattices of @xmath22 dimensions ) , there exists a simple algorithm for finding a short path between two given vertices , making use only of local information . for any other value of @xmath87 the problem of finding a short path is provably much harder . this result demonstrates that there is more to the small world effect than simply the existence of short paths . in this article we have given an overview of recent theoretical work on the `` small - world '' phenomenon . we have described in some detail the considerable body of recent results dealing with the watts strogatz small - world model and its variants , including analytic and numerical results about network structure and studies of dynamical systems on small - world graphs . what have we learned from these efforts and where is this line of research going now ? the most important result is that small - world graphs those possessing both short average person - to - person distances and `` clustering '' of acquaintances show behaviors very different from either regular lattices or random graphs . some of the more interesting such behaviors are the following : 1 . these graphs show a transition with increasing number of vertices from a `` large - world '' regime in which the average distance between two people increases linearly with system size , to a `` small - world '' one in which it increases logarithmically . this implies that information or disease spreading on a small - world graph reaches a number of people which increases initially as a power of time , then changes to an exponential increase , and then flattens off as the graph becomes saturated . disease models which incorporate a measure of susceptibility to infection have a percolation transition at which an epidemic sets in , whose position is influenced strongly by the small - world nature of the network . dynamical systems such as games or cellular automata show quantitatively different behavior on small - world graphs and regular lattices . some problems , such as density classification , appear to be easier to solve on small - world graphs , while others , such as scheduling problems , appear to be harder . some real - world graphs show characteristics in addition to the small - world effect which may be important to their function . an example is the world wide web , which appears to have a scale - free distribution of the coordination numbers of vertices . research in this field is continuing in a variety of directions . empirical work to determine the exact structure of real networks is underway in a number of groups , as well as theoretical work to determine the properties of the proposed models . and studies to determine the effects of the small - world topology on dynamical processes , although in their infancy , promise an intriguing new perspective on the way the world works . the author would like to thank luis amaral , marc barthlmy , rahul kulkarni , cris moore , cristian moukarzel , naomi sachs , steve strogatz , toby walsh , and duncan watts for useful discussions and comments . this work was supported in part by the santa fe institute and darpa under grant number onr n00014 - 95 - 1 - 0975 . # 1#2#3#4#5#6 # 1 # 2#3 # 4 * # 5 * , # 6 . # 1#2#3#4 # 1 # 2 # 4 . @xmath88citations of the form cond - mat / xxxxxxx refer to the online condensed matter physics preprint archive at http://www.arxiv.org/. das , r. , mitchell , m. , and crutchfield , j. p. 1994 a genetic algorithm discovers particle - based computation in cellular automata . in _ parallel problem solving in nature , _ davidor , y. , schwefel , h. p. and manner , r. ( eds . ) , springer ( berlin ) . kleinberg , j. 1999 the small - world phenomenon : an algorithmic perspective . cornell university computer science department technical report 991776 . also http://www.cs.cornell.edu/home/kleinber/swn.ps .
it is believed that almost any pair of people in the world can be connected to one another by a short chain of intermediate acquaintances , of typical length about six . this phenomenon , colloquially referred to as the `` six degrees of separation , '' has been the subject of considerable recent interest within the physics community . this paper provides a short review of the topic . # 1([#1 ] ) # 1[#1 ] # 1[#1 ] # 1[#1 ] # 1#1 o # 1 # 1 # 1
the deployment of a cloud - based radio access network ( cran ) is expected to be an integral part of 5 g networks @xcite . in cran , remote radio heads ( rrhs ) , which are small radio antennas responsible for transmission , cooperate to serve their wireless users . to achieve such coordination , the rrhs are connected through a fronthaul link to a control unit ( cu ) that performs efficient signal processing . however , in cran , the fronthaul links are often of finite capacity @xcite , and thus , the number of rrhs connected to the cu will be limited . thus , the cran will need to co - exist with the small cell base stations ( bss ) of an existing heterogeneous cellular network ( hetnet ) . in particular , in the downlink , the presence of a large number of rrhs can increase the interference on hetnet users . thus , it is imperative to devise new solutions for optimized interference management in such a hybrid cran - hetnet architecture @xcite . in the existing literature , the problem of downlink interference mitigation has been considered mainly within the context of hetnets @xcite . the objective is to reduce to the interference caused by the high - powered macro bss on the low - powered femto and pico bss . these techniques range from power control @xcite to the allocation of frequency and time resources such as in @xcite and @xcite . more recent works such as in @xcite considered interference mitigation for a cran using mimo techniques . the authors in @xcite studied the challenge of limited fronthaul capacity in cran , and derived the optimal power allocations that maximize the system s sum - rates under a constrained fronthaul . in @xcite , the authors proposed a centralized resource allocation technique for the downlink of a cran coexisting with a macro bs . however , this existing body of work is either focused on cran or hetnets , in isolation , or in a very restricted setting . thus , none of these works addresses the problem of interference management in hybrid cran / small cell architectures . the main contribution of this paper is to introduce a novel framework for power control and interference management that can ensure an efficient hetnet and cran co - existence . we formulate the problem as a noncooperative game between the cran and hetnet s bss . in this game , each base station seeks to choose its optimal power allocation to maximize the sum - rate of its users . given the heterogeneity of the system , we cast the game within the framework of _ cognitive hierarchy theory _ @xcite in which players can be organized into multi - tier hierarchies depending on their capabilities , unlike in classical stackelberg games that allow only for a two level pre - determined hierarchy @xcite . for our model , the hierarchy pertains to the fact that computationally capable and high - powered bss , such as the cran or the macrocell bss can have a higher hierarchy in the network than smaller bss , such as femtocells . such a hierarchical organization eventually allows reducing the interference on low - powered bss as they no longer need to consider the actions of the devices with higher power capabilities . in contrast , by being at the top of the hierarchy , high - powered bss such as the cran and macrocell bss will be more conservative in their power control decisions . for both the conventional game and the cognitive hierarchy game , we characterize the equilibrium solutions and study their properties . simulation results show that the proposed cognitive hierarchy model yields a two - fold improvement in the total system rate , when compared with a classical game - theoretic model . the results also show considerable interference reduction on the low - powered bss . the paper is organized as follows : section ii presents the system model and the power allocation problem . section iii presents the game theoretic formulation of the problem as well as the nash equilibrium and the cognitive hierarchy solutions respectively . section iv presents the simulation results . finally , conclusions are drawn in section v. consider the downlink of a cran composed of a set @xmath0 of remote radio heads ( rrh ) serving cooperatively a set @xmath1 of users . all rrhs are connected to a central cloud - based control unit ( cu ) via finite capacity fronthaul link . the cu is responsible for the resource allocation and the signal processing of the rrhs , while the rrhs are responsible only for data transmission . due to the finite capacity of the fronthaul , there is a maximum number of rrhs that could be connected to the cu . hence , the cran must co - exist with a hetnet composed of a set @xmath2 of macro bss , a set @xmath3 of pico bss , and a set @xmath4 femto bss . each bs @xmath5 @xmath6 has a set @xmath7 of users to serve . each user @xmath8 is located at a distance @xmath9 from bs @xmath10 or from rrh @xmath11 . ofdma transmissions of @xmath12 subcarriers are also assumed at each bs in @xmath13 and rrh in the cran . the channel gain @xmath14 between each bs / rrh @xmath5 and user @xmath15 over subcarrier @xmath16 is distributed according to an independent rayleigh fading . the received power of a user @xmath15 over subcarrier @xmath16 is @xmath17 where @xmath18 is the path loss exponent and @xmath19 is the transmission power of bs / rrh @xmath5 over subcarrier @xmath16 . additive white gaussian noise of variance @xmath20 is present at each receiver . in our model , each bs / rrh @xmath5 can transmit with a maximum power @xmath21 , and each mobile user has a single receive antenna . in the cran , the cu assigns subcarrier @xmath16 to a single user in @xmath1 and all rrhs transmit simultaneously to that user . the cu uses beamforming to coordinate the transmissions of the rrhs over subcarrier @xmath16 . thus , the received rate of user @xmath22 over subcarrier @xmath16 is given by the multiple input single output capacity with power constraint on each antenna as in @xcite:@xmath23where @xmath24 is the vector of power allocations of all rrhs over subcarrier @xmath16 and @xmath25 is the transmission bandwidth . it has been shown in @xcite that the optimal signaling to achieve this capacity is to use beamforming weight vector @xmath26 where each entry @xmath27 is the weight of rrh @xmath5 and is given by : @xmath28based on this transmission scheme and the rate equation in ( [ ratecloud ] ) , and in order to ensure fairness , the cu assigns subcarrier @xmath16 to user @xmath29 such that : @xmath30assuming that transmissions occur within time frames and @xmath31 is the average rate over the previous time slots . since each bs assigns each subcarrier to one of its users , hereinafter , we drop the user index @xmath15 for brevity . the cu finds the optimal power allocations @xmath32 for all rrhs that maximize the sum of rates over @xmath12 subcarriers : @xmath33 as for the bss in @xmath13 , the received rate at each user @xmath15 in @xmath7 from bs @xmath5 transmission will be : @xmath34 based on ( [ bsrate ] ) , bs @xmath5 allocates subcarrier @xmath16 to user @xmath35 in @xmath7 such that : @xmath36 each bs @xmath5 is required to find its transmission powers @xmath19 so that the sum of rates over all subcarriers is maximized:@xmath37 the bss in the hetnet operate independently of each other and of the cran . hence , the cu and the bss must find their optimal transmission powers in a distributed manner . however , the rrhs and the bss will mutually interfere which , in turn , affects their users received rates . this motivates the adoption of a game - theoretic approach , as explained next . this problem is formulated as a static noncooperative game defined by the triplet @xmath38 where the @xmath39 is the set of players . the strategy vector @xmath40 of each player @xmath5 is to choose its transmission power values over all the subcarriers . hence , the strategy set for bs @xmath5 in @xmath13 is : @xmath41^{l } \hspace{0.1 cm } \textrm{s.t . } \hspace{0.1 cm } \sum_{k=1}^lp_{ik}=p_{i,\textrm{max } } \}$ ] , and the strategy set for the cu in the cran is @xmath42 \hspace{0.1 cm } \textrm{s.t . } \hspace{0.1 cm } \sum_{k=1}^lp_{ik}=p_{i,\textrm{max } } \forall i \in \mathcal{k}_c\}$ ] . the utility for each bs @xmath5 will be : @xmath43where @xmath44 and @xmath45 is the distance between bs @xmath5 and the user assigned to subcarrier @xmath16 . similarly , the cu s utility is : @xmath46 the conventional solution for this formulated game is the so - called nash equilibrium @xcite . in particular , a strategy profile @xmath47 is said to constitute a nash equilibrium ( ne ) if : @xmath48where @xmath49 is the vector of strategies of all players except @xmath5 . to derive the ne for this game , we first determine the best response strategy for each player @xmath50 . for each bs @xmath10 , the best response is the well known water - filling solution @xcite . _ given a fixed power policy @xmath49 , the optimal strategy ( best response ) of bs @xmath5 is given by : @xmath51 where @xmath52 is given by : @xmath53 and @xmath54 satisfies : @xmath55_. this follows from the fact that utility function is concave in @xmath40 since it is a sum of log functions such that each term is a log function of @xmath19 ( for @xmath56 ) . also , the constraint function of @xmath19 in ( [ bsopt ] ) is a linear function of @xmath19 ( for @xmath56 ) . thus , the optimal solution is found using the kkt conditions . _ the best response of the cu is the solution of the nonlinear equations : @xmath57 where @xmath52 and @xmath58 are given by respectively : _ @xmath59 @xmath60 . we first show that @xmath61 is concave in the variables @xmath19 ( @xmath62 ) by showing that the function @xmath63 is concave . this function can be written as:@xmath64 in the above equation , the first sum is a linear sum of @xmath19 for @xmath65 . also , each term of the second sum can be easily seen to be concave in @xmath19 and @xmath66 since in general the function @xmath67 is concave in @xmath68 and @xmath69 . thus , the function @xmath63 is a sum of concave functions in @xmath19 for @xmath65 and is therefore concave . also in general , the function @xmath70 is increasing concave function of @xmath68 . thus , the function @xmath61 is an increasing concave function of @xmath63 . thus , it is concave using the general property that a monotonic concave transformation of a concave function is concave . it follows that the utility function of the cu in ( [ ucloud ] ) is also concave since it is the sum of concave functions in @xmath19 , @xmath65 . thus , the optimal solution can be found using kkt conditions . the existence of the ne is presented next . _ there exists at least one pure nash equilibrium for the heterogeneous cloud power allocation game . _ this follows directly from ( * ? ? ? * proposition 20.3 ) . at the ne , the cu or each bs considers the strategies made by all other devices when choosing its own strategy . this can be detrimental to low - powered bss who will need to respond to a high level of interference caused by the cran and macro bss , and hence , will choose a more conservative power allocation . to overcome this challenge , we will next develop a new game - theoretic model based on _ cognitive hierarchy theory _ @xcite . under cognitive hierarchy ( ch ) , devices are organized into a hierarchy such that the cran and macro bss are placed at the higher hierarchy levels , while low - powered bss are placed at the lower hierarchy levels . each player assumes that the other players belong to equal or lower hierarchy levels when selecting its strategy . this allows low - powered bss to choose less conservative actions than in a conventional ne . we adopt the poisson ch model in @xcite with the following considerations . first , players are distributed across hierarchy levels according to a poisson distribution @xmath71 of rate @xmath72 . moreover , a player at level @xmath16 knows the true distributions of players which are at lower levels . this is known as the overconfidence assumption and it is a realistic assumption for cases in which players are humans for example . using hierarchical approaches for power allocation in hetnets has been earlier investigated using stackelberg games @xcite . stackelberg games , however , rely on only a two level hierarchy composed of leaders and followers , while ch allows for multiple levels of hierarchy , which is more suitable to our setting . further , in stackelberg games , both leaders and followers consider the actions of each others when choosing their actions . this may not be very effective for reducing interference on low - powered bss . in the hetnet tier , the maximum transmit power of a bs is dependent on its type . for instance , a larger bs has a higher maximum transmit power , and thus , it can cause higher interference on smaller bss . to mitigate this high interference , we group the bss based on their type , using the ch level approach . in particular , femtocell bss , who have the lowest power , are assumed to belong to level 1 while pico bss belong to level 2 , macro bss belong to level 3 , and the cran belongs to the last level . in order to reduce the interference among bss of the same type , the overconfidence assumption of the poisson ch model is modified so that bss consider the actions of bss of the same type . however , in ch , each bs takes its action based on its beliefs and does not learn the power values of the bss of the same type . thus , we assume that each player at level @xmath16 believes that all other players of the same type act using the same strategy . consequently , the utility of bs @xmath5 at level @xmath73 is : @xmath74 ) } \bigg),\label{bsch } \vspace{-0.3 cm}\ ] ] where @xmath75&= & g_m(m)\sum_{l \in \mathcal{v } , l \neq i}|h_{lk}|^2p_{ik}d_{lk}^{-\alpha}\nonumber\\ & & + \sum_{l \in \mathcal{v } , l \neq i}\sum_{h=0}^{m-1}g_m(h)|h_{lk}|^2p_{lk}(h)d_{lk}^{-\alpha},\nonumber \vspace{-0.6 cm}\end{aligned}\]]and @xmath76 is the proportion of players at level @xmath77 ( @xmath78 ) . it is given by : @xmath79 . for the cran , we assume that the cu is aware that there are no other players at the same level . this can be achieved by selecting a proper value of the poisson distribution parameter @xmath72 that satisfies @xmath80 in the ch model . one suitable value here is : @xmath81 . the utility of the cu at level 4 is thus given by:@xmath82 } \bigg),\vspace{-0.2 cm}\ ] ] where @xmath83=\sum_{l = 0}^{\mathcal{h}}\sum_{h=0}^{3}g_4(h)|h_{lk}|^2p_{lk}(h)d_{lk}^{-\alpha}.\vspace{-0.1 cm}\ ] ] the equilibrium for the ch case is defined next . a strategy profile @xmath47 is said to constitute a _ cognitive hierarchy equilibrium ( che ) _ iff : @xmath84 where @xmath73 is the ch level of player @xmath15 . for this ch game , the equilibrium strategy for each bs and for the cu are given respectively by the following two propositions . _ the che strategy for each bs @xmath10 at ch level @xmath73 is achieved by solving the following system of equations : @xmath85 _ _ where _ @xmath86 , @xmath87 @xmath88 it can be easily shown that the function @xmath89 is concave in @xmath68 @xmath90 . thus , the term @xmath91 ) } \bigg)$ ] is concave in @xmath19 . thus , the utility in ( [ bsch ] ) is a sum of concave functions . _ the che strategy of the cu is the solution of the nonlinear equations : @xmath92 where @xmath52 and @xmath58 are given by respectively : _ @xmath93 @xmath94 the utility of the cu in the ch case is still the same function of @xmath19 as in ( [ ucloud ] ) . hence , the function is concave and the above equations are obtained by applying the kkt conditions . based on ( [ bsche ] ) and ( [ cuche ] ) , each player at level @xmath73 finds the level @xmath77 che power @xmath95 for each other player @xmath96 and for all @xmath97 in order to find its che strategy . to assess the performance achieved by both ne and ch , the following values are considered : @xmath98 , @xmath99 , @xmath100 , @xmath101 @xmath102 , @xmath103 , @xmath104 @xmath105 , @xmath106 , @xmath107 @xmath108 , @xmath109 , @xmath110 dbm , and @xmath111 mhz . also , the bss of the same type are assumed to have the same maximum transmit power such that : @xmath112 dbm , @xmath113 dbm , @xmath114 dbm , @xmath115 dbm . the rrhs , the bss , and the cran users are deployed uniformly within a @xmath116 km rectangular grid . the hetnet users are deployed uniformly within a circle centered at each bs of radius equal to the coverage of each bs . the coverage radius of the macro , pico , and femto bss are set to : @xmath117 m , @xmath118 m , and @xmath119 m respectively . all channel coefficients are generated according to a rayleigh distribution with variance @xmath120 dbm . statistical results are averaged over a large number of independent runs . for comparisons , we consider the ne , the proposed che , and an equal power allocation scheme . [ rrhrate ] [ rrhpowrate ] [ rrhtotal ] [ rrhpowtotal ] fig . [ rrhrate ] shows that average rate of the cran increases with the number of rrh while the average rate of the bss in @xmath13 decreases due to interference . in this figure , we can see a drastic increase in the sum - rate of femtocells and picocells under ch as opposed to the ne . the average sum - rate of the pico and femto bs increases from @xmath121 bits / sec to @xmath122 and @xmath123 bits / sec , respectively . as for the macro cells and the cran , the average sum - rate decreases slightly with ch . the average decrease is only @xmath124 for macro bss and @xmath125 for the cran . [ rrhpowrate ] shows the sum - rate of the cran varies with the maximum rrh transmit power . clearly , the average cran sum - rate increases while that of the bss decreases due to the increased interference . in fig . 2 , we can also see , analogously to fig . 1 , that ch yields significant overall rate improvements , particularly for pico and femto bss . figs . 3 and 4 show the average total system rate as the number of rrhs and their maximum power vary , respectively . fig . [ rrhtotal ] shows a considerable improvement in the total system rate when ch is used compared to both cases of ne and equal power policy . the improvement in the total system rate is on the average double the total sum - rate of the ne and the equal power policy . [ rrhpowtotal ] further corroborates such rate improvements at all rrh transmit power values . clearly , the proposed ch power allocation approach improves considerably the performance of the low - powered ( pico and femto ) bss without degrading the performance of the cran and the macro bss . it also results in a considerable improvement in the overall system performance . in this paper , we have studied the problem of power control in a hybrid cellular network in which a cran and a small cell heterogeneous network co - exist . we have formulated the problem as a noncooperative game and we have studied the solution under a conventional nash equilibrium and under a case in which the base stations are grouped in hierarchies according to the framework of cognitive hierarchy theory . simulation results have shown a considerable system performance in terms of the total sum - rate compared to conventional solutions . the results have also shown that the proposed ch based approach reduces considerably the interference on the low - powered pico and femto cells without jeopardizing the performance of macro base stations and the cran . a. checko , h. christiansen , y. yan , l. scolari , g. kardaras , m. berger , and l. dittmann , `` cloud ran for mobile networks a technology overview , '' _ ieee communication surveys and tutorials _ 17 , no . 1 , pp . 405 - 426 , march 2015 . m. peng , y. li , j. jiang , j. li , and c. wang `` heterogeneous cloud radio access networks : a new perspective for enhancing spectral and energy efficiencies , '' _ ieee wireless commun . _ , vol . 21 , no . 6 , pp . 126135 , dec . t. ran , s. sun , b. rong , and m. kadoch,``game theory based multi - tier spectrum sharing for lte - a heterogeneous networks , '' _ in proc . of ieee international conference on communications ( icc ) _ , london , uk , june 2015 . s. cheng , s. lien , f. chu , and k. chen,``on exploiting cognitive radio to mitigate interference in macro / femto heterogeneous networks , '' _ ieee wireless communications _ , vol . 18 , no . 3 , pp . 40 - 47 , june 2011 . c. shen , jie xu , and m. van der schaar `` , silence is gold : strategic interference mitigation using tokens in heterogeneous small cell networks , '' _ ieee journal on selected areas in communications _ , vol . 33 , no . 6 , pp . 1097 - 1111 , june 2015 . h. zhang , c. jiang , and j. cheng , `` cooperative interference mitigation and handover management for heterogeneous cloud small cell networks , '' _ ieee wireless communications _ , vol . 22 , no . 3 , pp . 92 - 99 , june 2015 . l. liu , s. bi , and r. zhang , `` joint power control and fronthaul rate allocation for throughput maximization in ofdma - based cloud radio access network , '' _ ieee trans . on communications _ , 63 , no . 11 , pp . 4097 - 4110 , november 2015 . m. peng , k. zhang , j. jiang , j. wang , and w. wang , `` energy - efficient resource assignment and power allocation in heterogeneous cloud radio access networks , '' _ ieee trans . on vehicular technology _ , vol . 64 , no . 11 , pp . 5275 - 5287 , november 2015 . l. yanhui , w. chunming , y. changchuan , and y. guangxin , `` downlink scheduling and radio resource allocation in adaptive ofdma wireless communication systems for user - individual qos , '' _ in proceedings of the world academy of science , engineering and technology _ , vol . 12 , pp . 221 - 225 , march 2006 . s. guruacharya , d. niyato , d. in kim , and e. hossain , `` hierarchical competition for downlink power allocation in ofdma femtocell networks , '' _ ieee trans . on wireless commun . _ , vol . 12 , no . 4 , pp . 1543 - 1553 , april 2014 . z. han , d. niyato , w. saad , t. baar , and are hjrungnes , _ game theory in wireless and communication networks : theory , models , and applications _ , cambridge university press , 2012 . g. bacci , s. lasaulce , w. saad , and l. sanguinetti,game theory for networks : a tutorial on game - theoretic tools for emerging signal processing applications , ieee signal processing magazine , vol . 33 , no . 1 , 94 - 119 , january 2016 .
in this paper , the problem of distributed power allocation is considered for the downlink of a cloud radio access network ( cran ) that is coexisting with a heterogeneous network . in this multi - tier system , the heterogeneous network base stations ( bss ) as well as the cran remote radio heads seek to choose their optimal power to maximize their users rates . the problem is formulated as a noncooperative game in which the players are the cran s cloud and the bss . given the difference of capabilities between the cran and the various bss , the game is cast within the framework of cognitive hierarchy theory . in this framework , players are organized in a hierarchy in such a way that a player can choose its strategy while considering players of only similar or lower hierarchies . using such a hierarchical design , one can reduce the interference caused by the cran and high - powered base stations on low - powered bss . for this game , the properties of the nash equilibrium and the cognitive hierarchy equilibrium are analyzed . simulation results show that the proposed cognitive hierarchy model yields significant performance gains , in terms of the total rate , reaching up to twice the rate achieved by a classical noncooperative game s nash equilibrium .
the mechanism of quark confinement stands as a key concept in our understanding of the fundamental interactions in high energy physics @xcite . as a quark and an anti - quark are pulled apart , the energy stored in the gluon string connecting them grows linearly with distance , due to the confining nature of strong nuclear forces described by quantum - chromodynamics ( qcd ) . in gauge theories hosting dynamical charges , there exists a critical length scale at which the confining string breaks , creating particle - antiparticle pairs which reduce the energy density in the string and give rise to the hadrons at the string edges @xcite . the _ static properties _ of string breaking have been widely explored using a variety of lattice methods , wherein the effective string potential separating static charges can be extracted by the polyakov or wilson loops @xcite . however , the _ real - time dynamics _ of gauge theories are usually biased by a severe sign problem , and as such can not be accessed using lattice montecarlo simulations @xcite . in this paper , we apply tensor network ( tn ) methods in order to study the real - time dynamics of a lattice gauge theory ( lgt ) with dynamical charges and quantum gauge degrees of freedom in one dimensional systems . in particular , we investigate the real - time string - breaking dynamics in abelian u(1 ) lgts in ( 1 + 1)d , which share with qcd the basic feature of confinement . in recent years , efficient numerical methods based on tns have found widespread application to the real - time dynamics of strongly correlated low - dimensional systems @xcite . they are nowadays routinely used to tackle a variety of condensed matter and atomic physics problems , such as the evaluation of spectral functions of low - dimensional magnets and the quench or controlled dynamics of ultra - cold quantum gases in optical lattices @xcite . while tn methods have been extensively applied to spin and hubbard - type models , only recently it has been shown how tns can provide an ideal platform for the investigation of gauge theories , for example in the study of the static properties of the schwinger model , the low - energy mass excitation spectrum , and the dynamics of deconfinement in 2d @xmath0 lgts @xcite . in particular , the quantum link model ( qlm ) formulation of lgt @xcite gauge theories whose link hilbert space is finite - dimensional has been used to develop efficient general - purpose tn algorithms to describe static and real - time dynamical properties of abelian and non - abelian lgts including generic forms of matter fields , and to present different possible quantum simulator implementations on different platforms : trapped ions , cold atoms in optical lattices and circuit quantum electrodynamics ( qed ) @xcite ( see also @xcite for recent reviews and references therein ) . tn methods are based on variational tensor structure ansatze for the many - body wave function of the quantum system of interest : the tensor structure is chosen to best accommodate some general system properties , e.g. , dimensionality , boundary conditions and symmetries , while a controlled approximation is introduced in such a way that one can interpolate between a mean field and an exact representation of the system . being a wave function based method , one has direct access to all relevant information of the system itself , including quantum correlations , i.e. , entanglement . in one - dimensional systems , an efficient tensor structure is given by the matrix product state ( mps ) ansatz @xcite , defined as , @xmath1 where the tensor @xmath2 contains the variational parameters needed to describe the system wave - function , @xmath3 characterize the local hilbert space , and @xmath4 account for quantum correlations or entanglement ( schmidt rank ) between different bipartitions of the lattice . indeed , setting @xmath5 results in a mean field description , while any @xmath6 allows for the description of correlated many - body states . given the tensor structure , the tensor dimensions and coefficients are then optimized to efficiently and accurately describe the system properties by means of algorithms that scale polynomially in the system size and @xmath7 . usually , these algorithms exploit the system hamiltonian tensor structure , naturally arising from the few - body and local nature of the interactions , to efficiently describe the system ground state or low - lying eigenstates , or to follow the real- or imaginary - time evolution of the system itself . indeed , in the tn approach , real- and imaginary - time evolution have no fundamental differences at the computational level as there is no sign problem , and limitations arise only in some scenarios depending on the specific dynamics of interest as witnessed by the fast increasing literature appearing based on this approach @xcite from the amount of quantum correlations present in the system wave function . here , we show how tn algorithms allow for the study of the real - time dynamics of lgts , focusing on the string breaking in a paradigmatic confining theory - the schwinger model @xcite in a quantum link formulation . we characterize the real - time dynamics of the primary and secondary string breaking and show that string breaking is intimately related to entanglement production in the system . a qualitative picture for string breaking in our models , together with a typical result for our time - dependent simulations on a system of @xmath8 lattice sites , is illustrated in fig . [ summary : fig ] . even more importantly , our simulations allow us to track the entanglement evolution during string breaking : as we will show below , string breaking and the so - called schwinger mechanism are intimately connected to entanglement propagation , which we address by evaluating the so called von neumann entanglement entropy . finally , we show that tn methods can be used to study scattering processes between bound states of lgts : we develop a scheme to engineer _ meson _ collisions @xcite , and we show how , very surprisingly , the scattering not only reflects into an enhanced rate of particle - antiparticle creation , but it does affect drastically the entanglement properties of the system , which stays significantly correlated well beyond the scattering time - window . the paper is structured as follows : in sec . [ sec_model ] we present the system hamiltonian and recall the tn algorithm we are using throughout this work . in sec . [ sec : strbreaking ] we present the results on string breaking and mass production dynamics , including a discussion on how this phenomenon can be observed using a quantum simulation platform . in sec . [ sec : ent ] we show how entanglement follows the string breaking dynamics , providing a quantitative picture which underlines how entanglement entropies are directly tied to string - breaking . finally , we present our result on scattering in sec . [ sec : scatt ] , and draw a summary of our results in sec . [ sec : concl ] . qed in ( 1 + 1)d , also known as the schwinger model @xcite , represents an ideal testing - ground for the benchmarking and development of new computational methods . despite its relative simplicity , this model captures fundamental aspects of gauge theories such as , e.g. , the presence of a chiral symmetry undergoing spontaneous symmetry breaking @xcite . even more importantly , this theory , like qcd , displays confinement : differently from ( 3 + 1)d qed , in ( 1 + 1)d electrons and positrons are confined , and interact via a long - range potential which increases linearly with distance . due to the large energy cost associated with the electric flux between charges at large inter - charge distances , the electric flux string is unstable to particle - antiparticle creation , as in qcd , and string breaking takes place @xcite . while this phenomenon , directly connected to the schwinger mechanism of mass production out of a vacuum , has long been debated , and notable insights have been provided using a variety of approximate methods , a full quantum mechanical understanding of the complex real - time dynamics taking place during string breaking is lacking due to the computationally complexity of the many - body problem @xcite . in the hamiltonian formulation , its dynamics are defined by the following form : @xmath9 \nonumber\\ & & + m\sum_x ( -1)^x\psi^{\dagger}_x\psi_x + \frac{g^2}{2}\sum_xe^2_{x , x+1}. \label{eq : hamiltonian}\end{aligned}\ ] ] where @xmath10 are fermionic creation / annihilation operators describing kogut - susskind ( staggered ) fermions ( see fig . [ summary : fig ] ) , @xmath11 are the gauge fields residing on the @xmath12 link , and we denote the strength of fermion - hopping ( the kinetic energy of electrons and positrons ) with @xmath13 , the staggered mass of the fermions with @xmath7 , and the electric coupling strength with @xmath14 , where @xmath15 is the electric - field operator . the gauge generator is given by @xmath16 so that @xmath17=0 $ ] and all physical states @xmath18 satisfy the gauss law @xmath19 . while in the wilson formulation @xmath11 are parallel transporters acting on an infinite dimensional hilbert space , we focus here on a formulation based on qlms , where the gauge fields are represented by spin-1 operators , @xmath20 and , as such , act on a finite - dimensional link hilbert space @xcite . in particular , the electric field operator allows three possible states for the electric flux , constraining the physical states per site as described in fig . [ summary : fig ] . a detailed discussion of the quantum link formulation can be found in ref . @xcite , while in ref . @xcite it was shown how such quantum link formulation reproduces the phase diagram and quantum criticality of the continuum theory . string breaking is the process of cutting and shortening the electric flux string that connects a particle - antiparticle pair by creating a new charge - anticharge pair @xcite . within our framework , a string consists of two charges creating non - zero electric flux between them . the charges are represented by appropriate boundary conditions ( static charges ) or by excitations of the mass field at the site of the fermion ( dynamical charges ) . this is realized by an effective jump of a fermion from the site of one charge to the site of the second charge satisfying the gauss law . the string of electric flux then follows from gauss law . the charges force the links into a non - zero flux state , according to the configuration of the charges in either one direction or the other . before embarking in a full quantum mechanical investigation of string breaking , we now discuss the classical ( @xmath21 ) static picture which provides a simple yet informative illustration of the different stages of the string breaking mechanism . a set of cartoons of the classical states is provided in fig . [ summary : fig ] : _ vacuum.- _ in the vacuum ( a ) , neither mass nor electric field excitations are present . the vacuum energy is thus @xmath22 . _ string.- _ in the string state ( b ) , two mass excitations are present at the boundaries , and all electric fields connecting the two are also in the @xmath23 state . the resulting string energy then takes the form @xmath24 _ pairs.- _ in the pairs state ( c ) all the masses are excited forming charge - anticharge pairs with an energy @xmath25 . _ mesons.- _ in a confined phase , particle - antiparticle pair production can favor the establishment of a vacuum state between two static charges , with a pair of mesons at the boundary of the string ( see ( d ) ) . the resulting energy is : @xmath26 at the static level , string breaking takes place at a critical distance @xmath27 , above which the meson state is energetically favored over the string state ( @xmath28 ) : @xmath29 _ antipairs and antistring.- _ the antipair - state ( e ) and the antistring ( f ) denote the pair - state and the string with the electric flux having the opposite sign . at a dynamical level , string breaking takes place as a consequence of the schwinger mechanism : in ( 1 + 1)d , the vacuum between two charges of opposite sign is unstable against particle - antiparticle creation @xcite , eventually leading to the electric field in the system flipping sign and to mass production . in real - time , this process develops following intermediate consecutive steps , and is schematically illustrated in fig . [ summary : fig ] : at very short timescales , particle - antiparticle creation takes place in the middle of the string , creating the _ pair _ state depicted in ( c ) . subsequently , the electric field in the center of the string relaxes to 0 , and the external charges get screened , effectively forming mesons ( d ) . at this point the process reverses , first establishing a state with _ anti - pairs _ ( e ) , which finally decays into a string of oppositely signed electric field with respect to the initial state , an _ anti - string _ ( f ) . string - breaking is a direct consequence of confinement : while in a deconfined phase such as qed in ( 3 + 1)d , it is possible to separate opposite charges at large distances due to coulomb s law , in a confined phase the corresponding electric field string breaks due to the effective potential between charges increasing as a function of distance . however , the exact real - time dynamics of string breaking are inaccessible to classical simulations based on montecarlo sampling due to a severe sign - problem . while classical - statistical approaches can provide remarkable insights in some parameter regimes @xcite such as small masses , unbiased numerical simulations for arbitrary parameter regimes have been lacking . in the following , we present a systematic study of the string breaking dynamics using mps techniques . tensor network algorithms are one of the paradigms for simulating quantum many - body systems in low - dimensions , both in and out of equilibrium , via a representation of the quantum state with a variational ansatz for wave - functions and/or density matrices @xcite . for one - dimensional pure states , on which we focus here , the starting point is to consider a class of states of the form given in eq . and depicted in fig . [ mpsplot ] with some fixed auxiliary dimension @xmath7 and physical dimension @xmath30 : for example , for spin one - half systems one has @xmath31 , while the dimension @xmath7 depends on the states studied and on the desired accuracy . for ground states of one - dimensional gapped hamiltonians , @xmath7 is in general independent of the system size @xmath32 , while for critical systems , due to area - law logarithmic violations , the auxiliary bond dimension scales as @xmath33 , where @xmath34 is the central charge of the system ( see ref . for a review ) . the case of out - of - equilibrium dynamics as considered here is more challenging , and no general picture is known , thus the bond dimension @xmath7 has to be adapted to each specific case and convergence has to be checked comparing the results at increasing bond dimension @xcite . global symmetries of the system of interest , such as particle number or global magnetization and even more complex non - abelian global symmetries , can be embedded in the wave function ansatz given in eq . in an elegant way by promoting each tensor @xmath35 to a symmetry - sector - preserving tensor to a tensor that preserves every symmetry sector ; that is , every index of the tensor is dressed with the corresponding symmetry charge number @xcite . this symmetric formulation of tensor networks allows one to describe wave functions exactly and more efficiently with the desired quantum numbers , addressing each symmetry sector separately . recently , it has also been shown that local symmetries namely gauge symmetries can be embedded in such description by generalizing the wave function ansatz to a gauge invariant one @xcite . we briefly recall such a construction in the rest of this section , and tailor it to the model we will study in the rest of the paper . + + the hilbert space of a gauge invariant system is given by the direct sum of every sector with different values of the `` gauss law , @xmath36 and @xmath37 . due to the gauge symmetry , there is no physical ( e.g. , gauge invariant ) operator , including the hamiltonian , that connects any two different sectors . hence , starting with a quantum state partially defined by the values of the gauss law , or a set of local ( gauge ) constants of motion , the time - evolved wave - function will remain in this sector under the action of the hamiltonian . in the qlm formulation of lgts , a gauge invariant tensor description is immediately obtained if we use a ' ' rishon @xcite or schwinger representation of the gauge degrees of freedom ( independent of the nature of the local continuous symmetry - abelian or non - abelian - and the dimensionality of the lattice @xcite ) . in our case , the @xmath38 qlm in ( 1 + 1)-d with a spin-1 per link , the spin operator or electric field @xmath39 is represented by a pair of schwinger bosons @xmath40 with a total occupation of two bosons per link , i.e. @xmath41 . the first step to build an efficient tensor network representation of such states is to identify a local hilbert space spanned by the states @xmath42 defined on the tensor product of the fermionic matter field on a lattice site and of the rishon states on its left and right , that is @xmath43 . this allows for the projection of the state into the gauge invariant subspace , restricting the local bases only to the " physical states : in our model , this process results in the five gauge invariant states ( uniquely identified in terms of the schwinger rishons for even and odd lattice sites respectively and depicted in fig.