article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
the study of the solvation of ionic species in aqueous clusters is of great importance in many areas of environmental and atmospheric chemistry.@xcite many chemical reactions occurring in the earth atmosphere involve ions embedded at the surface of icicles .
protonated water clusters are perhaps one of the most abundant of these species in the stratosphere and can host a large variety of chemical processes ; most notably are those controlling atmospheric nucleation .
for example , yang et al.@xcite have analyzed the importance of these aggregates in the earliest stages of the formation of noctilucent clouds at high altitudes , at temperatures as low as 150 k. from a more fundamental perspective , the analysis of the equilibrium and dynamical characteristics of chemical reactivity in these clusters continues to draw considerable attention from experimental and theoretical perspectives.@xcite one of the reasons for this interest is the fact that , by a careful control of the cluster sizes , one can bridge the gap between analyses of reactivity in two important limiting behaviors , @xmath4 gaseous and condensed phases . in the present paper
we examine the solvation of an excess proton in mesoscopic clusters of the type h@xmath5[h@xmath0o]@xmath1 , with sizes intermediate between @xmath6 and @xmath7 .
this size range covers aggregates with incipient three dimensional structures up to nanoclusters where it is possible to establish a clear distinction between inner , bulk - like " , domains and surface states .
our analysis will be based on results obtained from molecular dynamics simulation experiments and includes the study of the equilibrium solvation structures and dynamical aspects related to proton transfer processes as well . to tackle the problem we resorted on an multistate empirical valence bond ( ms - evb ) model hamiltonian.@xcite this approach has been repeatedly used in the past and has proved to be sufficiently versatile so as to be successfully applied for the examination of aqueous protons in a wide variety of environments : bulk,@xcite clusters,@xcite confined environments,@xcite air - water interfaces,@xcite and water at extreme conditions of temperature and pressure,@xcite to cite a just a few important examples .
the organization of the paper is as follows : in section ii we present a brief overview of the model and the simulation procedure .
the main results are presented in section iii .
the last section contains the concluding remarks .
the systems under investigation consisted of protonated clusters of the type h@xmath5[h@xmath0o]@xmath1 , with @xmath6 , 21 and 125 .
the molecular dynamics experiments were performed using an ms - evb scheme .
this model has been thoroughly described in previous studies , so we will present here a brief overview of its main features and refer the interested reader to refs .
@xcite , for more complete presentations .
the cornerstone of the model is the following ms - evb hamiltonian : @xmath8 where @xmath9 represents a basis set of diabatic states .
each of these states denotes the spatial localization of the excess proton in a different water molecule .
the matrix elements of the ms - evb hamiltonian , @xmath10 , are parametrized in terms of the nuclear coordinates and are adjusted so that evb predictions for the energetics and geometrical details of small protonated water clusters agree with those obtained from highly sophisticated ab initio calculations.@xcite at each step of the molecular dynamics runs , the simulation procedure involved the construction of a connectivity pattern of hydrogen bonds ( hb ) , linking the water molecule exhibiting strongest h@xmath11o@xmath5 character ( hereafter referred to as the pivot water and denoted h@xmath0o@xmath12 ) with its first , second and eventually third solvation shells .
this pattern allowed the identification of the instantaneous diabatic states included in the construction of @xmath13 . for @xmath6 and @xmath14 , the number of evb states considered was intermediate between 8 and 15
; this number went up to typically 20 , for simulations of the largest aggregates .
the dynamics of the classical nuclei was generated from the following newton s equations of motions : @xmath15 where @xmath16 represents the lowest eigenvalue of the evb hamiltonian , namely : @xmath17 is the ground state eigenstate of @xmath13 , whose expression in terms of the diabatic states can be written as : @xmath18 the index of the largest expansion coefficient , @xmath19 , identifies the instantaneous pivot water , which can be eventually updated , in the advent of a proton translocation episode .
diagonal elements @xmath20 included inter- and intra - molecular interactions involving the h@xmath11o@xmath5 group and the rest of the water molecules,@xcite that were modeled using the flexible tip3 model.@xcite all simulation experiments began with a @xmath21 ns preliminary thermalization stage , at temperatures close to @xmath22 k. in this thermal regime , the clusters exhibited structural and dynamical characteristics similar to those observed in liquidlike phases .
subsequent to this equilibration period , statistics were collected from constant energy trajectories , lasting typically @xmath23 ns .
mesoscopic clusters represent inhomogeneous environments at the nanometer scale .
perhaps the first important question to be answered in connection with the solvation of an excess proton in a water cluster concerns the average localization of the excess charge . in this context
, the consideration of two local density fields with respect to the center of mass of the cluster will be useful : the first one , @xmath24 , involves the position of the pivot water and is defined as : @xmath25 note that @xmath24 represents the probability distribution of finding the pivot at a distance @xmath26 from the center of mass of the aggregate . in the previous equation , @xmath27 and @xmath28 denote the positions of the pivot and the center of mass of the cluster , respectively .
the second density , @xmath29 , involves the rest of the water molecules and is defined as : @xmath30 where @xmath31 represents the coordinate of oxygen site in the @xmath32-th molecule and the sum involves all the water molecules , except the one acting as pivot . results for both functions are depicted in fig .
1 . the profiles of the bottom panel contrast sharply to those shown in the upper ones : for the smallest size considered , @xmath6
, it is evident that the excess charge localizes closer to the center of mass of the aggregate .
a snapshot of a typical cluster configuration for a protonated clusters of this size is depicted fig .
2 . note that the overall structure looks as a mostly planar , star - like , arrangement of hydrogen bonded water molecules , with the three - coordinated hydronium lying at a central position .
the situation changes at a qualitative level in larger aggregates ( see middle and top panels of figs . 1 ) . under these circumstances ,
the water local density at the central part looks much more uniform and its value does not differ substantially from the usual one for bulk water at ambient conditions : @xmath33 @xmath34 .
snapshots for cluster configurations in this size range are also included in fig . 2 and reveal that , in both cases , the aggregates exhibit a three - dimensional hydrogen - bonded structure , in which it is possible to discriminate bulk from surface states environments . as for the position of the proton
is concerned , there are clear indications that the regions for stable proton solvation in both cases are located at the corresponding cluster boundaries .
this observation has been previously reported@xcite and has also been found in simulations of protons at macroscopic water - air interfaces.@xcite it is also of interest to investigate the closest solvation structures of the excess charge . in bulk water ,
the microscopic description of such environments includes a whole series of arrangements which are intermediate between two limiting moieties , with well defined geometry : we are referring to the symmetric zundel ( zdl ) dimer@xcite [ h@xmath0o - h - oh@xmath0]@xmath5 and three - coordinated eigen ( egn ) cation@xcite [ h@xmath11o@xmath36(h@xmath0o)@xmath11]@xmath5 .
interconversions between these structures take place in the picosecond time scale and seems to be triggered by changes in the intramolecular connectivity between the proton closest solvation shells
. the relative prevalence of these structures can be conveniently monitored by considering probability densities associated to an appropriate order parameter identifying egn and zdl complexes . in the ms - evb model
, such order parameter can be expressed in terms of the two largest expansion coefficients ( see eq .
3 ) as:@xcite @xmath38 zdl and egn solvation environments are characterized by values of @xmath39 close to 0 and 1 , respectively . in fig .
3 we present results for @xmath40 , the free energy associated to the order parameter : @xmath41 where @xmath42 is the inverse of boltzmann constant times the temperature and @xmath43 represents a statistical average . for the sake of comparison ,
we have also included in the figure the results for bulk water reported in refs .
[ 14 , 16 ] in all cases , the most stable structures which correspond to the global minimum of @xmath44
are characterized by values of @xmath45 .
the magnitude of the free energy differences between egn - like and zdl - like structures though , does depend on the particular environment considered .
results for @xmath46 in aggregates with @xmath47 and @xmath48 are practically identical , revealing that the characteristics of the surface dynamical modes triggering egn - zdl interconversions are comparable for both cluster sizes . on the other hand , for @xmath49 , egn - zdl interconversions become much more frequent and are dominated exclusively by polarization fluctuations in the aggregate .
the differences in temperature preclude the direct comparison of these results to those observed in bulk water at ambient conditions ; anyhow , our simulations show that the free energy barrier for the isotropic bulk phases is @xmath21 kcal mol@xmath50 higher than the one observed at the surface of the largest clusters .
to gain additional insight on the local coordination of the excess charge , we also examined pair correlations involving the o@xmath12 site . in the top panel of fig .
4 we show plots for pair correlations between the pivot o@xmath12 and oxygen ( top panel ) and hydrogen ( bottom panel ) sites in the rest of the aqueous domains . correlations with oxygen sites are dominated by main peaks located at @xmath51 which involves @xmath52 water - oxygen sites acting as hb acceptors . the magnitude and position of these peaks do not change with the cluster size and are similar to those observed in bulk water.@xcite the overall shapes of the pivot - hydrogen profiles are also dominated by main peaks shifted @xmath53 outward and include six hydrogen atoms belonging to water molecules in the pivot first solvation shell .
consequently , we are led to conclude that , in all aqueous environments investigated , the closest solvation structures of the proton are preserved , being characterized by a first solvation shell composed by three water molecules , acting as hb acceptors with no evidence of pivot hb of the type o - h@xmath54o@xmath12 .
having established the main features of the equilibrium solvation structures of the proton , we now move to the consideration of the dynamical characteristics related to transfer processes in aqueous clusters . in order to acquire a preliminary notion of the timescales characterizing proton translocation events in water clusters , in fig .
5 we present the time evolutions of the pivot label during a 1 ns time interval . at a first glance , it is selfevident that proton jumps become much more frequent as we move to larger aggregates . at @xmath22
k and for @xmath6 , the dynamics of the pivot label can be pictured as a sequence of three well differentiated classes of episodes : @xmath55 on one hand , one observes fast resonances in the subpicosecond time domain , between zdl pairs interrupted by @xmath56 stages during which the pivot label remains practically unchanged for several hundreds of picoseconds . @xmath57
the stabilization of the proton in a different water pivot is a much more rare event since it is normally preceded by major rearrangements of the overall cluster geometry .
these modifications involve changes in the localization of the potential donor and acceptor molecules , requiring the migration of the acceptor molecule towards a more central position within the aggregate .
this is operated by the breakage and regeneration of new hb , that take place over lengthscales comparable to the overall size of the aggregate .
although the time evolutions of the pivot label for larger aggregates present similar qualitative characteristics , translocation episodes do occur much more frequently .
concerning possible mechanisms that might trigger transfer episodes at cluster surfaces , the direct inspection of the close vicinity of the acceptor - donor pair revealed that the transfer dynamics is operated by subtle changes in the hb structure in the second and third solvation shells . in this respect , although we have not performed a detailed trajectory analysis of reactive processes , we are led to believe that the surface mechanisms that regulate translocations resemble very much those observed in bulk macroscopic phases.@xcite a more precise estimate for the proton transfer rates can be extracted from population relaxation time correlation functions of the type : @xmath58 where @xmath59 denotes the instantaneous fluctuation of the population of the @xmath32th reactant away from its equilibrium value . in eq .
[ ct ] , the dynamical variable @xmath60 is unity if the proton is localized in the @xmath32-th diabatic state at time @xmath61 and zero otherwise .
provided onsager s
regression hypothesis@xcite remains valid , the exponential decay at long times of @xmath62 should yield an estimate of the proton transfer rate , @xmath63 .
plots for @xmath64 are presented in fig . 5 ; after fast transients ascribed to zdl resonances , the three curves present single exponential decays .
compared to the bulk result @xmath65 ps , the estimates for the characteristic timescales in clusters are intermediate between one and two orders of magnitude larger : @xmath66 ps and @xmath67 ps .
the structural and dynamical characteristics of excess protons embedded in liquid - state nanoclusters of the type h@xmath5[h@xmath0o]@xmath68 described in this paper illustrate distinctive behaviors , depending on the cluster size considered . more specifically , we focussed attention to @xmath6 , 21 and 125 .
taking into consideration that the cluster radii scale as @xmath69 , we tried to cover different scenarios : on one hand , we considered @xmath70 aggregates , in which it is impossible to discern surfance from bulk states ; on the other , @xmath71 clusters with @xmath72 of the order of three molecular diameters perhaps the smallest ones in which such distinction can be clearly established . in small aggregates , the overall structures of h@xmath5[h@xmath0o]@xmath73 clusters can be pictured as dendritic - like arrays of hydrogen bonded water molecules , with the excess proton occupying a central position . at larger sizes , there is a clear tendency for surface solvation of the excess charge .
note at in all cases , these two solvation structures preserve two important characteristics of the closest aqueous environments of the excess charge : @xmath55 the three - coordinated , hb donor structure of hydronium with respect to its first solvation shell and @xmath56 the absence of hb acceptor structure of the type o - h@xmath54o@xmath12 .
we remark that this anisotropies have also been reported in previous cluster studies,@xcite at water - air interfaces,@xcite and confined water within membranes.@xcite the observed proton dynamics dynamics in small clusters can be pictured as a sequence of three different episodes : @xmath55 first , one observes resonance stages , during which the proton seems to be delocalized over a tagged zdl pair .
these fast dynamical modes of the pivot label seem to be dictated by polarization fluctuations originated in the closes solvation environments of the excess charge ; @xmath56 the latter fluctuations seem to be also the controlling agent for the localization stages , during which the resonances cease and the proton gets localized in a single water molecule , for periods that may last typically @xmath74 ps ; @xmath57 finally , one also observes much rare proton translocation episodes , which are preceded by global rearrangements of the hydrogen bond connectivity of the the cluster .
these dynamical characteristics are maintained at a qualitative level for @xmath3 and @xmath7 aggregates , although the translocations become much more frequent and the triggering mechanisms require much more subtle modifications in the hb connectivity that those observed in smaller aggregates .
dl and jr are staff members of conicet - argentina .
p. kebarle , annu .
( 1997 ) 445 .
a. w. castleman , jr .
, r. g. keese , chem . rev . 86 ( 1986 ) 589 .
a. w. castleman , jr .
, r. g. keese , annu .
37 ( 1986 ) 525 .
a. w. castleman , jr .
, r. g. keese , science 241 ( 1988 ) 36 .
s. s. iyengar , m. k. petersen , j. f. day , c. j. burnham , v. e. teige , g. a. voth , j. chem .
( 2005 ) 084309 . c. j. burnham , m. k. peterser , j. f. dany , s. s. iyengar , g. a. voth , j. chem .
124 ( 2006 ) 024327 . | structural and dynamical characteristics pertaining to the solvation of an excess proton in liquid - like nanoclusters of the type [ h@xmath0o]@xmath1 are investigated using molecular dynamics experiments .
three different aggregate sizes were analyzed : @xmath2 21 and 125 .
the simulation experiments were performed using a multistate empirical valence bond hamiltonian model . while in the smallest aggregates the proton occupies a central position , the stable solvation environments for @xmath3 and 125
are located at the cluster boundaries . in all cases ,
the structure of the closest solvation shell of the excess charge remains practically unchanged and coincides with that observed in bulk water .
compared to results obtained in bulk , the computed rates for proton transfer in clusters are between one and two orders of magnitude slower , and tend to increase for larger cluster sizes . |
relaxation to thermal boltzmann - gibbs equilibrium in @xmath0-particle hamiltonian systems with long - range interactions has been recently the subject of an intense debate @xcite . in some cases
, the relaxation time has been shown to be extremely long and to increase with the number of particles .
hence , the study of these out - of - equilibrium conditions is of paramount importance for physical applications .
the dynamics of systems with long - range interactions shows that some states are attained quickly , on time scales of @xmath4 , and that afterwards the system evolves very slowly , on time scales diverging with @xmath0 , towards boltzmann - gibbs equilibrium .
we call states that evolve on time scales that diverge with @xmath0 `` quasi - stationary '' .
some of them are characterized by a particle distribution in the @xmath5-space , @xmath6 ( with @xmath7 the position and conjugate momentum of the @xmath8-th particle and @xmath9 the dirac function ) , which remains close to a slowly varying smooth distribution for times that increase with @xmath0 .
it should be remarked that most of the numerical evidences of this behavior are for @xmath10 and @xmath11 systems .
the theoretical explanation we propose in this paper , which is developed for mean - field models , extends to any dimension , as soon as the two body interaction is sufficiently smooth .
many recent studies of such quasi - stationary states are performed for the so - called hamiltonian mean field ( hmf ) model ( for a review see ref .
this model describes the motion of @xmath0 rotators under the action of an attractive or repulsive infinite range cosine interaction . in this paper
we will consider the attractive case .
the model then displays a second order phase transition , which is related to the development of a dynamical instability of the spatially homogeneous initial state with gaussian distribution of momenta , at a given value of the energy @xcite .
the analysis developed in this paper is based on a theorem due to braun and hepp @xcite , according to which the dynamics of a classical @xmath0-particle system interacting via a two body and sufficiently regular long - range potential is well approximated by the associated vlasov dynamics of the density in @xmath5-space . in the case in which the interaction force derives from a two - body smooth potential , vlasov equation writes @xmath12 the mean field macroscopic potential @xmath13 is a functional of the probability distribution function @xmath14 , which makes the equation nonlinear in @xmath15 .
more precisely , braun - hepp s theorem states that , for a mean - field microscopic two - body smooth potential , the distance between two initially close weak " solutions of the vlasov equation increases at most exponentially in time .
the theorem applies also to singular distributions , e.g. distributions that have a support on a set of dimension smaller than the one of the @xmath5-space ( for instance , a line in the two dimensional @xmath5-space of the hmf model ) .
if we apply this result to a large @xmath0 particle approximation of a continuous distribution the error at @xmath16 is typically of order @xmath17 , thus for any small " @xmath18 and any large enough " particle number @xmath0 , there is a time @xmath19 up to which the dynamics of the original hamiltonian and its vlasov description coincide within an error bounded by @xmath18 .
the theorem implies that this time @xmath19 increases at least as @xmath20 .
extensions to wave particle dynamics of such a result have also been reported recently @xcite .
since quasi - stationary states evolve on time scales that diverge with @xmath0 , this result suggests that these states might gain their stability from being `` close '' to some stable stationary states of the vlasov dynamics . besides the conserved quantities of the hamiltonian dynamics ( energy , momentum , angular momentum ... ) , the vlasov description introduces additional integrals : the so - called casimirs @xmath21=\int s \left(f(\mathbf{r},\mathbf{p},t)\right ) \ , d\mathbf{r } d\mathbf{p}\ , , \ ] ] for any function @xmath22 .
these conservation laws are responsible for the existence of an infinity of stable stationary states for the vlasov dynamics . in this paper
, we will argue that the existence of this infinity of stationary states is a possible explanation for the generic existence of quasi - stationary states in the finite @xmath0-dynamics .
this interpretation rises the following questions .
may one predict the quasi - stationary states that emerge after a complex unstable vlasov dynamics ? among these stationary states , are there some `` statistically preferred '' states ? what governs the relaxation towards equilibrium of the stationary states , and which is the scaling of the relaxation time with the number of particles ?
we will address in this paper the first and the third question , while the issues related to the second question will be only briefly mentioned .
before the slow relaxation phase sets in , a fast evolution takes place on a time scale that is independent of @xmath0 .
this phenomenon is denoted as _ violent relaxation _ in the astrophysical context @xcite , and is a consequence of the nonlinear complex dynamics of the vlasov equation .
after this very rapid stage , vlasov dynamics produces thinner and thinner filamentations of the density @xmath15 , which lead to an apparent equilibrium described by a coarse grained density function @xmath23 .
a statistical mechanics interpretation of this process has been proposed : for instance , for an initial two level density function , the equilibrium density is of fermi - dirac type @xcite . even if this may be already a partial answer to our second question above , we will not address in detail the very delicate point of the convergence of the vlasov dynamics towards _ its _ statistical equilibrium in the present paper .
some authors @xcite have advanced a challenging interpretation of the quasi - stationary states , suggesting that they should be true equilibrium " states , obtained by maximizing the tsallis @xcite entropy of the single particle distribution .
we think that a landmark of such an interpretation would be the assessment of the existence of power law tails in the single particle momentum ( or energy ) distributions .
although we will not enter into a detailed fitting of such distributions using tsallis @xmath3-exponentials , we will present a strong evidence of the absence of power law tails in the single particle momentum distribution at any stage of the time evolution , for the whole class of initial conditions we investigate in this paper , that are all homogeneous in space ( see section [ sec : long - time ] ) . for an initial condition in which the particles are concentrated in a point and momentum
is uniform in an interval around zero , the authors of ref .
@xcite have been able to fit a tsallis @xmath3-exponential to the central part of the momentum distribution , after imposing an arbitrary cut - off to the tails .
however , these authors do not exhibit any evidence of existence of power law tails , even for such a special initial state .
moreover , as mentioned above , our analysis applies also to this initial condition although it does nt correspond to a stationary state of the vlasov equation . at best
, tsallis statistics could describe the quasi - stationary states obtained from such a special initial state , but certainly not all of them , in particular those originated from the large class of homogeneous states that are studied in this paper . moreover , the steadily progressing dynamical evolution observed for our class of initial conditions does not show any intermediate `` statistically preferred '' state .
as stated above , the time evolution follows a sequence of stationary vlasov states until , asymptotically , boltzmann - gibbs equilibrium is attained .
it should be however mentioned that the special initial state studied in @xcite is very interesting from a dynamical point of view , since it shows long - time correlations that are absent for the homogeneous initial states studied here ( see also @xcite ) .
similar initial states produce fractal structures in the @xmath5-space for a self - gravitating sheet model @xcite . as discussed above
, braun - hepp s theorem suggests that the similarity between hamiltonian @xmath0-particle dynamics and vlasov dynamics persists for times that increase as @xmath20 .
these times are linked with the fastest possible instability of the vlasov dynamics . for stable solutions ,
the appropriate timescale is , however , the one associated with the fluctuations of the mean field . in agreement with this theoretical remark , we present in sec . [ sec : long - time ] numerical results that indicate that the persistence of quasi - stationary states is present up to times that are much longer than braun - hepp s @xmath20 .
this time scale increases as a power law in @xmath0 with a nontrivial exponent .
similar time scales have been found in gravitational systems .
this is the case of chandrasekhar s collisional " time scale , which is of order @xmath24 @xcite .
although such time scale is similar to those we find in the hmf model , because it signals the final process of relaxation to boltzmann - gibbs equilibrium , its origin in our model is certainly different . in order to reformulate
the timescale hierarchy sketched above , we can expect the following scenario : 1 .
an initial and fast evolution , well described for all initial conditions by the vlasov dynamics , takes place on a timescale independent of the particle number @xmath0 .
the system is then always trapped close to one of the numerous stable stationary states of the vlasov equation .
this state may be the statistical equilibrium of the vlasov equation ( the most probable state with constraints given by the vlasov invariants ) .
the system slowly evolves , on a much longer time scale , due to `` collisions '' , or due to the fluctuations around this vlasov stationary state .
consequently this time scale will be a function of @xmath0 .
one can expect that this slow evolution takes place passing through different stable vlasov stationary states .
finally , the system reaches a particular vlasov stable state , corresponding to the boltzmann - gibbs equilibrium state .
we will try to give support to this scenario in the remaining of the paper .
the plan is the following .
we first introduce in section [ model ] the hamiltonian mean field model .
we will then show in section [ vlasov ] that the vlasov dynamics has an infinity of stationary states and we propose a new general stability criterion for any homogeneous distribution ( including non boltzmann - gibbs ones ) .
we will compare these analytical results with numerical simulations of the finite-@xmath0 hamiltonian dynamics .
the rapid convergence towards a stable stationary state of the vlasov equation is described in section [ sec : short - time ] . in the case of unstable stationary states , we show that the exponential destabilization may be investigated taking advantage of the existence of unstable modes of the vlasov dynamics , in accordance with braun - hepp s theorem .
the slow evolution towards equilibrium , passing through different stable vlasov stationary states is described in section [ sec : intermediate - time ] .
we will consider the hamiltonian mean field ( hmf ) model , whose hamiltonian is @xmath25,\ ] ] where @xmath26 is the position ( angle ) of the @xmath8-th particle on a circle and @xmath27 the corresponding momentum .
this system can be seen as representing particles moving on a unit circle interacting via an infinite range attractive cosine potential , or as classical xy - rotators with infinite range ferromagnetic couplings @xcite .
the magnetization , defined as @xmath28 or more precisely its modulus , @xmath29 equilibrium statistical mechanics can be derived exactly both in the canonical and in the microcanonical ensembles @xcite . in the ferromagnetic case , that we consider here , a minimal free energy ( maximum entropy ) state with a nonvanishing magnetization appears when lowering temperature ( resp .
energy per particle ) below @xmath30 ( resp .
@xmath31 ) . a discontinuity at @xmath32 ( @xmath33 ) in the second derivative of the free energy ( entropy ) with respect to magnetization signals a second order phase transition . this transition is between a low energy phase with particles forming a cluster ( rotators pointing towards a preferred direction ) , and a high energy phase with particles evenly distributed on the circle ( no preferred direction for rotators ) .
this theoretical result , valid in the @xmath34 limit , is also confirmed by direct numerical simulations of the equations of motion ( [ eq : canonical ] ) , which moreover allows a careful analysis of finite @xmath0 corrections and give access to the study of non equilibrium features .
it s in this context that , within the energy region @xmath35 $ ] , a class of initial states has been found that displays an extremely slow relaxation towards boltzmann - gibbs equilibrium @xcite , with a relaxation time that increases with @xmath0 .
similar phenomena occur for other particle systems with long - range interaction ( self - gravitating stars or point vortices @xcite ) .
indeed , in the context of two - dimensional fluid - dynamics @xcite and plasma physics @xcite , the existence of an infinity of stationary states is known since a long time .
these slow relaxation processes have recently attracted considerable attention , since the hmf model can be considered as a simple paradigmatic model of long - range interactions , without the two additional difficulties of gravitational dynamics : singularity at short range and particle evaporation . as briefly recalled in the introduction , latora _ et al . _
@xcite have carefully analyzed an initial condition where all particles are located at the same position on the circle ( giving initially @xmath36 ) and momentum is uniformly distributed over a finite range , symmetrically around zero .
the system shows a fast relaxation towards a small magnetization state which persists for an extremely long time , that increases with @xmath0 .
the authors compare the momentum distribution of such a quasi - stationary state with tsallis distributions , obtaining some convincing fit of the central part of the distribution only after they impose a cut - off to momentum tails .
montemurro and zanette @xcite , analyzing the same initial condition , have even criticized the existence of a small magnetization plateau in time , by presenting some numerical evidence that magnetization first evolves towards a minimum , and then take off again towards the higher equilibrium value .
we will avoid this controversial point by using the definition of quasi - stationary state given in the introduction , i.e. we shall call quasi - stationary a state which still evolves , but on a time scale that diverges with @xmath0 .
hence , macroscopic properties are well defined over a sufficiently wide time span to allow local running time averages , even though the system slowly and continuously evolves towards equilibrium . at variance with most previous numerical experiments @xcite
, we choose here an initial state where the particles are uniformly distributed on the circle ( hence @xmath37 is initially close to zero , @xmath38 ) and momentum has a uniform distribution centered around zero , as above .
we make this choice for two reasons : _ i _ ) this state is a stationary state of the vlasov equation , that describes the hmf model in the @xmath34 limit , _ ii _ ) the @xmath39 state plays a relevant role also in the previous numerical experiments @xcite and we then thought that it is better to start directly from it . finally , we think that the analysis of the @xmath39 initial state can also clarify several aspects of the phenomenology of the @xmath36 initial state used by the authors above .
as explained in the introduction , we believe that the vlasov description is a useful and appropriate tool to understand the slow relaxation process . in the following section , we introduce and discuss the vlasov equation for the hmf model .
in the continuum limit , that is keeping the volume ( here the interval @xmath40 ) and the energy per particle fixed as the number of particles @xmath1 , the dynamics governed by eqs .
( [ eq : canonical ] ) is described by the vlasov equation .
the state of the finite @xmath0 system can be described by a single particle time - dependent density function @xmath41 where @xmath9 is the dirac function .
when @xmath0 is large , it is natural to approximate the discrete density @xmath42 by a continuous one @xmath43 . using this density , also called @xmath5-space distribution , it is possible to rewrite the two components of the magnetization @xmath37 given by eq .
( [ eq : magnetization ] ) .
they read @xmath44 & \equiv & \int f(\theta , p , t)\ , \cos \theta\ { \text{d}}\theta { \text{d}}p\quad,\\ \overline{m}_y\left [ f\right ] & \equiv & \int f(\theta , p , t)\,\sin \theta \ { \text{d}}\theta { \text{d}}p\quad .
\label{eq : magnetization_vlasovb}\end{aligned}\ ] ] within this approximation , one can write the potential that affects all the particles as @xmath45 = -\overline{m}_x\left [ f\right]\ , \cos \theta - \overline{m}_y\left [ f\right]\ , \sin \theta\quad.\ ] ] this potential enters the expression of the vlasov equation @xmath46 { \dfrac{\partial f}{\partial p } } = 0\quad,\ ] ] which governs the spatiotemporal evolution of the density @xmath15 . in the remaining of the paper , we will omit the over - bar on @xmath47 and @xmath48 for the sake of simplicity . + it is important to notice that the discrete distribution @xmath42 , a sum of dirac peaks , contains exactly the true dynamics of the system and is also a solution of the vlasov equation ( [ eq : vlasov ] ) . then , introducing a suitably defined distance on the space of probability measures on @xmath49 , it is possible to show @xcite that the distance between two solutions @xmath50 and @xmath51 of eq .
( [ eq : vlasov ] ) grows at most exponentially in time @xmath52 with a growth rate @xmath53 which is independent of the initial conditions .
this result , which is the essence of braun - hepp s theorem , heavily relies on the mean field character of the underlying hamiltonian dynamics and on the genericity of exponential instability of trajectories .
choosing then @xmath54 and taking for @xmath55 a continuous approximation of @xmath42 , one immediately obtains that @xmath56 as @xmath0 grows . as a consequence
, the previous result implies that the discrete particle dynamics converges to vlasov dynamics when @xmath34 , uniformly over all fixed time intervals @xmath57 $ ] .
however , for all fixed @xmath0 there is a typical time @xmath58 over which the two dynamics diverge . from this analysis ,
it is therefore natural to expect that particle and vlasov dynamics coincide during a time that diverges as @xmath20 , if @xmath59 and @xmath53 is independent of @xmath0 . however , for initial conditions corresponding to stable stationary solutions of the vlasov equation , this time may be much longer , actually of order @xmath0 @xcite .
all this explains why the properties of the vlasov equation are of particular interest for the study of the particle dynamics . in the next subsection
, we will study the vlasov dynamics of the hmf model and its stationary states , with the aim of getting useful insights on particle dynamics .
the vlasov equation inherits from the particle dynamics the conservation of the energy @xmath60
@xmath61 = \int \frac{p^{2}}{2}\ , f(\theta , p , t)\,{\text{d}}\theta { \text{d}}p + \frac{1}{2 } - \frac{m_{x}^{2}+m^{2}_{y}}{2}\ ] ] and of the total momentum @xmath62=\int p\,f(\theta , p , t)\,{\text{d}}\theta { \text{d}}p\quad.\ ] ] however , the casimir s functionals ( [ eq_casimir ] ) , defined for any continuous function @xmath22 , yield an infinity of additional conserved quantities , linked with the labeling symmetry when following a fluid particle in the @xmath63-space .
these new conserved quantities play of course a major role in the dynamics of the vlasov equation and thus of the particle dynamics . for a non stationary initial distribution , the dynamics of the vlasov equation
is known to give rise to a very complex nonlinear evolution , characterized by stretching and folding of the initial distribution , the details of this evolution being usually unpredictable .
however , one may predict the final evolution using statistical mechanics arguments , in the spirit of the statistical mechanics of two dimensional conservative flows @xcite or of the vlasov - poisson equation @xcite .
one ends up with the most probable coarse - grained distribution @xmath64 , which takes into account the dynamical invariants .
unfortunately , the dynamical mixing of the distribution is likely to be incomplete : the reason lies in the existence of infinitely many stable stationary solutions of the vlasov equation , in the neighborhood of which the system may be trapped
. therefore , vlasov dynamics quickly converges ( weak convergence ) to a stable stationary state which should be studied in detail .
equations ( [ eq : magnetization_vlasov ] ) and ( [ eq : magnetization_vlasovb ] ) show that the magnetization @xmath65 is constant for a stationary solution @xmath66 .
this implies that the potential @xmath67 is constant .
the equation for the stationary states of the vlasov equation may thus be considered as a linear first order partial differential equation .
solutions are then given by densities @xmath15 that are constant on the characteristics of the equation , corresponding to the level sets of the individual particle energy @xmath68 which corresponds to the energy of a pendulum .
it is important to observe that the total energy @xmath69 given by eq .
( [ eq : energy ] ) , is different from the sum of the individual energies ( [ eq : individual - energy ] ) .
smooth stationary solutions of the vlasov equation are thus given by @xmath70 where @xmath71 is any real function .
moreover , the values of @xmath72 , @xmath73 and the function @xmath71 must be self - consistent .
the boltzmann - gibbs equilibrium density @xmath74 is a particular case , although very important .
one may also prove that stationary states of the vlasov equation in a moving frame with constant velocity @xmath75 are given by @xmath76 .
let us note that the function @xmath71 may be multi - valued for individual energies @xmath77 greater than the energy of the separatrix ( one branch for particles with positive momentum and the other for negative momenta ) .
we will assume that this does not happen in the remaining of the paper , for the sake of simplicity ( a generalization would be straightforward ) .
as discussed above , the stationary states of the vlasov equation are not true stationary states of the particle dynamics .
if they are stable , however , they may explain the long lifetime of quasi - stationary states in the particle dynamics .
linear stability results for the stationary states of the vlasov equation have been already reported in the case of spatially homogeneous distributions for both gaussian @xcite and water bag @xcite momentum distributions . in this section
, we will show that it is possible to derive stability results for arbitrary spatially homogeneous stationary states , using a method developed in the context of two dimensional fluid dynamics and in plasma physics @xcite , based on original ideas introduced by arnold @xcite .
the authors of ref .
@xcite actually study _ nonlinear stability _
, a concept that we would like to briefly distinguish from other stability concepts . for a generic dynamical system ,
any extremum @xmath78 of a conserved quantity @xmath79 $ ] is a stationary point of the dynamics .
it is said to be _ formally stable _ if the second variations @xmath80 $ ] of @xmath81 is positive definite ( @xmath82 is then a minimum ) or negative definite ( @xmath82 is then a maximum ) . in the case of the linearized dynamics around a formally stable point @xmath83 , as the second variations of @xmath84 at @xmath83 are conserved , a small perturbation of @xmath85 remains bounded in the norm provided by the second variations : this state is _ linearly stable .
_ since this implies that the spectrum of the linearized dynamics does not have any negative value , the system is also _ spectrally stable .
_ it is however not true in general that spectral stability implies linear stability , and that linear stability implies formal stability .
finally , nonlinear stability corresponds to the case where a small perturbation , evolving according to the real dynamics , remains bounded in some norm .
it can be shown that nonlinear stability implies spectral stability , the converse being wrong in general , whereas formal stability implies nonlinear stability only in finite dimensional systems . in this section we prove that any stationary state of the vlasov equation , defined by eq .
( [ statstatevlasov ] ) with @xmath86 strictly decreasing , corresponds to a critical point of some invariant functional .
computing the second variations of this functional , we can therefore exhibit a _ necessary and sufficient condition of formal stability _ for such a stationary state .
let us consider the maximum of the functional @xmath87 = c_{s}\left [ f\right ] -\beta h_v\left [ f\right ] -\mu \int f(\theta , p , t)\ , { \text{d}}\theta { \text{d}}p \quad , \ ] ] where @xmath88 is the energy ( [ eq : energy ] ) , @xmath89 is a casimir functional ( [ eq_casimir ] ) corresponding to a strictly concave function @xmath22 and @xmath90 is positive .
performing the first variations of this functional , we obtain the equation @xmath91 which defines the critical points @xmath83 .
this yields @xmath92 with @xmath93 given by ( [ eq : individual - energy ] ) and @xmath94 the inverse function of @xmath95 , the derivative of @xmath96 .
the condition that @xmath96 is strictly concave is equivalent to the fact that @xmath94 is strictly decreasing .
the computation of the second variation of @xmath84 gives : @xmath97 = \int s''\left ( f_{0 } \left ( \theta , p\right ) \right ) \left [ \delta f\left ( \theta , p\right ) \right ] ^{2 } { \text{d}}\theta { \text{d}}p + \beta \left ( \left ( m_{x}\left [ \delta f\right ] \right ) ^{2 } + \left ( m_{y}\left [ \delta f\right ] \right ) ^{2}\right)~,\ ] ] where @xmath47 and @xmath48 are given by eqs .
( [ eq : magnetization_vlasov ] ) and ( [ eq : magnetization_vlasovb ] ) .
as @xmath22 is strictly concave , @xmath98 is negative .
the first term of @xmath99 $ ] is thus clearly negative whereas the second one is positive .
we will consider now only homogeneous states , corresponding to @xmath100 and @xmath101 .
let us introduce the fourier series of the perturbation @xmath102 from eq .
( [ eq : second_variation ] ) , after integration on the spatial variable @xmath103 , one obtains the second variations @xmath104 = \int s^{\prime \prime } \left ( f_{0}\left ( p\right ) \right ) \sum_{n>1}\left ( c^{2}_{n}\left ( p\right ) + s^{2}_{n}\left ( p\right ) \right ) { \text{d}}p + 2g\left ( c_{1}\left ( p\right ) \right ) + 2g\left ( s_{1}\left ( p\right ) \right)~,\ ] ] where we have introduced @xmath105 the terms involving @xmath106 and @xmath107 , @xmath108 , in eq .