[summary : fig ] left ) given by @xcite : @xmath44 these are the only states allowed locally by gauge invariance ( notice that the local hilbert space is different on even and odd sites as a consequence of the staggered nature of the fermions ) . a gauge invariant many - body state clearly exists in the tensor product of such local basis states . however , not all tensor product combinations are compatible with the spin representation @xmath45 : with our choices , only the states with @xmath41 are allowed . this additional constraint can be satisfied by modifying the tensor structure of the many - body state ansatz , that is by introducing a projector which acts on nearest - neighbor lattice sites and enforces the correct number of rishons on the link @xcite . hence , the gauge structure of systems with a local continuous symmetry can be encoded in a matrix product operator ( mpo ) as depicted in fig . [ mpsplot ] , such that : @xmath46 for the model we study , the mpo has bond dimension three ( @xmath47 ) and the non - zero elements of the tensors @xmath48 for the odd sites can be expressed as : @xmath49 the tensors for the even sites can be computed in a similar way . given the gauge invariant tensor structure introduced above , one can reformulate the standard algorithms for tensor networks to compute the ground state of the model , or , as we will do in the next sections , compute the real time evolution of some initial state , e.g. , by means of a suzuki - trotter decomposition of the time evolution operator acting on a pair of neighboring sites @xcite . the parameters used in our calculation , together with a discussion of the errors involved in the approximations , are presented in appendix [ app ] . finally , we mention that one can further simplify the tensor structure in eq . to increase the algorithmic efficiency and that this construction is completely transparent with respect to the abelian or non - abelian nature of the gauge symmetry , resulting in a drastic simplification of numerical analysis of non - abelian lattice gauge theories @xcite . in this section we present our results of the lattice gauge tn numerical simulations for the real - time dynamics of string breaking : we focus on the time evolution of the electric and matter fields , quantitatively studying the properties of string breaking and of the schwinger mechanism . in the following section we analyze the time evolution of quantum correlations during this process an analysis enabled by the tn approach and show that the two figures of merit are intimately related . the setup for our simulations is a dynamical string surrounded by the vacuum . the total lattice length @xmath50 is chosen such as boundary effects are negligible for the timescales we investigate . some typical results of the electric field time evolution are shown in the left column of fig . [ overview : fig ] ( column a ) for different values of the fermion mass @xmath7 and electric field coupling strength @xmath14 ( hereafter we set @xmath51 and times are given in units of @xmath52 ) . in the top row , the string freely breaks as for @xmath53 no energy - cost is needed for such process to occur and the system evolves according to the free hopping hamiltonian : the electric field starts to oscillate between string and anti - string displaying primary , secondary ( marked by vertical lines ) and subsequent string breakings . moreover , the string propagates into the vacuum creating two diverging electric field excitation wavefronts . in the middle row , representing the same scenario for non - zero mass and electric field coupling strengths , string breaking still occurs ; however , the string evolution is damped and stabilizes the mean electric flux around zero . finally , for large mass @xmath7 and @xmath14 , the large mass suppresses vacuum fluctuations while the large electric coupling prevents the occurrence of string breaking and thus the string does not decay . however , some dynamics still occur as we will see in more details in subsection [ sec : schwing ] , as fermion - antifermion pairs are created using the energy of the electric field and then annihilated , thus restoring the electric field excitations . ( 200,230 ) ( -15,0 ) to perform a quantitative analysis of string breaking , we repeat the same simulation for different values of @xmath7 and @xmath14 and analyze the mean electric field of the central six lattice sites of the string as a function of time . the results are reported in fig . [ stringbreakingevolution : fig ] where one can clearly see that different scenarios might occur : either the string breaks when the mean electric field drops below zero , or the electric flux remains positive throughout the whole evolution and no string breaking occurs . in the limit of @xmath54 the oscillations show the highest amplitude , which is reduced by changing at least one of the two parameters ; as previously noticed , for high values of either system parameter the string never breaks . in the regime between these two extreme cases , we observe a third type of behavior : the electric flux tends to zero , however no anti - string is formed : the oscillation is strongly damped and the system remains in the broken string state ( compare with the middle row in fig . [ overview : fig ] ) . these findings are summarized in fig . [ stringbreakingmap : fig ] : for @xmath55 we observe the full string breaking dynamics with at least the partial formation of a string with opposite electric field flux after reaching the broken string state ( red area ) . for @xmath56 string breaking was not observed and the dynamics are dominated by the interplay of the state of maximum pair creation and the original string ( green area ) . finally , the white area in between represents the region of parameters where we observe the string breaking with over - damped oscillations . during the string breaking process , a wavefront of electric flux spreads outwards , as can be clearly seen in fig . [ overview : fig ] ( panel a1 and a2 ) . in this section , we quantitatively characterize the wavefront spreading by analyzing its spreading velocity and the oscillation intensity . in fig . [ spreadingef : fig ] , we show the wavefront propagation as a function of time for different electric coupling @xmath14 for the zero mass case . the lower inset illustrates how we calculated such propagation : we follow the electric field excitation on one side of string by means of tracking the difference @xmath57 between the gauge field at some position @xmath58 and the next nearest neighbor site @xmath59 . further , we define the time when this difference displays a maximum s arrival of the wave front . as can be seen from the lower inset of fig . [ spreadingef : fig ] , where different colors represent different coordinates @xmath58 , a wave - like propagation can be clearly identified . following this scheme , we plot the position of the wavefront as a function of time in the main panel of fig . [ spreadingef : fig ] . the result shows an approximatively linear spreading after an initial transient time of about @xmath60 , with a velocity almost independent of the values of the electric coupling for @xmath61 . increasing @xmath14 starting from @xmath62 , the velocity increases as well , until for @xmath63 , where the results are inconclusive as increasing @xmath14 leads to smaller wavefront amplitudes and consequently the errors bars prevent an accurate analysis to be carried out . however , for small @xmath14 , the spreading velocity can be extracted directly from fig . [ spreadingef : fig ] . by fitting the values for @xmath64 and @xmath65 we obtain a value of @xmath66 . finally , in the upper inset of fig . [ spreadingef : fig ] , we repeated this analysis for different masses and @xmath67 . the results clearly show that , for sufficiently large @xmath68 , the wave front spread velocity has an inverse linear dependence on the mass . all these results are in agreement with a theoretical estimate obtained assuming the ends of the string as sources of excitations : in a quasi - free or weak coupled model , the speed is related with the band - width of the kinetic term resulting on an excitation spreading velocity proportional to @xmath69 . during string breaking , the schwinger model dynamics exhibit particle - antiparticle pair production as a consequence of the energy released from the external electric field string . this phenomenon is usually referred to as the schwinger mechanism , and has been studied extensively since its first presentation in 1951 @xcite . in the following , we provide a systematic investigation of the schwinger mechanism in the context of the u(1 ) qlm . in our analysis , in the initial state defining the string , the only two mass excitations present are the two dynamical charges which create the string itself . however , during the dynamics , the energy of the string is transformed into mass excitations . when the maximum mass production is reached , the particles start to annihilate , either to break the string or to restore it . the time needed to reach the maximum mass production @xmath70 depends on the mass and electric coupling as shown in fig . [ schwingertimes : fig ] . it clearly displays two different behaviors , depending whether the electric coupling @xmath14 is greater or less than one . for small-@xmath14 , @xmath70 is maximal for @xmath65 and decreases monotonically with @xmath7 . this occurs , recalling the results of fig . [ stringbreakingmap : fig ] , in the regime where string breaking is observable . indeed , these are the cases comparable to the top row in fig . [ overview : fig ] . on the contrary , for larger values of @xmath14 , we observe the maximum to occur for @xmath71 . in particular , the maximum is obtained at a point where the energy of the electric field matches approximately the energy needed to fully convert the string into particle pairs , i.e. , @xmath72 . this corresponds to the dynamics as displayed in the bottom row in fig . [ overview : fig ] . in the regime of high masses we can use second - order perturbation theory to estimate the mass - dependence of the mass - production time - scale analytically , as the model can be approximated as decoupled double - wells potentials , resulting in @xmath73 . we checked this approximation comparing the analytical estimate with the numerical results in fig . [ schwingertimeshigh : fig ] : the best fit results in @xmath74 , in good agreement with the theoretical prediction . recently , the implementation of abelian quantum link models has been envisaged on different platforms , such as ultra cold atom gases in optical lattices @xcite , trapped ions @xcite , and circuit qed architectures @xcite . in this context , the possibility of investigating the real - time dynamics using tn methods provides an invaluable tool to benchmark experiment against theory in ( 1 + 1)d , and to address the role of possible imperfections in quantum simulators . unavoidable imperfections which will be present in any implementation can be detrimental to the observation of both ground state physics properties and real - time dynamics . the former case has recently been investigated in the context of adiabatic state preparation , where it was shown how gauge - variant perturbations weakly affect the fidelity of the loading process @xcite . here , we focus instead on the effect of gauge - invariant imperfections on the string - breaking dynamics . different from gauge - variant terms , the role of gauge - invariant imperfections can not be systematically addressed in an experiment using , e.g. , post - selection over the experimental data . following the implementation schemes in ref . @xcite , one of the most common form of gauge - invariant imperfections are nearest - neighbor interactions between matter and gauge fields : @xmath75 with @xmath76 . this form of the imperfection is usually generated as a resonant term in perturbation theory to next - to - leading order with respect to @xmath13 . while this implies @xmath77 , for realistic implementations the difference in magnitudes between these two terms can not be made arbitrary large : this will require small absolute energy scales , thus making other sources of more detrimental imperfections such as temperature ( for the cold atom implementation ) and disorder ( in the circuit qed implementation ) dominant . at a qualitative level , this interaction term can freeze the system into a configuration where the matter fields remain pinned . this is due to the effective attraction generated by the nearby electric field configuration . to estimate the effects of these imperfections on a quantum simulation of the dynamics considered in this work , we repeat the numerical simulations including realistic imperfections expected in a first generation of experiments . in the top row of fig . [ imperf : fig ] we show the string breaking evolution of the electric field , in the presence of @xmath78 , with the same system parameters as used for fig . [ overview : fig ] , with imperfections of the order of 10% @xmath79 . the results including the imperfections still clearly exhibit the physics observed in the imperfection - free results . in general , even the quantitative dynamics is very well captured up to long - timescales , as illustrated in the bottom row of fig . [ imperf : fig ] . the only exception are intermediate @xmath14 and @xmath7 values , where significant discrepancies ( up to 50% ) are observed for intermediate timescales ( middle panel of the last row ) . in all other regimes , we could observe deviations up to a maximum of 15% caused by typical imperfections @xmath79 . one of the key aspects of mps - based methods is that they give full access to the wave - function during the real - time dynamics . by considering the dynamics of quantum correlations embodied by entanglement , this enables us to tackle the string - breaking problem from a fully complementary viewpoint with respect to the electric field and mass generation studies undertaken in the previous sections . the main question we want to address in this section is whether and to what extent entanglement plays a role in the string - breaking dynamics . in the last decade it has been shown that entanglement play a fundamental role in many - body quantum processes , from quantum critical phenomena , to quantum information theory and other fundamental aspects of quantum physics . moreover , several aspects of quantum field theories , such as the static properties of conformal field theories , have also been extensively studied using entanglement measures @xcite . additionally , entanglement was shown to play a crucial role in the limits of classical simulations of quantum systems calling for the need of the development of quantum simulators to overcome such limitations @xcite . a common way to quantify entanglement for pure quantum states is by using the so - called renyi entanglement entropies , and , in particular , the von neumann entropy . given a pure state with the density matrix @xmath80 , the entanglement entropy is given by the von neumann entropy of the reduced density matrix @xmath81 @xcite : @xmath82 @xmath83 is a measure of the entanglement of a bipartition at the lattice site @xmath58 . the entanglement entropy takes values between @xmath84 for a separable state ( product state ) and @xmath85 , with @xmath30 being the size of the hilbert space , for a maximally entangled state . in fig . [ overview : fig ] ( column c ) we plot the time evolution of the entanglement entropy for the three different cases considered before ( panels 1 - 3 ) : as it can be clearly seen , the entanglement evolution resembles the mass and electric field excitations , indicating how the two phenomena are related . firstly , the vacuum fluctuations for small @xmath14 and @xmath7 generate not only mass and electric field fluctuations , but also a high amount of correlations . moreover , the electric field wavefront is mimicked by the entanglement behavior , once again showing that not only do excitations propagate from the string , but also correlations do as well . secondly , the string breaking process is clearly a two - steps process : first a correlated pair is created in between odd - even sites , and later in between the even - odd sites ( blue - yellow checkerboard pattern inside the string in panels c1 and 2 ) . finally , for large @xmath14 and @xmath7 , when the string does not break , the correlations behavior drastically changes as well . there , within the string , the correlations are built periodically only between odd - even sites , while in the vacuum the correlations are drastically suppressed ( panel c3 ) . a transparent picture on how entanglement is generated during the quench dynamics can be gathered by monitoring the entanglement growth at different points in space . in fig . [ entropyprofile : fig ] we show the entanglement time evolution for a set of system bipartitions and parameters , e.g. , cutting the system in the middle of the string or in the vacuum . the solid lines are results obtained for @xmath65 and @xmath67 , while the dashed lines correspond to the case @xmath86 and @xmath87 . the orange line represents the entanglement growth for a partition at lattice site @xmath88 , therefore in the center of the lattice region starting in the vacuum . in the zero - mass case , one can see that the entanglement entropy grows almost linearly for most part of the evolution - signaling the vacuum instability against resonant particle - antiparticle pair production . towards the end of the evolution , the linear growth breaks down as boundary effects start to play a role . the remaining solid lines in fig . [ entropyprofile : fig ] represent the behavior of a partition between an even - odd lattice site ( violet ) and between an odd - even lattice site ( blue ) and display a more complex behavior . these results were obtained in the center of the string while it breaks up : the counter - phase oscillations indicate the competition of two states , together with an overall growth of the entanglement entropy , though not as fast as in the vacuum . as we have seen , in this regime the string breaking is a result of consecutive hopping processes : fermions hop on the lattice creating mass excitations followed by annihilations , which result in the string breaking with two remaining dynamical mesons . this dynamical process goes on and after two hopping processes the anti - string is created . [ entropyprofile : fig ] shows that these dynamics are well captured by the oscillations of the entanglement entropy . each maximum represents one hopping process : at the first maximum of the blue line the fermions are about to hop for the first time , while at the first maximum of the violet line the second hopping process is at its peak resulting in the string broken state . at around @xmath89 , the blue line does not display a maximum : this depleted region signals the anti - string state , where the last hopping event creating the anti - string is again the first hopping to break the negative string ( and thus , it takes twice as long for the entropy to reach the next maximum ) . differently from before , in the massive scenario ( dashed lines ) , we see that the entanglement entropy for the vacuum stays close to zero as the large mass and electric coupling strongly suppress the particle - pair creation that triggered the strong growth of the entropy in the previous case . moreover , in the middle of the string the entanglement entropy is drastically affected : the blue dashed line initially behaves as the solid line in the massless case , reflecting the same mass excitation by pair creation . however , the violet dashed line always remains close to zero as further evolution into the string broken state is energetically forbidden : the state evolves back into the string and the correlations between the even - odd sites can not be created . the system is then oscillating between two almost degenerate states : the initial string state and the state made out of pairs . this results in the oscillating behavior of the entanglement entropy between zero and one . finally , the third case with @xmath90 and @xmath91 ( dot - dashed lines ) lies between the two previous limiting cases : here the string breaks , but does not evolve into an anti - string . in the vacuum , the entanglement evolution is very similar to the first case ( @xmath92 ) as the entropy grows almost linearly after a transient regime . however , the slope is reduced by the nonzero mass . the entanglement in the center of the string initially evolves as for the massless case , but after the first two hopping processes for @xmath93 , the oscillation turns into vacuum - like growth . this is a strong indication for non - periodic string breaking : the dynamics , although being unitary , resemble a dissipative process where the electric field energy irreversibly disperses into the vacuum . this irreversible behavior directly resembles what we observe in the electric field dynamics , which does not display any clear periodic signature . even more remarkably , the real - space particle creation and the entanglement dynamics are quantitatively tied . we concentrate on the signatures of the wavefront of the string imprinted on the evolution of the entanglement entropy . we consider the case @xmath54 as it is characterized by the most pronounced wavefront , where the string with its slow entanglement growth is embedded in the fast growing vacuum ( see fig . [ overview : fig ] , panel c1 ) . to characterize the entanglement spreading due to the wavefront , we exploit the fact that the entanglement entropy in the vacuum is constant in space even though it evolves in time . therefore , far enough from both sides of the string there is a plateau of constant entropy much higher than the entropy in the middle of the string . thus , to define the wavefront of entanglement spreading due to the string , one can look for the lattice site at which the entropy plateau starts to decrease . we identify this point computing the difference of entropy between nearest neighbor bipartitions : tracking when this quantity become bigger than a given threshold allows us to characterize the entanglement wavefront spreading . in fig . [ spreadingee : fig ] we show the estimated spreading velocity for different values of the threshold : the limit for the threshold value going to zero gives an estimate of the spreading velocity . a power law fit results in a spreading velocity of @xmath94 in very good agreement with the analytic estimate of @xmath95 and the result from the electric field of @xmath96 demonstrating the intimate connection between entanglement and electric field spreading . bound states are a fundamental component of gauge theories , and understanding their complex internal structure is one of the most ambitious goals of computational physics @xcite . in experiments , such internal structure is usually explored by colliding heavy ions , so that the energy released during the process can be released via particle - antiparticle creation . this makes the _ ab initio _ numerical simulations of such scattering processes challenging , as mc simulations suffer from a severe sign problem when tackling real - time dynamics . here , we show how tn simulations allow to investigate meson scattering for the ( 1 + 1)d qlm using tns . first , we present a general procedure to implement scattering processes between composite particles , and discuss the electric field dynamics after the collision . then , we present results for the entanglement dynamics during and after the collisions showing that the meson collision is accompanied by the creation of entanglement between the mesons themselves . indeed , as we will show , the entanglement is bounded by the propagation wavefronts of the particles after collision , and is characterized by a constant plateau of the entanglement entropy within the region . in order to produce the scattering process , we shall start with two composite particles . each particle is charge / anti - charge pair , and divided only by one link , namely a meson , with opposite momentum . for the two - meson problem , there is a simple picture in the strong coupling limit : the massless theory is a free massive boson ( meson ) theory that is expected to become weakly interacting once a small mass term is included . hence , in the strong coupling region , a possible two - meson bound state is loosely bound , while in the weak coupling region it is tightly bound . we start the numerical simulation with the state represented in the cartoon ( d ) in fig . [ summary : fig ] : two mesons separated by a vacuum state of ten sites , which can be straightforwardly be written in a simple , separable matrix product state with @xmath21 . we provide momentum to the mesons by adiabatically moving them from the boundaries toward the center of the system : this is done by introducing a deep box - shaped potential that decouples the mesons from the rest of the system , with the only dynamics allowed being the oscillate between its position and a neighboring site . the box - potential is removed at time @xmath97 when the meson is exactly at half oscillation : from that point on the mesons evolve freely with an effective momentum mostly in one direction , one towards the other and eventually colliding , the meson wave packets still spread slightly in both directions : however the centers of mass of the two mesons are clearly propagating with constant and opposite momentum until the scattering process occurs and the particles bounce back . ] . in order to avoid vacuum fluctuations during the process , we choose a large value of @xmath98 . [ scatt : fig ] shows an example of such a scattering process . in particular , it shows the absolute value of the electric field of two mesons approaching each other , colliding in the center and there parting again . while before the collision the meson are tightly bound , after the scattering process the electric field diffuses , and the corresponding wavefront has a significantly attenuated signal . in the lower panel of fig . [ scatt : fig ] , we monitor the time - evolution of the total particle number ( blue ) , clearly indicating that this quantity is approximately conserved over the entire time - evolution , due to the large electric field strength , which suppresses particle - antiparticle creation . a classical - like picture of the scattering process presented above , would have that two particles move against each other and then bounce back as there is not enough energy available to generate a more complex inelastic scattering . however , this picture is oversimplified , as this is a fully quantum process . indeed one can , once more , monitor the quantum correlations generated during the scattering process . this is done in fig . [ escheat : fig ] , where we show the evolution of the bipartite entanglement entropy : one sees that entanglement is created and that it is mostly carried by the two mesons . in this parameter regime , the vacuum does not generate entanglement due to the very large value of @xmath99 . studying the bipartite entanglement entropy for different bipartitions and times , one clearly sees that there are two regimes : before the scattering occurs , the entanglement is present only in the bipartition that cuts the mesons wave packets , indicating two electron - positron wave packets internally correlated but not sharing any quantum correlations among them . to the contrary , after the scattering , the two wave packets become highly correlated even when their two centers of mass are clearly separated ( see fig . [ scatt : fig ] for times @xmath100 ) . the values of the entanglement entropy indicate that one ebit of quantum information has been created during the scattering process . in the right panel of fig . [ escheat : fig ] , we present various cuts of the entanglement entropy profile taken at different times , together with a comparison with the entanglement generated by a single meson moving through the lattice ( dashed line ) . the difference of @xmath101 between the two cases ( highlighted in fig . [ escheat : fig ] , green bar ) clearly shows that one additional _ ebit _ of entanglement has been generated during the scattering process : that is , that a singlet state has been created between the two indistinguishable mesons . the entropy has increased as the information on which process occurs ( either mesons bouncing and moving back , or each one moving freely through the other one and keeping moving with its own momentum ) is completely lost . notice that this is a direct consequence of the wave nature of the mesons : the classical case of two scattering particles in one dimension - with the constraint that no double occupancy might occur in a single matter site - would necessary results only in backscattering , since the two hard classical particles can not pass through each other . we presented a detailed tensor network study on the real - time dynamics of a lattice gauge theory in the presence of dynamical charges and quantum gauge fields . within this approach , we have shown that one has direct access to all local quantities of interests - the time evolution of the mass , charge and gauge fields - and to the quantum correlation between bipartitions of the system by means of the von neumann entropy . we investigated the primary and secondary string breaking in qed in ( 1 + 1)d represented by an @xmath102 quantum link model with staggered fermions . in this context , we studied the real - time evolution of the schwinger mechanism leading to mass creation and annihilation by means of the interplay with the electric energy released by the string . we quantified key properties of these effects such as the mass production rate of the schwinger mechanism and the velocity of the electric field spreading . moreover , we unveiled the relation between string breaking dynamics and the entanglement spreading in the systems and we showed that it is possible to study scattering dynamics , characterizing not only mass and charge real - time evolution but also the creation of quantum correlations among scattered particles . finally , we showed that the presented results can be in principle verified experimentally in possible future quantum simulations , as they appear to be robust with respect to the most common sources of gauge - invariant imperfections appearing in most of the proposed implementations . this work paves the way to systematic studies of real time dynamical phenomena in abelian and non - abelian lgts in low dimensional systems . indeed , the present approach can be straightforwardly generalized to more complex lgt and geometries , e.g. , ladders or cylinders , and also can be studied in presence of an external environment by means of , for example , the tensor network approach presented in @xcite . moreover , one can also study the continuum limit of the lgt ( as already discussed for wilson theories @xcite ) : for qlms , this can either be done using dimensional reduction @xcite , or by increasing the quantum link representation , similarly to what has been done in @xcite . notice that the gauge invariant tensor network formulation behaves favorably as the speed - up it grants scales as the number of rishons per link squared @xcite . the unprecedented access to the entanglement dynamics in lgts will allow investigations on the role of quantum correlations in different contexts enabling a deeper understanding of the quantum real - time dynamics of lattice gauge theories . finally , exploiting the capability to prepare a wide class of complex states granted from recent developments in quantum optimal control of many - body quantum systems @xcite , more complex dynamics could be investigated . one example would be to perform extensive studies of the scattering at different energies . alongside , these methods can also be applied to study condensed matter systems to compute , e.g. , response functions of antiferromagnets described by lgt ( spin ices , resonating valence bond models , etc . ) which are very hard to evaluate using mc due to analytic continuation @xcite . moreover , real time dynamics of gauge theories is fundamentally interesting to ab initio investigations of scattering equilibration and pre - thermalization @xcite . finally to benchmark and verify small quantum simulations whose proposal have recently appeared for different platforms ( ranging from cold atoms to superconducting circuits and trapped ions ) and that can be foreseen to be experimentally implemented in the next years @xcite . while writing this manuscript , we became aware of a related work on string breaking using tn methods @xcite . we acknowledge useful discussions with d. banerjee , j. berges , p. hauke , f. hebenstreit , v. kasper , e. martinez , t. monz , s. pascazio , and u .- j . wiese , and thank c. laflamme and m. rider for proofreading the manuscript . md , sm and pz would like to acknowledge the hospitality of the institute of nuclear theory - university of washington , where this work was finalized . work in innsbruck is partially supported by erc synergy grant uquam , siqs , and the sfb foqus ( fwf project no . f4016-n23 ) , in ulm by the eu projects siqs and rysq , and by the dfg project sfb / trr21 , while in bilbao , we acknowledge financial support from basque government grants it472 - 10 and it559 - 10 , spanish mineco fis2012 - 36673-c03 - 02 , upv / ehu ufi 11/55 , promisce and scaleqit european projects . we investigated string - breaking in a one - dimensional system with up to 100 lattice sites . the simulation was performed using a matrix - product - state ( mps ) algorithm using a second order suzuki - trotter decomposition for the time - evolution . in this appendix we present the parameters we used for our simulations and provide a discussion on the relative errors . the main sources of numerical error in our calculations are the finite bond dimension of the mps - state representation and the suzuki - trotter time - step . as a bond dimension we used a value up to @xmath103 , which ensures a truncation error on the corresponding wave function of maximum order @xmath104 and @xmath105 for @xmath106 and @xmath107 respectively for the string breaking calculations , and @xmath108 for @xmath109 for the scattering processes . examples of the change of the truncation error during the evolution can be seen in the inset of the left panel in fig . [ convergence : fig ] . in order to make sure that these errors lead to small changes in our main observables , we tested the convergence of the mean electric field in the center of the chain ( cf . [ stringbreakingevolution : fig ] ) for different values of @xmath110 and typical system parameters . the results can be seen in the left panel of fig . [ convergence : fig ] where we plot the mean electric field at the end of the evolution ( @xmath111 ) as a function of the inverse bond dimension @xmath112 . the mean electric field was subtracted by a fitted offset @xmath113 to allow a better comparison between the two sets of system parameters . as we see from the plot , even for rather small bond dimensions @xmath114 the change to the largest bond dimension used ( @xmath115 ) is in the order of @xmath116 . for the bond dimension used in our simulations ( @xmath103 ) the difference to the extrapolated correct value is in the order of @xmath117 . + as said before , we used a second order suzuki - trotter decomposition with a time - step of @xmath118 to simulate the time - evolution . in the right panel of fig [ convergence : fig ] we report the convergence test obtained repeating the same simulation with different time - steps . as expected , we find a clear @xmath119 dependence : this clean power - law behavior allows us to give a very good estimate of the correct value of @xmath113 . in summary , the error from the suzuki - trotter decomposition using a time - step of @xmath118 is of the order of @xmath117 , that is , of the same order as the error introduced by the truncated bond dimension . from the earliest discussion of the schwinger model , the screening of static charges due to vacuum polarization has attracted a lot of interest @xcite . recent semiclassical calculations have reproduced this process including also finite masses where it could be shown that the main dynamics for @xmath120 is mostly independent from the mass @xcite . here , we show that it is possible to study such process using quantum link models implemented in a tensor network algorithm , already with very small spin representation , e.g. @xmath102 . as the screening of charges can be observed only in asymptotically large times , the large correlations building up during the process have carefully to be kept under control , otherwise they would undermine the efficiency of our approach . thus , here we limit our system size to @xmath121 lattice sites and encode the two static charges which build up the initial string in the boundary conditions on both sides of the lattice . in such a way , we avoid to simulate the vacuum regions where quantum correlations build up very quickly ( see fig . [ overview : fig ] ) and are less interesting for the phenomena under study . thus , we simulate the time evolution of the string of electric field among the charges with a small , but finite mass : after a long time compared to other timescales of the system ( @xmath122 ) when the largest fluctuations in the lattice damped out , and report the final charge distribution @xmath123 ( averaged over the last @xmath124 to remove remaining oscillatory effects ) in fig . [ screening : fig ] . the net results is an exponential decay of the charge density with the distance from the charges , as discussed in ref . @xcite where the authors showed that the screening also at @xmath125 is equal to that experienced by the massless schwinger model , as @xmath126 with a decay rate @xmath127 @xcite . the parameter @xmath128 is derived from the normalization condition @xmath129 . it can be clearly seen that our findings are compatible with the theoretical results from ref . @xcite ( black solid line in fig . [ screening : fig ] ) . we stress that the goal of the presented analysis is to show that it is possible to perform a study of the screening of charges using tensor networks , not to perform a thorough discussion which will be presented elsewhere . indeed , an extensive numerical analysis is needed to individuate or rule out possible deviations of the massive from the massless case in a full quantum mechanical model , the influence of the spin representation in the quantum link formulation , possible finite size effect , and also more challenging regimes of parameters , e.g. @xmath130 as shown in @xcite . however , the agreement between the presented results and the semiclassical approach , indicates that both analysis are capable of describing the physics of the system and might be alternative approaches which can complement and benchmark each other . f. halzen , a.d . martin , _ quarks and leptons : an introductory course in modern particle physics_. wiley ( 1984 ) . t. degrand , c. detar , _ lattice methods for quantum chromodynamics _ , world scientific ( 2006 ) . p. langacker , _ the standard model and beyond . _ series in high energy physics , cosmology and gravitation . taylor & francis ( 2011 ) . rothe , _ lattice gauge theories : an introduction _ , world scientific ( 2012 ) . o. philipsen and h. wittig , _ string breaking in non - abelian gauge theories with fundamental matter fields _ , phys . lett . * 81 * , 4056 ( 1998 ) . f. knechtli and r. sommer , _ string breaking in su(2 ) gauge theory with scalar matter fields _ , phys . b * 440 * , 345 ( 1998 ) . bali , h. neff , t. dssel , t. lippert , k. schilling _ observation of string breaking in qcd _ d * 71 * , 114513 ( 2005 ) . k.g.wilson , _ confinement of quarks _ , phys . d * 10 * , 2445 ( 1974 ) . j. kogut , _ an introduction to lattice gauge theory and spin systems _ , rev . 51 * , 659 ( 1979 ) . j. kogut , _ the lattice gauge theory approach to quantum chromodynamics _ , rev . phys * 55 * , 775 ( 1983 ) . white , _ density matrix formulation for quantum renormalization groups _ lett . * 69 * , 2863 ( 1992 ) . m. fannes , _ finitely correlated states on quantum spin chains _ , b. nachtergaele , r.f . werner , comm . phys . * 144 * , 443 ( 1992 ) . j. dukelsky , m.a . martin - delgado , t. nishino , g. sierra , _ equivalence of the variational matrix product method and the density matrix renormalization group applied to spin chains _ , europhys . * 43 * , 457 ( 1998 ) . f. verstraete , d. porras , j. i. cirac , _ density matrix renormalization group and periodic boundary conditions : a quantum information perspective _ , phys . lett . * 93 * , 227205 ( 2004 ) . u. schollwoeck , _ the density - matrix renormalization group _ , rev . phys . * 77 * , 259 ( 2005 ) . a. osterloh , g. palacios , s. montangero , _ enhancement of pairwise entanglement via z(2 ) symmetry breaking _ , phys . lett * 97 * , 257201 ( 2006 ) . f. verstraete , m.m . wolf , d. prez - garca , j.i . cirac , _ criticality , the area law , and the computational power of projected entangled pair states _ , . lett . * 96 * , 220601 ( 2006 ) . d. prez - garca , f. verstraete , m. m. wolf , j.i . cirac , _ matrix product state representations _ * 7 * , 401 ( 2007 ) . g. de chiara , m. rizzi , d. rossini , s. montangero , _ density matrix renormalization group for dummies _ , j. comp . nanosci . * 5 * , 1277 ( 2008 ) . s. montangero , r. fazio , p. zoller , g. pupillo , _ dipole oscillations of confined lattice bosons in one dimension _ phys . rev . a * 79 * , 041602 ( 2009 ) m. sanz , m. m. wolf , d. prez - garca , j. i. cirac , _ matrix product states : symmetries and two - body hamiltonians _ , a * 79 * , 042308 ( 2009 ) . v. murg , f. verstraete , . legeza , r.m . noack , _ simulating strongly correlated quantum systems with tree tensor networks _ , phys . b * 82 * , 205105 ( 2010 ) . p. silvi , g. de chiara , t. calarco , g. morigi , s. montangero , _ full characterization of the quantum linear - zigzag transition in atomic chains _ , ann . phys . * 525 * , 827 ( 2013 ) . m. gerster , p. silvi , m. rizzi , r. fazio , t. calarco , s. montangero , _ unconstrained tree tensor network : an adaptive gauge picture for enhanced performance _ b * 90 * , 125154 ( 2014 ) . byrnes , p. sriganesh , r.j . bursill , c.j . hamer , _ density matrix renormalization group approach to the massive schwinger model _ , d * 66 * , 013002 ( 2002 ) . j. casanova , l. lamata , i. l. egusquiza , r. gerritsma , c. f. roos , j. j. garca - ripoll , e. solano , _ quantum simulation of quantum field theories in trapped ions _ , phys . . lett . * 107 * , 260501 ( 2011 ) . bauls , k. cichy , j.i . cirac , k. jansen , _ the mass spectrum of the schwinger model with matrix product states _ , jhep * 11 * , 158 ( 2013 ) . bauls , k. cichy , j.i . cirac , k. jansen , h. saito , _ matrix product states for lattice field theories _ proc . sci . lattice 2013 , 332 , ( 2013 ) . e. rico , t. pichler , m. dalmonte , p. zoller , s. montangero , _ tensor networks for lattice gauge theories and atomic quantum simulation _ lett . , * 112 * , 201601 ( 2014 ) . b. buyens , j. haegeman , k. van acoleyen , h. verschelde , f. verstraete , _ matrix product states for gauge field theories _ , phys , * 113 * , 091601 ( 2014 ) . s. khn , j.i . cirac , m.c . bauls , _ quantum simulation of the schwinger model : a study of feasibility _ , phys . a , * 90 * , 042305 ( 2014 ) . p. silvi , e. rico , t. calarco , s. montangero , _ lattice gauge tensor networks _ , new j. phys . , * 16 * , 103015 ( 2014 ) . l. tagliacozzo , a. celi , m. lewenstein , _ tensor networks for lattice gauge theories with continuous groups _ , x , * 4 * , 041024 ( 2014 ) . j. haegeman , k. van acoleyen , n. schuch , j.i . cirac , f. verstraete , _ gauging quantum states : from global to local symmetries in many - body systems _ , phys . x , * 5 * , 011024 ( 2015 ) . l. garca - lvarez , j. casanova , a. mezzacapo , i. l. egusquiza , l. lamata , g. romero , e. solano , _ fermion - fermion scattering in quantum field theory with superconducting circuits _ , lett . * 114 * , 070502 ( 2015 ) . d. horn , _ finite matrix models with continuous local gauge invariance _ b , * 100 * , 149 ( 1981 ) . p. orland , d. rohrlich , _ lattice gauge magnets : local isospin from spin _ , nucl . b , * 338 * , 647 ( 1990 ) . s. chandrasekharan , u - j . wiese , _ quantum link models : a discrete approach to gauge theories _ b , * 492 * , 455 ( 1997 ) . h. weimer , m. mller , i. lesanovsky , p. zoller , h.p . buchler , _ a rydberg quantum simulator _ , nat.phys . * 6 * 382 ( 2010 ) . e. kapit , e. mueller , _ optical - lattice hamiltonians for relativistic quantum electrodynamics _ , a , * 83 * , 033625 ( 2011 ) . e. zohar , b. reznik , _ confinement and lattice quantum - electrodynamic electric flux tubes simulated with ultracold atoms _ , phys . lett . , * 107 * 275301 ( 2011 ) . e. zohar , j. i. cirac , b. reznik , _ simulating compact quantum electrodynamics with ultracold atoms : probing confinement and non - perturbative effects _ , lett . , * 109 * , 125302 ( 2012 ) . d. banerjee , m. dalmonte , m. mller , e. rico , p. stebler , u .- j . wiese , p. zoller , _ atomic quantum simulation of dynamical gauge fields coupled to fermionic matter : from string breaking to evolution after a quench _ , phys . lett . , * 109 * , 175302 ( 2012 ) . e. zohar , j.i . cirac , b. reznik,_simulating ( 2 + 1)-dimensional lattice qed with dynamical matter using ultracold atoms _ , phys . lett . , * 110 * , 055302 ( 2013 ) . d. marcos , p. rabl , e. rico , p. zoller , _ superconducting circuits for quantum simulation of dynamical gauge fields _ , phys . lett . , * 111 * , 110504 ( 2013 ) . p. hauke , d. marcos , m. dalmonte , p. zoller , _ quantum simulation of a lattice schwinger model in a chain of trapped ions _ , phys . x , * 3 * , 041018 ( 2013 ) . l. tagliacozzo , a. celi , a. zamora , m. lewenstein , _ optical abelian lattice gauge theories _ , phys . , * 330 * , 160 ( 2013 ) . e. zohar , j.i . cirac , b. reznik , _ quantum simulations of gauge theories with ultracold atoms : local gauge invariance from angular - momentum conservation _ , a * 88 * , 023617 ( 2013 ) . d. marcos , p. widmer , e. rico , m. hafezi , p. rabl , u .- j . wiese , p. zoller , _ two - dimensional lattice gauge theories with superconducting quantum circuits _ , ann . phys . * 351 * , 634 ( 2014 ) . k. stannigel , p. hauke , d. marcos , m. hafezi , s. diehl , m. dalmonte , p. zoller , _ constrained dynamics via the zeno effect in quantum simulation : implementing non - abelian lattice gauge theories with cold atoms _ , phys . 112 * , 120406 ( 2014 ) . a. w. glaetzle , m. dalmonte , r. nath , i. rousochatzakis , r. moessner , p. zoller , _ quantum spin - ice and dimer models with rydberg atoms _ , phys . x * 4 * , 041037 ( 2014 ) . s. notarnicola , e. ercolessi , p. facchi , g. marmo , s. pascazio , f. v. pepe , _ discrete abelian gauge theories for quantum simulations of qed _ , arxiv . 1503.04340 . a. bazavov , y. meurice , s .- w . tsai , j. unmuth - yockey , j zhang , _ gauge - invariant implementation of the abelian higgs model on optical lattices _ , arxiv.1503.08354 . wiese , _ towards quantum simulating qcd _ a * 931 * , 246 ( 2014 ) . e. zohar , j.i . cirac , b. reznik , _ quantum simulations of lattice gauge theories using ultracold atoms in optical lattices _ , arxiv:1503.02312 ( 2015 ) . j. schwinger , _ on gauge invariance and vacuum polarization _ , phys . rev . * 82 * , 664 ( 1951 ) . j. schwinger , _ gauge invariance and mass _ , phys . , * 125 * , 397 , ( 1962 ) . j. schwinger , _ gauge invariance and mass . ii _ , phys . rev . , * 128 * , 2425 ( 1962 ) . j.h . lowenstein , j.a . swieca , _ quantum electrodynamics in two dimensions _ * 68 * , 172 ( 1971 ) . a. casher , j. kogut , l. susskind , _ vacuum polarization and the absence of free quarks _ , d , * 10 * , 732 ( 1974 ) . s. coleman , _ more about the massive schwinger model _ , * 101 * , 239 ( 1976 ) . a.m. polyakov , _ quark confinement and topology of gauge theories _ , b , * 120 * , 429 ( 1977 ) . t. banks , r. myerson , j. kogut , _ phase transitions in abelian lattice gauge theories _ nucl . b , * 129 * , 493 ( 1977 ) . drell , h.r . quinn , b. svetitsky , m. weinstein , _ quantum electrodynamics on a lattice : a hamiltonian variational approach to the physics of the weak - coupling region _ d , * 19 * , 619 ( 1979 ) . s. ben - menahem , _ confinement in compact qed for low couplings _ d , * 20 * , 1923 ( 1979 ) . s. coleman , _ aspects of symmetry : selected erice lectures _ , cambridge university press ( 1985 ) . n. szpak , r. schtzhold , _ optical lattice quantum simulator for quantum electrodynamics in strong external fields : spontaneous pair creation and the sauter - schwinger effect _ , new j. phys . , * 14 * , 035001 ( 2012 ) . v. kasper , f. hebenstreit , j. berges , _ fermion production from real - time lattice gauge theory in the classical - statistical regime _ , d * 90 * , 025016 ( 2014 ) . c. holzhey , f. larsen , f. wilczek , _ geometric and renormalized entropy in conformal field theory _ b424 , 443 ( 1994 ) . p. calabrese , j. cardy , b. doyon , j. phys . a : math . and theor . * 42 * , 50 ( 2009 ) . g. vidal , _ efficient classical simulation of slightly entangled quantum computations _ lett . * 91 * , 147902 ( 2003 ) . a.j . daley , c. kollath , u. schollwck , g. vidal , _ time - dependent density - matrix renormalization - group using adaptive effective hilbert spaces _ , j. stat . p04005 ( 2004 ) . s. stlund , s. rommer , _ thermodynamic limit of density matrix renormalization _ , phys . lett . * 75 * , 3537 ( 1995 ) . s. singh , r. n. c. pfeifer , g. vidal , _ tensor network decompositions in the presence of a global symmetry _ , phys . a * 82 * 050301(r ) ( 2010 ) . b. bauer , p. corboz , r. orus , m. troyer , _ implementing global abelian symmetries in projected entangled - pair state algorithms _ , phys . b * 83 * , 125106 ( 2011 ) . a. weichselbaum , _ non - abelian symmetries in tensor networks : a quantum symmetry space approach _ , annals of physics * 327 * , 2972 ( 2012 ) . s. drr , z. fodor , j. frison , c. hoelbling , r. hoffmann , s. d. katz , s. krieg , t. kurth , l. lellouch , t. lippert , k. k. szabo , g. vulvert , _ ab initio determination of light hadron masses _ , science * 322 * , 1224 ( 2008 ) . p. doria , t. calarco , s. montangero , _ optimal control technique for many - body quantum dynamics _ , lett . * 106 * , 190501 ( 2011 ) . s. lloyd and s. montangero , _ information theoretical analysis of quantum optimal control _ , lett . * 113 * 010502 ( 2014 ) .
tensor network algorithms provide a suitable route for tackling real - time dependent problems in lattice gauge theories , enabling the investigation of out - of - equilibrium dynamics . we analyze a u(1 ) lattice gauge theory in ( 1 + 1 ) dimensions in the presence of dynamical matter for different mass and electric field couplings , a theory akin to quantum - electrodynamics in one - dimension , which displays string - breaking : the confining string between charges can spontaneously break during quench experiments , giving rise to charge - anticharge pairs according to the schwinger mechanism . we study the real - time spreading of excitations in the system by means of electric field and particle fluctuations : we determine a dynamical state diagram for string breaking and quantitatively evaluate the time - scales for mass production . we also show that the time evolution of the quantum correlations can be detected via bipartite von neumann entropies , thus demonstrating that the schwinger mechanism is tightly linked to entanglement spreading . to present the variety of possible applications of this simulation platform , we show how one could follow the real - time scattering processes between mesons and the creation of entanglement during scattering processes . finally , we test the quality of quantum simulations of these dynamics , quantifying the role of possible imperfections in cold atoms , trapped ions , and superconducting circuit systems . our results demonstrate how entanglement properties can be used to deepen our understanding of basic phenomena in the real - time dynamics of gauge theories such as string breaking and collisions .