( [ eq : second_variation_fourier ] ) , are clearly negative definite , since @xmath109 is strictly negative . consequently
, the second variations of @xmath84 are negative definite if and only if @xmath110 is negative definite .
the sign of the function @xmath110 can be studied using the cauchy - schwartz inequality : @xmath111 this inequality leads therefore to @xmath112 recalling that @xmath109 is strictly negative , we conclude that if the quantity @xmath113 is positive , the function @xmath110 is negative .
conversely , when it is negative , considering the particular function @xmath114 demonstrates that @xmath110 may be positive . differentiating with respect to the variable @xmath115 the equality @xmath116 obtained from eq .
( [ derivs ] ) in the case of homogeneous states , yields : @xmath117 as @xmath118 , the quantity ( [ criterai ] ) can be written as @xmath119 = 1 + \dfrac { 1}{2 } \int ^{+\infty } _ { -\infty } \dfrac { f'_{0}(p)}{p}{\text{d}}p\quad.\ ] ] this leads to the following equivalence : @xmath120 > 0\quad .\ ] ] this condition will of course be an excellent criterion to predict the stability of several initial conditions , and therefore the lifetime of the corresponding quasi - stationary states .
this is what we will consider now .
let us note that inagaki and konishi @xcite have found a dispersion relation for the linearized vlasov equation which leads to the above stability criterion .
hence , in this particular case , linear stability and formal stability criteria coincide .
[ [ waterbag - distribution ] ] waterbag distribution + + + + + + + + + + + + + + + + + + + + + one of the most widely used initial condition in previous numerical studies of the hmf model is the so called waterbag distribution @xmath121 if @xmath122 ( homogeneous state ) , the relation between energy @xmath123 and @xmath124 is @xmath125 .
computing the first derivative of @xmath126 @xmath127\quad,\ ] ] one obtains the following expression @xmath128=1-\dfrac { 1}{2}\dfrac { 1}{\bar{p}^{2}}\quad,\ ] ] exhibiting that the critical width of the distribution , above which the waterbag is formally unstable , is @xmath129 .
this corresponds to the critical energy density @xmath130 as reported earlier @xcite .
[ [ gaussian - distribution ] ] gaussian distribution + + + + + + + + + + + + + + + + + + + + + as a second example , let us consider a gaussian distribution , @xmath131 characteristic of an equilibrium canonical distribution . in this case , the quantity ( [ eq : testf ] ) can be easily reduced to @xmath132=1-\dfrac { \beta}{2 } \quad,\ ] ] emphasizing that the critical inverse temperature is @xmath133 and consequently the critical energy density @xmath134 .
this coincides with the critical point of the second order phase transition @xmath135 .
[ [ mixed - distribution ] ] mixed distribution + + + + + + + + + + + + + + + + + + finally let us consider a more general distribution , namely a mixed distribution between @xmath136 and @xmath137 , defined as @xmath138 thanks to the linearity of the quantity ( [ eq : testf ] ) with respect to the distribution , the critical energy density for this mixed distribution @xmath139 is obtained as a linear combination of both previous results : @xmath140 such a result allows to define a phase diagram of the dynamical critical energy that we will be able to confirm using numerical simulations .
we aim in the following at showing how these considerations about the vlasov equation and its stationary states are useful to understand the dynamics of the discrete particle system .
we begin in section [ sec : short - time ] by studying the short time evolution , and turn in section [ sec : intermediate - time ] to the intermediate and long time behaviour .
we have numerically integrated the canonical equations of motion ( [ eq : canonical ] ) , by using symplectic 4th - order integrators , the mclachlan - atela s @xcite or yoshida s @xcite algorithms .
the timestep @xmath141 was chosen to obtain an energy conservation with a relative accuracy @xmath142 better than @xmath143 .
we will consider initial conditions with uniform distribution with respect to @xmath144 as explained in section [ sec : stability ] .
the magnetization being consequently zero , these states are stationary solutions of the vlasov equation ( [ eq : vlasov ] ) . however , the numerical calculations correspond to simulations with a _
finite _ number of degrees of freedom , and these initial states are not anymore stationary a priori .
the spatial coordinates @xmath145 are randomly chosen in the interval @xmath40 , which leads to a magnetization of order @xmath17 initially .
the momenta @xmath146 are also randomly chosen from the given distribution @xmath147 satisfying the conditions for the energy @xmath148 whereas the conserved total momentum is set to zero @xmath149 we will present how the numerical results allow us to detect the critical energy for dynamical instability , using first the most widely used initial conditions , the waterbag distribution , and then the mixed ones .
figure [ fig : short - time - m ] presents the initial temporal evolution of the magnetization @xmath150 for different values of the energy density , but with the same number of particles , for a system initialized with a waterbag distribution in momentum and a homogeneous one for the angles(@xmath37 is close to zero since initially @xmath38 ) .
averages over a set of initial conditions ( samples ) are taken .
these results ( already partially reported elsewhere @xcite ) show that the initial time evolution of the magnetization , starting from such a non equilibrium initial condition , is quite different depending on whether @xmath13 is larger or smaller than @xmath151 .
if @xmath152 , but still below the second order phase transition energy @xmath33 , the magnetization remains close to the @xmath39 initial value and does not show any tendency towards the nonvanishing equilibrium value ( the long time relaxation to equilibrium will be discussed in section [ sec : long - time ] ) . for @xmath153 ,
instead , the magnetization shows a fast relaxation to a non vanishing value which is close to equilibrium .
relaxation proceeds through repeated oscillations that damp after a relatively short time . in order to characterize quantitatively this behavior ,
let us focus on the first peak of @xmath150 , by studying its height and its time of occurrence as a function of the energy density @xmath13 , as presented in fig .
[ fig : first - peak - a0.0 ] for increasing particle numbers . for different values of the energy density @xmath13 , when the number of particles is @xmath154 .
the values of @xmath13 are from @xmath155 to @xmath156 with @xmath157 step size .
the curves correspond to averages over 100 samples.,width=264 ] .
the curves correspond to different particle numbers @xmath158 and @xmath159 from top to bottom in @xmath152 for panel ( a ) and from bottom to top in @xmath153 for panel ( b ) , keeping in mind that every curve is obtained averaging over @xmath160 ( resp .
@xmath161 ) samples for @xmath162 ( resp .
@xmath163 ) .
the vertical line represents the theoretical critical energy density @xmath164 , given by eq .
( [ criticalwb ] ) .
[ fig : first - peak - a0.0],title="fig:",width=264 ] .
the curves correspond to different particle numbers @xmath158 and @xmath159 from top to bottom in @xmath152 for panel ( a ) and from bottom to top in @xmath153 for panel ( b ) , keeping in mind that every curve is obtained averaging over @xmath160 ( resp .
@xmath161 ) samples for @xmath162 ( resp .
@xmath163 ) .
the vertical line represents the theoretical critical energy density @xmath164 , given by eq .
( [ criticalwb ] ) .
[ fig : first - peak - a0.0],title="fig:",width=264 ] figure [ fig : first - peak - a0.0](a ) emphasizes that the first peak height vanishes in the energy region above @xmath165 as @xmath0 increases , in agreement with the theoretical predictions of the vlasov equation , which imply that the @xmath39 state becomes stable in this energy range in the continuum ( @xmath34 ) limit . the initial fluctuations of the magnetization , which are of order @xmath17
, do not grow when @xmath152 , because the waterbag state is vlasov stable . on the contrary ,
these fluctuations rapidly grow for @xmath153 , leading in a short time to a nonvanishing magnetization state which is close to equilibrium .
similar indications come from the first peak time , shown in fig .
[ fig : first - peak - a0.0](b ) . for @xmath152
this quantity clearly shows a convergence to an asymptotic value as @xmath0 increases , but this value sharply increases as one approaches @xmath165 from above , signaling the instability of the @xmath39 initial state .
the behavior below @xmath165 is less clear since the first peak time does not yet show the saturation expected on the basis of the vlasov equation for the values of @xmath0 considered here .
this might perhaps be due to the limited validity in time of the vlasov description when starting from an initially unstable state .
we will however show in the next section that at least the early exponential instability is well reproduced .
both quantities , the first peak height and the first peak time , are therefore useful indicators of the presence of a dynamical critical energy @xmath165 , as predicted theoretically in the continuum limit .
to summarize , we could say that on a @xmath4 time scale , and when starting from such a waterbag initial state , one would observe the instability of the @xmath39 state at @xmath165 instead of @xmath33 , the phase transition energy .
before extending these findings to other types of initial distributions in section [ sec : extended - initial - distribution ] , we will first show in the next section that the predictions of the vlasov equation are sufficiently precise to give the growth rate of the instability . as shown in fig .
[ fig : first - peak - a0.0](b ) , a typical time scale of the system ( the one of the first peak in magnetization ) shows a tendency to diverge at @xmath165 .
we show in this section that vlasov equation not only predicts this divergence , but is also quantitative in determining the time scale of the initial exponential instability of the magnetization for @xmath166 .
let us therefore define the exponential growth rate @xmath167 of an initial perturbation @xmath168 of a vlasov stationary unstable state as follows @xmath169 the perturbation we consider is around a homogeneous state , whose density function @xmath15 depends only on the variable @xmath115 .
however , the perturbation @xmath170 depends on all variables @xmath171 .
the linearized equation that governs the time evolution of the perturbation @xmath170 , for vanishing magnetization , is @xmath172 -\cos \theta\ , m_y[\delta f ] \right ) \frac{\partial f}{\partial p}=0\quad.\ ] ] using the development in fourier series of the perturbation @xmath170 given in eq .
( [ dvlpfourierdeltaf ] ) , we obtain the following equations for each fourier component @xmath173 one can easily show that the fourier components with @xmath174 can not be unstable .
indeed , by introducing the quantity @xmath175 , it is straightforward to show that its time derivative is @xmath176 .
hence , the generalized eigenvalues are all pure imaginary numbers , @xmath177 ( with @xmath178 ) , and the corresponding eigenvectors are dirac delta functions .
let us then concentrate on the @xmath179 components for waterbag initial conditions . using expressions ( [ eq : waterbag_deri ] ) for the derivative with respect to @xmath115 of the distribution ( [ eq : waterbag ] ) , eqs .
( [ eq : fourier2c ] ) and ( [ eq : fourier2d ] ) become @xmath180\int du\ , s_1(u , t ) \\
\frac{\partial s_1}{\partial t } & = & \ \
\ pc_1+\frac{1}{8\bar{p}}\left [ \delta(p+\bar{p})-\delta(p-\bar{p})\right]\int du\ , c_1(u , t ) \quad.\end{aligned}\ ] ] to solve this infinite dimensional eigenvalue problem , we restrict to functions @xmath181 and @xmath182 that are linear combinations of @xmath183 and @xmath184 .
this yields a four dimensional problem , whose eigenvalues can be calculated explicitly as functions of @xmath124 , and therefore as functions of the energy per particle @xmath13 .
one finds that above @xmath185 all eigenvalues are purely imaginary , indicating that the vlasov equation is linearly stable for such perturbations , in full agreement with the stability analysis developed in sec .
[ applicationsnonlinearstability ] . below @xmath165
, the largest eigenvalue @xmath167 that controls the growth rate is @xmath186 the exponential growth of perturbations of the initial distribution implies an exponential growth of magnetization @xmath150 .
this is indeed confirmed in fig .
[ fig : semi - log - m](a ) where the magnetization is plotted versus time in semi - logarithmic scale .
the comparison of the theoretical estimate ( [ eq:1st - ev ] ) of the growth rate with the numerical results , reported in fig .
[ fig : semi - log - m](b ) , shows a very good agreement .
moreover , the time scale @xmath187 diverges at @xmath165 . for @xmath163 particles , the number of samples being @xmath161 .
the values of @xmath13 are from @xmath155 to @xmath188 from top to bottom with @xmath189 step size .
the exponential growth rates of @xmath150 , @xmath167 , are estimated from the fitting solid lines .
( b ) comparison of the theoretical ( solid curve ) and numerical ( crosses ) growth rate @xmath167.,title="fig:",width=264 ] for @xmath163 particles , the number of samples being @xmath161 .
the values of @xmath13 are from @xmath155 to @xmath188 from top to bottom with @xmath189 step size .
the exponential growth rates of @xmath150 , @xmath167 , are estimated from the fitting solid lines .
( b ) comparison of the theoretical ( solid curve ) and numerical ( crosses ) growth rate @xmath167.,title="fig:",width=264 ] we present in this section the numerical results for the mixed distributions in formula ( [ eq : fa ] ) using the indicators introduced in section [ sec : first - peak ] . the first peak height , shown in fig .
[ fig : first - peak - height ] , shows a dependence on @xmath13 which is perfectly consistent with the theoretical prediction of the existence of a critical energy density @xmath190 , given in formula ( [ critmixed ] ) , above which the @xmath39 state is stable . however , the numerical results emphasize that the transition is much less abrupt for non vanishing values of the parameter @xmath191 than for @xmath192 ( the waterbag case previously analyzed ) .
the explanation of this effect is possibly twofold . on one hand , the stability criterion derived in sec .
[ sec : stability ] gives no information on the time evolution that begins from an unstable initial state : it may well be that the system evolves initially to states with smaller magnetization than for @xmath192
. on the other hand , this weaker instability below @xmath165 may be also due to the linear vanishing of the growth rate at @xmath165 characteristic of the gaussian initial distribution @xcite , instead of the sharper square root behaviour of the waterbag initial distribution , as expressed by eq .
( [ eq:1st - ev ] ) . for waterbag ( panel ( a ) ) , mixed with @xmath193 ( panel ( b ) ) and gaussian ( panel ( c ) ) initial conditions .
the vertical line represents the theoretical critical energy density @xmath194 , given by eq .
( [ critmixed ] ) .
the different curves correspond to the following values of @xmath0 : @xmath195 , @xmath196 , @xmath197 and @xmath159 from top to bottom for @xmath198 .
, title="fig:",width=177 ] for waterbag ( panel ( a ) ) , mixed with @xmath193 ( panel ( b ) ) and gaussian ( panel ( c ) ) initial conditions . the vertical line represents the theoretical critical energy density @xmath194 , given by eq .
( [ critmixed ] ) .
the different curves correspond to the following values of @xmath0 : @xmath195 , @xmath196 , @xmath197 and @xmath159 from top to bottom for @xmath198 .
, title="fig:",width=177 ] for waterbag ( panel ( a ) ) , mixed with @xmath193 ( panel ( b ) ) and gaussian ( panel ( c ) ) initial conditions . the vertical line represents the theoretical critical energy density @xmath194 , given by eq .
( [ critmixed ] ) .
the different curves correspond to the following values of @xmath0 : @xmath195 , @xmath196 , @xmath197 and @xmath159 from top to bottom for @xmath198 .
, title="fig:",width=177 ] the first peak time data , presented in fig .
[ fig : first - peak - time ] as a function of the energy density @xmath13 , do not follow sometimes a smooth curve
. this may be due to the fact that magnetization @xmath150 is almost flat in the region @xmath199 , and hence the first peak time is strongly affected by slight variations in the shape of the function @xmath150 .
nevertheless , the behavior of the first peak time is qualitatively the same as the one shown in fig .
[ fig : first - peak - a0.0](b ) for @xmath192 .
the same comments made there apply also to this case .
for waterbag ( panel ( a ) ) , mixed ( panel ( b ) ) and gaussian ( panel ( c ) ) initial conditions .
the vertical line represents the theoretical critical energy density @xmath194 , given by eq .
( [ critmixed ] ) .
the different curves correspond to the following values of @xmath0 : @xmath195 , @xmath196 , @xmath197 and @xmath159 from bottom to top in @xmath200.,title="fig:",width=177 ] for waterbag ( panel ( a ) ) , mixed ( panel ( b ) ) and gaussian ( panel ( c ) ) initial conditions .
the vertical line represents the theoretical critical energy density @xmath194 , given by eq .
( [ critmixed ] ) .
the different curves correspond to the following values of @xmath0 : @xmath195 , @xmath196 , @xmath197 and @xmath159 from bottom to top in @xmath200.,title="fig:",width=177 ] for waterbag ( panel ( a ) ) , mixed ( panel ( b ) ) and gaussian ( panel ( c ) ) initial conditions .
the vertical line represents the theoretical critical energy density @xmath194 , given by eq .
( [ critmixed ] ) .
the different curves correspond to the following values of @xmath0 : @xmath195 , @xmath196 , @xmath197 and @xmath159 from bottom to top in @xmath200.,title="fig:",width=177 ] we have shown in this section that the _ formal stability _ criterion of vlasov stationary states , stated in section [ sec : stability ] , allows to characterize the short time ( @xmath4 time scale ) behavior of the finite @xmath0 dynamics of the hmf model in the @xmath34 limit . in particular , it describes the behavior of the magnetization as a function of the energy per particle @xmath13 and the presence of several different instability thresholds depending on the detailed properties of the chosen initial distribution .
the critical energy for the instability of the gaussian distribution coincides with the thermodynamical phase transition point @xmath33 , as already remarked by inagaki @xcite
. moreover , the linear stability analysis of the vlasov equation gives results for the growth rates of the instability that are perfectly confirmed by numerical simulations .
several aspects remain to be ascertained . an important
one concerns the intermediate and long time evolution of an initial vlasov unstable state . in the following section , after introducing suitable indicators of stationarity we will follow the time evolution of such states for the hmf model .
a further important question is the ultimate fate of vlasov stable @xmath39 states for the finite @xmath0 dynamics of the hmf model . in the next section
, we will show that these states display indeed a much slower `` instability '' , that becomes evident on a time scale that increases with a nontrivial power law in @xmath0 .
these are examples of the _ quasi - stationary _ states that are the main object of study of this paper .
this section is devoted to the study of the intermediate and long time evolution of both initially unstable and stable vlasov stationary states .
as discussed in the previous section , the initial evolution is well described by vlasov dynamics .
unstable states quickly evolve ( on times of order 1 ) , until they approach , and are trapped , close to stationary states of the vlasov equation .
the system then evolves slowly through these stationary states , until it reaches boltzmann - gibbs statistical equilibrium . in this section
, we will construct indicators to assess if a @xmath0 particle state is close to a stationary state ( necessary conditions ) . using these indicators
, we will show that in all cases , during the relaxation towards equilibrium , the system always remains close to vlasov stationary states . in order to check these features
, we need to introduce necessary conditions for stationarity , and to perform numerical tests of non stationarity for the hmf model .
the tests guarantee non stationarity only , they do not guarantee stationarity . in order to obtain supporting evidences of stationarity , we introduce several indicators of non stationarity and observe whether all of them fail .
we take this fact as a good indication of stationarity . if @xmath201 is stationary , so will be the single particle energy distribution , @xmath202 .
this implies the stationarity of its moments @xmath203 the logical implications @xmath204 guarantees that the stationarity of the energy moments is a necessary condition for the stationarity of the distribution @xmath205 . in other words ,
the system is not stationary if at least one of the moments @xmath206 is not .
the computation of the time derivatives @xmath207 for the first four moments will be our test i for non - stationarity .
the previous necessary condition ( [ genstati ] ) is valid for all hamiltonian models .
it is possible to derive , using symmetry arguments , another necessary condition which is specific of the hmf model . indeed , using the individual particle energy ( [ eq : individual - energy ] ) , the vlasov equation ( [ eq : vlasov ] ) for a stationary distribution can be written as @xmath208 f_{st}(\theta , p ) = 0\quad.\ ] ] the solution of eq .
( [ eq : stationary - condition ] ) is constant on the characteristic curves @xmath209constant .
moreover , if one always resets to zero the phase of the magnetization vector @xmath210 ( i.e. @xmath211 ) , all characteristic curves will be symmetric with respect to a sign reversal of @xmath144 . hence
, if @xmath15 is stationary and the previous condition on the phase of @xmath210 is respected , then the stationary distribution obeys the following symmetry condition @xmath212 consequently , an asymmetry of the distribution @xmath205 with respect to the spatial variable @xmath144 , implies that the system is non stationary .
it is important to notice that the symmetry with respect to sign changes in momentum @xmath115 , @xmath213 , is not a necessary consequence of the stationarity of @xmath15 .
two separate characteristic curves for @xmath214 , where @xmath215 is energy of the separatrix , may correspond to different values of @xmath15 since they are _ not _ connected .
this symmetry is therefore not required even if the system is stationary .
a violation of this symmetry is not therefore an acceptable test of non stationarity .
obtaining the distribution @xmath205 numerically is , however , not easy since it is defined on the @xmath216-dimensional @xmath5-space at all fixed times @xmath19 , and therefore a huge number of samples would be necessary to obtain a good statistics .
a trick to easily check asymmetries in @xmath205 would be to verify the associated symmetry of the marginal distribution function @xmath217 however , unfortunately , @xmath218 may be symmetric even if the symmetry ( [ eq : symm - p ] ) is not satisfied .
typical examples of asymmetric distributions @xmath205 and corresponding symmetric marginal distributions @xmath218 are shown in figs .
[ fig : asymmetry](a ) and ( c ) , respectively . in order to recognize the asymmetry of @xmath205 with respect to @xmath144 , one must consider partially integrated distributions @xmath219 these latter distributions , besides being symmetric if the system is stationary , are able to correctly detect the asymmetries of @xmath11 distributions in @xmath220 , as shown in figs .
[ fig : asymmetry](a ) and ( c ) .
we quantitatively observe the time evolution of the asymmetry by introducing the two following quantities @xmath221^{2 } \,{\text{d}}\theta = \int { \text{d}}\theta \left [ \int_{p>0 } { \text{d}}p~ ( f(\theta , p , t)-f(-\theta , p , t ) ) \right]^{2}\quad,\\ a_{-}(t ) & = \int \left [ \tilde{f}_{-}(\theta , t ) - \tilde{f}_{-}(-\theta , t ) \right]^{2 } \,{\text{d}}\theta = \int { \text{d}}\theta \left [ \int_{p<0 } { \text{d}}p~ ( f(\theta , p , t)-f(-\theta , p , t ) ) \right]^{2}\quad . \end{split}\ ] ] both @xmath222 and @xmath223 exactly vanish when the distribution is symmetric , but they are affected by numerical errors and finite @xmath0 effects when estimated from numerical simulations . in order to reduce such effects
, we first define the quantity @xmath224^{2}.\ ] ] for initial conditions that are symmetric under the transformation @xmath225 , since this symmetry is conserved during the time evolution , @xmath226 for both stationary or non stationary distributions . of course ,
also this quantity is affected by numerical errors . to factor out such errors , we introduce the quantities @xmath227 and @xmath228 , which are large for asymmetric distributions similar to those in fig .
[ fig : asymmetry](a ) , and are instead @xmath4 for symmetric distributions ( actual values taken for the cases shown fig .
[ fig : asymmetry ] are @xmath229 for fig .
[ fig : asymmetry](a ) and @xmath230 for fig .
[ fig : asymmetry](b ) ) .
large values of the quantities @xmath231 will be our test ii for non stationarity . now that we have defined these two indicators , let us check non stationarity by considering both vlasov unstable and vlasov stable initial states with waterbag distribution , @xmath192 in eq.([eq : fa ] ) .
typical results of test - i and -ii are shown in fig .
[ fig : test - a0.0 ] . for vlasov unstable cases ( fig .
[ fig : test - a0.0](a ) ) , the quantity @xmath207 takes large values only in the time region @xmath232 .
similarly , the indicator @xmath233 is large in the time region @xmath234 .
we have reported only @xmath235 and @xmath227 , but we have checked that the behaviour is totally identical for higher moments @xmath236 , 3 and 4 and for @xmath228 , respectively .
this time region of non stationarity perfectly coincides with the region where the magnetization @xmath150 rapidly grows or largely fluctuates . for vlasov stable cases ,
@xmath207 exhibits small fluctuations only in the time interval @xmath237 , while @xmath233 remains small all the time .
we guess that the fluctuations in @xmath207 are due to finite size effects and would vanish in the @xmath34 limit .
note the logarithmic scale for the horizontal time axis : the increase in @xmath150 in fig.[fig : test - a0.0](b ) is much slower than the one in fig.[fig : test - a0.0](a ) .
the behavior shown in fig .
[ fig : test - a0.0](b ) is very important , because it shows that a variation in @xmath37 is compatible with the fact that the distribution @xmath15 may remain stationary ; i.e. the system can relax to equilibrium ( notably in the time region @xmath238 $ ] in fig .
[ fig : test - a0.0](b ) ) passing through stationary states of the vlasov equation .
it is natural to suppose in addition that these stationary states are stable .
particles and @xmath195 samples .
note the logarithmic scale for the time variable .
panels ( a ) ( resp .
( b ) ) presents the time evolution of the magnetization @xmath150 ( crosses ) , test i @xmath235 ( open squares ) and test ii @xmath227 ( full squares ) in the unstable case @xmath239 ( resp .
stable case @xmath240 ) .
the quantities @xmath241 ( resp .
@xmath233 ) are multiplied by a factor @xmath242 ( resp .
@xmath243 ) for graphical purposes .
the horizontal line represents the canonical value of @xmath37 .
[ fig : test - a0.0],title="fig:",width=283 ] particles and @xmath195 samples .
note the logarithmic scale for the time variable .
panels ( a ) ( resp .
( b ) ) presents the time evolution of the magnetization @xmath150 ( crosses ) , test i @xmath235 ( open squares ) and test ii @xmath227 ( full squares ) in the unstable case @xmath239 ( resp .
stable case @xmath240 ) .
the quantities @xmath241 ( resp .
@xmath233 ) are multiplied by a factor @xmath242 ( resp .
@xmath243 ) for graphical purposes .
the horizontal line represents the canonical value of @xmath37 .
[ fig : test - a0.0],title="fig:",width=283 ] figure [ fig : test - a1.0 ] reports results for two unstable gaussian initial conditions , @xmath244 in eq.([eq : fa])with @xmath239 and @xmath240 .
the behavior in panel ( a ) is very similar to the unstable waterbag case of figs .
[ fig : test - a0.0](a ) .
the case in panel ( b ) exhibits a small peak around @xmath245 for both test i and ii .
this presumably indicates that the system is weakly non stationary at this time , meaning that the temporal variations of the distributions @xmath15 are small .
it is also possible to explain the weakness of the non stationarity for gaussian initial conditions with the same arguments that we used when commenting fig .
[ fig : first - peak - height ] .
this state is unstable below the critical energy @xmath246 , but the equilibrium value for @xmath37 goes to zero when @xmath13 tends to @xmath135 .
hence , the initially unstable state may be very close to stationary stable states , and can be therefore immediately trapped into a stable state .
we expect the non stationarity to become weaker and weaker as @xmath13 approaches @xmath135 . , for a gaussian initial condition .
[ fig : test - a1.0],title="fig:",width=283 ] , for a gaussian initial condition .
[ fig : test - a1.0],title="fig:",width=283 ] the initial conditions used above correspond to stationary solutions of the vlasov equation . what happens with non stationary initial conditions ?
on the basis of what we have found above , we expect that the system reaches a stationary stable state , just like it does when starting from initially unstable stationary states ( figs .
[ fig : test - a0.0 ] and [ fig : test - a1.0 ] ) . to confirm this expectation ,
we have prepared the initial condition @xmath247 , and we have reported the results of tests i and ii in fig .
[ fig : test - ic1 ] . in the short time region @xmath248 ,
both @xmath241 and @xmath249 are large but they vanish for longer times , indicating that the system reaches some stationary state of the vlasov equation . summarizing the results of this section ,
we have found a strong support to the guess that , generically , the states corresponding to stationary stable solutions of the vlasov equation _ attract _ stationary unstable and non stationary solutions .
the system evolves in time passing through a series of stationary solutions of the vlasov equation . , for waterbag initial conditions in momentum @xmath115 whereas all the phases @xmath145 are set at @xmath250 , i.e. @xmath247 . the quantity @xmath227 is multiplied by @xmath251 for graphical purposes .
[ fig : test - ic1],title="fig:",width=283 ] , for waterbag initial conditions in momentum @xmath115 whereas all the phases @xmath145 are set at @xmath250 , i.e. @xmath247 . the quantity @xmath227 is multiplied by @xmath251 for graphical purposes .
[ fig : test - ic1],title="fig:",width=283 ] let us now consider the long time behaviour of the system .
we first discuss the behavior of the time evolution of the marginal distribution in angle @xmath252 , shown in fig .
[ fig : pdftheta - a0.0 ] . for @xmath239 , the initial condition is vlasov unstable ,
thus @xmath253 drastically changes from @xmath254 to @xmath255 ( see figs . [
fig : pdftheta - a0.0](a ) and [ fig : test - a0.0 ] ) . from @xmath255 to @xmath256 , the distribution is almost frozen and finally relaxes to the equilibrium distribution at @xmath257 . on the contrary , in the vlasov stable case , @xmath240 , ( see fig . [
fig : pdftheta - a0.0](b ) ) the distribution is almost frozen from @xmath254 up to @xmath258 : _ this is the crucial test of the presence of a quasi - stationary state associated to the vlasov stable initial condition_. meanwhile , the magnetization @xmath150 remains constant , well below the equilibrium value .
finally the distribution changes from @xmath256 to @xmath159 , and correspondingly the magnetization @xmath150 increases to reach equilibrium ( similar results have been obtained for gaussian initial distributions ) . for a waterbag initial condition , identical to the one used for fig .
[ fig : test - a0.0 ] . @xmath154 and the number of samples is @xmath195 .
the distributions at times @xmath254 ( crosses ) , @xmath255 ( open squares ) , @xmath259 ( full squares ) , @xmath260 ( open circles ) , are reported for @xmath239 ( a ) , and at @xmath254 ( crosses ) , @xmath261 ( open squares ) , @xmath259 ( full squares ) , @xmath260 ( open circles ) , for @xmath240 ( b ) . the solid curve represents the equilibrium distribution , @xmath262/c$ ] , where @xmath263 , @xmath264 and @xmath265 for @xmath239 and @xmath266 , @xmath267 and @xmath268 for @xmath240 .
[ fig : pdftheta - a0.0],title="fig:",width=264 ] for a waterbag initial condition , identical to the one used for fig .
[ fig : test - a0.0 ] . @xmath154 and the number of samples is @xmath195 . the distributions at times @xmath254 ( crosses ) , @xmath255 ( open squares ) , @xmath259 ( full squares ) , @xmath260 ( open circles ) , are reported for @xmath239 ( a ) , and at @xmath254 ( crosses ) , @xmath261 ( open squares ) , @xmath259 ( full squares ) , @xmath260 ( open circles ) , for @xmath240 ( b ) .
the solid curve represents the equilibrium distribution , @xmath262/c$ ] , where @xmath263 , @xmath264 and @xmath265 for @xmath239 and @xmath266 , @xmath267 and @xmath268 for @xmath240 .
[ fig : pdftheta - a0.0],title="fig:",width=264 ] in order to characterize the relaxation timescale with respect to the particle number @xmath0 in the case of the vlasov stable waterbag initial conditions ( @xmath240 ) , we have studied the temporal evolution of the magnetization @xmath150 ( a preliminary study has been already reported in ref .
this can be fitted by the function @xmath269~c(n ) + d(n),\ ] ] as shown in fig .
[ fig : m - apptanh](a ) .
the parameters @xmath270 and @xmath271 represent the half width between the initial and the equilibrium levels of @xmath150 and the initial level of @xmath150 , respectively ( we further comment about such parameters in the appendix ) .
the product @xmath272 is the slope at @xmath273 , i.e. @xmath274 , and @xmath275 is the timescale .
the most important parameter , @xmath276 , presented in fig .
[ fig : m - apptanh](b ) as a function of @xmath0 , is shown to be proportional to @xmath2 .
the fit is very good and excludes both the @xmath0 and the @xmath277 trivial scalings .
this nontrivial power law emerges unexpectedly , since theoretical arguments as well as previous studies of the hmf model suggest trivial divergences of the timescale ( typically as @xmath0 ) ( see ref .
quite interestingly however , zanette and montemurro @xcite have also investigated the timescale for the hmf model for @xmath240 with a waterbag initial distribution of momenta , but with a fully ordered initial state , i.e. @xmath247 .
they have shown that @xmath150 presents a minimum at @xmath278 ( already present in the simulation by latora et al @xcite ) which is also proportional to @xmath2 .
although our initial condition is different and hence different scalings could be found , we consider this finding as a confirmation of the presence of nontrivial timescales in the hmf model .
other , apparently unrelated , nontrivial timescales have been found , associated to the vanishing lyapunov exponent , for instance the well verified @xmath279 scaling @xcite . for different particles numbers : @xmath280 , @xmath281 , @xmath282 , @xmath283 , @xmath284 and @xmath285 from left to right , the number between brackets corresponding to the number of samples .
the horizontal line represents the equilibrium value of @xmath37 .
panel ( b ) shows the logarithmic timescale @xmath286 as a function of @xmath0 , whereas the dashed line represents the law @xmath287 .
[ fig : m - apptanh],title="fig:",width=264 ] for different particles numbers : @xmath280 , @xmath281 , @xmath282 , @xmath283 , @xmath284 and @xmath285 from left to right , the number between brackets corresponding to the number of samples .
the horizontal line represents the equilibrium value of @xmath37 .
panel ( b ) shows the logarithmic timescale @xmath286 as a function of @xmath0 , whereas the dashed line represents the law @xmath287 .
[ fig : m - apptanh],title="fig:",width=264 ] we were led by this discovery to rescale the momentum distributions @xmath288 by the nontrivial time - scale @xmath289 , see fig .
[ fig : momentumpdf ] .
the superposition of the curves at different times , in the quasi - stationary state , is quite impressive and we consider this as a strong test of the nontrivial timescale . at times
rescaled by @xmath2 for the vlasov stable case @xmath240 : @xmath154 ( pluses ) , @xmath290 ( crosses ) , @xmath291 ( stars ) and @xmath292 ( open squares ) .
the distributions are plotted for @xmath293 , for instance , @xmath294 for @xmath295 .
the four curves , once time - scaled , superpose almost perfectly.,width=264 ] there have been recently claims that momentum distributions in quasi - stationary states , obtained from the @xmath247 initial conditions , could be fitted with single particle distributions inspired by tsallis statistics @xcite . in this study
, we could not obtain any reasonable fit of the momentum distribution in fig .
[ fig : momentumpdf ] using several values of the tsallis @xmath3 parameter with @xmath296 .
moreover , the tails of the distribution are rather sharp , as shown by the inset in lin - log scale , and we can numerically exclude the presence of power law tails . it should be however mentioned that in another physical context , i.e. in the dynamics of self - gravitating @xmath0-body systems confined in an adiabatic wall , the long term time evolution has been argued to sweep through _ stellar polytropes _
, that are peculiar stationary solutions that maximize tsallis entropy @xcite .
however , chavanis @xcite , using the minimization of functionals built on conserved quantities ( called h - functions in this context ) , showed that tsallis entropy can be considered as a particular choice among the infinitely many possible h - functions : their maximization lead to particular stationary solutions of the vlasov dynamics .
the success of tsallis statistics in describing quasi - stationary regimes in n - body dynamics remains therefore doubtful .
further studies should fix this puzzling issue .
it should be moreover mentioned that in connection with some aging properties of the @xmath36 initial state , @xmath3-exponential have been used to fit quite efficiently numerical data of correlation functions @xcite .
in this paper , we have proposed a general framework , based on vlasov equation , to study the relaxation to equilibrium of the hmf model . in the short time regime ,
the behaviour of the system is well described by the vlasov equation , which is a valid approximation for finite times .
the stability of the initial conditions is therefore adequately characterized by considering the stability of vlasov stationary states .
we have checked the accuracy of this criterion in comparison with finite @xmath0 simulations .
in addition , a simple analytical derivation of the largest eigenvalue of the linearized vlasov operator gives an excellent estimate of the numerical timescale in the short time regime of the finite @xmath0 system . in the intermediate and long time regimes ,
two situations have been analyzed separately .
the first one , corresponding to unstable stationary vlasov states , shows a complex short time evolution before reaching some stationary stable state of the vlasov equation that is situated close to boltzmann - gibbs equilibrium .
the latter is finally reached on very long times , @xmath297 .
the second one , which begins from vlasov stable stationary states , evolves through other _ quasi - stationary _ states that are far from equilibrium and are among the many other stationary vlasov states .
the final relaxation to boltzmann - gibbs equilibrium takes place on times that increase with a nontrivial power law in the number of particles @xmath289 .
interestingly , this exponent is consistent with the one recently reported in ref .
@xcite for a different initial condition .
we have analyzed the momentum distributions in such a _ quasi - stationary _ regime and found that they scale properly with the nontrivial exponent .
as in @xcite , we have been unable to fit the momentum distributions using @xmath3-exponentials , as it had instead been done for other initial states in ref .
this is a negative indication that _ quasi - stationary _ states can always be described by tsallis statistics .
an analogy between the slow time evolution appearing in _ quasi - stationary _ states and the aging phenomenon in glasses and spin - glasses has been proposed in ref .
@xcite and further analyzed in ref .
however , none of these studies has been performed using stationary states of the vlasov equation as initial conditions .
it would be extremely interesting to repeat the study of correlation functions and of other glassy behavior indicators for such states , as already partially done in ref .