the accumulating astronomical evidence for an accelerating cosmic expansion has stimulated many investigations into the nature of the dark energy , or other possible deviant gravitational effects , which might be responsible for this unexpected dynamics ( for a review see , _ e.g. _ , @xcite ) . besides proposing to add some new ( and purely theoretical ) matter species into the energy budget of the universe , many investigators have also focused their attentions on modifying general relativity ( gr ) on the largest scales , so as to introduce significant modifications in the behaviour of gravity at late times when it is comparatively weak . one example of the latter sort is provided by the family of @xmath3 gravity models , which had also been considered before the discovery of cosmic acceleration ( see for example refs . @xcite ) with reference to alternative forms of inflation and the existence of singularities . in refs . @xcite , the authors discuss a specific model where the correction to gr is a polynomial function of the @xmath4 and @xmath5 quadratic curvature invariants ( here , @xmath6 and @xmath7 are respectively the ricci scalar , ricci tensor and riemann tensor calculated in the standard way from the physical metric @xmath8 ) and showed that there exist late - time accelerating attractors in friedmann cosmological solutions to the theory . barrow and clifton established the general existence conditions for de sitter , einstein static , gdel universes in theories where the lagrangian is an arbitrary function of these three invariants @xcite . when the ricci scalar @xmath9 in the einstein - hilbert action is replaced by some general functions of @xmath9 and @xmath10 , it becomes necessary to distinguish between two different variational approaches to deriving the field equations . in the metric approach , as in refs . @xcite , the metric components @xmath8 are the only variational quantities and the field equations are generally of fourth - order , which makes the theories phenomenologically richer but more stringently constrained in most cases . within the palatini variational approach , on the other hand , we treat the metric @xmath8 and the connection @xmath11 as independent variables and extremize the action with respect to both of them ; the resulting field equations are second order and easier to solve . the palatini @xmath3 gravity is also proposed as an alternative to dark energy in a series of works @xcite . there has since been growing interest in these modified gravity theories : for the local tests of the palatini and metric @xmath3 gravity models see @xcite ; for the cosmologies of these two classes of models see @xcite . both approaches to modifying gravity are far from problem - free . in the metric @xmath3 gravity models , the theory is conformally related to standard gr plus a self - interacting scalar field @xcite , which generally introduces extra forces inconsistent with solar system tests @xcite . the palatini approach , on the other hand , generally leads to a large ( or even negative ) sound - speed - squared term in the growth equation of the non - relativistic matter perturbations on small scales . this induces effects on the cosmic microwave background ( cmb ) and the matter clustering power - spectra which deviate unacceptably from those which are observed @xcite . again , these examples highlight the difficulties encountered when trying to make modifications to standard gr which are compatible with observations . in this work we will focus on the ricci - squared gravity models within the palatini variational approach , which we also denote by the @xmath12 gravity . it turns out that the ricci tensor , @xmath13 and ricci scalar , @xmath14 , appearing in the field equations in the palatini approach are not the ones calculated from the physical metric , @xmath8 , ( we consider the metric @xmath8 as the physical one because it is this metric which the matter lagrangian density @xmath15 depends on and the energy - momentum conservation law holds with respect to ) as in gr , and we denote the gr equivalents by @xmath16 and @xmath9 respectively to distinguish them from the palatini quantities ) . such a modification of gravity has indeed been considered in @xcite and shown to give an accelerating cosmology . however , our work differs from @xcite in that we replace the ricci scalar in the gravitational action with @xmath17 rather than simply @xmath18 and we concentrate more on the cosmology at the first - order perturbation level , especially the late - time cold - dark - matter ( cdm ) density perturbation growth . we emphasize the similarity to the palatini @xmath19 gravity models also . our presentation is organized as follows . in sec . [ sect : equations ] we briefly introduce the model and outline the methods used to derive the background and first - order covariant and gauge - invariant ( cgi ) perturbed field equations . in sec . [ sect : background ] , we present the modified friedmann equation and apply it to a specific family of theories with @xmath20 ; the constraints on the parameter space @xmath21 from cosmological data are also given . then , in sec . [ sect : perturbations ] , we analyse the growth of cdm density perturbations at late cosmological times for this model . since this analysis shares some similarities with the palatini @xmath19 gravity models , we also present a similar discussion of the latter for comparison . our discussion and conclusions are presented in sec . [ sect : conclusions ] . throughout this work our convention is chosen as @xmath22u^{c}=r_{ab\ d}^{\ \ c}u^{d},r_{ab}=r_{acb}^{\ \ \ c}$ ] where @xmath23 run over @xmath24 and @xmath25 ; the metric signature is @xmath26 and the universe is assumed to be spatially flat and filled with cdm and black body radiation . in this section we briefly introduce the main ingredients of @xmath12 gravity and outline the strategy for deriving the general perturbation equations that govern the dynamics of small inhomogeneities in the cosmological models that arise in this theory . we will start our discussion with the modified einstein - hilbert action in the present model , @xmath27,\end{aligned}\ ] ] in which @xmath28 with @xmath29 the newtonian gravitational constant . here , @xmath30 is given by @xmath31 and @xmath32 ; note that @xmath33 is a new and independent variable with respect to which we extremise the action , and is different from the christoffel symbol @xmath11 calculated using the metric @xmath8 . @xmath34 is assumed to be a symmetric tensor ( if it contains an antisymmetric part then the field equation will be spoiled as discussed in @xcite ) and @xmath8 could be used to raise or lower its indices . varying the action eq . ( [ eq:1 ] ) with respect to the metric @xmath8 ( note that @xmath35 as they are independent ) gives the modified einstein equations : @xmath36 & = & \kappa t_{ab}^{f}\ \\end{aligned}\ ] ] where @xmath37 with @xmath38 and @xmath39 is the energy - momentum tensor of the fluid matter ( cdm and radiation ) . on the other hand , varying the action with respect to the new variable @xmath40 with the relation @xmath41 , one arrives at another field equation @xmath42 & = & 0,\end{aligned}\ ] ] where @xmath43 represents the covariant derivative compatible to @xmath44 ( the covariant derivative compatible to @xmath8 is denoted , as conventionally , by @xmath45 ) . just like in the palatini @xmath46 models , this equation implies some relation between the physical metric @xmath8 and the metric @xmath47 whose christoffel symbol is @xmath40 . however , because of the presence of the second term in the parentheses this relation is nontrivial and some further algebra will be needed to explicate it . before doing that , we will present some preliminary definitions and expressions , one of which is the notation of @xmath48 decomposition . the main idea of @xmath48 decomposition @xcite is to make spacetime splits of physical quantities with respect to the 4-velocity @xmath49 of an observer . the projection tensor @xmath50 is defined as @xmath51 and can be used to obtain covariant tensors perpendicular to @xmath49 . for example , the covariant spatial derivative @xmath52 of a tensor field @xmath53 is defined as @xmath54 the energy - momentum tensor and covariant derivative of the 4-velocity are decomposed respectively as @xmath55 in the above , @xmath56 is the projected symmetric trace - free ( pstf ) anisotropic stress , @xmath57 the heat flux vector , @xmath58 the isotropic pressure , @xmath59 the pstf shear tensor , @xmath60}$ ] , is the vorticity , @xmath61 ( @xmath62 is defined here as the mean expansion scale factor ) the volume expansion rate scalar , and @xmath63 is the fluid acceleration ; the overdots denote time derivatives expressed as @xmath64 , brackets mean antisymmetrisation , and parentheses symmetrization . the velocity normalization is chosen to be @xmath65 . the quantities @xmath66 are referred to as _ dynamical _ quantities and @xmath67 as _ kinematical _ quantities . note that the dynamical quantities can be obtained from the energy - momentum tensor @xmath68 through the relations @xmath69 decomposing the riemann tensor and making use the einstein equations , we could obtain , after linearization , the perturbed ( constraint and propagation ) equations @xcite . here , we shall not list all of them because most are irrelevant for the following discussion ; rather we will use the linearised raychaudhuri equation @xmath70 the linearised conservation equations for the energy density : @xmath71 and the linearised friedmann equation @xmath72 the above equations are derived and presented for standard general relativity , and so the @xmath73 variables describe imperfect fluid matter . for general modified gravity theories , such as those presented here , the modification to gr might be parameterized as an effective energy - momentum tensor . in this case the formalism of these equations is preserved and one just needs to replace @xmath74 by the total effective quantities of the same sort : @xmath75 @xcite . in eq . ( [ eq:4 ] ) , we see that @xmath76 is a symmetric @xmath77 tensor density of weight 1 , and so we can introduce a new metric @xmath47 by means of the following relation @xmath78 where the levi - civita connection of the metric @xmath47 is just @xmath40 , as we referred to above . to go further , we need to express @xmath34 explicitly . this is easy to do in principle , because eq . ( [ eq:3 ] ) is just an algebraic equation for @xmath34 . to see this , let us write the symmetric tensor @xmath34 in a general way as @xmath79 where @xmath80 is the 4-velocity of the observer referred to above . substituting eqs . ( [ eq:6 ] , [ eq:13 ] ) into eq . ( [ eq:3 ] ) , we get @xmath81 \nonumber \\ & & -\frac{1}{2}\left ( \delta + 3\xi + f\right ) u_{a}u_{b}-\frac{1}{2}\left ( \delta + 3\xi + f\right ) h_{ab } \nonumber \\ & = & \kappa \left(\rho^{f}u_{a}u_{b}-p^{f}h_{ab } + 2u_{(a}q_{b)}^{f } + \pi_{ab}^{f}\right ) , \nonumber\end{aligned}\ ] ] which leads to the following four equations : @xmath82 \upsilon _ { a } & = & \kappa q_{a}^{f } , \\ \label{eq:17 } ( 1 + 4f\xi ) \sigma _ { ab } & = & \kappa \pi _ { ab}^{f},\end{aligned}\ ] ] where @xmath83 are functions of @xmath84 . thus given the specified form of @xmath85 and the values of @xmath86 , the quantities @xmath87 can be obtained from eqs . ( [ eq:14 ] , [ eq:15 ] ) , at least numerically . then , @xmath88 and @xmath89 can also be calculated from eqs . ( [ eq:16 ] , [ eq:17 ] ) provided the values of @xmath90 and @xmath91 are given . note that @xmath88 and @xmath92 are nonzero only at first order in perturbation . taking the time derivatives of eqs . ( [ eq:14 ] , [ eq:15 ] ) , and using the background values of @xmath87 , we could easily obtain @xmath93 and @xmath94 by solving the two linear algebraic equations . similarly , @xmath95 and @xmath96 could be worked out ( here , @xmath52 is the spatial derivative ) . in what follows , we shall assume that @xmath97 and their derivatives have been calculated . the next step is to find out the relation between @xmath8 and @xmath47 . we could rewrite eq . ( [ eq:12 ] ) as @xmath98 taking the determinants of both sides and equating we get @xmath99 with @xmath100 thus , we conclude from eq . ( [ eq:18 ] ) that @xmath101 obviously , @xmath102 and @xmath103 need to be evaluated respectively . for @xmath102 , we have @xmath104\nonumber.\end{aligned}\ ] ] to calculate this , let us write @xmath105 $ ] and @xmath106 where @xmath107 run over @xmath108 , @xmath109 and @xmath110 are first order metric variables of which @xmath110 is traceless , @xmath111 is the spatial component of @xmath80 , and @xmath112 is the metric of 3 dimensional flat space . as a result @xmath113 and @xmath114 . from these expressions the components of @xmath115 can also be obtained and one can substitute all these quantities into the above equation to get @xmath116 . since @xmath88 and @xmath117 are only of first order and because @xmath117 is traceless , it is then not difficult to see that up to first order ( note that the facts @xmath118 and @xmath119 indicate that @xmath120 and @xmath121 ) @xmath122 for @xmath123 , we know that it is symmetric as the inverse matrix of a symmetric matrix , and so could be written as @xmath124 using @xmath125 and @xmath126 it is then easy to obtain , to first order , that @xmath127 as a result , we have now the relations between the two metrics @xmath47 and @xmath8 and their inverses as @xmath128 where @xmath129.\end{aligned}\ ] ] a discussion of how the two ricci tensors @xmath130 and @xmath131 are related to one another is given in the appendix , with the help of which the einstein equation eq . ( [ eq:3 ] ) can be rewritten as @xmath132 where @xmath133 and ( see the appendix for a definition of the tensor @xmath134 ) @xmath135 with the aid of eqs . ( [ eq:8 ] , [ eq:33 ] , [ eq:34 ] , [ eq:35 ] ) one could identify @xmath136 and @xmath137 and express them in terms of @xmath138 and @xmath139 which are functions of @xmath140 and @xmath141 ( c.f . ( [ eq:14 ] , [ eq:15 ] , [ eq:28 ] , [ eq:29 ] ) ) , and @xmath142 which are also functions of @xmath143 ( c.f . , ( [ eq:16 ] , [ eq:17 ] ) ) . however , from eqs . ( [ eq:26 ] - [ eq:31 ] , [ eq : a6 ] , [ eq : a7 ] ) one can see that this process will involve a lot of calculation . in the present work we will not perform detailed numerical calculations of the perturbation equations of the @xmath12 model ; instead , in the next two sections of the paper we will : \1 . study the background evolutions of general ricci - squared gravity models . as an example , we will consider a specific family of theories with @xmath1 , and constrain the allowed @xmath144 parameter space using data sets supernovae ( sne ) luminosity distances and the cmb shift parameter . present a simple argument to show that this class of modified gravity theory , like those arising in the palatini @xmath19 theory , generally possesses a scale - dependent effective sound - speed - squared term which affects the growth of cdm density perturbations and thus influences the matter power spectrum @xcite on small scales . in order to analyse the cosmological background evolution we can neglect the @xmath90 and @xmath91 terms , and hence the quantities @xmath145 . as a result , the equations are greatly simplified . we are interested in the modified friedmann equation in the present model . from eqs . ( [ eq:8 ] , [ eq:11 ] , [ eq:33 ] ) we have @xmath146 in which @xmath147 is the hubble expansion rate and @xmath148 is expressed as @xmath149 with @xmath150 and @xmath151 given in eqs . ( [ eq:34 ] , [ eq:35 ] ) . after a lengthy calculation , and using eq . ( [ eq:9 ] ) to eliminate the term @xmath152 which appears in eq . ( [ eq:37 ] ) , we obtain the following simple result @xmath153^{2 } & = & \frac{1}{6}(\delta -3\omega \xi ) .\end{aligned}\ ] ] there are two interesting points regarding eq . ( [ eq:38 ] ) . firstly , we see that only @xmath154 , and not its time derivatives @xmath155 or @xmath156 enter the equation . secondly , the second - order derivative of @xmath157 does not appear either ; to see the consequence of this , note that since @xmath158 ( with @xmath159 for matter and @xmath160 for radiation ) , the dependence of @xmath161 on @xmath141 could be expressed in terms of @xmath162 , _ e.g. _ , in radiation - dominated era @xmath163/3 and in matter - dominated era @xmath164 , and so we have @xmath165 . consequently , eq . ( [ eq:38 ] ) has the form @xmath166 where @xmath167 is a complicated function of @xmath140 and @xmath141 ( at late times @xmath168 and it becomes a function of @xmath140 alone ) . model . for clearness we only show the cases of @xmath169 ( the dashed curve ) and @xmath170 ( @xmath171 , the solid curve ) . the other parameters defining the cosmology are @xmath172 and @xmath173 . clearly an accelerating phase following the standard matter - dominated era can be realized within this model . note that the redshift range @xmath174 corresponds to the future of the present epoch , where the universe evolves into a de sitter stage . ] as we discussed above , knowing @xmath140 and @xmath141 means that we know @xmath175 and @xmath154 . thus , given a specific form for @xmath176 eq . ( [ eq:38 ] ) completely determines the background cosmological evolution of the model . as a particular example let us consider the case of @xmath177 where @xmath178 and @xmath179 are the model parameters . note that here @xmath180 is always non - negative , and @xmath181 corresponds to a picking the standard @xmath171 cosmology of gr . for convenience , we shall define the following dimensionless quantities @xmath182 then eqs . ( [ eq:14 ] , [ eq:15 ] ) could be rewritten as @xmath183 where @xmath184 and than eq . ( [ eq:38 ] ) reduces to @xmath185^{2}\frac{h^{2}}{h_{0}^{2}}=\frac{1}{6}\left ( \tilde{\delta}-3\omega \tilde{\xi}\right)\end{aligned}\ ] ] where @xmath186 . in this paper we will set @xmath173 , so today we have @xmath187 and there are 3 equations ( eqs . ( [ eq:42 ] , [ eq:43 ] , [ eq:46 ] ) ) for the 5 parameters @xmath188 and @xmath189 . therefore , we are able to express all the other quantities in terms of @xmath179 and @xmath190 , which can therefore be treated as the two independent degrees of freedom of our model . note that @xmath191 is a constant , and once evaluated at the present day , it could be used all through the cosmic history , which helps determine @xmath192 at arbitrary times . in figure [ fig : figure1 ] we have plotted the effective equation of state , defined by @xmath193 ( where a star - superscript denotes the derivative with respect to @xmath194 ) , as a function of the redshift . the values of @xmath179 are indicated beside the curves . at early times , when the @xmath12 corrections are negligible , the models all mimic the @xmath195 evolution of @xmath196 , and the same thing happens in the future . this is because during this era the matter ( relativistic and non - relativistic ) is greatly diluted so that the right - hand sides of eqs . ( [ eq:42 ] , [ eq:43 ] ) both vanish ; consequently , we can solve them to show that @xmath197 and so @xmath12 is also constant . the deviation from @xmath171 occurs mainly at intermediate times , that is , in the recent past and future . we now use the observational data on the background cosmology to constrain the parameter space ( in the @xmath198 plane ) of the present model . for this we jointly use the 157 measurements on sne luminosity distance in the gold data sets of riess _ et al_. @xcite and the cmb shift ( r ) parameter . the sne luminosity distance is expressed as @xmath199 where @xmath200 . the measurements supply the extinction - corrected distance modulus @xmath201 ( with @xmath202 in units of mega - parsecs ) and its uncertainty , @xmath203 , for individual sne , so that the standard @xmath204 minimization , defined by @xmath205 ^{2}}{\sigma _ { i}^{2}}\end{aligned}\ ] ] is easy to implement , where @xmath206 is the theoretically predicted distance modulus . as @xmath207 appears only as it does in eq . ( [ eq:47 ] ) , we could marginalize over it by integrating the probability density @xmath208 for all values of @xmath207 . for the cmb r - parameter , defined as @xmath209 we adopt the observational value @xmath210 at @xmath211 from @xcite . note that this does not depend on the specified value of @xmath207 . our constraining result is shown in figure [ fig : figure2 ] , where we have shown the 68% and 95% confidence regions respectively . the constrained intervals are roughly @xmath212 and @xmath213 at the 95% confidence level , with the best fitting values being @xmath214 with @xmath215 . also note that the concordance @xmath216 model ( the white star ) lies within the 68% confidence region of our constraints . and @xmath217 in the present model from joint sne and cmb shift parameter data sets . the grey and light grey regions represent the 68.3% and 95.4% confidence contours respectively . the white circle ( @xmath218 ) is the best - fitting parameter of our model , and the star is the concordance @xmath216 model . ] thus , we see that the background cosmological data is able to constrain @xmath219 to be of order @xmath220 . in the next section we will briefly investigate the possible constraint from the growth of dark - matter density perturbations , and show that this may provide a potentially more stringent restriction on @xmath221 . however , considering that this latter limit depends on the properties of the dark matter , our background constraints given in this section are less model dependent . in this section we study the effects of the @xmath12 corrections to gr on the cdm density perturbation growth . we start by recalling the case of @xmath19 gravity because it shares some similarities with the @xmath12 one , while being technically simpler than the latter , and because a similar analysis for the former is still missing from refs . @xcite ( see however @xcite for a slightly different treatment ) . recall that in our simplified model the universe is filled with cdm and radiation , and at later times the radiation energy density is negligible , so @xmath222 . taking the spatial derivative of the raychaudhuri equation eq . ( [ eq:9 ] ) ( with the @xmath223 there being replaced by @xmath224 ) , and working in the cdm frame ( where the observer is comoving with cdm particles and thus @xmath225 ) @xcite , we have @xmath226 where @xmath227 is the cdm density perturbation contrast that is defined through @xmath228 , and @xmath229 is the hubble expansion rate with respect to conformal time ( note that a prime denotes the conformal time derivative , and a dot the cosmic comoving proper - time derivative ) ; @xmath230 are respectively the harmonic expansion coefficients for @xmath231 and @xmath232 ( defined via @xmath233 and @xmath234 @xcite ) . clearly we need to know about @xmath235 and @xmath236 which arise from the @xmath12 modifications to gr ( c.f . ( 33 ) ) . in the palatini @xmath19 model , in which the ricci scalar @xmath14 in the gravitational action is replaced with @xmath237 , eqs . ( [ eq:26 ] , [ eq:27 ] ) still hold , but with ( see for example @xcite ) @xmath238then , with the help of the calculations in the appendix , it is straightforward to show that @xmath239where @xmath240 , and the modified einstein equation @xcite , @xmath241 can be rewritten as @xmath242 in which the effective total energy - momentum tensor is given by @xmath243using eq . ( [ eq:8 ] ) , we can now identify @xmath244 \nonumber \\ & & -\frac{3\dot{\lambda}^{2}}{2\lambda^{2}}+\frac{\ddot{\lambda}}{\lambda } + % \frac{1}{2\lambda } \box \lambda , \\ \label{eq:55 } \kappa p^{\mathrm{tot } } & = & \frac{1}{\lambda } \kappa p^{f}-\frac{1}{2\lambda } ( \mathbf{r}+f ) \nonumber \\ & & + \frac{1}{2}\left [ \mathbf{r}+\frac{3}{\lambda } \box \lambda -\frac{3}{2\lambda ^{2}}\nabla ^{a}\lambda \nabla _ { a}\lambda \right ] \nonumber \\ & & -\frac{1}{3\lambda } ( \theta \dot{\lambda}+\hat{\nabla}^{2}\lambda)-\frac{1}{2\lambda } \box \lambda .\end{aligned}\]]thus @xmath245 in which @xmath246 represent the terms _ not _ involving second order ( time and spatial ) derivatives . the reason why we keep only two second derivative terms explicitly on the right - hand side of eq . ( [ eq:56 ] ) is that , after taking the spatial covariant derivative , the first term contributes a @xmath247 piece to eq . ( [ eq:50 ] ) while the second term contributes a @xmath248 piece . none of the remaining terms in @xmath246 contribute these two pieces to eq . ( [ eq:50 ] ) . to be more explicit , recall that @xmath249 in the model , so @xmath250 where we have used eq . ( [ eq:10 ] ) to background order . as a result , @xmath251and eq . ( [ eq:50 ] ) can be recast into the form @xmath252 \delta _ { \mathrm{cdm}}^{\prime \prime } + [ \cdots ] \delta _ { \mathrm{cdm}}^{\prime } \nonumber \\ + \left [ \cdots + \frac{\dot{\lambda}}{6\lambda h}k^{2}\right ] \delta_{\mathrm{cdm } } & = & 0 \nonumber\end{aligned}\]]which , after rearrangement , gives @xmath253\delta_{\mathrm{cdm}}^{\prime } \nonumber \\ + \left[\cdots + \frac{\dot{\lambda}}{3(2\lambda h+\dot{\lambda})}k^{2}\right ] \delta_{\mathrm{cdm } } & = & 0.\end{aligned}\]]here , @xmath246 denotes complicated terms that are determined completely by the background evolutions of the model , and are unimportant for our analysis here . what is essential in eq . ( [ eq:58 ] ) is that it tells us that , as long as the quantity @xmath254 does not vanish , in general there will appear an effective sound - speed - squared term for the growth of matter density perturbations . depending on the sign of @xmath254 , this sound - speed - squared term could be either positive or negative , in both cases the small - scale density perturbation growth becomes extremely scale - dependent , altering the shape of the matter power spectra significantly @xcite . notice that the terms in @xmath255 $ ] could also modify the evolution of density contrasts differently as compared with the prediction in standard general relativity , but in a scale - independent manner , and at small scales their effects are subdominant . one more comment on the modified friedmann equation in the palatini gravity models is appropriate . using eq . ( [ eq:11 ] ) with @xmath256 replaced by the @xmath257 given in eq . ( [ eq:54 ] ) , we can see that only the @xmath258 terms are involved and the @xmath259 terms cancel ( we only consider terms to background order here ) . since @xmath260 we see that @xmath261 and could be moved to the left - hand side of eq . ( 11 ) ; the remaining terms on the right - hand side are also functions of @xmath140 only and so eq . ( [ eq:39 ] ) is also realized . to be more explicit , the modified friedmann equation in palatini @xmath19 gravity is @xcite @xmath262 ^{2 } & = & \frac{1}{6\lambda}\left [ \kappa ( \rho^{f}+3p^{f})-(\mathbf{r}+f)\right ] \nonumber\end{aligned}\ ] ] which can be shown to be just eq . ( [ eq:38 ] ) if @xmath263 there , as expected , because the metrics then take the same form ( of course , the definitions of @xmath161 are different in the two cases ) . now consider the @xmath12 gravity model . as discussed in the last section , the detailed forms of @xmath264 are very complicated , but fortunately we need not evaluate the full formulae explicitly . our experience of the simpler theory described in the last subsection shows that what is most relevant for our analysis are the second order ( time and space ) derivative terms ( note that the term @xmath265 , if exists , can also contribute to @xmath266 because @xmath267 contains @xmath268 where @xmath269 is the harmonic expansion coefficient of @xmath270 via @xmath271 , and because @xmath272 ; however it turns out no @xmath265 terms appears in @xmath273 ) , which are straightforward to identify . we shall formally repeat the procedure of the last subsection . note that the quantities @xmath274 are determined by @xmath275 which in our case are due to the radiation matter species . at late times the radiation energy density is negligible so that , to a good approximation , @xmath275 and thus @xmath274 , vanish . as a result , eqs . ( [ eq:26 ] , [ eq:27 ] ) become @xmath276with @xmath277 defined in eqs . ( [ eq:28 ] , [ eq:29 ] ) . meanwhile , since @xmath278 we have @xmath168 at late times ; @xmath279 , and hence @xmath161 and @xmath280 become functions of @xmath281 only , _ i.e. _ , @xmath282 , @xmath283 . the analysis in eq . ( [ eq:57 ] ) then also applies to @xmath161 and @xmath154 here , so @xmath284 now , from eq . ( [ eq:33 ] ) we obtain @xmath285where the relation @xmath286 and eq . ( [ eq:8 ] ) are used ; @xmath151 is given in eq . ( [ eq:34 ] ) . note that @xmath287 does not appear in this formula , and what we need to evaluate is just the collection of second - order derivative terms in @xmath288 . after some manipulation we obtain a similar expression to eq . ( [ eq:56 ] ) for the palatini @xmath19 model : @xmath289 the following analysis then completely parallels that for the palatini @xmath19 model and the cdm density perturbation growth equation can be shown to be ( like eq . ( [ eq:58 ] ) ) @xmath290\delta _ { \mathrm{cdm}}^{\prime } \nonumber \\ + \left[\cdots + \frac{(\lambda \omega ) ^{\cdot } } { 3(2\lambda h+\dot{\lambda})}k^{2}\right ] \delta _ { \mathrm{cdm } } & = & 0.\end{aligned}\ ] ] again the @xmath291 @xmath292 $ ] in eq . ( [ eq:63 ] ) denotes the terms which are completely determined by the background evolution and are not of interests to us here . thus we see that , similar to the case of palatini @xmath19 gravity model , a scale - dependent sound - speed - squared term also appears in the palatini @xmath12 gravity model , whose sign depends on @xmath293 . note that in the case of @xmath263 the metric @xmath47 ( eq . ( [ eq:59 ] , [ eq:60 ] ) ) has the same form as that in the palatini @xmath19 model , and eq . ( 63 ) reduces to eq . ( 58 ) as expected . in the general @xmath12 gravity models it is possible that @xmath294 , thus the scale - dependent effective sound - speed - squared term influences the matter perturbation growth and could alter the shape of the predicted matter power spectrum . our previous knowledge derived from refs . @xcite suggests that this effect might allow observational data to place stringent constraints on the parameter space of these theories . in fact , we can give a rough estimate of how stringent the constraint can be . consider first eq . ( [ eq:58 ] ) for the palatini @xmath19 model : since on small scales the @xmath295 term dominates the other terms in front of @xmath227 ( the @xmath296 terms ) , it is obvious that the magnitude of the quantity @xmath297 determines the deviation from @xmath195 results . the observational constraint on the model parameter @xmath179 ( recall that @xmath298 ) , as shown in refs . @xcite , is @xmath299 ; we take @xmath300 and @xmath301 for illustrative purposes , and find that @xmath302 for the relevant redshift range of @xmath303 . in the analogous case of palatini @xmath12 gravity , eq . ( [ eq:63 ] ) , we have @xmath304 as the dominant term instead of @xmath305 , and so the constraint should be @xmath306 in the same redshift range . again , taking @xmath301 , for the theories with @xmath20 we find that to satisfy the above constraint @xmath307 must also be limited to be @xmath308 or even smaller . thus , we conclude that the palatini @xmath12 models may be constrained by the observational data on the matter power spectrum just as stringently as are the palatini @xmath19 models . yet , we should note that a more exact quantitative constraint can only be obtained by exploiting a full parameter - space search as done in @xcite , and that the above conclusion depends on the assumption that the cdm particles have vanishing anisotropic stress . if , in contrast , the dark matter particles admit an anisotropic stress , in a manner similar to that prescribed in @xcite , then the effective sound - speed - squared terms might be canceled and leave no significant traces . we considered a general class of modified gravity models where the ricci scalar in the gravitational action of gr is replaced by a function @xmath17 and the field equations are derived using the palatini variational approach , _ i.e. _ , treating the metric @xmath8 and connection @xmath40 as independent variables so that the action is varied with respect to both of them . the strategy for deriving the cosmological equations at both the background ( zero - order ) and the first - order perturbation levels is outlined . the main step in this process is to determine the metric , @xmath47 , whose levi - civita connection is @xmath309 , relate it to the physical metric @xmath8 and thereby fix the relation between @xmath34 and @xmath16 . then , the correction to gr is treated as a new effective energy - momentum tensor while the field equations take the same form as in gr . the formulae laid down here might be useful for the numerical implementations of such modified gravity models . we also investigated in detail a power - law correction to the usual einstein - hilbert action given by @xmath20 . we used the snia luminosity distance and cmb shift parameter data to constrain its ( independent ) two - parameter space @xmath310 . we found that at 95% confidence level @xmath311 $ ] and @xmath312 $ ] . a slightly positive value of @xmath179 ( @xmath313 ) is preferred by the data used however , the standard @xmath171 model ( equivalent to @xmath181 ) with @xmath314 is still within the 68% confidence contour . hence , although the best fit to the data is a non - null leading - order correction to einstein gravity , the current data exhibits no significant preference over the concordance @xmath171 model of gr . the late - time growth of matter density perturbations in general palatini @xmath315 gravity models was also studied . it was shown that the equations governing this class of models look very similar to that in the palatini @xmath19 models . in particular , there exists a scale - dependent effective sound - speed - squared term in the perturbation growth equation which may be either positive or negative , depending on the background evolution in both models . in the @xmath19 case it is well known that these terms can lead to strong scale - dependence of the matter power spectrum , which is highly constrained by observational data @xcite , and we expect a similar feature to exist in the @xmath12 case . we estimate that this will produce a strong observational bound of @xmath316 unless some exotic properties are added to the dark matter candidate @xcite . as the final remark , we give a brief comment about the static and spherically - symmetric solutions of the present model . the analogue for the palatini @xmath19 model was considered in @xcite and the authors found that the exterior spherically - symmetric vacuum solutions are unique . here we just want to point out the @xmath317 model also shares this feature . in fact , in the vacuum where @xmath318 it is easy to show that for our model @xmath319 are uniquely determined by eqs . ( [ eq:14 ] , [ eq:15 ] ) and so is @xmath34 . the full consideration of a static system also needs the interior solution and its matching the exterior , which is beyond the scope of this paper and will be left for further investigation . in this appendix we present the relation between @xmath130 and @xmath131 if the two metrics @xmath47 and @xmath320 satisfy the following relations @xmath321where @xmath161 is a scalar function and @xmath322 symmetric tensors . firstly , the requirement @xmath323 implies that @xmath324 then , with some algebra , and using eq . ( [ eq : a4 ] ) , we can easily show that ( here a comma denotes the ordinary derivative ) @xmath325 \nonumber \\ & = & \gamma _ { bc}^{a}(g_{ab})+\gamma _ { bc}^{a},\end{aligned}\]]where the difference between @xmath40 and @xmath326 , denoted @xmath327 is defined by @xmath328 \nonumber \\ & & + \frac{1}{2}\left [ \zeta _ { b}^{a}\nabla _ { c}\lambda + \zeta_{c}^{a}\nabla_{b}\lambda -g_{bc}\zeta ^{ad}\nabla _ { d}\lambda \right ] \nonumber \\ & & + \frac{1}{2\lambda } \left [ \nabla _ { b}\xi _ { c}^{a}+\nabla_{c}\xi_{b}^{a}-\nabla ^{a}\xi _ { bc}\right ] \nonumber \\ & & + \frac{1}{2}\zeta^{ad}[\nabla_{c}\xi_{bd}+\nabla_{b}\xi_{cd}-\nabla_{d}\xi_{bc}]\end{aligned}\]]and is a true tensor , as expected . from the definition in eq . ( [ eq:2 ] ) we can thus derive that @xmath329 note that @xmath34 differs from @xmath16 by a rank-2 symmetric tensor . we thank andrzej borowiec , john miritzis , thomas sotiriou and the referee for helpful discussions . bl is supported by the overseas research studentship , cambridge overseas trust and the department of applied mathematics and theoretical physics ( damtp ) at the university of cambridge . dfm acknowledges the a. humboldt foundation . , * * , ( ) ; and , * * , ( 2005 ) ; and , * * , ( 2007 ) ; , and ( 2006 ) , ; and ( 2006 ) , ; , * * , ( ) ; , * * , ( ) ; , * * , ( ) ; ( 2007 ) , [ arxiv : astro - ph/0701662 ] ; and ( 2007 ) , arxiv : 0705.1158 [ astro - ph ] ; , , and , * * , ( ) . arxiv : 0704.1733 [ astro - ph ] ; and ( 2007 ) , arxiv : 0706.1378 [ hep - th ] ; , and , * * , ( 2006 ) ; , and , arxiv : astro - ph/0610854 ; , and , arxiv:0709.0046 [ gr - qc ] ; , and , arxiv:0709.0891 [ gr - qc ] ; and ( ) , arxiv:0708.0432 [ hep - th ] . , and , * * , ( ) ; , , and , * * , ( ) ; , * * , ( ) ; , and ( 2007),[arxiv : astro - ph/0701479 ] ; , and , * * , ( ) ; , and ( 2007 ) , arxiv : 0705.0232 [ gr - qc ] ; , * * , ( 2006 ) ; , * * , ( 2007 ) ; and , * * , ( 2007 ) ; , * * , ( 2006 ) ; and , * * , ( 2006 ) . , * * , ( ) ; , and ( 2006 ) , ; and , * * , ( 2006 ) ; , , and , * * , ( 2007 ) ; ( 2007 ) , arxiv : gr - qc/0703044 ; and ( 2007 ) , arxiv : 0704.0680 [ astro - ph ] ; and , * * , ( 2005 ) ; and , * * , ( 2006 ) ; and ( 2007 ) , arxiv : gr - qc/0702098 ; , * * , ( 2006 ) ; , , and , * * , ( ) . ( 2006 ) , arxiv : astro - ph/0601672 ; and , * * , ( 2006 ) ; and ( 2007 ) , ; , , and , * * , ( ) ; and , arxiv : hep - th/0601213 ; * * , ( ) ; see also arxiv : hep - th/0611071 ; , and , * * , ( ) ; , and , ; , , , and ( 2006 ) , ; , , and ( 2006 ) , ; , , and ( 2006 ) , [ arxiv : astro - ph/0612569 ] ; and ( 2007 ) , arxiv : 0705.0396 [ astro - ph ] ; ( 2007 ) , arxiv : 0705.1032 [ astro - ph ] ; , , and ( 2007 ) , arxiv : 0704.1932 [ astro - ph ] ; and , * * , ( ) ; , and * * , ( ) ; , , , and * * , ( ) ; , , , and ( ) , arxiv : hep - th/0611198 ; , and ( ) , arxiv : gr - qc/0701138 ; , and , * * , ( ) ; and , * * , ( ) . [ arxiv : gr - qc/0701111 ] ; , and ( 2007 ) , arxiv : 0706.0452 [ gr - qc ] ; and ( 2007 ) , arxiv : 0706.1146 [ astro - ph ] ; ( ) , arxiv : 0706.2041 [ astro - ph ] ; and ( ) , arxiv : 0705.3375 [ astro - ph ] ; , and ( ) , arxiv : 0706.1197 [ astro - ph ] ; , and ( ) , , * * , ( ) [ arxiv:0705.3795 [ gr - qc ] ] ; , and ( ) , arxiv : 0706.1663 [ gr - qc ] ; , and ( ) , arxiv : 0706.2399 [ astro - ph ] ; , and ( ) , arxiv : 0707.0106 [ gr - qc ] ; ( ) , arxiv:0707.0942 [ gr - qc ] ; and ( ) , arxiv:0707.1941 [ hep - th ] ; and ( ) , arxiv:0709.0296 [ astro - th ] ; ( ) , arxiv:0709.1391 [ astro - ph ] ; and ( ) , arxiv:0709.2375 [ astro - ph ] ; and ( ) , arxiv:0708.0924 [ hep - th ] ; , , and ( ) , arxiv:0708.0411 [ astro - ph ] ; , and ( ) , arxiv:0710.0814 [ gr - qc ] .
we consider the cosmology of the ricci - tensor - squared gravity in the palatini variational approach . the gravitational action of standard general relativity is modified by adding a function @xmath0 to the einstein - hilbert action , and the palatini variation is used to derive the field equations . a general method of obtaining the background and first - order covariant and gauge - invariant perturbation equations is outlined . as an example , we consider the cosmological constraints on such theories arising from the supernova type ia and cosmic microwave background observations . we find that the best fit to the data is a non - null leading - order correction to einstein gravity , but the current data exhibit no significant preference over the concordance model . the growth of non - relativistic matter density perturbations at late times is also analyzed , and we find that a scale - dependent ( positive or negative ) sound - speed - squared term generally appears in the growth equation for small - scale density perturbations . we also estimate the observational bound imposed by the matter power spectrum for the model with @xmath1 to be roughly @xmath2 so long as the dark matter does not possess compensating anisotropic stresses .
in the past years , experimentalist has made big progress on the search for the charmonium - like states , the so - called xyz states , in the @xmath14 meson decay , the @xmath15 collision , the @xmath16 fusion process , which have aroused extensive interest in revealing the underlying properties of the observed charmonium - like states ( see refs . @xcite for a review ) . the study of charmonium - like states is a research field full of challenges and opportunities in hadron physics . very recently the belle collaboration @xcite reported two charged @xmath17 structures around 10610 mev and 10650 mev by studying the @xmath18 ( @xmath19 ) and @xmath20 ( @xmath21 ) invariant mass spectra of @xmath22 hidden - bottom decay channels ( see refs . @xcite for theoretical progress ) . in ref . @xcite , we proposed the initial single pion emission ( ispe ) mechanism to explain the observed @xmath17 structures . by emitting a pion , @xmath23 decays into @xmath24 and @xmath25 mesons with low momentum . then , @xmath24 and @xmath25 mesons interact with each other by exchanging @xmath24 meson and transit into @xmath18 or @xmath20 . here , two structures near the @xmath26 and @xmath27 thresholds appear in the @xmath18 and @xmath20 invariant mass spectra , which could correspond to @xmath28 and @xmath29 @xcite . just indicated in ref . @xcite , if the ispe mechanism is an universal mechanism existing in heavy quarkonium decay , we can naturally extend such physical picture to study hidden - charm decays of higher vector charmonia due to the similarity between charmonium and bottomonium families , and predict some novel phenomena similar to the @xmath17 structures . in particle data book @xcite , six vector charmonia are established well , which are @xmath30 , @xmath31 , @xmath32 , @xmath2 , @xmath3 and @xmath4 with @xmath33 . among these charmonia , only @xmath2 , @xmath3 and @xmath4 are higher than the thresholds of @xmath34 , @xmath35 and @xmath36 . thus , we study the hidden - charm decays of @xmath2 , @xmath3 and @xmath4 via the ispe mechanism . from the analysis of such modes , one indicates that enhancement structures similar to the charged @xmath17 also exist in the charm case . in addition , in this work we will study the hidden - charm decays of @xmath5 , which is an important charmonium - like state observed by the babar collaboration in the @xmath37 process @xcite . its mass , width and @xmath38 are @xmath39 mev , @xmath40 mev and @xmath41 @xcite . the study presented in ref . @xcite indicates that @xmath5 can be related to charmonia @xmath3 and @xmath4 , where the @xmath5 structure can be reproduced by the interference of production amplitudes of the @xmath42 processes via direct @xmath15 annihilation and through intermediate charmonia @xmath3 and @xmath4 @xcite . we naturally apply the ispe mechanism existing in @xmath3 and @xmath4 decays to discuss the hidden - charm dipion decays of @xmath5 . thus , studying @xmath5 hidden - charm decays through the ispe mechanism is an intriguing issue , where we will also predict some enhancement structure similar to the charged @xmath17 . as announced by babar @xcite , @xmath5 was first observed in its @xmath9 decay channel . searching enhancement structure in the @xmath43 invariant mass spectrum of @xmath44 will be accessible in future experiments , which could provide a direct test to the non - resonant explanation for @xmath5 proposed in @xcite . this work is organized as follows . after the introduction , we illustrate the hidden - charm dipion decays of higher charmonia under the ispe mechanisms . in sec . iii , the numerical results are presented . the last section is the discussion and conclusion . with @xmath45 as an example , we first illustrate the possible decay mechanisms in the dipion hidden - charm decay of higher charmonium . @xmath2 can directly decay into @xmath9 . in refs . @xcite , the qcd multipole expansion method was proposed and can be applied to calculate such direct decay process . the second mechanism is that the dipion is from the intermediate scalar ( @xmath46 ) or tensor ( @xmath47 ) meson , where the hadronic loops constructed by the @xmath48 mesons could be as a bridge to connect @xmath2 and @xmath9 ( see ref . @xcite for more details ) . the remaining decay mechanism existing in the hidden - charm dipion decays of higher charmonia is the ispe mechanism , which was first proposed in ref . @xcite . by the quark - level diagram we give an explicit description ( left - side diagram in fig . [ ipe ] ) of the ispe mechanism in @xmath45 decay . the physical picture is that with a pion emission @xmath2 first dissolves into @xmath48 and @xmath49 mesons with low momentum , which further turn into @xmath6 . here , @xmath50 transition occurs via exchanging @xmath48 meson @xcite . an equivalent hadron - level description is also presented in the right - side diagram of fig . [ ipe ] , which can be as an effective approach for dealing with the practical calculations . [ cols="^,^,^,^ " , ] we also present the decay amplitude of @xmath51 via the intermediate @xmath52 . @xmath53_{d^\ast\bar{d}^\ast}=2\sum_{\alpha=1,\cdots,4}m_{d^\ast\bar{d}^\ast}^{(\alpha)}.\label{a1}\end{aligned}\ ] ] we list all diagrams contributing to @xmath45 in fig . [ dsds ] . among these eight eight diagrams , [ dds ] ( @xmath54 ) can be obtained by fig . [ dds ] ( @xmath55 ) ( @xmath56 ) if making the transformations @xmath57 and @xmath58 , which results in factor 2 in eq . ( [ a1 ] ) due to @xmath59 symmetry . the decay amplitudes @xmath60 and @xmath61 are expressed as @xmath62[ig_{d^\ast d \pi } ( -ip_{4 \lambda } ) ] \nonumber \\ & & \quad \times [ -g_{j/\psi d^\ast d } \varepsilon_{\delta \nu \theta \phi } ( ip_{5}^{\delta } ) \epsilon_{j/\psi}^{\nu } ( -ip_{2}^{\theta } ) ] \frac{-g_{\rho}^{\lambda } + p_{1 \rho } p_1^\lambda / m_{d^\ast}^2 } { p_1 ^ 2-m_{d^\ast}^2}\nonumber \\ & & \quad \times \frac{-g_{\beta}^ { \phi } + p_{2 \beta } p_2^\phi / m_{d^\ast}^2 } { p_2 ^ 2-m_{d^\ast}^2 } \frac{1}{q^2-m_d^2 } \mathcal{f}^2(q^2),\label{hb1}%\\\end{aligned}\ ] ] @xmath63 [ -g_{d^\ast d^\ast \pi } \varepsilon^{\delta \tau \theta \phi } ( -ip_{1 \delta } ) ( iq_{\theta } ) ] \nonumber\\ & & \quad \times [ -ig_{j/\psi d^\ast d^\ast } \epsilon_{j/\psi}^\nu ( ( -iq_{\nu } + ip_{2\nu } ) ) g_{\omega \lambda } + ( ip_{5 \omega } + iq_{\omega})g_{\nu \lambda } \nonumber\\ & & \quad + ( -ip_{2\lambda } -ip_{5 \lambda } ) g_{\nu \omega}]\frac{-g^{\rho \tau } + p_{1\rho } p_{1\tau}/m_{d^\ast}^2}{p_1 ^ 2-m_{d^\ast}^2 } \nonumber\\ & & \quad \times \frac{-g_{\beta}^ { \omega } + p_{2\beta } p_2^\omega / m_{d^\ast}^2 } { p_2 ^ 2-m_{d^\ast}^2 } \frac{-g_{\phi}^{\lambda } + q_\phi q^\lambda / m_{d^\ast}^2 } { q^2-m_{d^\ast}^2 } \mathcal{f}^2(q^2).\label{hb2}\end{aligned}\ ] ] thus , by eqs . ( [ hb1 ] ) and ( [ hb2 ] ) we can easily obtain decay amplitudes @xmath64 and @xmath65 corresponding to fig . [ dsds ] ( 3 ) and ( 4 ) , where the transformation @xmath66 is performed . in the following , we extend the same framework to study the dipion transition between @xmath2 and @xmath67 . by replacing @xmath30 with @xmath67 in fig . [ dds ] ( 1 ) , ( 3 ) , ( 4 ) , ( 6 ) , ( 7 ) , ( 9 ) , ( 10 ) and ( 12 ) and fig . [ dsds ] , we obtain all diagrams relevant to @xmath68 decay . the total decay amplitudes of @xmath69 via @xmath70 and @xmath52 are @xmath71_{d^\ast\bar{d}+h.c.}&=&2\sum_{\beta=1,\cdots,4}a_{d^\ast\bar{d}+h.c.}^{(\beta)},\label{c1}\\ \mathcal{m}[\psi(4040)\to h_c(1p)\pi^+\pi^-]_{d^\ast\bar{d}^\ast}&=&2\sum_{\kappa=1,\cdots,4}a_{d^\ast\bar{d}^\ast}^{(\kappa)},\label{c2}\end{aligned}\ ] ] respectively , where the concrete amplitude expressions are @xmath72 \nonumber\\ & & \quad \times [ ig_{d^\ast d \pi } ( -ip_4^\rho ) ] [ ig_{h_c d^\ast d^\ast } \varepsilon_{\delta \nu \theta \phi } ( ip_5^\delta ) \epsilon_{h_c}^\nu ] \frac{1}{p_1 ^ 2-m_d^2 } \nonumber \\ & & \quad \times \frac{-g_{\mu}^ { \phi } + p_{2\mu } p_2^\phi / m_{d^\ast}^2}{p_2 ^ 2 -m_{d^\ast}^2 } \frac{-g_{\rho}^ { \theta}+ q_\rho q^\theta / m_{d^\ast}^2 } { q^2-m_{d^\ast}^2 } \mathcal{f}^2(q^2 ) , % \\\end{aligned}\ ] ] @xmath73 [ -g_{d^\ast d^\ast \pi } \varepsilon_{\theta \phi \delta \tau } ( iq^\theta ) \nonumber\\ & & \quad \times ( -ip_1^\delta ) ] [ g_{h_c d^\ast d } \epsilon_{h_c \nu } ] \frac{-g_{\mu } { \phi } + p_{1 \mu } p_1^\phi / m_{d^\ast}^2 } { p_1 ^ 2-m_{d^\ast}^2}\nonumber\\ & & \quad \times \frac{1}{p_2 ^ 2-m_d^2 } \frac{-g^{\nu \tau } + q^\nu q^\tau / m_{d^\ast}^2}{q^2-m_{d^\ast}^2 } \mathcal{f}^2(q^2),\end{aligned}\ ] ] and @xmath74 \nonumber\\ & & \quad \times [ ig_{d^\ast d \pi } ( -ip_{4 \lambda})][g_{h_c d^\ast d } \epsilon_{h_c \nu } ] \frac{-g^{\rho \lambda}+ p_1^\rho p_1^\lambda / m_{d^\ast}^2 } { p_1 ^ 2-m_{d^\ast}^2 } \nonumber\\ & & \quad \times \frac{-g^{\beta \nu}+p_2^\beta p_2^\nu / m_{d^\ast}^2}{p_2 ^ 2-m_{d^\ast}^2 } \frac{1}{q^2-m_d^2}\mathcal{f}^2(q^2),\end{aligned}\ ] ] @xmath75 \nonumber\\ & & \quad \times[-g_{d^\ast d^\ast \pi } \epsilon_{\delta \tau \theta \phi } ( -ip_1^\delta ) ( iq^\theta ) ] [ ig_{h_c d^\ast d^\ast } \varepsilon_{\kappa \nu \lambda \omega } ( ip_5^\kappa ) \epsilon_{h_c}^\nu]\nonumber\\ & & \quad \times \frac{-g^{\rho \tau } + p_1^\rho p_1^\tau / m_{d^\ast}^2}{p_1 ^ 2-m_{d^\ast}^2 } \frac{-g^{\beta \omega}+p_2^\beta p_2^\omega / m_{d^\ast}^2 } { p_2 ^ 2- m_{d^\ast}^2 } \nonumber\\ & & \quad \times \frac{-g^{\phi \lambda}+ q^\phi q^\lambda / m_{d^\ast}^2}{q^2-m_{d^\ast}^2}\mathcal{f}^2(q^2).\end{aligned}\ ] ] after performing the transformation @xmath66 , @xmath76 , @xmath77 , @xmath78 and @xmath79 can be transferred into @xmath80 , @xmath81 , @xmath82 and @xmath83 respectively . the differential decay width for @xmath84 reads as @xmath85 with @xmath86 and @xmath87 , where the overline indicates the average over the polarizations of the @xmath2 in the initial state and the sum over the polarization of @xmath88 in the final state . replacing @xmath89 with @xmath90 or @xmath91 , we obtain the differential decay width for @xmath92 or @xmath93 . when studying the hidden - charm dipion decay of other higher charmonia @xmath3 , @xmath4 and charmonium - like state @xmath5 , we only need to replace the relevant coupling constants and the masses in the formulism of the @xmath2 decays . in this work , we are mainly concerned with the line shapes of the differential decay widths of @xmath2 , @xmath3 , @xmath4 and charmonium - like state @xmath5 decays into @xmath9 , @xmath10 and @xmath94 , which are dependent on the invariant mass spectra of @xmath6 , @xmath7 and @xmath12 . thus , we set the coupling constants of @xmath95 as 1 in our calculation . besides these coupling constants listed in sec . [ h ] , other input parameters are the masses involved in our calculation , which are taken from particle data book @xcite . cc & + & + + & + & in fig . [ result ] , we present the results of @xmath96 , @xmath97 and @xmath98 of @xmath2 , @xmath3 , @xmath4 , @xmath5 decays into @xmath9 , @xmath10 , @xmath94 . 1 . there exit sharp peak structures close to the @xmath99 threshold and the corresponding reflections in the distributions of @xmath96 , @xmath97 and @xmath98 of @xmath45 , @xmath100 and @xmath68 decays . we notice that this structure appearing in @xmath101 is not obvious comparing with the structure in @xmath102 or @xmath103 distribution . two sharp peaks appear in the @xmath104 and @xmath105 distributions , which are close the @xmath99 threshold . the structure in the @xmath6 invariant mass spectrum is more narrow than that in the @xmath12 invariant mass spectrum . 3 . in the hidden - charm dipion decays of @xmath4 , we find two sharp peak structures around the @xmath99 and @xmath106 thresholds appearing in the @xmath6 invariant mass spectra . in addition , a sharp peak close the @xmath106 threshold is observed in the @xmath12 invariant mass spectrum distribution . in the @xmath107 distribution , a peak near @xmath1 with its reflection form a broad structure . under the ispe mechanism , the intermediate @xmath108 can result in a very broad structure in the @xmath12 invariant mass spectrum distribution . 4 . there exist the sharp peaks close to @xmath99 threshold in the @xmath109 and @xmath110 distributions , the structures around @xmath106 threshold in the @xmath111 and @xmath110 distributions . the peak close the @xmath99 threshold and its reflection overlap with each each to form a broad structure in the @xmath112 invariant mass spectrum . we need to specify that the result presented in fig . [ result ] is obtained by taking @xmath113 . our study shows that the line shapes in fig . [ result ] are weakly dependent on the values of @xmath114 . with @xmath115 as an example , in fig . [ beta ] we illustrate the @xmath114 dependence of @xmath116 distribution , where the line shapes corresponding to @xmath117 remain almost unchanged . in this work , we study the line shapes of the differential decay widths of @xmath2 , @xmath3 , @xmath4 and charmonium - like state @xmath5 decays into @xmath9 , @xmath10 and @xmath94 , where the ispe mechanism is introduced . furthermore , we predict the sharp peak structures close to @xmath99 and @xmath106 thresholds appearing the corresponding @xmath6 , @xmath7 and @xmath12 invariant mass spectra . the ispe mechanism plays crucial role to form these novel charged charmonium - like structures in the hidden - charm dipion decays of higher charmonia . to some extent , these predicted structures are the charmonium analogue of two newly observed @xmath17 structures in the hidden - bottom dipion decays of @xmath23 @xcite . we suggest further experimental search for these predicted charmonium - like structures close to the @xmath99 and @xmath106 thresholds . recently , besiii has stated accumulating @xmath2 data with an aim to search for higher charmonia and the charmonium - like states @xcite . our result shows the charged structures around the @xmath99 threshold in the @xmath6 , @xmath7 and @xmath8 invariant mass spectra of @xmath2 decays into @xmath9 , @xmath10 and @xmath11 , which are accessible at besiii and could be considered in future studies . since these charged charmonium - like structures also exist in the @xmath6 , @xmath7 and @xmath8 invariant mass spectra of @xmath13 hidden - charm dipion decays , carrying out the search for them will be an important and intriguing research topic , especially at belle and babar . if these predicted enhancement structures can be found in the higher charmonium and y(4260 ) hidden - charm decays , it will provide direct test to the ispe mechanism existing in the higher charmonium or y(4260 ) hidden - charm dipion decays . note added : after completion of this work , we notice a measurement reported by cleo - c @xcite . recently the cleo - c collaboration announced the measurement of the @xmath118 mass distribution from @xmath119 at @xmath120 mev ( points with errors in fig . 4 ( b ) of ref . thus , we compare the predicted @xmath121 mass distribution of @xmath122 ( see fig . [ result ] ) with the cleo - c result . we notice that the predicted theoretical line shape of the @xmath118 mass distribution of @xmath122 in this work is consistent with the one measured by cleo - c , which is listed in fig . [ com ] , where there indeed exist a broad structure around @xmath35 threshold and its reflection . to some extent , this fact provides a direct test to our prediction presented here . x.l . would like to thank chang - zheng yuan for suggestive discussion . this project is supported by the national natural science foundation of china under grants nos . 10705001 , no . 11005129 , no . 11035006 , no . 11047606 , the ministry of education of china ( fanedd under grant no . 200924 , dpfihe under grant no . 20090211120029 , ncet under grant no . ncet-10 - 0442 , the fundamental research funds for the central universities ) , and the west doctoral project of chinese academy of sciences . e. s. swanson , phys . rept . * 429 * , 243 ( 2006 ) [ arxiv : hep - ph/0601110 ] . s. l. zhu , int . j. mod . e * 17 * , 283 ( 2008 ) [ arxiv : hep - ph/0703225 ] . s. godfrey and s. l. olsen , ann . nucl . part . sci . * 58 * , 51 ( 2008 ) [ arxiv:0801.3867 [ hep - ph ] ] . m. nielsen , f. s. navarra and s. h. lee , phys . rept . * 497 * , 41 ( 2010 ) [ arxiv:0911.1958 [ hep - ph ] ] . n. brambilla _ et al . _ , eur . j. c * 71 * , 1534 ( 2011 ) [ arxiv:1010.5827 [ hep - ph ] ] . i. adachi _ et al . _ [ belle collaboration ] , arxiv:1105.4583 [ hep - ex ] . y. r. liu , x. liu , w. z. deng and s. l. zhu , eur . j. c * 56 * , 63 ( 2008 ) [ arxiv:0801.3540 [ hep - ph ] ] . x. liu , z. g. luo , y. r. liu and s. l. zhu , eur . j. c * 61 * , 411 ( 2009 ) [ arxiv:0808.0073 [ hep - ph ] ] . a. e. bondar , a. garmash , a. i. milstein , r. mizuk and m. b. voloshin , arxiv:1105.4473 [ hep - ph ] . d. y. chen , x. liu and s. l. zhu , arxiv:1105.5193 [ hep - ph ] . j. r. zhang , m. zhong and m. q. huang , arxiv:1105.5472 [ hep - ph ] . y. yang , j. ping , c. deng and h. s. zong , arxiv:1105.5935 [ hep - ph ] . d. v. bugg , arxiv:1105.5492 [ hep - ph ] . i. v. danilkin , v. d. orlovsky and yu . a. simonov , arxiv:1106.1552 [ hep - ph ] . t. guo , l. cao , m. z. zhou and h. chen , arxiv:1106.2284 [ hep - ph ] . z. f. sun , j. he , x. liu , z. g. luo and s. l. zhu , arxiv:1106.2968 [ hep - ph ] . d. y. chen and x. liu , arxiv:1106.3798 [ hep - ph ] . k. nakamura _ et al . _ [ particle data group ] , j. phys . g * 37 * , 075021 ( 2010 ) . b. aubert _ et al . _ [ babar collaboration ] , phys . lett . * 95 * , 142001 ( 2005 ) [ arxiv : hep - ex/0506081 ] . d. y. chen , j. he and x. liu , phys . d * 83 * , 054021 ( 2011 ) [ arxiv:1012.5362 [ hep - ph ] ] . kuang and t. m. yan , phys . d * 24 * , 2874 ( 1981 ) . t. m. yan , phys . d * 22 * , 1652 ( 1980 ) . v. a. novikov and m. a. shifman , z. phys . c * 8 * , 43 ( 1981 ) . d. y. chen , j. he , x. q. li and x. liu , arxiv:1105.1672 [ hep - ph ] . y. s. oh , t. song and s. h. lee , phys . c * 63 * , 034901 ( 2001 ) [ arxiv : nucl - th/0010064 ] . r. casalbuoni , a. deandrea , n. di bartolomeo , r. gatto , f. feruglio and g. nardulli , phys . rept . * 281 * , 145 ( 1997 ) [ arxiv : hep - ph/9605342 ] . p. colangelo , f. de fazio and t. n. pham , phys . b * 542 * , 71 ( 2002 ) [ arxiv : hep - ph/0207061 ] . a. anastassov _ et al . _ [ cleo collaboration ] , phys . rev . d * 65 * , 032003 ( 2002 ) [ arxiv : hep - ex/0108043 ] .
in this work , we predict two charged charmonium - like enhancement structures close to the @xmath0 and @xmath1 thresholds , where the initial single pion emission mechanism is introduced in the hidden - charm dipion decays of higher charmonia @xmath2 , @xmath3 , @xmath4 and charmonium - like state @xmath5 . we suggest besiii to search for these structures in the @xmath6 , @xmath7 and @xmath8 invariant mass spectra of the @xmath2 decays into @xmath9 , @xmath10 and @xmath11 . in addition , the experimental search for these enhancement structures in the @xmath6 , @xmath7 and @xmath12 invariant mass spectra of the @xmath13 hidden - charm dipion decays will be accessible at belle and babar .
extrasolar planets are currently detected via indirect methods . the radial velocity technique has proved most successful in recent years . however , photometric measurements of planets transiting in front of their central star through the line - of - sight to earth have already provided more than 30 planet detections in the past years . many more transiting planet detections are expected in the very near future by the space missions corot @xcite and kepler @xcite . combining both methods allows the derivation of planetary radii , true masses and thus , the mean density of extrasolar planets . transiting planets are therefore of particular importance to investigate basic planetary parameters . the detection of extrasolar planets via the transit method requires long duration , high photometric quality lightcurves of a large number of stars . the major difficulties for ground - based transit searches are interruptions of the time series , e.g. due to the diurnal cycle or weather conditions , and the earth s atmosphere limiting the photometric signal - to - noise - ratio @xcite . located on the antarctic plateau at @xmath0 southern latitude , provides the potential for long , uninterrupted high - quality photometric time series observations from ground during its winter period . the maximum number of astronomical nights at in comparison to mauna kea has already been calculated by @xcite , including the influence of the sun , the moon and cloud coverage . while has less astronomical nighttime , the percentage of cloud free observing hours is similar to mauna kea . kenyon & storey do not discuss optimizations for long time series observations of target fields . their paper , however , presents a detailed discussion of the expected sky brightness at . they estimate that atmospheric scattering should be exceptionally low , and even the sky brightness due to the moon should be reduced . has reduced zodiacal light , no light pollution and only minor disturbances due to aurora . furthermore , scintillation noise as a major limiting factor of photometric quality is expected to be about 3.6 times lower at dome c compared to chilean sites for long integrations @xcite , which is particularily important for the search of small planets . may therefore be the best site on earth for stable high - accuracy photometry . alternatively , networks distributed favorably in longitude can also provide high duty cycles . a well known example is the global oscillation network group ( gong ) for asteroseismological studies which include six network sites ( four in the northern and two in the southern hemisphere ) . an estimate of the expected duty cycle for spectroscopic asteroseismology from in comparison to gong has been presented by @xcite . they conclude that provides a better or similar performance than gong for observing periods up to 100 days , attaining a duty cycle of up to 87% . the network performs better for observing periods of half a year or more , but with a duty cycle of only about 77% . in this paper , we include the factors already investigated by mosser & aristidi , like the sun and cloud coverage . in addition we include the moon and the varying extinction of the target fields during observations . furthermore , we discuss not only a comparison of to low - latitude networks , but also a combination of with other sites . a comparison between and the corot mission has already been presented @xcite . the resulting potentials for long - duration time - series photometric observations are then discussed in the context of maximizing the probability to detect extrasolar photometric transits . observations from the ground are limited for various reasons . the main restriction is imposed by the sun for observations in the visible spectral range . dark time can be reduced further by periods of full moon , because during such periods the sky background noise level reduces the photometric accuracy . selected target fields move across the sky , rising and setting and thus further limiting any observation campaign from a single site . after a brief discussion of the underlying time - frame in [ period ] , the actual determination of observation times from these astronomical constraints will be the subject of section [ geometricalconstraints ] . in practice , local weather might likewise limit observations , e.g. due to cloud coverage . section [ clouds ] describes the model used for simulating this effect . finally , the total observing time on a chosen target field as well as the photometric quality obtained crucially depend on its coordinates . in [ targetoptimization ] we therefore present a simple model designed to provide a quantitative solution to the target field selection process . the selected time - frame for our simulation limits possible observations to a certain period and therefore affects the results systematically . the most appropriate choice is to consider a one year period for each site studied , as the earth s orbit around the sun has the most important impact on observability . after one year has elapsed , time intervals of possible observations would basically show again the same pattern for any given target field . the results presented here correspond to one year ( @xmath1 ) , starting on 1 march 2008 , 0:00h ut . furthermore , only one field is considered per site to determine the maximum possible time coverage for planet detections . thus , planet detections at large orbital distance was given highest priority , rather than optimizing for detections of very short - period planets by maximizing the number of fields . the sun and moon affect the background noise level depending on their position relative to the observer . in order to decide whether a target is observable at a certain time or not , some criteria are needed . the most common and efficient way is to limit the three local altitudes @xmath2 , i.e. to apply @xmath3 for example @xmath4 calculates observation times without direct sun- or moonlight , as they would be restricted to stay always below the horizon . however , since scattered light might likewise disturb observations significantly , the common limit @xmath5 was used in this work , at which the contamination of the photometry is negligible . furthermore , target fields below @xmath6 altitude are not considered observable due to their poor photometric quality , i.e. @xmath7 . moreover , the moon s contribution to sky brightness is a function of its phase @xmath8 . therefore the moon criteria in eq . ( [ limit_h ] ) was relaxed by additionally allowing for situations corresponding to @xmath9 i.e. times when the moon altitude is above its limit , but the corresponding phase stays below a specified parameter @xmath10 . furthermore , the distance between moon and target field can be restricted to stay above a certain value , that is @xmath11 where @xmath12 denotes the angle between moon and target on their great circle . based on our observational experience , these last two limits were chosen to be @xmath13 and @xmath14 at which the influence on photometric accuracy starts to be significant . compared to low - latitude sites , the sky background is expected to be darker at dome c for the same local positions of sun , moon and target due to the low aerosol content of the air @xcite . nevertheless , we refrained from adjusting the limits as this effect is still to be examined more quantitatively . the average cloud pattern at a given site was simulated using a model by @xcite , which requires two input parameters , @xmath15 and @xmath16 , derived from weather statistics . @xmath15 stands for the average fraction of clear skies , whereas @xmath16 denotes the mean duration of cloudless periods . both can vary seasonally . according to the model , a simple pattern is created that consists of periods that either allow observations ( clear sky ) or not ( clouds ) for any given site and season . the duration of a _ single _ cloudless period is governed by a gamma distribution of the form @xmath17 so that @xmath18 denotes the probability to find an uninterrupted interval of length @xmath19 . in order to meet the given fraction @xmath15 of clear sky , the mean duration of cloudy periods is then clearly given by @xmath20 . substituting this value for @xmath16 in eq . ( [ clouds_p(t ) ] ) , one obtains the corresponding probability distribution of cloudy weather . an alternating sequence of clear and cloudy periods will therefore provide an approximation of real sky coverage . so , periods of potential observations ( calculated using the geometrical constraints ) are further reduced by only selecting clear sky times . as these are based upon random processes , 500 monte carlo runs are performed for each configuration in order to achieve statistically significant results . for the purpose of site comparisons , the calculation of observational time series is often derived from astronomical night times only , i.e. without including a specific target field . these results form an upper limit on the duty cycle , but are usually not achieved by any _ single _ target . for instance , most fields can only be observed for a fraction of a year due to the earth s annual motion around the sun . however , the outcome of transit search programmes largely relies on how much time a chosen field is being observed . in particular , only long time - series might provide a sufficient detection rate for planets with earth - like orbits . including target field coordinates into the simulation is therefore necessary for an accurate analysis of duty cycles . to select optimized target field coordinates , we posed the question : for a given site and time interval , which target field will provide the longest time coverage ? two parameters are relevant here : * observation time @xmath21 * mean airmass @xmath22 considering the observation time , fields observable over long periods are favoured . however , the photometric quality is reduced with increasing airmasses . therefore our calculations seek to minimize the mean airmass . for that , both parameters are calculated as a function of potential equatorial target coordinates @xmath23 throughout the whole sky for each site . in order to optimize them simultanously , they must subsequently be weighted against each other by defining a function @xmath24 . using such a weighting mechanism then provides a sky map from which we can easily derive a maximum value . although the choice of a specific function @xmath25 is somewhat arbitrary , it should in general show a reasonable dependency on its two variables @xmath21 and @xmath22 : firstly , @xmath25 should be monotonically decreasing in @xmath22 , as the photometric quality can only be reduced with larger airmasses . secondly , only a linear dependency on the observing time @xmath21 is meaningful . finally , we have already excluded target fields with airmasses @xmath26 by setting @xmath27 , thus it is consequent to demand @xmath28 for @xmath29 . these requirements are well fulfilled by the following approach : @xmath30 the parameters @xmath31 and @xmath32 have been chosen to reproduce the dependency of photometric accuracy on the airmass of observations , as illustrated in fig . [ fig : am_weighting ] : in the region @xmath33 we get @xmath34 , which accounts for the usually negligible variation of the signal - to - noise ratio of photometric data recorded at these airmasses . beyond that the photometric quality is significantly reduced , resulting in a strong devaluation of observing times with an average airmass close to 2 . this procedure yields an efficient choice of target coordinates for site networks . we then maximized the total time of observation from at least one site and decided to observe from whichever site provided the smallest airmass whenever simultaneous observations were possible . target coordinates found by this procedure obviously take into account astronomical configurations only . additional factors are required to optimize the field for an exoplanet transit search programme . in particular , the target field must contain a large number of dwarf stars . the required high number has to be weighted against crowding effects in dense star fields . therefore , this optimization step depends on the fov and the spatial scale of the instrument chosen . it has to be performed for a particular instrument and is therefore not within the scope of this paper which concentrates on the astronomical boundary conditions in terms of optimized duty cycle . as already mentioned , we aim to compare the window function obtained from antarctica with one or more low - latitude sites as well as a combination of with such locations . in addition to , three sites with similar low - latitudes are considered for a potential southern hemisphere network . the location of the best ii telescope ( erikson _ et al . _ , in prep . ) at cerro armazones was chosen in chile . in africa a site next to the gamsberg in namibia was included in our calculations , whereas the anglo - australian observatory ( aao ) was selected on the australian continent . the geographical coordinates are listed in table [ localcoordinates ] . .local coordinates of considered sites [ cols="<,<,<,<",options="header " , ] the two parameters used by the weather model , @xmath15 and @xmath16 , can exhibit strong seasonal and geographical variability . for the simulation of free - sky windows , monthly mean values are used for each site in table [ localcoordinates ] . figure [ fig : weather_compare ] shows the dependency of the assumed average clear sky fraction @xmath15 on time and location . for , winter observations from @xcite with mean values of 6.8 consecutive clear days and 84% overall clear skies were extrapolated to the whole year . in the case of chile , cloudiness data from the european southern observatory ( eso ) paranal were averaged for 1983 - 2006 , whereas publicly available _ surface synoptic observations _ ( synop ) from 2004 - 2007 were used for the aao . finally , an eso site survey provided values for @xmath15 at close - by gamsberg in namibia @xcite . for australia , the average duration of clear periods in the data set was found to be @xmath35 . for chile and namibia we assumed @xmath36 according to @xcite . however , the parameter @xmath16 only affects the duration of individual observations and is thus of minor importance . for example , using values @xmath37 for chile does neither change the total observation time @xmath21 nor the time coverage of planetary transits . we first validated the model against previous studies . calculated celestial coordinates agree with results from the publicly available _ _ jpl horizons _ _ ephemeris software to within 30 arcseconds for the years 1900 - 2100 . in addition to this , the results of @xcite for available dark time at and mauna kea were confirmed when we adopted their slightly differing limits for sun- and moonlight . then , optimal target fields have been simulated for all single sites as well as for three possible combinations of the sites . the various configurations are summarized in table [ tab : results ] . as a first step , we investigated the limitations imposed on time - series observations by each of the constraining factors considered individually . figure [ fig : effects ] shows calculated values of @xmath21 taking into account firstly only sunlight , then gradually including moonlight , target field visibility and finally cloud coverage . at a single site like best ii , the maximum dark time is 163 days per year due to the diurnal cycle ( @xmath38 , dark bars in fig . [ fig : effects ] ) . at , the dark time per year is somewhat less ( 135 days ) due to its high latitude . on the contrary , a low - latitude network can in principle observe all year round . however , effects like moonlight , target field selection and clouds further reduce the maximum available observing time . the target field selection does not impose a strong restriction on available observing time from , because many fields are above @xmath39 almost continuously during the polar night . at low latitude sites , the full available dark time can only be filled with observations of several target fields over the year . for one field it is , of course , significantly reduced ( compare third bars from left for each site in fig . [ fig : effects ] ) . in addition , there are also severe effects by clouds and the final performance is therefore significantly reduced in comparison to the ideal case ( fourth bar ) . due to the unequal impact of limitations upon different sites it is therefore necessary to consider all four effects simultaneously . lcccrrrcc dome c & @xmath40 & @xmath41 & without & @xmath42 & @xmath43 & @xmath44 & & 1.03 + & & & with & @xmath45 & @xmath46 & @xmath47 & & 1.03 + best ii & @xmath48 & @xmath49 & without & @xmath50 & @xmath51 & @xmath52 & & 1.33 + & & & with & @xmath53 & @xmath54 & @xmath55 & & 1.41 + network & @xmath56 & @xmath57 & without & @xmath58 & @xmath59 & @xmath60 & & 1.33 + & & & with & @xmath61 & @xmath62 & @xmath63 & & 1.40 + dome c + best ii & @xmath64 & @xmath65 & without & @xmath66 & @xmath67 & @xmath44 & & 1.33 + & & & with & @xmath68 & @xmath69 & @xmath70 & & 1.34 + dome c + network & @xmath71 & @xmath72 & without & @xmath73 & @xmath74 & @xmath44 & & 1.33 + & & & with & @xmath75 & @xmath76 & @xmath77 & & 1.34 + in table [ tab : results ] potential target coordinates @xmath23 with maximum @xmath78 are given for each combination of sites . also the corresponding values of total observation time @xmath21 and mean airmass @xmath22 are given with and without clouds , respectively . furthermore , the mean and maximum duration of single continuous periods of observations is shown as @xmath79 and @xmath80 . the window function itself is illustrated in fig . [ fig : domec ] for . the long observing periods during the polar night are clearly visible . however , continuous observations do not exceed 23 hours due to the sun rising above @xmath81 at noon even around winter solstice . nevertheless , the task of observing the _ same _ target from a single site as long as possible can best be fulfilled from antarctica . even when weather statistics are taken into account , clearly outcompetes even an excellent low - latitude site such as best ii . under realistic conditions , the mean total duration of observations of a target field with cloud coverage is 85 days from antarctica which is more than twice as long as from chile ( 38 days ) . in order to achieve a similar performance without , the combination of at least three low - latitude sites would be necessary . figure [ fig : network1 ] shows the performance of a network consisting of chile , namibia and australia . this would in principle allow for very long continuous observations taken in turns , peaking in mid - winter at a maximum period close to one month between full moons . but interruptions due to cloud coverage limit long runs and reduce the total observation time to 98 days . however , the network would still yield about two weeks more than in terms of observing duration . beyond the direct comparison of with one or more low - latitude sites , network configurations including antarctica deliver very interesting prospects . as is restricted to observations during the winter months only , low - latitude sites can also monitor targets located on the summer night sky around @xmath82 . with the additional choice of @xmath83 , such a field would always remain more than @xmath6 above the horizon at . the consequence is a significantly increased time statistics , as the overlap of observations from antarctica with the other sites would be marginal . figure [ fig : network2 ] illustrates the excellent duty cycle expected from a network consisting of all four considered sites including , whereas fig . [ fig : target_plot ] shows the functional dependency of the total observation time on field coordinates . combining only and best ii , it would be possible to observe for a total of 123 days per year under realistic conditions , thus exceeding the duty cycle of the three - site low - latitude network by 25% . finally , increasing the total obervation time further to 165 days by including namibia and australia into the network would provide an outstanding performance for an earth - bound project . in the following , the efficiency of transit detection is examined . figure [ fig : cov3_cld ] shows a comparison of observational coverage based on the simulated window functions . simulations have been performed as a function of the orbital period and for three or more transit events , which is usually the minimum requirement for detection . that is , a coverage of 100% means at least three events of a potential transiting planet will be observed for a given orbital distance . note that we only take into account the time sampling here , so the actual detection probabilities will be reduced further e.g. by the signal - to - noise ratio or geometric orientation . all curves in fig . [ fig : cov3_cld ] show peaks and dips caused by interruptions in the data sampling e.g. due to the diurnal cycle , full moon or cloud coverage . periodic data gaps lead to a reduced observational coverage for orbital periods which are multiple of interruption frequencies . can provide continuous observational coverage for planets with up to about two weeks orbital period , extending to 30 days with a probability of 50% . in contrast to that , it is very unlikely to detect planets with periods significantly longer than 15 days for a single low - latitude site like best ii in chile within one year . however , an almost equal performance to could be achieved by the combination of three low - latitude sites . combining a telescope at with one in chile extends the full observational coverage to about three weeks and leads to a significantly improved performance for larger periods compared to alone . finally , the combination of with a network including chile , namibia and australia could , in addition to a good coverage of small orbits , in principle still detect planets with periods of more than 60 days at a reasonable coverage . we have investigated the prospects for long - duration uninterrupted photometric time series from and other low - latitude sites . contrary to previous studies we _ simultaneously _ took into account the limitations of the sun and moon , clouds and target field extinction variations . due to the long winter , has an outstanding performance in observing a target field over a long time period from a single site . furthermore , the effective observing time period can be extended if is combined with other sites which can provide data during the summer period . for example , the combination of with a site in chile extends the observational coverage significantly . combining with a low - latitude network of sites at chile , namibia and australia provides prospects for observational duty cycles that can be improved upon only by observations from space . to predict quantitative transit detection rates , the photometric quality of must be investigated directly . however , for this we have to await the results of photometric site testing projects like the antarctica search for transiting extrasolar planets ( astep , @xcite ) for a detailed study . nevertheless , the extensively long observational time baseline for photometric variability studies clearly demonstrates the extremely encouraging prospects of this site in antarctica . baglin , a. , _ et al . _ 2006 , in : _ the corot mission : pre - launch status _ , esa sp-1306 , 33 - 38 borucki , w. j. , _ et al . _ 2003 , in : _ earths : darwin / tpf and the search for extrasolar terrestrial planets _ , esa sp-539 , 69 - 81 fressin , f. , _ et al . _ 2007 , eas pub . ser . , 25 , 225 - 232 hill , f. & newkirk , g . , jr . 1985 , , 95 , 201 kenyon , s.l . & storey , j.w.v . 2006 , , 118:489 - 502 kenyon , s.l . , _ et al . _ 2006 , , 118:924 - 932 mosser , b. & aristidi , e. 2007 , , 119 , 127 - 133 rauer , h. & erikson , a. 2007 , in : _ extrasolar planets _ , ed . r. dvorak , wiley - vch , 207 - 240 rauer , h. , fruth , t. & erikson , a. 2008 , _ proceedings of second arena conference _ , eas pub . ser . , submitted sarazin , m. 1994 , _ technical note on namibian gamsberg site _ , part i , eso - mpia 1994 - 1995 site survey
the detection of transiting extrasolar planets requires high - photometric quality and long - duration photometric stellar time - series . in this paper , we investigate the advantages provided by the antarctic observing platform for planet transit detections during its long winter period , which allows for relatively long , uninterrupted time - series . our calculations include limiting effects due to the sun and moon , cloud coverage and the effect of reduced photometric quality for high extinction of target fields . we compare the potential for long time - series from with a single site in chile , a three - site low - latitude network as well as combinations of with chile and the network , respectively . is one of the prime astronomical sites on earth for obtaining uninterrupted long - duration observations in terms of prospects for a high observational duty cycle . the duty cycle of a project can , however , be significantly improved by integrating into a network of sites .
the phenomenon of radicalization @xcite is of central interest in the context of criminality and terrorism . it is currently spreading all over the world including european countries . the recent unprecedented terrorists attacks in paris ( november 13 , 2015 ) and brussels ( march 22 , 2016 ) took life of respectively 130 and 32 persons with over 300 wounded in each case @xcite . it puts at a very high level the burden on making substantial progress in the mastering of the issue . over the years sociologists and social - psychologists have contributed a good deal of work to the phenomenon @xcite . however an understanding , which could lead to some practical curbing of radicalization is unfortunately still lacking as dramatically demonstrated by the recent series of terrorist attacks in france ( 2015 @xcite ) and in brussels ( 2016 @xcite ) . one promising direction is the prospect to access the huge amount of data ( big data ) which exists in the world wide web . it could open a valuable source of surveillance and forecasting to prevent some aspects of radicalization spreading . however , efficient data - mining tools are still to be constructed yet within the constraints related to the preservation of individual privacies . accordingly , under the current risk of loosing control of the situation any new attempt to tackle the issue of radicalization is valuable in itself . to identify some hints to implement novel adequate policies towards at least the hindrance of radicalization spreading is of particular importance . along this line it happens that the modern field of sociophysics @xcite where models inspired from physics are developed to describe a large spectrum of social behaviors , may contribute to the challenge . among others , sociophysics includes the study of opinion dynamics @xcite , language dynamics @xcite , crowd behavior @xcite , criminal activities @xcite , and cultural dynamics @xcite . our work , focusing on a formal modeling of radicalization ( see also @xcite ) from the viewpoint of opinion dynamics , subscribes to this trend @xcite . therefore , according to the analytical approaches developed in sociophysics the proposed model adopts some assumptions that allow to simplify the scenario of reference . the complexity underlying terrorism phenomena is thus reduced to a series of more simple local interactions monitored by two parameters , which tune the global dynamics of the system . the focus on local interactions to reach the global equilibrium state constitutes one major trend of statistical physics , i.e. , the branch of physics from which sociophysics developed . more specifically we consider a mixed population made up of two subpopulations , each one sharing a peculiar way of life . first one is a core population locally rooted in the country . in contrast , the other one is an immigrant ( two , three generations ) subpopulation whose way of life is rooted in another territory . differences between the two ways of life may be strong , numerous and contradictory . however , in case of a disagreement about some specific cultural habit like for instance wearing the islamic veil , both subpopulations do not stand at the same level of resilience . core agents consider that it is not up to them to modify their way of life or accept from newcomers behaviors perceived as contrary to their long time country rooted cultural habits . core agents behave here as inflexible agents . for them it is up to newcomers including immigrants even at second or third generation to adjust to the country prevailing way of life . it is thus up to newcomers to either choose to live peacefully with the core population adjusting part of their habits to the local constraints or to maintain the integrality of their habits at a cost of creating conflicts with the core population . accordingly , the newcomers can be considered as sensitive agents . they can choose between two individual states either peaceful or opponent . sensitive agents are entitled to shift state from peaceful to opponent and vice - versa . in addition we make the assumption that being in an opponent state may lead the corresponding agent to take part or to support violent activities . in principle , the latter choice can be linked to the appearance of local terrorist groups . we are dealing with a mixture of inflexible and sensitive agents in given fixed proportions @xmath0 and @xmath1 with @xmath2 . however , @xmath1 is made up of two time dependent parts @xmath3 and @xmath4 , which are the respective proportions of peaceful and opponent sensitive agents . at any time @xmath5 @xmath6 . the time dependence is driven by an internal dynamics among sensitive agents . it is the result of pairwise interactions both among themselves between peaceful and opponent agents and with inflexible agents . an opponent may drive a peaceful agent to opponent and an inflexible may drive peaceful an opponent agent . the associated dynamics is studied using a lotka - volterra - like ordinary differential equation . given an initial tiny minority of opponents we investigate the role of their activeness @xcite in turning peaceful agents to opponents via pairwise interactions . the effectiveness of their activism is materialized in the degree of radicalization of the sensitive population against the core population . it creates a social basis for passive supporters @xcite to emerge in support to terrorists @xcite . in parallel , the mechanism behind the dynamics of radicalization enlightens by symmetry a potential role core inflexible agents could have in the launching of an eventual counter radicalization . by individual counter activeness core agents can contribute substantially to both curb the radicalization spreading and in certain conditions make it shrink down to an equilibrium state where inflexible , peaceful and opponent agents co - exist . the associated required minimum core individual involvement is calculated . it is found to be a function of both the majority or minority status of the sensitive subpopulation with respect to the core subpopulation and the degree of activeness of opponents . it is worth to stress that different mathematical frameworks could be used to describe our dynamics . for instance , approaches based on evolutionary game theory @xcite allow to perform both computational and analytical ( e.g. , @xcite ) investigations . it requires to define a payoff matrix and rules for local interactions to monitor the updating . in this work we use stochastic processes based on opinion dynamics @xcite . local interactions reduce to contact processes , which make updating rules to depend on the relative densities of the various agent states . the choice of the current approach in the modeling arises from the aim to evaluate to which extent the heterogeneity of a population in cultural and behavioral terms may lead to critical and complex social phenomena as radicalization . furthermore , it is important to emphasize that the attribute ` inflexible ' adopted to describe the core population stand , refers to cultural habits and traditions which allow to peacefully coexist with individuals coming from abroad provided they share the fundamental features of the local cultural frame . it happens that opinion dynamics constitutes one of the most investigated topics in sociophysics and in computational social science . for instance , its dynamics have been recently studied using the framework of multiplex networks @xcite considering different social behaviors @xcite . it allows to understand phenomena recorded in huge social network datasets @xcite . opinion dynamics allows to analyze and to model the spreading of ideas , opinions , and feelings by reducing the study of complex social scenarios to the analysis of few variables @xcite . even terrorism and criminal activities may be studied by the same approach , i.e. , reducing the related process to a problem of opinion dynamics . to conclude , our results may contribute to shed a new light on the instrumental role core agents could play to curb radicalization and establish a coexistence with the sensitive population . some hints at novel public policies towards social integration are obtained . in the last years several authors have worked on opinion dynamics models to analyze various underlying behaviors , which produce social phenomena , e.g. , group polarization , conformity and extremism . in this section , we briefly review some of these investigations , which are connected to our work along the topic of extreme social phenomena , especially radicalization . a computational model for tackling political party competitions is introduced in @xcite . the authors investigate different possible occurrences of fragmentation according to variations in the amount of important political issues and their current relevance . different interaction patterns among voters are considered using an analytical approach . the focus is on the role of extremism in opinion dynamics with a qualitative analysis of real scenarios . the complex social phenomenon of group polarization is described in @xcite in the context of politics . in particular , the authors propose a model based on probability theory to drive the emergence of group polarization . the emergence of risks is shown to be related to the group polarization in a wide range of scenarios related to political and economical issues ( e.g. , immigration , religion , welfare state , human rights ) . the results highlight the necessity to a better understanding of the emergence of extreme opinions . the connection between contradictory public opinions , heterogeneous beliefs and the emergence of extremism is analyzed in @xcite . an agent - based model considers a population with different socio - cultural classes to describe the process of opinion spreading with calculations performed on small groups of individuals ( e.g. , composed of @xmath7 and @xmath8 agents ) . the model constitutes a useful reference for defining models related to complex social phenomena . moreover , the related results suggest that the direction of the inherent polarization effect , which occurs in the formation of a public opinion driven by a democratic debate , is biased due to the existence of common beliefs within a population . opinion dynamics is also studied using computational approaches , e.g. , by agent - based models on continuous or discrete spaces . such approaches require a careful attention during the implementation phase . for instance , in the work @xcite authors focus on the role of activation regimes . more precisely they compare different asynchronous updating schemes ( e.g. , random and uniform ) . the activation regime refers to the order or scheme adopted to let agents express their opinion . as a result , the activation regime is found to affect opinion dynamics processes in some cases ( i.e. , @xcite ) . it is therefore of importance to clearly state which activation regime is selected to implement a dynamical model . the role of conformity in the q - voter model by arranging agents on heterogeneous networks has been also investigated @xcite . the authors showed that different steady states may be reached by tuning the ratio of conformists versus that of nonconformists in an agent population , which evolves according to the dynamics of the q - voter model . in our model we do not consider complex topologies . however , the influence that may arise from different interaction patterns may constitute the topic of future investigations . in order to study the emergence of radicalization in an heterogeneous population we consider a system with @xmath9 interacting agents distributed among inflexible ( @xmath10 ) , peaceful ( @xmath11 ) and opponent ( @xmath12 ) agents . each category refers to a different behavior or feeling . inflexible and opponent agents have behaviors mapped respectively to states @xmath13 . peaceful agents have a behavior mapped to the state @xmath14 . inflexible agents never change state ( see also @xcite ) while peaceful and opponent agents may shift state from one to another over time . opponents may become peaceful and peaceful may become opponents . hence , neither peaceful nor opponent agents may assume the state of inflexible agents . inflexible agents interact with sensitive agents both peaceful and opponents . during these pairwise interactions when an inflexible agent meets an opponent it may well turn the opponent to peaceful via different paths . among those paths most are spontaneous through normal social and friendship practices . but as it will appear latter , exchanges could become intentional as to promote coexistence with sensitive agents via monitored informal exchanges . to account for all interacting pairs a parameter @xmath15 is introduced to represent on average the rate per unit of time of encounters where opponents become peaceful agents . in parallel and in contrast we introduce the parameter @xmath16 to account on average for the rate of success of opponents in convincing peaceful agents to turn opponents . contrary to inflexible agents opponents are acting intentionally to increase the support to their radical view within the sensitive population . the value of @xmath16 is a function of the power of conviction of opponents . it also takes into account the activeness of opponent agents since opponents are activists . it is not the case of the core inflexible agents who interact spontaneously with sensitive agents without an a priori goal . it is worth to stress that both @xmath15 and @xmath16 may in principle vary over time . however , the corresponding time scale for variation is expected to be much longer than the time scale of the dynamics driven by pairwise interactions . this is why at the present stage of our work @xmath15 and @xmath16 are assumed to be fixed and constant . analyzing their time dependence , which might be of great interest to get further insights on equilibriums among people belonging to different cultures is left for future work . we emphasize that our analytical approach entails to consider the system as if it was continuous , i.e. , analyzing the relative densities of agents in the various states . a similar approach is usually followed in other contexts as epidemic dynamics @xcite . a compartmental approach to the studying of epidemics entails to analyze the spreading of a disease by modifying the state of agents . for instance , the sis @xcite model considers a two - state population where agents may vary their state from @xmath17 ( i.e. , susceptible ) to @xmath10 ( i.e. , infected ) and vice - versa over time . considering the probability to get infected or to heal the dynamics can be studied analytically defining odes as if the underlying system were continuous . going to the analytical details of our model we defined the following system of equations @xmath18 where @xmath0 is the constant density of inflexible agents , while @xmath4 and @xmath3 are the respective densities of peaceful and opponent agents at time @xmath5 . dealing with densities the third equation of system [ eq : evolution ] allows to reduce the number of odes to one equation . in particular , choosing the peaceful agents density @xmath3 we get @xmath19 the equilibrium state of the population can be obtained from @xmath20 , which reads @xmath21 the two solutions of equation [ eq : evolution_equilibrium ] read @xmath22^{2 } - 4 \beta \alpha \sigma_i ( 1 - \sigma_i)}}{2 \beta}\ ] ] where @xmath23 is the equilibrium value of peaceful agents . those values simplify to @xmath24 which implies the two associated equilibrium opponent values @xmath25 indeed equation [ eq : evolution_reduced ] can be solved analytically to yield @xmath26 fig 1 shows the evolution of the system on varying the initial conditions . we analyze the respective stability ranges for @xmath27 and @xmath28 : @xmath29 where @xmath30 and @xmath31 , we obtain @xmath32 + 2\beta \sigma_p\ ] ] therefore , for respective values @xmath33 we obtain @xmath34 stability being achieved for @xmath35 , equation [ eq : lambda_value_solutions ] shows that @xmath36 is stable when @xmath37 . accordingly we get two stable regimes : @xmath38 with @xmath39 . these two regimes yield the respective equilibrium values for peaceful and opponent agents as from [ eq : equilibrium_solution_value ] and [ eq : equilibrium_solution_value_op ] @xmath40 the first equation of system [ eq : final_stability ] highlights that in some conditions the amount of opponent agents is equal to zero . hence , we perform a further investigation to study under which conditions it is possible to avoid the phenomenon of radicalization ( i.e. , by reaching the equilibrium state @xmath41 ) . in terms of opinion dynamics these results indicate that under appropriate conditions it is possible to remove one opinion from the system . given the relevance of this outcome in the related context , i.e. , criminal activities and terrorism , we explore in more details this result . from the above results radicalization can be totally thwarted if @xmath42 . accordingly , given @xmath0 and @xmath16 the individual involvement for the inflexible population in striking up with individual opponents must be at least at a level @xmath43 therefore , as seen from equation [ eq : alpha ] the larger @xmath0 the less effort is required from the inflexible population . however , the more active are the opponents ( i.e. , larger @xmath16 ) the more involvement is required . to visualize the multiplicative factor by which @xmath15 must overpass @xmath16 it is worth to draw the curve @xmath44 as a function of @xmath0 as shown in fig 2 . from equation [ eq : alpha ] it is seen that to prevent radicalization inflexible agents s involvement must be either lower ( @xmath45 ) or larger ( @xmath46 ) than that of opponents depending on the magnitude of the multiplicative factor @xmath44 . when @xmath47 , i.e. , @xmath48 core agents do not need to much individual engagement as could be expected in the case of a coexistence of a core majority population with a sensitive minority subpopulation . more precisely , the engagement depends on the opponent activism but the core population benefiting from its majority status . in this case its requirement is always lower than the opponent involvement . however , the situation turns difficult when the initial sensitive minority turns to a majority status as it occurred in some specific urban areas . in that case to avoid a radicalization requires a very high individual engagement from the core agents , which may be rather hard to implement . in particular since no collective information is available about the situation . we thus have three different cases : * 1 ) * @xmath48 , * 2 ) * @xmath49 , and * 3 ) * @xmath50 to consider to determine the respective level of individual core involvement to avoid the phenomenon of radicalization . + * case 1 . * for @xmath51 core agents need little involvement to thwart totally the radicalization of the sensitive subpopulation with values of @xmath15 much lower than @xmath16 . indeed , opponent agents need to produce very high values of @xmath16 ( compared to @xmath15 ) to survive , precisely the condition @xmath52 must be satisfied . however , very large values of @xmath16 can shrink to zero the amount of peaceful agents yielding a fully radicalized sensitive population , which although in a small minority status may produce substantial violence against inflexible agents . + * case 2 . * for @xmath53 the opponent activism must be counter with an equal core counter activism since @xmath54 makes opponent agents to extinct . instead , for @xmath45 peaceful and opponent agents coexist and the former disappear for large values of @xmath16 with again a fully radicalized sensitive population with @xmath55 . + * case 3 . * for values @xmath50 , if @xmath56 the equilibrium condition entails that @xmath57 ( and @xmath58 ) . if @xmath46 , we can reach the extinction of opponent agents as @xmath59 . in contrast when @xmath45 opponent agents strongly prevail in the population . + in order to asses the degree of radicalization in a population we can introduce two parameters : @xmath60 and @xmath61 . the former is defined to evaluate the fraction of opponent agents among flexible agents while the latter ( i.e. , @xmath61 ) evaluates the ratio between opponent and inflexible agents . therefore , @xmath60 represents the relative ratio of opponents among flexible agents and @xmath61 gives a measure about the real power of opponents agents in a population . an high value of @xmath60 ( i.e. , close to @xmath62 ) in a population with @xmath63 indicates that strategies to fight radicalization are too weak but at the same time opponents are few . therefore , in this case governments should take an action even if the situation seems still under control . on the other hand , a low value of @xmath60 ( i.e. , close to 0 ) together with a high value of @xmath61 represent an alarming situation . indeed , even if there are only a few opponents among flexible agents their amount is bigger than that of inflexible ones @xcite . to evaluate these measures , @xmath60 and @xmath61 have been defined as follows @xmath64 hence , recalling that @xmath65 and having solved analytically @xmath3 ( see [ eq : solution_neutral ] ) we are able to compute values of both parameters @xmath60 and @xmath61 at equilibrium and on varying the initial conditions see fig 3 . it is worth to note that the parameter @xmath60 as defined in [ eq : radicalization ] has a range in @xmath66 $ ] . at equilibrium @xmath67 means that there are no opponent agents in the population while @xmath68 means that all flexible agents became opponents . on the other hand , the parameter @xmath61 has potentially an unlimited range from @xmath69 to @xmath70 ( in the case @xmath0 is very close to @xmath69 and @xmath71 to @xmath62 ) . to conclude , we want to emphasize the meaningful role of the two parameters @xmath60 and @xmath61 . they represent a way to quantify in which extent radicalization phenomena are taking place in a population . moreover , in more general terms we envision a further utilization of these parameters in opinion dynamics since they clearly indicate the prevalence of one opinion / state over another one . the recent anti - western terrorist attacks @xcite in europe have brought the question of radicalization at a top priority of policy maker agenda of the different european governments . in particular , most of the terrorists involved in the various killings which took place in several european capitals were either national citizens or legal residents . this very fact points to the direct link existing between terrorism and radicalization @xcite . indeed , various institutions are faced with the difficult issue to implement innovative procedures to stop if not eradicate radicalization . the task turns out to be rather hard since radicalization has been prospering quietly in different areas of european countries for now many years without any substantial barrier . it has been a sensitive political issue and most officials had preferred the laissez - faire instead of addressing the problem in solid terms . the dramatic scores of @xmath72 paris and @xmath73 brussels attacks have now prompted the necessity to face the problem and start implementing counter measures . the burden is on european governments to find ways to tackle the radicalization . almost everyone is expecting action from the states . but the states seem to have no solid scheme to apply . one direction has been along the education side with the setting up of so - called de - radicalization programs . however , such an approach concerns identified radicals who have been arrested . all efforts and thoughts are focused on acting on radicalized citizens . coercive measures are implemented against known associations and active leaders . even to contain radicalization appears to be a challenging task . our model , although rather simple puts light on the process by which the phenomenon of radicalization spreads over within a sensitive population . it articulates around the capacity of radicalized agents to turn radical otherwise peaceful agents who had chosen to coexist with the native population sharing their habitat . this capacity is embedded in the coefficient @xmath16 . in addition , the main novelty of our model is to account for the possible capacity of native agents to overturn radical agents in making them choose the peaceful state quantified with the coefficient @xmath15 . moreover , the ratio of native versus sensitive populations ( @xmath74 ) was found to be a critical parameter . in the past this ratio was rather stable over time with slow evolution . it made feasible to evaluate the activeness of radicals , which has been not meaningful for decades . however , rather quick changes may occur in the demography of the sensitive populations especially with the substantial increase of recent years immigration . on this basis , our results show how the overall situation can be totally put upside down ( @xmath75 ) with respect to the extent of current radicalization while not much seems to have happened with respect to radical activities . keeping the same level of activeness from radicals , a slow change in the population ratio may produce a sudden spreading of radicalization . therefore the knowledge of the evolution of the current ratio of populations is a key parameter to evaluate the associated potential of radicalization spreading . and yet , in many countries like france , ethnic statistics are forbidden . most of curbing radicalization still involves the state and diverse official institutions . in contrast the phenomenon of radicalization results from informal interactions among sensitive and radical agents . at this point our results unveiled a new and unexpected promising path to fight radicalization . an innovative strategy could be implemented by launching a citizen counter radicalization movement mapped from the path used by opponents to spread radicalization within the sensitive population . instead of being the sole prerogative of national authorities de - radicalization would become a citizen matter . the same way a radical tries to turn a peaceful sensitive agent to hostility towards the natives , natives can try to bring back opponents to peaceful coexistence . normal citizens would have to engage in personal interactions with sensitive agents to establish a solid ground for coexistence . the required degree of efficient citizen involvement can be clearly identified using the degree of activeness of the radicals and the ratio of subpopulations , native versus sensitive . while this ratio is at the exclusive hand of national authorities , the citizen involvement is a citizen prerogative . in addition , the centrality of the ratio of subpopulations within a given territory emphasizes the importance of avoiding a de - mixing of the subpopulations . in case of a different discriminating distribution of the subpopulations within distinct sub - territories , radicalization would be enhance at once with the same proportion of radical due the large value of @xmath74 within the sub - territory where the sensitive population is mostly confined . it is worth to notice that the proposed model may in principle be applied also to criminal and terrorist scenarios in homogeneous populations as it occurred in the cases of italian red brigades and french revolution . these two cases are concerned with homogeneous populations as both inflexible , peaceful and opponent agents belong to the local core population . in the former case ( i.e. , red brigades ) inflexible agents represent individuals who respect laws and believe in institutions and governments . individuals having a different behavior can fall in the mild category of peaceful agents or in the extreme category of opponents ( i.e. , criminals ) . instead , in the case of the french revolution inflexible agents represent the small proportion of french nobility . the remaining part of the population is represented by peaceful and opponent agents . there , the extremely difficult life conditions fed opponent ideals and the wide proportion of the sensible subpopulation became completely opponent giving rise to a revolution . to summarize we have identified the equilibrium state of a mixed population in terms of order or disorder phases . we have also identified the ratio between social strategies and the strength of opponents ideal . since we refer to the concept of social strategies it is worth to emphasize that although the considered scenario can be modeled in various ways as those based on evolutionary games . there the concept of `` strategy '' acquires a particular meaning . here we develop a model based on opinion dynamics processes . as a result social strategies are embodied in a parameter while updating rules depend on the density of different opinions in the population . moreover , in this context opinions refer to the different cultural extractions and behaviors that can be observed in an heterogeneous population . we remark that today the question of de - radicalization has became a key priority issue of internal security in european countries . yet the challenge is intact with no ready to use solution . different state agencies are launching a series of experimental treatments but all are concerned with institutional managing of the issue . given the acute current terrorist threat people are expecting and requesting policy makers to take initiative to curb the current phenomenon of radicalization within sensitive local populations . unlike this heavy policy trend our study has enlightened the crucial role so called `` normal citizens '' could play to stop the spreading of radicalism . it could even shrink it back with a serious perspective to eventually eradicate the actual growing threat set in european cities . from our results it appears that an efficient action should not be limited to state involvement but also to call on individual voluntary engagement within their respective neighborhoods towards the sensitive individuals . given an evaluation of radical activeness within some sensitive neighborhoods we were able to calculate the required degree of `` normal citizen '' counter - activeness to curb radicalization . this degree of engagement was also found to depend on the ration of native to sensitive populations . focusing on local interactions in the modeling of these dynamics underlines the instrumental role neighborhood compositions can have in the shaping of the social behavior of the corresponding subpopulations . today it often happens that people of different cultures are fostered to coexist together in the same district of a city occupying each a series of connected blocks . the connection with natives is thus drastically reduced jeopardizing opportunities of real integration even after a few generations . in addition , within some delimited urban areas the majority group is no longer the native subpopulation . at the same time it is still majority in a close by other area . accordingly given the same tiny proportion of opponents with the same degree of activeness in the two neighborhoods , one ends up highly radicalized while the other stays very peaceful . local interactions and the degree of mixing are key factors to undermine the spreading of radicalization . a free and uncontrolled ( by authorities ) settling of people often leads to a geographical concentration of sensitive subpopulations . as a result this process may spontaneously develop a natural ground for the emergence of hate towards native individuals . people are thus lead towards the strengthening of the initial culture differences , which results in the establishment of social distances with native individuals despite being physically very close to them . to conclude , we want to highlight that our work creates a first step to envision new policies to support campaigns promoting the daily life sharing among people from different cultural backgrounds . in particular , we focus on methods that potentially may lead `` radical neighbors '' to the choice of coexistence , i.e. , renouncing to fight against the native population . at least we hope our results will trigger more research along this path of individuals engaging to establish a peaceful coexistence with sensitive agents . at this stage further studies along this direction are required , in particular from a computational social science perspective . it should be possible to identify earlier traces ( i.e. , big data ) and seeds of radical behavior in social networks . suitable tools to quantify their strength are also required . last but not least , we would like to stress that although we have been mentioning criminal activities we are not judging neither the motivations nor the ideal of opponent agents . indeed , they can be considered negative ( as in the case of current anti - western terrorism ) or positive ( as today in the case of the french revolution ) depending both on the side taken and the chosen epoch . our aim was to study the conditions of emergence or vanishing of radicalization as a social phenomenon independently of a moral judgment . maj would like to thank fondazione banco di sardegna for supporting his work . this work was supported in part by a convention dga-2012 60 0013 00470 75 01 .