@xcite , where non stationary stretched exponential correlation functions have been found . after completing this paper
, we became aware of a further , and more complete , study by choi and choi @xcite of the linear stability of the homogeneous state in both the canonical ( fokker - planck ) and the microcanonical ( vlasov ) ensemble .
however , this paper does not give any answer to the question of the instabilities caused by the finiteness of the number of particles .
we warmly thank f. baldovin , a. campa , p - h .
chavanis , j - p .
eckman , y. elskens , d. escande , a. giansanti , v. latora , m. nauenberg , a. rapisarda , f. tamarit , a. taruya , c. tsallis , r. vallejos , c. villani for helpful discussions .
we also thank p - h .
chavanis , m. nauenberg and c. tsallis for their critical reading of the first version of this manuscript .
this work has been partially supported by eu contract no .
hprn - ct-1999 - 00163 ( locnet network ) , the french ministre de la recherche grant aci jeune chercheur-2001 n@xmath298 21 - 31 , and the rgion rhne - alpes for the fellowship n@xmath298 01 - 009261 - 01 .
this work is also part of the contract cofin00 on _ chaos and localization in classical and quantum mechanics _ and of the firb n. rbne01cw3m_01 project on synchronization .
is supported by e.u .
network _ stirring and mixing _
( rtn2 - 2001 - 00285 ) .
let us recall the fitting function we have used in section [ sec : long - time ] for the magnetization @xmath150 @xmath299~c(n ) + d(n)~.\ ] ] we have already commented about the scaling behavior of @xmath275 . according to fig .
[ fig : m - scaling ] , all the parameter set scales as follows : @xmath300 where @xmath301 is the equilibrium level of @xmath150 . and @xmath271 in panel ( a ) , and @xmath270 in panel ( b ) .
the symbols represent numerical results whereas the curves correspond to the fitted functions , given by eqs .
( [ eq : parameter - scaling]).,title="fig:",width=264 ] and @xmath271 in panel ( a ) , and @xmath270 in panel ( b ) .
the symbols represent numerical results whereas the curves correspond to the fitted functions , given by eqs .
( [ eq : parameter - scaling]).,title="fig:",width=264 ] by using the scaling functions ( [ eq : parameter - scaling ] ) , we can predict when @xmath150 reaches a given threshold , @xmath302 , as a function of @xmath0 .
the corresponding threshold time , @xmath303 , with @xmath304 , has the following expression @xmath305 let us consider , for numerical purposes , two threshold times for both threshold levels , @xmath306 and @xmath307 .
these two times roughly represent the beginning and ending times when @xmath150 grows towards the equilibrium value @xmath301 @xcite .
numerical results have fluctuations , particularly in the early time region , and hence we have defined @xmath308 and @xmath309 as follows : @xmath310 in fig . [
fig : t - mth ] , the comparison between the numerical results and the theoretical prediction ( [ eq : t_th ] ) for @xmath311 shows an excellent agreement , except in the small @xmath0 region .
we remark that both threshold times are approximately proportional to @xmath2 in the thermodynamic limit .
this further confirms the nontrivial scaling law . | we perform a detailed study of the relaxation towards equilibrium in the hamiltonian mean - field ( hmf ) model , a prototype for long - range interactions in @xmath0-particle dynamics . in particular , we point out the role played by the infinity of stationary states of the associated @xmath1 vlasov dynamics . in this context
, we derive a new general criterion for the stability of any spatially homogeneous distribution , and compare its analytical predictions with numerical simulations of the hamiltonian , finite @xmath0 , dynamics .
we then propose and verify numerically a scenario for the relaxation process , relying on the vlasov equation .
when starting from a non stationary or a vlasov unstable stationary initial state , the system shows initially a rapid convergence towards a stable stationary state of the vlasov equation via non stationary states : we characterize numerically this dynamical instability in the finite @xmath0 system by introducing appropriate indicators .
this first step of the evolution towards boltzmann - gibbs equilibrium is followed by a slow quasi - stationary process , that proceeds through different stable stationary states of the vlasov equation .
if the finite @xmath0 system is initialized in a vlasov stable homogenous state , it remains trapped in a quasi - stationary state for times that increase with the nontrivial power law @xmath2 .
single particle momentum distributions in such a quasi - stationary regime do not have power - law tails , and hence can not be fitted by the @xmath3-exponential distributions derived from tsallis statistics . , , , and hamiltonian dynamics , long - range interactions , vlasov equation , nonlinear stability , relaxation times . _
pacs numbers : _ + 05.20.-y classical statistical mechanics , + 05.45.-a nonlinear dynamics and nonlinear dynamical systems , + 05.70.ln nonequilibrium and irreversible thermodynamics , + 52.65.ff fokker - planck and vlasov equation . |
we begin by reviewing the formulation of the quark and gluon propagator on the lattice . in the continuum ,
the tree - level @xmath1 quark propagator is identified with the ( euclidean space ) fermionic greens function , @xmath2 where @xmath3 is the bare quark mass . in momentum space
this equation is solved straightforwardly , @xmath4 denote @xmath5 to be the tree - level propagator in momentum space . then in the presence of gauge field interactions , define @xmath6 to be the fourier transform of the ( interacting ) fermionic green s function , @xmath7 we define the mass function @xmath8 and the bare renormalization function @xmath9 such that the bare quark propagator has the form @xmath10 then for the renormalization point @xmath11 the renormalised quark propagator is given by @xmath12 where @xmath13 is the ( @xmath14-dependent ) renormalization function , and @xmath15 is the wave function renormalization constant , which depends on @xmath14 and the regulator parameter @xmath16 the @xmath17-dependence of @xmath18 is implicit .
@xmath15 is chosen such that @xmath19 as @xmath20 is multiplicatively renormalisable , all of the @xmath14-dependence is contained within @xmath21 that is , the mass function is @xmath14-independent .
the continuum tree - level gluon propagator in landau gauge is associated with the following greens function , @xmath22 in momentum space the tree - level gluon propagator takes the form , @xmath23 the non - perturbative gluon propagator is defined as the following two - point function , @xmath24 the scalar propagator @xmath25 is related to the full propagator ( in momentum space ) by @xmath26 the renormalised scalar propagator @xmath27 is chosen such that at some momentum scale @xmath28 @xmath29 we refer to @xmath30 as the gluon dressing function .
the overlap formalism @xcite in the vector - like case leads to the following definition of the massless overlap - dirac operator @xcite , @xmath31 here , @xmath32 is the matrix sign function applied to the overlap kernel @xmath33 the kernel can be any hermitian version of the dirac operator which represents a single fermion species of large negative mass .
the standard choice is the hermitian wilson - dirac operator ( setting @xmath34 ) , @xmath35 where @xmath36 is the central covariant finite difference operator , and @xmath37 is the lattice laplacian , or wilson term . in this work , as in previous studies , we use the flic action @xcite as the overlap kernel@xcite .
the flic fermion action is a variant of the clover action where the irrelevant operators are constructed using ape - smeared links @xcite , and mean field improvement @xcite is performed .
the hermitian flic operator is given by @xmath38 where the presence of fat ( smeared ) links and/or mean field improvement has been indicated by the super- and subscripts .
we choose @xmath39 $ ] and use a standard one - loop @xmath40 @xmath41 where @xmath42 is the elementary plaquette in the @xmath43 direction .
the ape smeared links @xmath44 constructed from @xmath45 by performing @xmath46 smearing sweeps , where in each sweep we first perform an ape blocking step , @xmath47 followed by a projection back into @xmath48 in this work , the projection is defined by first performing a projection into @xmath49 @xmath50^{-{\frac{1}{2}}},\ ] ] and then projection into @xmath51 @xmath52{\det u'(v ) } } u'(v).\ ] ] as it is only the product @xmath53 that matters @xcite , we fix @xmath54 and only vary @xmath55
mean field improvement is performed by making the replacements @xmath56 where @xmath57 and @xmath58 are the mean links for the standard and smeared gauge fields .
we calculate the mean link via the fourth root of the average plaquette , @xmath59
it is easily seen that the continuum massless quark propagator anti - commutes with @xmath60 @xmath61 a straightforward consequence of the ginsparg - wilson relation is the inverse of the overlap dirac operator satisfies @xmath62 noting that the mean link is a function of @xmath17 and that @xmath63 as @xmath64 we obtain @xmath65 it is then natural to define the ( external ) massless bare overlap propagator on the lattice as @xcite @xmath66 as we then have that @xmath67 as in the continuum case .
the massive overlap dirac operator is given by @xcite @xmath68 with @xmath69 representing fermions of mass @xmath70 the massive ( external ) bare overlap propagator is defined as @xcite @xmath71 and with the identification @xmath72 satisfies @xmath73 in order to construct @xmath8 and @xmath9 on the lattice , we first define @xmath74 such that @xmath75 then @xmath76 , \\ { \cal{b}}(p)&= \frac{1}{n_{\rm s}n_{\rm c}}\operatorname{tr}[s^{\rm bare}(p)],\end{aligned}\ ] ] where the trace is over color and spinor indices only , and @xmath77 specify the dimension of the spinor and color vector spaces .
now , define the functions @xmath78 such that the inverse of the bare ( lattice ) quark propagator has the form @xmath79 then it is easily seen that @xmath80 where @xmath81 the kinematical lattice momentum @xmath82 is defined such that at tree - level @xmath83 that is @xmath84 we note that the simple form of these relations is one of the advantages of overlap fermions , as the absence of additive mass renormalization prevents the need for having to perform any tree - level correction @xcite ( necessary most other fermions actions @xcite ) , outside of identifying the correct momentum variable @xmath85 now , we define @xmath86 such that @xmath87 the mass function @xmath8 and renormalization function @xmath9 may then be straightforwardly constructed , @xmath88
we use a tadpole - improved plaquette plus rectangle gluon action , @xmath89.\end{gathered}\ ] ] the lattice gauge field may be related to a continuum gauge field through a path ordered exponential , @xmath90 we can recover the lattice gluon field through the following `` mid - point '' definition , @xmath91 as in the continuum , the lattice gluon propagator in coordinate space is then given by @xmath24 a fourier transform takes us to momentum space , and we calculate the scalar propagator @xmath25 directly .
tree level improvement of the scalar gluon propagator is performed by identifying the appropriate kinematical momenta @xmath92 such that @xmath93 we denote the two different ( gluonic and fermionic ) kinematical momentum by @xmath94 and use context to distinguish them .
.parameters for the different lattices .
all use a tadpole improved luscher weisz gluon action .
lattice spacings are set via a string tension analysis incorporating the lattice coulomb term .
shown is the lattice volume , gauge coupling @xmath95 lattice spacing , pion mass ( for the dynaical quarks ) in mev , the mean link and the physical volume . [
cols="^,^,^,^,^,^,^",options="header " , ] each lattice ensemble consists of 50 configurations .
the lattice fermionic green s function is obtained by inverting the flic overlap dirac operator on each configuration using a multi - mass cg inverter @xcite .
we use periodic boundary conditions in the spatial directions and anti - periodic in the time direction .
the fourier transform is taken to convert to momentum space , and the the bare quark propagator is obtained from the ensemble average .
the quark propagator is calculated for 15 different masses .
the details of the flic overlap parameters used are presented in table [ tab : flicoverprop ] .
the matrix sign function in the flic overlap is evaluated using the zolotarev rational polynomial approximation @xcite , of degree ( typically 8) chosen to give an accuracy of @xmath96 within the spectral range of the kernel , after projecting out low - lying modes .
the tree level propagator is calculated by setting the links and the mean link to unity ( the free theory ) .
the kinematical lattice momentum @xmath92 is obtained numerically from the tree level propagator , although it could equally well have been obtained from the analytic form for @xmath92 derived in ref .
@xcite .
the gluon propagator is obtained by performing a fourier transform of the lattice gluon field @xmath97 using the midpoint definition and then calculating the scalar propagator directly .
we first examine the gluon propagator results .
figure [ fig : gluondata ] shows the bare gluon dressing function @xmath98 on each of the lattices .
the squares represent momenta entirely in a spatial cartesian direction while triangles indicate points where the momentum is entirely along the temporal axis .
the separation between these is due to the use of an asymmetric lattice and is a clear indication of finite volume effects .
this is to be expected on such small physical volumes .
the quenched and dynamical renormalised gluon dressing functions @xmath99 are compared directly in figure [ fig : gluoncomp ] .
we see that the presence of dynamical quarks causes the suppression of the dressing function in the infrared .
this effect is clear even with the relatively heavy sea quarks used in this study .
the same effect was seen in work which compared the gluon propagator on quenched and dynamical lattices with a staggered fermion action@xcite .
we now turn to the quark propagator results .
the mass function @xmath8 and the renormalization function @xmath13 are calculated on each of the lattices .
a cylinder cut @xcite is applied to all the data , to reduce the effects of rotational symmetry violation .
the full results for the dynamical lattices are displayed in fig .
[ fig : massfunc ] for @xmath8 and fig .
[ fig : zfunc ] for @xmath100 the renormalization point @xmath14 for @xmath13 is chosen to be approximately @xmath101 the full results for the quenched flic overlap quark propagator presented in previous work@xcite . in order to compare the quenched and dynamical results , for each fixed momenta value we perform a quartic fit of both @xmath8 and @xmath9 as a function of the pion mass ( excluding the lightest two masses for finite volume reasons ) . the data and fits for the smallest 10 momenta on each of the lattices are shown in fig .
[ fig : mqchifit ] for @xmath8 and fig .
[ fig : zqchifit ] for @xmath102 having obtained the fit functions @xmath103 and @xmath104 we then choose @xmath105 and compare the functions at these values of the pion mass . in this way we can compare the quenched and dynamical results with matched pion masses .
results are shown in figure [ fig : flicvsflic ] , to which we now turn our attention .
first we examine the mass function .
the results for @xmath106 show little effect due to the presence of sea quarks , just a slight suppression in the infrared for the lighter masses .
the heavier masses show essentially no difference between the quenched and dynamical results , execpt in the extreme ultraviolet , where the discretisation errors give some difference .
the discretisation errors are not present in the @xmath107 results , but now we can clearly see infrared suppression of @xmath8 in the dynamical results for all masses
. turning to the renormalization function , both the lattices show infrared suppresssion of @xmath108 in the presence of dynamical quarks , and agreement in the far ultraviolet . at @xmath107
the quenched and dynamical results also agree in the intermediate momentum regime , whereas there is a slight discrepancy in this region at @xmath109
the gluon propagator and flic overlap quark propagator are compared on quenched and dynamical flic lattices at @xmath107 and @xmath109 we observe the the presence of dynamical quarks causes suppression in the infrared in all quantities examined : the gluon dressing function @xmath110 the mass function @xmath8 and the quark renormalization function @xmath111 some differences between the results for the two lattice spacings indicate we are not quite in the scaling region at the coarser spacing .
an alternative source of systematic error in the the quark propagator results is the simplistic fitting functions used , which do not take into account the possibility of non - analytic behaviour at light quark mass .
however , our results are consistent with those obtained elsewhere@xcite .
we thank both the south australian partnership for advanced computing ( sapac ) and the australian partnership for advanced computing ( apac ) for generous grants of supercomputer time which have enabled this project .
this work is supported by the australian research council .
wk is supported by sfi basic research grant 04/br / p0266 .
jbz is partly supported by chinese nsfc - grant no .
10675101 . c. morningstar and m. j. peardon , phys .
d * 69 * , 054501 ( 2004 ) [ arxiv : hep - lat/0311018 ] .
j. m. zanotti , b. lasscock , d. b. leinweber and a. g. williams , phys .
d * 71 * , 034510 ( 2005 ) [ arxiv : hep - lat/0405015 ] . s. boinepalli , w. kamleh , d. b. leinweber , a. g. williams and j. m. zanotti , phys .
b * 616 * , 196 ( 2005 ) [ arxiv : hep - lat/0405026 ] .
r. narayanan and h. neuberger , phys .
b302 * , 62 ( 1993 ) , hep - lat/9212019 . | in this work we examine the fat - link irrelevant clover ( flic ) overlap quark propagator and the gluon propagator on both dynamical and quenched lattices .
the tadpole - improved luscher - weisz gauge action is used in both cases .
the dynamical gauge fields use the flic fermion action for the sea quark contribution .
we observe that the presence of sea quarks causes a suppression of the mass function , quark renormalization function and gluon dressing function in the infrared .
the ultraviolet physics is unaffected .
quark and gluon propagators are fundamental quantities in qcd encoding the rich nonperturbative and perturbative properties of qcd . in previous studies , the fat - link irrelevant clover ( flic )
overlap quark propagator and the gluon propagator were examined on quenched lattices@xcite . in this work
we study the effects of two flavor dynamical flic fermions @xcite on these quantities for the first time . the use of smeared or `` fat '' links in lattice fermion actions
has been of interest for some time @xcite and flic fermions @xcite have shown a number of promising advantages over standard lattice actions .
the flic fermion action is a wilson - clover type fermion action in which the irrelevant operators of the wilson and clover terms are constructed using fat links , while the relevant operators use the untouched ( thin ) gauge links . by smearing only the irrelevant , higher dimensional terms in the action , and leaving the relevant dimension - four operators untouched ,
short distance quark and gluon interactions are retained .
fat links may be created by any number of smearing prescriptions including ape smearing @xcite , hypercubic smearing @xcite , stout - link smearing @xcite or the unit - circle smearing @xcite used herein .
the use of fat links in the flic action minimizes the effect of renormalization on the action improvement terms . scaling studies indicate flic fermions provide a new form of nonperturbative @xmath0 improvement @xcite where near - continuum results are obtained at finite lattice spacing .
access to the light quark mass regime is enabled by the improved chiral properties of the lattice fermion action @xcite . in particular , the histogram of the additive mass renormalization encountered in chiral - symmetry breaking wilson - type fermion actions is seen to narrow upon introducing fat - links in the irrelevant operators .
these benefits facilitate the generation of the dynamical fermion configurations examined herein . |
the nuclei around @xmath7 region recently attracted considerable experimental and theoretical attention @xcite .
the excitation of one neutron from @xmath4 to @xmath6 orbital is responsible for high - spin level structure of as , sr , y , zr , mo , tc , ru and rh nuclei , with @xmath8 @xmath9 50 @xcite .
the importance of monopole fitted effective interaction , inclusion of higher orbitals across @xmath7 shell and seniority quantum number are discussed in refs .
@xcite .
recently , high - spin level structure of the semi - magic nucleus @xmath10nb has been investigated via @xmath11se(@xmath12n,@xmath13)@xmath10nb reaction @xcite .
the importance of neutron particle - hole excitation across @xmath7 shell gap is reported to describe the states above @xmath9 @xmath14 .
the high - spin level structures of near - spherical nuclei @xmath0zr have been reported in @xcite . as we move from @xmath10zr to @xmath15zr , a possible reduction in the energy gap between @xmath2 and @xmath3 orbitals
are indicated , since one more neutron is excited across the @xmath7 gap . for @xmath15zr ,
et al . _ , reported the pure neutron configuration to the @xmath16 state , while the second excited quadrupole state shows the signatures of the one - phonon mixed - symmetric @xmath17 state @xcite .
the results of an in - beam study of high - spin structure of @xmath18mo at gasp , legnaro was reported in @xcite . in this work
the authors also reported shell model results in @xmath19 and @xmath20 space . with this space
the yrast states @xmath21 and @xmath22 in @xmath23mo and those up to @xmath24 and @xmath25 in @xmath26mo are explained .
the importance of possible contribution from the @xmath27 , @xmath28 core , especially that of neutron excitations from the @xmath4 orbital has been reported in this work . in the present paper ,
we report a systematic study of shell - model results for @xmath0zr .
the aim of present work is to explain the structure of recently available high - spin experimental data for these two isotopes @xcite .
the work is organized as follows : details of the calculation and effective interaction are given in section 2 , the shell model results and a comprehensive comparison with experimental data and wave - function analysis are given in section 3 . finally , concluding remarks are drawn in section 4 .
in this work the results of recently available experimental data of @xmath0zr have been interpreted with large scale shell model calculations .
the model space consists of full @xmath1 , @xmath2 , @xmath3 , @xmath4 orbitals for protons and valence neutrons in @xmath3 , @xmath4 , @xmath5 , @xmath6 , @xmath29 , @xmath30 orbitals .
since the dimension of matrices are very large , we allowed neutrons to occupy only @xmath31 orbital .
we have completely filled @xmath3 orbital for neutrons and allowed one neutron excitation from @xmath4 orbital to @xmath6 and @xmath5 orbitals .
the calculations have been carried out with gwbxg effective interaction .
the single - particle energies ( in mev ) for protons orbital are @xmath1= -5.322 , @xmath2= -6.144 , @xmath3= -3.941 , @xmath4 = -1.250 , and for neutrons orbital are @xmath3 = -0.696 , @xmath4 = -2.597 , @xmath5 = + 5.159 , @xmath6= + 1.830 , @xmath29 = + 4.261 , and @xmath30 = + 1.741 .
we have performed shell model calculation with shell model code nushellx @xcite .
the gwbxg effective interaction is constructed with different interactions .
the original 974 two - body matrix elements ( tbmes ) were taken from bare g - matrix of h7b potential @xcite .
the 65 tbmes for proton orbitals were replaced with the effective values of ji and wildenthal @xcite . the tbmes connecting the @xmath32 and the @xmath33 orbitals were replaced by those from the work of gloeckner @xcite , and those between the @xmath34 and the @xmath35 orbitals were replaced with those of serduke _ et al .
previously , shell model results with this interaction for the positive - parity yrast states of the neutron - rich @xmath36rb , @xmath15y , and @xmath37y nuclei were reported in ref .
@xcite .
, width=650 ] zr into their @xmath38 components .
the percentage above 10% are written inside the squares , drawn with an area proportional to it .
percentage below 5% are not written .
, width=650 ]
results for energy levels of @xmath10zr are shown in fig .
the energy levels of positive parity states @xmath39 - @xmath40 , are compared with the available experimental data .
the overall agreement with experimental data is good .
the first excited state @xmath41 is 282 kev below than the experimental data , while @xmath42 is 139 kev above .
the majority of states are in very good agreement with experimental data with a difference of only few kev . in the case of negative parity ,
the calculation predicts a flip of levels for @xmath43 and @xmath44 .
the experimentally observed plausible state ( @xmath44 ) at 2765 kev predicted at 2924 kev .
the calculated @xmath45 state is 181 kev lower than the experimental value .
the ground state @xmath46 of @xmath10zr is showing single - particle character with @xmath47 configuration .
the @xmath48 state has @xmath49 configuration .
the states @xmath50- @xmath51 mainly come from @xmath52 ) configuration .
this configuration gives fully aligned @xmath51 state .
the states beyond this come from the breaking of @xmath53 subshell and also with neutron excitation from @xmath6 @xmath54 @xmath5 orbital .
the structure of @xmath55-@xmath56 states come from @xmath57 configuration .
the @xmath58 state has @xmath59 configuration . from @xmath58 to @xmath60 , proton empties from @xmath3 to @xmath4 and configuration for @xmath60 becomes @xmath61 .
the states , @xmath62 come from @xmath63 , while @xmath64 has @xmath65 configuration
. the @xmath66 state at 6711 kev has @xmath65 configuration .
analysis of wave functions also helps us to identify which nucleon pairs are broken to obtain the total angular momentum of the calculated states .
the two components for neutrons and protons are @xmath67 and @xmath68 , respectively .
these components are coupled to give the total angular momentum of each state . in the fig .
2 , we have shown results of positive and negative parity states of @xmath10zr .
the positive parity states @xmath69 - @xmath70 come from @xmath71 @xmath72 @xmath73 .
the dominant components are 84% ( @xmath69 ) , 92% ( @xmath70 ) .
the @xmath74 comes from @xmath71 @xmath72 @xmath75 ( 90% ) , while for @xmath55 , the configuration is @xmath76 @xmath72 @xmath73 ( 69% ) and @xmath77 @xmath72 @xmath73 ( 26% ) .
decomposition of the total angular momenta for the negative parity states @xmath78 - @xmath79 are also shown in fig .
2 . the @xmath79 state comes from @xmath80 @xmath72 @xmath73 ( 89% ) . the high - spin plausible positive parity states ( 27/2 ) , ( 29/2 ) , ( 31/2 ) and ( 33/2 ) at 7211.4 , 7663.7 , 8358.2 , 9129.8 kev , are predicted by shell model at 6164 , 7125 , 8166 , 8465 kev , respectively .
the configuration of 27/2@xmath81 is @xmath77 @xmath72 @xmath73 ( 94.4% ) ; 29/2@xmath81 is @xmath77 @xmath72 @xmath75 ( 94.3% ) ; 31/2@xmath81 is @xmath82 @xmath72 @xmath75 ( 79% ) ; 33/2@xmath81 is @xmath71 @xmath72 @xmath83 ( 41% ) .
the configuration for 33/2@xmath81 is @xmath84 ) .
, width=650 ] zr into their @xmath38 components .
the percentage above 10% are written inside the squares , drawn with an area proportional to it .
percentage below 5% are not written .
, width=650 ] the results of @xmath15zr are shown in fig .
the dominant configuration of @xmath85 , @xmath17 , and @xmath86 states is @xmath87 .
the @xmath88 and @xmath89 states are at a large gap of 1.5 mev from @xmath90 state because we need more energy to excite protons from @xmath91 to @xmath92 orbital to form these two states ( [ @xmath93_{6^+,8^+}$ ] ) . the higher - angular momentum states @xmath94 associated with @xmath95 configuration . beyond @xmath96 ,
other high - spin states arise from excitation of one proton / neutron across the @xmath53/@xmath97 shell , i.e. one proton excitation from @xmath98 @xmath54 @xmath4 and one neutron from @xmath6 @xmath54 @xmath5 .
the @xmath99 state at 6591 kev is due to @xmath100 @xmath72 @xmath101 configuration . after looking the structure and occupancy of three eigen values of calculated @xmath102 states .
we found that @xmath21 and @xmath103 have same occupancy of @xmath5 and @xmath6 orbitals [ also @xmath104 value for this transition is large ] .
thus we have taken @xmath103 as first state and next excited state is @xmath105 after this .
the structure of @xmath103 state at 7063 kev is changed from @xmath99 state and it becomes @xmath106 @xmath72 @xmath101 .
the average occupancy of @xmath6 orbital for @xmath21 and @xmath99 states is @xmath9 1.09 , while for @xmath103 state it is 1.83 .
this is also reflected from the structure change from @xmath99 to @xmath103 .
the present study predicts that the @xmath107 to @xmath108 are fully aligned states of the @xmath106 @xmath72 @xmath101 configuration .
the negative parity state @xmath109 comes from @xmath110 configuration .
the @xmath111 to @xmath112 states come from @xmath110 @xmath72 @xmath87 .
the high - spin isomer @xmath113 observed in @xmath15zr is reported at 8.041 mev @xcite , we have calculated this state at 7.781 mev .
the main component of the wave function of this state is @xmath114 .
the wave function analysis for @xmath15zr is shown in fig 4 , the @xmath102 ( 4618 kev ) comes from @xmath71 @xmath72 @xmath115 ( 87% ) ; the @xmath102 ( 6591 kev ) comes from @xmath71 @xmath72 @xmath115 ( 69% ) ; @xmath107 comes from @xmath116 - @xmath117 @xmath72 @xmath115 ; @xmath118 comes from @xmath76 - @xmath119 @xmath72 @xmath115 and @xmath108 comes from @xmath77 @xmath72 @xmath115 ( 94% ) .
the composition of negative parity state @xmath109 comes from @xmath120 ( 64% ) configuration .
the @xmath111 comes from , @xmath120 @xmath72 @xmath121 ( 82% ) and @xmath122 @xmath72 @xmath123 ( 11% ) .
the @xmath124 comes from , @xmath125 @xmath72 @xmath123 ( 82% ) , while @xmath112 comes from @xmath120 @xmath72 @xmath123 ( 95% ) .
the experimentally observed high - spin plausible positive parity states ( 18 ) and ( 19 ) at 9129.6 , 9724.2 kev , predicted by shell model at 9296 , 10412 kev , respectively .
the configuration of 18@xmath81 is @xmath82 @xmath72 @xmath115 ( 77% ) . for 19@xmath81
the major configuration is @xmath126 @xmath72 @xmath115 ( 52% ) and other dominant configuration is @xmath127 @xmath72 @xmath115 ( 35% ) .
there are two - ways to generate high - spin angular momentum due to single - particle effects .
first is by pair - breaking and second is due to particle - hole excitations . thus creating one - hole
( @xmath128 ) configuration is important for generating high - spin states .
the calculated energy levels of low - lying states for @xmath15zr are more compressed .
thus , we have also performed second calculation ( calc .
ii ) for @xmath10zr and @xmath15zr without neutron excitation from @xmath129 to @xmath47 orbital , i.e. we have completely filled the @xmath130 orbital .
the results are shown in fig .
5 . in the case of @xmath15zr ,
the results for low - lying states 6@xmath81 - 8@xmath81 - 10@xmath81 - 12@xmath81 show good agreement with the experimental data .
previously , shell model results for @xmath15zr with the same effective interaction ( tbmes ) for the proton ( @xmath3 , @xmath4 ) and neutron ( @xmath6 , @xmath30 ) orbitals are reported in ref .
the low - lying states in this restricted model space show good agreement with the experimental data .
although , high - spin states are not in good agreement without neutron excitation from @xmath129 to @xmath47 orbital , as we have shown in figs . 1 and 3 . the second set of calculation predicting @xmath88 and @xmath131 at 3355 and 3944 kev , respectively .
thus reported @xmath132 state is the first state . in the table 1 , we have shown the decomposition of the total angular momentum for low - lying states of @xmath15zr .
the two calculations giving different configuration for @xmath132 states .
zr and @xmath15zr . first calculation with one neutron excitation from @xmath4 orbital to @xmath6 and @xmath5 orbitals ( calc .
i ) and second calculation with completely filled @xmath130 orbital ( calc .
ii).,title="fig:",width=529,height=665 ] zr and @xmath15zr .
first calculation with one neutron excitation from @xmath4 orbital to @xmath6 and @xmath5 orbitals ( calc .
i ) and second calculation with completely filled @xmath130 orbital ( calc .
ii).,title="fig:",width=529,height=665 ] cc states & configuration + @xmath16 calc .
i & @xmath133(82% ) + .. + calc .
ii & @xmath133(91% ) + .. + @xmath134 calc .
i & @xmath133(42% ) + @xmath135(33% ) + @xmath133@xmath72@xmath135 ( 12%)+ .. + calc .
ii & @xmath133(35% ) + @xmath135(40% ) + @xmath133@xmath72@xmath135 ( 13%)+ .. + @xmath136 calc .
i & @xmath133(46% ) + @xmath137@xmath72@xmath135 ( 25% ) + .. + calc .
ii & @xmath133(28% ) + @xmath135 ( 41% ) + .. + @xmath90 calc .
i & @xmath138(93% ) + .. + calc .
ii & @xmath138(97% ) + .. + @xmath139 calc .
i & @xmath138(12% ) + @xmath140(25% ) + @xmath133@xmath72@xmath135 ( 43%)+ .. + calc .
ii & @xmath133@xmath72@xmath135 ( 46%)+ @xmath140(29% ) + .. + @xmath141 calc .
i & @xmath138(54% ) + @xmath140(12% ) + @xmath138@xmath72@xmath135 ( 10%)+ .. + calc .
ii & @xmath138(47% ) + @xmath140(19% ) + @xmath133@xmath72@xmath135 ( 13%)+ .. + @xmath88 calc .
i & @xmath142(51%)+ @xmath142@xmath72@xmath135 ( 37%)+ .. + calc .
ii & @xmath138@xmath72@xmath135 ( 10%)+ @xmath133@xmath72@xmath140 ( 33%)+@xmath143 ( 44%)+ .. + @xmath131 calc .
i & @xmath138@xmath72@xmath135 ( 13%)+@xmath133@xmath72@xmath140 ( 34%)+@xmath143 ( 38%)+ .. + calc .
ii & @xmath138@xmath72@xmath135 ( 24%)+ @xmath133@xmath72@xmath140 ( 19%)+ + & @xmath143 ( 32%)+@xmath133@xmath72@xmath143 ( 16%)+ .. + @xmath144 calc .
i & @xmath138@xmath72@xmath135 ( 33%)+@xmath133@xmath72@xmath140 ( 16%)+@xmath143 ( 30%)+ .. + calc .
ii & @xmath138@xmath72@xmath135 ( 29%)+ @xmath133@xmath72@xmath140 ( 18%)+@xmath143 ( 14%)+ + & + @xmath138@xmath72@xmath140 ( 13%)+@xmath133@xmath72@xmath143 ( 22% ) .. + @xmath89 calc .
i & @xmath145(48%)+ @xmath133@xmath72@xmath146 ( 37%)+ .. + calc .
ii & @xmath145(53%)+ @xmath133@xmath72@xmath146 ( 35%)+ .. + @xmath147 calc .
i & @xmath138@xmath72@xmath140 ( 17%)+@xmath133@xmath72@xmath143 ( 50%)+@xmath146 ( 13%)+ @xmath133@xmath72@xmath146 ( 10%)+ .. + calc .
ii & @xmath138@xmath72@xmath140 ( 16%)+@xmath133@xmath72@xmath143 ( 52%)+@xmath146 ( 13%)+ @xmath133@xmath72@xmath146 ( 12%)+ .. + @xmath148 calc .
i & @xmath142@xmath72@xmath135 ( 30%)+@xmath142@xmath72@xmath140 ( 42%)+ @xmath142@xmath72@xmath143 ( 12%)+ .. + calc .
ii & @xmath133@xmath72@xmath143 ( 23%)+@xmath145(37%)+ @xmath133@xmath72@xmath146 ( 19%)+ .. + @xmath149 calc .
i & @xmath138@xmath72@xmath143 ( 15%)+ @xmath133@xmath72@xmath146 ( 59%)+ .. + calc .
ii & @xmath138@xmath72@xmath143 ( 14%)+ @xmath133@xmath72@xmath146 ( 63%)+ .. + @xmath150 calc .
i & @xmath138@xmath72@xmath143 ( 73%)+ @xmath138@xmath72@xmath146 ( 11%)+ .. + calc .
ii & @xmath138@xmath72@xmath143 ( 79%)+ .. + @xmath151 calc .
i & @xmath142@xmath72@xmath140 ( 33%)+ @xmath142@xmath72@xmath143 ( 35%)+ .. + calc .
ii & @xmath133@xmath72@xmath152 ( 27%)+ @xmath153 ( 40%)+ .. + + -@xmath3 shell gap between shell model and experimental @xcite data . in the case of shell model calculation
the configuration is not dominant one . while experimental plausible configuration from ref .
, width=498 ] from the energy levels of @xmath10zr and @xmath15zr it is possible to estimate the energy gap between proton @xmath2-@xmath3 shell gap , once one more neutron goes into the @xmath6 orbital . in the case of @xmath15zr the @xmath96 state at 4728 kev
have @xmath95 configuration , while @xmath102 at 7063 kev have @xmath154 @xmath72 @xmath87 configuration ( although it is less dominant ) .
thus energy gap between the proton @xmath2 and @xmath3 orbitals is 2335 kev . for @xmath10zr ,
the @xmath70 at 3154 kev have @xmath155 configuration , while the @xmath156 at 5769 kev have @xmath154 @xmath72 @xmath157 configuration .
the difference in energy for the @xmath51 and the @xmath156 is 2615 kev ( i.e. energy gap between the @xmath2 and @xmath3 orbital ) .
thus , as we move from @xmath10zr to @xmath15zr , the proton @xmath2-@xmath3 shell gap is lower by 280 kev . in the fig . 6 , we have shown the comparison of @xmath2-@xmath3 shell gap between the experimental value as reported in ref . @xcite and present shell model calculation .
experiment is predicting this gap as 180 kev while corresponding shell model prediction is 280 kev .
present calculation also predicts reduction of the neutron @xmath6-@xmath5 shell gap when a proton pair is added into the @xmath4 orbital . from the experimental data @xcite of @xmath10zr and @xmath158sn , the single - particle excitation energy for @xmath49
is decreased from @xmath9 2.19 mev ( @xmath10zr ) to @xmath9 0.17 mev ( @xmath158sn ) , thus for each proton pair there is reduction of 0.4 mev in excitation energy ( in total we have five pairs in @xmath4 from @xmath10zr and @xmath158sn ) .
thus , excitation energy for @xmath159 to @xmath160 configuration is @xmath9 1.79 mev [ instead of starting value of @xmath9 2.19 mev ] . from the present calculation for @xmath10zr ,
the @xmath70 [ @xmath159 ] and @xmath74[@xmath160 ] are at 3154 , and 4333 kev , respectively .
thus the gap of 1.179 mev between these two states is lower than starting value of 1.790 mev [ if we take the difference in the single - particle energies of the present gwbxg effective interaction then the starting gap is 3.329 mev ] .
these results reflect the reduction of the neutron @xmath6-@xmath5 shell gap when a proton pair is added into the @xmath4 orbital .
the neutron @xmath6-@xmath5 shell gap from the experimental value reported in ref .
@xcite to be @xmath9 1.79 mev for @xmath159 @xmath54 @xmath160 transition . to test further the quality of our sm wave functions ,
we have calculated several @xmath163 transitions , which are listed in table 2 .
first , we have performed the calculations with the standard effective charges , i.e. , @xmath164 and @xmath165 .
since the calculation without truncation is not feasible in the present model space , thus to account for the missing mixing of configurations , we have also performed calculations with the increased effective charges @xmath166 and @xmath167 . in table 2
, we have listed both results .
these results indicating deviation with experimental data even with the enhanced values of effective charges . the calculated @xmath168 )
values with two set of effective charges are small in comparison to the experimental data . in table 3 , we have listed quadrupole and magnetic moment for the ground state and few excited states .
the results are in good agreement with the experimental data for @xmath10zr . from table 3 , it is also possible to compare recently available experimental data for the magnetic moment with the calculated values . the experimental value for @xmath169 is -0.360(20 ) , while corresponding calculated value is -0.745 . for @xmath134 state
the experimental value for @xmath170 is + 1.5(10 ) , while corresponding calculated value is + 0.962 .
similarly for @xmath90 state the experimental value for @xmath171 is -2.0(4 ) , while corresponding calculated value is -2.122 .
thus , the predicted result for the magnetic moment with shell model is close to the experimental data for @xmath90 .
.experimental @xcite and calculated @xmath104 values in w.u . with different set of effective charges .