the phenomenon of radicalization is investigated within a mixed population composed of core and sensitive subpopulations . the latest includes first to third generation immigrants . respective ways of life may be partially incompatible . in case of a conflict core agents behave as inflexible about the issue . in contrast , sensitive agents can decide either to live peacefully adjusting their way of life to the core one , or to oppose it with eventually joining violent activities . the interplay dynamics between peaceful and opponent sensitive agents is driven by pairwise interactions . these interactions occur both within the sensitive population and by mixing with core agents . the update process is monitored using a lotka - volterra - like ordinary differential equation . given an initial tiny minority of opponents that coexist with both inflexible and peaceful agents , we investigate implications on the emergence of radicalization . opponents try to turn peaceful agents to opponents driving radicalization . however , inflexible core agents may step in to bring back opponents to a peaceful choice thus weakening the phenomenon . the required minimum individual core involvement to actually curb radicalization is calculated.it is found to be a function of both the majority or minority status of the sensitive subpopulation with respect to the core subpopulation and the degree of activeness of opponents . the results highlight the instrumental role core agents can have to hinder radicalization within the sensitive subpopulation . some hints are outlined to favor novel public policies towards social integration .
miniboone @xcite is a neutrino oscillation experiment at fermilab , designed to confirm or rule out the hypothesis that the lsnd @xmath0 excess @xcite is due to @xmath1 oscillations . a general description of the experiment can be found elsewhere @xcite . heretofore , miniboone has been taking data in neutrino mode , searching for @xmath2 oscillations . however , in some scenarios involving cp and cpt violation , oscillations may occur only in antineutrinos . thus , searching for oscillations in antineutrinos is a crucial test @xcite . furthermore , the search for cp violation in the neutrino sector by future off - axis experiments @xcite requires @xmath1 oscillation measurements , as well as @xmath2 . the signature for cp violation is an asymmetry in these oscillation probabilities , but this can only be confirmed if the precision of the @xmath3 and @xmath4 cross sections are smaller than the observed asymmetry . there are few @xmath3 cross section data published @xcite to date , but even fewer measurements of low energy @xmath4 cross sections . we will need more and better data if we hope to find cp violation in the neutrino sector . .__event rates expected in miniboone @xmath4 running with @xmath5 pot assuming a 550 cm fiducial volume , before cuts . listed are the expected right - sign ( rs ) and wrong - sign ( ws ) events for each reaction channel . _ _ [ cols="^,^,^",options="header " , ] in order to use this technique to constrain the ws background , the angular resolution for muons must be sufficiently good to separate the two populations . miniboone s angular resolution is measured using the cosmic muon calibration system , which consists of a muon tracker hodoscope placed above the detector . figure [ fig : mutr_delmt ] shows the 4@xmath6 angular resolution of muons between 400 and 500 mev . this resolution is sufficient to allow exploitation of the angular differences in the ccqe outgoing muons distributions . the angular distributions can be fitted to extract the wrong - sign contribution . analysis of our monte carlo data sets indicates that the wrong - sign content can be measured using this technique , with a statistical uncertainty of 5% of itself @xcite . this is illustrated in figure [ fig : muang_recon ] , which shows the reconstructed angle of the out - going lepton with respect to the neutrino beam axis for both @xmath7 and @xmath8 cc qe events generated with the miniboone monte carlo , after all selection cuts have been applied . including systematic uncertainties and ( non - qe ) backgrounds increases the uncertainty on the ws extraction only to @xmath9 . a second constraint results from measuring the rate at which muons decay in the miniboone detector . due to an @xmath10 @xmath11 capture probability in mineral oil , negatively and positively charged muons exhibit different effective lifetimes ( @xmath12 for @xmath11 @xcite and @xmath13 for @xmath14 @xcite ) . to extract the ws fraction , one simply fits the muon lifetime with a sum of two exponentials , with the extracted fraction of each exponential term giving the fraction of rs or ws events . for ccqe events , we find that the wrong - sign contribution can be extracted with a @xmath15 statistical uncertainty based solely on this lifetime difference and negligible systematic uncertainties . while not as precise as fits to the muon angular distributions , this particular constraint is unique , as it is independent of kinematics . our third wrong - sign constraint employs the the fact that antineutrinos do not create cc@xmath16 events in the detector these all stem from neutrinos ( table [ table : nubar - event - stats ] ) . miniboone identifies cc@xmath16 events by tagging the two decay electrons that follow the primary neutrino interaction , one from the @xmath11 and one from the @xmath17 decay @xcite . however , cc@xmath18 events do not pass this requirement because all the emitted @xmath19 s which stop in the detector descend into atomic orbits around carbon nuclei and are instantly captured , leaving no decay electrons @xcite . the @xmath19 decay in flight rate is much smaller than the rate of @xmath17 decays at rest @xcite . thus , applying the two michel cut to the full sample , which is @xmath20 antineutrino ( rs ) interactions , yields an @xmath21 pure sample of ws neutrino events . assuming conservative uncertainties for the antineutrino background events and the cc@xmath16 cross section , which is currently being measured by miniboone , we expect a @xmath22 uncertainty on the wrong - sign content in the beam given @xmath5 pot . this constraint is complementary to the muon angular distributions , because cc@xmath16 events stem mainly from resonance decays , thus constraining the wrong - sign content at larger neutrino energies . miniboone expects more than 40,000 qe interactions in antineutrino mode with @xmath5 pot before cuts . using the same qe event selection criteria as the previously reported miniboone neutrino analysis @xcite yields a sample of @xmath23 events , with @xmath24 qe purity with both ws and rs events . assuming the above wrong - sign constraints and conservative errors on the @xmath3 flux , backgrounds , and event detection , we expect a miniboone measurement of the @xmath8 qe cross section to better than @xmath25 with @xmath5 pot . @xmath8 neutral current ( nc ) @xmath26 production is one of the largest backgrounds to future @xmath1 oscillation searches . there has been only one published measurement of the absolute rate of this channel , with @xmath27 uncertainty at 2 gev @xcite . applying miniboone s @xmath7 nc @xmath26 cuts @xcite with no modifications leaves a sample of antineutrino nc @xmath26 events with a similar event purity and efficiency . after this selection , we expect 1,650 @xmath8 resonant nc @xmath26 events and 1,640 @xmath8 coherent nc @xmath26 events assuming @xmath5 pot and the rein and sehgal model of coherent pion production @xcite . coherent pion production has a characteristic pion angular distribution that allows it to be distinguished from resonant production , as illustrated in figures [ fig : ncpi0-nu ] and [ fig : ncpi0-nubar ] . moreover , the previously mentioned figures illustrate that the distinctiveness of the angular distributions should be even more marked in antineutrino running , which increases the value of antineutrino data for understanding coherent production . the background of @xmath28 1000 ws events will be determined by the constraints on the wrong - sign content in the beam as described in section [ sec : ws - constraints ] and the measurement of the @xmath7 nc @xmath26 cross section from miniboone neutrino data . miniboone expects roughly 7,000 resonant cc @xmath18 with @xmath5 pot before cuts . as discussed above , almost all of the emitted @xmath19 s will be absorbed by carbon nuclei , and will therefore not be selected by the cc1@xmath17 cuts . nevertheless , these events still have a signature : two cherenkov rings ( one each from the @xmath14 and @xmath19 ) and one michel electron in the vicinity of the @xmath11 . the selection efficiency and purity of such events is unknown at this time . further investigation is currently underway . although miniboone has been searching for oscillations in neutrino mode , the lsnd oscillation signal was actually an excess of antineutrino events . because of the potential for finding cp violation in the neutrino sector , it is imperative that miniboone test the lsnd oscillation hypothesis with antineutrino data @xcite . figure [ fig : nuebar - app ] shows an estimate of the miniboone sensitivity to @xmath1 oscillations under the assumption that @xmath2 oscillations do not occur , i.e. assuming miniboone sees no @xmath29 appearance signal . here , we compare the sensitivity to the joint karmen - lsnd region ( @xmath4 only ) @xcite , not the full lsnd allowed region ( @xmath30 ) . this is a statistics - limited search ; further running is needed beyond the @xmath5 pot assumed in the preceding sections in order to test the lsnd hypothesis . this sensitivity shown assumes @xmath31 pot . we have developed three techniques for determining the wrong - sign background in antineutrino mode . the resulting systematic error on any given @xmath4 cross section measurement due to the wrong sign contamination should be less than 2% averaged over the entire flux , which is remarkable for a detector which does not possess event - by - event sign selection . given this redundant approach , the wrong - sign contamination should not be considered prohibitive to producing meaningful antineutrino cross section @xcite and oscillation measurements @xcite at miniboone . these techniques may also be useful for other experiments without magnetized detectors which have plans to study antineutrino interactions ( _ e.g. _ t2k , no@xmath3a , super - k ) . miniboone began running in antineutrino mode on 19 january , 2006 , and is currently approved to run for one more year . in order to truly confirm or rule out the lsnd oscillation hypothesis , miniboone needs 6@xmath32 pot , which will require additional years of running . the author is pleased to acknowledge the collaborative efforts of j.m . link , h.a . tanaka , and g.p . zeller in developing the ideas in this work . the author would also like to express gratitude to the organizers of nuint05 for their generous travel support . the miniboone collaboration gratefully acknowledges support from various grants and contracts from the department of energy and the national science foundation . the author was supported by grant number de - fg02 - 91er0617 from the department of energy . tanaka , `` estimating ws content in negative polarity horn data '' , boone memo ( 2004 ) . for more information on the boone memos and technical notes , contact the miniboone spokespeople , j. m. conrad or w. c. louis .
miniboone began running in antineutrino mode on 19 january , 2006 . we describe the sensitivity of miniboone to lsnd - like @xmath0 oscillations and outline a program of antineutrino cross - section measurements necessary for the next generation of neutrino oscillation experiments . we describe three independent methods of constraining wrong - sign ( neutrino ) backgrounds in an antineutrino beam , and their application to the miniboone antineutrino analyses .
one can find that there exist many reasons and motivations to consider alternative theories of gravity to the general relativity . some theories are motivated by the modifications in infrared regime and they mainly aim to resolve a question about the dark energy . for instance , @xmath0 gravity @xcite can explain the accelerating expansion of current universe without the cosmological constant and avoid the hierarchy problem . others are motivated by the modifications in ultraviolet regime and they are often associated with the effects of quantum gravity @xcite . higher - curvature theories typified by curvature - squared and the gauss - bonnet terms are induced from quantum corrections . naturally , challenges to the theory beyond general relativity themselves are important because there is no fundamental reason to choose the einstein - hilbert action or einstein s equation over many kind of alternatives . recently , much attention has been paid to bi - metric theory or what we call bigravity , which includes two independent metric tensor fields , @xmath1 and @xmath2 @xcite . bigravity contains massive spin-@xmath3 propagating mode in addition to ordinary massless spin-@xmath3 mode corresponding to the graviton . this theory has been successfully constructed as the generalization of drgt massive gravity in recent years @xcite . some people expect that the new degrees of freedom introduced by another metric can solve remaining problems in cosmology , that is , dark energy @xcite and dark matter @xcite problems . interactions between two metric tensors produce effective cosmological constant , furthermore , the massive spin-@xmath3 fields and matter fields coupled with the metric @xmath2 can be candidates of dark matter . when we intend to view the bigravity to be an alternative theory of gravity , it is also interesting that we apply this theory to other phenomena in cosmology or astrophysics and find the differences from general relativity . in our work , we focus on the nature of black holes . it is well known that the horizon radius of the black hole in the vacuum usually decreases by the hawking radiation . however , bousso and hawking have observed a phenomenon where the black hole radius increases by the quantum correction for the specific nariai black hole @xcite . this phenomenon is called anti - evaporation of black holes . note that the schwarzschild - de sitter black holes and the nariai black holes can be primordial ones , thus , they are not expected to appear at the final stage of star collapse . in order that the anti - evaporation could occur in the einstein gravity , we need to include the quantum correction from the matters . it might be remarkable , however , that the anti - evaporation may occur even on the classical level in @xmath0 gravity theories @xcite . this could be because that the equations for the @xmath0 gravity are more complicated than those for the einstein gravity , but at present , it is not so clear what could be essential for the anti - evaporation on the classical level . then it might be interesting if the anti - evaporation in the classical level might be a general phenomena in the modified gravity . in this paper , we consider the possibility of the anti - evaporation in bigravity on the classical level because the contribution from the interaction between two metric tensors is not so trivial . we will give a classical analysis in the stability of the nariai black hole in the bigravity and study if the anti - evaporation could occur in the classical level , which may clarify what could be necessary for the anti - evaporation to occur . this paper is organized as follows : first , we explain about the nariai black hole . the nariai black hole is defined as a subset of the schwarzschild - de sitter black hole where the radii of the cosmological and black hole horizons are degenerate . second , we give a brief review about the bigravity . in order to show the existence of the nariai black hole as an exact solution in the bigravity , we specify some parameters and solutions , then , we give a proof that the asymptotically de sitter solutions can be realized . finally , we consider the perturbations around the black hole and evaluate their stability and investigate if the anti - evaporation could occur on the classical level . at first , we introduce the nariai space - time as a family of the schwarzschild - de sitter space - time . the schwarzschild - de sitter solution is expressed in the following form : @xmath4 where the function @xmath5 is defined by @xmath6 here , @xmath7 is a mass parameter and @xmath8 is a positive cosmological constant . for @xmath9 , @xmath5 has two positive roots @xmath10 and @xmath11 , corresponding to the cosmological and black hole horizon , respectively . in the limit @xmath12 , the radius of the black hole horizon coincide with that of the cosmological horizon . here , the coordinate system in eq . ( [ polar coordinate ] ) becomes inappropriate because @xmath13 between the two horizons . then it is useful to introduce new coordinate system as follows : @xmath14 where @xmath15 is the parameter defined as @xmath16 , and @xmath17 corresponds to the degeneracy of two horizons . in above coordinate , the black hole horizon corresponds to @xmath18 and the cosmological horizon corresponds to @xmath19 , and the metric takes the following form : @xmath20 in the degenerate case , @xmath21 , the metric is given by @xmath22 and this space - time is called the nariai black hole . note that the topology of the space - like sections of the schwarzschild - de sitter space - time ( and the nariai space - time ) is @xmath23 while that of the ordinary black hole solution is @xmath24 in four dimensions . in this coordinate system , the radius of two - sphere , @xmath25 , varies along the one - sphere coordinate , @xmath26 ; the minimal two - sphere corresponds to the black hole horizon and the maximal one corresponds to the cosmological horizon . in this section , we give a brief review of the anti - evaporation in general relativity . first of all , we begin with the hawking radiation from the black holes . it is well known that there is radiation by the quantum effects of matter fields around the black hole horizon , which is called the hawking radiation . this quantum corrections leads to the trace anomaly of the energy - momentum tensor although the trace of the energy - momentum tensor should classically vanish , @xmath27 . when we consider the massless scalar as the hawking radiation , the effective action corresponding to the trace anomaly is written by a covariant form @xcite , @xmath28\ , .\end{aligned}\ ] ] here , the effective action is reduced to two dimensional form , and @xmath29 is the redundancy parameter corresponding to the renormalization scheme . the above effective action leads to the modification for the equation of motion . in general relativity , specific perturbations around the nariai black holes shrink from its initial values . then , the size of black hole horizon increases at least initially . this phenomena is called anti - evaporation . note that if we do not include the quantum correction , the anti - evaporation does not occur and the horizon size remains in that of the initial perturbation . we should note that in case of the @xmath0 gravity , there occurs the anti - evaporation even on the classical level , that is , without quantum correction @xcite . in the @xmath0 gravity , the equations are complicated , which may generate the the anti - evaporation on the classical level . because the equations in the bigravity are also pretty complicated , we may expect that the anti - evaporation could occur on the classical level , and therefore , it could be interesting to investigate the anti - evaporation in the bigravity even if on the classical level . in this section , we give a brief review of the bigravity and show that the nariai space - time is an exact solution in this theory . the action of the bigravity @xcite is given by @xmath30 here , @xmath31 and @xmath32 are dynamical variables and rank - two tensor fields which have properties as metrics , @xmath33 and @xmath34 are the ricci scalars for @xmath1 and @xmath2 , respectively , @xmath35 and @xmath36 are the two planck mass scales for @xmath1 and @xmath2 as well , and the scale @xmath37 is the effective planck mass scale defined by @xmath38 the quantities @xmath39s and @xmath40 are free parameters , and the former defines the form of interactions and the latter expresses the mass of the massive spin-@xmath3 field . the matrix @xmath41 is defined by the square root of @xmath42 , that is , @xmath43 for general matrix @xmath44 , @xmath45s are polynomials of the eigenvalues of @xmath46 : @xmath47\ , , \nonumber \\ e_{2}(\mathbf{x } ) = & \frac{1}{2 } \left ( [ \mathbf{x}]^{2 } - [ \mathbf{x}^{2 } ] \right)\ , , \nonumber \\ e_{3}(\mathbf{x } ) = & \frac{1}{6 } \left ( [ \mathbf{x}]^{3 } - 3[\mathbf{x}][\mathbf{x}^{2 } ] + 2[\mathbf{x}^{3 } ] \right)\ , , \nonumber \\ e_{4}(\mathbf{x } ) = & \frac{1}{24 } \left ( [ \mathbf{x}]^{4 } - 6[\mathbf{x}]^{2}[\mathbf{x}^{2 } ] + 3[\mathbf{x}^{2}]^{2 } \right . \nonumber \\ & \left . \qquad + 8[\mathbf{x}][\mathbf{x}^{3 } ] - 6[\mathbf{x}^{4 } ] \right ) \nonumber \\ = & \mathrm{det}(\mathbf{x})\ , , \nonumber \\ e_{k}(\mathbf{x } ) = & 0 \quad \mbox{for } \ \ k>4 \ , , \label{e_n}\end{aligned}\ ] ] where the square brackets denote the traces of the matrices , that is , @xmath48=x^{\mu}_{\mu}$ ] . for conventional notation , we explicitly denote the determinant of matrix @xmath49 as @xmath50 , and @xmath51 represents a matrix which is the square root of @xmath49 . now we consider the variation of the action ( [ bigravity action ] ) with respect to @xmath1 . the equation of motion for @xmath1 is given by @xmath52 here , for a matrix @xmath44 , @xmath53s are defined by @xmath54 or explicitly , @xmath55 \ , , \nonumber \\ y_{2}(\mathbf{x } ) = & \mathbf{x}^{2 } - \mathbf{x}[\mathbf{x } ] + \frac{1}{2 } \mathbf{1 } \left ( [ \mathbf{x}]^{2 } - [ \mathbf{x}^{2 } ] \right ) \ , , \nonumber \\ y_{3}(\mathbf{x } ) = & \mathbf{x}^{3 } - \mathbf{x}^{2}[\mathbf{x } ] + \frac{1}{2 } \mathbf{x } \left ( [ \mathbf{x}]^{2 } - [ \mathbf{x}^{2 } ] \right ) \nonumber \\ & - \frac{1}{6 } \mathbf{1 } \left ( [ \mathbf{x}]^{3 } - 3[\mathbf{x}][\mathbf{x}^{2 } ] + 2[\mathbf{x}^{3 } ] \right)\ , . \label{interaction_terms}\end{aligned}\ ] ] we also obtain the equation of motion for @xmath2 , @xmath56 in this case , we do not consider the energy - momentum tensor for the ordinary matter fields . the constraints for the conservation law appear if the minimal couplings to the matter are introduced , and we find @xmath57 \ , , \\ 0=&\nabla^{\mu}_{(f ) } \left [ \sum^{3}_{n=0}(-1)^{n}\beta_{4-n } \left \ { f_{\mu \lambda}y^{\lambda}_{(n ) \nu}(\sqrt{f^{-1}g } ) \right . \nonumber \\ & \hphantom { \nabla^{\mu}_{(f ) } [ \sum^{3}_{n=0}(-1)^{n}\beta_{4-n } \ { f } \left . + f_{\nu \lambda}y^{\lambda}_{(n ) \mu}(\sqrt{f^{-1}g } ) \right \ } \right ] \ , .\end{aligned}\ ] ] here , @xmath58 and @xmath59 are covariant derivatives which are defined in terms of @xmath1 and @xmath2 , respectively . in order to discuss the anti - evaporation in the bigravity , we need to confirm that the asymptotically de - sitter solutions are realized in this theory . however , it is not so easy to investigate this problem for all combinations of the included parameters , thus , we impose specific assumptions to make discussion simpler . one of the authors considered a particular class of solutions where the two metric tensors are proportional to each other @xcite , @xmath60 this proportional relation leads to the einstein s equation with a cosmological constant because @xmath41 and @xmath61 turn to be proportional to unity . the two equations of motion are given by @xmath62 note that the dynamics of two metric tensors @xmath1 and @xmath2 are separated from each other and the constraints derived from the preservation of energy - momentum tensor are automatically satisfied . and , we have not assumed any symmetries in the space - time , thus , we can impose the spherical symmetry to the solutions later . furthermore , we consider a specific parametrization for the interacting parameter @xmath39s , @xmath63 this combination of two parameters , @xmath64 and @xmath65 , are required by the existence of the solution corresponding to the flat space - time in massive gravity , which is often used in bigravity . with a assumption @xmath66 , it has been shown that the de sitter solution can be realized in some parameter regions ( fig.[fig ] ) . and @xmath65 are given @xcite . we obtain only flat solutions in the region `` only trivial solution '' , although asymptotically non - flat ( de sitter and/or anti - de sitter ) solutions are realized in `` non - trivial solutions '' . ] for instance , the minimal model @xmath67 has only asymptotically flat solutions although the next - to minimal models @xmath68 have asymptotically de sitter solutions . therefore , the schwarzschild - de sitter black hole solutions are realized in the bigravity , and we can obtain the nariai black hole solution by the limit @xmath12 . note that it is recently implied that the de sitter solution is an attractor with homothetic relation , which may support us to consider our setting . under the assumption that @xmath69 , the dynamics of two metric tensors @xmath1 and @xmath2 are separated from each other , described by einstein s equations here , this property is just for the background solution and perturbations can be independent ; degrees of freedom of bigravity does not descend to that of general relativity . when we consider the perturbations from a background space - time , the interaction terms give non - trivial contribution to the evolution compared with the case of general relativity . therefore , it is important to analyze the stability of perturbation even on the classical level , and we need to investigate if the anti - evaporation could be realized on the classical level . we now consider the perturbation from the nariai black hole . according to the topology , @xmath23 , we make spherically symmetric metric ansatz as follows , @xmath70 here , two - dimensional metric , corresponding to @xmath71 and @xmath72 components , is written in the conformal gauge and @xmath72 is the coordinate on the one - sphere and has the period of @xmath73 . we should also note that the black hole and cosmological horizons are located at same place @xcite , respectively . under the above ansatz , we can calculate each components of the einstein tensor @xmath74 , @xmath75 here , @xmath76 and @xmath77 , and sub indices are omitted for simplicity . furthermore , we need to calculate the interaction terms in eqs . ( [ geq ] ) and ( [ feq ] ) . the interaction terms in each equations of motion are written in terms of @xmath41 and @xmath61 . defining @xmath78 and @xmath79 for convention , these two matrices are expressed as follows : @xmath80 where we define @xmath81 . after short calculation , we obtain @xmath82s which take the following forms : @xmath83 @xmath84 next , we consider the equations of motion . when we choose @xmath85 , the effective planck mass scale is given by @xmath86 now we do not restrict the combinations of parameters in the interaction terms but consider general case , that is , the case that there are independent five parameters @xmath39s . two equations of motion for @xmath1 and @xmath2 take the following form : @xmath87 g_{\mu \lambda}\ , , \\ 0 = & r_{\mu \nu}(f ) - \frac{1}{2}r(f)f_{\mu \nu } \nonumber \\ & + \frac{1}{2 } m_{0}^{2 } \left [ \beta_{4 } y^{\lambda}_{(0 ) \nu}(\mathbf{b } ) - \beta_{3 } y^{\lambda}_{(1 ) \nu}(\mathbf{b } ) \right . \nonumber \\ & \hphantom{+ \frac{1}{2 } m_{0}^{2 } [ \beta_{4 } } \left . + \beta_{2 } y^{\lambda}_{(2 ) \nu}(\mathbf{b } ) - \beta_{1 } y^{\lambda}_{(3 ) \nu}(\mathbf{b } ) \right ] f_{\mu \lambda}\ , .\end{aligned}\ ] ] as we have discussed , one can show that we obtain asymptotically de sitter solution for the specific combinations of parameters under the ansatz ( [ proportional relation ] ) . furthermore , when we impose the spherical symmetry on the solutions , the schwarzschild - de sitter space - time can be solutions in our setting . so , we will consider the condition to obtain the nariai space - time as a background solution . in the coordinate system of eqs .([gansatz ] ) and ( [ fansatz ] ) , the nariai solutions are expressed as follows : @xmath88 therefore , the corresponding @xmath89s and @xmath90s in the nariai solutions are given by @xmath91 that is , @xmath92 now , we substitute these solutions into the equations of motion . substituting ( [ nariai components ] ) into ( [ einstein tensors ] ) , ( [ interaction gf ] ) , and ( [ interaction fg ] ) , the einstein tensors and the interaction terms take the following forms : @xmath93 @xmath94 @xmath95 and , we find that two equations of motion are given by @xmath96 \ , , \\ 0 = & 1 - \frac{c^{2}}{2\lambda } m_{0}^{2 } \left [ \beta_{4 } + 3\beta_{3}c^{-1 } + 3\beta_{2 } c^{-2 } + \beta_{1 } c^{-3 } \right]\ , . \end{aligned}\ ] ] here , one can identify the two cosmological constants as follows : @xmath97\ , , \label{lambdag}\\ & \lambda_{f}(c ) = \frac{1}{2 } m_{0}^{2 } \left [ \beta_{4 } + 3\beta_{3}c^{-1 } + 3\beta_{2 } c^{-2 } + \beta_{1 } c^{-3 } \right ] \label{lambdaf}\ , .\end{aligned}\ ] ] then we obtain the quartic equation of @xmath98 , @xmath99 by using the two equations of motion . when one choose the interacting parameters @xmath39s , eq . ( [ ceq ] ) are determined and the consistent @xmath98 can be given as a solution which reproduces the positive cosmological constants in eqs . ( [ lambdag ] ) and ( [ lambdaf ] ) . therefore , in order to obtain the nariai solutions , all we have to do is to find the suitable interacting parameters . in the following discussion , we assume that the @xmath39s are chosen to realize the asymptotically de sitter space - time . next , we define the perturbations as follows : @xmath100 here , @xmath101s and @xmath102s correspond to the unperturbed nariai space - time and @xmath103s and @xmath104s are the perturbations . by substituting the above expressions into ( [ gansatz ] ) and ( [ fansatz ] ) , we find the metric perturbations of @xmath1 and @xmath2 in the first order , @xmath105 @xmath106 where , we define @xmath107 @xmath108 we now evaluate the equations of motion for the perturbation . at first , we calculate the perturbation of the einstein tensor in the first order . when we substitute the metric perturbations into ( [ einstein tensors ] ) , we obtain the deviations of @xmath109 , @xmath110 then , we define the deviations of the einstein tensor from the nariai space - time . note that the deviations of the einstein tensor for @xmath2 are obtained by changing @xmath111 and @xmath112 , because @xmath113 and @xmath114 is constant . compared with eq . ( [ nariai einstein tensors ] ) , we find the deviations of the einstein tensor are given by @xmath115 next , we evaluate the interaction terms . we define the deviation of @xmath116 and @xmath117 as follows : @xmath118 then , we can calculate the deviations of the interaction terms @xmath82s from the nariai space - time , and they are given by @xmath119 @xmath120 where we define the tensor @xmath121 as follows , @xmath122 finally , we consider the equations for the perturbations . for the convention , we express the equations of motion as follows : @xmath123 where @xmath124s are the sum of @xmath82s . when we consider the perturbation up to first order , above equations are divided by background part and deviation part , and the equations for the deviation take the following forms : @xmath125 here , we define @xmath126 \mathbf{1 } \nonumber \\ = & \lambda \mathbf{1}\ , , \\ i(\mathbf{b } ) = & \frac{1}{2 } m_{0}^{2 } \left [ \beta_{4 } + 3\beta_{3}c^{-1 } + 3\beta_{2 } c^{-2 } + \beta_{1 } c^{-3 } \right ] \mathbf{1 } \nonumber \\ = & \frac{\lambda}{c^{2 } } \mathbf{1 } \ , , \\ \delta i(\mathbf{a } ) = & - \frac{1}{2 } m_{0}^{2 } \left [ \beta_{1}c + 2\beta_{2}c^{2 } + \beta_{3}c^{3 } \right ] \mathbf{z } \nonumber \\ = & - c_{1 } \mathbf{z } \ , , \label{nariai int a } \\ \delta i(\mathbf{b } ) = & \frac{1}{2 } m_{0}^{2 } \left [ \beta_{3}c^{-1 } + 2\beta_{2 } c^{-2 } + \beta_{1 } c^{-3 } \right ] \mathbf{z } \nonumber \\ = & c^{-4}c_{1}\mathbf{z}\ , , \\ c_{1 } \equiv & \frac{1}{2 } m_{0}^{2 } \left [ \beta_{4 } + 3\beta_{3}c^{-1 } + 3\beta_{2 } c^{-2 } + \beta_{1 } c^{-3 } \right ] \ , . \label{nariai int b}\end{aligned}\ ] ] in order to describe the evolution of black holes due to the perturbations , we need to know where the horizons are located for @xmath1 and @xmath2 . in the following , we consider the black hole horizon for @xmath1 at first . let us specify the form of perturbations so that the two - sphere radius @xmath127 varies along the one - sphere coordinate @xmath72 : @xmath128 note that above form of perturbation is consistent with eq . ( [ gperturb ] ) . substituting the above form of perturbation into the @xmath129 component of ( [ gperturb ] ) , we obtain @xmath130 with the boundary condition , @xmath131 at @xmath132 , the solution is @xmath133 for this solution , we consider the time evolution of the black hole horizon . the condition for a horizon is @xmath134 , which is required that the gradient of the two - sphere size is null . here , eq . ( [ gperturb2 ] ) yields @xmath135 from above conditions , locations of the black hole horizon @xmath136 and cosmological horizon @xmath137 are defined as follows : @xmath138 therefore , the size of the black hole horizon , @xmath11 , is given by @xmath139 where the we define the perturbation for the horizon @xmath140 , @xmath141 then , substituting eq . ( [ classical sigma ] ) into eq . ( [ horizon ] ) , we obtain @xmath142 this means that no anti - evaporation takes place and horizon size remains that of the initial perturbation . this is just a static schwarzschild - de sitter black holes of nearly maximal mass . note that , if we define the same form of perturbation for @xmath143 as that for @xmath144 , we obtain same results because the equations have same form as that of @xmath145 . then , one can find that anti - evaporation does not occur for two metrics @xmath1 and @xmath2 on classical level . in the previous section , we found that the anti - evaporation is not realized in bigravity on classical level , which is not changed from the result in the general relativity . in this section , we focus on the problem how we can identify the difference between the case in general relativity and in bigravity . when we substitute the perturbations into the @xmath146 and @xmath147 components of ( [ gperturb ] ) , we obtain @xmath148 thus , deviations of the interaction terms ( [ nariai int a ] ) , ( [ nariai int b ] ) vanish if @xmath149 or @xmath150 . when we define the perturbation for @xmath2 as @xmath151 @xmath152 vanishes in the case where the amplitude of the perturbations are identical , @xmath153 . this means that the two sets of metric perturbations are proportional to each other and the relation between the perturbations is not changed from the background , @xmath154 . in this case , whole metrics including the perturbations are proportional and it does not lead to difference from general relativity . therefore , we can not distinguish bigravity theory from general relativity . note that , regarding the perturbation , one can introduce the different forms between @xmath144 and @xmath143 . for instance , we may assume the following form , @xmath155 which is consistent with the equations eqs . ( [ gperturb ] ) and ( [ fperturb ] ) . in this case , the amplitude and phase can take independent values and these are difference from the general relativity . we have studied the possibility of the anti - evaporation on classical level in the bigravity . for the assumption @xmath156 and particular parameters @xmath39s and the planck mass scales @xmath66 , we obtained the asymptotically de sitter space - time . when we considered the perturbations around the nariai space - time , the size of black hole horizon does not increase . and we have found that the anti - evaporation does not take place on the classical level although the equations of motion are different from general relativity . when we assume the perturbation ( [ gperturb2 ] ) , eq . ( [ classical sigma ] ) is derived from the @xmath129 component of the eq . ( [ gperturb ] ) . however , the non - diagonal components of the eq . ( [ gperturb ] ) take the forms identical with that in general relativity because the interaction terms do not modify the non - diagonal components . therefore , the result that the size of black hole horizon does not increase is not changed from that in the general relativity . in the @xmath0 gravity , however , the anti - evaporation can occur because the equations of motion are modified in different way . in order to realize the anti - evaporation , we need to take the quantum corrections into account in a similar way to general relativity . the effective action due to the quantum correction of matter fields is given in the same manner in general relativity . interesting problem is probably to study if we need to introduce the quantum corrections for only one metric or both metrics . for instance , when one regards @xmath1 as the metric which describes our world and finds that the anti - evaporation occurs with introducing the quantum corrections only to @xmath2 sector , the black holes grow although our world is exactly classical . there is another way to realize the anti - evaporation by modification to @xmath0 bigravity theory @xcite . this theory modify the kinetic terms of bigravity , from the ricci scalar to the function of it . in @xmath0 bigravity , we find similar problem to introducing the quantum corrections . that is , we need to study if the modification is required for only one metric or both metrics . the authors are deeply indebted to sergei . d. odintsov for constructive advices and discussion in the early stage of this research . t.k is partially supported by the nagoya university program for leading graduate schools funded by the ministry of education of the japanese government under the program number n01 . this work is also supported by the jsps grant - in - aid for scientific research ( s ) # 22224003 and ( c ) # 23540296 ( s.n . ) . the connections and curvature in the conformal gauge are given as follows : @xmath157 @xmath158 @xmath159 here , @xmath76 and @xmath77 . the equations for the perturbations are given by as follows : * @xmath146 component of ( [ gperturb ] ) @xmath160 * @xmath129 component of ( [ gperturb ] ) @xmath161 * @xmath147 component of ( [ gperturb ] ) @xmath162 * @xmath163 , @xmath164 component of ( [ gperturb ] ) @xmath165 * @xmath146 component of ( [ fperturb ] ) @xmath166 * @xmath129 component of ( [ fperturb ] ) @xmath167 * @xmath147 component of ( [ fperturb ] ) @xmath168 * @xmath163 , @xmath164 component of ( [ fperturb ] ) @xmath169 s. nojiri and s. d. odintsov , phys . d * 68 * ( 2003 ) 123512 [ hep - th/0307288 ] . s. nojiri and s. d. odintsov , phys . d * 74 * ( 2006 ) 086005 [ hep - th/0608008 ] . g. t hooft and m. j. g. veltman , annales poincare phys . theor . a * 20 * ( 1974 ) 69 . j. f. donoghue , phys . d * 50 * ( 1994 ) 3874 [ gr - qc/9405057 ] . s. f. hassan and r. a. rosen , phys . * 108 * ( 2012 ) 041101 [ arxiv:1106.3344 [ hep - th ] ] . s. f. hassan and r. a. rosen , jhep * 1107 * ( 2011 ) 009 [ arxiv:1103.6055 [ hep - th ] ] . s. f. hassan and r. a. rosen , jhep * 1202 * ( 2012 ) 126 [ arxiv:1109.3515 [ hep - th ] ] . t. damour , i. i. kogan and a. papazoglou , phys . d * 66 * ( 2002 ) 104025 [ hep - th/0206044 ] . m. s. volkov , jhep * 1201 * ( 2012 ) 035 [ arxiv:1110.6153 [ hep - th ] ] . m. s. volkov , phys . d * 86 * ( 2012 ) 104022 [ arxiv:1207.3723 [ hep - th ] ] . m. von strauss , a. schmidt - may , j. enander , e. mortsell , s. f. hassan and , jcap * 1203 * ( 2012 ) 042 [ arxiv:1111.1655 [ gr - qc ] ] . y. akrami , t.s . koivisto , m. sandstad , jhep * 03 * ( 2013 ) 099 [ arxiv:1209.0457[astro - ph.co ] ] . m. berg , i. buchberger , j. enander , e. mortsell and s. sjors , jcap * 1212 * ( 2012 ) 021 [ arxiv:1206.3496 [ gr - qc ] ] . m. s. volkov , class . grav . * 30 * ( 2013 ) 184009 [ arxiv:1304.0238 [ hep - th ] ] . y. akrami , t. s. koivisto , d. f. mota and m. sandstad , jcap * 1310 * ( 2013 ) 046 [ arxiv:1306.0004 [ hep - th ] ] . n. tamanini , e. n. saridakis and t. s. koivisto , jcap * 1402 * ( 2014 ) 015 [ arxiv:1307.5984 [ hep - th ] ] . m. fasiello and a. j. tolley , jcap * 1312 * ( 2013 ) 002 [ arxiv:1308.1647 [ hep - th ] ] . a. r. solomon , y. akrami , t. s. koivisto [ arxiv:1404.4061[astro - ph.co ] ] . f. knnig , y. akrami , l. amendola , m. motta , a. r. solomon [ arxiv:1407.4331[astro - ph.co ] ] . n. rossi , eur . j. st * 163 * ( 2008 ) 291 [ arxiv:0902.0072 [ astro-ph.co ] ] . m. banados , a. gomberoff , d. c. rodrigues and c. skordis , phys . d * 79 * ( 2009 ) 063515 [ arxiv:0811.1270 [ gr - qc ] ] . k. aoki and k. i. maeda , phys . d * 89 * ( 2014 ) 064051 [ arxiv:1312.7040 [ gr - qc ] ] . k. aoki and k. i. maeda , arxiv:1409.0202 [ gr - qc ] . s. nojiri and s. d. odintsov , class . * 30 * ( 2013 ) 125003 [ arxiv:1301.2775 [ hep - th ] ] . l. sebastiani , d. momeni , r. myrzakulov and s. d. odintsov , phys . d * 88 * ( 2013 ) 10 , 104022 [ arxiv:1305.4231 [ gr - qc ] ] . s. nojiri and s. d. odintsov , phys . b * 735 * ( 2014 ) 376 [ arxiv:1405.2439 [ gr - qc ] ] . t. katsuragawa , phys . d * 89 * ( 2014 ) 124007 [ arxiv:1312.1550 [ hep - th ] ] . s. nojiri , s. d. odintsov and n. shirai , jcap * 1305 * ( 2013 ) 020 [ arxiv:1212.2079 [ hep - th ] ] . s. nojiri and s. d. odintsov , phys . b * 716 * ( 2012 ) 377 [ arxiv:1207.5106 [ hep - th ] ] . j. kluso , s. nojiri and s. d. odintsov , phys . lett . b * 726 * ( 2013 ) 918 [ arxiv:1309.2185 [ hep - th ] ] .
we study the stability under the perturbation and the related anti - evaporation of the nariai space - time in bigravity . if we impose specific condition for the solutions and parameters , we obtain asymptotically de sitter space - time , and show the existence of the nariai space - time as a background solution . considering the perturbation around the nariai space - time up to first order , we investigate the behavior of black hole horizon . we show that the anti - evaporation does not occur on the classical level in the bigravity .