[ cols=">,^,^,^,^,^ " , ]
structure of high spin levels of @xmath0zr have been studied with large - basis shell model calculations with neutrons excitation across @xmath7 shell . following broad conclusions
can be drawn from the present work .
* the present study reveals the importance of inclusion of @xmath6 and @xmath5 orbitals in the model space . *
the structure of @xmath55 in @xmath10zr is mainly from @xmath172 configuration .
the structure of @xmath62 states come from @xmath63 configuration , while @xmath64 have @xmath65 configuration . * for @xmath15zr , @xmath107 to @xmath108 states are fully aligned state of the @xmath106 @xmath72 @xmath101 configuration . * from the energy levels of @xmath10zr and @xmath15zr , we have also estimated reduction of energy gap between proton @xmath173 and @xmath174 orbitals once one more neutron goes into the @xmath6 orbital .
further , we have also shown that once a proton pair is added into the @xmath4 orbital there is a reduction of the neutron @xmath6-@xmath5 shell gap .
* we have calculated electromagnetic properties between different transitions for few high - spin states , wherever experimental data is available . these results indicating deviation with experimental data even with the enhanced values of effective charges .
* further theoretical work is needed by enlarging model space with protons in full @xmath175 shell and neutrons in @xmath176 space .
the author is very grateful to prof .
b a brown , michigan state university , for providing me with the computer program nushellx . thanks are due to y.h .
zhang and y. h. qiang for useful suggestions during this work .
i would like to thank mirshod ermamatov and v.k.b .
kota for a careful reading of the manuscript .
99 h. nakada , prog .
theor . phys .
( kyoto ) , suppl . * 196 * , 371 ( 2012 ) .
srivastava , j. phys .
g : nuclear and particle physics * 39 * , 015102 ( 2012 ) .
srivastava , mod .
a * 27 * , 1250061 ( 2012 ) .
srivastava , r. sahu and v.k.b .
kota , eur .
a : hadrons and nuclei , * 51 * , 3 ( 2015 ) .
et al . _ ,
a * 893 * , 1 ( 2012 ) .
c * 89 * , 034318 ( 2014 ) .
p. federman and s. pittel , phys .
b * 69 * , 385 ( 1977 ) .
n.s . pattabiraman _
et al . _ ,
c * 65 * , 044324 ( 2002 ) .
m. varga pajtler _
et al . _ ,
a * 941 * , 273 ( 2015 ) .
bingham and m.l .
halbert , phys .
c * 2 * , 2297 ( 1970 ) . m .-
porquet_et al . _ ,
c * 84 * , 054305 ( 2011 ) .
e. a. stefanova et al . , nucl .
a * 669 * , 14 ( 2000 ) .
e. a. stefanova et al . , phys .
c * 62 * , 054314 ( 2000 ) .
j. reif , g. winter , r. schwengner , h. prade , and l. kubler , nucl .
a * 587 * , 449 ( 1995 ) .
e. k. warburton , j. w. olness , c. j. lister , r. w. zurmhle , and j. a. becker , phys .
c * 31 * , 1184 ( 1985 ) . n. s. pattabiraman et al . ,
c * 65 * , 044324 ( 2002 ) .
s. s. ghugre and s. k. datta , phys .
c * 52 * , 1881 ( 1995 ) .
i. p. johnstone and l. d. skouras , phys .
c * 55 * , 1227 ( 1997 ) .
h. a. roth et al . , phys . rev .
c * 50 * , 1330 ( 1994 ) .
k. sieja , f. nowacki , k. langanke , and g. mart ' inez - pinedo , phys .
c * 79 * , 064310 ( 2009 ) .
p. federman and s. pittel , phys .
c * 20 * , 820 ( 1979 ) . r.k .
bansal and j.b .
french , phys .
lett . * 11 * , 145 ( 1964 ) .
t.otsuka , r. fujimoto , y. utsuno , b.a .
brown , m. honma , and t. mizusaki , phys .
lett . * 87 * , 082502 ( 2001 ) .
t.otsuka , t. suzuki , r. fujimoto , h. grawe , and y. akaishi , phys .
* 95 * , 232502 ( 2005 ) .
t. faestermann , m. grska , h. grawe , prog .
phys . * 69 * , 85 ( 2013 ) .
i. talmi , simple models of complex nuclei , harwood academic publishers , switzerland ( 1993 ) . v. werner et al .
b * 550 * , 140 ( 2002 ) .
y.h . zhang _
et al . _ ,
c * 79 * , 044316 ( 2009 ) .
b.a . brown , w.d.m .
rae , e. mcdonald and m. horoi , nushellx@msu .
a. hosaka , k .- i .
kubo , and h. toki , nucl .
phys . a * 444 * , 76 ( 1985 ) .
x. ji and b. h. wildenthal , phys .
c * 37 * , 1256 ( 1988 ) .
d. h. gloeckner , nucl .
a * 253 * , 301 ( 1975 ) .
f. j. d. serduke , r. d. lawson , and d. h. gloeckner , nucl .
a * 256 * , 45 ( 1976 ) .
d. bucurescu _ et al .
c * 76 * , 064301 ( 2007 ) .
g. korschinek , m. fenzl , h. hick , a. kreiner , w. kutschera , e. nolte , and h. morinaga , proceedings of the international conference on nuclear structure , tokyo , 1977 .
baglin , nucl .
data sheets * 113 * , 2187 ( 2012 ) .
d. seweryniak et al . , phys . rev .
lett . * 99 * , 022504 ( 2007 ) . | in the present work we have interpreted recently available experimental data [ z.g .
wang et al . , phys .
rev .
c , 89 ( 2014 ) 044308 ] for high - spin states of the near - spherical nuclei @xmath0zr , using the shell - model calculations within the full @xmath1 , @xmath2 , @xmath3 , @xmath4 model space for protons and valence neutrons in @xmath4 , @xmath5 , @xmath6 orbits .
we have employed a truncation for the neutrons due to huge matrix dimensions , by allowing one neutron excitation from @xmath4 orbital to @xmath6 and @xmath5 orbitals .
results are in good agreement with the available experimental data .
thus , theoretically we have identified the structure of many high - spin states , which were tentatively assigned in the recent experimental work . |
the time - dependent hartree - fock ( tdhf ) approach can be used to determine the expectation values of single - particle observables , such as fragment mass , in nuclear reactions and decays but is known to underestimate the fluctuations in these values @xcite .
this is due to the 1-body nature of tdhf and the fact that it neglects 2-body correlations .
this problem has previously been studied by balian and vnroni @xcite , who derived a general variational theory for the determination of expectation values , correlations and fluctuations .
they found that , given the state of a system described , at the time @xmath2 , by the 1-body density matrix , @xmath3 ( a slater determinant satisfying @xmath4 ) , the fluctuation , @xmath5 , in a 1-body observable , @xmath6 , at some later time @xmath7 , is given by @xmath8 , \label{eqn : bv}\ ] ] where @xmath9 is a 1-body density matrix related to @xmath10 through the boundary condition @xmath11 and where the time evolution of @xmath10 and @xmath9 is given by the usual tdhf equation .
this result is significantly different from the standard tdhf result @xmath12 , \label{eqn : tdhf}\end{aligned}\ ] ] in that it depends on the initial time , @xmath2 , with the final time , @xmath7 , entering only through the boundary condition ( [ eqn : sigma ] ) . the other key feature of this result is that it contains , through ( [ eqn : sigma ] ) , the observable @xmath6 such that this method is specifically tuned to the determination of the fluctuation of the observable of interest . a practical implementation of ( [ eqn : bv ] ) requires that a hartree - fock calculation be performed to determine the initial state , @xmath3 .
the system is then excited by a suitable external excitation , and a tdhf calculation performed from @xmath13 to determine @xmath14 .
this is used to obtain @xmath15 using ( [ eqn : sigma ] ) and a second tdhf calculation is then performed with the tdhf code run backwards , @xmath16 , to obtain @xmath17 .
the transformation ( [ eqn : sigma ] ) and the second tdhf calculation must be repeated for a range of values of @xmath18 to allow @xmath19 to be determined by extrapolation to @xmath20 .
the large number of computations required to evaluate ( [ eqn : bv ] ) and the complexity of these calculations means that only a handful of calculations have been performed using this method and those calculations which have been performed have used simplified interactions and made use of symmetries ( either spherical @xcite , or axial @xcite ) to render the problems tractable .
however , modern advances in computing power mean that this approach can now be implemented using fully 3-dimensional tdhf codes with full skyrme interactions @xcite .
we consider the mass dispersion in a bounded region of space around a giant dipole resonance ( gdr ) which decays through particle emission and calculate the mass ( number of nucleons ) in the nucleus according to @xmath21 where @xmath22 is the cutoff radius used to define the bounded region of space .
the nucleus was excited by multiplying the ground state wavefunctions from the hf calculation by a dipole boost given by @xmath23 with @xmath24 and where , for protons , @xmath25 , and for neutrons , @xmath26 , where @xmath27 is the atomic mass number of the nucleus under investigation and @xmath28 is its charge .
@xmath29 , @xmath30 and @xmath31 determine the strength of the boost applied to the nucleus .
written in terms of the single particle wavefunctions ( [ eqn : bv ] ) becomes @xcite @xmath32 @xmath33 the wavefunctions @xmath34 were obtained from the results of a static hartree - fock calculation , whilst the wavefunctions @xmath35 result from the backwards tdhf calculations and are related to the wavefunctions @xmath36 through the boundary condition @xmath37
we consider first a gdr in @xmath0s calculated using the skyrme interaction with the sly6 @xcite parametrisation .
all calculations were performed in a cubic model space of size @xmath38 fm discretised in steps of @xmath39 fm .
the initial hf calculation gave a @xmath0s ground state with a total binding energy of @xmath40 mev ( compared with the experimental value of @xmath41 mev @xcite ) and a prolate deformation ( @xmath42 ) . at the beginning of the dynamic calculation
the ground state wavefunctions were boosted in accordance with ( [ eqn : gdr - boost ] ) and with @xmath43 @xmath44 . the simulation was allowed to run from an initial time @xmath45 fm / c to @xmath46 fm / c in steps of @xmath47 fm / c .
the emitted nucleons were reflected back from the boundary of the box and would , were the simulation allowed to run long enough , re - enter the region occupied by the de - exciting nucleus causing unphysical interactions .
an analysis of the density and of @xmath48 as a function of time was used to verify that the number of nuclei in the nucleus had stabilised well in advance of the time @xmath7 and that the radiated flux had not had enough time to be reflected back and to interact with the nucleus .
the dipole moments , @xmath49 , @xmath50 and @xmath51 , were obtained as a function of time using @xcite @xmath52 where @xmath53,@xmath54,@xmath55 denotes @xmath56 , @xmath57 and @xmath58 and @xmath59 and @xmath60 are the expectation values for position calculated using the proton and neutron single particle states respectively .
this is shown in figure [ fig : s32](a ) . due to the prolate deformation of the @xmath0s nucleus ,
the @xmath50 and @xmath51 values are identical and differ from the @xmath49 values .
the periodicity of @xmath49 , @xmath50 and @xmath51 allow the excitation energies for the oscillations along each of the three primary axes to be estimated .
in this instance we obtain , for @xmath49 , a period of @xmath61 fm / c giving an excitation energy @xmath62 mev and , for @xmath50 and @xmath51 , a period of @xmath63 fm / c giving an excitation energy @xmath64 mev .
the final state gave @xmath65 with @xmath66 using @xmath67 fm which represents the emission of @xmath68 nucleons .
@xmath22 was chosen so that the bounded region fully enclosed the nucleus but omitted , as much as possible , the extended ( or dissipated ) components of the wavefunctions .
the transformation ( [ eqn : sp - psi ] ) was then applied and the tdhf code was run in reverse .
this process was repeated for @xmath18 values in the range @xmath69 in steps of @xmath70 . at the end of each time - reversed calculation the fluctuation , @xmath71 ,
was estimated using ( [ eqn : sp - bv ] ) .
these values were plotted ( see figure [ fig : s32](b ) ) and a straight line was fitted to the linear section of the graph and extended back to @xmath72 to obtain @xmath73 which represents a @xmath74% increase in @xmath75 using the bv approach compared with the standard tdhf result .
this graph is typical of those obtained using this approach and is linear for larger values of @xmath18 increasing asymptotically as @xmath20 due to the @xmath76 term in ( [ eqn : bv ] ) .
often , as in this case , the curve decreases for intermediate values of @xmath18 where the reduced value of @xmath18 means that the transformation ( [ eqn : sp - psi ] ) only has a small effect making the numerator in ( [ eqn : bv ] ) numerically approximately zero and dominant over the @xmath77 denominator .
these calculations have been repeated for the doubly magic nucleus @xmath1sn .
all the calculations were carried out using the same model space and interaction as the @xmath0s calculation .
the hf calculation produced an undeformed ground state with a binding energy of @xmath78 mev ( compared with the accepted value of @xmath79 mev @xcite ) .
the ground state single particle wavefunctions were boosted at the start of the tdhf calculation in accordance with ( [ eqn : gdr - boost ] ) and with @xmath80 @xmath44 and the calculation was run from @xmath45 fm / c to @xmath46 fm / c as in the previous calculation .
the dipole moments were plotted as a function of time and are shown in figure [ fig : sn132](a ) .
the graph shows @xmath49 , @xmath50 and @xmath51 to be identical as expected for a spherical nucleus and gives the periodicity of the dipole moments as @xmath81 fm / c which corresponds to a resonance energy of @xmath82 mev .
this is close to the experimentally measured value of @xmath83 mev @xcite .
the standard thdf calculation gave , at the time @xmath7 , @xmath84 and @xmath85 representing the emission of 11 nucleons .
a series of transformations and time - reversed tdhf calculations were carried out as previously . the resulting graph , and linear fit ,
are shown in figure [ fig : sn132](b ) which gives @xmath86 and represents a @xmath87% increase in @xmath75 compared with the standard tdhf result .
the balian - vnroni approach has been implemented for the first time using a 3-dimensional tdhf code with the full skyrme interaction .
calculations have been performed for gdrs in @xmath0s and @xmath1sn and have demonstrated that the bv approach does produce quantitatively larger results for the fluctuations of 1-body operators .
this approach is now being applied to heavy ion collisions .
the authors are pleased to acknowledge the assistance of , and discussions with , ph .
chomaz , r. balian , j. s. al - khalili and e. b. suckling .
this work was supported by epsrc grant no .
ep / p501679/1 .
11 p .- g . reinhard , r. y. cusson and k. goeke , nuc .
a*398 * , 141 - 188 ( 1983 ) r. balian and m. vnroni , annals of physics * 281 * , 65 - 142 ( 2000 ) r. balian and m. vnroni , annals of physics * 187 * , 29 - 78 ( 1988 ) r. balian and m. vnroni , annals of physics * 216 * , 351 - 430 ( 1992 ) t. troudet and d. vautherin , phys .
c * 31*(1 ) , 278 - 279 ( 1985 ) j. b. marston and s. e. koonin , phys .
54*(11 ) , 1139 - 1141 ( 1985 ) p. bonche and h. flocard , nucl . phys .
a*437 * , 189 - 207 ( 1985 ) j. a. maruhn , p .-
reinhard , p. d. stevenson and m. r. strayer , phys .
c * 74 * , 027601 ( 2006 ) a. s. umar and v. e. oberacker , phys .
c * 74 * , 024606 ( 2006 ) c. simenel , ph .
chomaz and d. de france , phys .
c * 76 * , 024609 ( 2007 ) t. nakatsukasa and k. yabana , nucl .
a*788 * , 349c-354c ( 2007 ) e. chabanat , p. bonche , p. haensel , j. meyer and r. schaeffer , nucl . phys .
a*635 * , 231 - 256 ( 1998 ) g. audi , a. wapstra and c. thibault , nucl .
a*729 * , 337 - 676 ( 2003 ) p. adrich _ et al .
_ , phys .
lett . * 95 * , 132501 ( 2005 ) | the balian - vnroni variational approach has been implemented using a 3-dimensional time - dependent hartree - fock ( tdhf ) code with realistic skyrme interactions and used to investigate the mass dispersions from giant dipole resonances in @xmath0s and @xmath1sn decaying through particle emission .
the fluctuations obtained are shown to be quantitatively larger than the standard tdhf results . |
let @xmath0 be a @xmath1-dimensional lvy process without diffusion component , that is , @xmath2.\ ] ] here @xmath3 , @xmath4 is a poisson random measure on @xmath5 with intensity @xmath6 satisfying @xmath7 and @xmath8 denotes the compensated version of @xmath4 .
we study the case when @xmath9 , that is , there is an infinite number of jumps in every interval of nonzero length a.s .
further , let @xmath10 be an @xmath11-valued adapted stochastic process , unique solution of the stochastic differential equation @xmath12 , \label{pt.sde}%\end{aligned}\ ] ] where @xmath13 is an @xmath14 matrix .
in this article we are interested in the numerical evaluation of @xmath15 $ ] for a sufficiently smooth function @xmath16 by monte carlo , via discretization and simulation of the process @xmath10 .
we propose new weak approximation algorithms for and study their rate of convergence .
the traditional method to simulate @xmath10 is to use the euler scheme with constant time step @xmath17 this method has the convergence rate @xcite @xmath18-e[f(\hat x^n_1)]|\leq \frac{c}{n}\ ] ] but suffers from two difficulties : first , for a general lvy measure @xmath6 , there is no available algorithm to simulate the increments of the driving lvy process and second , a large jump of @xmath0 occurring between two discretization points can lead to an important discretization error . a natural idea due to rubenthaler @xcite ( in the context of finite - intensity jump processes , this idea appears also in @xcite ) , is to approximate @xmath0 with a compound poisson process by replacing the small jumps with their expectation @xmath19 and then place discretization dates at all jump times of @xmath20 .
the computational complexity of simulating a single trajectory using this method becomes a random variable , but the convergence rate may be computed in terms of the _ expected _ number of discretization dates , proportional to @xmath21 . when the jumps of @xmath0 are highly concentrated around zero , however , this approximation is too rough and the convergence rates can be arbitrarily slow .
in @xcite , the authors proposed a scheme which builds on rubenthaler s idea of using the times of large jumps of @xmath0 as discretization dates but achieves better convergence rates .
their idea is , first , to approximate the small jumps of @xmath0 with a suitably chosen brownian motion , in order to match not only the first but also the second moment of @xmath0 , and second , to construct an approximation to the solution of the continuous sde between the times of large jumps .
similar ideas of gaussian correction were recently used in @xcite in the context of multilevel monte carlo methods for the problem . however , although diffusion approximation of small jumps improves the convergence rate , there are limits on how well the small jumps of a lvy process can be approximated by a brownian motion . in particular , the brownian motion is a symmetric process , while a lvy process may be asymmetric . in this paper
we develop new jump - adapted discretization schemes based on approximating the lvy process @xmath0 with a finite intensity lvy process @xmath20 without diffusion part .
contrary to previous works , instead of simply truncating jumps smaller than @xmath22 , we construct efficient finite intensity approximations which match a given number of moments of @xmath0 .
these approximations are superior to the existing approaches in two ways .
first , given that @xmath20 is a finite intensity lvy process , the solution to with @xmath0 replaced by @xmath20 is easy to compute , either explicitly or with a fast numerical method , making it straightforward to implement the scheme .
second , by choosing the parameters of @xmath20 in a suitable manner , one can , in principle , match an arbitrary number of moments of @xmath0 and obtain a discretization scheme with an arbitrarily high convergence rate .
the paper is structured as follows . in section [ pt.mmatch ] ,
we present the main idea of moment matching approximations and provide a basic error bound for such schemes . in section [ pt.3mom ] , we introduce our first scheme which is based on matching 3 moments of @xmath0 and can be used for general lvy processes . for lvy processes with stable - like behavior of small jumps near zero , the scheme is shown to be rate - optimal .
finally , section [ pt.highorder ] shows how schemes of arbitrary order can be constructed by matching additional moments , once again , in the context of lvy processes with stable - like behavior of small jumps .
let @xmath20 be a finite intensity lvy process without diffusion part approximating @xmath0 in a certain sense to be defined later : @xmath23 where @xmath24 is a poisson random measure with intensity measure @xmath25 such that @xmath26 . in this paper
we propose to approximate the process by the solution to @xmath27 which can be computed by applying the euler scheme at the jump times of @xmath20 and solving the deterministic ode @xmath28 explicitly ( or by a runge - kutta method ) between these jump times .
the following proposition provides a basic estimate for the weak error of such an approximation scheme .
we impose the following alternative regularity assumptions on the functions @xmath16 and @xmath13 : * @xmath29 * : : @xmath30 , @xmath31 @xmath32 and @xmath33 are bounded for @xmath34 and @xmath35 . * @xmath36 * : : @xmath30 , @xmath31 , @xmath33 are bounded for @xmath34 , @xmath32 have at most polynomial growth for @xmath34 and @xmath37 for all @xmath38 . [ pt.basicbound ]
let @xmath0 and @xmath39 be lvy processes with characteristic triplets @xmath40 and @xmath41 respectively , and let @xmath10 and @xmath42 be the corresponding solutions of sde .
assume @xmath43 , @xmath44 on @xmath45 , either @xmath46 or @xmath47 for @xmath48 and @xmath49 then @xmath50| \leq c \int_{\mathbb r^d } \|x\|^n|d\nu -d \hat \nu|,\ ] ] where the constant @xmath51 may depend on @xmath16 , @xmath52 , @xmath53 and @xmath6 but not on @xmath54 . to simplify notation
, we give the proof in the case @xmath55 .
let @xmath56 $ ] .
by lemma 13 in @xcite , @xmath57\times \mathbb r)$ ] and satisfies @xmath58 applying it formula under the integral sign and using and lemma 11 in @xcite ( bounds on moments of @xmath59 ) yields @xmath60 = e[u(1,\hat x_1 ) - u(0,x_0 ) ] \\ & = e\left[\int_0 ^ 1 \int_{\mathbb r } \left\{u(t,\hat x_t + h(\hat x_t ) z )
- u(t,\hat x_t ) - h(\hat x_t ) z \frac{\partial u}{\partial x}\right\}(d\nu_\varepsilon - d\nu ) dt \right ] \\ & + e\left[\int_0 ^ 1 \int_{\mathbb r } \left\{u(t,\hat x_{t- } + h(\hat x_{t- } ) z ) - u(t,\hat x_{t-})\right\}\hat n(dz , dt ) \right]\\ & = e\left[\int_0 ^ 1 \int_{\mathbb r } \sum_{k=2}^{n-1 } \frac{\partial^k u ( t,\hat x_t)}{\partial x^k } h^k(\hat x_t ) z^k ( d\nu_\varepsilon - d\nu ) dt + \text{remainder}\right],\\ & = e[\text{remainder}],\end{aligned}\ ] ] where in the last line we used the moment matching condition and the remainder coming from the taylor formula can be estimated as @xmath61 from the lipschitz property of @xmath13 and lemma 13 in @xcite , @xmath62 for some @xmath63 , where @xmath64 under @xmath46 and @xmath65 under @xmath66 . following the arguments in the proof of lemma 11 in @xcite
, we get @xmath67 \leq c(1+|x|^p)\exp\left[c\left(|\bar \gamma|^p + \int_{\mathbb r } |z|^p \hat\nu(dz ) + \left(\int_{\mathbb r}z^2 \hat\nu(dz)\right)^{p/2}\right)\right]\ ] ] for different constants @xmath51 and @xmath68 , where @xmath69 by our assumptions .
since @xmath70 by assumption , and @xmath71 it is clear that @xmath72\leq c$ ] for some constant @xmath51 which does not depend on @xmath54 .
our first scheme is based on matching the first 3 moments of the process @xmath0 .
let @xmath73 be the unit sphere in the @xmath1-dimensional space , and @xmath74 be a lvy measure on @xmath75 written in spherical coordinates @xmath76 and @xmath77 and satisfying @xmath78 .
denote by @xmath79 the reflection of @xmath6 with respect to the origin defined by @xmath80 .
we introduce two measures on @xmath73 : @xmath81 the _
3-moment scheme _ is defined by @xmath82\times s^{d-1 } } r\theta\ , \nu_\varepsilon(dr , d\theta ) , \label{pt.3mom2}\end{aligned}\ ] ] where @xmath83 denotes a point mass at @xmath22 .
[ pt.3mom.prop ] for every @xmath84 , @xmath85 is a finite positive measure satisfying @xmath86 where the last inequality is an equality if @xmath87 .
the positivity of @xmath85 being straightforward , let us check .
let @xmath88 be the coordinate vectors .
then , @xmath89 the other equations can be checked in a similar manner .
let @xmath90 .
then the 3-moment scheme can be written as @xmath91 assume @xmath92 or @xmath93
. then the solution @xmath42 of with the characteristics of @xmath20 given by satisfies @xmath94| = o(\lambda_\varepsilon^{-1}).\ ] ] as @xmath95 . by proposition [ pt.basicbound ]
we need to show that @xmath96 by proposition [ pt.3mom.prop ] , @xmath97 where in the last line the dominated convergence theorem was used . in many parametric or semiparametric models ,
the lvy measure has a singularity of type @xmath98 near zero .
this is the case for stable processes , tempered stable processes @xcite , normal inverse gaussian process @xcite , cgmy @xcite and other models .
stable - like behavior of small jumps is a standard assumption for the analysis of asymptotic behavior of lvy processes in many contexts , and in our problem as well , this property allows to obtain a more precise estimate of the convergence rate .
we shall impose the following assumption , which does not require the lvy measure to have a density : * @xmath99 * : : there exist @xmath100 and @xmath101 such that if @xmath102 and @xmath103 if @xmath104 . ]
@xmath105 where @xmath106 .
assume @xmath99 and @xmath92 or @xmath93 .
then the solution @xmath42 of with the characteristics of @xmath20 given by satisfies @xmath94| = o\left(\lambda_\varepsilon^{1-\frac{4}{\alpha}}\right).\ ] ] under @xmath99 , by integration parts we get that for all @xmath107 , @xmath108 therefore , under this assumption , @xmath109 from which the result follows directly .
[ [ rate - optimality - of - the-3-moment - scheme ] ] rate - optimality of the 3-moment scheme + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + from proposition [ pt.basicbound ] we know that under the assumption @xmath110 or @xmath111 , the approximation error of a scheme of the form can be measured in terms of the @xmath112-th absolute moment of the difference of lvy measures .
we introduce the class of lvy measures on @xmath75 with intensity bounded by @xmath4 : @xmath113 the class of lvy measures with intensity bounded by @xmath114 is then denoted by @xmath115 , and the smallest possible error achieved by any measure within this class is bounded from below by a constant times @xmath116 .
the next result shows that as @xmath117 , the error achieved by the 3-moment scheme @xmath85 differs from this lower bound by at most a constant multiplicative factor .
assume @xmath99 and let @xmath85 be given by .
then , @xmath118 _ step 1 .
_ let us first compute @xmath119 for @xmath120 , let @xmath121 where @xmath122 is absolutely continuous with respect to @xmath6 and @xmath123 is singular .
then @xmath124 and @xmath125 which are absolutely continuous with respect to @xmath6 , or , in other words , @xmath126 where the @xmath127 is taken over all measurable functions @xmath128 such that @xmath129 . by a similar argument
, one can show that it is sufficient to consider only functions @xmath130 $ ] .
given such a function @xmath131 , the spherically symmetric function @xmath132 leads to the same values of the intensity and the minimization functional .
therefore , letting @xmath133 , @xmath134 for every @xmath135 , @xmath136 the @xmath127 in the right - hand side can be computed pointwise and is attained by @xmath137 for any @xmath138 $ ] .
let @xmath139 and @xmath140 be such that @xmath141 such a @xmath139 can always be determined uniquely and @xmath140 is determined uniquely if @xmath142 .
it follows that @xmath143 is a minimizer for and therefore @xmath144 where @xmath145 and @xmath146 are solutions of @xmath147 _ step 2 .
_ for every @xmath84 , let @xmath148 and @xmath149 be solutions of @xmath150 it is clear that @xmath151 as @xmath95 and after some straightforward computations using the assumption @xmath99 we get that @xmath152 then , @xmath153 under @xmath99 the three limits are easily computed and we finally get @xmath154 the constant @xmath155 appearing in the right - hand side of can not be interpreted as a `` measure of suboptimality '' of the 3-moment scheme , but only as a rough upper bound , because in the optimization problem the moment - matching constraints were not imposed ( if they were , it would not be possible to solve the problem explicitly ) .
on the other hand , the fact that this constant is unbounded as @xmath156 suggests that such a rate - optimality result can not be shown for general lvy measures without imposing the assumption @xmath99 . [ [ numerical - illustration ] ] numerical illustration + + + + + + + + + + + + + + + + + + + + + + we shall now illustrate the theoretical results on a concrete example of a sde driven by a normal inverse gaussian ( nig ) process @xcite , whose characteristic function is @xmath157 = \exp\left\{-\delta t \left(\sqrt{\alpha^2 - ( \beta - iu)^2 } - \sqrt{\alpha^2 - \beta^2}\right)\right\},\ ] ] where @xmath158 , @xmath159 and @xmath160 are parameters .
the lvy density is given by @xmath161 where @xmath162 is the modified bessel function of the second kind .
the nig process has stable - like behavior of small jumps with @xmath163 , @xmath164 ( which means that @xmath99 is satisfied with @xmath165 ) , and exponential tail decay .
the increments of the nig process can be simulated explicitly ( see ( * ? ? ?
* algorithms 6.9 and 6.10 ) ) , which enables us to compare our jump - adapted algorithm with the classical euler scheme . for the numerical example
we solve the one - dimensional sde @xmath166 where @xmath0 is the nig lvy process ( with drift adjusted to have @xmath167=0 $ ] ) .
the solution of the corresponding deterministic ode @xmath168 is given explicitly by @xmath169 figure [ pt.nig.fig ] presents the approximation errors for evaluating the functional @xmath170 $ ] by monte - carlo using the 3-moment scheme described in this section ( marked with crosses ) , the diffusion approximation of @xcite ( circles ) and the classical euler scheme ( diamonds ) .
the parameter values are @xmath171 , @xmath172 , @xmath173 , @xmath174 and @xmath175 . for each scheme
we plot the logarithm of the approximation error as function of the logarithm of the computational cost ( time needed to simulate @xmath176 trajectories ) .
the curves are obtained by varying the truncation parameter @xmath22 for the two jump - adapted schemes and by varying the discretization time step for the euler scheme .
the approximation error for the euler scheme is a straight line with slope corresponding to the theoretical convergence rate of @xmath177 .
the graph for the 3-moment scheme seems to confirm the theoretical convergence rate of @xmath178 ; the scheme is much faster than the other two and the corresponding curve quickly drops below the dotted line which symbolizes the level of the statistical error .
in this section , we develop schemes of arbitrary order for lvy processes with stable - like behavior of small jumps . throughout this section , we take @xmath90 and let @xmath0 be a lvy process with characteristic triplet @xmath40 satisfying the following refined version of @xmath99 : @xmath179 : : there exist , @xmath180 , @xmath181 with @xmath182 and @xmath183 such that @xmath184 introduce the probability measure @xmath185 let @xmath186 and @xmath84 . the high - order scheme for the stochastic differential equation based on @xmath187 moments and truncation level @xmath22 is constructed as follows : 1 . [ pt.step1 ] find a discrete probability measure @xmath188 with @xmath189 such that @xmath190 , @xmath191 for all @xmath192 and @xmath193 for all @xmath192 .
2 . compute the coefficients @xmath194 by solving the linear system @xmath195 3 .
the high - order scheme is defined by @xmath196 the first step in implementing the scheme is to solve the moment - matching problem for measure @xmath197 .
the existence of at least one solution to this problem with @xmath198 and @xmath199 for all @xmath192 is guaranteed by the classical caratheodory s theorem , but this problem admits , in general , an infinite number of solutions .
here we impose the additional condition @xmath200 and @xmath201 for all @xmath192 , which should be checked on a case by case basis in concrete realizations of the scheme ( see example [ pt.momentex ] ) .
it is easy to see that the measure @xmath202 matches the moments of orders @xmath203 of @xmath204 , where @xmath205 is the measure given by @xmath206 that is , @xmath205 satisfies the assumption @xmath179 with equalities instead of equivalences .
the idea of the method is to replace the coefficients @xmath207 with a different set of coefficients while keeping the same points @xmath208 to obtain a measure which matches the moments of @xmath209 .
therefore , the points @xmath208 do not depend on the truncation parameter @xmath22 while the coefficients @xmath194 depend on it .
[ pt.momentex ] as an example we provide a possible solution of the moment matching problem for @xmath210 , which leads to a 5-moment scheme ( matching 3 moments of @xmath197 or @xmath211 moments of the lvy process ) .
we assume that @xmath197 has mass both on the positive and the negative half - line : @xmath212 .
the moments of @xmath197 are given by @xmath213 it is then convenient to look for the discrete measure matching the first 3 moments of @xmath197 in the form @xmath214 where @xmath215 , @xmath216 are parameters to be identified from the moment conditions @xmath217 for the purpose of solving this system of equations , let @xmath218 be a random variable such that @xmath219 = p = 1- p[\mathcal e = \varepsilon_1]$ ] . from the moment conditions , we get : @xmath220 = \frac{2-\alpha}{3-\alpha}\label{pt.epsbar3},\qquad \sigma^2 : = \text{var}\ , \mathcal e = \frac{(2-\alpha)}{(4-\alpha)(3-\alpha)^2},\\ s & : = \frac{e[(\mathcal e - e[\mathcal e])^3]}{\sigma^3 } = 2\frac{\alpha-1}{5-\alpha}\sqrt{\frac{4-\alpha}{2-\alpha}}. \label{pt.skew3}\end{aligned}\ ] ] on the other hand , the skewness @xmath221 can be directly linked to the weight @xmath222 : @xmath223 and the parameters @xmath224 and @xmath225 can be linked to @xmath226 and @xmath222 : @xmath227 the dependence of @xmath224 , @xmath225 and @xmath222 on @xmath228 is shown in figure [ pt.moment.fig ] : it is clear from the graph that the constraints @xmath215 and @xmath229 are satisfied for all @xmath183 : therefore , equations ( [ pt.mubar3][pt.eps123 ] ) define a @xmath112-atom probability measure which matches the first 3 moments of @xmath197 . )
.,scaledwidth=70.0% ] let @xmath208 be fixed according to .
there exists @xmath230 such that for all @xmath231 , @xmath85 is a positive measure satisfying @xmath232 there exist positive constants @xmath233 and @xmath234 such that @xmath235 [ pt.ho.prop ] assume @xmath236 or @xmath237 . then the solution @xmath42 of with characteristics of @xmath20 given by satisfies @xmath94| = o\left(\lambda_\varepsilon^{1-\frac{n+3}{\alpha}}\right).\ ] ] the moment conditions hold by construction . using integration by parts ,
we compute @xmath238 therefore , @xmath239 since the matrix @xmath240 , @xmath241 , @xmath242 is invertible ( vandermonde matrix ) , this implies that @xmath243
. therefore , there exists @xmath244 such that for all @xmath245 , @xmath246 for all @xmath192 and @xmath85 is a positive measure .
we next compute : @xmath247 @xmath248
this research is supported by the chair financial risks of the risk foundation sponsored by socit gnrale , the chair derivatives of the future sponsored by the fdration bancaire franaise , and the chair finance and sustainable development sponsored by edf and calyon . | we propose new jump - adapted weak approximation schemes for stochastic differential equations driven by pure - jump lvy processes .
the idea is to replace the driving lvy process @xmath0 with a finite intensity process which has the same lvy measure outside a neighborhood of zero and matches a given number of moments of @xmath0 . by matching 3 moments
we construct a scheme which works for all lvy measures and is superior to the existing approaches both in terms of convergence rates and easiness of implementation . in the case of lvy processes with stable - like behavior of small jumps ,
we construct schemes with arbitrarily high rates of convergence by matching a sufficiently large number of moments .
= 1 key words : lvy - driven stochastic differential equation , euler scheme , high order discretization schemes , jump - adapted discretization , weak approximation .