continuing disagreement as to whether massive compact halo objects ( machos ) have been detected by microlensing experiments looking towards the magellanic clouds highlights the need for other microlensing targets ( e.g. , kerins 2001 ) . the andromeda galaxy ( m31 ) presents an opportune target in this respect . the disk of m31 is highly inclined ( @xmath2 ) , with the consequence that lines of sight to disk stars in the north - west or near side of m31 are shorter than those to the south - east or far side . microlensing by a spheroidal dark halo will have a characteristic signature with an excess of events on the far side of the m31 disk ( crotts 1992 ; baillon et al.1993 ) . this asymmetric signal is absent for variable stars or stellar microlenses in the disk of m31 . a number of groups ( e.g. , aurire et al . 2001 ; riffeser et al . 2001 ; calchi - novati et al . 2002 ; crotts et al . 2001 ) are now carrying out large - scale surveys of m31 to search for this near - far disk asymmetry . this is a mammoth task as the individual stars in m31 are not resolved , so that new techniques based on the super - pixel method ( ansari et al . 1997 ) or difference imaging ( e.g. , crotts & tomaney 1996 ) have been exploited to measure the flux changes on unresolved stars . nonetheless , convincing candidate events are now being discovered , for example by the point - agape collaboration ( e.g. , aurire et al . 2001 ; paulin - henriksson et al . 2002 , 2003 ) . therefore , this is a timely moment to consider what factors affect the near - far disk asymmetry and how many events are likely to be needed for a convincing detection . the aim of this paper is to estimate the size of candidate event samples needed to detect asymmetry for different halo models . in sections 2 and 3 of the paper we show how the magnitude of the asymmetry signal is affected by the velocity anisotropy and the flattening of the m31 baryonic dark halo respectively . there are few ways known to us for measuring the properties of the orbits of dark objects in the halo of any galaxy or for ascertaining the flattening of the baryonic dark component of the halo . hence , any clues gleaned from pixel lensing experiments will be invaluable . in section 4 , we present some simple , non - parametric statistical estimators of asymmetry and calculate how many candidates are needed to give convincing detections . here , we investigate the microlensing properties of haloes in which the velocity distribution of the machos is anisotropic . we use models in which the halo density @xmath3 is isothermal and the rotation curve is flat : @xmath4 for these models , the velocity dispersions in the spherical polar coordinate system are given by ( see e.g. , white 1981 ; evans , hfner & de zeeuw 1997 ) @xmath5 where @xmath6 is the anisotropy parameter . if @xmath7 , then the velocity distribution is referred to as ` radially anisotropic ' ; if @xmath8 , then it is ` tangentially anisotropic . ' whilst @xmath9 in equation ( [ vel ] ) diverges as @xmath10 , velocity dispersion ratios are rarely observed to be more extreme than 3:1 , implying @xmath11 . we assume a circular velocity of @xmath12 for m31 s halo ( emerson 1976 ) and we compute the cutoff radius to give a total halo mass of @xmath13 ( e.g. , evans & wilkinson 2000 ) . the sources are drawn from the m31 disk which is adequately modelled as a sheet inclined at 77 to the line of sight . the source velocity is assumed to be dominated by the disk rotation speed of 235 @xmath14 . for each halo model , we calculate a theoretical estimate of the pixel - lensing rate @xmath15 . unlike the classical ( resolved star ) microlensing rate , @xmath15 depends additionally on the surface brightness of the m31 disk and the luminosity function of the m31 sources ( c.f . , kerins et al . the calculations here are performed for a @xmath16-band luminosity function and surface brightness distribution . the surface brightness of m31 is tabulated in walterbos & kennicutt ( 1987 ) . the m31 disk luminosity function is assumed to be the same as for the milky way and we use the data of wielen , jahrei & kruger ( 1983 ) to characterise the faint end ( @xmath17 ) and that of bahcall & soneira ( 1980 ) for the bright end . for each source star of flux @xmath18 at sky position ( @xmath19,@xmath20 ) on the m31 disk , we compute the maximum impact parameter , @xmath21 , needed to ensure that the magnified source star noticeably enhances the local flux contribution from the background galaxy and sky background . motivated by the point - agape experiment ( e.g. , paulin - henriksson et al . 2002 , 2003 ) , we assume that detection is performed on a `` super - pixel '' array of pixels of size @xmath22 which typically encloses 40% of the total flux from a point source . in addition to the background due to the m31 surface brightness we allow for a sky background of 19.5 @xmath23 . we assume an event is detectable if the flux change caused by microlensing is @xmath24 of the background flux on the super - pixel . the pixel lensing rate is then @xmath25 where @xmath26 is the classical microlensing rate ( paczyski 1986 ; kiraga & paczyski 1994 ) and @xmath27 is the maximum impact parameter averaged over the luminosity function @xmath28 of the source stars . explicitly , we can write that @xmath29 this is an upper limit to the observed pixel lensing rate for any real experiment , as it does not take into account the effects of sampling , changing observing conditions or event identification algorithms . these effects may alter the spatial distribution of events , but they can be corrected via the calculated detection efficiency . in fact , the efficiency is largely controlled by the local surface brightness and so is approximately symmetric with respect to the major axis . the ratio of the number of far - disk to near - disk events therefore does not depend on the efficiency to lowest order . finally , the spatially - averaged pixel - lensing rate @xmath30 is obtained by weighting the rate with the m31 disk surface brightness . the central portions of the m31 disk are omitted , partly because stellar lenses in the m31 bulge dominate here and partly because the halo model is singular at the centre . so , a central region of 5 radius is excised from the m31 disk before performing the spatial averaging separately for events above and below the m31 major axis . none of the current experiments is surveying the entire m31 disk , but their fields do span the large majority of the minor axis and we therefore expect any underlying asymmetry in their fields to be representative of the globally - averaged asymmetry computed here . figure [ fig : pix ] shows the spatially - averaged theoretical pixel - lensing rate , @xmath30 , for lenses of mass @xmath31 , normalised to the value @xmath32 events per star per year . when @xmath7 , the velocity distribution is radially anisotropic and the rate @xmath33 . when @xmath34 , the velocity distribution is tangentially anisotropic and the rate @xmath35 , so the total rate in radially anisotropic models is somewhat higher . however , we are primarily interested in the differences between the near and far disk . such asymmetries may manifest themselves in the numbers , locations and time - scales of the events . accordingly , figure [ fig : pix ] also shows the far - to - near disk ratios for the pixel - lensing rates ( @xmath36 ) , the mean einstein crossing times ( @xmath37 ) and the ratio of mean projected distances of events to the major axis ( @xmath38 ) . all models show a strong excess of far - disk events , with @xmath39 increasing from 5.3 to 7 as the models go from tangential to radial anisotropy . it is also evident from @xmath38 in figure [ fig : pix ] that far - disk events lie systematically farther from the major axis than near - disk events , providing a second signature of asymmetry . however @xmath38 is nearly constant across the range of @xmath40 so this spatial signature does not provide a probe of the degree of velocity anisotropy . there is also an asymmetry in the einstein crossing times of m31 machos . if machos have radially distended orbits , then their motion tends towards being parallel to the line of sight for the near disk but orthogonal to it for the far disk . events on the near side therefore last longer . in other words , the ratio of the typical time - scales of far - disk to near - disk events decreases with increasing radial anisotropy . consequently , the ratio of the numbers of events in the far - disk to the near - disk is enhanced in radially anisotropic models compared to isotropic model . this can be seen from the fact that the spatially - averaged microlensing optical depth , @xmath41 , is independent of the velocity distribution . from figure [ fig : pix ] , the expected enhancement in the number asymmetry @xmath39 is @xmath130% . for the isotropic model ( @xmath42 ) , the time - scales are typically shorter in the far disk than in the near disk . the reason for this is that the typical separation between lens and source is larger for near - disk machos . for far - disk events , the typical lens - source separation is biased towards the location where the density peaks along the line of sight , which is at a distance @xmath43 in front of the sources , where @xmath20 is the projected distance along the minor axis . the situation is different for near - disk machos , where the line of sight density is always a monotonically decreasing function of lens - source separation . here , the typical separation is @xmath12030 kpc for @xmath44 , so @xmath37 is less than unity for an isotropic model , as shown in figure [ fig : pix ] . unfortunately , the einstein crossing time is not generally measurable for pixel - lensing events . instead experiments measure the full - width half - maximum ( @xmath45 ) of the lightcurve , which is additionally correlated with the source luminosity and background surface brightness distribution . the resulting @xmath45 distributions are therefore predicted to be broad and any asymmetries in the time - scale distribution are unlikely to be easily observable . self - consistent solutions of the self - gravitation equations for the density , potential and velocity distributions of flattened halo models are rare . even solving the jeans equations for the second velocity moments can lead to cumbersome results . baltz , gyuk & crotts ( 2003 ) have computed pixel - lensing rates explicitly for simple analytic flattened halo models . however , if we are merely interested in comparing the effects of flattening on the ratios of quantities in the near- and far - disk , then there is a quick alternative to carrying out computations with a fully axisymmetric halo model . figure [ fig : schematic ] shows lines of sight passing through an elliptical halo with axis ratio @xmath46 and eccentricity @xmath47 . the lines strike the disk of m31 at an angle @xmath48 . the ratios of the optical depths in the near and far disk are proportional to the line segments np and fp . also shown is an equivalent spherical halo . by extending n vertically to n and f to f , we can construct lines of sight that pass through the spherical halo and strike the disk at a different angle @xmath49 . similarly , the ratios of the optical depths are proportional to the line segments np and fp . from the elementary properties of the ellipse , it follows that the two ratios are the same ( np / fp = np / fp ) . by straightforward trigonometry , one has @xmath50 . in other words , the asymmetry signal of a flattened halo with axis ratio @xmath46 is the same as that of an equivalent spherical halo , provided the disk is viewed not at angle @xmath48 but at an angle @xmath49 . this transformation takes into account the geometric effects of flattening . the first - order changes in the velocity distribution can be computed using the tensor virial theorem ( binney & tremaine 1987 ; han & gould 1995 ) . strictly speaking , the tensor virial theorem applies globally and relates the components of the total kinetic energy tensor @xmath51 to the components of the total potential energy tensor @xmath52 . if we assume that the virial theorem holds at each spot , then it follows that @xmath53 here , we have assumed that the figure is oblate spheroidal , with the the short axis being in the @xmath54 direction and the @xmath55 plane being equatorial . although this is only an approximate relation , valid for small flattening , it shows that the first - order changes in the velocity distribution are also accounted for by the transformation . this means that a quick way to study the asymmetry properties of flattened haloes is to take isotropic spherical models and vary the inclination angle of the m31 disk . we stress that the transformation does not allow us to calculate absolute quantities like the rate , but only the ratios of such quantities in the near and far disk . figure [ fig : flatpix ] shows the variation in the far - to - near - disk ratios for the pixel lensing rates @xmath39 , the means of vertical distances @xmath38 and the time - scales . the asymmetry signal @xmath39 for a halo with flattening @xmath46 is linearly related to the asymmetry signal for a spherical halo @xmath56 , that is @xmath57 the asymmetry signal clearly diminishes with flattening ; it is obvious that , in the completely flat limit where the halo becomes a razor - thin disk , the asymmetry must vanish . the change in @xmath39 with flattening is almost entirely caused by the change in the ratio of optical depths . therefore the ratio of the average time - scales is largely unaffected by flattening . the distance asymmetry @xmath38 also decreases with increasing flattening , as is again obvious in the razor - thin limit . the distance asymmetry arises in a spherical model because of two effects . first , lines of sight are longer as we move from near to far disk . second , the line of sight with greatest column density goes through the centre in the near - disk but lies away from the centre in the far - disk . so the distribution of distances of events is monotonically decreasing in the near disk , but rises to a maximum and then decreases in the far disk . as the flattening increases , all lines of sights become shorter and the density becomes more concentrated towards the centre . the distributions of distances in both the near and far disk shrink and the maximum moves towards the centre in the far disk . this latter effect is the dominant one , and so the distance asymmetry signals falls with increasing flattening . in this section , we ask whether one can detect the asymmetry signals in the presence of contaminating events and how the velocity anisotropy and the halo flattening affect this detectability . that is , we ask how many candidates are needed to detect the asymmetry at a certain confidence level . first , we consider the number asymmetry @xmath39 between the far and near side . one problem is that none of the experiments will be able to obtain pure samples of m31 macho events ; some microlensing contamination from milky way machos and m31 stars , as well as non - microlensing contamination from variable stars and supernovae in background galaxies will be inevitable . contamination by periodic variables can be minimized by observing over a sufficiently long time . colour information can also be used to eliminate variable stars . fortunately , the other contaminants are equally likely to occur in the near and far regions of the m31 disk . whether or not an asymmetry can be detected is therefore a question of the size of the sample , the magnitude of the underlying m31 macho asymmetry and the level of contamination . let us denote the numbers of m31 macho events in the near and far disk by @xmath58 and @xmath59 respectively , and the number of contaminants by @xmath60 . the condition for a detection of the asymmetry signal at the @xmath61-@xmath62 level is that the difference between the far and the near counts be greater than @xmath61 times the poisson error , which is given by the square root of the total number of the events . therefore @xmath63 where @xmath64 and @xmath65 are the total number of candidates on the far and near sides , respectively , including contaminant events . in the case where the contaminants are distributed evenly between the far and near disk , @xmath66 . straightforward manipulation of the above condition leads to a condition on the total number of candidate events , @xmath67 where @xmath68 is the underlying asymmetry of the m31 macho events , @xmath69 is the contamination factor , and @xmath70 the total number of candidate events . figure [ fig : crit ] shows the expected number of events ( including the contaminants ) required to detect an asymmetry at the 99% confidence level ( @xmath71 ) for various contamination factors . we see immediately that the size of the asymmetry signal is crucial . if @xmath72 , then even with low contamination the number of m31 microlensing events needed to give a convincing detection of asymmetry exceeds @xmath73 . on the other hand , if the asymmetry is @xmath15 , then more than 15 candidates are required . reassuringly , on referring to figures [ fig : pix ] and [ fig : flatpix ] , we see that @xmath39 typically lies between 5 and 7 for spherical models , and only becomes as small as 2 for models with @xmath0 . in fact , haloes flatter than this are not likely on dynamical grounds , as they are susceptible to bending instabilities ( e.g. , merritt & sellwood 1994 ) . samples of @xmath115 candidate events are well within reach of the current surveys if machos contribute significantly to the dark matter budget . however , this number is only the contribution of the m31 halo , as we have cut out the central parts of the m31 disk . if the contribution of machos to the dark matter mass budget is significant in m31 , we expect a similar macho contribution for the milky way . for interesting macho fractions , the foreground milky way machos may well provide the dominant contribution to @xmath74 . if the typical macho mass and halo density contributions are universal , then the magnitude of @xmath74 will be determined , to first order , by the relative masses of the m31 and milky way haloes . if the m31 halo is twice as massive as the milky way s , then @xmath75 ( kerins et al . 2001 ) , whilst we should expect @xmath74 to be closer to unity if the two haloes are equally massive ( evans & wilkinson 2000 ) . in any case , the total number of candidates required to confirm asymmetry scales as @xmath76 from equation ( [ asymm ] ) , so we require four times as many candidates to detect the asymmetry when @xmath77 as when @xmath78 . motivated by the variation in mean distances of events between the near and far disk seen in figures [ fig : pix ] and [ fig : flatpix ] , we can go beyond the asymmetry in number counts by considering a distance asymmetry signature . to quantify differences between the near - disk and far - disk event positions , we apply the mann - whitney rank - sum test ( mann & whitney 1947 ) . this is a non - parametric test for differences between the medians of two samples . the mann - whitney test exploits the fact that two samples drawn from identical distributions exhibit the property that , if one combines them and then ranks the elements by size , the two samples are uniformly intermingled on average within the ranked combined sample . for two samples of large enough size @xmath79 and @xmath80 , the sum of the rank numbers for sample @xmath81 is normally distributed about a mean @xmath82 with a variance @xmath83 . the null hypothesis of similarity ( or , in our case , symmetry ) is therefore straightforward to quantify . in the case of m31 pixel lensing , we can apply the mann - whitney test to the distribution of @xmath20 , the projected distance from the m31 major axis , for candidates in the near- and far - disk sub - samples . we can also combine the mann - whitney test with the number asymmetry test , since the mann - whitney statistic probes the spatial distribution of events , not the relative sizes of the samples . if the far - disk event positions are designated as sample @xmath81 and the sum of their rank numbers within the combined near- and far - disk sample is @xmath84 , then we can define @xmath85 . taking @xmath86 to be the significance of the number asymmetry statistic , as defined by the left - hand side of equation ( [ asymm_c ] ) , then the overall significance of the combined sample is @xmath87 . a value of @xmath88 would indicate a near - far asymmetry favoured at the 99% confidence level . we have performed monte carlo simulations to determine the number of candidates required to secure a 99% confidence detection of asymmetry . the simulations test a range of anisotropic and flattened halo models and a range of contaminations . for a given model , event position realisations are generated using the theoretical pixel - lensing rate of equation ( [ prate ] ) , weighted by source number density . for each new event realisation , the overall significance , @xmath61 , of the cumulative sample is computed from both the mann - whitney statistic ( @xmath89 ) and the number asymmetry statistic ( @xmath86 ) . when there are fewer than five events in either the far- or near - disk sub - samples , only the number asymmetry statistic is used because the distribution of rank sums for small data - sets can deviate strongly from gaussianity . we assume the contaminating populations are symmetrically distributed about the m31 major axis , though we adopt the most difficult case when they have a comparable spatial dispersion to the m31 macho events . the extent to which this is true depends upon whether milky way machos , on the one hand , or variable stars and stellar microlenses , on the other hand , provide the dominant contribution to @xmath60 . if the latter population dominates , then contaminants are likely to be more spatially concentrated than m31 machos , making the macho asymmetry easier to measure for a given @xmath74 . so for each pixel - lensing realisation , a uniform random number in the interval [ 0,1 ] is chosen . if it is less than @xmath90 $ ] , then the position of the event is flipped about the m31 major axis . this means that , on average , a fraction @xmath91 of the total sample is symmetrically distributed , as required . a full trial , @xmath48 , is terminated either when the cumulative sample of @xmath92 candidates provides at least a 99% confidence detection of asymmetry , or when the program estimates that the required sample is likely to exceed 500 events . this whole process is repeated 1000 times for each model and the median value of @xmath93 ( excluding trials which are prematurely terminated ) is adopted as the estimate for @xmath94 . the upper panel of figure [ fig : monte ] shows the median number of candidate events ( including the contaminants ) required to detect an asymmetry with 99% confidence , as a function of the halo velocity anisotropy parameter @xmath40 . in the absence of contamination , a sample size between 11 to 14 events is typically sufficient to detect asymmetry . the raggedness of this line is partly due to monte carlo noise but also partly due to the fact that @xmath94 necessarily takes only a few discrete values in the limit of small data - sets . from equation ( [ asymm_c ] ) , the smallest sample needed give rise to a 99% confidence detection of asymmetry is 7 events , all of which must be located in the far disk ( assuming the asymmetry is caused by m31 machos ) . when one candidate lies in the near disk we need @xmath95 , and when two candidates lie in the near disk we require @xmath96 . the leaps in the thick solid line between 11 and 14 events reflect this discrete behaviour , though the oscillation back and forth for @xmath97 is due to monte - carlo noise . in the worst case considered in figure [ fig : monte ] , where contaminants outnumber m31 machos by 3:1 , a median sample of @xmath1120 events is required to detect asymmetry . for @xmath42 and @xmath98 , we find that the expectation value of @xmath94 is around 190 . using @xmath88 , @xmath98 and an asymmetry @xmath99 when @xmath42 ( figure [ fig : pix ] ) , equation ( [ asymm ] ) indicates that we should require @xmath100 if we use number - count information alone . therefore the addition of distance information allows around a 15% reduction in the required number of candidates in this case . overall , @xmath94 does not appear to be particularly sensitive to @xmath40 . this is to be expected when contamination levels are high because the 30% contrast in @xmath39 between the radially- and circularly - anisotropic models ( see figure [ fig : pix ] ) is strongly diluted . the lower panel of figure [ fig : monte ] shows the situation for flattened halo models . the larger range in @xmath39 for the flattened models ( @xmath101 ) means that the median @xmath94 shows a greater sensitivity to flattening than to velocity anisotropy . in the absence of contaminants , @xmath102 is typically needed to confirm asymmetry for models rounder than @xmath0 . a sample of 100 candidates would permit asymmetry to be detected even if @xmath103 . adding distance information is particularly effective at reducing the required size of candidate samples for highly flattened models . as an extreme example , when @xmath104 and @xmath78 , the expectation of @xmath94 is around 80 if distance information is used along with number counts . from equation ( [ asymm ] ) , 750 candidates are required if only number counts are used . this shows the value of the mann - whitney statistic . it is of course much more difficult to measure an asymmetry than to detect one . suppose an experiment has gathered @xmath105 events ( including contaminants ) . with reference to figure [ fig : crit ] , if there has been no detection of asymmetry at the 99% confidence level , then as @xmath106 for all models we have considered with @xmath107 we can infer that the signal has been overwhelmed by contaminants ( @xmath108 ) . the degree of contamination must be greater than a critical value which is given by computing the curve which passes through the point ( @xmath109 ) . so , a null signal can be used to give a constraint on the contamination . on the other hand , if there has been a detection , then the measured signal @xmath110 is merely a lower limit to the true signal . it should be possible to estimate the contamination fraction statistically even if we do not know the individual contaminating events , so the true asymmetry signal can be matched to models using standard bayesian likelihood methods ( kerins et al 2001 ) . current surveys should therefore be able to discriminate between halo models with different degrees of flattening and may , if the spatial distribution of contaminant populations is well characterized , be able to distinguish between radially - anisotropic halo models and tangentially - anisotropic or isotropic models . additional confusion of the asymmetry signal may come from microlensing by stars which belong to streams or tidal debris from disrupted satellite galaxies cannibalized by m31 and/or by stars belonging to the intervening dwarf elliptical m32 . ibata et al . ( 2001 ) have traced out a giant stream in the m31 stellar halo in red giant branch star counts . the stream is @xmath11 wide in projection . it seemingly originates from the satellite galaxy m32 and possibly also incorporates ngc 205 . the average surface brightness of the stream is @xmath130 @xmath23 in the @xmath16-band . it has been suggested by ferguson et al . ( 2002 ) that this may confuse the detection of the near - far disk asymmetry in the microlensing experiments . for sources in m31 and lensing populations at roughly the same distance from the sources , the optical depth of the stream is @xmath111 where @xmath112 is the surface brightness in magnitude per arcsec@xmath113 , @xmath114 the separation of the stream from m31 disk along the line of sight and m / l is the macho mass - to- light ratio of the stream . this ratio must exceed 100 for the optical depth of the stream to be comparable to the other lensing populations . accordingly , the presence of such streams is not likely to be problematic for asymmetry detection . more problematic may be the intervening dwarf elliptical galaxy m32 . the point - agape collaboration has already found a candidate which lies @xmath13 in projection away from the centre of m32 and argued that the lens most probably lies in m32 itself ( paulin - henrikkson et al . the optical depth of m32 is estimated to be ( paulin - henrikkson et al . 2003 ) @xmath115 where @xmath114 is the separation of m32 from m31 along the line of sight and @xmath116 is the stellar mass - to - light ratio of m32 . this is comparable to the signal expected from the baryonic halo ( assuming a 20% macho fraction ) . although microlensing associated with m32 is a potentially significant contaminent the affected region can be readily excluded from the statistics using an appropriate mask . pixel - lensing experiments targeting m31 are hoping to exploit the favourably high disk inclination in order to detect an asymmetry in the spatial distribution of microlensing events . if such a signal is found , it will provide powerful evidence for the existence of machos . for a spherically symmetric and isotropic halo , the numbers of m31 machos in the far - disk outnumber those in the near - disk by more than 5 to 1 . this asymmetry is increased by about 30% if m31 machos occupy radial orbits rather than tangential ones . the signal is diminished if m31 machos lie in a flattened halo . however , even for haloes as flat as @xmath0 , the numbers of m31 machos in the far side will outnumber those in the near side by a factor of @xmath12 . the key to detecting asymmetry is to isolate microlensing events solely due to m31 machos . there is likely to be significant contamination from other microlensing populations , as well as from variable stars and supernovae mistaken for microlensing , which will dilute the observed signal . the combination of number - count and distance information permits asymmetry to be detected for a wide range of halo models , even in the presence of significant levels of contamination . for models with high levels of asymmetry , such as spherical haloes or haloes with a high degree of radial velocity anisotropy , number count information alone provides a sensitive diagnostic . the addition of distance information allows @xmath115% reduction in the size of samples needed to confirm asymmetry . for models with low levels of asymmetry , such as strongly flattened haloes , distance information can reduce the required size of candidate samples by a factor 2 or more . a sample of 50 events is typically sufficient to detect asymmetry in the m31 macho distribution within spherical haloes , even if only half the sample is due to m31 macho events . for flattened halo models , a sample of 50 candidates would likely allow asymmetry to be seen , provided that the halo axis ratio @xmath117 and the contaminants do not contribute more than a third of the sample . the term `` contaminants '' covers milky way machos , m31 disk stellar lensing events and variable star populations , all of which are assumed to be symmetrically distributed with respect to the major axis . for comparison , paulin - henriksson et al . ( 2003 ) have already found 362 lightcurves compatible with microlensing from the first two years of the point - agape survey , though the contamination factor may still be very large . samples of 50 events with modest contamination are easily achievable with the current generation of pixel - lensing surveys . ( _ solid line _ ) as a function of anisotropy parameter @xmath40 and normalised to the value for the isotropic model @xmath118 . also shown is the m31 macho number asymmetry @xmath39 ( _ dashed line _ ) , the ratio of near - disk to far - disk average durations @xmath37 ( _ dot - dashed line _ ) , and the ratio of projected distances to the major axis @xmath38 ( _ dotted line _ ) ] ( _ dashed line _ ) , the mean vertical distance @xmath38 ( _ dotted line _ ) and the time - scales ( _ dot - dashed line _ ) , shown as a function of flattening . this diagram is drawn using the transformation introduced in section [ tilt ] . ] . _ upper panel _ shows @xmath94 as a function of halo velocity anisotropy for a spherical halo . _ lower panel _ shows @xmath94 for flattened halo models , using the equivalence between the halo flattening parameter @xmath46 and disk inclination @xmath48 discussed in section [ tilt].,title="fig : " ] . _ upper panel _ shows @xmath94 as a function of halo velocity anisotropy for a spherical halo . _ lower panel _ shows @xmath94 for flattened halo models , using the equivalence between the halo flattening parameter @xmath46 and disk inclination @xmath48 discussed in section [ tilt].,title="fig : " ]
the point - agape collaboration is currently searching for massive compact halo objects ( machos ) towards the andromeda galaxy ( m31 ) . the survey aims to exploit the high inclination of the m31 disk , which causes an asymmetry in the spatial distribution of m31 machos . here , we investigate the effects of halo velocity anisotropy and flattening on the asymmetry signal using simple halo models . for a spherically symmetric and isotropic halo , we find that the underlying pixel - lensing rate in far - disk m31 machos is more than 5 times the rate of near - disk events . we find that the asymmetry is increased further by about 30% if the machos occupy radial orbits rather than tangential orbits , but is substantially reduced if the machos lie in a flattened halo . however , even for haloes with a minor - to - major axis ratio @xmath0 , the numbers of m31 machos in the far - side outnumber those in the near - side by a factor of @xmath12 . there is also a distance asymmetry , in that the events on the far - side are typically further from the major axis . we show that , if this positional information is exploited in addition to number counts , then the number of candidate events required to confirm asymmetry for a range of flattened and anisotropic halo models is achievable , even with significant contamination by variable stars and foreground microlensing events . for pixel - lensing surveys which probe a representative portion of the m31 disk , a sample of around 50 candidates is likely to be sufficient to detect asymmetry within spherical haloes , even if half the sample is contaminated , or to detect asymmetry in haloes as flat as @xmath0 provided less than a third of the sample comprises contaminants . we also argue that , provided its mass - to - light ratio is less than 100 , the recently observed stellar stream around m31 is not problematic for the detection of asymmetry .
we live in exciting times for particle physics . the tevatron run ii program is producing new physics results , which will be followed by the startup of the lhc in three years . these machines have a rich physics menu of new physics searches , precision measurements , qcd studies , @xmath1 , @xmath2 , and @xmath3 quark physics , and much more . we are in an enviable position . we have an electroweak theory which is consistent with all data , but which predicts new physics waiting to be discovered at the tevatron and the lhc . when the lhc begins accumulating data , the physics landscape will change dramatically : the number of observed top quark pairs will jump from @xmath4 at the tevatron to @xmath5 in the first @xmath6 at the lhc . similarly , the number of @xmath7 pairs in our world data set will increase from @xmath8 from the @xmath9 factories at slac and kek to @xmath10 at the lhc . we are confident in our predictions that the wealth of new data at the lhc will guarantee new discoveries . in this note , we begin by reviewing new measurements of the top quark and @xmath11 boson masses and their implications for electroweak physics . we continue by discussing the emergence of the tevatron as a machine for precsion studies . finally , we conclude by emphasizing our certainty that new physics discoveries are just around the corner . two of the most interesting results of the last year were the new values for the top quark and the @xmath11 boson masses . a review of other electroweak measurements at the tevatron can be found in ref . @xcite . in 2003 , we had for the world average for the top quark mass@xcite , @xmath12 the d0 collaboration improved their run i analysis to include matrix element calculations , and this analysis dominates the new world average . in 2004 , we have the new combination of run i results@xcite , @xmath13 this value for @xmath14 is most sensitive to the d0 lepton plus jets result . as the tevatron run ii program continues , we expect ever more accurate measurements of the top quark mass@xcite . the top quark mass plays a special role in the standard model . in qed , the running of @xmath15 at a scale @xmath16 is not affected by heavy quarks with @xmath17 . the decoupling theorem tells us that diagrams with heavy virtual particles do nt contribute to experimental observables at scales @xmath18 if two conditions are met : * coupling constants do nt grow with @xmath19 , and * the gauge theory with the heavy quark removed is still renormalizable . the spontaneously broken @xmath20 gauge theory violates both conditions and we expect large effects from virtual diagrams involving the top quark , with effects from virtual top quarks growing quadratically with @xmath21 . a small change in the top quark mass therefore has large effects in the inferred predictions of the standard model . the @xmath11 boson is produced at the tevatron through the partonic sub - process , @xmath22 , and the leptonic decays of the @xmath11 can be used to determine the @xmath11 mass and width to good precision through a measurement of the transverse @xmath11 mass or the @xmath23 of the lepton . the run i value for the @xmath11 mass from a combination of cdf and d0 data is @xmath24@xcite . at lep2 , the @xmath11 boson was pair produced through the process , @xmath25 , ( see fig . [ eeww_fig ] ) , and a value for the @xmath11 mass was found of @xmath26 . the run i and lep2 values have been combined to give a new world average for the @xmath11 mass of@xcite @xmath27 this is slightly lower than the previous value , @xmath28 the result of the new top quark mass is to shift the best fit value for the higgs boson mass as can be seen in fig . [ fg : higsfit]@xcite , @xmath29 the small shift in the top quark mass , @xmath30 shifted the @xmath31 confidence level limit on the higgs mass by roughly @xmath32 . the best fit for the higgs mass is now and @xmath14 from the tevatron are included in the fit . a fit to the higgs mass without the @xmath11 mass measurement gives a best fit value of @xmath33 , while a fit which does not include @xmath14 or @xmath34 as measured at the tevtraon gives @xmath35.]@xcite @xmath36 it is reassuring to see that the best fit value is not in the region excluded by the direct search experiments at lep . ( the best fit value of 2003 was @xmath37 , a value which is excluded by the direct search experiments . ) the global fit of the experimental data to the parameters of the standard model is shown in fig . [ fg : globfit ] and is in spectacular agreement with the predictions of the standard model . the fits have been redone using the new value of @xmath14 and only the high @xmath38 data is included in the 2004 fits . the results of the low energy experiments such as atomic parity violation in cesium , neutrino - nucleon scattering ( nutev ) , and moller scattering ( @xmath39 ) are now predicted results . the prediction of the global fit for @xmath40 is in good agreement with the value of eq . [ topeq ] , while the best fit value for @xmath34 ( with the experimental value of @xmath14 included in the fit ) is slightly lower than the experimental value@xcite , @xmath41 the puzzles of the global fit from previous years remain . the extracted value of @xmath42 still has the problem that measurements from experiments with leptons and from experiments with hadrons disagree by @xmath43@xcite . using the value of @xmath42 found from leptons at sld ( @xmath44 ) , @xmath45 a value of the higgs boson mass in conflict with the direct search experiments is predicted . on the other hand , using @xmath42 from the forward backward @xmath2 asymmetry at lep ( @xmath46 ) , @xmath47 one obtains a prediction in disagreement with the measured value of @xmath34 . this can be seen in fig . [ fg : swfit ] . it is extremely difficult to construct models which explain this feature . measurements at low energy test our understanding of the energy scaling of the theory . the best fit values for the parameters of the standard model which are given in fig . [ fg : globfit ] can be used to make predictions for moller scattering , for the value of @xmath48 which is extracted from nutev experiment , and for atomic parity violation in cesium , and then compared with the experimental values . the energy dependence of the weak mixing angle @xmath48 is predicted in the standard model and is shown as the solid line in fig . [ ee ] and we see that the results of the low energy experiments are in reasonable , although not perfect , agreement with the predictions of the standard model except for the nutev result@xcite . the nutev collaboration has published its result for the ratio of neutral currents to charged currents in neutrino - nucleon scattering . when the measurement is interpreted in terms of @xmath48 , it is @xmath49 away from the global fit@xcite : @xmath50 the nutev experiment is under intense theoretical scrutiny since connecting the experimental observables with theoretical predictions requires a detailed understanding of theory . a number of possible solutions to the discrepancy between theory and experiment have been proposed , including an asymmetry in the strange / anti - strange parton distributions@xcite , and the effects of higher order @xmath51 corrections@xcite the situation with atomic parity violation has improved since the 2002 global fit showed a @xmath52 deviation of the experiment from the global fit . there is a new calculation of qed corrections which is now in good agreement with the best fit value@xcite @xmath53 moller scattering ( @xmath54 ) provides a clean measurement of the weak mixing angle since it is a purely leptonic process . a cominbination of the e158 experiment at slac run i , ii and iii data gives the result@xcite , @xmath55 in good agreement with the theoretical prediction , @xmath56 the inclusion of the low @xmath38 results into the fit has very little effect on the higgs mass limits . the conclusion from the electroweak fits is that electroweak physics is in even better shape this year than last year ! the standard model continues to be tested at the tevatron . in the coming few years , we can expect that the properties of standard model particles will be measured with high precision . @xmath11 and @xmath57 gauge bosons have large cross sections at the tevatron , so high statistics and precision measurements will dominate : @xmath11 and @xmath57 masses and width measurements , production cross sections , and gauge boson pair production cross sections . measurements of the top quark mass and top quark properties will continue to unfold , along with higgs searches , new physics searches , @xmath2 physics , and qcd studies , to name just a few areas . this conference saw a large number of new results in these areas from the d0 and cdf collaborations . the top quark plays a leading role in many models with physics beyond the standard model , so measurements of top quark properties play an especially important role in constraining these models . one particularly important measurement which we look forward to in the future is single top production . single top production at the tevatron has a total rate of roughly @xmath58@xcite , approximately half that of top pair production , and can serve to measure the ckm mixing parameter , @xmath59 . d0 has looked for single top production in @xmath60 and @xmath61 of run ii data and finds the @xmath62 limits@xcite , @xmath63 while cdf finds the corresponding @xmath62 limits@xcite , @xmath64 the production of @xmath11 and @xmath57 bosons tests the qcd production mechanism , while the ratio of the production cross section times the leptonic branching ratios can be used to extract the @xmath11 boson total width . figs . [ figw ] and [ figz ] show the results of @xmath65 of data from d0 , and @xmath66 from cdf , for @xmath11 production with the decay @xmath67 and for @xmath57 production followed by the leptonic decay , @xmath68 , along with a comparison to the standard model prediction . there is good agreement with the standard model predictions . a particularly interesting set of results concerns two gauge boson production , which is sensitive to the three gauge boson vertices , and hence to the non - abelian nature of the theory . the results from lep2 for @xmath69 ( fig . [ eeww_fig ] ) show that the three gauge boson couplings are very close to their standard model values . in an effective theory where there is new physics at a higher energy scale , the non - standard model contributions to the three gauge boson couplings would grow like @xmath70@xcite and we can expect enhanced couplings at the tevatron and lhc . preliminary measurements@xcite of the @xmath71 pair production cross section at the tevatron show good agreement with next - to - leading order ( nlo ) qcd predictions and no indication of non - standard model gauge couplings@xcite , @xmath72 advances in theory go hand in hand with higher statistics measurements at the tevatron . the strong coupling constant is not small : @xmath73 emission of soft gluons and multi - particle states are important , along with large effects from higher order perturbative qcd corrections . understanding precision measurements at the tevatron and the lhc requires new tools and the systematic inclusion of higher order qcd effects . this in turn requires new techniques for computing the higher order corrections . many processes are now calculated at next - to - next - to leading order ( nnlo ) . not only total cross sections@xcite , but also kinematic distributions for a few processes are becoming available at nnlo . for example , predictions for the drell - yan rapidity computed to nnlo in perturbative qcd at the lhc are shown in fig . [ drellyan_fig]@xcite , we see the very small error bar due to the scale dependence at nnlo . this suggests that it may be possible to use drell yan production to measure parton distribution functions at nnlo . understanding the precise data from the tevatron and future data from the lhc will require monte carlo programs which incorporate physics at next - to - leading order ( nlo ) in perturbative qcd . there are several programs on the market . the mc@nlo program matches parton showering calculations at low @xmath23 with exact nlo matrix element calculations at high @xmath23@xcite . an example of the importance of properly incorportating higher order effects is the @xmath2 production cross section at the tevatron , which has been a long standing puzzle . by including the higher order qcd corrections in both the production and the fragmentation functions , the theoretical prediction now agrees quite well with the data , as can be seen in fig . [ bfig]@xcite . another approach to including higher order qcd corrections is the mcfm monte carlo program@xcite . this program includes exact matrix element calculations for both the signal and background processes for a number of processes at hadron colliders . for most processes , the matrix elements are included to next - to - leading order and the full set of spin correlations is included for the decays . despite the successes of the standard model , the arguments for new physics at the @xmath0 scale have never been stronger . many of the questions particle physicists seek to answer become even more pressing as accelerators approach the @xmath0 energy scale : * the question of naturalness : why is @xmath34 so much smaller than @xmath74 ? * dark matter : what is it ? why is there so much of it ? * neutrino masses : where do they come from ? why are they so small ? * what mechanism restores unitarity to the @xmath75 scattering amplitude : is there a light higgs ? is there some other mechanism ? extensions of the standard model which have been proposed in various attempts to answer these questions almost uniformly predict physics beyond that of the standard model at the @xmath0 energy scale@xcite . so even though current results from lep and the tevatron support the experimental validity of the standard model , we are confident that there is new physics just around the corner . furthermore , the higgs sector of the standard model is unsatisfactory to theorists . light scalar particles , such as the higgs boson , are unnatural in the sense that the scalar higgs mass depends quadratically on new physics which may exist at a higher energy scale . in order to have a higgs boson mass below @xmath76 ( as suggested by the precision electroweak measurements ) , large cancellations between various contributions to the higgs boson mass are required . attempts to avoid this problem have led to an explosion of model building in recent years@xcite . models of new physics are highly constrained by precision measurements . new physics effects can be parameterized in a model independent fashion by adding a tower of dimension six operators to the standard model : @xmath77 precision measurements place significant limits on many possible operators . for example , current data requires@xcite , @xmath78 a more complete list of constraints on dimension 6 operators is given in ref . operators which violate flavor conservation are even more tightly constrained , often requiring scales @xmath79 . the fact that physics at the @xmath80 scale may already be excluded experimentally is termed the `` little hierarchy '' problem . * remove the higgs boson completely : models of this type include models with dynamical symmetry breaking@xcite and higgsless models in extra dimensions@xcite . * lower the cut - off scale @xmath81 of eq . [ ops ] : examples of this class of models are models with large extra dimensions@xcite . * force cancellations between contributions to the higgs boson mass : supersymmetric models are the favorite of this type , while little higgs models are a new entrant into the theory game ( although they have difficulty being consistent with precision measurements@xcite . ) z. sullivan , hep - ph/0408049 ; b. harris , e. laenen , l. phaf , z. sullivan , and s. weinzierl , _ phys . _ * d66 * ( 2002 ) 054024 , hep - ph/0207055 ; t. stelzer , z. sullivan , and s.willenbrock , _ phys . rev . _ * d58 * ( 1998 ) 094021 , hep - ph/9807340 . r. harlander and w. kilgore , _ phys . * 88 * ( 2002 , 201801 , hep - ph/0201206 ; c. anastasiou and k. melnikov , _ nucl . _ b646 ( 2002 ) , 220 , hep - ph/0207004 ; c. catani , d. de florian , and m. grazzini , * jhep05 * ( 2001 ) , 025 , hep - ph/0102227 ; o. brein , a. djouadi , and r. harlander , _ phys . lett . _ * b579 * ( 2004 ) 149 , hep - ph/0307206 ; r. harlander and m. steinhauser , _ phys . _ * d68 * ( 2003 ) 111701 , hep - ph/0308210 . g. kribs , hep - ph/0305157 and references therein;j . hewett , f. petriello , t. rizzo , * jhep0310 * ( 2003 ) 062 , hep - ph/0211218 ; t. han , h. logan , b. mcelrath , and l.wang , _ phy . _ * d67 * ( 2003 ) 095004 , hep - ph/0301040 ; m. chen and s. dawson , _ phys . rev . _ * d70 * ( 2004 ) 015003 , hep - ph/0311032 .
we discuss recent results from global electroweak fits and from the tevatron and review the motivation for physics at the @xmath0 energy scale .