2010 mathematics subject classification : primary 60h35 , secondary 65c05 , 60g51 . |
in this supplementary notes , we want to explicitly outline how , starting from the master equation associated with the stochastic may - leonard model [ defined by the reactions ( 1 ) ] , the set of stochastic partial differential equations ( 3 ) can be obtained via system size expansion . for the sake of illustration , here we focus on the role of internal noise stemming from the reactions ( 1 ) , and ignore spatial degrees of freedom . as detailed in a forthcoming publication
[ 18 ] , and following the reasoning presented in [ 16 ] , in a proper continuum limit , the spatial dispersal of individuals is accounted by diffusive terms in the spde ( 3 ) . + as in the main text , the overall number of individuals is denoted @xmath29 and @xmath101 stands for the frequencies ( or densities ) of the species @xmath6 , @xmath7 , and @xmath2 in the population ( i.e. @xmath102 and @xmath103 ) .
the master equation giving the time - evolution of the probability @xmath104 of finding the system in the state @xmath105 at time @xmath106 then reads @xmath107 where @xmath108 denotes the transition probability from state @xmath105 to the state @xmath109 within one time step ( loss term ) , @xmath110 is the analogous gain term , and the summation extends over all possible changes @xmath111 . as an example , the relevant changes @xmath112 in the density @xmath113 resulting from the basic reactions ( 1 ) are @xmath114 in the third reaction , @xmath115 in the fourth , and zero in all others .
we also choose the unit of time such that , on average , every individual reacts once per time step .
the transition rates resulting from the reactions ( 1 ) then read @xmath116 for the reaction @xmath117 ( the prefactor of @xmath29 enters due to our choice of time scale , where @xmath29 reactions occur in one unit of time ) and @xmath118 for @xmath119 .
transition probabilities associated with all other reactions ( 1 ) follow similarly .
the system size , or kramers - moyal , expansion ( sze ) [ 16 ] of the master equation is an expansion in the increment @xmath111 , which is proportional to @xmath47 .
therefore , the sze may be understood as an expansion in the inverse system size @xmath47 . to second order in @xmath111
, it yields the ( generic ) fokker - planck equation [ 16 ] : @xmath120+\frac{1}{2}\partial_i\partial_j[\mathcal{b}_{ij}({\bm s})p({\bm s},t ) ] ~. \label{fokker_planck}\ ] ] for the system under consideration , in the above the indices @xmath121 and the summation convention in ( [ fokker_planck ] ) implies sums carried over them . according to the kramers - moyal expansion ( or sze ) , the quantities @xmath122 and @xmath123 are given by [ 16 ] @xmath124 note that @xmath125 is symmetric . for the sake of clarity
, we outline the calculation of @xmath126 : the relevant changes @xmath127 result from the third and fourth reactions in ( 1 ) , as described above .
the corresponding rates respectively read @xmath116 and @xmath118 , resulting in @xmath128 .
the other quantities are computed analogously .
all explicit expressions for @xmath129 and @xmath130 will be derived and given in detail in [ 18 ] .
the well - known correspondence between fokker - planck equations and ito calculus [ 16 ] implies that ( [ fokker_planck ] ) is equivalent to the following set of ito stochastic differential equations ( with the above summation convention ) : @xmath131 where the matrix @xmath132 is defined from @xmath125 via the relation @xmath133 [ 16 ] , and the @xmath134 s denote ( uncorrelated ) gaussian white noise terms .
note that for the model under consideration @xmath125 is diagonal and one can therefore always choose a diagonal matrix for @xmath132 [ see eq .
( 5 ) ] , with only @xmath135 s contributing to the right - hand side of eqs . ( 11 ) . in ref.[18 ] , we demonstrate that ( in a proper continuum limit ) spatial degrees of freedom and exchange processes simply yield additional diffusive terms @xmath136 in ( [ stoch_part_eq ] ) .
this leads to the spde ( 3 ) , given and discussed in the main text , in which the @xmath137s still denote gaussian white noise contributions .
the spde ( 3 ) and the cgle ( 6 ) , as well as the analytical predictions ( 7 ) , are valid in all spatial dimensions . while in the text , for the sake of specificity
, we mainly focus on the two - dimensional situation , here we comment on some features of the one - dimensional version of the system , as well as on some properties of the general situation in higher dimensions . in one dimension and in the absence of exchange processes
( mixing ) , our model is expected to exhibit coarsening like the cyclic lotka - volterra model ( see e.g. [ 12 ] ) .
the same phenomenon still occurs in the presence of very slow mixing rate ( for the model under consideration , this means very low values of the exchange rate @xmath11 ) . on the other hand , and as shown in the main text , in the presence of ( finite ) mixing the systems behavior is aptly described by the spde ( 3 ) .
the underlying cgle ( 6 ) predicts the propagation of traveling waves , with velocity and wavelength still given by the analytical expression ( 7 ) .
similarly to what has been found in the two - dimensional system [ 6 ] , if the exchange rate is `` moderate '' , i.e. below a certain mobility threshold but still finite [ 6,15 ] , the system confirms these predictions and the species coexist .
however , due to stochastic events and the presence of absorbing boundaries , some domains will occasionally merge , resulting in growing domains .
this coarsening phenomenon will happen on a much longer time - scale than in the absence of the mixing processes . above the threshold value for the diffusivity ,
the system is well - mixed , the underlying spatial structure plays no role , and the description in terms of the rate equations ( 2 ) is valid . in this mean - field scenario ,
the system approaches an absorbing steady state and no patterns emerge . in higher dimensions ( i.e. in dimensions @xmath138 ) ,
the description of the stochastic spatial system in terms of the spde ( 3 ) and of the cgle ( 7 ) is again valid for `` moderate '' ( or intermediate ) mixing ( i.e. for a finite mobility rate which is below a certain critical threshold , see [ 6 ] ) . in this situation , and as discussed in the main text , the description in terms of the cgle ( 6 ) is qualitatively valid and predicts the emergence of moving spiral waves in two dimensions ( which is the case discussed in detail in the main text , see also [ 6,15 ] ) , and to `` scroll waves '' , i.e. vortex filaments , in three dimensions [ 24 ] . | noise and spatial degrees of freedom characterize most ecosystems .
some aspects of their influence on the coevolution of populations with cyclic interspecies competition have been demonstrated in recent experiments [ e.g. b. kerr et al . ,
nature * 418 * , 171 ( 2002 ) ] . to reach a better theoretical understanding of these phenomena , we consider a paradigmatic spatial model where three species exhibit cyclic dominance .
using an individual - based description , as well as stochastic partial differential and deterministic reaction - diffusion equations , we account for stochastic fluctuations and spatial diffusion at different levels , and show how fascinating patterns of entangled spirals emerge .
we rationalize our analysis by computing the spatio - temporal correlation functions and provide analytical expressions for the front velocity and the wavelength of the propagating spiral waves .
understanding the combined influence of spatial degrees of freedom and noise on biodiversity is an important issue in theoretical biology and ecology .
this implies to face the challenging problem of studying complex nonequilibrium structures , which form in the course of nonlinear evolution @xcite .
more generally , self - organized nonequilibrium patterns and traveling waves are ubiquitous in nature and appear , for instance , in chemical reactions , biological systems , as well as in epidemic outbreaks @xcite . among the most studied types of patterns
are spiral waves , which are relevant to autocatalytic chemical reactions , aggregating slime - mold cells and cardiac muscle tissue @xcite . in all these _ nonequilibrium
_ and _ nonlinear _ processes , as well as in population dynamics models @xcite , pattern formation is driven by diffusion which , together with internal noise , act as mechanisms allowing for stabilization and coevolution of the reactants . in this work , we consider a paradigmatic spatially - extended @xmath0 species population system with cyclic competition , which can be regarded as a simple food - chain model @xcite .
in fact , such a system is inspired by recent experiments on the coevolution of @xmath0 species of bacteria in cyclic competition @xcite .
using methods of statistical physics , we study the influence of spatial degrees of freedom and internal noise on the coevolution of the species and on the emerging spiral patterns . in particular , we compute the correlation functions and provide analytical expressions for the spreading speed and wavelength of the propagating fronts . to underpin the role of internal noise , the results of the stochastic description are compared with those of the deterministic equations . in this letter
, we investigate a stochastic spatial variant of the _ rock - paper - scissors game _ @xcite ( also referred to as cyclic lotka - volterra model ) .
these kinds of systems have been studied both from a game - theoretic perspective , see e.g. @xcite and references therein , and within the framework of chemical reactions @xcite , revealing rich spatio - temporal behaviors ( e.g. emergence of rotating spirals ) . while our methods have a broad range of applicability , they are illustrated for a prototypical model introduced by may and leonard @xcite where @xmath0 species , @xmath1 and @xmath2 undergo a cyclic competition ( codominance with rate @xmath3 ) and reproduction ( with rate @xmath4 ) , according to the reactions @xmath5 hence , an individual of species @xmath6 will consume one of species @xmath7 ( @xmath8 ) with rate @xmath3 and will reproduce with rate @xmath4 if an empty spot , denoted @xmath9 , is available ( @xmath10 , i.e. there is a _
finite _ carrying capacity ) .
in addition , to mimic the possibility of migration , it is realistic to endow the individuals with a form of mobility . for the sake of simplicity , we consider a simple exchange process , with rate @xmath11 , among any _ nearest - neighbor _ pairs of agents : @xmath12 .
if one ignores the spatial structure and assumes the system to be well - mixed ( with an infinite number of individuals ) , the population s mobility plays no role and the dynamics is aptly described by the deterministic rate equations ( re ) for the densities @xmath13 of species @xmath14 and @xmath2 , respectively . introducing @xmath15 ,
the re read : @xmath16 , \quad i\in \{1,2,3\}\end{aligned}\ ] ] where the index @xmath17 is taken modulo 3 and @xmath18 is the total density . as shown by may and leonard @xcite
( see also @xcite ) , these equations possess 4 absorbing fixed points , corresponding to a system filled with only one species and to an empty system .
in addition , there is a reactive fixed point @xmath19 , corresponding to a total density @xmath20 .
a linear stability analysis shows that @xmath21 is _ unstable_. the absorbing steady states @xmath22 , @xmath23 and @xmath24 are heteroclinic points .
the existence of a lyapunov function @xmath25 allows to prove that , within the realm of the above re , the phase portrait is characterized by flows spiraling outward from @xmath26 , with frequency @xmath27 $ ] in its vicinity . approaching the boundaries of the phase portrait , the trajectories form ( heteroclinic ) cycles indefinitely close to the edges ( without ever reaching them ) , with densities approaching in turn the value one . despite its mathematical elegance
, this behavior has been recognized to be unrealistic @xcite .
in fact , for finite populations , fluctuations arise and _ always _ cause the extinction of two species in finite time ( see e.g. ref @xcite ) . in this work ,
considering the spatial version of the above model in the presence of internal noise , we show that a robust ( and , arguably , more realistic ) scenario for the evolution arises .
the reaction schemes ( [ ml_react ] ) and the exchange events are considered to occur on a @xmath28dimensional regular lattice of @xmath29 sites , labeled @xmath30 . each lattice site has @xmath31 neighbors at a distance @xmath32 ( e.g. @xmath33 and @xmath34 for hypercubic lattices of linear size @xmath35 ) and is either empty or occupied by at most one individual . on the lattice , the binary reactions ( [ ml_react ] ) and exchanges only occur among pairs of _ nearest - neighbors_. in the situation of large system sizes , the continuum limit reveals that for the exchange process to be an efficient driving mechanism , the rate @xmath11 has to scale as @xmath36 , with @xmath37 and @xmath38 .
in fact , if @xmath39 the system is dominated by the local reactions ( [ ml_react ] ) among neighboring individuals ; while effective diffusion renders locality irrelevant when @xmath40 . only when @xmath37 , there is an effective competition between the stirring process and the local reactions ( [ ml_react ] )
it is therefore useful to introduce the _ effective diffusion constant _ @xmath41 . because of the discreteness of the number of individuals involved in the reactions , internal fluctuations arise in the system .
the latter originate from ( i ) the interspecies reactions ( [ ml_react ] ) and ( ii ) the exchange processes . in the continuum limit , where @xmath42 with @xmath43 ( and finite @xmath44 ) , there is a separation of time scales and the pair exchanges occur much faster than the reactions ( @xmath45 ) . actually , the fluctuations associated with ( [ ml_react ] ) and the agents mobility scale respectively as @xmath46 and @xmath47 , with the former dominating over the latter and being the only relevant contribution .
this result is revealed by a system size , also called kramers - moyal ( see e.g. ref .
@xcite , chap . 8) , expansion ( sze ) of the master equation underlying the exchange processes and the reactions ( [ ml_react ] ) @xcite .
furthermore , the sze yields a proper fokker - planck equation , which is equivalent to a set of ( ito ) stochastic partial differential equations ( spde ) with white noise .
the derivation , obtained in the continuum limit from the master equation , is outlined in the supplementary epaps document @xcite and will be detailed elsewhere @xcite . here
, we quote the expression of the spde : @xmath48 where @xmath49 is the laplacian operator ; @xmath50 , @xmath51 and @xmath52,\\ \mathcal{c}_{ij}\big({\bm s}\big)&=&\delta_{ij}\sqrt{n^{-1}s_i\big[\mu(1- \rho)+\sigma s_{i+2}\big]}.\end{aligned}\ ] ] again , the indices are taken modulo @xmath0 and now @xmath53 .
as explained in @xcite , these spde have to be interpreted in the sense of ito calculus . while eqs .
( [ spde ] ) and our approach are valid in any dimension @xcite , for specificity , we now analyze the spatio - temporal properties of the system in two dimensions with periodic boundary conditions . on the one hand we have solved
numerically the spde ( [ spde ] ) using the open software from the xmds project @xcite .
on the other hand , we have carried out individual - based simulations of the reactions ( [ ml_react ] ) for mobile ( exchange process ) particles on lattices of size @xmath54 , with @xmath55 .
this allows to carefully study the convergence towards the continuum limit , where the description in terms of ( [ spde ] ) is expected to be accurate . .
each color ( level of grey ) indicates one species ( black dots correspond to vacancies ) . in ( a ) and
( b ) results are from lattice simulations for @xmath56 ( a ) and @xmath57 ( b ) , i.e. different @xmath11 .
spiral structures emerge for sufficiently large exchange rate ( b ) . numerical solution of the spde ( [ spde ] ) and drde are shown in ( c ) , resp .
( d ) ; see text . in
( a ) , ( b ) and ( c ) , initially @xmath58 . [ snap ] ] as other spatially - extended dynamical systems @xcite , the model under consideration displays fascinating nonequilibrium patterns emerging in the course of the evolution . in fig .
[ snap ] ( a ) and ( b ) , we report typical long - time snapshots of the system for low ( a ) and high ( b ) exchange rates ( but keeping @xmath44 fixed ) , as obtained from lattice simulations . in both cases
we notice that intriguing patterns form .
for slow exchange rate , the system displays non - geometrical patches , similarly to what happens in systems with self - organized criticality @xcite . when the exchange rate is raised , the patterns display spiral structures .
in fact , starting from a spatially homogeneous initial condition , @xmath59 , the system is randomly perturbed by the internal noise and the resulting spatial inhomogeneities grow and form wavefronts moving through the system .
the emergence of spiral patterns is a feature shared by other _ excitable systems _ ( see e.g. @xcite ) and corresponds to the ability of the system to sustain the propagation of oscillating waves . for sufficiently large @xmath11
, one observes a striking resemblance between the size and structure of the patterns obtained from the lattice simulations [ fig .
[ snap ] ( b ) ] and those from the spde ( [ spde ] ) [ fig . [ snap ] ( c ) ] . to further compare the predictions of the spde ( [ spde ] ) with the lattice simulations , and to gain additional information on the structure of the the emerging patterns , we have computed the correlation functions , @xmath60 in two dimensions . in fig .
[ corr_comp ] ( red and blue curves ) , we report the results for @xmath61 in the steady state and notice an excellent agreement between the results of the lattice simulations and the predictions of the spde ( [ spde ] ) . the inset of fig
. [ corr_comp ] , displays the correlation length @xmath62 from their maximal value . ] as function of @xmath11 ( @xmath44 is kept fixed , @xmath35 varies ) obtained in the lattice simulations , which is found to coincide with the prediction of the spde already for @xmath63 .
we have also computed the autocorrelation function @xmath64 and found , both in the lattice simulations and from the solutions of the spde , an oscillating behavior with a similar characteristic frequency , markedly different from @xmath65 @xcite .
this confirms that , even for _ finite exchange rates _ , the solution of the spde ( [ spde ] ) provides an excellent approximation of the lattice simulations of the system .
this is rather surprising since eqs .
( [ spde ] ) have been derived in the continuum limit , where @xmath29 and @xmath66 .
a comparable influence of finite exchange rate in a predator - prey system has been reported recently @xcite .
, @xmath57 ) , from the solution of the spde ( [ spde ] ) ( dark blue , squares ) and of the drde ( green , triangles ) , see text .
the reaction rates are @xmath67 and @xmath68 .
inset : the correlation length @xmath69 , for @xmath70 , as function of @xmath11 ( i.e. for different lattice sizes ) compared to the prediction of the spde ( black line ) .
the latter is in excellent agreement with lattice simulations already for @xmath63 ( i.e. @xmath71 ) .
[ corr_comp ] ] according to the spde ( [ spde ] ) , @xmath62 scales as @xmath72 , so that by raising the diffusion one increases the size of the spirals . as we have shown in ref .
@xcite , this happens up to a critical value @xmath73 ( e.g. @xmath74 for @xmath67 ) : above that threshold , the spiral structures outgrow the system size and only one species survives , corresponding to an absorbing steady state predicted by eqs .
( [ re ] ) . as the properties of the lattice simulations are well captured by the spde ( [ spde ] ) , where the strength of the noise scales as @xmath46 , with @xmath75 , it is natural to investigate the actual influence of this internal noise on the steady state of the system . to address this issue ,
we have solved numerically ( in 2d , with periodic boundary conditions ) the deterministic reaction - diffusion equation ( drde ) obtained from ( [ spde ] ) by dropping the noise terms , i.e. @xmath76 of course , to obtain a nontrivial steady state for the drde one has to assume spatially inhomogeneous initial conditions . in fig .
[ snap ] ( d ) , we have reported a snapshot of the long - time behavior predicted by the drde starting from @xmath77 . in this case
, the dynamics evolves towards a reactive steady state which also exhibits spiral waves .
however , the latter do not form entangled structures , but ordered geometrical patterns . as an example , only four spirals cover the system in fig .
[ snap ] ( d ) [ while noise leads to 106 entangled spirals in fig .
[ snap ] ( c ) ] .
the correlation functions associated with the drde therefore exhibit only weakly damped spatial oscillations ( see fig
. [ corr_comp ] , green triangles ) . by analyzing typical snapshots like those of fig . [ snap ] ( d )
, we have noted that in the deterministic and stochastic [ i.e. lattice simulations with `` large '' @xmath11 and solutions of eqs .
( [ spde ] ) ] descriptions , the spiral waves share the same propagation velocity , frequency and wavelength .
however , a major difference between these descriptions lies on the crucial dependence of the drde on initial conditions , which determine the overall number of spirals and their size . on the contrary , because the internal noise acts a random source of spatial inhomogeneities , the lattice stochastic system and the spde display _
robust features_. in particular , we have found noise to induce a universal spiral density of about @xmath78 per square wavelength .
analytical expressions for the spreading velocity and the wavelength of the propagating fronts of the drde can be obtained by considering the dynamics on the invariant manifold of the re @xcite , given by @xmath79 , with @xmath80 .
on @xmath81 , up to @xmath82 order , the drde can be recast in the form of a forced complex ginzburg - landau equation ( cgle ) @xcite . by performing the nonlinear transformation
@xmath83 $ ] and @xmath84 $ ] , upon ignoring nonlinearities like @xmath85 , one is left with the following cgle in the variable @xmath86 @xcite : @xmath87 with @xmath88 , @xmath89 and @xmath90 .
the general theory of front propagation @xcite predicts that eq .
( [ cgle ] ) always admits traveling waves as stable solutions ( i.e. no benjamin - feir or eckhaus instabilities occur ) .
we have determined such periodic solutions by computing , from the dispersion relation of ( [ cgle ] ) , the spreading velocity @xmath91 and the spirals wavelength @xmath92 ( details will be given in @xcite ) : @xmath93 in the stochastic version of the model , the wavelength and velocity of the wavefronts have been found to agree with those of the deterministic treatment .
hence , the expressions ( [ res ] ) also apply ( for large @xmath11 , with @xmath94 ) to the results of lattice simulations ( rescaled by a factor @xmath35 ) and to the solution of the spde ( [ spde ] ) . for instance , on a square grid with @xmath67 , lattice simulations and eqs .
( [ spde ] ) yield @xmath95 , in good agreement with the prediction of ( [ res ] ) : @xmath96 . for the spirals wavelength ,
numerical results ( lattice simulations and spde ) yield @xmath97 as predicted by ( [ res ] ) . in fig .
[ wlength ] , the analytical prediction ( [ res ] ) for @xmath92 is compared with the values obtained from the spde ( [ spde ] ) , yielding a remarkable agreement for the functional dependence on the parameter @xmath4 . yet , as eq .
( [ cgle ] ) does not account for all nonlinearities , the analytical and numerical values differ by a prefactor @xmath98 ( considered in fig .
[ wlength ] ) @xcite .
it can still be noted that ( [ cgle ] ) and the predictions ( [ res ] ) are valid in all dimensions @xcite . , where @xmath99 is the spirals wavelength of the propagating spiral waves .
analytical results ( red curve , rescaled by a factor @xmath100 ; see text ) are compared with the solution of the spde ( [ spde ] ) ( black circles ) .
[ wlength ] ] motivated by recent experiments @xcite , we have considered a spatially - extended model with three species in cyclic competition , and focused on the spatial and stochastic effects . the local character of the reactions and internal noise allow mobile populations to coexist and lead to pattern formation .
we have shown that already for finite mobility the lattice model can be described by spde . with the latter and lattice simulations ,
we have studied how entanglement of spirals form and obtained expressions for their spreading velocity and wavelength .
the size of the patterns crucially depends on the diffusivity : above a certain threshold the system is covered by one species @xcite . in the absence of noise , the equations still predict the formation of spiral waves , but their spatial arrangement depends on the initial conditions .
support of the german excellence initiative via the program
nanosystems initiative munich " is gratefully acknowledged .
m. m. is grateful to the humboldt foundation for support through the grant iv - scz/1119205 . + currently at : mathematics institute & centre for complexity science , university of warwick , coventry cv4 7al , u.k .
26 natexlab#1#1bibnamefont # 1#1bibfnamefont # 1#1citenamefont # 1#1url # 1`#1`urlprefix[2]#2 [ 2][]#2 , * * , ( ) . ,
_ _ ( , ) ; , _ _ ( , ) . , _ _ ( , ) . , * * , ( ) . , * * , ( ) ; * * , ( ) . , , , * * , ( ) . in _ _ , edited by ( , ) ; , * * , ( ) ; , , , * * , ( ) . , * * , ( ) , _ _ ( , ) . , * * , ( ) . , * * , ( ) . , , , * * , ( ) . ,
* * , ( ) . ,
* * , ( ) . , , , * * , ( ) . ,
_ _ ( , ) , ed . , , ,
.
http://www.xmds.org . ,
* * , ( ) .
, , , * * , ( ) . , _ _ ( , ) , ed . , * * , ( ) . ,
* * , ( ) . + |
signals transmitted over wireless channels undergo severe degradations due to effects such as path loss , shadowing , fading , and interference from other transmitters , in addition to thermal noise at the receiver . one major way to combat static fading is to provide diversity in either time , frequency , or space @xcite .
for this purpose , multiple - antenna systems that provide high orders of spatial diversity and high capacity have been extensively studied @xcite . however , due to limited terminal sizes , the implementation of two or more antennas may be impossible . based on the seminal works in @xcite and @xcite
, the authors in @xcite@xcite set up a framework for _ cooperative communications _
, where multiple terminals use the resources of each other to form a virtual antenna array .
following these works , many researchers have proposed distributed communication schemes and analyzed their outage probability behavior such as in @xcite@xcite@xcite@xcite@xcite .
the main protocols that have been proposed are the _ amplify - and - forward _ , where the relay only amplifies the signal received from the source , before transmitting it to the destination , and the _ decode - and - forward _ , where the relay decodes the received signal before transmitting it to the destination . in this paper
, we study the performance of distributed space - time bit - interleaved coded modulations ( d - st - bicm ) schemes for non - orthogonal amplify - and - forward protocols .
furthermore , we focus on situations where the transmitter to relay and inter - relay links quality is highly better than the transmitter to receiver link quality .
this situation occurs for example when deploying professional relays on top of buildings in a way to improve the link reliability in low coverage zones of a multi - cellular system .
the paper is organized as follows : section [ matryoshka ] defines the matryoshka block - fading channel , a channel that characterizes the cooperative protocol considered in this paper . in section [ system_model ] , we describe the system model and all the parameters involved in our study .
we then derive bounds on the diversity of d - st - bicm for the minimum cooperation frame length in section [ div1 ] , and section [ div2 ] extends these results for any length .
section [ simulations ] shows simulation results for different network topologies , while section [ conclusions ] gives the concluding remarks .
in this paper , we consider the block - fading channel model in which a d - st - bicm codeword undergoes a limited number of fading channel realizations , namely one fading coefficient per spatial path . for the sake of analysis , we introduce a block - fading channel model where the set of random variables of a higher diversity block always includes the set of random variables of a lower diversity block , in a way similar to nested matryoshka dolls .
let us consider @xmath0 independent fading random variables providing a total diversity order of @xmath0 .
let @xmath1 be a channel built from the concatenation of @xmath2 blocks , where @xmath3 and @xmath4 are respectively the sets of diversity orders and lengths of each block . as usual ,
the integer @xmath5 denotes the cardinality of the set @xmath6 .
the @xmath7-th block has a diversity order equal to @xmath8 and its fading set is @xmath9 with @xmath10 , @xmath11 fading random variables , such that @xmath12 .
thus , we have @xmath13 and @xmath14 or equivalently @xmath15 is the maximum diversity order . this channel defined by nested fading sets
is referred to as a matryoshka channel and it is illustrated in fig .
[ fig : matr ] .
let us now transmit a bpsk - modulated and interleaved codeword of a rate-@xmath16 code over the @xmath1 channel .
first , let us focus on the pairwise error probability ( pep ) of two given binary codewords @xmath17 and @xmath18 .
due to the channel model , the diversity order of this pep is equal to the diversity order of the lowest index block observing a non - zero part of @xmath19 .
the performance of the coded modulation has a diversity order upper - bounded by @xmath20 defined as follows : the diversity observed after decoding a rate-@xmath16 linear code transmitted over a @xmath1 channel is upper - bounded by @xmath21 where @xmath7 is given by the following inequalities : @xmath22 and @xmath20 is achievable for any systematic linear code .
+ [ prop1 ] _ proof : _ this proof is inspired from the singleton bound s one .
the code @xmath23 has length @xmath24 and dimension @xmath25 , where @xmath26 and @xmath27 . if @xmath28 , whatever the code is , a puncturing of the last @xmath29 bits leads to a zero minimum hamming distance because @xmath30
this means that there always exists two codewords @xmath17 and @xmath18 such that the last @xmath29 bits of @xmath19 are null , and involves that @xmath31 .
let us now suppose that the code is linear and systematic . if the information bits are transmitted on the blocks of highest diversity order and if @xmath32 , the hamming distance after puncturing the last @xmath33 bits remains strictly positive and induces that @xmath34.@xmath35 + it is straightforward to show that the bound on the diversity order applies to any discrete modulation . as a remark , in order to achieve the upper - bound on the diversity of a block - fading channel , non - zero bits of word @xmath19
should be placed in as many independent blocks as given by the singleton bound . for matryoshka channels ,
the bound is achieved as soon as one non - zero bit of any word @xmath19 is placed in a block of diversity higher than @xmath20 .
where @xmath64 is the length-@xmath36 vector of received signals and @xmath65 is the length-@xmath36 vector of @xmath66-qam symbols .
@xmath67 is a @xmath68 precoding matrix , and @xmath69 is upper - triangular as shown in ( 4 ) .
+ @xmath70\ ] ] @xmath71\ ] ] with : @xmath72 finally , the vector @xmath73 is a length-@xmath36 colored gaussian noise vector as given by ( 5 ) .
we set : @xmath74= 2 n_0 { \bf \theta}\ ] ] where the @xmath75 operator denotes transpose conjugate . by performing a cholesky decomposition on @xmath76 ,
we get : @xmath77 thus the equivalent channel model becomes : @xmath78 where @xmath79 is a white gaussian noise vector . + digital transmission is made as follows : uniformly distributed information bits are fed to a binary convolutional encoder .
coded bits @xmath80 are then interleaved and gray mapped into qam symbols .
the qam symbols are then rotated via @xmath81 and transmitted on the ssaf channel defined by @xmath69 given in ( [ vector_form ] ) .
the coherent detector at the destination computes an extrinsic information @xmath82 based on the knowledge of @xmath69 , the received vector @xmath64 , and independent _ a priori _
information @xmath83 for all coded bits .
the channel decoder then computes _ a posteriori _ probabilities ( app ) based on the de - interleaved extrinsic information coming from the detector using the forward - backward algorithm @xcite .
the transmitted information rate is equal to @xmath84 bits per channel use , where the cardinality of the qam constellation is @xmath66 . as a remark ,
one precoded symbol at the output of @xmath81 is transmitted over a row of the channel matrix @xmath69 and thus experiences a set of random variables @xmath85 .
if we assume that the quality of the source to relays and inter - relays links is much better than the source to destination or relay to destination links , we can then focus on the @xmath86 or @xmath87 random variables to understand the diversity behavior of such a system . indeed , in the context of professional relay deployment on top of buildings
, we may assume that the relays are placed and have their antennas tuned to ensure a good link quality with the base station .
furthermore , in the case of detect - and - forward or decode - and - forward protocols , this assumption is still relevant .
finally , one precoded symbol transmitted on the @xmath7-th row of the channel matrix sees a set of @xmath88 fading variables included in the set seen by a symbol sent on the @xmath89-th row .
hence , we will see in the sequel that the equivalent channels obtained by the use of a sequential slotted amplify and forward protocol fall into the class of matryoshka channels .
the maximum diversity inherent to the ssaf channel is @xmath91 , and it can be collected by an app detector ( at the destination ) if a full - diversity linear precoder is used at the transmitter .
the precoder mixes the @xmath90 constellation symbols being transmitted on the channel providing full diversity with uncoded systems and without increasing the complexity at the detector . using precoders that process spreading among more than @xmath90 time slots
can further improve the performance . from an algebraic point of view , a linear precoder of size @xmath92 is the optimal configuration to achieve good coding gains ( without channel coding ) @xcite at the price of an increase in detection complexity ( the complexity of an exhaustive app detector grows exponentially with the number of dimensions ) .
on the other hand , for coded systems transmitted on block - fading channels , the channel decoder is capable of collecting a certain amount of diversity that is however limited by the singleton bound @xcite . in @xcite@xcite ,
the modified singleton bound taking into account the rotation size over a mimo block - fading channel is used to achieve the best tradeoff between complexity and diversity .
for this purpose , we derive hereafter an upper - bound on the diversity order of a coded transmission over a precoded @xmath90-slot ssaf channel , and then deduce the precoding strategy to follow in order to achieve full diversity .
we will first assume that the interleaver of the bicm is ideal , which means that for any pair of codewords @xmath93 , the @xmath94 non - zero bits of @xmath19 are transmitted in different blocks of @xmath95 time periods , which means that no inter - slot inter - bit interference is experienced .
the interleaving , modulation and transmission through the channel convert the codewords @xmath17 and @xmath18 onto points @xmath96 and @xmath97 in a euclidean space . for a fixed channel , the performance
is directly linked to the euclidean squared distance @xmath98 , that can be rewritten as a sum of @xmath94 squared euclidean distances associated to the @xmath94 non - zero bits of @xmath19 .
for each of the @xmath94 squared euclidean distances , we can build an equivalent channel model which corresponds to the transmission of a bpsk modulation over one row of the channel matrix @xmath99 .
thus , several squared euclidean distances appear to be transmitted on the same equivalent channel and the squared distance @xmath98 can be factorized as follows : @xmath100 where @xmath101 is linearly dependent on the norm of the @xmath102-th row of @xmath99 . in other words , at the output of the app detector ,
an equivalent block - fading channel is observed and the constituent blocks do not have the same intrinsic diversity order : a soft output belonging to the @xmath103-th block carries the attenuation coefficients @xmath104 . as a remark ,
blocks are sorted such that the @xmath103-th block carries a diversity order of @xmath105 and the subset of realizations of random variables observed in the @xmath7-th row of @xmath99 is included in the subset of random variables observed in the @xmath89-th row of @xmath99 . as the same modulation is used on each time slot of the relaying protocol
, each block length is equal to @xmath106 .
finally , the equivalent @xmath90-slot ssaf channel at the output of the app detector is a matryoshka @xmath107 $ ] , @xmath108)$ ] channel , where @xmath24 is the number of coded bits per codeword . with this observation
, we can conclude that the upper - bound on the diversity order of a non - precoded ssaf channel is @xmath109 which is equal to the classical singleton bound on the diversity order of block - fading channels @xcite , with the difference that it can be achieved by any systematic code . for the sake of generalization
, we now suppose that modulations with different spectral efficiencies are sent over the @xmath90 slots of the cooperation frame .
we define @xmath110 as the number of bits carried by one symbol of the modulation transmitted on the @xmath102-th time slot . in this case , the block fading channel is a @xmath111 , \left[\frac{m_{1}n}{\sum_{k=1}^{\beta + 1 } m_k } , \frac{m_{2}n}{\sum_{k=1}^{\beta + 1 } m_k } , ... , \frac{m_{\beta+1}n}{\sum_{k=1}^{\beta + 1 } m_k } \right ] \right ) $ ] matryoshka channel . by applying ( [ matryoshkabound ] )
, we obtain that if : @xmath112 then the achievable diversity order is @xmath113 . for a given distribution of spectral efficiencies
, it is better to choose @xmath114 , as a higher diversity order might be achieved for a given coding rate .
it is also clear that higher coding rates than in ( [ d1 ] ) can be attained for a given target diversity .
however , the bound on the diversity does not give any information on the coding gain of the coded scheme .
we will see later that a fine tuning of the choice of the spectral efficiencies might be needed to optimize the coding gain .
for example , the orthogonal amplify - and - forward protocol leads to @xmath115 , which provides full diversity whatever the code rate is but exhibits a poor coding gain @xcite .
let us now introduce a linear precoder that rotates symbols of @xmath116 different diversity blocks together .
first of all , let us focus on two different scenarios : * the linear precoder size is lower than ( or equal to ) @xmath90 . in this case
, the dimension of the received vector @xmath64 remains unchanged , thus there is no increase in detection complexity when an exhaustive app detector is used , and no delay is introduced to the protocol .
the authors in @xcite considered the design of such precoders for uncoded systems . *
the linear precoder size is lower than ( or equal to ) @xmath117 , where @xmath118 is the delay ( i.e. the source broadcasts for @xmath119 time slots before the relays start to cooperate ) . in this case
, the complexity of the detector increases exponentially with @xmath118 .
as mentionned previously , these precoders are mandatory to achieve optimal performance for uncoded systems . as we focus on channel coding issues in this work , delay - precoders
will not be considered in the sequel .
we will now present two precoding strategies and compute the bound ( [ matryoshkabound ] ) for these two particular cases .
[ [ first - strategy - a - single - precoder ] ] first strategy : a single precoder + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + first , let us assume that @xmath116 diversity blocks of size @xmath106 are linearly precoded together , then the diversity order of the new @xmath120-length block is the maximum diversity order of the precoded blocks . as the other blocks keep their own diversity , it seems natural to maximize their diversity orders in a way to increase the coding gain at the output of the decoder ( the best performance is achieved for a block - fading channel with diversity orders as equal as possible . ) .
the length of the precoder input vector is @xmath121 .
we propose to precode the first block with the @xmath122 last blocks , i.e. the highest diversity order with the @xmath122 lowest ones . at the output of the app detector ,
the channel model is a matryoshka @xmath123 channel where @xmath124 $ ] and @xmath125 $ ] , which leads to the following upper - bound on the diversity order : @xmath126 indeed , by replacing @xmath124 $ ] and @xmath125 $ ] in ( [ matryoshkabound ] ) , we observe that if @xmath127 then @xmath128 and @xmath129 .
else , if @xmath130 , then @xmath131 which leads to @xmath132 .
note that , in the representation of fig .
[ fig : matr ] , we have that @xmath133 in this case . if @xmath134 , then @xmath135 is equal to the singleton bound on the diversity order of an uncorrelated block - fading channel with equal per - block diversity . if @xmath136 , @xmath135 is greater than the upper - bound on the diversity order for block - fading channels .
for example , the full diversity order can not be achieved for the transmission of a @xmath137-precoded bicm with rate @xmath138 on a block - fading channel with diversity order @xmath139 ( the diversity is upper - bounded by @xmath140 ) . for the ssaf channel
, the full diversity order can be achieved in that case , as shown in fig .
[ diversity ] .
[ table1 ] and [ table2 ] show the values of @xmath135 for different coding rates with respect to the number of relays and the value of @xmath116 .
we can notice that full diversity is obtained with @xmath141 in all configurations . [ [ second - strategy - beta1s - precoders ] ] second strategy : @xmath142 precoders + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + let us assume that @xmath116 divides @xmath90 , we can then use @xmath142 precoders : the first precodes the highest diversity order block with the @xmath122 lowest ones . the second , if any , precodes the second highest diversity order block with the @xmath122 lowest non - precoded ones , and so on . by using this precoding strategy that includes several independent precoders , we further increase the diversity of the extrinsic probabilities at the input of the decoder , and consequently the diversity at the output of the decoder .
indeed , the equivalent @xmath123 channel has parameters @xmath143 $ ] and @xmath144 $ ] , which leads to the following upper - bound on the diversity order : @xmath145 it can be easily shown that @xmath146 however , the maximum diversity order @xmath147 is achieved for the same @xmath141 .
the advantage of @xmath148 over @xmath135 is for non - full diversity schemes . in addition
, it is important to note that the bounds in ( [ oneprecoder ] ) and ( [ severalprecoders ] ) have straight - forward applications to systems employing delay precoders .
now we reconsider the scenario of section [ diff_mod ] , in which different modulation sizes are sent over the blocks .
in addition , we consider that a space - time precoder with spreading @xmath116 combines the symbol having maximum diversity with those having the least @xmath122 diversity orders .
we thus obtain a @xmath149 \left[\frac{\left(m_{1}+ \sum_{\gamma=1}^{s-1 } m_{\beta+2-\gamma } \right)n}{\sum_{k=1}^{\beta + 1 } m_k } , \frac{m_{2}n}{\sum_{k=1}^{\beta + 1 } m_k}, ... ,\frac{m_{\beta+2-s}n}{\sum_{k=1}^{\beta + 1 } m_k } \right ] \right)$ ] matryoshka channel . by applying ( [ matryoshkabound ] )
, we obtain that if : @xmath150 with : @xmath151 then the achievable diversity order is @xmath113 .
thus , the parameters @xmath152 allow for a fine tuning of the target diversity for a given coding rate .
this tuning allows to further improve the coding gain .
unfortunately , the theoretical of coding gains for coded modulations on block fading channels is difficult and often by extensive computer simulations .
, the analysis of such a design is out of the scope of this paper , which mainly focuses on diversity orders optimization .
so far , we have considered the @xmath37-relay ssaf protocol with length-@xmath121 cooperation frames . in @xcite ,
the authors consider a cooperation scheme ( for @xmath140-relay ssaf and higher ) in which the cooperation frame is stretched in a way to protect more symbols .
in other words , we consider the @xmath36-slot @xmath37-relay ssaf protocol with : @xmath154 where @xmath155 is the number of additional slots .
the goal of this extension is to increase the number of coded bits that experience full diversity .
the first @xmath156 symbols in * x * from ( [ channel_model ] ) will have maximum diversity , which reduces to the first symbol having maximum diversity in the @xmath90-slot ssaf scenario .
however , this additional protection entails an increase in the size of * x * , thus complexity at the app detector increases as well .
+ an illustration of this scheme is provided in fig .
[ m - slot ] for the @xmath157-slot @xmath140-relay ssaf protocol ; the source always transmits a constellation symbol , and starting from the second time slot , the relays cooperate in a round robin way ; in this case , the first @xmath158 out of a total of @xmath157 constellation symbols have a maximum diversity @xmath159 .
it is then clear that this protocol allows to achieve full diversity with higher coding rates . in the sequel
we will provide bounds on the diversity order of coded modulations under this cooperative protocol .
we first consider the @xmath160-slot @xmath37-relay ssaf protocol without precoding .
we thus obtain a matryoshka block - fading channel as @xmath161 with @xmath162 $ ] and @xmath163 $ ] .
this means that in a cooperation frame of length @xmath160 , there are @xmath164 symbols that have maximum diversity @xmath165 .
this makes clear the fact that higher coding rates can be attained with this scheme .
the diversity of a non - precoded bicm over this protocol is given by : @xmath166 hence , we attain the maximum diversity order if : @xmath167 which implies that we can - theoretically - achieve full diversity with a coding rate getting close to @xmath168 as @xmath155 increases , but at the price of an app detection complexity increase . to illustrate this bound on diversity , fig .
[ rate ] gives examples of the @xmath140-relay , @xmath139-relay , and @xmath169-relay ssaf channels .
we noticed that with increasing the number of slots , the maximum coding rate has a logarithmic - like growth , while the complexity at the detector increases exponentially ( as the cardinality of the received vector @xmath64 is @xmath170 ) .
this means that only few additional slots can be practically added to the cooperation frame in order to provide a reasonable rate / complexity tradeoff . if we precode the first symbol with maximum diversity with the @xmath122 symbols having the lowest diversity orders we obtain a @xmath171 block - fading channel where @xmath162 $ ] and @xmath172 $ ]
it is clear then that we provide @xmath173 symbols having maximum diversity with precoding .
the bound on diversity with a single precoder is given by : @xmath174 full diversity is obtained for @xmath175 which , again , shows that linear precoding can be used to increase the obtained diversity without increasing the complexity of an optimal app detector .
in this section , word error rate performances of different d - st - bicm schemes are compared to information outage probability for different system configurations .
we consider the single - relay ( fig .
[ 1 - 64qam ] ) , two - relay ( fig .
[ 2 - 16qam ] ) , and three - relay ( fig . [ 3-qpsk ] ) half - duplex ssaf cooperative channels with different coding rates and constellation sizes .
we use interleavers designed as in @xcite with an additional constraint to transmit the systematic bits on the higher diversity blocks of the equivalent matryoshka channels .
we set the values of @xmath176 , and @xmath177 , so that the received average energy over all the time slots is invariant .
the space - time precoders are built using algebraic rotations from @xcite ( see appendix [ appen ] for further details ) , and the number of iterations between the detector and the decoder is fixed to @xmath178 . fig .
[ 1 - 64qam ] shows the performance of st - bicm over the single - relay ssaf channel using 64-qam modulation and half - rate coding .
following @xmath179 , no rotation is needed , as the channel decoder with optimized interleaving is capable of recovering the maximum available diversity . for small to moderate signal - to - noise ratios , and due to noise amplification at the relay , precoding the signal constellation does not affect the performance .
from moderate to high signal - to - noise ratios , a rotation yields a severe performance degradation .
this is due to the fact that interference between symbols ( due to the rotation ) becomes too heavy for the decoder and thus affects the coding gain . in fig .
[ 2 - 16qam ] , various coding strategies using rsc codes for the @xmath140-relay ssaf protocol , all at an information rate of @xmath180 b / s / hz , are compared to gaussian input outage probability .
the first observation is that orthogonal coded schemes suffer from weak coding gains @xcite , although providing full diversity . for the curves employing the ssaf protocol , coding strategies following @xmath179 in ( [ d1 ] ) , @xmath181 in ( [ oneprecoder ] ) , and the bound on the coding rate derived in section [ diff_mod ] .
the best strategy is shown to be the @xmath182 code with an @xmath137 rotation with qpsk modulation in the three slots , following @xmath181 .
[ 3-qpsk ] shows the performance of the ssaf with three relays using qpsk modulation and the half - rate @xmath183 rsc code .
the three strategies following @xmath181 and @xmath184 achieve full diversity with @xmath185 . in case
no precoder is available at the source and we want to transmit at the same coding rate , another option is to follow @xmath186 from ( [ d4 ] ) , thus extending the cooperation frame with @xmath187 slots .
this strategy allows to achieve full diversity without precoding , as shown with the dashed blue curve .
we studied coding strategies for the non - orthogonal amplify - and - forward half - duplex cooperative fading channel .
finally , performances close to outage probabilities for different number of relays , coding rates , and constellation sizes are shown .
the real @xmath188 cyclotomic rotation from @xcite can be written as : @xmath189\ ] ] with @xmath190 .
suppose now we have to transmit a half - rate code over the @xmath139 relay ssaf channel . according to @xmath181 , one rotation with @xmath137 is sufficient .
this gives the following space - time precoder : @xmath191\ ] ] according to @xmath184 , we need two rotations with @xmath137 each .
this gives the following space - time precoder : @xmath192\ ] ] e. viterbo , `` table of best known full diversity algebraic rotations , '' available at : link : www1.tlc.polito.it/~ viterbo / rotations / rotations.html[www1.tlc.polito.it/~ viterbo / rotations / rotations.html ] .
-slot ssaf protocol with two relays .
the cooperation frame has length @xmath197 , a light gray rectangle means that the terminal is emitting , a dark blue rectangle means the terminal is receiving .
a white rectangle means the terminal is inactive . ] | in this work , we consider the problem of coding for the half - duplex non - orthogonal amplify - and - forward ( naf ) cooperative channel where the transmitter to relay and the inter - relay links are highly reliable .
we derive bounds on the diversity order of the naf protocol that are achieved by a distributed space - time bit - interleaved coded modulation ( d - st - bicm ) scheme under iterative app detection and decoding .
these bounds lead to the design of space - time precoders that ensure maximum diversity order and high coding gains .
the word error rate performance of d - st - bicm are also compared to outage probability limits . |
the basic concept of any obstruction theory is to produce an invariant associated to a specific construction in such a way that the nature of the invariant determines whether the construction can or can not be performed .
an ( equivariant ) obstruction theory considers two basic problems . for a finite group @xmath4 ,
consider a relative @xmath4-cellular complex @xmath5 such that the @xmath4-action on @xmath6 is * free .
* let @xmath7 be a @xmath4-space .
_ extension problem .
_ let @xmath8 be a @xmath4-map .
is there a @xmath4-map @xmath9 such that @xmath10 ? here @xmath11 denotes the inclusion .
_ homotopy problem .
_ let @xmath12 and @xmath13 be @xmath4-maps such that there is a @xmath4-homotopy @xmath14 from @xmath15 to @xmath16 .
is there a @xmath4-homotopy @xmath17 which extends @xmath18 , i.e , @xmath19 , @xmath20 and @xmath21 ? the answer which obstruction theory provides is a sequence of obstruction elements living in equivariant cohomology . for the details about ( equivariant )
obstruction theory one can consult the expositions in ( * ? ? ?
* section ii .
3 ) , ( * ? ? ?
* chapter 7 ) and ( * ? ? ?
* section v. 5 ) .
let @xmath5 be a relative @xmath4-cellular complex with a free action on @xmath6 .
let @xmath22 denote the integral cellular chain complex .
the cellular free @xmath4-action on every skeleton of @xmath6 induces a free @xmath4-action on the chain complex @xmath22 .
therefore , the chain complex @xmath22 is actually a chain complex of free @xmath23$]-modules .
let @xmath5 be a relative @xmath4-cellular complex such that the @xmath4-action on @xmath6 is free .
let @xmath24 be a @xmath23$]-module .
\(a ) the chain complex@xmath25}m
\label{eq:0}\]]is the equivariant chain complex of @xmath5 with coefficients in @xmath24 , and its homology @xmath26 is called the * equivariant homology * of @xmath5 with coefficients in @xmath24 .
\(b ) the cochain complex @xmath27}(c_{\ast } ( x , a),m ) \label{eq:1}\]]is the equivariant cochain complex of @xmath5 with coefficients in @xmath24 , and its homology @xmath28 is called the * equivariant cohomology * of @xmath5 with coefficients in @xmath24 .
a central example of a @xmath23$]-module in our applications will come from a path connected , @xmath29-simple
@xmath4-space @xmath7 .
being @xmath29**-simple * * means that @xmath30 acts trivially on @xmath31 for every @xmath32 .
the action of @xmath4 on @xmath7 easily produces an action on the set of homotopy classes @xmath33 $ ] .
since @xmath7 is @xmath29-simple , the action can be extended to the homotopy group @xmath34 $ ] .
consequently there exists a @xmath23$]-module structure on @xmath35 .
the equivariant homology and cohomology groups of @xmath5 can be interpreted as the ordinary homology and cohomology groups of the quotient pair @xmath36 with appropriate local coefficients @xcite . for @xmath24
interpreted as a local coefficient system @xmath37 on @xmath38:@xmath39@xmath40 let @xmath41 be a fixed integer and @xmath7 a path - connected @xmath29-simple @xmath4-space . for every @xmath4-relative cell complex @xmath5 with * free * action of @xmath4 on @xmath6 , there exists an obstruction exact sequence@xmath42_{g}\longrightarrow \mathrm{im}\left ( [ x_{n},y]_{g}\rightarrow \lbrack x_{n-1},y]_{g}\right ) \overset{\mathfrak{o}_{g}^{n+1}}{\longrightarrow } \mathfrak{h}_{g}^{n+1}(x , a;\pi _ { n}y ) , \label{obstructionexactsequence}\]]where @xmath43 denotes the @xmath44-skeleton of @xmath45 .
the sequence is natural in @xmath45 and @xmath7 .
this exact sequence should be understood in the following way : \(a ) every @xmath4-map on the @xmath46-skeleton @xmath47 which can be equivariantly extended to the @xmath29-skeleton @xmath48 defines an unique element @xmath49 living in @xmath50 , called the * obstruction element * ; \(b ) the exactness of the sequence means that the obstruction element @xmath49 is zero if and only if there is a map in the homotopy class of the restriction @xmath51 which can be extended to the @xmath52-skeleton @xmath53 . the obstruction element can be introduced on the cochain level by a universal geometrical construction .
let @xmath54\in \lbrack x_{n},y]_{g}$ ] , let @xmath55 be an attaching map and @xmath56 the generator associated to the cell @xmath57 .
then the * * obstruction cochain * * _ _ _ _ @xmath58 of the map @xmath18 is defined on @xmath59 ( with a little abuse of the notation ) by@xmath60\in \lbrack s^{n},y ] .
\label{defobstructioncocycle}\]]it can be proved that the cohomology class of the obstruction cocycle is the obstruction element defined by the exact sequence ( obstructionexactsequence ) .
let @xmath41 be a fixed integer .
the following proposition holds for an @xmath46-connected , @xmath29-simple @xmath4-space @xmath7 .
[ prop : primalobstruction]let @xmath5 be a relative @xmath4-cell complex with the action of @xmath4 free on @xmath6 , and @xmath8 any @xmath61-map .
\(a ) there exists a @xmath4-map @xmath62 extending @xmath63 , i.e. , @xmath64 .
\(b ) every two @xmath4-extensions @xmath18 and @xmath65 of @xmath63 are @xmath4-homotopic @xmath66 on @xmath67 , that is @xmath68_{g}\rightarrow \lbrack x_{n-1},y]_{g}\right ) = \{\ast \}.\ ] ] \(c ) if @xmath69 is a @xmath4-homotopy between @xmath70 and @xmath71 , and @xmath72 are the extensions of @xmath63 and @xmath73 , then there is a @xmath4-homotopy @xmath74 extending @xmath75 between @xmath76 and @xmath77 .
\(a ) we extend @xmath78 starting with the @xmath79-skeleton and go up to the @xmath29-skeleton . the obstruction for lifting @xmath80 from the @xmath81-skeleton to the @xmath82-skeleton lies in @xmath83 .
since @xmath7 is @xmath46-connected and @xmath29-simple , @xmath84hence , there is a @xmath4-map @xmath62 extending @xmath63 .
\(b ) let @xmath72 be @xmath4-extensions of @xmath63 .
a @xmath4-map@xmath85can be defined by @xmath86now consider a relative @xmath4-cell complex @xmath87 and extend @xmath88 .
the assumptions on @xmath7 provide a @xmath4-map @xmath75 from the @xmath29-skeleton @xmath89 of @xmath90 to @xmath7 which extends @xmath88 .
the restriction @xmath91 is the required @xmath4-homotopy @xmath66 .
\(c ) considering a relative @xmath4-cell complex @xmath92 instead of @xmath5 the statement becomes a direct consequence of the property ( a ) .
when@xmath68_{g}\rightarrow \lbrack x_{n-1},y]_{g}\right ) = \{\ast \},\]]the obstruction sequence ( [ obstructionexactsequence ] ) becomes @xmath42_{g}\longrightarrow \{\ast \}\overset{\mathfrak{o}_{g}^{n+1}}{\longrightarrow } \mathfrak{h}_{g}^{n+1}(x,\pi _ { n}y ) .
\label{primaryobstrsequence}\ ] ] the element @xmath93 is called the * primary obstruction * and does not depend on the map of the @xmath29-th skeleton .
if @xmath94 and @xmath95 are @xmath4-actions on @xmath96 and @xmath94 is free , then there exists a @xmath4-map @xmath97 such that @xmath98for all @xmath99 and @xmath100 .
the corollary is a direct consequence of the previous proposition statement ( a ) .
let @xmath45 be a compact @xmath29-dimensional free @xmath4-manifold .
then there is a version of the poincar duality isomorphism for the equivariant homology and cohomology of@xmath101 with the coefficients in any @xmath4-module @xmath24 .
[ th : eqpoincare]let @xmath45 be a compact @xmath29-dimensional , simply connected , free @xmath4-manifold , @xmath102 , a @xmath4-module , @xmath103 , and @xmath24 a @xmath4-module .
then there is an isomorphism of the groups@xmath104for every @xmath105 .
when @xmath24 is interpreted as the local coefficient system @xmath37 , the isomorphism ( [ eq:4 ] ) implies : @xmath106from ( * ? ? ?
* theorem 4.18 .
p. 196 ) and hatcher we know that @xmath107 is a compact manifold with fundamental group @xmath108 .
since all the assumptions for applying theorem 2.1 @xcite are met , @xmath109 careful reading of the definition of the modified homology @xmath110.@xmath111 , @xcite implies that@xmath112the isomorphism ( [ eq:2 ] ) concludes the proof@xmath113 let @xmath24 be a connected , @xmath52-dimensional , compact free @xmath4-manifold , @xmath115 a @xmath116-dimensional smooth @xmath4-manifold , and @xmath117 the union of a finite @xmath4-invariant arrangement @xmath118 of the @xmath119-dimensional smooth submanifolds .
let us also assume that \(a ) the complement @xmath120 is @xmath29-simple , paracompact space , \(b ) the complement @xmath120 is @xmath46-connected , \(c ) the tangent spaces of the submanifolds @xmath121 in any mutual intersection point do not coincide , and \(d ) @xmath122 . _ the question we consider is whether there is a _
@xmath4-_map _ @xmath114__. _ _ [ [ section ] ] * 1 . * since the complement @xmath120 is @xmath46-connected by assumption , the problem of the existence of a @xmath4-map @xmath123 depends only on primary obstruction .
the obstruction exact sequence , like in ( [ primaryobstrsequence ] ) , has the form@xmath124_{g}\longrightarrow \{\ast \}\overset{\mathfrak{o}_{g}^{n+1}}{\longrightarrow } \mathfrak{h}_{g}^{n+1}(m,\pi _ { n}\left ( w\backslash \sigma \right ) ) .\]]the assumptions ( a ) and ( b ) on the space @xmath120 and the hurewicz theorem imply that @xmath125 as @xmath4-modules .
thus , the obstruction element @xmath126 lives in the group @xmath127 , where the natural @xmath4-structure on @xmath128 is assumed .
consider the equivariant cohomology group @xmath127 .
since @xmath24 is a free @xmath4-manifold by assumption , theorem th : eqpoincare provides an isomorphism @xmath129where @xmath102 is the @xmath4-module @xmath130 .
the isomorphism ( [ eq:2 ] ) implies that@xmath131since @xmath132 an application of proposition 5.14 .
of @xcite ( or alternatively ( * ? ? ?
* exercise 1 , p.44 ) and ( * ? ? ? * ( 1.5 ) , p.57 ) ) provides an isomorphism @xmath133thus the obstruction element lives in a group of coinvariants of the first non - trivial reduced homology group of the target space @xmath120 , @xmath134 [ [ section-1 ] ] * 2 .
* the situation where the primary obstruction is the only obstruction , as in our case , has the advantage of not depending on the particular @xmath4-map on the @xmath29-th skeleton .
[ def : mapingeneralposition]let a @xmath4-map @xmath135 satisfy the following conditions _ ( a ) _ @xmath136 , _ ( b ) _
@xmath137 is a finite set of points , _ ( c ) _
@xmath138 @xmath139 @xmath140 intersection is transversal at @xmath141 , _ ( d ) _ @xmath142@xmath143 .
we then say that @xmath135 is a map in * general position * with respect to @xmath117
. condition ( d ) allows the intersection points @xmath137 to belong to the lower strata of the arrangement @xmath144 .
this forces the introduction of broken point classes along point classes as possible results of evaluation of the obstruction cocycle .
[ [ section-2 ] ] * 3 . *
the notion of point and broken point classes was introduced in @xcite and @xcite for the complements of arrangements of linear spaces .
we extend this definition to the present setting .
consider @xmath145 .
there are elements @xmath146 in @xmath144 such that @xmath147 and @xmath148 .
let @xmath149 denote disks in fibers at the point @xmath141 of tubular neighborhoods of the submanifolds @xmath146 such that for all @xmath150we have @xmath151 .
the smoothness assumptions on @xmath115 and elements of the arrangement @xmath144 guarantee the existence of the above construction bredon .
the fundamental class of the pair @xmath152 determines a homology class in @xmath153 that we denote by @xmath154 $ ] and call the _ _ _ _ * point class * of @xmath141 determined by @xmath155 .
the homotopy axiom implies that @xmath154 $ ] does not change if @xmath141 is moved inside the connected component of @xmath156 . because of the assumption ( d ) , at the beginning of the section , that @xmath122 , the epimorphism @xmath157 determines the class @xmath158 $ ] which is also called the * * point class * * _ _ _ _ of _ _ _ _ @xmath141 determined by @xmath155 .
thus all point classes @xmath159 are born as @xmath154 $ ] classes , but @xmath159 can be zero while @xmath154\neq 0 $ ] .
[ prop : singularset]consider a @xmath4-map @xmath135 in general position with respect to @xmath117 .
the obstruction element @xmath160 is the equivariant poincar dual of the point set @xmath161 understood as a chain in the group @xmath162with the appropriate coefficients in the group of coefficients .
[ [ section-3 ] ] * 4 . * to compute the obstruction cocycle and the obstruction element
, we have to choose at least one equivariant cell structure on @xmath24 compatible with the given action .
we choose two equivariant cell structures connected by a cellular map . to simplify the exposition , from now on
we assume that @xmath102 is a trivial @xmath4-module .
usually , the first equivariant cell structure induced on @xmath24 is a simplicial one .
it is used to define a piecewise affine @xmath4-map in general position @xmath135 .
the advantage of the simplicial structure is that the map is completely determined by the images of the vertex orbits and the requirement that the map is piecewise affine .
the second equivariant cell structure should satisfy the requirement that the top dimensional group of chains is generated equivariantly by a single cell @xmath59 , and therefore @xmath163e.\]]such a structure will be called an * economic * @xmath4-structure of @xmath24 .
the economic @xmath4-cell structure does not have to exist .
[ [ section-4 ] ] 5 .
the obstruction cocycle @xmath49 is computed using the simplicial cell structure and the geometric definition of the obstruction cocycle .
[ th : obcocycle]let @xmath135 be a map in general position in respect to @xmath117 .
then for a @xmath52-simplex @xmath164 of @xmath24 the following formula holds @xmath165here @xmath166 denotes the intersection number of the image @xmath167 and the appropriate oriented element @xmath168 of the arrangement @xmath144 .
a cellular map between two cell structures allows evaluation of the obstruction cocycle in both structures . in an economic @xmath4-structure
, every maximal cochain is determined by its value on the equivariant generator @xmath59 .
therefore the obstruction cocycle can be treated as an element of the coefficient group @xmath169 expressed as a linear combination of the point classes .
the obstruction element is the class @xmath170 $ ] , along the quotient homomorphism @xmath171 .
there exists a @xmath4-map @xmath114 if and only if @xmath172\in h_{n}(w\backslash \sigma , \mathbb{z } ) _ { g}$ ] is zero . the first proof of the lovsz conjecture was given by babson and kozlov @xcite . a revealingly simple proof was given by schultz , @xcite , @xcite .
kozlov gave another proof in kozlov .
let @xmath173 be a graph , @xmath174 and @xmath175 .
if @xmath176 is @xmath46-connected , then @xmath173 is not @xmath177-colorable .
here @xmath178 is a circular graph with @xmath179 edges .
let us assume that @xmath173 is @xmath177-colorable .
this means that there exists a graph homomorphism @xmath180 , and consequently a map @xmath181 for every graph @xmath75 .
when we put @xmath178with a @xmath182-action instead of @xmath75 , we end up with the @xmath182-equivariant map@xmath183the assumption that @xmath176 is @xmath46-connected implies the existence of a @xmath182 map@xmath184where @xmath96 is equipped with the antipodal action .
thus to prove the lovas conjecture it is enough to prove that there is no @xmath182-map @xmath185 in @xcite , c. schultz proved the nonexistence of the map ( [ map3 ] ) by comparing the complex @xmath186 with the complement of the torus arrangement and then performing some characteristic class computations .
let us reproduce this beautiful construction and substitute characteristic class computations with obstruction theory .
let @xmath187 be a torus , and @xmath188 the union of the arrangement of the following @xmath189 subtoruses@xmath190if we define a @xmath191 action on @xmath192 by@xmath193it is apparent that @xmath188 is a @xmath182-invariant subspace of @xmath192 .
proposition 2.9 of @xcite says that there exists a @xmath182-map@xmath194thus , the assumption that @xmath173 is @xmath177-colorable implies the existence of a @xmath182-equivariant map@xmath195the following theorem provides a contradiction which implies lovsz conjecture .
there is no @xmath182-equivariant map @xmath196 .
the codimension of @xmath188 inside @xmath187 is @xmath29 .
since @xmath188 is compact , locally contractible space and @xmath197 for @xmath198 then poincar - lefschetz duality , ( ( * ? ? ?
* corollary 8.4 .
p.352 ) , ( * ? ? ?
* proposition 3.46 .
p.254 ) ) , combined with long exact sequence in cohomology , implies that the complement @xmath199 is @xmath200 connected .
thus , the existence of a @xmath182-map @xmath196 is determined by the primary obstruction .
let @xmath201 . a map @xmath202 defined by@xmath203is a @xmath182-map in general position and does not hit lower strata of the arrangement .
indeed , the intersection contains two points@xmath204moreover , @xmath205and @xmath206 .
proposition prop : singularset implies that the obstruction element living in @xmath207is the equivariant poincar dual of the ( orbit of the ) point in @xmath208the isomorphisms ( * ? ? ?
* exercise 1 , p.44 ) , ( * ? ? ? * ( 1.5 ) , p.57 ) , and the fact that @xmath209imply@xmath210since the orbit of a point is a generator of the @xmath79-homology , the obstruction element is not zero and the map does not exist . in this example
we were fortunate , because the intersection of the image of the general position map and the forbidden set was just an orbit of a point . that allowed a direct application of proposition [ prop : singularset ]
this chapter is contains an extension of makeev s result @xcite . a * proper * borel measure on the sphere @xmath0 is a borel measure @xmath211 such that \(a ) @xmath212)=0 $ ] for any circular arc @xmath213\subset s^{2}$ ] , and \(b ) @xmath214 for each nonempty open set @xmath215 .
let @xmath216 , @xmath217 and @xmath218 be three planes through the origin in @xmath219 .
planes are in a * fan position * if they intersect along a common line .
the planes in a fan position cut the sphere @xmath0 in six parts @xmath220 naturally oriented up to a cyclic permutation .
[ problem1]find all the six - tuples @xmath221 such that @xmath222 , and that for any proper borel probability measure @xmath211 on the sphere @xmath0 there exist three planes in a fan position , with angular sectors having the prescribed amount of the measure , i.e. , for all @xmath223 , @xmath224the six - tuples which satisfy these conditions are * solutions * of the problem .
the existence of the equipartition solution was proved by v. v. makeev @xcite . modifying the configuration space and the test map , we prove in section [ sec : thmain1 ] that beside @xmath225 there exist at least two more solutions .
[ th : main1]let @xmath211 be a _ proper _ borel probability measure on the sphere @xmath0 . then there are three planes intersecting along a line such that the ratio of the measure @xmath211 in the angular sectors cut by the planes is@xmath226 we use the configuration space / test map scheme to reduce the partition problem to an equivariant one .
the basic idea , which we modify , comes from the papers of imre brny and jii matouek bama2001 , @xcite .
the @xmath44**-fan * * @xmath227 in @xmath219 ( or on @xmath0 ) is formed from an oriented line through the origin @xmath228 and @xmath44 closed half planes @xmath229 which intersect along the common boundary @xmath230 .
the intersection of a @xmath44-fan with the sphere @xmath0 is also equally called a @xmath44-fan .
thus , the collection @xmath231 of a point @xmath232 and @xmath44 great semicircles @xmath233 emanating from @xmath141 is also a @xmath44-fan .
sometimes instead of great semicircles we use : \(a ) open angular sectors @xmath234 between @xmath235 and @xmath236 ; or \(b ) tangent vectors @xmath237 which are determined by the great semicircle curves @xmath235 , @xmath238 . here
@xmath239 denotes the tangent space at a point @xmath232 .
so there are four equally useful notations for a single @xmath44-fan , and we prefer the tangent vector notation @xmath240 .
the space of all @xmath44-fans in @xmath219 will be denoted by @xmath241 .
* the configuration space . * for a proper borel probability measure @xmath211 on @xmath0 and @xmath242 , the configuration space is defined by @xmath243since every @xmath29-fan @xmath244 of the configuration space @xmath245 is completely determined by the pair @xmath246 and the measure @xmath211 , there is a homeomorphism @xmath247 . *
the test map .
* let us fix the `` symmetric '' six - tuple @xmath248 such that @xmath249 .
the test map for our problem is defined by @xmath250where @xmath251 is the angle between tangent vectors @xmath252 and @xmath253 in the tangent plane @xmath239 . here
we assume that @xmath254 .
* the action . * the dihedral group @xmath255 acts both on the configuration space @xmath245 and on the hyperplane @xmath256 in the following way @xmath257for @xmath258 and @xmath259 .
the action of @xmath260 on @xmath245 is * free * and the test map @xmath261 is equivariant . * the test space . *
the test space in this symmetric problem is the union @xmath262 of the smallest @xmath260-invariant arrangement @xmath263 , which contains a linear subspace @xmath264 defined by the equalities @xmath265 we have proved the basic proposition of the configuration space / test map scheme .
[ prop : vezaproblem - ekvivarijantan]let @xmath266 be a symmetric @xmath267-tuple such that @xmath268 . if there is no @xmath260-equivariant map@xmath269then for every proper borel probability measure on the sphere @xmath0 there exist three planes in a fan position with angular sectors such that@xmath270 the extension of scalars from homological algebra ( as shown in @xcite and @xcite ) allows us to prove the following equivalence
. let @xmath271 denote the generalized quaternion group @xmath272 , ( section [ sec : q4n ] ) .
[ prop : prelazaknanovugrupu]following maps jointly exist or do not exist:@xmath273the group @xmath271 acts on @xmath274 as a subgroup and on @xmath256 via the quotient homomorphism @xmath275 ( see appendix ) .
according to propositions [ prop : vezaproblem - ekvivarijantan ] and [ prop : prelazaknanovugrupu ] it is enough to prove that there is no @xmath271-map @xmath276 .
here @xmath263 is the minimal @xmath271(=@xmath260 ) arrangement containing the subspace @xmath277 defined \(a ) for @xmath278 and @xmath279 , by the equations:@xmath280 \(b ) for @xmath281 and @xmath282 , by the equations:@xmath283the codimension of @xmath263 inside @xmath256 in both cases is @xmath1 , so the complement @xmath284 is @xmath285-connected ( by poincar - lefschetz duality ( * ? ? ?
* corollary 8.4 , p.352 ) , ( * ? ? ?
* proposition 3.46 , p.254 ) or the goresky - macpherson formula ) .
therefore the primary obstruction is responsible for the existence of @xmath271-maps @xmath276 , and the obstruction exact sequence has the form@xmath286_{\mathbb{q}_{4n}}\longrightarrow \{\ast \}\overset{\mathfrak{o}_{\mathbb{q}_{4n}}^{3}}{\longrightarrow } \mathfrak{h}_{\mathbb{q}_{4n}}^{3}(s^{3},h_{2}\left ( w_{n}\setminus \bigcup \mathcal{a},\mathbb{z } \right ) ) .\]]since @xmath287 acts on @xmath274 as a subgroup and @xmath288 is connected , the @xmath271-module @xmath102 is trivial .
the equivariant poincar duality isomorphism ( theorem [ th : eqpoincare ] ) implies@xmath289thus , in both cases we are going to define a map in general position , compute the obstruction cocycle and identify its obstruction element inside a group of coinvariants .
all computations are done by the mathematica 5.0 package .
let us define a map @xmath291 on the vertex @xmath292 ( @xmath293 in appendix ) by @xmath294and extend it equivariantly .
for example @xmath295 . for the subspace @xmath277 defined by @xmath296
the arrangement @xmath263 is the minimal @xmath297-arrangement containing @xmath277 .
it has four maximal elements @xmath277 , @xmath298 , @xmath299 and @xmath300 .
this can easily be seen from the set equalities @xmath301 and @xmath302 . the intersection @xmath303 is a codimension @xmath285 subspace of @xmath277 , @xmath298 , @xmath304 , @xmath300 .
thus , the hasse diagram of the intersection poset of the arrangement @xmath263 with the reversed order is as in figure [ fig : fig22 ] .
now we intersect the image under @xmath63 of the maximal cell @xmath305\cup .. \cup \lbrack
\epsilon ^{7}t,\epsilon ^{8}t]\right ) \ast \lbrack jt,\epsilon jt]$ ] with the test space @xmath306 .
the results of the @xmath307 intersections @xmath308\ast \lbrack jt,\epsilon jt]\right ) \cap
\epsilon ^{r}l$ ] are summarized in the following table : @xmath309\ast \lbrack jt,\epsilon jt]\right ) \cap
\epsilon ^{r}l$ } \\
\hline $ \mathbf{i / r}$ & { \small preimage of the intersection point } & { \small intersection point in } $ w_{8}$ & \\
\hline $ 1/2 $ & $ \frac{3}{7}jt+\frac{1}{14}\epsilon jt+\frac{1}{14}\epsilon t+\frac{3}{7}\epsilon ^{2}t$ & $ \left ( -\tfrac{1}{2},\frac{15}{14},-\frac{2}{7},-\frac{2}{7},\frac{15}{14},-\tfrac{1}{2},-\frac{2}{7},-\frac{2}{7}\right ) $ & $ p_{1}$ \\
\hline $ 2/2 $ & $ \frac{20}{51}jt+\frac{23}{153}\epsilon jt+\frac{62}{153}\epsilon ^{2}t+\frac{8}{153}\epsilon ^{3}t$ & $ \left ( -\tfrac{4}{9},\tfrac{16}{17},-\frac{1}{3},-\frac{25}{153},1,-\tfrac{77}{153},-\frac{1}{3},-\tfrac{25}{153}\right ) $ & $ p_{2}$ \\ \hline $ 2/1 $ & $ \frac{8}{153}jt+\frac{63}{153}\epsilon jt+\frac{23}{153}\epsilon ^{2}t+\frac{20}{51}\epsilon ^{3}t$ & $ \left ( 1,-\frac{25}{153},-\frac{1}{3},\frac{16}{17},-\frac{4}{9},-\frac{25}{153},-\frac{1}{3},-\frac{77}{153}\right ) $ & $ p_{3}$ \\
\hline $ 4/1 $ & $ \frac{1}{20}jt+\frac{11}{40}\epsilon jt+\frac{1}{10}\epsilon ^{4}t+\frac{23}{40}\epsilon ^{5}t$ & $ \left ( -\frac{11}{20},\frac{1}{2},-\frac{1}{5},-\frac{3}{2},\frac{6}{5},\frac{1}{2},-\frac{1}{5},\frac{1}{4}\right ) $ & $ p_{4}$ \\
\hline $ 4/0 $ & $ \frac{23}{40}jt+\frac{1}{10}\epsilon jt+\frac{11}{40}\epsilon ^{4}t+\frac{1}{20}\epsilon ^{5}t$ & $ \left ( -\frac{1}{5},\frac{1}{2},-\frac{11}{20},\frac{1}{4},-\frac{1}{5},\frac{1}{2},\frac{6}{5},-\frac{3}{2}\right ) $ & $ p_{5}$ \\ \hline $ 5/0 $ & $ \frac{1}{3}jt+\frac{1}{6}\epsilon jt+\frac{1}{6}\epsilon ^{5}t+\frac{1}{3}\epsilon ^{6}t$ & $ \left ( \frac{1}{3},\frac{1}{3},-\frac{5}{3},1,\frac{1}{3},\frac{1}{3},1,-\frac{5}{3}\right ) $ & $ p_{6}$ \\
\hline $ 7/3 $ & $ \frac{1}{14}jt+\frac{3}{7}\epsilon jt+\frac{3}{7}\epsilon ^{7}t+\frac{1}{14}t$ & $ \left ( \frac{1}{7},-\frac{25}{14},\frac{3}{2},\frac{1}{7},\frac{1}{7},\frac{3}{2},-\frac{25}{14},\frac{1}{7}\right ) $ & $ p_{7}$ \\
\hline \end{tabular}\ ] ] the formula for the obstruction cocycle ( [ obcocycle ] ) implies that there are signs @xmath310 such that@xmath311the point classes in ( [ cocycleproblem1 ] ) do not depend on the embedding of the associated simplices because each intersection point @xmath312 is contained in just one element of the arrangement .
in this situation , in contrast to the lovas conjecture , we can not just apply proposition prop : singularset . to prove that the obstruction element does or does not vanish , the cohomology class of the obstruction cocycle inside @xmath313 has to be computed . however , to prove that the obstruction element does not vanish , we are not compelled to completely identify the obstruction element . with a help of the poincar - alexander duality isomorphism and the universal coefficient isomorphism we have
@xmath314where @xmath315 denotes the one - point compactification of the arrangement @xmath263 .
the calculations of @xmath316 and @xmath317 can be carried out by the ziegler - ivaljevi formula zz .
for example , there is a decomposition @xmath318where @xmath319 is the intersection poset of the arrangement @xmath263 . by convention , @xmath320 .
the decomposition ( [ z - z - decomposition ] ) is not a decomposition of @xmath271-modules .
this fact is illustrated in bl - vr - ziv . if considered with the appropriate field coefficients ( z - z - decomposition )
is a decomposition of @xmath271-modules ( * ? ? ?
* proposition 2.3 ) . in order to use ( [ poincareduality ] ) and ( z - z - decomposition ) and
to compute the coinvariants @xmath313 , we have to keep in mind that poincar - alexander duality isomorphism is not an isomorphism of @xmath271-modules .
fortunately , it is a @xmath271-map up to an orientation character . on the other hand ,
the universal coefficient isomorphism , when the @xmath321 part vanishes , is a @xmath271-map , if the action on @xmath322 is given by@xmath323for @xmath324 , @xmath325 and @xmath326 .
let a modified @xmath271 action @xmath95 on @xmath327 be defined by@xmath328for @xmath324 , @xmath325 and @xmath326 .
if the @xmath321 part vanishes in ( [ poincareduality ] ) and the action on the @xmath322 part is assumed to be @xmath95 , the poincar duality isomorphism @xmath329becomes an isomorphism of @xmath271-modules .
consequently , @xmath330 now we identify every point class from the sum@xmath331(cocycleproblem1 ) using the isomorphism ( [ poincareduality ] ) @xmath332the isomorphism is an evaluation of the linking number ( when it is correctly defined ) . for point classes the following formula holds @xmath333where @xmath334
. finally we must determine whether the class of @xmath335 in the group of coinvariants @xmath336 is or is not zero .
in each case this is a different , and usually a very difficult problem . in the present situation , the exact sequence of @xmath271-modules@xmath337induces an exact sequence of @xmath271-modules @xmath338the geometric interpretation of the map @xmath339 is the computation of the linking numbers .
let us assume that all of the @xmath340 homology groups in the above sum are free .
the left exactness of the coinvariant functor implies that the sequence@xmath341is exact .
the geometric interpretation of the map @xmath342 is the summation of linking .
the map @xmath342 is a good test map for detecting whether an element @xmath343 is not zero .
we use the map @xmath342 ( [ projection ] ) and prove that @xmath344)\neq 0 $ ] .
let us identify groups in question ( [ hompartprojection ] ) and ( [ projection ] ) .