when compared to most stars with similar stellar parameters , the sun displays an anomalous chemical abundance pattern . to date , high - resolution spectroscopic chemical abundance analyses offer that only @xmath315% of solar type stars have abundance patterns similar to the sun ( melndez et al . 2009 ; ramrez et al . the disparity between the sun and other solar twins has been demonstrated in a number of studies ( e.g. , melndez et al . 2009 ; ramrez et al . 2010 , 2011 ; gonzalez et al . 2010a ; schuler et al . 2011 ) , and is possibly due to the formation of terrestrial planets . planetesimal and terrestrial planet formation may have imprinted a signature in the solar composition by locking - up refractory elements in the surrounding disk , and leaving behind material low in refractory elements to later accrete onto the sun . in this letter , we present elemental abundance ratios as a function of condensation temperature for the @xmath38.2 gyr old solar twin hip 102152 ( hd 197027 ) . this 1 m@xmath4 star ( 0.97 m@xmath4 ; current study ) that is about to leave the main sequence is of interest not only because it has an abundance pattern similar to the sun , thus making it an excellent older solar analog for comparative studies , but also because its solar abundance pattern may suggest that this star is a candidate host for terrestrial planets . we also include abundances of the younger 2.9 gyr solar twin 18 sco and examine systematic differences between refractory and volatile elements in this representative solar twin and the older hip 102152 . additionally , we report li abundances for the two stars and examine the tight li - age depletion trend of solar twins studied at high - resolution , bringing further insights to the debate of whether low lithium is a signature of planets , or just an effect of secular stellar depletion ( e.g. , takeda et al . 2007 ; takeda & tajitsu 2009 ; ramrez et al . 2012 ; baumann et al . 2010 ; ghezzi et al . 2010 ; gonzalez et al . 2010b ; israelian et al . 2009 ) . spectra of hip 102152 and 18 sco were obtained with the ultra - violet and visible echelle spectrograph ( uves ; dekker et al . 2000 ) on the 8.2-m very large telescope ( vlt ) at the eso paranal observatory . observations were carried out on 30 august 2009 , with air masses of 1.11.5 for hip 102152 and 1.61.7 for 18 sco . multiple exposures were obtained at three spectrograph configurations ( six exposures per configuration ) for each star to achieve redundancy in the observations around the li doublet at @xmath56708 , with individual exposure times of 1020 min for hip102152 and 4590 s for the brighter 18 sco . spectral coverage included 30603870 , 48005770 , 58508200 , and 843010200 , with incomplete coverage in the near - infrared due to gaps between echelle orders . the spectral resolving powers were r(@xmath5/@xmath6 ) = 65,000 at @xmath530603870 , and r = 110,000 for @xmath5480010200 . the asteroid juno was also observed with the same spectrograph setups , and served as the solar reference in our differential analysis . standard reductions were carried out in iraf , including bias subtraction , flat fielding , wavelength calibrations , spectral order extractions , and spectra co - additions . continuum normalization was performed on the combined one - dimensional spectra in idl . final composite spectra had signal - to - noise ( s / n ) ratios that varied across the spectral orders , but had a typical s / n@xmath3500 pixel@xmath7 . it should be noted that our spectrograph setups allowed us to achieve s / n@xmath31000 pixel@xmath7 in the continuum regions about the li feature at @xmath56708 . chemical abundances of most species were determined similarly to the methods outlined in melndez et al . equivalent width analyses were performed using the _ abfind _ driver in the local thermodynamic equilibrium ( lte ) line analysis code moog ( 2002 version ; sneden 1973 ) and kurucz ( atlas9 ) model atmospheres , without convective overshooting ( castelli & kurucz 2004 ) . lte abundances were determined for the following z@xmath830 elements in our analysis : c , n , o , na , mg , al , si , s , k , ca , sc , ti , v , cr , mn , fe , co , ni , cu , and zn ; and are included in table [ tab1 ] . the line list was mostly taken from melndez et al . equivalent width ( ew ) measurements were first made with the automated code ares ( sousa et al . lines were measured by hand in both the star and sun with the iraf task _ splot _ when the ew @xmath9 10 m , or if the species had five or fewer lines . additionally , lines that were found to be outliers from the mean abundance of a species were also measured by hand . measurements were performed by fitting a single gaussian profile to the line profile , or deblending multiple gaussian profiles when necessary . to aid in continuum placement , the observed solar spectrum was overplotted on each stellar line . the measurements for the star and sun were then made in an identical manner , using the same regions of continuum . all elemental abundances were computed line - by - line differentially with respect to the sun , using the same line list . a strictly differential approach was taken to decrease the additional line - to - line scatter that can arise from uncertainties in the adopted transition probabilities . also , the differential approach minimizes systematic errors in the determinations of stellar parameters , as well as systematic uncertainties in the abundances due to deficiencies in the model atmospheres . for the odd - z elements of v , mn , co , and cu , we took into account effects of hyperfine splitting ( hfs ) . abundances of these elements were performed using our measured ews and the _ blends _ driver in moog , which allowed us to consider the numerous individual components and wavelength splittings of the measured lines . the hfs line lists used were adopted from melndez et al . additionally , abundances of the li resonance doublet feature at @xmath5=6708 were determined using spectrum synthesis , and will be discussed below . differential nlte corrections were shown to be very small in solar twins relative to the sun ( @xmath30.001 dex ) by melndez et al . ; therefore , we did not take them into account in this study , except for cr , mn , and co ( e.g. , bergemann & cescutti 2010 ) . the sensitivity of our differential chemical abundances to the choice of model atmosphere grid has been shown to be negligible ( melndez et al . ) . abundances were first determined for the sun using the standard solar atmospheric parameters of t@xmath10=5777 k and log g=4.44 . we determined our own value of the solar microturbulent velocity ( v@xmath11 ) by using the common spectroscopic technique of removing any trend between the abundances of fe i lines and reduced equivalent width . the resulting microturbulence was v@xmath11 = 0.86 km s@xmath7 , although this value was individually tweaked slightly for the analyses of both hip 102152 and 18 sco . an initial ew analysis was performed for the stars using the above atmospheric parameters adopted for the sun . effective temperatures ( t@xmath10 ) for the program stars were initially set at the solar value and then refined by employing excitation equilibrium on the differential abundances of the fe i lines . that is , all trends between the differential abundances of each line with excitation potential ( ep ) were removed . stellar microturbulences were refined by removing trends between the differential abundances and reduced ew . surface gravities ( log g ) were incrementally adjusted from solar to arrive at ionization balance between the average differential iron abundances derived from fe i and fe ii lines . final atmospheric parameters for the stars are presented in table [ tab1 ] . we determined the following values for hip 102152 : t@xmath10 = 5723@xmath15 k , log g = [email protected] dex , and v@xmath11 = [email protected] km s@xmath7 ( @xmath12v@xmath11 = 0.0 km s@xmath7 for v@xmath13 = 0.86 km s@xmath7 ) . the differential stellar parameters between hip 102152 and the sun are therefore -54 k in t@xmath10 , -0.09 dex in log g , and 0.0 km s@xmath7 in microturbulence , and thus confirm that this older dwarf is excellent for comparisons to the sun . our t@xmath10 is in good agreement with previous spectroscopic determinations by the lower resolution studies by ramrez et al . ( 2009 ) , who reported a temperature of 5746@xmath150 k , and baumann et al . ( 2010 ) , who report a t@xmath10 of 5737@xmath147 k. the previous estimates of surface gravities are also in excellent agreement with our determinations : log g = [email protected] ( ramrez et al . ) and [email protected] ( baumann et al . ) . our atmospheric parameters for 18 sco ( t@xmath10 = 5824@xmath15 k , log g = [email protected] dex , and v@xmath11 = [email protected] km s@xmath7)v@xmath11 = 0.02 km s@xmath7 for v@xmath13 = 0.87 km s@xmath7 . ] are also in agreement with independent determinations in melndez et al . ( 2012 ; 2013 ) . we estimate that the uncertainties in t@xmath10 , log g , microturbulence , and metallicity of the adopted model atmospheres are 5k , 0.02 dex , 0.01 km s@xmath7 , and 0.01 dex , respectively . these errors include consideration of measurement uncertainties and the degeneracies in the stellar parameters . abundance uncertainties for each element due to the errors in the stellar parameters are included in table [ tab1 ] . the last column in the table includes a total uncertainty for each element , obtained by adding in quadrature the uncertainties from the stellar parameters and observational standard error of the mean ( sem ) . the cumulative error from the stellar parameters is 0.004 dex for [ fe / h ] . the metallicities of hip 102152 and 18 sco were determined to be [ fe / h ] = -0.013 @xmath1 0.004 and 0.055 @xmath1 0.004 , respectively , as indicated in table [ tab1 ] . the solar metallicity of hip 102152 and slightly supersolar abundance of 18 sco were also reported in baumann et al . ( 2010 ) , who similarly found [ fe / h ] = -0.010 @xmath1 0.022 for hip 102152 and [ fe / h ] = 0.051 @xmath1 0.020 for 18 sco . elemental abundance ratios ( [ x/@xmath9c , o@xmath14 ) of species with z@xmath830 are plotted vs. equilibrium condensation temperature ( t@xmath15 ) for hip 102152 and 18 sco in figure [ f1 ] . the abundance ratios are relative to the average of the carbon and oxygen abundances ( @xmath9c , o@xmath16 ) , since those elements may be undepleted in the solar atmosphere ( melndez et al . 2009 , 2012 ) , and allow us to visually compare volatile ( t@xmath15 @xmath9 900 k ) and refractory ( t@xmath15 @xmath16 900 k ) species easily . the differential abundances of hip 102152 display a pattern that is similar to solar , with a slope ( solid line ) consistent with zero ( -3.7 @xmath17 10@xmath18 @xmath1 9.9 @xmath17 10@xmath18 k@xmath7 ) and an element - to - element scatter of 0.012 dex , after excluding three outliers ( n , na , and co ) that have abundances significantly lower than the rest . given the large age of hip 102152 , the low abundances of these three elements may be due to a lower contribution from type ia supernovae ( co ; figure 7 in melndez et al . 2012 ) and asymptotic giant branch ( agb ) stars ( n and na ) . we will re - examine these possible scenarios for our larger sample of solar twins of disparate ages . abundance ratios of neutron capture elements with z@xmath1630 also appear roughly solar for the r - process elements , but the s - process elements appear deficient , which is consistent with the low n and na abundances resulting from the lower agb contribution . these results will be presented elsewhere . in contrast , our abundances of 18 sco in figure [ f1 ] agree with previous studies ( e.g. , melndez et al . 2012 ) , which show the star has an abundance pattern similar to other solar twins . that is , the refractory elements are enhanced relative to the volatile species , as illustrated by the dashed line . a similar correlation has been demonstrated on larger samples of solar twins in a number of high - resolution studies ( e.g. , melndez et al . 2009 ; ramrez et al . 2009 ; schuler et al . 2011 ) . the element - to - element scatter for 18 sco is low , at 0.010 dex . as shown in figure [ f1 ] , the abundances of refractory elements between hip 102152 and 18 sco differ by as much as 0.07 dex . lte lithium abundances were determined for hip 102152 , 18 sco , and the sun with spectrum synthesis using the _ synth _ driver in moog . we estimate the solar li abundance to be log @xmath2 ( li ) = 1.03 @xmath1 0.02 , in strong agreement with the reported value of log @xmath2 ( li ) = 1.05 @xmath1 0.10 by asplund et al . as shown in figure [ f2 ] , the li doublet is weak in the spectrum of hip 102152 , compared to the also depleted solar line . the high s / n @xmath31000 pixel@xmath7 in the region about the line allows us to place bounds on the li abundance of hip 102152 from log @xmath2 ( li ) @xmath30.370.51 dex , with a best estimate of log @xmath2 ( li ) = 0.43 @xmath1 0.07 . in contrast , the line of the younger solar twin , 18 sco , is much stronger . we estimate the abundance of this star to be log @xmath2 ( li ) = 1.58 @xmath1 0.02 . nlte li abundance corrections were determined as in lind et al . ( 2009 ) , and estimated to be + 0.04 dex for 18 sco and the sun , and + 0.05 dex for hip 102152 . ischronal ages and masses were estimated for hip 102152 and 18 sco using yonsei - yale isochrones ( kim et al . 2002 ) based on their stellar parameters and associated errors using probability distribution functions as described in melndez et al . in figure [ f3 ] we present nlte li abundances vs. isochronal stellar age for a small group of solar twins studied at high - resolution ( r@xmath1660,000 ) and high s / n ( @xmath16400 ) , along with four theoretical li abundance models for a solar - mass , solar metallicity star ( charbonnel & talon 2005 ; do nascimento et al . 2009 ; xiong & deng 2009 ; and denissenkov 2010 ) . from the published studies , we chose a model that best represented the li abundance data . in the interest of better characterizing and constraining the timescale of li depletion models , we are concentrating on studies with high quality data for two reasons . ( 1 ) for stars younger than the sun , these constraints limit uncertainties in age , which have unfortunately impeded previous studies ( e.g. , baumann et al . 2010 ) , and ( 2 ) for stars older than the sun , the high s / n ratios are crucial for obtaining accurate measurements using very weak li lines ( see figure 2 ) , which have restricted previous attempts ( e.g. , takeda et al . 2007 ) . in figure [ f3 ] li abundances of hip 102152 , 18 sco , and the sun are from the current study , while the abundances and ages of hip 56948 and 16 cyg b were taken from melndez et al . ( 2012 ) and ramrez et al . ( 2011 ) , respectively . age determinations for all the stars were determined using the same method . recently , do nascimento et al . ( 2013 ) reported an old 6.7 gyr solar twin with a low li abundance of log @xmath2 ( li ) @xmath8 0.85 dex , but this value is unfortunately limited by low s / n observations . as predicted and demonstrated previously , older stars show significantly more depletion than younger stars , with the abundance of hip 102152 being the lowest li abundance reported for a solar twin to date . this low abundance along with the abundance of the widely separated binary star 16 cyg b may be useful in distinguishing between the depletion models , as well as helping to constrain the models for the upper end of the age range of solar - like main sequence stars , where the models start to diverge or are incomplete . overall the models represent the data well for the parameter space covered . given this surprisingly tight preliminary correlation , we are currently analyzing a larger set ( @xmath315 ) of solar twins of disparate ages observed with uves to further fill in this diagram and hopefully confirm and strengthen the trend . the solar twin hip 102152 is notably similar to the sun in mass , effective temperature , surface gravity , microturbulence , and metallicity . the solar abundance pattern of hip 102152 that we report is the closest to the solar pattern published to date , using high precision abundances . the abundance ratios of hip 102152 show no enhancements of refractory elements compared to volatile species , unlike other solar twins such as 18 sco . the abundances of the old star hip 102152 are even closer to solar than the abundances of hip 56948 , a solar twin previously found to be the most similar to the sun ( melndez et al . 2012 ) . the sun appears deficient in refractory elements relative to volatile species , when compared to other solar twins ( melndez et al . 2009 ) . assuming the solar twins and the sun formed under similar conditions , from gas of similar composition , then the solar refractory elements must be tied - up somewhere . a reasonable inference by many authors is that the observed pattern is due to the accretion of dust - depleted gas onto the pre - main sequence sun ( e.g. , nordlund 2009 ; chambers 2010 ) . perhaps one of the more tantalizing mechanisms that has been suggested to produce dust - depleted gas in a natal disk is the formation of planetesimals and terrestrial planets , since the abundance patterns of the inner solar - system planets and meteorites are conversely enhanced in refractory elements ( ciesla 2008 ; alexander et al . if the solar abundance pattern is in fact a consequence of terrestrial planet formation , then other stars with abundance patterns similar to the sun may be identified as potential rocky planet hosts . in fact , the solar abundance pattern of hip 102152 makes it a prime potential candidate host for terrestrial planets . this star is included in our long - term radial velocity monitoring survey of solar twins being carried out with harps on the eso 3.6-m telescope ; however , current observations show no evidence of velocity variations that may be attributed to giant planets within 2 au ( see figure [ f4 ] ) . while most stellar parameters of hip 102152 are virtually identical to the sun , the age of the star is significantly larger , by @xmath33.6 gyr . the star s old age and physical similarities to the sun make hip 102152 an excellent older analog for the sun . in particular , this star is ideal for exploring the potential evolution of the sun s lithium abundance , by probing the older end of the lithium - age trend both observationally and theoretically . lithium is severely depleted in hip 102152 , as predicted . this star along with other solar twins firmly demonstrate that the sun s li has depleted as predicted for a star with its age , mass , and metallicity . previously , the apparently low solar li abundance has been of concern for investigations that have compared the sun to other solar analogs , with a broader range of stellar parameters ( e.g. , lambert & reddy 2004 ) . however , when comparisons of the sun s li abundance are made strictly to other solar twins , the sun appears normal ( melndez et al . the small amount of scatter found about the li - age relation for the sample of stars considered here is encouraging , and worthy of followup . this may be accomplished with the ongoing spectroscopic analysis of our larger sample of solar twins . the high - resolution and high s / n of our uves spectra are allowing us to determine more precise stellar parameters for the sample , and thus more precise isochronal ages . it is expected that our remaining data set will address the longstanding problem of li depletion and provide tight constraints for non - standard models of temporal li evolution . some authors have claimed enhanced li depletion in planet hosts , but li - age trends have been largely ignored or not detected due to age determination uncertainties . here we have derived reliable ages and highly - precise li abundances , leading to a well - defined li - age correlation that demonstrates the importance of the age bias in li - planet studies . trm and jm acknowledge support from fapesp ( 2010/19810 - 4 and 2012/24392 - 2 ) . ir s work was performed under contract with caltech , funded by the nasa sagan fellowship program . we thank bengt gustafsson and stefan dreizler for useful comments . alexander , c. m. o. , boss , a. p. , & carlson , r. w. 2001 , science , 293 , 64 asplund , m. , grevesse , n. , sauval , a. j. , & scott , p. 2009 , , 47 , 481 baumann , p. , ramrez , i. , melndez , j. , asplund , m. , & lind , k. 2010 , , 519 , a87 bergemann , m. , & cescutti , g. 2010 , , 522 , a9 castelli , f. , & kurucz , r. l. 2004 , arxiv : astro - ph/0405087 chambers , j. e. 2010 , , 724 , 92 charbonnel , c. , & talon , s. 2005 , science , 309 , 2189 ciesla , f. j. 2008 , meteoritics and planetary science , 43 , 639 dekker , h. , dodorico , s. , kaufer , a. , delabre , b. , & kotzlowski , h. 2000 , , 4008 , 534 denissenkov , p. a. 2010 , , 719 , 28 do nascimento , j. d. , jr . , castro , m. , melndez , j. , et al . 2009 , , 501 , 687 do nascimento , j .- d . , jr . , takeda , y. , melndez , j. , et al . 2013 , , 771 , l31 ghezzi , l. , cunha , k. , smith , v. v. , & de la reza , r. 2010 , , 724 , 154 gonzalez , g. , carlson , m. k. , & tobin , r. w. 2010a , , 407 , 314 gonzalez , g. , carlson , m. k. , & tobin , r. w. 2010b , , 403 , 1368 israelian , g. , delgado mena , e. , santos , n. c. , et al . 2009 , , 462 , 189 kim , y .- c . , demarque , p. , yi , s. k. , & alexander , d. r. 2002 , , 143 , 499 lambert , d. l. , & reddy , b. e. 2004 , , 349 , 757 lind , k. , asplund , m. , & barklem , p. s. 2009 , , 503 , 541 melndez , j. , asplund , m. , gustafsson , b. , & yong , d. 2009 , , 704 , l66 melndez , j. , ramrez , i. , casagrande , l. , et al . 2010 , , 328 , 193 melndez , j. , bergemann , m. , cohen , j. g. , et al . 2012 , , 543 , a29 nordlund , a. 2009 , arxiv:0908.3479 nehag , a. , korn , a. , gustafsson , b. , stempels , e. , & vandenberg , d. a. 2011 , , 528 , a85 ramrez , i. , melndez , j. , & asplund , m. 2009 , , 508 , l17 ramrez , i. , asplund , m. , baumann , p. , melndez , j. , & bensby , t. 2010 , , 521 , a33 ramrez , i. , melndez , j. , cornejo , d. , roederer , i. u. , & fish , j. r. 2011 , , 740 , 76 ramrez , i. , fish , j. r. , lambert , d. l. , & allende prieto , c. 2012 , , 756 , 46 schuler , s. c. , cunha , k. , smith , v. v. , et al . 2011 , , 737 , l32 sneden , c. a. 1973 , ph.d . thesis , sousa , s. g. , santos , n. c. , israelian , g. , mayor , m. , & monteiro , m. j. p. f. g. 2007 , , 469 , 783 takeda , y. , kawanomoto , s. , honda , s. , ando , h. , & sakurai , t. 2007 , , 468 , 663 takeda , y. , & tajitsu , a. 2009 , , 61 , 471 xiong , d. r. , & deng , l. 2009 , , 395 , 2013 + c & -0.006 & 0.006 & 4 & -0.005 & 0.005 & 0.000 & -0.001 & 0.007 & 0.010 + n & -0.060 & 0.013 & 1 & -0.005 & -0.001 & -0.003 & -0.009 & 0.011 & 0.017 + o & 0.003 & 0.004 & 3 & -0.007 & 0.002 & -0.001 & 0.000 & 0.007 & 0.008 + na & -0.057 & 0.006 & 3 & 0.002 & 0.000 & 0.000 & 0.001 & 0.003 & 0.007 + mg & 0.013 & 0.012 & 3 & 0.002 & -0.002 & -0.001 & 0.001 & 0.003 & 0.012 + al & 0.002 & 0.005 & 6 & 0.002 & -0.002 & 0.000 & 0.001 & 0.003 & 0.005 + si & -0.012 & 0.002 & 13 & 0.000 & 0.001 & -0.001 & 0.001 & 0.002 & 0.003 + s & -0.017 & 0.012 & 6 & -0.004 & 0.004 & 0.000 & -0.001 & 0.006 & 0.013 + k & -0.001 & 0.003 & 1 & 0.004 & -0.008 & -0.002 & 0.003 & 0.009 & 0.010 + ca & -0.003 & 0.002 & 10 & 0.003 & -0.003 & -0.002 & 0.002 & 0.005 & 0.006 + sc & -0.016 & 0.004 & 9 & 0.000 & 0.007 & -0.002 & 0.003 & 0.008 & 0.009 + ti & -0.005 & 0.003 & 21 & 0.005 & 0.000 & -0.001 & 0.001 & 0.005 & 0.006 + v & -0.035 & 0.001 & 9 & 0.005 & 0.001 & -0.001 & 0.001 & 0.005 & 0.005 + cr & -0.016 & 0.002 & 16 & 0.004 & -0.002 & -0.002 & 0.002 & 0.005 & 0.005 + mn & -0.026 & 0.004 & 5 & 0.003 & -0.003 & -0.002 & 0.002 & 0.005 & 0.007 + fe & -0.013 & 0.001 & 87 & 0.003 & -0.001 & -0.002 & 0.002 & 0.004 & 0.004 + co & -0.076 & 0.003 & 8 & 0.004 & 0.002 & -0.001 & 0.002 & 0.005 & 0.006 + ni & -0.039 & 0.002 & 17 & 0.003 & 0.000 & -0.002 & 0.002 & 0.004 & 0.004 + cu & -0.021 & 0.007 & 4 & 0.003 & 0.000 & -0.002 & 0.002 & 0.004 & 0.008 + zn & -0.007 & 0.013 & 2 & -0.001 & 0.002 & -0.002 & 0.002 & 0.003 & 0.013 + + c & -0.012 & 0.009 & 4 & & & & & & 0.012 + n & -0.003 & 0.013 & 1 & & & & & & 0.017 + o & -0.008 & 0.006 & 3 & & & & & & 0.009 + na & 0.019 & 0.003 & 3 & & & & & & 0.004 + mg & 0.043 & 0.002 & 1 & & & & & & 0.004 + al & 0.031 & 0.009 & 6 & & & & & & 0.009 + si & 0.048 & 0.002 & 13 & & & & & & 0.003 + s & 0.005 & 0.006 & 6 & & & & & & 0.008 + k & 0.039 & 0.004 & 1 & & & & & & 0.010 + ca & 0.058 & 0.003 & 9 & & & & & & 0.006 + sc & 0.036 & 0.002 & 9 & & & & & & 0.008 + ti & 0.055 & 0.002 & 21 & & & & & & 0.005 + v & 0.041 & 0.002 & 9 & & & & & & 0.006 + cr & 0.059 & 0.001 & 16 & & & & & & 0.005 + mn & 0.042 & 0.004 & 5 & & & & & & 0.006 + fe & 0.055 & 0.001 & 85 & & & & & & 0.004 + co & 0.024 & 0.003 & 8 & & & & & & 0.006 + ni & 0.041 & 0.001 & 17 & & & & & & 0.004 + cu & 0.039 & 0.007 & 3 & & & & & & 0.008 + zn & 0.009 & 0.012 & 2 & & & & & & 0.012 +
we present the first detailed chemical abundance analysis of the old 8.2 gyr solar twin , hip 102152 . we derive differential abundances of 21 elements relative to the sun with precisions as high as 0.004 dex ( @xmath01% ) , using ultra high - resolution ( r = 110,000 ) , high s / n uves spectra obtained on the 8.2-m very large telescope . our determined metallicity of hip 102152 is [ fe / h ] = -0.013 @xmath1 0.004 . the atmospheric parameters of the star were determined to be 54 k cooler than the sun , 0.09 dex lower in surface gravity , and a microturbulence identical to our derived solar value . elemental abundance ratios examined vs. dust condensation temperature reveal a solar abundance pattern for this star , in contrast to most solar twins . the abundance pattern of hip 102152 appears to be the most similar to solar of any known solar twin . abundances of the younger , 2.9 gyr solar twin , 18 sco , were also determined from uves spectra to serve as a comparison for hip 102152 . the solar chemical pattern of hip 102152 makes it a potential candidate to host terrestrial planets , which is reinforced by the lack of giant planets in its terrestrial planet region . the following non - local thermodynamic equilibrium li abundances were obtained for hip 102152 , 18 sco , and the sun : log @xmath2 ( li ) = 0.48 @xmath1 0.07 , 1.62 @xmath1 0.02 , and 1.07 @xmath1 0.02 , respectively . the li abundance of hip 102152 is the lowest reported to date for a solar twin , and allows us to consider an emerging , tightly constrained li - age trend for solar twin stars .
we consider the standard problem of denoising grayscale images that are corrupted with additive white gaussian noise @xcite . in this setup , we are given the _ corrupted _ ( or _ noisy _ ) image @xmath0 where @xmath1 is some finite rectangular domain of @xmath2 , @xmath3 is the unknown _ clean _ image , @xmath4 are independent and distributed as @xmath5 , and @xmath6 is the noise level . the goal is find a _ denoised _ estimate @xmath7 of the clean image from the corrupted samples . the denoised image should visually resemble the clean image . one way to quantify the resemblance is using the mean - squared - error ( mse ) which is defined to be @xmath8 later in the paper , we will use the peak signal - to - noise ratio ( psnr ) which is given by @xmath9 . the image denoising problem has been extensively studied and a survey of even a fraction of this literature is beyond the scope of this paper . we instead refer the interested reader to @xcite - @xcite and the references therein . our present interest is in the image denoising applications of the edge - preserving bilateral filter @xcite . the denoised image in this case is set to be @xmath10 where @xmath11 the gaussian kernels in are respectively referred to as the _ spatial _ and _ range _ kernels . in practice , the domain @xmath12 is restricted to some neighbourhood of the origin . typically , @xmath12 is a square neighbourhood , @xmath13 \times [ -w , w]$ ] , where @xmath14 @xcite . we refer the reader to @xcite for a detailed exposition on the working and in particular the edge - preserving property of the filter . the bilateral filter has received renewed attention in the image processing community in the context of image denoising @xcite . it is well - known that , while the filter is quite effective in removing modest levels of additive noise , its denoising performance drops at large noise levels @xcite . it was demonstrated in @xcite that a patch - based extension of the filter can be used to bring the denoising performance of the filter at par with state - of - the - art methods . these , and other advanced patch - based methods @xcite , are however much more computation - intensive than the bilateral filter . in this work , we present a couple of ideas for improving the denoising performance of the classical bilateral filter , and give fast algorithms for the same . the contributions ( and the organization ) of the paper are as follows : @xmath15 in section [ sec : rbf ] , we demonstrate how the denoising performance of the bilateral filter can be improved at large noise levels ( at almost no additional cost ) by incorporating a simple pre - processing step into the framework . @xmath15 the modified filter is designed to be robust at large noise levels , and tends to perform poorly below a certain noise threshold . to get the best of the either filters at all noise levels , we propose to optimally combine them in a weighted fashion in section [ sec : owf ] . the optimality is in terms of a certain surrogate of the mean - squared - error ( mse ) known as stein s unbiased risk estimate ( sure ) @xcite . this mse - estimate is known to be quite accurate in practice @xcite and , as a result , the combined filter is almost always guaranteed to provide a lower mse than the original filter . @xmath15 following @xcite , we present an approximation for the proposed filtering in section [ sec : fastsure ] that has a fast implementation . we derive the sure estimator for this approximation , and demonstrate how it can be efficiently computed in the process of approximating the bilateral filter . the overall cost of computing the estimator and the optimally - weighted filters is about twice the cost of computing a single bilateral filtering using the fast algorithm in @xcite . we provide both visual and quantitative denoising results on standard test images in section [ sec : experiments ] which demonstrate that the improvement over the original filter is significant both visually and in terms of psnr . notice that the range filter in operates on the noisy samples . in other words , the corrupted image is used not just for the averaging but also to control the blurring via the range filter . what if the range filter could directly operate on the clean image ? it can be verified that the denoising image obtained using this `` oracle '' filter is visibly better and has higher psnr , which is not surprising . of course , we do not have access to the clean image in practice , and thus the oracle bilateral filter can not be realized . nevertheless , one could consider some form of proxy for the clean image . our proposal is simply to use the box - filtered version of the noisy image as a proxy . in particular , we propose the following robust bilateral filter ( rbf ) : @xmath16 where @xmath17 the amount of smoothing induced by the box filter is controlled by @xmath18 . we performed exhaustive simulations in this direction . the simulation results suggest that @xmath19 ( @xmath20 blur ) is optimal for most settings . a possible way to explain this is that this small blur is able to suppress the noise without excessively blurring the image features . the denoising results obtained using the standard and the robust filter a couple of test images are shown in table [ table1 ] . we note that the filters have been independently tuned with respect to @xmath21 and @xmath22 to get the best psnr . these results ( and the results on other test images that are not shown here ) suggest that the robust filter starts to perform better beyond a certain noise level depending on the type of image . this is not surprising since , at low noise levels , the box filtering introduces more blurring than noise suppression which brings down the overall signal - to - noise ratio . indeed , the corrupted image is already a good proxy for the clean image when the noise floor is small . on the other hand , notice that the improvement in psnr is often as large as @xmath23 db at large noise levels . the results in table [ table1 ] suggest that we could possibly perform better denoising by combining and . a particularly simple idea would be to take a linear combination of these estimates . that is , we could set the denoised image to be @xmath24 notice that this includes and as special cases . this bring us to the question of setting the weights @xmath25 and @xmath26 . this can hypothetically be done by minimizing the mse between and the clean image . however , we do not have access to the clean image . this is precisely where the classical stein s unbiased risk estimate ( sure ) @xcite is useful . this is given by @xmath27 sure has the property that its expected value equals that of for the gaussian noise model in @xcite . this makes it an useful surrogate for the mse , which can be computed without the knowledge of the clean image . we note that the idea of taking linear ( or affine ) combinations of estimators and tuning the weights to get the optimal sure has been tried in different contexts in the literature ; e.g. , see @xcite . note that the added computation required for sure are the computation of the partial derivatives in . we propose to find the optimal weights in by minimizing sure . in particular , on substituting in , we see that sure is quadratic in @xmath25 and @xmath26 . thus , by convexity , a necessary and sufficient condition for optimality is that that the gradient of must vanish at the optimal weights . in particular , it can be verified that the resulting gradient equations can be written as @xmath28 , where @xmath29 \sum_{\i \in i } \hat{f}_1(\i)\hat{f}_2(\i ) & \hspace{2 mm } \sum_{\i \in i } \hat{f}_2(\i)^2 \end{pmatrix},\ ] ] and @xmath30 \sum_{\i \in i } f(\i ) \hat{f}_2(\i ) - \sigma^2 \sum_{\i \in i } \partial_{\i } \hat{f}_2(\i ) \end{pmatrix},\ ] ] and @xmath31 are the optimal weights . to simply notations , we will henceforth use the shorthand @xmath32 in place of the operator @xmath33 appearing in . in summary , by computing @xmath34 and @xmath35 , and solving the @xmath36 linear equation @xmath28 , we get the optimal weights . the weights are then plugged into to get the final denoised image . .comparison of the standard bilateral filter ( sbf ) and the robust bilateral filter ( rbf ) at @xmath37 . for a fixed image and noise level , we tuned @xmath21 and @xmath22 to individually optimize the psnr obtained using both methods . [ cols="<,^,>,>,>,>,>,>",options="header " , ] [ table2 ] we now present some denoising results on standard test images . our objective is to compare the denoising results obtained using the proposed modification with the standard bilateral filter , both visually and in terms of psnr . in this work , we assume that @xmath6 is provided ; this has to be estimated in practice from the noisy image . in this regard , we note that it has been demonstrated in @xcite that the sure is quite robust to standard data - based estimates of @xmath6 . we remark that one can optimize the sure for the proposed filter with respect to @xmath21 and @xmath22 as done in @xcite ; however , we do not investigate this possibility in this paper . l*6cr & @xmath38 & ( 4,20 ) & @xmath39 & @xmath40 & @xmath41 & @xmath42 + direct & 32s & 120s & 70s & 185s & 74s & 125s + fast & 1.2s & 0.8s & 0.7s & 0.6s & 0.5s & 0.6s + [ table3 ] in table [ table2 ] , the denoising performance of the proposed weighted filter at different noise level is compared with the standard and the robust bilateral filter , as well the patch - based non - local means filter @xcite . we have used @xmath19 in for all the experiments . to have a fair comparison for a given image and noise level , we independently tuned @xmath21 and @xmath22 to optimize the psnrs of the respective filters . we also tuned the parameters of nlm to get the optimal psnr at each noise level . as remarked earlier , the robust filter starts to perform better than the standard filter beyond a certain noise level ( @xmath43 ) . notice that the psnr obtained using the optimally - weighted filters is consistently higher than that of the constituent filters at all noise levels ; the improvement is often as large as @xmath44 db . on the other hand , the improvement in psnr over the standard bilateral filter is often as high as @xmath23 db . this does not come as a surprise since it is well - known that the bilateral filter has a lot of scope of improvement at high - noise levels @xcite . however , what does come as a surprise is that proposed filter is competitive with the non - local means filter @xcite that uses patches ( groups of neighboring pixels ) instead of single pixels for denoising . for a visual comparison , the results of a particular denoising experiment are shown in figure [ images ] . notice that the denoised image obtained using the proposed filter looks much more sharp than that obtained using the standard filter , and has a significantly higher psnr . also , notice that the noise in the background is much less in ( d ) compared to ( c ) . in table [ table3 ] , we compare the run times of the direct and the fast implementation of the proposed filter for different parameter settings . notice that the fast implementation is significantly faster than the direct implementation . we demonstrated how a simple pre - processing step can substantially improve the denoising performance of the bilateral filter . to consistently get the best of the standard and the pre - processed filter at all noise levels , we proposed to optimally weight them using stein s unbiased estimate of the mse . the optimal weights were found by solving a small linear system . a fast algorithm for implementing the optimally - weighted filters was also described . we reported visual and psnr results on test images which confirmed the improvement over the original bilateral filter . an interesting finding was that the weighted bilateral filter is competitive with the non - local means filter . j. portilla , v. strela , m. j. wainwright , and e. p. simoncelli , `` image denoising using scale mixtures of gaussians in the wavelet domain , '' _ ieee transactions on image processing _ 1338 - 1351 , 2003 . s. p. awate and r. t. whitaker , `` unsupervised , information - theoretic , adaptive image filtering for image restoration , '' _ ieee transactions on pattern analysis and machine intelligence _ , vol . 3 , pp . 364 - 376 , 2006 .
the bilateral filter is known to be quite effective in denoising images corrupted with small dosages of additive gaussian noise . the denoising performance of the filter , however , is known to degrade quickly with the increase in noise level . several adaptations of the filter have been proposed in the literature to address this shortcoming , but often at a substantial computational overhead . in this paper , we report a simple pre - processing step that can substantially improve the denoising performance of the bilateral filter , at almost no additional cost . the modified filter is designed to be robust at large noise levels , and often tends to perform poorly below a certain noise threshold . to get the best of the original and the modified filter , we propose to combine them in a weighted fashion , where the weights are chosen to minimize ( a surrogate of ) the oracle mean - squared - error ( mse ) . the optimally - weighted filter is thus guaranteed to perform better than either of the component filters in terms of the mse , at all noise levels . we also provide a fast algorithm for the weighted filtering . visual and quantitative denoising results on standard test images are reported which demonstrate that the improvement over the original filter is significant both visually and in terms of psnr . moreover , the denoising performance of the optimally - weighted bilateral filter is competitive with the computation - intensive non - local means filter . image denoising , bilateral filter , unbiased risk estimator , fast algorithm .
synchronization in coupled chaotic oscillators has been extensively studied in the past twenty years @xcite , including complete synchronization @xcite , generalized synchronization @xcite , and phase synchronization ( ps ) @xcite , etc . in cs , the dynamics of two coupled systems totally coincide with each other ; in gs , certain functional relation exists between the dynamics of two coupled systems which are usually nonidentical . moreover , ps is a weaker synchronization form in which the phases of two oscillators can be locked while their amplitudes remain uncorrelated and chaotic . recently , the study of synchronization has been extended to the area of complex networks @xcite . for example , synchronization on small - world networks @xcite , scale - free networks @xcite , modular networks @xcite , weighted networks @xcite , and gradient networks @xcite , have been investigated . these studies aim to explore the interplay between network topology and dynamics on network . they are important for us to understand the real situations in complex systems comprising interacting elements in nature and human society . in literatures , there are two main approaches studying synchronization in complex networks . one approach is to study synchronization of phase oscillators on complex networks , i.e. , the generalized kuramoto models @xcite . in this model , the node dynamics is very simple which is governed by an ideal phase oscillators , @xmath0 , where @xmath1 is the frequency . the heterogeneity in the node dynamics can be modeled by assigning different , usually random , frequencies to different phase oscillators . the generalized kuramoto models in complex networks have the advantage that they can still be treated analytically in many aspects @xcite . the second approach is to study the synchronizability of complex networks through the master stability function @xcite . to apply this approach , the node dynamics in complex networks are assumed to be identical , and then the theory of master stability function provides a general mathematical framework to relate the synchronizability of a network to the spectral properties of the corresponding coupling matrix @xcite . the above two approaches allows an analytical and therefore relatively more complete understanding of synchronization in complex networks . nevertheless , in realistically physical or biological situations the node dynamics can conceivably be much more complicated , and in general be different from each other . the key point is that the possible collective behaviors for coupled nonidentical oscillators should be of the type of gs rather than cs and ps . it is thus of great importance to investigate the collective behaviors on complex networks where the local node dynamics might be different from each other . in this paper , we present a systematical study on gs phenomenon in various complex networks . our particular attention is paid to the development of gs in networks . for typical complex networks , such as scale - free networks , small - world networks , random networks and modular networks , interestingly , we find that gs generally occurs in these networks , regardless of whether the node dynamics are identical or nonidentical . for coupled identical oscillators on complex networks , a general bifurcation path toward the final global cs is shown as : non - synchronization @xmath2 pgs @xmath2 gs @xmath2 pcs @xmath2 cs . here , pgs and pcs stand for partial gs and partial cs , respectively . we further study how the network topology competes with the local dynamics to affect the development of gs on networks . it is shown that for linearly coupled oscillators on heterogeneous networks , gs usually starts from the hub nodes and then spreads to other nodes ; while on homogeneous networks , gs generally starts from the nodes whose local dynamics are less chaotic in the sense that the largest lyapunov exponents have relatively smaller values . moreover , we analyze how different coupling strategies can affect the development of gs on complex networks , a point which has not been addressed before . in processing this paper , we noticed a recent work which has reported gs phenomenon in scale - free networks @xcite . compared with the results in ref . @xcite , the present work is different in the following aspects . firstly , in our work gs is systematically investigated on various complex networks , including scale - free networks , random networks , small - world networks and modular - networks ; while in ref . @xcite , gs is mainly studied on a very special network , i.e. , the scale - free networks with tree - like structure . secondly , the present work characterizes the complete bifurcation path for coupled identical oscillators on complex networks , i.e. , from non - synchronization to global cs , including the transition regimes ; while in ref . @xcite , global cs has not be achieved . thirdly , the present work investigates the development of gs for coupled nonidentical oscillators , either parametrically different or physically different ; while ref . @xcite does not consider this general setting where the occurrence of gs is naturally justified . finally , in the present work , we discuss the effect induced by different coupling strategies on the development of gs on network . especially , we analyze and explain the different observations regarding the gs evolution on networks in this work and in ref . in fact , in many aspects the current study deepens and widens the work in ref . @xcite , and can offer more thorough and comprehensive insight in understanding gs phenomenon on complex networks . this paper is organized as follows . in sec . ii , the methods to characterize gs and cs on complex networks are described . in sec . iii , gs of coupled identical oscillators is studied on complex networks . the complete bifurcation path from non - synchronization to global cs is studied and the dynamical features in the transition regimes are characterized . in sec . iv , gs of coupled nonidentical oscillators is further studied on complex networks . our interest is focused on how the development of gs on networks can be affected by network topology and the local dynamics . in sec . v , the effect of different coupling strategies on the development of gs on complex networks is analyzed . especially , gs is observed in coupled system with hybrid oscillators even when their local dynamics are physically different . finally , a summary ends this paper in sec . the auxiliary - system approach has been extensively used to detect gs in two coupled chaotic systems @xcite . here , we can extend it to detect gs on complex networks . the key observation is that , for any given node in a network , the coupling from other nodes can be regarded as a kind of `` driving . '' in particular , we consider the following linearly coupled identical oscillators on a network : @xmath3 for @xmath4 , where @xmath5 denotes the dynamical variables of node @xmath6 , @xmath7 is the local vector field governing the evolution of @xmath5 in the absence of interactions with other nodes , @xmath8 is the element of the network adjacency matrix @xmath9 ( @xmath10 if there is a link between node @xmath6 and node @xmath11 , @xmath12 otherwise , and @xmath13 ) , and @xmath14 is the coupling strength . to apply the auxiliary - system approach , we consider a replica for each oscillator in the original network : @xmath15 for @xmath4 . note that the driving variable @xmath16 is identical for both eqs . ( [ eq : network_equation ] ) and ( [ eq : network_auxiliary_equation ] ) . if , for initial conditions @xmath17 , we have @xmath18 as @xmath19 , node @xmath20 then is entrained in the sense that its dynamics is no longer sensitive to initial conditions . in other words , there is gs relation between @xmath5 and @xmath21 for @xmath22 . numerically , we can examine the following local distance of gs between a node and its auxiliary counterpart : @xmath23 where @xmath24 is chosen to be larger than the typical transient time of the local dynamics @xmath25 . for oscillators on complex networks , gs may be gradually developed with the increase of the coupling strength . to characterize the development of gs on networks , we can define the distance of global gs as @xmath26 . here , @xmath27 denotes the spatial average over all nodes . if @xmath28 , global gs has been achieved between any two pair of oscillators on the whole network . for coupled identical oscillators on complex networks , cs is generally expected to take place as long as the network connectivity is not too sparse . to characterize cs , we can define @xmath29 as the distance of global cs , which measures the distance between the dynamics of all oscillators and their average , i.e. , @xmath30 where the meaning of @xmath24 and @xmath31 is the same as that in eq . ( [ eq : local_syn_distance ] ) . if @xmath32 , global cs has been achieved on the whole network . for individual node dynamics , we choose the chaotic lorenz oscillator : @xmath33^t,\ ] ] where @xmath34 are the state variables of the lorenz oscillator . note that in studying gs , the local dynamics could be different from each other . this can be modelled by setting different @xmath35 values for lorenz oscillators in the network . without losing generality , the coupling between two nodes in eq . ( [ eq : network_equation ] ) is through the @xmath36 variable . for convenience , in the present work , we use two kinds of node indices : @xmath37 and @xmath38 . in the first index , we order the degrees of the network so that @xmath39 denotes the node with the largest degree , @xmath40 is for the node with the second largest degree , and so on . in the second index , we rank the lorenz oscillators in the network according to their @xmath35 values , i.e. , @xmath41 denotes the node with the largest parameter @xmath42 , @xmath43 is for the node with the second largest parameter @xmath42 , and so on . these two indices of nodes are used throughout the paper . previously , in the study of synchronization of two coupled chaotic oscillators , it is found that cs generally takes place when the coupled oscillators are identical , while gs usually occurs in systems of coupled nonidentical oscillators . this observation is accurate for most cases except for some specially designed coupling strategies @xcite . however , this generally accepted view turns out to be not true when synchronization is studied on complex networks . in this section , we investigate the occurrence of gs for coupled identical chaotic oscillators on various complex networks . our particular interest is focused on the complete bifurcations from initial non - synchronization towards the final global cs with the increase of the coupling strength . by applying auxiliary - system approach , interestingly , we find that gs generally exists in coupled identical oscillators on various complex networks , including scale - free networks , random networks , small - world networks and modular networks . as an example , we first study the bifurcation route towards global cs for 300 coupled chaotic lorenz oscillators on a scale - free network . in this case , @xmath35=28 in eq . ( [ eq : lorenz ] ) for all nodes , i.e. , all oscillators are identical . intuitively , with the increase of coupling , we expect that the coupled chaotic oscillators will finally achieve global cs . the surprising finding here is that before the system achieves the final global cs , there exists another synchronization regime , i.e. , the gs regime , which usually occurs with much smaller coupling strength compared with cs . to illustrate the gs phenomenon on scale - free network , we plot the color - map of the distance matrix @xmath44 in fig . 1 . from fig . 1 , three distinct regimes of coupling can be identified from the viewpoint of gs . when the coupling strength is small , @xmath45 are greater than 0 for all nodes , showing that the system is in the non - synchronous state . on the other hand , when the coupling strength is large enough , @xmath45 are 0 for all nodes , showing that all oscillators have been entrained and the coupled system is in the global gs state according to the auxiliary system approach criterion . between these two regimes , it is the regime of partial gs , where part of the oscillators have been entrained but the others have not . obviously , this regime is a transition stage between non - synchronous state and the global gs state . to quantitatively characterize the gs regime for coupled chaotic oscillators on scale - free network , in fig . 2(a ) we plot the distance of global gs @xmath46 vs the coupling strength . from this figure , three regimes of coupling can numerically be identified . on the other hand , for coupled identical oscillators , cs is generally expected to occur as long as the connections of the network are dense enough . it is found that usually cs requires larger coupling strength than gs . this implies that gs is a weaker synchronization form compared with cs on complex networks , as in the situation of coupled two - oscillator systems . to characterize cs in the network , we plot the distance of global cs @xmath47 vs the coupling strength in fig . we find that before cs is achieved , there is also a transition stage , i.e. , the partial cs regime . combining all the results in fig . 2 , we can identify five dynamical regimes on the path towards global cs on the specific scale - free network as follows . i : @xmath48 , the non - synchronization regime ; ii : @xmath49 , the pgs regime ; iii : @xmath50 , the global gs regime ; iv : @xmath51 , the pcs regime ; v : @xmath52 , the global cs regime . these regimes thus characterize the complete bifurcations from non - synchronization state to global cs state for coupled identical oscillators on scale - free networks . in the above we have shown that there are two transition regimes , i.e. , the pgs regime and pcs regime , before the coupled system achieves global gs and cs , respectively . in fact , these two regimes are important to obtain information about how synchronization is generated and spread on a network . as shown in fig . 1 , with the increase of coupling strength , usually the hubs of scale - free network become entrained first . then more and more nodes are entrained with further increase of the coupling strength . at @xmath53 , a giant cluster of synchronous nodes in the sense of gs has been formed . the most important dynamic feature in the pgs regime is that nodes may oscillate between the entrained state and the unentrained state with the increase of the coupling strength . as shown in fig . 3 , if we focus on a specific node , we find that its path toward global gs is generally oscillatory . especially , for those nodes with small degrees , the path could be extremely oscillatory as shown in fig . as a consequence , in the pgs regime , the size of the synchronous giant cluster does not increase monotonically , instead it increases in an oscillatory manner with the increase of the coupling strength , as shown in fig . similar phenomenon has been observed in the pcs regime . for example , in fig . 4(b ) , we plot the number of nodes which have achieved cs with the hub ( with the largest degree ) . this figure actually characterizes the dynamical features for regime iv shown in fig . it is found that the size of the synchronous cluster oscillates drastically on the way toward global cs with the increase of the coupling strength . recently , the path towards global ps in generalized kuramoto model has been studied in ref . it is found that for phase oscillators with the simplest local dynamics , once a phase oscillator becomes frequency locked , it will never escape from the synchronous cluster . thus in the ps situation the size of the synchronous cluster increases monotonically with the increase of the coupling strength . such a path toward global synchronization can be called as _ static and monotonic_. in the present work , we demonstrate a fundamentally different scenario towards global synchronization for coupled chaotic oscillators on complex networks . the most important feature of the path is that the development and spreading of synchronization on networks is a dynamic process during which oscillators may frequently oscillate between the non - synchronous state and the synchronous state . this implies that the size of the synchronous cluster might be sensitive to the perturbation of the coupling strength in the transition regimes . to visualize the dynamic feature of the development of synchronization on networks , a schematic figure has been plotted in fig . 5 for a scale - free network . our results suggest that in more realistic situations , the path towards global synchronization on complex networks could be _ dynamic and oscillatory _ rather than _ static and monotonic_. in the above we have characterized the path towards global cs for coupled identical chaotic oscillators on scale - free network , which is characterized by a heterogeneous degree distribution . we have also investigated synchronization path for coupled identical chaotic oscillators on other typical complex networks . specifically , the networks studied include the following : ( 1 ) a random network consisting of 300 nodes . the average degree is 10 . ( 2 ) a small - world network , which is obtained by rewinding 20 links in a regular network consisting of 100 nodes and each node has 10 nearest neighbor connections @xcite . ( 3 ) a modular network consisting of 100 nodes which is evenly divided into 5 modules . inside each module nodes are fully connected . any two nodes in different modules have probability @xmath54 to connect each other . ( 4 ) as a special case of complex networks , a regular network consisting of 100 nodes and each node has 10 nearest neighbor connections . the pathes towards global cs on these typical complex networks have been characterized as for scale - free networks . here , we only show one example in the case of random network . in fig . 6 , the bifurcation route to global cs for 300 identical lorenz oscillators on a random network is characterized . comparing fig . 6(a ) with fig . 1 , we can see that the development of gs on random network is different from the situation on scale - free network . mainly , most nodes are entrained at approximately the same coupling strength . this is consistent with the fact that random network has approximately homogeneous degrees and thus no extraordinary hubs exist as in scale - free networks . although the development of gs on different networks depends on the network topologies , qualitatively , the path toward global gs is the same as shown in fig . 1 and fig . 6 . in fact , for other types of complex networks , we have similar observations , as in the scale - free network and random network . based on these studies , we can draw the following conclusions . ( 1 ) for coupled identical chaotic oscillators , whether the network is homogeneous or heterogeneous , a general bifurcation path towards global cs exists with the increase of coupling strength , i.e , non - synchronization @xmath2 pgs @xmath2 gs @xmath2 pcs @xmath2 cs . ( 2 ) usually , global gs needs relatively smaller coupling strength to achieve compared with global cs . sometimes , the latter requires ten times larger coupling strength to achieve than the former . this shows that gs is a weaker synchronization form on complex networks compared with cs , and is consistent with the observation in coupled two - oscillator systems . ( 3 ) typically , the coupled system undergoes a _ dynamic and oscillatory _ transition to achieve global synchronization , such as global gs or global cs . during the pgs or pcs regime , oscillators may frequently join in and escape from the giant synchronous cluster . this leads to that the size of synchronous giant cluster increases in an oscillatory manner with the increase of the coupling strength . in some circumstances , the transition regimes cover quite wide range of coupling strength , reflecting that there exists intense interaction among nodes in achieving global synchronization on complex networks . so far , most of the works on synchronization in complex networks study the coupled identical oscillators , or coupled simple phase oscillators of kuramoto type . however , in real - world networks , the interacting components are in principle different from each other . for example , in neuron networks any two pair of neurons can not be exactly the same . this raises an important question : is it possible for complex networks with different local dynamics achieve gs , or certain coherence ? if so , how network topologies affect the gs ? in this section , we will investigate synchronization of coupled nonidentical oscillators on complex networks . our particular interest is to reveal how the development of gs on networks is dominated by the interplay between network topology and the local dynamics . as in the case of coupled system with two different oscillators , we find that cs is generally forbidden for coupled nonidentical oscillators on complex networks . instead , we observe that gs generally occurs in such coupled systems on networks . in our study , the lorenz oscillators , i.e. , eq . ( [ eq : lorenz ] ) , are made different on network by randomly setting parameter @xmath35 in the interval @xmath55 $ ] . by applying auxiliary system approach , we can plot the color - map of the local synchronization distance @xmath44 in the two - dimensional parameter plane @xmath56 . as an example , fig . 7 illustrates the development of gs for lorenz oscillators with parameter mismatches on a scale - free network . it is shown that with the increase of coupling strength , nodes in the network are entrained first from the hubs and gradually spread to other nodes . when the coupling strength is large enough , all nodes in the network are entrained indicating global gs has been achieved . this situation is similar to the case of coupled identical oscillators due to the fact that the parameter mismatches here are small . moreover , for coupled nonidentical oscillators , the dynamical feature of the route toward global gs is essentially the same as in the case of identical oscillators . the interplay between network dynamics and network topology is the key to understand collective behaviors on complex networks . previously , many works have revealed how network topology affects the network synchronizability under the setting of coupled identical oscillators . for nonidentical oscillators , there are two factors affecting the development of gs on complex networks , i.e. , network topology and heterogeneity in the local dynamics . the effect of heterogeneity in the local dynamics , if they are chaotic , on the development of gs remains to be a challenging problem . in the two preceding sections , we have shown that for scale - free networks , gs typically starts from the hubs , and then spreads to others nodes with relatively smaller degrees . in this case , heterogeneity in the degree distribution appears to be the main factor governing the development of gs . an interesting issue is then that , for networks with homogeneous degree distributions , such as regular , small - world , or certain modular networks , how does the heterogeneity in the local dynamics affect gs ? in the following , we study this question numerically . we first consider a regular network of @xmath57 nodes . each node in the network has @xmath58 connections to its nearest neighbors . as a special case , this is an exactly homogeneous network with a regular degree sequence . the local dynamics are those of lorenz chaotic oscillators with different parameter @xmath35 randomly distributed in the interval [ 28 , 38 ] for this example and for the following two examples . figure 8 shows the color - maps of the local synchronization distances @xmath44 and @xmath59 , where @xmath37 and @xmath38 are the node indices arranged according to the decreasing node degree and decreasing values of parameter @xmath35 , respectively . for the lorenz oscillator under the parameter setting in our study , we find that the larger the value of parameter @xmath42 , the larger the largest lyapunov exponent of the chaotic attractor . thus , in fig . 8 , the index @xmath38 actually corresponds to the decreasing value of the largest lyapunov exponent . specifically , larger value of @xmath38 corresponds to a small value of @xmath35 so that the corresponding local dynamics is less chaotic in the sense that its largest lyapunov exponent has a relatively smaller value . comparing figs . 8(a ) with 8(b ) , we find that , when transient behaviors are disregarded , nodes whose dynamics are less chaotic require smaller value of the coupling strength to be entrained with other nodes in the network . entrainment of more chaotic nodes requires stronger coupling . thus the development of gs on regular network is dominated by the local dynamics . we next consider the occurrence and development of gs on small - world networks . a representative example is shown in fig . we see from fig . 9(a ) that there is no apparent synchronization sequence of nodes according to the degree index @xmath37 . however , as can be seen from fig . 9(b ) , gs starts from nodes with smaller values of the largest lyapunov exponent . this demonstrates that for a small - world network , heterogeneity in the local dynamics plays a dominant role in the development of gs , which is similar to the situation in regular networks . lastly , we investigate gs in a type of modular network . a modular network is characterized by a number of sparsely connected subnetworks , each with dense internal connections . the modular structure has been identified in biological , social , and certain technological networks @xcite . for such a network , synchronization within each individual cluster can usually be achieved readily due to the dense internal connections , so the occurrence of global synchronization is particularly interesting @xcite . an example of the development of gs on modular network is shown in fig . 10 . we see that global gs can be achieved despite the sparse inter - cluster connections . an interesting phenomenon is that , the development of gs does not seem to strongly depend on @xmath37 or @xmath38 . nevertheless , comparatively , it still can be found that nodes with less chaotic local dynamics are easier to be entrained with smaller coupling strength . this can be explained by noting that the degree distribution is approximately homogeneous , as a result of the sparse inter - cluster connectivity . in addition , nodes with different values of @xmath42 are uniformly distributed in different synchronous clusters . as a result , the effect of heterogeneity in the local dynamics is suppressed to some extent . based on the above studies , we can conclude that the development of gs on complex networks is the result of interaction between network topology and local dynamics when they both have heterogeneity . for heterogenous networks , such as scale - free networks , the network topology plays a leading role ; while for approximately homogeneous networks , such as small - world networks and modular networks , the local dynamics is the dominant factor affecting the development and spreading of gs on networks . we have seen that the hub nodes behave as `` seeds '' to develop gs in scale - free networks . the phenomenon that synchronization starts from the hub nodes in a heterogeneous network is in fact quite general , which holds for both gs and cs . qualitatively , this can be understood by regarding the interactions among nodes as a kind of `` noisy '' driving within the network . in particular , since the dynamics of each individual node is chaotic , the coupling term in eq . ( [ eq : network_equation ] ) for node @xmath6 is also chaotic , which can be regarded effectively as correlated noisy driving . it has been known that correlated noise can induce and/or enhance synchronization @xcite , ps @xcite , and gs as well @xcite . that is , correlated noise of larger amplitude can cause oscillators under such driving to be better synchronized . for a heterogeneous network , hub nodes have significantly more links than other nodes . since the drivings from different nodes are not entirely uncorrelated , the equivalent driving force to the hub nodes are generally stronger , facilitating their synchronization with other nodes in the network . however , in a recent publication studying gs on networks @xcite , it is reported that only for a kind of special scale - free network with tree - like structure , gs is observed to develop from the hubs and then gradually spread to other nodes in the network ; while for usual scale - free networks , the heterogeneity of network seems to have little effect on the development of gs in the network . in attempting to find the reason that leads to different observations in the present work and ref . @xcite , we notice that a different coupling strategy is used in ref . take a coupled one - dimensional map network as an example , the coupled system in ref . @xcite is : @xmath60 where @xmath61 is the state variable of node @xmath6 at time step @xmath62 , @xmath63 is the local map , and @xmath64 is the degree of node @xmath6 . this coupling scheme with each node in the network driven by the local mean field is essentially the same as the following normalized coupling strategy : @xmath65 where notations are the same as that in eq . ( [ eq : network_equation ] ) and eq . ( [ eq : hung ] ) . in this coupling scheme , the coupling strength for each node @xmath6 is normalized by its degree . therefore , the effective coupling strength each node received from its network neighbors is of the same order of magnitude , regardless it is a hub or a node with very small degree . obviously , due to the normalization , the effect of network topology on the development of gs has been almost suppressed . as a consequence , there should be no distinct difference among the coupling strength when nodes in the network achieve gs . quantitatively , we can find that in order to obtain the same effective coupling strength , system ( [ eq : network_equation2 ] ) should require @xmath66 times larger coupling strength than system ( [ eq : network_equation ] ) . here , @xmath66 denotes the average degree of the network . thus on an average , it can be expected that system ( [ eq : network_equation2 ] ) needs @xmath67 times larger coupling strength to achieve gs in network compared with system ( [ eq : network_equation ] ) . the above heuristical considerations can be further analytically explained . by examining the coupled system ( [ eq : network_equation ] ) and its corresponding auxiliary system ( [ eq : network_auxiliary_equation ] ) . letting @xmath68 and subtracting eq . ( [ eq : network_equation ] ) from eq . ( [ eq : network_auxiliary_equation ] ) , we obtain @xmath69\cdot \delta { \bf x}_i,\end{aligned}\ ] ] where @xmath70 is the @xmath71 identity matrix with @xmath72 the dimension of local dynamics . we see that , for node @xmath6 , the effective coupling strength is proportional to its degree @xmath64 . for a fixed value of @xmath14 , the coupling between a hub node and its counterpart in the auxiliary system can be significantly larger than the coupling for nodes with smaller degree , leading to `` earlier '' synchronization between the hub nodes in the original and in the auxiliary systems . therefore , for usual linear coupled oscillator system , the general observation is that , in a complex network with heterogeneous degree distribution , the set of hub nodes provides a skeleton around which synchronization is developed . the above analysis is also suitable for coupled system ( [ eq : network_equation2 ] ) . in this case , there will be no factor @xmath64 as in eq . ( [ eq : delta_x_evolution ] ) . this implies that the effect of network topology on the development of gs in networks is eliminated . in the following of this section , we further present two examples where the oscillators in a scale - free network are interacted with the normalized coupling strategy . the first example is coupled lorenz oscillators described by eq . ( [ eq : network_equation2 ] ) , and the second example is the following coupled logistic maps : @xmath73.\ ] ] for both cases , the local dynamics are identical , i.e. , @xmath35=28 for all lorenz oscillators and @xmath74 for all logistic maps . figure 11 illustrates the development of gs for the above two systems on a scale - free network . the effect of different coupling strategies on the development of gs can be verified by comparing fig . 11 with fig . 1 . for system ( [ eq : network_equation ] ) , gs first occurs on the hubs and then gradually spread to other nodes ; while for systems ( [ eq : network_equation2 ] ) and ( [ eq : logistic_map ] ) , gs almost simultaneously takes place on the hubs and the other nodes . actually , there is no significant difference among all nodes on networks . by examining fig . 1 and fig . 11(a ) , it is found that the giant gs cluster forms at @xmath75 and @xmath76 , respectively , in the two cases . in particular , the latter is approximately four times larger than the former . note that the scale - free networks we used in fig . 1 and fig . 11(a ) are ba model with @xmath77 , so the average node degree @xmath78=4 . this is consistent with the theoretical analysis above . in sec . iv , we have studied how network topology competes with local dynamics to dominate the development of gs in complex networks when there are heterogeneity in both network topology and local dynamics . in fact , for coupled nonidentical oscillators on complex networks , competition may also occur among local dynamics if the parameter mismatches of local dynamics are large enough , or the local dynamics are physically different . by adopting the normalized coupling strategy , we can conveniently investigate this issue . in the following , we present two examples showing how gs is developed for coupled hybrid oscillator system in networks . by hybrid we mean that the oscillators in the network can be classified into different types which are parametrically different or physically different . in the first example , we consider a system coupled with hybrid lorenz oscillators . to be specific , we arbitrarily select 5% lorenz oscillators in the network and make their parameter @xmath35 be randomly distributed in the interval [ 30 , 40 ] . for the rest oscillators , they are the same with @xmath35=28 , which are significantly smaller than that of the 5% oscillators . in fig . 12 , the development of gs for such a hybrid system is illustrated on a scale - free network . it is seen that most oscillators in the network are entrained at @xmath79 . however , a small number of oscillators require significantly larger coupling strength to be entrained . a careful examination of the locations of these nodes reveals that they just correspond to the 5% oscillators with larger @xmath42 values . in the second example , the coupled hybrid system consists of two kinds of oscillators which have different dynamical equations . similar to the first example , 5% nodes are randomly selected to be the identical rossler oscillators , while the rest 95% percent nodes are the identical lorenz oscillators with @xmath35=28 . the dynamical equations of the rossler oscillator read : @xmath80^t,\ ] ] where @xmath34 are the state variables of the rossler oscillator . in fig 13 , we plot the local distance of gs vs the coupling strength for all nodes in a scale - free network . it is clearly shown gs can be achieved for such a coupled hybrid system . the development of gs in network has two distinct stages : the rossler oscillators are much easier to be entrained compared with the lorenz oscillators . the latter needs quite larger coupling strength to achieve gs . note that the largest lyapunov exponent of the lorenz oscillator is proportional to its @xmath42 value , and the rossler oscillator has much smaller lyapunov exponent than that of the lorenz oscillator , we can conclude from the above two examples that gs usually develops from the nodes in the network where the local dynamics are less chaotic in the sense that the largest lyapunov exponents of the local dynamics have relatively smaller values . although our numerical simulations are carried out on a scale - free network , we believe this conclusion can still hold for other network topologies when the normalized coupling strategy is applied . in summary , we have investigated the occurrence and development of gs on various complex networks , including scale - free networks , small - world networks , random networks and modular networks . it is shown that gs generally takes place in such networks for both coupled identical oscillators and nonidentical oscillators . for coupled identical oscillators , we reveal a general bifurcation path to global cs , i.e. , non - synchronization @xmath2 pgs @xmath2 gs @xmath2 pcs @xmath2 cs . prior to the onset of global gs ( cs ) , there exists a coupling regime of partial gs ( cs ) characterized by the occurrence of a cluster of synchronous nodes . in the transient regimes , the size of the synchronous cluster usually oscillates with the increase of coupling strength , showing a dynamic scenario when the coupled system is approaching global synchronization . we find that the development of gs on complex networks depends on both network topology and local dynamics . moreover , the specific coupling strategy also plays an important role during the evolution of synchronization . we show that under the usual linear dissipative coupling scheme , for heterogeneous networks , gs generally starts from a small number of hub nodes , and then spread to the rest nodes in the network ; while for homogeneous networks , gs usually starts from the nodes whose local dynamics are less chaotic in the sense that the largest lyapunov exponents have relatively smaller values . we further show that the effect of network topology on the development of gs can be suppressed if the coupling strengths of nodes in the network are normalized . under such coupling scheme , the development of gs is essentially determined by the competition among different local dynamics . we demonstrate that gs can also occur in coupled systems with hybrid oscillators on complex networks , and the development of gs has distinct stages corresponding to physically different local dynamics . we note that , while there are extensive works on synchronization in complex networks , prior to this work the bifurcation path toward global cs via gs , the dynamical features of the development of gs , especially the interplay between the network topology and local dynamics on typical complex networks have not been investigated . our work reveals that on complex networks , coupled oscillators may present fundamentally different synchronization regimes which deserves further study . this work is supported by temasek laboratories at national university of singapore . l. m. pecora , t. l. carroll , g. a. johnson , d. j. mar , and j. f. heagy , chaos * 7 * , 520 ( 1997 ) ; a. pikovsky , m. rosenblum , and j. kurths , in _ synchronization : a universal concept in nonlinear science _ ( cambridge university press , cambridge , 2001 ) ; s. boccaletti , j. kurths , g. osipov , d. l. valladares , and c. s. zhou , phys . rep . * 366 * , 1 ( 2002 ) . l. m. pecora and t. l. carroll , phys . lett . * 64 * , 821 ( 1990 ) ; k. pyragas , phys . a * 181 * , 203 ( 1993 ) . l. kocarev and u. parlitz , phys . lett . * 77 * , 2206 ( 1996 ) . y. jiang and p. parmananda , phys . e * 57 * , 4135 ( 1998 ) . n. f. rulkov , m. m. sushchik , l. s. tsimring , and h. d. i. abarbane , phys . e. * 51 * , 980 ( 1995 ) ; l. kocarev and u. parlitz , phys . 76 * , 1816 ( 1996 ) ; k. pyragas , phys . e * 54 * , r4508 ( 1996 ) ; j. m. gonzlez - miranda , _ ibid _ , * 65 * , 047202 , ( 2002 ) ; b. blasius , e. montbri , and j. kurths , _ ibid _ , * 67 * , 035204(r ) ( 2003 ) ; z. liu , y .- c . lai , and m. a. matas , _ ibid _ , * 67 * , 045203(r ) ( 2003 ) ; w .- h . kye , d .- s . lee , s. rim , c .- m . kim , and y .- j . park , _ ibid _ , * 68 * , 025201(r ) ( 2003 ) ; a. uchida , r. mcallister , r. meucci , and r. roy , phys . * 91 * , 174101 ( 2003 ) ; s. guan , c .- h . lai , and g. w. wei , phys . e * 71 * , 036209 ( 2005 ) . m. g. rosenblum , a. s. pikovsky , and j. kurths , phys . lett . * 76 * , 1804 ( 1996 ) ; e. rosa , jr . , e. ott and m. h. hess , ibid , * 80 * , 1642 ( 1998 ) ; s. guan , c .- h . lai , and g. w. wei , phys . e * 72 * , 16205 ( 2005 ) ; s. guan , x. wang , and c .- h . lai , phys . a * 354 * , 298 ( 2006 ) . l. f. lago - fernandez , r. huerta , f. corbacho , and j. a. siguenza , phys . . lett . * 84 * , 2758 ( 2000 ) ; p. m. gade and c .- k . hu , phys . e * 62 * , 6409 ( 2000 ) ; x. f. wang and g. chen , int . j. of bifur . chaos * 12 * , 187 ( 2002 ) ; ieee trans . , part i * 49 * , 54 ( 2002 ) . a. e. motter , c. zhou , and j. kurths , europhys . lett . * 69 * , 334 ( 2005 ) , phys . e * 71 * , 016116 ( 2005 ) ; c. zhou , a. e. motter , and j. kurths , phys . rev . lett . * 96 * , 034101 ( 2006 ) ; m. chavez , d .- u . hwang , a. amann , h. g. e. hentschel , and s. boccaletti , phys . lett . * 94 * , 218701 ( 2005 ) . d. j. watts , p. s. dodds , and m. e. j. newman , science * 296 * , 1302 ( 2002 ) ; r. milo , s. shen - orr , s. itzkovitz , n. kashtan , d. chklovskii , and u. alon , science * 298 * , 824 ( 2002 ) ; a. e. motter , t. nishikawa , and y .- c . lai , phys . e * 68 * , 036105 ( 2003 ) . see , for example , h. herzel and j. freund , phys . e * 52 * , 3238 ( 1995 ) ; p. m. gade and c. basu , phys . a * 217 * , 21 ( 1996 ) ; e. sanchez , m. a. matias , and v. perez - munuzuri , phys . e * 56 * , 4068 ( 1997 ) ; c .- h . lai and c. zhou , europhys . * 43 * , 376 ( 1998 ) ; l. kocarev and z. tasev , phys . e * 65 * , 046215 ( 2002 ) .