( a)@xmath345 , as abelian groups .
\(b ) @xmath346 .
statement ( a ) follows from the hasse diagram of the arrangement and the ziegler - ivaljevi formula .
statement ( b ) is a consequence of the set equality @xmath301 and the following orientation computation : the element @xmath347 acts on @xmath348 by preserving its orientation . on the orthogonal complement @xmath349 of @xmath277 the operator @xmath347 , for the basis @xmath350 of @xmath349 , has the matrix@xmath351since @xmath352 , the element @xmath347 changes the orientation of @xmath349 and consequently it changes the orientation on @xmath277 .
if @xmath353 is the generator associated with the subspace @xmath277 , then the sum @xmath354 is generated as a free abelian group by elements @xmath228 , @xmath355 , @xmath356 , @xmath357 .
the set equality @xmath301 implies the homology equality @xmath358 .
let @xmath359 @xmath360 be given by @xmath361 and @xmath362 for @xmath363 . then by ( [ eq : modifiedaction ] ) @xmath364and @xmath365for @xmath366 .
thus , there is a relation @xmath367which in coinvariants becomes @xmath368=0 $ ] .
the statement ( b ) follows .
the proof of case ( a ) of theorem [ th : main1 ] is a consequence of the following lemma .
\(a ) the element @xmath369 is given by@xmath370
\(b ) @xmath344)\in \mathrm{hom}\left ( \underset{v\in p:\dim v=4}{\bigoplus } \tilde{h}_{-1}(\delta ( p_{<v});\mathbb{\mathbb{z } } \mathbf{)},\mathbb{\mathbb{z } } \right ) _ { \mathbb{q}_{32}}$ ] is the generator of the group @xmath371 .
statement ( a ) follows from formula ( [ cocycleproblem1 ] ) and the associated table of intersections .
the second statement is a consequence of the geometric interpretation of the map @xmath342 - summation of the intersection numbers .
thus the obstruction element @xmath372 $ ] is not zero , and we have proved case ( a ) of theorem [ th : main1 ] .
since the proof of ( b ) follows the steps of the previous case , we will just outline the computational parts which differ .
let @xmath374 be given by @xmath375 .
the arrangement @xmath263 is now a minimal @xmath376-arrangement containing the subspace @xmath277 defined by@xmath377the arrangement @xmath263 has five maximal elements @xmath277 , @xmath298 , @xmath304 , @xmath300 and @xmath378 .
this follows from the set equality @xmath379 .
let us determine the intersection of the image under @xmath63 of the maximal cell@xmath380\cup \lbrack \epsilon t,\epsilon ^{2}t]\cup \lbrack \epsilon ^{2}t,\epsilon ^{3}t]\cup ~ ... ~\cup \lbrack
\epsilon ^{8}t,\epsilon ^{9}t]\right ) \ast \lbrack jt,\epsilon jt]\]]with the test space @xmath381 .
the results of @xmath382 intersections @xmath308\ast \lbrack jt,\epsilon jt]\right ) \cap \epsilon ^{r}l$ ] can be summarized in the following way:@xmath383\ast \lbrack jt,\epsilon jt]\right ) \cap
\epsilon ^{2}l=\{q_{1}\ } , & & f\left ( [ \epsilon ^{5}t,\epsilon ^{6}t]\ast \lbrack jt,\epsilon jt]\right ) \cap
\epsilon ^{3}l=\{q_{2}\ } , \\ f\left ( [ \epsilon ^{5}t,\epsilon ^{6}t]\ast \lbrack jt,\epsilon jt]\right ) \cap \epsilon ^{4}l=\{q_{3}\ } & & \end{array}\]]and consequently @xmath384
. there are @xmath385 such that@xmath386 as in the previous case we use the map @xmath342 ( projection ) and prove that @xmath387)\neq 0 $ ] .
the set equality @xmath379 implies that@xmath388the element @xmath389 acts on @xmath390 by changing its orientation . on the orthogonal complement @xmath349 of @xmath277 the operator @xmath389 , for the basis @xmath391 of @xmath349 , has the matrix@xmath351since @xmath352 , the element @xmath389 changes the orientation of @xmath349 and consequently does not change the orientation on @xmath277 . let @xmath392 be the generator associated to the subspace @xmath277 , then the sum @xmath393 is generated as a free abelian group by elements @xmath228 , @xmath355 , @xmath356 , @xmath357 , @xmath394 .
the set equality @xmath379 implies the homology equality @xmath395 .
let @xmath396 @xmath360 be defined by @xmath361 and @xmath362 for @xmath397 .
then @xmath398and @xmath399for @xmath400 .
thus , there is a relation @xmath401which implies the isomorphism ( [ iso-1 ] ) we indicated . as before @xmath402)$ ] is a generator of the group @xmath182 .
therefore case ( b ) of theorem [ th : main1 ] is proved .
we state the following conjecture .
all symmetric six - tuples @xmath403 are solutions of problem problem1 .
the problem discussed in this chapter will be managed by an extension of obstruction theory methods used in the paper @xcite .
our result , although obtained along the lines of the same method , differs from the result of the paper @xcite at the most critical phase of the proof .
the search for the target extension space and the identification of the obstruction element is only remotely similar .
the appendix contains the ancillary material referred to in this section .
let @xmath404 be proper borel probability measures on @xmath0 .
let @xmath405 be a vector where @xmath406 .
the general problem stated in @xcite is : determine all triples @xmath407 such that for any collection of @xmath408 measures @xmath409 , there exists a @xmath44-fan @xmath231 with the property@xmath410such a @xmath44-fan @xmath231 is called an @xmath411__-partition _ _ for the collection of measures @xmath412 . in this chapter
we prove the existence of a class of solutions of the above problem in the case @xmath413 / @xmath414 .
[ th : main2]let us choose @xmath415 such that @xmath416 . then any two _ proper _ borel probability measures @xmath211 and @xmath417 on the sphere @xmath0 admit an @xmath411-partition by a @xmath1-fan @xmath418 .
the cs / tm setting is the same as in the previous chapter except for the test space .
therefore we give only an outline . for details
one can also consult @xcite , @xcite and @xcite . * the configuration space .
* let @xmath211 and @xmath417 be two proper borel probability measures on @xmath0 .
the configuration space @xmath419 associated with the measure @xmath211 is @xmath420as before , @xmath419 is a stiefel manifold @xmath421 of all orthonormal @xmath3-frames in @xmath219
. * test maps .
* let @xmath411-vectors have the form @xmath422 , where @xmath423 .
the test map for this fan problem is defined by@xmath424*the action . *
the dihedral group @xmath255 acts both on the possible solution space @xmath419 and the linear subspace @xmath425 , as in the previous chapter .
observe that @xmath419 is @xmath260-homeomorphic to the manifold @xmath421 , where @xmath421 is a @xmath260-space given by @xmath426and @xmath427 is the rotation around the axes determined by @xmath141 through the angle @xmath428 . * the test space .
* given @xmath429 , let @xmath430 be defined by the equations@xmath431the test space for this problem is the union @xmath432 of the smallest @xmath260-invariant linear subspace arrangement @xmath433 , containing @xmath434 . since the test map @xmath435 is @xmath260-equivariant , the following proposition holds .
[ prop : fansvezaproblem - ekvivarijantan]let @xmath429 be a vector such that @xmath423 .
if there is no @xmath260-equivariant map@xmath436then for any two measures @xmath211 and @xmath417 on @xmath0 , there exists an @xmath411-partition @xmath437 of measures @xmath211 and @xmath417 .
as we have seen in proposition [ prop : prelazaknanovugrupu ] and in @xcite and @xcite the stiefel manifold @xmath421 can be substituted with the sphere @xmath274 .
thus we prove that there is no @xmath271-map@xmath438 theorem [ th : main2 ] will be proved by showing that there is no @xmath271-map @xmath439 . the arrangement @xmath440 is a minimal @xmath260-invariant arrangement containing the subspace @xmath264 given by @xmath441the codimension of the arrangement @xmath440 inside @xmath256 is two , and so the complement is connected .
the fundamental group of the complement is far from trivial , so the obstruction theory can not be applied directly .
we use the idea of a target extension scheme introduced in bl .
let us denote by @xmath24 the complement @xmath442 .
the basic idea of the target extension scheme is to find a @xmath271-space @xmath443 which contains the target space @xmath24 , and to prove that there is no @xmath271-map _ _ _ _ @xmath444 .
this would imply that there is no @xmath271-map _ _ _ _ @xmath445 . in our case
the target space @xmath24 is the complement of the arrangement , and so we can refine the idea as follows : * _ _ increase the codimension . _
_ let @xmath75 be an arbitrary hyperplane in @xmath256 and let @xmath446 be the minimal @xmath271-invariant arrangement containing the subspace @xmath447 .
the inclusion @xmath448 implies that @xmath449the dimension of maximal elements of the arrangement @xmath446 is @xmath450 .
let us denote by @xmath443 the `` new '' complement @xmath451 .
* _ apply the general position map scheme . _
the codimension of the arrangement @xmath446 inside @xmath256 is three and so the complement @xmath452 is 1-connected .
therefore the question of the existence of a @xmath271-map _ _ _ _ @xmath444 depends only on the primary obstruction .
we can now use the general position map scheme .
unfortunately , the target extension scheme only provides hope that we can apply , once more , the mechanisms we already developed .
the main problem is just shifted to the question of _ how to find a hyperplane _ @xmath453 in a way that there is no @xmath271-map @xmath444 .
there are two rough heuristics we can use .
first , if we define a @xmath271-map _ _ _ _ @xmath454 , we can introduce @xmath75 in such a way that \(a ) @xmath63 becomes a map in general position for the new arrangement @xmath446 , and \(b ) the cardinality of the set @xmath455 is as small as can be achieved .
the second requirement for introduction of the hyperplane is that the group of coinvariants @xmath456 has torsion .
nevertheless , the choice of the hyperplane @xmath75 is strongly connected with the properties of the action of the group on the arrangement @xmath440 . in this situation
we are going to exploit the following symmetry @xmath457 .
observe that element @xmath458 _ changes _ the orientation of the orthogonal complement @xmath459 .
let us define a @xmath271-map _ _ _ _ @xmath454 and introduce @xmath75 so that @xmath63 becomes a map in general position .
let @xmath274 be the simplicial complex @xmath460 , where @xmath461 is the regular @xmath462-gon .
let @xmath292 be a fixed vertex in one of the copies of @xmath461 .
let @xmath463 , @xmath464 , where @xmath465 are elements of the standard basis of @xmath466 .
we define @xmath454 on the vertex @xmath467 and then extend it equivariantly .
this implies @xmath468 in the future all the indexes in @xmath256 will be calculated @xmath469 .
every 3-simplex of the sphere @xmath460 , determined by four vertices , has form @xmath470\]]and will be denoted by @xmath471 .
the list of all simplices of the form @xmath472 $ ] which intersect the subspace @xmath277 is given in the following table @xmath473 $ & $ \frac{1}{n}{\small ( a,\beta , \gamma , a);~\beta + \gamma = b}$ & \\
\hline { \small 2 . \ } $ [ u_{a},u_{a+1};u_{r},u_{r+1}]$ & $ \frac{1}{n}{\small ( a , b,\gamma , \delta ) ; ~\gamma + \delta = a}$ & $ { \small [ a+b+1,n-3]}$ \\ \hline { \small 3 .
\ } $ [ u_{a},u_{a+1};u_{n-2},u_{n-1}]$ & $ \frac{1}{n}{\small ( a , b,\gamma , \delta ) ; ~\gamma + \delta = a}$ & \\
\hline { \small 4 . \ } $ [ u_{a},u_{a+1};u_{n-1},u_{n}]$ & $ \frac{1}{n}{\small ( a , b,\gamma , \delta ) ; ~\gamma + \delta
= a}$ & \\
\hline { \small 5 . \ } $ [ u_{r},u_{r+1};u_{a+b},u_{a+b+1}]$ & $ \frac{1}{n}{\small ( \alpha , \beta , b ,
a);~\alpha + \beta = a}$ & $ { \small [ 2,a-2]}$ \\ \hline { \small 6 . \ } $ [ u_{1},u_{2};u_{a+b},u_{a+b+1}]$ & $ \frac{1}{n}{\small ( \alpha , \beta , b ,
a);~\alpha + \beta = a}$ & \\
\hline { \small 7 . \ } $ [ u_{n},u_{1};u_{a},u_{a+1}]$ & $ \frac{1}{n}{\small ( \alpha , \beta , b ,
a);~\alpha + \beta = a}$ & \\
\hline { \small 8 . \ } $ [ u_{n},u_{1};u_{a+b},u_{a+b+1}]$ & $ \frac{1}{n}{\small ( a , b,\gamma , \delta ) ; ~\gamma + \delta = a}$ & \\
\hline { \small 9 . \ } $ [ u_{n},u_{1};u_{r},u_{r+1}]$ & $ \frac{1}{n}{\small ( a,\beta , \gamma , a);~\beta + \gamma = b}$ & $ { \small [ a+1,a+b-2]}$ \\ \hline { \small 10 . } $ [ u_{n},u_{1};u_{a+b-1},u_{a+b}]$ & $ \frac{1}{n}{\small ( a,\beta , \gamma , a);~\beta + \gamma = b}$ & \\
\hline \end{tabular}\ ] ] if we introduce the hyperplane @xmath75 by @xmath474the list of simplices now intersecting @xmath475 reduces to \(a ) two simplices , for @xmath476 , @xmath477\text { and } \rho _ { 2}=[u_{n},u_{1};u_{a},u_{a+1}],\]]with barycentric coordinates of intersection points @xmath478 and @xmath479 \(b ) one simplex , for @xmath480 , @xmath481 $ ] , with barycentric coordinates of intersection points @xmath478 . in order to simplify the exposition let us assume that @xmath480 .
case ( a ) is actually a consequence of case ( b ) .
therefore,@xmath482 we have made extensive use of the first heuristic .
in is not easy to see that we also used the second heuristic .
let us just say that the following property will be a significant ingredient in the interpretation of the obstruction element.@xmath483 there are 8 simplices in @xmath460 which are in the inverse image under @xmath63 of the simplex @xmath484 $ ] .
these simplices are@xmath485all eight simplices are in the orbit of the following two simplices from the maximal cell @xmath59 ( consult appendix [ sec : simplicial]):@xmath486the points @xmath487 and @xmath488 , with barycentric coordinates @xmath489 and @xmath490 , are mapped inside the arrangement @xmath446 , precisely@xmath491therefore , formula ( [ obcocycle ] ) implies that the obstruction cocycle is @xmath492since we are interested in the cohomology / coinvariant class of the obstruction cocycle ( [ coinvariantisomorphism ] ) , instead of @xmath493 we can look at the element @xmath494 in contrast to the previous chapter , the point class @xmath495 * depends * on the placement of the simplex @xmath496 $ ] with respect to the arrangement @xmath446 . in terms of paper @xcite , @xmath497 is the broken point class .
it remains to prove that @xmath498 does not vanish when we pass to coinvariants .
the brief proof of this fact would be : there is a symmetry@xmath499=[u_{a},u_{a+1};u_{a+b},u_{a+b+1}]\]]which does not change the orientation of the simplex and therefore gives the relation @xmath500 and not @xmath501 .
this is the only relation which could provide in coinvariants that @xmath502 @xmath503 .
thus @xmath498 does not vanish in coinvariants and the obstruction element is not zero .
we proved theorem [ th : main2 ] for all rational triples @xmath2 .
just to be safe , let us analyze the point class @xmath504 , its placement with the respect to the arrangement @xmath446 and the indicated @xmath505 symmetry .
observe that we are not forced to determine the complete group of coinvariants @xmath506 .
we only have to prove that @xmath502 @xmath507 . translating the simplex
@xmath484 $ ] by a small generic vector and then intersecting it with the arrangement , we find that the boundary ( of the shaded region ) links with the elements of the arrangement as in the figure [ fig : fig2 ] .
figure [ fig : fig2 ] indicates two set relations @xmath508 and @xmath509 . since element @xmath505 does not change the orientation of the subspace @xmath510 , the geometric generators of the group @xmath511 ( the dual of @xmath512 ) can only be given by sets@xmath513this fact is a consequence of the following observations * the orientation on the arrangement should be consistent with the group action , * the boundaries of the two pieces which are glued together have opposite orientations , * the element @xmath514 is not a generator because it is a linear combination of already introduced generators .
the isomorphism composition of the poincar duality isomorphism and the universal coefficient isomorphism for the arrangement @xmath446 , @xmath515is the computation of the linking number with the elements of the basis ( homologybasis ) .
thus,@xmath516where @xmath517 @xmath518 are elements of a basis determined by sets @xmath519 and @xmath520 . since @xmath521 , the class @xmath522 $ ] does not vanish .
the obstruction element is not zero and we have proved theorem [ th : main2 ] for all rational triples @xmath2 .
let @xmath523 be the space of all triples @xmath524 , @xmath525 such that for @xmath526 there exists an @xmath411-partition of measures @xmath211 and @xmath417 by a @xmath1-fan . since we assumed that our measures @xmath211 and @xmath417 are two proper borel probability measures
, the space @xmath527 is a closed subset of @xmath528 . in this notation , so far we have proved the inclusion @xmath529therefore @xmath530where @xmath531 denotes the closure of the subset in @xmath532 . the theorem is proved .
the sphere @xmath533 can be seen as the group of all unit quaternions .
consider a root of unity @xmath534 .
the group generated by @xmath535 is a subgroup of @xmath536 of order @xmath462 .
the generalized quaternion group _ _ , _ _ @xcite _ _ , _ _ is a subgroup of order @xmath537 generated by @xmath535 and @xmath505 .
the group @xmath271 acts on @xmath274 as a subgroup , and on @xmath256 via the already defined @xmath260-action by the quotient homomorphism @xmath538 .
the @xmath271-action on @xmath274 is * free*. [ [ secsimplicialthe - natural - mathbbq_4n - cellular - simplicial - structure - on - s3 . ] ] [ sec : simplicial]the natural @xmath271 cellular / simplicial structure on @xmath274 . ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
let the circle @xmath544 be represented by the simplicial complex of the regular @xmath462-gon @xmath461 .
then the sphere @xmath274 , as a simplicial complex , is the join @xmath545 of two copies of @xmath461 .
denote the vertices of @xmath546 and @xmath547 by @xmath548 and @xmath549 , respectively .
the action of the group @xmath550 on @xmath274 is defined on vertices by@xmath551and it extends equivariantly to the upper skeletons .
the associated chain complex @xmath552 has the form@xmath553 it comes from the minimal resolution of @xmath554 by free @xmath271-modules described in @xcite .
the associated cellular complex has one @xmath271 @xmath79-cell @xmath555 , two @xmath271 @xmath285-cells @xmath556 and @xmath557 , two @xmath271 @xmath3-cells @xmath558 and @xmath559 , and then finally one @xmath271 @xmath1-cell @xmath59 .
the associated chain complex @xmath560 has the form @xmath561e\overset{\partial } { \rightarrow } \mathbb{\mathbb{z } } [ \mathbb{q}_{4n}]c\oplus \mathbb{\mathbb{z } } [ \mathbb{q}_{4n}]c^{\prime } \overset{\partial } { \rightarrow } \mathbb{\mathbb{z } } [ \mathbb{q}_{4n}]b\oplus \mathbb{\mathbb{z } } [ \mathbb{q}_{4n}]b^{\prime } \overset{\partial } { \rightarrow } \mathbb{\mathbb{z } } [ \mathbb{q}_{4n}]e\rightarrow 0\]]where @xmath562thus , it will be enough to look at the obstruction cocycle @xmath563 on the maximal cell @xmath59 , and to prove that its image is or is not zero , when we pass to the cohomology .
there exists a cellular map @xmath565 .
since we are interested in @xmath1-cochains we only need to know the concrete expression for the cellular map on the top dimensional cell @xmath59,@xmath566\ast \lbrack b_{1},\epsilon b_{1}]+[\epsilon a_{1},\epsilon ^{2}a_{1}]\ast \lbrack b_{1},\epsilon
b_{1}]+ ... +[\epsilon ^{n-1}a_{1},\epsilon ^{n}a_{1}]\ast \lbrack b_{1},\epsilon b_{1}]\text{,}\]]where the simplices on the right hand side are appropriately oriented . | a significant group of problems coming from the realm of combinatorial geometry can be approached fruitfully through the use of algebraic topology . from the first such application to kneser s problem in 1978 by lovsz @xcite through the solution of the lovsz conjecture @xcite , @xcite , many methods from algebraic topology have been developed .
specifically , it appears that the understanding of equivariant theories is of the most importance .
the solution of many problems depends on the existence of an elegantly constructed equivariant map .
for example , the following problems were approached by discussing the existence of appropriate equivariant maps .
a variety of results from algebraic topology were applied in solving these problems .
the methods used ranged from well known theorems like borsuk - ulam and dold theorem to the integer / ideal - valued index theories .
recently equivariant obstruction theory has provided answers where the previous methods failed .
for example , in papers @xcite and @xcite obstruction theory was used to prove the existence of different mass partitions . in this paper
we extract the essence of the equivariant obstruction theory in order to obtain an effective _ general position map _
scheme for analyzing the problem of existence of equivariant maps .
the fact that this scheme is useful is demonstrated in this paper with three applications : * ( a ) * a `` half - page '' proof of the lovsz conjecture due to babson and kozlov @xcite ( one of two key ingredients is carsten s map @xcite ) , * ( b ) * a generalization of the result of v. makeev @xcite about the sphere @xmath0 measure partition by @xmath1-planes ( section [ result1 ] ) , and * ( c ) * the new @xmath2 , class of @xmath1-fan @xmath3-measures partitions ( section [ result2 ] ) .
these three results , sorted by complexity , share the spirit of analyzing equivariant maps from spheres to complements of arrangements of subspaces . |
neutrino oscillations experiments search for a distortion in the neutrino flux at the detector positioned far away from the source . by comparing both near and far neutrino energy spectra ,
one gains information about the oscillation probability and with that about mixing angles and mass squared differences . however , the neutrino energy , that enters critically into the oscillation probability , is not directly measurable but has to be reconstructed from the reaction products .
a proper understanding of neutrino - nucleus interactions is , therefore , essential for the interpretation of current neutrino oscillations experiments .
present @xmath5 disappearance experiments use the ccqe reaction both as signal event and to reconstruct the neutrino energy from the outgoing muon .
ccqe is defined as @xmath6 on a single nucleon ; in the nucleus , ccqe is masked by final - state interactions ( fsi ) .
thus , the correct identification of ccqe events is directly related to the question of how fsi influence the event selection .
the main background to ccqe is cc@xmath7 production . if the pion is absorbed in the nucleus and/or not seen in the detector , these events can be misidentified as ccqe . consequently , a proper understanding of ccqe _ and _ cc@xmath7 is essential for the reconstruction of the neutrino energy . the main task in a @xmath8 appearance experiment like miniboone is to detect electron neutrinos in a ( almost ) pure @xmath5 beam .
the signal event , the @xmath8 ccqe interaction , is dominated by background .
a major problem comes from misidentified events , mainly because of the fact , that the miniboone detector can not distinguish between a photon and an electron .
thus , @xmath5 induced neutral current @xmath9 production , where the @xmath9 decays into two @xmath10s , is the major source of background when one of the photons is not seen or both cherenkov rings overlap .
as all of the present oscillations experiments use nuclear targets , it is mandatory to consider fsi , i.e. , pion rescattering , with and without charge exchange , and absorption in the nuclear medium .
a realistic treatment of the final state interactions ( fsi ) can be achieved in the framework of a coupled - channel transport theory the gibuu model .
after a brief review of our model , we first discuss the impact of pion production on ccqe measurements .
we further investigate the influence of nuclear effects on cc@xmath11 and nc@xmath12 cross sections , and , where possible , we confront our model to recent data measured at miniboone . in the second part
we extend the description to higher energies using duality arguments .
we treat neutrino - nucleus scattering as a two - step process . in the initial - state step , the neutrinos
interact with nucleons embedded in the nuclear medium . in the final - state step ,
the outgoing particles of the initial reaction are propagated through the nucleus using a hadronic transport approach . in the energy region
relevant for miniboone , sciboone and k2k , the elementary @xmath13 reaction is dominated by two processes : quasielastic scattering and the excitation of the @xmath1 resonance ( p@xmath14(1232 ) ) . additionally , our model includes the excitation of 12 @xmath15 and @xmath1 resonances with invariant masses less than 2 gev and also a non - resonant single - pion background .
details are given in @xcite .
the excitation of the resonances ( r ) is described within the isobar model with the help of the nucleon , @xmath16 and @xmath17 form factors .
vector form factors are derived from maid helicity amplitudes @xcite extracted from electron scattering experiments .
cross section as a function of the neutrino energy compared to the pion production data of of anl ( refs .
@xcite ( @xmath18 ) , @xcite ( @xmath19 ) ) and bnl ( @xcite ( @xmath20 ) ) .
the solid line has been obtained with a form factor fitted to the anl data , the dashed one is fitted to the bnl data . ] experimental information on the n - r axial form factors is very limited .
goldberger - treiman relations have been derived for the axial couplings , but there is no information about the @xmath21 dependence .
we apply pcac and pion pole dominance to derive one of the axial couplings for each resonance and to relate it to the pseudoscalar coupling .
the @xmath21 dependence of the @xmath1 axial form factors are fitted to either anl or bnl bubble chamber neutrino - scattering @xmath22 data for the @xmath23 reaction .
[ fig : delta_sigmatot ] shows the integrated cross section together with the data .
we note already here , that the solid curve fits the anl data , while the dashed curve fits the bnl data .
thus , the latter would obviously lead to higher pion production cross section also on the nucleus .
the single-@xmath24 non - resonant background cross section @xmath25 includes vector , axial and also interference contributions @xmath26 the vector part is fully determined by electron scattering data , as described in @xcite .
the axial and the interference term collected under the label `` non - v '' are only present in neutrino scattering and are fitted to the available neutrino data for both , @xmath27 and @xmath28 .
the final results are shown in fig .
[ fig : xsec_pionprod_wbg ] . ) and @xcite ( @xmath19 ) ) and bnl ( @xcite ( @xmath20 ) ) .
the solid lines denote the our full result including the non - resonant background .
furthermore , we show the results for pion production only through the excitation and the subsequent decay of all resonances ( dashed lines ) or through the @xmath1 alone ( dash - dotted lines ) .
note that the anl based fit for the @xmath1 axial form factor has been used here . ] on at @xmath29 gev .
the solid lines are obtained using the tracking detector identification , while the dashed lines are for cherenkov detectors .
[ fig : ccqelike_over_trueccqe_thresholds],title="fig : " ] on at @xmath29 gev .
the solid lines are obtained using the tracking detector identification , while the dashed lines are for cherenkov detectors .
[ fig : ccqelike_over_trueccqe_thresholds],title="fig : " ] the neutrino - nucleon cross sections are modified in the nuclear medium . bound nucleons are treated within a local thomas - fermi approximation .
they are exposed to a mean - field potential which depends on density and momentum .
we account for this by evaluating the above cross sections with full in - medium kinematics , i.e. , hadronic tensor and phase - space factors are evaluated with in - medium four - vectors .
we also take pauli blocking and collisional broadening of the outgoing hadrons into account . our model for neutrino-(bound)nucleon scattering
is described in detail in @xcite .
, title="fig : " ] , title="fig : " ] after the initial neutrino - nucleon interaction , the produced particles propagate through and out of the nucleus . during propagation
they undergo fsi which are simulated with the coupled - channel semi - classical gibuu transport model ( for details , see ref .
@xcite and references given there ) .
it is based on the buu equation which describes the space - time evolution of a many - particle system in a mean - field potential including a collision term .
nucleons and resonances acquire medium - modified spectral functions and are propagated off - shell .
herby we ensure , that vacuum spectral functions are recovered after leaving the nucleus .
the collision term accounts for changes ( gain and loss ) in the phase - space density due to elastic and inelastic collisions between the particles , and also to particle decays into other hadrons .
baryon - meson two - body interactions ( e.g. , @xmath30 ) are described by resonance contributions and a small non - resonant background term ; baryon - baryon cross sections ( e.g. , @xmath31 , @xmath32 , @xmath33 , @xmath34 ) are either fitted to data or calculated , e.g. , in pion exchange models .
the three - body channels @xmath35 and @xmath36 are also included .
the buu equations for all particle species are thus coupled through the collision term and also through the potentials .
such a coupled - channel treatment is required to account for side feeding into different channels .
the treatment of pion final state interactions in gibuu has been widely tested both with @xmath37 and @xmath38 data @xcite .
the latter reaction is quite similar to the @xmath39 reaction in that the incoming particle interacts with all target nucleons and the vector couplings are the same .
one challenge is to identify _ true _ ccqe events in the detector , i.e. , muons originating from an initial qe process .
the difficulty comes from the fact that the true ccqe events are masked by fsi in a detector built from nuclei . in general , at cherenkov detectors such as miniboone ccqe - like events are all those where no pion is detected while in tracking detectors such as k2k - scibar / scifi ccqe - like events are those where a single proton track is visible and at the same time no pions are detected . in both detector types the fsi lead to misidentified events , e.g. , an initial @xmath1 whose decay pion is absorbed or which undergoes `` pion - less decay '' can count as ccqe event we call this type of background events `` fake ccqe '' events .
we denote every event which looks like a ccqe event by `` ccqe - like '' . to investigate the relation between the ccqe - like and true ccqe cross section , we show their ratio as a function of proton and pion momentum thresholds in fig .
[ fig : ccqelike_over_trueccqe_thresholds ] .
as the proton is not at all relevant for the ccqe identification in cherenkov detectors , the ratio is independent of the proton momentum detection threshold ( dashed line in left panel ) .
this is very different in tracking detectors which rely on the detected proton here the efficiency is reduced to @xmath4010% at a proton momentum threshold of 0.5 gev ( solid line in left panel ) .
even at @xmath41 the efficiency does not exceed 80% because of charge - exchange processes that lead to the emission of undetected neutrons and because of secondary proton knockout that leads to multiple - proton tracks .
focussing on the right panel of fig .
[ fig : ccqelike_over_trueccqe_thresholds ] we find that the ccqe - like cross section increases for both detector types as @xmath42 increases . in this case
even more events with pions in the final state appear as ccqe - like because these pions are below threshold and thus not detected .
fixing the flux normalization with harp s pion - production data , the miniboone collaboration has presented their first , preliminary absolutely normalized total , differential and double differential cross sections for ccqe and finds an excess of about 35% compared to the total cross section measured at nomad , anl and bnl @xcite .
we emphasize that these absolute cross sections depend directly on the pion background subtraction which again is based on the monte carlo prediction ( cf .
@xcite ) . in fig .
[ fig : miniboone_doublediff_qe ] , we show our prediction for the double differential cross section at miniboone in muon observables , all calculated with @xmath43 gev .
the left panel shows the true ccqe events . to emphasize the role of `` fake '' ccqe events ,
we show the ratio ccqe - like / true ccqe in the right panel .
this ratio shows that fakes contribute mainly at high energy transfers ( low @xmath44 ) and forward angles . unlike for monochromatic beams , the qe and @xmath1 peaks
are not distinguishable any more but strongly overlap .
this fact makes a model - independent cut based on muon variables to subtract the background impossible .
in fig . [ fig : mb_cc ] we give our results for the single-@xmath45/qe ratio for cc interactions on mineral oil ch@xmath46 .
the solid lines denote the cc@xmath11-like / ccqe - like result , the dashed lines stand for the true cc@xmath11/true ccqe result , and the dash - dotted lines give the vacuum expectation . note that we have applied the cherenkov detector identification criteria .
we emphasize that nuclear corrections cancel out in the ratio , only as long as fsi are not considered ( `` true '' vs. `` free '' ) . in general , the complexity of fsi prevent such cancellations as one can infer from the result denoted with `` like '' which does not coincide with the `` true '' and `` free '' ones . /qe
cross section ratio for cc interactions vs. neutrino energy on ch@xmath46 together with recent data from miniboone @xcite ( upper data set : corrected for fsi , lower data set : uncorrected for fsi ) .
the solid lines denote the cc@xmath11-like / ccqe - like result ( cherenkov detector definitions ) , the dashed lines stand for the true cc@xmath11/true ccqe result , and the dash - dotted lines give the vacuum expectation , i.e. , the sum of the nucleon cross sections ( with two additional protons in the miniboone case ) .
[ fig : mb_cc ] ] production on as a function of the pion kinetic energy averaged over the miniboone flux .
right panel : same on averaged over the k2k flux .
the dashed lines show our calculation without fsi or spectral functions , both included in the full calculation denoted with the solid lines .
the dash - dotted lines indicate the contribution from the @xmath1 resonance to the full calculation .
[ fig : miniboone_k2k_nc],title="fig : " ] production on as a function of the pion kinetic energy averaged over the miniboone flux .
right panel : same on averaged over the k2k flux .
the dashed lines show our calculation without fsi or spectral functions , both included in the full calculation denoted with the solid lines .
the dash - dotted lines indicate the contribution from the @xmath1 resonance to the full calculation .
[ fig : miniboone_k2k_nc],title="fig : " ] we further compare to very recent miniboone data @xcite ( fig .
[ fig : mb_cc ] ) .
let us first focus on the data denoted with the triangles ( upper data set ) .
these are corrected for fsi using a specific monte carlo generator , i.e. , they give the cross sections for bound nucleons `` before fsi '' .
as this procedure introduces a model dependence in the data , a fully consistent comparison is not possible . ignoring this inconsistency ,
our calculation denoted with `` true '' should be the one to compare with .
the agreement is perfect for energies up to 1.5 gev , and still within their error bars for higher @xmath47 .
the miniboone data denoted with bullets ( lower data set ) is their result for the ratio of cc@xmath11-like to ccqe - like .
as these data are not corrected for fsi within a specific monte carlo event generator , this observable is less model dependent .
still , the energy reconstruction requires specific assumptions as well as the detector simulation .
we find that our calculation clearly underestimates the uncorrected data .
however , the perfect agreement with their corrected distribution indicates a significant difference between the pion absorption models .
the underestimate of the pion / quasielastic ratio in particular at higher energies could be due to , among other possibilities , an underestimate of the pion production cross section or an overestimate of the ccqe - like cross section .
both depend directly on the input at the nucleon level , i.e. , in particular on the axial form factor @xmath48 of the @xmath1 resonance ( see fig . [
fig : delta_sigmatot ] ) . for the results presented here , we have used the anl data as a reference .
however , we have shown in ref .
@xcite that an increase of the total pion production cross section on the nucleon compatible with the bnl data also seems to be insufficient to describe this ratio at all energies .
we note that a similar result for the ratio has been recently obtained by athar _
et al . _
in the left panel of fig . [ fig : miniboone_k2k_nc ] , we show our results for nc single-@xmath9 production off as a function of the pion kinetic energy . we have averaged over the miniboone energy flux which peaks at about 0.7 gev neutrino energy @xcite . in nc reactions the total pion yield is dominated by @xmath9 production , while @xmath45 dominate in cc processes ( for details , see ref .
@xcite and references therein ) . comparing the dashed with the solid line ( results without fsi and spectral function vs. full calculation ) ,
one finds a considerable change .
the shape is caused by the energy dependence of the pion absorption and rescattering cross sections .
pions are mainly absorbed via the @xmath1 resonance , i.e , through @xmath49 followed by @xmath50 .
this explains the reduction in the region around @xmath51 gev .
pion elastic scattering @xmath52 reshuffles the pions to lower momenta and leads also to charge exchange scattering into the charged pion channels .
the vast majority of the pions comes from initial @xmath1 excitation ( dash - dotted line ) , their production in the rescattering of nucleons is not significant at these energies .
production on ch@xmath46 as a function of the pion momentum averaged over the miniboone flux .
the dashed and the solid lines denote the calculation with a modified dipole form factor ( fitted to anl data ) for the @xmath1 resonance ( identical to the result in the left panel of fig . [
fig : miniboone_k2k_nc ] ) .
the dash - dotted and dotted lines are obtained with a dipole form for the form factor ( fitted to the bnl data ) .
the miniboone data are taken from ref .
[ fig : miniboone_nc_ff ] ] the right panel of fig .
[ fig : miniboone_k2k_nc ] shows the results for nc single-@xmath9 production off averaged over the k2k energy flux which peaks at about 1.2 gev neutrino energy @xcite .
compared to the left panel , the spectrum is broader and extends to larger @xmath53 due to the higher neutrino energy .
again , pion production through initial qe scattering is not sizable . in fig .
[ fig : miniboone_nc_ff ] we show also the latest nc@xmath9 differential cross sections from miniboone @xcite .
it can be seen that the full line , depicting the results of our calculation after fsi with the fit of the elementary cross section to the anl data lies nearly a factor 2 below the data .
the dashed line gives the result of the same calculation before fsi , the dash - dotted line shows the same quantity , but using a fit to the bnl data for the elementary cross section .
we find it remarkable that these two curves ( without fsi ) are comparable to the data ( with fsi ) , even though pion fsi have a major effect on the spectra ( see fig .