in this paper , we investigate the development of generalized synchronization ( gs ) on typical complex networks , such as scale - free networks , small - world networks , random networks and modular networks . by adopting the auxiliary - system approach to networks , we show that gs can take place in oscillator networks with both heterogeneous and homogeneous degree distribution , regardless of whether the coupled chaotic oscillators are identical or nonidentical . for coupled identical oscillators on networks , we find that there exists a general bifurcation path from initial non - synchronization to final global complete synchronization ( cs ) via gs as the coupling strength is increased . for coupled nonidentical oscillators on networks , we further reveal how network topology competes with the local dynamics to dominate the development of gs on networks . especially , we analyze how different coupling strategies affect the development of gs on complex networks . our findings provide a further understanding for the occurrence and development of collective behavior in complex networks .
supernova ( sn ) 2006gy is an extremely luminous sn found by @xcite near the nucleus of the host galaxy ngc 1260 [ @xmath10 mag ] . its absolute magnitude reaches @xmath2 @xmath11 mag ( ofek et al . 2007 ; smith et al . 2007 ) , being more luminous than typical type ia supernovae by a factor of @xmath12 . the sn was referred to as `` the most luminous sn '' at that time . in addition to the high peak luminosity , the light curve ( lc ) of sn 2006gy evolves very slowly , peaking at @xmath13 70 days after the explosion ( hereafter , @xmath14 means time after the explosion ) . the total radiation energy exceeds @xmath15 ergs , i.e. , comparable to the kinetic energy of typical sne . the optical spectra of sn 2006gy show type iin - like features , i.e. , narrow emissions of h balmer lines . for a model of sn 2006gy , a ` type iia ' scenario ( hybrid of type iin and ia sn , like sne 2002ic and 1997cy ; deng et al . 2004 ) , i.e. , type ia sn in a dense circumstellar medium ( csm ) , has been suggested . however , the radiation energy of sn 2006gy is comparable to ( or overwhelming ) the kinetic energy of type ia sne so that this scenario seems unlikely ( ofek et al . 2007 ; smith et al . the progenitor of sn 2006gy should be a massive star . the source of the huge optical luminosity is under debate . possible mechanisms are ( i ) thermal emission from hydrogen recombination front , ( ii ) interaction between csm and sn ejecta , and ( iii ) radioactive decay of @xmath4ni . ofek et al . ( 2007 ) suggested that the luminosity comes from the csm interaction , as in type iin / iia sne . on the contrary , smith et al . ( 2007 ) suggested that radioactive decay is the primal ( or least problematic ) heating source , because only weak x - rays were detected from sn 2006gy ( suggesting csm interaction was not strong ) and the observed expansion velocity of hydrogen envelope was unusually slow ( suggesting too small emitting radius for thermal radiation ) . agnoletto et al . ( 2009 ) gave an idea that the massive csm was clumpy and both the csm interaction and the radioactive decay contributed to the enormous luminosity . furthermore , smith et al . ( 2007 ) suggested that sn 2006gy is a pair - instability sn ( pisn ) because a huge amount of @xmath4ni ( @xmath16 ) is required to explain the peak luminosity . however , nomoto et al . ( 2007 ) shows that the lc of the pisn model evolves too slowly to reproduce the observed lc . umeda & nomoto ( 2008 ) explored a possibility that core - collapse sne could synthesize such a large amount of @xmath4ni . woosley , blinnikov & heger ( 2007 ) showed that the collision between the shells ejected by pulsational pair instability in massive stars initially having @xmath17 could reproduce the peak luminosity of sn 2006gy . smith & mccray ( 2007 ) also presented a simple `` shell - shocked '' model involving shock - deposited energy as the heat source . in these two models , the shell is so thick that x - rays are absorbed . portegies zwart & van den heuvel ( 2007 ) suggested that the sn 2006gy was an outcome of merging between two massive stars in a dense and young cluster , being consistent with presence of a large amount of hydrogen . recently , smith et al . ( 2008a ) presented results of late phase observations ( @xmath18@xmath19 days ) . the sn was clearly seen in nir images obtained with the adaptive optics on the keck ii telescope , while any optical observation failed to detect the sn component even for h@xmath7 emission . the nir flux would be attributed to newly - formed dust particles or an ir echo . agnoletto et al . ( 2009 ) also successfully detected this sn in @xmath20 band at @xmath21@xmath22 days , but failed to detect the sn in the optical bands at @xmath23@xmath24 days . in this paper , we report on the results of our late time observation ( @xmath25 days ) of sn 2006gy with the subaru telescope . the late phase luminosity and spectra may give clues to distinguish the underlying scenarios as also mentioned by smith et al . ( 2008a ) . in [ sec : obs ] , we describe our observations and data reduction . in [ sec : lc ] , the lc of sn 2006gy is compared with those of other sne and results of model calculations . in [ sec : spec ] , the spectra of sn 2006gy are analyzed . features in the late phase spectrum is further investigated by comparing it with spectra of sne of various types ( [ sec : discussion ] ) . finally we give conclusions in [ sec : conclusions ] . the observations were carried out with the faint object camera and spectrograph ( focas , kashikawa et al . focas is an optical versatile instrument attached to the cassegrain focus of the 8.2-m subaru telescope . the images were recorded on two mit / ll ccds ( 2k@xmath264k , 15 @xmath27 m pixel@xmath6 ) . to minimize the harmful effect due to atmospheric dispersion in observation at low altitude , we used the atmospheric dispersion corrector ( adc ) in front of focas , which reduces the chromatic elongation to less than 01 within wavelengths 350011000 at airmass @xmath28 . we took all images at airmass less than 1.9 ( table [ tbl : obs ] ) , which promises that the effect was negligible in our observations . for spectroscopy , we obtained high resolution spectra ( @xmath29 ) on 2006 dec 25.4 ( @xmath30 days ) and 2007 jan 24.4 ( @xmath31 days ) and a low resolution spectrum ( @xmath32 ) on 2007 sep 18.5 ( @xmath25 days ) . for the high resolution spectroscopy , we used a @xmath33 width slit , a 665 lines mm@xmath6 volume - phase holographic grism and sy47 order - cut filter , which gives a wavelength coverage of 53007700 and a spectral resolution of @xmath34 . for the low resolution one , we used a @xmath35 width offset - slit , a 300 lines mm@xmath6 blue grism and sy47 filter , giving a wavelength coverage of 48009000 and a spectral resolution of @xmath36 . the total exposure time was 1200 s on each night . the direction of the slit was set to position angle ( pa ) @xmath37 to take spectra of both the nucleus of the host galaxy ngc 1260 and the sn simultaneously . flux calibration and the correction for atmospheric absorption bands were performed using spectra of a spectrophotometric standard star obtained on the same night ; bd @xmath38 in 2006 dec , feige 34 in 2007 jan and g191b2b in 2007 sep . the log of observation is shown in table [ tbl : obs ] . for photometry , we took @xmath39 and @xmath3 band images just before the spectroscopy . the exposure times of @xmath39 and @xmath3 bands are 3 s and 4 s in 2006 dec , 5 s and 5 s in 2007 jan and 10 s and 10 s in 2007 sep , respectively . we used @xmath40 on - chip binning in the ccd readout and the resultant scale was 0208 per a binned - pixel . the magnitude of the sn is calibrated with landolt field stars ( landolt 1992 ) . since the sn was very close to the bright nucleus of the host galaxy ( @xmath41 ) , the subtraction of the background component should be carefully performed , especially for late phase data . in the earlier phase data , the sn was sufficiently bright and the artificial error originated in the adopted method for the background subtraction is not significant . for photometry , we subtracted psf - matched 394 days image in the same band as the background , and then performed psf - photometry . although this method gives an over - subtraction due to a non - negligible sn component at @xmath42 394 days , the error should be small ( @xmath43 mag ) because the sn faded by about 4 mag between @xmath42 127 days and 394 days . the photometric error was derived from the root sum square of the magnitude transformation error using the landolt field stars and the photon noises of the sn and the background . for spectroscopy , we assumed that the two - dimensional ( 2d ) spectral image of the host galaxy is symmetric across the nucleus , and simply subtracted the mirrored spectral image across the nucleus . for the late phase data , we should mention at first that the sn component was successfully detected in both imaging ( fig . [ fig : flipsubim ] ) and spectroscopic ( fig . [ fig:2dspecim]a - d ) data , although smith et al . ( 2008a ) and agnoletto et al . ( 2009 ) reported their null detection in optical data between @xmath44 and @xmath45 days . we assume that the better seeing condition at our observation enabled us to distinguish the sn component from the host galaxy nucleus . however , it is still a difficult task to derive the intrinsic sn component because the background galaxy has inhomogeneous structures . for photometry on @xmath42 394 days , we assumed a symmetry for the brightness distribution of the host galaxy with respect to its polar axis ( pa@xmath46 ) and adopted the mirrored image as the background for the lack of better alternatives . in the background - subtracted image , the sn is seen nearly as a point - like source embedded in a diffuse elongated component ( fig [ fig : flipsubim ] ) . the psf - photometry for the sn and nearby two stars ( c1 and c2 in fig . [ fig : flipsubim ] ) provided @xmath47 and @xmath48 for the sn , @xmath49 and @xmath50 for c1 and @xmath51 and @xmath52 for c2 . our preliminary aperture - photometry with several parameter sets ( radii of stellar aperture and sky annulus ) gave somewhat scattered values for the sn magnitudes , @xmath53 and @xmath54 , because of the inhomogeneous background . to check the reliability of the uncertainty , we added an artificial star having the same brightness at several positions near the sn and performed photometry of it . the result of this test was consistent with above photometry and we adopt @xmath53 and @xmath54 as the sn magnitude . it is noted that these magnitudes might be contaminated by unresolved background components , e.g. , regions and/or stellar clusters in the host galaxy , and thus they should be considered as the brighter limits . in the spectroscopic data , we estimated the background component using the neighboring regions on both sides of the sn position , bg1 and bg2 along the slit ( fig . [ fig : slitpos ] and [ fig:2dspecim ] ) . since the simple mean of bg1 and bg2 spectra clearly leads to over - estimation of the background component ( i.e. , spatial distribution of the background component is no longer linear along the slit ) , we scaled the bg1 and bg2 spectra to match the sn flux at apparently line - free region , 78147856 rest wavelength before the averaging . the scaling factors and the scaled spectra of bg1 , bg2 and some other regions are shown in figures [ fig : specbackg]a . four background spectra ( bg1 , bg2 , galaxy core and mirror position ) resemble to one another and they are well explained by a slightly reddened bulge spectrum ( bruzual & charlot 2003 ) . from these spectra ( including 2d spectral images in figure [ fig:2dspecim ] ) we recognize the existence of [ ] 7155 , [ ] 7300 and ir triplet emission lines in the sn spectrum . the enlarged spectra around h@xmath7 line ( fig . [ fig : specbackg]b ) also implies the presence of an intermediate - width ( 1530 @xmath55 7001400 km s@xmath6 ) h@xmath7 emission line . they are discussed later . finally , we adopted the mean of the scaled bg1 and bg2 spectra as the background component , and then subtracted it from the sn position spectrum to derive the pure sn spectrum . this method leads to complete elimination of the continuum component of the sn , at least , around the line - free regions . the resultant sn spectrum provides the faint - end limit of the magnitude of the sn as @xmath56 and @xmath57 ( as far as we ignore the contamination by the possible unresolved component as mentioned above ) , which are consistent with the results of our photometry . it is noted that there is a non - negligible difference between the normalized bg1 and bg2 spectra , especially at bluer wavelengths , which may lead to a large uncertainty for the resultant sn flux at @xmath25 days up to @xmath58% ( fig . [ fig : specobs2 ] ) . this uncertainty is larger than the error of estimated extinction toward the sn ( see [ sec : spec ] ) . in this paper we mainly discuss the emission line components at longer than 5800 for the late phase spectrum and therefore this uncertainty does not affect our discussion . figure [ fig : lcabs ] shows the @xmath3 band light curve ( lc ) of sn 2006gy constructed from the published data ( smith et al . 2007 ; agnoletto et al . 2009 ) and our subaru ones . it is compared with bright type ic sn 1998bw ( patat et al . 2001 ) and type iia sn 2002ic ( deng et al . 2004 ) and type iin sn 1999el ( di carlo et al . 2002 ) . there exists a large variety among the lcs of type iia and iin sne . the lc of sn iia 2002ic declines very slowly . the slope is flatter than that the lc powered by the @xmath4co @xmath59 @xmath4fe decay with fully trapped @xmath60-rays . this behavior is also seen in other sne , e.g. , sn iia 1997cy and sn iin 1988z ( e.g. , germany et al . 2000 ; turatto et al . 1993 ; turatto et al . on the other hand , the lc of sn 1999el declines faster than sne 2002ic and 1998bw . similar behavior is also seen in the lc of type iin sn 1998s ( fassia et al . 2000 ) . in another type iin sn 1994w , a sudden decline of the lc was observed after @xmath61 days plateau ( sollerman , cumming & lundqvist 1998 , chugai et al . the main heating source of type iin and iia sne is thought to be the interaction between sn ejecta and csm ( e.g. , turatto et al . 1993 ; chugai et al . 2004 ) , and this heterogeneity in the lc may reflect the structure , the mass and the kinetic energy of the sn ejecta as well as the structure ( density and its slope ) of csm . in contrast to type iin sne , the lcs of sne where @xmath4ni is the dominant heating source show similar shapes . the lc declines faster than the @xmath4co decay line because a fraction of @xmath60-rays escape without interacting with the sn ejecta . the deviation from the @xmath4co decay line gradually becomes larger because the expanding sn ejecta gets thinner against @xmath60-rays . the lc of type ic sn 1998bw is fully explained by this scenario ( patat et al . 2001 ; maeda et al . 2003 ) . the early lc of sn 2006gy tends to be leveled around @xmath62 days . our subaru data shows that the lc fades by @xmath63 mag from @xmath64 to @xmath65 days , which is consistent with the results of nir photometry at @xmath66 and 468 days ( smith et al . the decline rate during this period is clearly faster than that of sn 2002ic and the @xmath4co decay line , while it seems comparable to that of sn 1998bw . in other words , the lc of sn 2006gy does not show any particularly peculiar behavior that is not explained by the @xmath4co decay . however , we can not exclude the possibility that the lc is almost leveled at @xmath67 days and suddenly declines just before @xmath68 days . in the right panel of figure [ fig : lcabs ] , the @xmath4ni@xmath69 @xmath4co decay model presented by nomoto et al . ( 2007 ) is also shown . they computed a bolometric lc and showed that the model with @xmath70 , @xmath71 erg and @xmath72 explains the observed lc up to @xmath64 days . here @xmath73 , @xmath74 and @xmath75 are the mass of the ejecta , the kinetic energy of the ejecta , and the mass of @xmath4ni ejected by the explosion , respectively . the model lc is in reasonably good agreement with our late phase observation . if we assume that the radioactive decay is the main energy source of emission at @xmath25 days , @xmath76 is needed to explain the @xmath3 band flux . although the model lc for sn iia 2002ic ( nomoto et al . 2005 ) is not consistent with that of sn 2006gy , we can not preclude the csm interaction model because of the large diversity of lcs in type iin / iia sne . we note that the model by woosley et al . ( 2007 ) is also consistent with our observation . however , the diffusion model by smith & mccray ( 2007 ) , predicting a fast decline after the maximum ( @xmath77 mag at @xmath78 days ) , is inconsistent with our data . also , the lcs of the pisn model ( blue dashed line ) is not consistent with the observed lc . the spectra of sn 2006gy are shown in figure [ fig : specobs ] , and closed - up profiles of some lines are shown in figure [ fig : specvel ] . the redshift @xmath79 , derived from the diffuse h@xmath7 emission line of the host galaxy ( cf . @xmath80 derived at the nucleus ) , has been corrected for . we assumed the total extinction of @xmath81 ( or @xmath82 ) toward the sn , which is the sum of the extinction in the milky way [ @xmath83 ; schlegel , finkbeiner , & davis 1998 ] and that within the host galaxy [ @xmath84 , derived from @xmath85 and @xmath86 ; smith et al . 2007 , schlegel et al . 1998 ] . the equivalent width ( ew ) of the galactic d@xmath87d@xmath88 absorption lines in our high resolution spectrum ( ew@xmath89 ) is almost consistent with @xmath83 using an empirical relation of @xmath90( ) suggested by turatto , benetti , & cappellaro ( 2003 ) . although the ew of the host galaxy d line in the same spectrum suggests a larger extinction within the host galaxy ( @xmath91 ) , it should be partly circumstellar origin because some of other spectral features ( h@xmath7 and ) are accompanied with narrow absorption components in their p - cyg type profiles . our late time spectrum suffers from large uncertainty in the ew of d line because of the faint sn spectrum and the uncertain continuum level ( see [ sec : backsub ] ) . like other type - iin sne , our early phase spectra of sn 2006gy are characterized by a strong h@xmath7 emission line as well as many , and lines . as shown in figure [ fig : specvel ] , the h@xmath7 emission line in the spectra at @xmath30 days has a narrow p - cyg type component near the peak , of which the marginally resolved emission line has a width of @xmath92 km s@xmath6 fwhm , and a broad wing component exceeding @xmath93 km s@xmath6 . the broad component is clearly asymmetric , being brighter in the red wing than in the blue . smith et al . ( 2007 ) took a high - resolution spectrum at @xmath94 days ( @xmath95 ) and suggested that the asymmetric profile is produced by broad absorption component at the blue side . this component had a sharp blue edge at @xmath96 km s@xmath6 , which is considered to be an effective speed of the sn blast wave ( smith et al . this asymmetric profile was seen through @xmath31 days , although the absorption component became less significant . for some features , e.g. , 5535 and d at @xmath30 days , broad blue - shifted absorption components up to @xmath97 km s@xmath6 are also significant ( fig . [ fig : specobs ] ) . the narrow absorption component of the h@xmath7 peak considerably evolved from @xmath94 to 157 days . it splits into possible double components by 127 days , and the whole width between both edges of the absorption feature increased from @xmath98 km s@xmath6 to @xmath99 km s@xmath6 . these suggest an ongoing interaction between the ejecta and the dense csm . in the late phase spectrum ( fig . [ fig : specobs2 ] ) , we can see several emission lines , mainly in red wavelengths , including [ ] 7291 , 7323 ( 7302 in average ) , ir triplet and [ ] 7155 , which are listed in table [ tbl : unident_line ] . on the other hand , the bluer spectrum is likely to be dominated by continuum component and/or heavily blended emission lines . a broad absorption trough ( up to @xmath100 km s@xmath6 ) possibly belonging to d is significant in this region . additionally , as mentioned in [ sec : backsub ] , there is a possible h@xmath7 emission line having an intermediate - width ( 7001400 km s@xmath6 ) . if the flux excess at the minimum point ( 6574 ) between h@xmath7 and [ ] 6584 lines is actually due to the intermediate - width h@xmath7 component ( fig . [ fig : specbackg]b ) , its observed flux is derived to be @xmath101 erg s@xmath6 @xmath102 @xmath6 above the continuum level of @xmath103 erg s@xmath6 @xmath102 @xmath6 . this continuum flux is consistent with the upper - limit ( @xmath104 and @xmath105 erg s@xmath6 @xmath102 @xmath6 before and after the extinction - correction , respectively ) set by smith et al . ( 2008a ) . assuming that the flux excess is actually due to an h@xmath7 line of which the line width is @xmath106 , we can derive the luminosity of the line flux as @xmath8 erg s@xmath6 . in figures [ fig : speccomp1][fig : speccomp3 ] , the late time spectrum of sn 2006gy is compared with those of various classes of sne . unlike other slowly - declining type iin / iia sne , the h@xmath7 emission line is neither strong nor wide . the emission of ir triplet in sn 2006gy is much narrower than those in other type iin / iia sne , suggestive of a slow sn ejecta . the presence of forbidden lines [ ] 7155 and [ ] 7302 is also peculiar for a type iin / iia sn . on the other hand , this sn does not show strong [ ] 6300 , 6364 emission line , which is generally strong in type ii and ib / c sne having forbidden [ ] 7302 line ( fig . [ fig : speccomp1 ] and [ fig : speccomp3 ] ) line is also absent in type ia sne because of the low oxygen abundance in the ejecta ( see e.g. , kozma et al . 2005 ) ] . some of the intermediate - width emission lines , marked with dashed lines in figure [ fig : specobs2 ] , are unusual for any type of sne . the presence of the wide absorption trough at d is also unusual for type iin / iia sne . similar feature is seen in the late phase spectrum of a peculiar type ia sn 2005hk ( fig . [ fig : speccomp3 ] ) . we discuss these peculiar properties of sn 2006gy in the next section . the luminosity of sn 2006gy at the late phase is consistent with the radioactive decay model as shown in figure [ fig : lcabs ] . the lcs of type iin sne show a large variety and some of which might also be in fair agreement with sn 2006gy . thus , it seems difficult to discriminate the main heating source of sn 2006gy only from the lc . the late time spectrum gives some indications on this issue . sn 2006gy is classified as type iin from earlier spectra showing strong h@xmath7 emission line . however , as mentioned in [ sec : spec ] , the late time spectrum is quite atypical for type iin / iia sne ( fig . [ fig : speccomp2 ] ) . the emission lines of ir triplet are narrow and not blended . there are several forbidden lines , e.g. , [ ] 7155 and [ ] 7302 . the luminosity of the possible h@xmath7 emission line at 394 days ( @xmath8 erg s@xmath6 ) is considerably smaller than those of type iin sne 1988z ( @xmath107 erg s@xmath6 at 492 days ; turatto et al . 1993 ) , 1995z ( @xmath108 erg s@xmath6 at 716 days ; fransson et al . 2002 ) and 2006tf ( @xmath109 erg s@xmath6 at 445 days after discovery ; smith et al . 2008b ) and also type iia sn 2002ic at earlier phase [ @xmath110 erg s@xmath6 at 222 days ; deng et al . it is rather consistent with the pure radioactive decay model of a typical type ii sn ( @xmath111 m@xmath112 @xmath4co and @xmath12 m@xmath112 envelope ; chugai 1991 ) . thus , the observed h@xmath7 emission flux does not require any type iin - like csm interaction at @xmath25 days . therefore , the strong csm interaction expected from the conspicuous h@xmath7 emission and the flattened lc before @xmath2 200 days seems to have faded by 394 days . if the observed @xmath3 band luminosity is purely originated from radioactive decay , @xmath113 is necessary as discussed in [ sec : lc ] . however , the pure radioactive decay model also seems to have an inconsistency . core - collapse sne generally show the [ ] 6300 , 6364 line at late time except for type iin / iia sne , while sn 2006gy showed little or no [ ] emission ( figs . [ fig : speccomp1][fig : speccomp3 ] ) . as for fe lines , in type ic sn 1998bw , a blend of [ ] lines is seen around 5200 and the [ ] 7155 line has a dominant contribution to the emission line around 7200 , while only [ ] 7155 line is seen in sn 2006gy . the lc of sn 1998bw is fully explained by the heating of @xmath114 of @xmath4ni(iwamoto et al . 1998 ; nakamura et al . 2001 ; maeda et al . 2006 ) . if the peak luminosity of sn 2006gy is powered by the decay of @xmath4ni , the required mass of @xmath4ni is @xmath115 , more than 20 times larger than that of sn 1998bw . this would result in stronger fe lines in sn 2006gy . the same discussion holds true in comparison with type ia sne , which eject @xmath116 of @xmath4ni on average ( e.g. , stritzinger et al . both sne 1990n and 2004eo showed strong and broad [ ] features around 5200 and 7150 ( fig . [ fig : speccomp3 ] ) . thus , the weakness of [ ] and [ ] lines seem against the @xmath4ni heating scenario . however , this peculiarity in the line strength might be explained by a moderately high density of the line emitting region . we will discuss it in [ subsec : density ] . the presence of the forbidden [ ] 7155 and [ ] 7302 lines suggests the density of the emitting region in sn 2006gy is less than those in other type iin / iia sne . however , it is likely that the density is still higher than those of other type sne ; the line flux ratio @xmath117 of sn 2006gy is only @xmath118 ( table [ tbl : unident_line ] ) , while @xmath119 in sne of other types ( fig . [ fig : speccomp1 ] and [ fig : speccomp3 ] ; see also gmez & lpez 2000 ) . the value of @xmath120 suggests an electron density of the emitting region @xmath121@xmath122 @xmath123 ( ferland & persson 1989 ; fransson & chevalier 1989 ) . this is consistent with the existence of the [ ] 7155 because its critical density of electron is not so small ( a few times @xmath124 @xmath123 ; zickgraf 2003 ) . in such a high density region , the forbidden line [ ] 6300 , 6364 would be effectively quenched , because its critical density is only a few times @xmath125 @xmath123 . therefore , the low @xmath117 value and the absence of [ ] 6300 , 6364 in sn 2006gy can be qualitatively explained by a moderately high electron density of the emitting region , which would be less than those of type iin / iia sne and more than other type ii and ib / c sne . in such a high density environment , emission could occur in permitted multiplets ( e.g. , turrato et al . 2000 ; deng et al . 2004 ) , which might contribute to the blue continuum flux at @xmath126 . the absence of intermediate - width [ ] line around 5200 , which is seen together with [ ] 7155 in type ia sne as well as in the peculiar type ic sn 1998bw ( fig . [ fig : speccomp3 ] ) , might be explained by this higher density assumption . the high density of the ejecta is also consistent with the current understanding that the @xmath4ni heating model requires a massive progenitor and a large amount of ejecta ( umeda & nomoto 2008 ) . a large ejecta mass is also needed to reproduce the long duration of the lc around maximum . the high density of the ejecta would also be consistent with the lower expansion velocity ( @xmath127 km s@xmath6 ) derived from fwhm of the emission lines in the late time spectrum . low expansion velocity was also suggested from early spectroscopy ( ofek et al . 2007 ) . there are several unusual features in the late time spectrum of sn 2006gy ( marked with dashed lines in fig . [ fig : specobs2 ] ) . these lines are not clearly seen in other types of sne ( fig . [ fig : speccomp1][fig : speccomp3 ] ) . consulting line lists in literature ( e.g. , moore 1945 ) , we find possible candidates as shown in table [ tbl : unident_line ] . some of them could be lines of nickel ( [ ] 7380 , 8704 ; [ ] 7394 , 7908 , 8202 , 8843 ) and/or titanium ( [ ] 7917 , 8040 , 8060 , 8349 , 8371 ) . the emission line at 7716 could be due to iron ( [ ] 7709 or [ ] 7765 ) . if these identifications are correct , the innermost ejecta of sn 2006gy was visible at @xmath25 days because ni and ti are likely to exist only inner part of the ejecta . assuming that the line emitting region expands homologously with @xmath128 km s@xmath6 and all the elements are singly ionized , the high density ( @xmath124 @xmath129 @xmath123 ) implies that the mass of the hypothetical innermost region is @xmath20.88@xmath130 ( if dominated by fe - peak elements ) or @xmath20.33@xmath130 ( if dominated by ca ) . this would suggest that the assumption of @xmath131 @xmath123 is reasonable . in this case , the electron scattering optical depth is @xmath20.22 , which suggests that @xmath131 is also necessary for the innermost region to be optically thin . in sum , our hypothesis of the emergence of the innermost part requires @xmath131 @xmath123 , being consistent with the high density interpretation . however , the identification of these lines are still tentative . for example , we do not see the [ ] 7507 feature in the observed spectrum ( fig . [ fig : specobs2 ] ) , although it is expected to emit as strong as the [ ] 7908 . it is noted the feature around 7910 is also seen in some type ia sne ( fig . [ fig : speccomp3 ] ) . there is also an overall resemblance with the peculiar type ia sn 2005hk at 240 days , especially at 55006000 , 71007500 and 85008900 ( fig . [ fig : speccomp3 ] ) . to discuss this subject further , a more reliable identification of these lines will be needed . we presented the spectroscopic and photometric observation of luminous type iin sn 2006gy at @xmath25 days . the good seeing condition enabled us to detect it in the optical wavelengths . the overall lc is roughly consistent with the radioactive heating model . the deviation between the observed lc and the radioactive lc model around 200 days suggests that the strong csm interaction considerably contributed at @xmath2 200 days and then weakened by 394 days . the late time spectrum of sn 2006gy was unique . it showed some emission lines with intermediate width ( @xmath5 km s@xmath6 ) , including forbidden lines of [ ] and [ ] which are not seen in normal type iin / iia sne . the absence of [ ] 6300 , 6364 line and the weakness of [ ] lines would be a result of moderate electron density of @xmath132@xmath122 @xmath123 in the emission line region . this is consistent with the larger amount of ejecta and less expansion velocity suggested for this sn . although sn 2006gy exhibited possible h@xmath7 emission line , its flux was considerably lower than those of type iin / iia sne and rather comparable with that of a typical type ii sn . this suggests that the strong csm interaction has finished by @xmath25 days , being consistent with the prediction from the lc . several unusual emission features were present at 74008800 and some of them might be ni and/or ti lines . however , these identifications are still tentative and we need detailed modeling to explain the full aspects of the observation . we are grateful to an anonymous referee for many helpful comments . m.t . would like to thank alex filippenko , weidong li and sergei blinnikov for fruitful discussion and useful comments . we have utilized suspect database and would like to thank the managers of this database and all the contributors of the data used in the paper . this research has been supported in part by the grant - in - aid for scientific research ( 17684004 , 18104003 , 18540231 , 20740107 , 20840007 ) from the jsps , mext , and world premier international research center initiative , mext , japan . and n.t . are supported through the jsps ( japan society for the promotion of science ) research fellowship for young scientists . nomoto , k. , tominaga , n. , tanaka , m. , maeda , k. , & umeda , h. 2007 , `` supernova 1987a : 20 years after : supernovae and gamma - ray bursters '' , eds . s. immler , k. weiler , & r. mccray ( new york : aip ) , 412 ( astroph/0707.2187 ) ccccccccc 127 days & 2006 dec 25.4 & 54094.4 & 3600 & 1200 & @xmath133@xmath134 & 1.181.23 & @xmath135 & @xmath136 + 157 days & 2007 jan 24.4 & 54124.4 & 3600 & 1200 & @xmath137@xmath138 & 1.681.84 & & + 394 days & 2007 sep 18.5 & 54361.5 & 660 & 1200 & @xmath139@xmath140 & 1.121.15 & @xmath141 & @xmath142 + cccl 6563 ? & @xmath21 & @xmath106 & h@xmath7 6563 + 7059 & @xmath143 & @xmath144 & 7067 + 7155 & @xmath145 & @xmath146 & [ ] 7155 + 7302 & @xmath147 & @xmath148 & [ ] 7291 , 7323 , [ ] 7320 , 7330 , 7308 + 7387 & @xmath149 & @xmath150 & [ ] 7380 ? , + 7443 & @xmath151 & @xmath152 & 7462 + 7716 & @xmath153 & @xmath154 & 7712 , [ ] 7709 ? , [ ] 7765 ? + 7913 & @xmath155 & @xmath156 & [ ] 7908 ? , + 8052 & @xmath157 & @xmath158 & [ ] 8040 ? , [ ] 8060 ? + 8206 & @xmath159 & @xmath160 & [ ] 8202 ? , + 8325 & @xmath161 & @xmath162 & [ ] 8349 ? , [ ] 8348 ? + 8383 & @xmath161 & @xmath163 & [ ] 8371 + 8479 & @xmath164 & @xmath165 & 8498 , [ ] 8490 ? + 8544 & @xmath166 & @xmath167 & 8542 + 8681 & @xmath168 & @xmath169 & 8662 , [ ] 8704 ? + 8835 & @xmath170 & @xmath171 & [ ] 8843 ?
supernova ( sn ) 2006gy is an extremely luminous type iin sn characterized by the bright peak magnitude @xmath0 @xmath122 mag and its long duration . the mechanism giving rise to its huge luminosity is still unclear . we performed optical spectroscopy and photometry of sn 2006gy at late time , @xmath2 400 days after the explosion , with the subaru / focas in a good seeing condition . we carefully extracted the sn component , although there is an ambiguity because of the contamination by bright nucleus of the host galaxy . we found that the sn faded by @xmath2 3 mag from @xmath2 200 to @xmath2 400 days after the explosion ( i.e. , by @xmath2 5 mag from peak to @xmath2 400 days ) in @xmath3 band . the overall light curve is marginally consistent with the @xmath4ni heating model , although the flattening around 200 days suggests the optical flux declined more steeply between @xmath2 200 and @xmath2 400 days . the late time spectrum was quite peculiar among all types of sne . it showed many intermediate width ( @xmath5 km s@xmath6 fwhm ) emission lines , e.g. , [ ] , [ ] , and . the absence of the broad [ ] 6300 , 6364 line and weakness of [ ] and [ ] lines compared with ir triplet would be explained by a moderately high electron density in the line emitting region . this high density assumption seems to be consistent with the large amount of ejecta and low expansion velocity of sn 2006gy . the h@xmath7 line luminosity was as small as @xmath8 erg s@xmath6 , being comparable with those of normal type ii sne at similar epochs . our observation indicates that the strong csm interaction had almost finished by @xmath2 400 days . if the late time optical flux is purely powered by radioactive decay , at least @xmath9 should be produced at the sn explosion . in the late phase spectrum , there were several unusual emission lines at 74008800 and some of them might be due to ti or ni synthesized at the explosion .