[ fig : miniboone_k2k_nc ] ) .
the absolute height of the experimental cross section is thus hard to reconcile with present descriptions of the elementary cross section and with what we know from other reactions about the pion fsi .
at higher energies that correspond to invariant masses of the @xmath54 system @xmath55 the dis channel becomes essential .
a connection between the asymptotic dis region and the resonance dominated mass range is provided by the so - called bloom - gilman quark - hadron duality .
nearly forty years ago , bloom and gilman found @xcite that in electron scattering on protons the inclusive structure function @xmath3 in the resonance region oscillates around the dis scaling curve and , after averaging , closely resembles it .
furthermore , the resonance region data `` slide '' along the dis curve with increasing @xmath21 .
if duality is understood quantitatively , there may be various applications .
for example , the region of high bjorken variable @xmath56 is hardly experimentally investigated , because in the dis region it would require very high @xmath21 and thus huge luminosities .
if duality is satisfied with good accuracy , one would be able to use the data in the resonance region to reach high @xmath56 at reasonable @xmath21 .
the topic becomes even more interesting when turning to nuclear targets and neutrino sources .
the current precision measurements of the oscillation parameters require an efficient and accurate description of the neutrino
nucleus cross sections . of particular interest
is the resonance region and the possibility of linking it with the dis region .
a hadronic description of a neutrino - nucleus cross section at low @xmath21 requires a good knowledge of vector and axial transition form factors for each resonance . for the majority of the resonances ,
these transition form factors are not well constrained .
provided that one can establish that quark - hadron duality holds with a reasonable accuracy , one could think of using the dis results for estimating the neutrino - nucleus cross sections in the transition region .
so far , most theoretical studies of quark hadron duality in lepton scattering were dealing with nucleon targets @xcite .
the dis parts were considered as known , the structure functions in the scaling region being conventionally evaluated from leading twist ( lt ) parton distribution functions ( pdf ) : for example , @xmath57 , @xmath58 . for nucleons , several parametrizations of the pdfs
are generally available ( from the grv , cteq and mrst groups ) . in the region of moderate @xmath56 , which is of interest for our duality study ,
they provide nearly the same results .
the studies of the resonance region differ in the way the models treat the resonant contributions and the way they extract the structure functions .
the fact , that the resonance data for increasing @xmath59 `` slide '' along the dis curve , can be observed if a few @xcite or even only one @xcite resonance are taken into account .
the advantage of the model @xcite is that the structure functions are given as simple analytical functions of the momentum transfer squared @xmath21 and the energy transfer @xmath60 , provided that the resonance form factors are known . in this work
the first four resonances were considered .
generally , however , as it was argued by close @xcite , inclusion of several resonances of different parities is desirable .
a quark model for resonance excitation has been applied by @xcite for the investigation of the duality in electro- and neutrinoproduction . within the gibuu framework ,
as it was mentioned above , 13 resonances can be considered for both electron and neutrino reactions .
the cross section is calculated numerically and the structure function @xmath3 is extracted from the cross section in a convenient way : @xmath61 \displaystyle \left [ 1-\frac{q^2}{4ee'}+2\frac{q^2}{4ee'}\frac{q^2 \nu^2 + q^4}{q^4 ( 1+r ) } \right ] \end{array } \label{f2-from - xsec}\ ] ] the ratio @xmath62 , defined as @xmath63 , is the world average value @xmath64 \\[4 mm ] \displaystyle + \frac{0.5747}{q^2 } - \frac{0.3534}{q^4 + 0.09 } , \end{array}\ ] ] taken from @xcite . the coefficients @xmath65 are the mott cross sections for electron and neutrino reactions . for electroproduction on an isoscalar target , @xmath66 is half sum of electroproduction cross sections on proton and neutron . for
charged current neutrinoproduction , in order to eliminate the structure function @xmath67 , one should use the linear combination of the neutrino and anti - neutrino cross sections . for an isoscalar target it is sufficient to take @xmath68 . for a quantitative estimate of the validity of duality
it is convenient to introduce the ratio of the integrals of the resonance ( res ) and dis structure functions @xmath69 where @xmath70 denotes @xmath71 , @xmath3 or @xmath72 ( for neutrino scattering ) .
the value @xmath73 is taken as the actual @xmath21 value for a given parametrization of dis pdfs ( for nucleon , in our case , @xmath74 ) or a dis experimental data set ( for nuclei ) . under conditions of perfect quark hadron duality this ratio would be @xmath75 and independent of @xmath21 . thus , the degree to which the local duality is fulfilled can be estimated from the @xmath21 dependence and the deviation from @xmath75 of the computed @xmath76 . the isoscalar nucleon @xmath77 structure function , which includes both resonance and background contributions , is shown in fig . [ fig : nucleon - em ] versus the nachtmann variable @xmath78 .
notice , that @xmath78 decreases with increasing invariant mass @xmath79 . for a given @xmath21 value ,
the highest peak at the larger @xmath78 value corresponds to the @xmath80resonance peak , and the two lower peaks at smaller values of @xmath78 correspond to the second ( @xmath81 ) and the third ( @xmath82 ) resonance regions .
the general picture shows a reasonable agreement with the duality hypothesis . in fig .
[ fig : i - nucleon - em ] , the ratio of the integrals @xmath83 , defined in ( [ eq : int ] ) , is shown not only for the whole structure function ( resonance @xmath84 1-pion background ) , but also for the resonance contribution separately . for @xmath85 , the ratio @xmath83 for the resonance contribution only is at the level of @xmath86 , which is smaller and flatter in @xmath21 in comparison with the results @xcite of the dortmund group resonance model .
the difference is due to the different parametrization of the electromagnetic resonance form factors used in the two models .
the @xmath2 background gives a noticeable contribution and brings the ratio up to @xmath87 .
the fact , that the ratio is smaller than @xmath75 is no surprise , because additional non - resonant contributions like 2- and many - pion background are possible , but not taken into account here .
they are the subject of coming investigations .
indeed , they could be obtained from requirement @xmath88 .
the principal feature of neutrino reactions , stemming from fundamental isospin arguments , is that duality does not hold for proton and neutron targets separately .
the interplay between the resonances of different isospins allows for duality to hold with reasonable accuracy for the average over the proton and neutron targets .
we expect , that a similar picture emerges in neutrino reactions with nuclei . for neutrinoproduction
, the structure function @xmath89 is shown in fig .
[ fig : nucleon - weak ] for the resonance contribution only .
[ fig : nucleon - weak ] the ratio @xmath90 is shown in fig .
[ fig : i - nucleon - weak ] and appears to be at the level of @xmath91 , which is ( similar to the electron case ) smaller than @xmath92 , which has been calculated within the dortmund resonance model @xcite .
thus , one would expect a large contribution from the background .
the role of the background in the anti - neutrino channel is under investigation .
recent electron scattering measurements at jlab have confirmed the validity of the bloom gilman duality for proton , deuterium and iron structure functions .
further experimental efforts are required for neutrino scattering . among the upcoming neutrino experiments ,
miner@xmath60a@xcite and sciboone@xcite aim at measurements with carbon , iron and lead nuclei as targets .
one of the major issues for nuclear targets is the definition of the nuclear structure functions @xmath93 .
experimentally they are determined from the corresponding cross sections , using eq .
( [ f2-from - xsec ] ) .
we follow the same procedure , using the gibuu cross sections .
so , at the first step the inclusive double differential cross section @xmath94 is calculated within the gibuu model .
the fsi do not play any role for the inclusive cross section .
the in - medium corrections are of considerable importance . in fig .
[ fig : f2-c12-em ] , the resonance contribution to the @xmath95 structure functions for a carbon target are shown for several @xmath21 values .
they are compared to experimental data obtained by the bcdms collaboration @xcite in muon
carbon scattering in the dis region ( @xmath96 ) .
they are shown as experimental points connected by smooth curves . for different @xmath21 values ,
the experimental curves agree within @xmath97 in most of the @xmath78 region , as expected from bjorken scaling .
when investigating duality for a free nucleon , we took the average over free proton and neutron targets , thus considering the isoscalar structure function . since the carbon nucleus contains an equal number of protons and neutrons , averaging over isospin
is performed automatically . due to the fermi motion of the target nucleons ,
the peaks from the various resonance regions , which were clearly seen for the nucleon target , are hardly distinguishable for the carbon nucleus .
as expected from local duality , the resonance structure functions for the various @xmath21 values slide along a curve , whose @xmath78 dependence is very similar to the scaling
limit dis curve .
however , for all @xmath78 , the resonance curves lie below the experimental dis data .
to quantify this underestimation , we now consider the ratio of the integrals of the resonance ( res ) and dis structure functions , determined in eq .
( [ eq : int ] ) .
as it is explained in @xcite , the integration limits are to be determined in terms of the effective @xmath98 variable , experimentally ( see , for example , @xcite ) defined as @xmath99 . for a free nucleon @xmath98
coincides with the invariant mass @xmath79 . for a nucleus
, it differs from @xmath79 due to the fermi motion of bound nucleons , but still gives a reasonable estimation for the invariant mass region involved in the problem .
in particular , the resonance curves presented in all figures are plotted in the region from the pion
production threshold up to @xmath100 . for a free nucleon ,
the threshold value for 1-pion production ( and thus the threshold value of the resonance region ) is @xmath101 .
bound backward - moving nucleons in a nucleus allow lower @xmath79 values beyond the free nucleon limits . the threshold for the structure functions
is now defined in terms of @xmath60 or @xmath98 , rather than @xmath79 .
hence , we consider two different cases in choosing the @xmath78 integration limits for the ratio ( [ eq : int ] ) .
first , for a given @xmath21 , we choose the @xmath78 limits in the same manner as for a free nucleon : @xmath102 \displaystyle \xi_{\rm max } = \xi(\tilde{w}=1.1{{\;\mathrm{gev } } } , \ , q^2 ) \ . \end{array } \label{11 - 16-}\ ] ] we refer to this choice as integrating `` from 1.1 gev '' . the integration limits for the dis curve always correspond to this choice . as a second choice ,
for each @xmath21 we integrate the resonance curve from the threshold , that is from as low @xmath98 as is achievable for the nucleus under consideration .
this corresponds to the threshold value at higher @xmath78 and is referred to as integrating `` from threshold '' . with this choice
we guarantee that the extended kinematical regions typical for resonance production from nuclei are taken into account .
since there is no natural threshold for the @xmath103 , for both choices it is determined from @xmath104 , as defined in eq .
( [ 11 - 16- ] ) .
the results for the ratio ( [ eq : int ] ) are shown in fig .
[ fig : f2-c12-int ] .
the curve for the isoscalar free - nucleon case is the same as in ref .
@xcite with the `` grv '' parametrization for the dis structure function .
one can see that the carbon curve obtained by integrating `` from threshold '' lies above the one obtained by integrating `` from 1.1 gev '' , the difference increasing with @xmath21 .
this indicates that the threshold region becomes more and more significant , as one can see from fig .
[ fig : f2-c12-int ] .
recall , that the flatter the curve is and the closer it gets to 1 , the better local duality would hold .
our calculations for carbon show that the ratio is at the same level as that for the free nucleon or even higher . for neutrino iron scattering ,
the structure functions @xmath105 are shown in fig .
[ fig : fe56-nu ] . as for the electron - carbon results of fig .
[ fig : f2-c12-em ] , the resonance structure is hardly visible .
the resonance structure functions are compared to the experimental data in the dis region obtained by the ccfr @xcite and nutev @xcite collaborations .
it appears , that the resonance curves slide along the dis curve , as one would expect from local duality , but lie well below the dis measurements .
hence , the computed structure functions do not average to the dis curve .
the necessary condition for local duality to hold is thus not fulfilled .
as expected from local duality , the resonance structure functions for the various @xmath21 values slide along a curve , whose @xmath78 dependence is very similar to the scaling
limit dis curve .
however , for all @xmath78 , the resonance curves lie below the experimental dis data .
the ratio @xmath106 defined in eq.([eq : int ] ) is shown in fig .
[ fig : i - fe56-nu ] .
the curve for the isoscalar free nucleon case is also presented for comparison .
our results show , that 1 ) this ratio is significantly smaller than 1 for all @xmath21 ; 2 ) it is significantly smaller than the one for the free nucleon ; 3 ) @xmath76 is even lower than the corresponding ratio for electroproduction ; 4 ) @xmath76 slightly decreases with @xmath21 . to summarize
, we find that the resonance structure functions are consistently smaller than dis functions in the same region of the nachtmann variable @xmath78 .
this is in agreement with earlier work @xcite , which implements elementary resonance vertices and nuclear effects differently .
recall that in this analysis for nuclei we include the resonance structure functions and ignore the background ones .
to estimate their contribution and compare the results with the nucleon case is one of the primary tasks of coming investigations .
overall , the impact of nuclear effects impact on observables is dramatic .
final state interactions of produced pions lead to pions being absorbed in the nucleus or rescattered , thus reducing the cross section and shifting the observed pion distributions to lower values of the pion energy .
furthermore , such interactions can lead to knock - out nucleons which experiments tend to identify with qe events .
ccqe and pion production are thus closely entangled .
thus model - independent data are needed for a meaningful comparison of theory with current experiments .
initial state interactions reduce the cross sections and structure functions in the resonance region and change the form of the distributions , dissolving the peaks of individual resonances .
the bloom - gilman duality seems to be violated for nuclear targets , possibly indicating a major change of background amplitudes in the medium .
l. tiator and s. kamalov , arxiv : nucl - th/0603012 .
d. drechsel , s. s. kamalov and l. tiator , eur .
j. * a34 * , 69 ( 2007 ) , arxiv:0710.0306 .
s. j. barish _
et al . _ , phys
. rev . * d19 * , 2521 ( 1979 )
. g. m. radecky _
et al . _ ,
phys . rev . * d25 * , 1161 ( 1982 ) .
t. kitagaki _ et al .
_ , phys . rev . *
d34 * , 2554 ( 1986 ) .
gibuu , http://gibuu.physik.uni-giessen.de/gibuu .
b. krusche _ et al .
. j. * a22 * , 277 ( 2004 ) , arxiv : nucl - ex/0406002 .
o. buss , l. alvarez - ruso , p. muehlich and u. mosel , eur .
j. * a29 ( 2 ) * , 189 ( 2006 ) , arxiv : nucl - th/0603003 .
o. buss , l. alvarez - ruso , a. b. larionov and u. mosel , phys . rev .
c * 74 * , 044610 ( 2006 ) , arxiv : nucl - th/0607016 .
miniboone , t. katori , arxiv:0909.1996 [ hep - ex ] .
miniboone , a. a. aguilar - arevalo _
et al . _ ,
lett . * 103 * , 081801 ( 2009 ) , arxiv:0904.3159 .
t. leitner , o. buss , u. mosel and l. alvarez - ruso , arxiv:0909.0838 . m. s. athar , s. chauhan and s. k. singh , arxiv:0908.1442 .
miniboone , a. a. aguilar - arevalo _
et al . _ ,
rev . * d79 * , 072002 ( 2009 ) , arxiv:0806.1449
. t. leitner , o. buss , u. mosel and l. alvarez - ruso , phys .
rev . * c79 * , 038501 ( 2009 ) , arxiv:0812.1787 . c. anderson ,
neutral current 1@xmath9 production at miniboone , talk given at nuint 09 , may 17 - 22 , 2009 , sitges , spain , available online at http://nuint09.ifae.es/agenda.html .
k2k , s. nakayama _ et al .
_ , phys . lett .
* b619 * , 255 ( 2005 ) , arxiv : hep - ex/0408134 .
e. d. bloom and f. j. gilman , phys .
lett . * 25 * , 1140 ( 1970 ) .
k. matsui , t. sato and t. s. h. lee , phys .
rev . * c72
* , 025204 ( 2005 ) , arxiv : nucl - th/0504051 . k. m. graczyk , c. juszczak and j. t. sobczyk , nucl . phys . *
a781 * , 227 ( 2007 ) , arxiv : hep - ph/0512015 . k. m. graczyk , arxiv:0910.2355 [ hep - ph ] .
o. lalakulich , w. melnitchouk and e. a. paschos , phys .
* c75 * , 015202 ( 2007 ) , arxiv : hep - ph/0608058 . f. e. close and w. melnitchouk , phys . rev . * c68 * , 035210 ( 2003 ) , arxiv : hep - ph/0302013 . l. w. whitlow , e. m. riordan , s. dasu , s. rock and a. bodek , phys . lett . *
b282 * , 475 ( 1992 ) .
o. lalakulich _ et al .
_ , aip conf . proc . * 967 * , 243 ( 2007 ) . , http://minerva.fnal.gov
. d. bollini _ et al .
_ , phys . lett . *
b104 * , 403 ( 1981 ) .
bcdms , a. c. benvenuti _
. lett . * b195 * , 91 ( 1987 ) .
o. lalakulich , n. jachowicz , c. praet and j. ryckebusch , phys . rev . * c79 * , 015206 ( 2009 ) , arxiv:0808.0085 .
r. m. sealock _
et al . _ ,
lett . * 62 * , 1350 ( 1989 ) .
w. g. seligman _
lett . * 79 * , 1213 ( 1997 ) .
nutev , m. tzanov _ et al .
_ , phys . rev .
* d74 * , 012008 ( 2006 ) , arxiv : hep - ex/0509010 . | neutrino - induced pion production on nuclear targets is the major inelastic channel in all present - day neutrino - oscillation experiments .
it has to be understood quantitatively in order to be able to reconstruct the neutrino - energy at experiments such as miniboone or k2k and t2k .
we report here results of cross section calculations for both this channel and for quasielastic scattering within the semiclassical gibuu method .
this methods contains scattering , both elastic and inelastic , absorption and side - feeding of channels all in a unitary , common theoretical framework and code .
we find that charged current quasielastic scattering ( ccqe ) and @xmath0 production are closely entangled in actual experiments , due to final state interactions of the scattered nucleons on one hand and of the @xmath1 resonances and pions , on the other hand .
we discuss the uncertainties in the elementary pion production cross sections from anl and bnl .
we find the surprising result that the recent @xmath2 production cross section data from miniboone are well described by calculations without any fsi . for higher energies we study the validity of the bloom - gilman quark - hadron duality for both electron- and neutrino - induced reactions .
while this duality holds quite well for nucleon targets , for nuclear targets the average resonance contributions to the structure function @xmath3 are always lower than the dis values .
this result indicates a significant impact of nuclear effects on observables , reducing the cross section and structure functions by at least 30 - 40@xmath4 and changing the form of various distributions .
address = institut fr theoretische physik , universitt giessen , germany address = institut fr theoretische physik , universitt giessen , germany address = institut fr theoretische physik , universitt giessen , germany address = institut fr theoretische physik , universitt giessen , germany address = departamento de fsica , centro de fsica computacional , universidade de coimbra , portugal |
x - ray observations of accreting neutron stars and black holes necessarily probe regions close to the compact object .
the long held promise of these observations is that aspects of the compact object itself , and the innermost accretion flow , may be revealed .
notable early attempts to realize this promise took several forms .
for instance , many attempts were made to use the blackbody spectra of type - i x - ray bursts in accreting neutron star systems to infer the stellar radius ; however , derived radii were implausibly small , likely due to scattering effects ( see london , taam , & howard 1986 ) . in the case of black holes , careful efforts to measure the inner radius of the accretion disk ( see ,
e.g. , makishima et al . 1986 ) were similarly complicated by various observational uncertainties such as mass and distance , and also by scattering effects ( shimura & takahara 1995 ; merloni , fabian , & ross 2000 ) . at the most basic level , measurements of stellar or disk radii using blackbody or modified blackbody continua amount to counting photons under a curve to measure an absolute flux .
this is extremely difficult ( though not impossible ) , and easily complicated by additional continuum components and detector flux calibration uncertainties . partially owing to difficulties associated with continuum spectroscopy , and partially owing to the detection of broad , possibly relativistic fe k emission lines in cygnus x-1 ( barr , white , & page 1985 ) and 4u
1543@xmath1475 ( van der woerd , white , & kahn 1989 ; also see park et al .
2004 ) , parallel efforts to measure fundamental parameters focused on line spectroscopy .
theoretical line models for non - spinning ( schwarzschild ; @xmath2 , where @xmath3 ) black holes and maximal - spin ( kerr @xmath4 ; thorne et al .
1974 ) black holes have been calculated for some time ( fabian et al .
1989 ; laor 1991 ) .
self - consistent models for hard x - ray illumination of an accretion disk that could give rise to such emission lines also have a long history ( e.g. george & fabian 1991 ) .
these models predict that lines arising in the inner disk should have skewed , asymmetric shapes , including a strong blue wing that is enhanced by special relativistic beaming and a long red wing that arises through doppler shifts and gravitational red - shifts .
line shapes are very difficult to discern at the resolution afforded by gas proportial counter spectrometers ( e.g. @xmath5 ) .
confirmation of this line shape was first achieved in an _ asca _ observation of the seyfert-1 galaxy mcg-6 - 30 - 15 ( tanaka et al .
it was facilitated by the x - ray charge - coupled devices ( ccds ) aboard the mission , which delivered a resolution of @xmath6 in the fe k band .
relativistic iron lines are merely the most prominent part of the broad - band response of an accretion disk to hard x - rays , known as the disk reflection spectrum ( see , e.g. , george & fabian 1991 ; magdziarz & zdziarski 1995 ; dovciak , karas , & yaqoob 2004 ; ross & fabian 2005 ) .
x - ray ccds and relativistic spectroscopy of black holes and neutron stars are thus intimately linked .
the advanced ccd spectrometers aboard _
chandra _ , _ xmm - newton _ , and _ suzaku _ have confirmed relativistic lines in a number of seyfert-1 spectra ( for a review , see miller 2007 ) .
just as importantly , the ability of these spectrometers to handle high flux levels has made it possible to clearly detect asymmetric lines in the spectra of stellar - mass black holes , and to reject the possibility that broad lines are actually a collection of narrow lines ( see , e.g. , miller et al .
if the inner disk is assumed to be truncated at the innermost stable circular orbit ( isco ) , which is set by the spin of the black hole ( see bardeen , press , & teukolsky 1972 ) , relativistic lines can be used to infer black hole spin parameters . in a number of stellar - mass black holes ,
the lines observed are sufficiently broad that varying degrees of black hole spin may be required ( e.g. miller et al . 2002 , 2004 , 2008 ;
reis et al .
2009 ; hiemstra et al .
2010 ; for a self - consistent analysis of eight systems , see miller et al .
2009 ) .
very recently , x - ray ccd spectroscopy of transient and persistent neutron star x - ray binaries has revealed skewed , asymmetric disk lines in these systems ( see , e.g. , bhattacharyya & strohmayer 2007 , cackett et al .
2008 , 2009 , di salvo et al .
2009 , da et al .
2009 , da et al . 2010
; for a self - consistent analysis of 10 systems , see cackett et al . 2010 . ) in these sources , relativistic disk lines can be exploited to constrain the radius of the star , since the stellar surface ( if nothing else ) must truncate the disk wherein the lines arise .
the inner extent of the disk is also related to the alfven radius , and so can provide a constraint on stellar magnetic fields . in the case of the relativistic line observed in the millisecond x - ray pulsar sax j1808.4@xmath13658 ( cackett et al .
2009 , papitto et al.2009 ) , the resulting field limits are commensurate with those derived from x - ray timing ( cackett et al .
2009 ) .
new , parallel efforts to derive fundamental properties of compact objects and inner accretion flows have been fueled by new disk models .
new prescriptions explicitly incorporate inner torque conditions , radiative transfer through the disk atmosphere , and even black hole spin parameters ( zimmerman et al .
2005 , li et al .
2005 , davis & hubeny 2006 ) . using these models , black hole spin parameters have been constrained in a number of systems ( see , e.g. , mcclintock et al.2006 , shafee et al .
2006 ; also see zhang , cui , & chen 1997 ) . the improved low - energy range of x - ray ccd detectors ( often extending down to 0.3 kev ) relative to current gas detectors ( coverage below @xmath7 kev is not possible with _ rxte _ nor _ integral _ ) allows for improved measurements of the disk flux , and enables observers to separate disk emission from line - of - sight absorption in the interstellar medium ( though in many cases dispersive spectroscopy may be required for this purpose ; miller , reis , & cackett 2009 ) .
thus , while the improved energy resolution of x - ray ccd spectrometers is not necessarily critical for continuum spectroscopy , their energy range and spectral resolution are beneficial .
while measuring the width of lines is often easier than measuring an absolute flux , it still requires an accurate characterization of the underlying continuum flux , and an accurate knowledge of how the spectrometer reacts to high flux levels .
the latter issue is typically unimportant for gas proportional counter spectrometers , but it can be important for x - ray ccd spectrometers . indeed , the reaction of x - ray ccds is important for both forms of relativistic spectroscopy of compact objects : an instrumental failure to accurately record the flux level is problematic for disk continua , and distortions to the continuum shape and energy response are problematic for line spectroscopy .
photon pile - up occurs when two or more photons land within a detection cell within a single ccd frame time .
this causes a degeneracy between detecting @xmath8 photons with energy @xmath9 and a single photon with energy of @xmath10 .
thus when pile - up occurs , energy and flux information are lost .
it has recently been suggested that ineffective pile - up mitigation could produce spurious results concerning black hole spin , neutron star radii , and the radial extent of accretion disks as a function of the mass accretion rate ( e.g. yamada et al .
2009 ; done & diaz trigo 2010 , ng et al .
advanced statistical descriptions of photon pile - up in x - ray ccd detectors have been developed ( davis 2001 ; also see ballet 1999 ) , and implemented into fitting packages such as xspec and isis .
the davis et al . ( 2001 ) model has been applied in a number of regimes to correct for photon pile - up distortions , including spectra from : isolated neutron stars ( e.g. van kerkwijk et al .
2004 ) , ultra - luminous x - ray sources ( e.g. roberts et al .
2004 ) , transient galactic black hole candidates ( e.g. jonker et al . 2004 ) , low - mass x - ray binaries ( e.g. heinke et al .
2006 ) , and even agn ( e.g. wang et al .
most notably , nowak et al .
( 2008 ) employed the basic elements of the davis model to correct distortions to the _ chandra _ spectra of the black hole candidate 4u 1957@xmath1111 .
inner disk parameters consistent with those derived using spectra from other missions , including _ xmm - newton _ and _ rxte _ , were then recovered . a similar treatment of _ chandra _ spectra of cygnus x-1 by hanke et al .
( 2009 ) was able to bring different spectra into close agreement . with the aim of developing a rigorous understanding of how pile - up may affect relativistic spectroscopy
, we have conducted extensive simulations based on a range of assumed input spectra , flux levels , and detector properties .
this effort is timely as x - ray ccd spectrometers will be the standard in the field for the foreseeable future , at least for moderate resolution spectroscopy of bright targets . in the following sections
, we briefly review the operation of x - ray ccd detectors , review the pile - up model used in our simulations , describe our simulations and results , and discuss the impacts of our findings .
the primary goal of this analysis is not to understand how to best avoid photon pile - up , though the results offer some insights on this point .
rather , the goal of this exercise is to understand the nature and magnitude of distortions to relativistic disk parameters when photon pile - up is present but unknown to an observer , poorly quantified , or difficult to mitigate in a rigorous manner . in a subset of our simulated spectra , then , photon pile - up is quite severe . when faced with similar real data , an observer would likely be motivated to extract an annulus or to otherwise at least partially mitigate the effects of pile - up .
at the most basic level , a ccd is an array of coupled capacitors .
when an incident photon interacts in the semiconductor substrate , electrons are liberated .
the charge is collected and stored in pixels , and it can be transfered to neighboring pixels and eventually into an amplifier and read - out system by virtue of the coupling . in many respects ,
the operation of x - ray ccds is similar to that of optical ccds .
one important difference is that optical photons only liberate a small number of electrons ( typically zero or one ) in an interaction .
x - ray photons have much more energy , of course , and liberate many electrons . indeed , the number of electrons liberated in an interaction is proportional to the energy of the incident x - ray .
this fact means that x - ray ccds are not merely imaging instruments , but medium - resolution spectrometers as well .
the gas proportional counters aboard _
rxte _ have an energy resolution of @xmath12 ; current x - ray ccds ( e.g. _ chandra_/acis ) achieve an energy resolution of @xmath13 at 5.9 kev .
most x - ray ccds cover the 0.310.0 kev band .
these limits are set by a combination of factors , including the efficiency of capturing and clocking charge clouds generated by low energy x - rays ( this is partially set by the electrodes ) , the depletion layer thickness , the ability of current x - ray mirrors to focus hard x - rays , and the ability of the instrument to distinguish hard x - rays from non - x - ray events .
the detailed performance characteristics of a given x - ray ccd depend on several variables , including the nature of the device structure ( n- or p- substrate , mos or pn junction ) ; whether charge is stored and read - out close to the x - ray - illuminated face of the ccd , or far from the illuminated face ; the temperature at which the ccd is being operated ( this affects dark currents and noise ) , and a number of more subtle factors .
photons are detected and elementary screening of desirable ( e.g. x - ray ) versus undesirable ( e.g. cosmic ray ) events is achieved in part by assigning event `` grades '' .
interactions are characterized using event `` boxes '' , which are usually 3x3 cells of pixels .
the pattern of pulseheights that are consistent with photon interactions can be defined and differentiated from `` hot '' pixel and cosmic ray patterns .
efficient spectroscopy with an x - ray ccd depends on each event box recording zero or one photons per ccd frame time .
when the x - ray flux incident on a ccd is too high , multiple photons can be read as a single high energy photon ( see above ) .
the resulting flux is thus registered as falsely low , and the resulting spectrum as falsely hard .
grade migration provides a potential means of recognizing and diagnosing the degree of photon pile - up suffered in a given observation . when pile - up is important , charge patterns consistent with single photon interactions are reduced and patterns consistent with higher energy photon interactions or multiple photon interactions are increased .
various strategies can be deployed to deliver nominal spectroscopic response even at high x - ray flux levels .
the nominal ccd frame time can be greatly reduced , thus lowering the probability of two photons registering in a given event box .
this often comes at the expense of imaging information , observing efficiency , or both .
over - sampling a broad telescope psf with a high number of event boxes is another viable means of limiting the number of photons that may land in a given event box .
the expense of a broad psf is , of course , the loss of fine image quality .
observers can try to further prevent or limit photon pile - up by only extracting events from the wings of the psf in annuli , by placing a source off of the optical axis of the telescope to blur its flux over more pixels , or by adopting a rigid grade selection that permits only single - photon events . in practice , however , some spectra obtained with x - ray ccds will suffer from photon pile - up .
pile - up is a complex but well - understood statistical process : at a given flux level , a given event box has a fixed probability of getting 0 , 1 , or n photons .
the probability of a given event grade for each outcome is also well - determined . if the input spectrum is known , or if it can be estimated , the distribution of `` good '' event grades ( those likely to originate from actual x - ray photons ) can be used to infer the extent and severity of photon pile - up . on this basis ,
davis ( 2001 ) has developed a model of photon pile - up .
the model not only accounts for the energy shifts due to photon pile - up , but also for flux decrements due to grade migration .
the crux of the davis model is a new integral equation for the number of counts detected from a point source . the conventional equation for the number of counts detected in an x - ray spectrometer is linear , and discussed in detail in gorenstein , gursky , & garmire ( 1968 ) .
however , pile - up is a nonlinear process , and requires a more advanced treatment .
davis ( 2001 ) presents an alternative integral equation that accurately characterizes the counts observed in the presence of photon pile - up .
a fundamental assumption of the pile - up model is that a charge cloud in the ccd arising from n photons can be treated as the linear superposition of n individual charge clouds .
since the drift time for charge clouds in a ccd is typically much less than a microsecond , which is much smaller than typical readout times , and since a ccd amplifier is highly linear at x - ray signal levels , this assumption is valid for all of the ccd modes that will be considered in this work . in practice ,
the chief uncertainty in correcting for pile - up using this model is that the details of the input spectrum are usually not known a priori .
exceptions might include instances where different missions are simultaneously observing the same source , or an instance wherein a given ccd spectrometer alternates between different operational modes .
_ this uncertainty does not enter into our analysis .
_ simulating the effects of photon pile - up by convolving different input spectra with the pile - up model of davis ( 2001 ) is a well - controlled experiment the nature of the input spectrum is known perfectly .
the results of this procedure can be thought of as characterizing the nature and severity of systematic errors on relativistic disk parameters arising due to photon pile - up .
it is worth noting that although the pile - up model was developed for _ chandra _ , it is easily applicable to other missions .
the crucial modification in describing pile - up in other ccd spectrometers is to accurately account for the fraction of the telesope point spread function ( psf ) that is enclosed in each event box .
this can differ considerably from mission to mission . whereas a single event box captures a high fraction of _ chandra s _ psf , a single event box captures a very small fraction of _ suzaku s _ psf . in the latter case , to evaluate pile - up in a reasonable extraction region ,
many event boxes must be considered jointly . as implemented in the xspec spectral fitting package ( arnaud & dorman 2000 ) , the davis pile - up model has six parameters : the ccd frame time , the maximum number of photons to pile - up , the grade correction for a single photon , the grade morphing parameter ( the grade migration function is a probability assumed to be proportional to @xmath14 , @xmath15 , where @xmath16 is the grade morphing parameter and @xmath17 is the number of piled photons ) , the psf fraction considered , and the number of event regions .
this description of grade migration is based on the premise that photon arrival times in a given cell obey poission statistics .
it correctly captures the fact that the probability of recording an unwanted event grade increases with the number of piled photons .
to understand the specific effects of photon pile - up on disk continua and reflection parameters , we created a large number ( 325 ) of simulated spectra using the davis pile - up model .
realistic , multi - component spectral forms based on published results were employed to make the simulations as realistic as possible .
the spectra were generated using the `` fakeit '' command in xspec version 12 .
for simplicity , and because this exercise is necessarily concerned with bright sources , no backgrounds were used in making the source spectra . in all cases , an exposure time of 100,000 seconds was used when creating simulated spectra .
this exposure is longer than a typical observation , but allows for excellent photon statistics . in all of the simulations made in this work
, default values were assumed for the maximum number of photons to pile - up ( 5 ) and the grade correction for a single photon ( 1 ) .
the frame time , psf fraction , and number of event boxes were set according to the telescope and ccd combination used in each simulation . with these parameters fixed
, the grade morphing parameter @xmath16 effectively sets the severity of photon pile - up ( when the incident source flux in each event box leads to pile - up ) .
as implemented in xspec , the pile - up model is a convolution model .
input spectra were convolved with the pile - up kernel to produce a resultant spectrum that is distorted by photon pile - up effects . in practice , for a given detector and mode ,
the degree of pile - up is set by the incident flux level . in our simulations , the degree of photon pile - up was controlled by adjusting the grade morphing parameter @xmath16 .
steps of 0.1 were used for @xmath18 , and steps of 0.04 were used for @xmath19 .
the step size was reduced for high values of @xmath16 in order to provide better resolution when pile - up typically starts to introduce strong distortions . for each combination of input spectrum and detector mode ,
then , 13 simulated spectra were generated .
x - ray binaries display a wide variety of phenomena . periods of correlated multi - wavelength behaviors can be classified into `` states '' that may correspond to distinct changes in the accetion flow ( for a review , see , e.g. remillard & mcclintock 2006 ; also see belloni et al .
2005 ) . in this work
, we consider three commonly - recognized `` states '' for black hole x - ray binaries , and a single input spectral form for neutron star low - mass x - ray binaries ( the so - called `` z '' and `` atoll '' sources ) . in the interest of simplicity and reproduciblity , we have adopted phenomenological spectral models for each state with values consistent with those reported in the literature .
table 1 lists values for all of the input spectra considered in this work .
the paragraphs below offer some context and motivation for why certain models and values were used to generate the spectra . _ the very high state _ : this state can be extremely bright , and it has seldom been observed using ccd spectrometers .
continua observed in this state often consist of a hot disk and steep power - law component .
we assumed the continuum parameters measured in the very high state of gx 339@xmath14 using the _ xmm - newton_/epic - pn in `` burst '' mode ( miller et al .
this mode has a frametime of just 7 @xmath20sec and a deadtime of 97% , but it is effective at preventing photon pile - up for fluxes up to 6 crab .
the particular spectrum chosen as a template is broadly consistent with very high state spectra obtained from _
monitoring of other transients ( e.g. 4u 1543@xmath1475 , park et al
. 2004 ; gro j1655@xmath140 , sobczak et al . 1999 ; xte j1550@xmath1564 , sobczak et al .
2000 , miller et al.2003 ) .
the `` intermediate '' state is similar to the `` very high '' state in terms of its spectral form ( both thermal disk and non - thermal power - law - like emission are important ) , but typically has a lower flux .
the effects of photon pile - up in this state are likely to be similar to the effects on very high state spectra at low values of the grade migration parameter @xmath16 .
the `` intermediate '' state is not treated separately in this work . _ the high / soft state _
: this state can also be quite bright , but it typically persists for a longer period than the `` very high state '' . as a result , it is more commonly observed with ccd spectrometers .
continua observed in this state are strongly dominated by a hot disk component , accompanied by a weak , steep power - law . to model the high / soft state , we selected one of the spectrally softest observations of 4u 1543@xmath1475 obtained using _ rxte _ ( park et al .
this observation occurred on mjd 52475 , and is consistent with the spectra selected for relativistic disk spectroscopy by shafee et al .
( 2006 ) . here
again , this high / soft state is typical of spectra of other black hole transients in the same state ( see , e.g. , sobczak et al .
1999 , 2000 ) .
_ the low / hard state _ : the low / hard state has been a frequent target with ccd spectrometers recently , in an effort to understand the accretion flow at low mass accretion rates ( for a recent survey , see reis , fabian , & miller 2010 , also see tomisck et al .
spectral continua in this state are dominated by a hard power - law , sometimes with weak emission from the accretion disk .
the continuum spectrum assumed in our simulations is a composite of the parameters reported in observations of gx 339@xmath14 by miller et al .
( 2006 ) , tomsick et al .
( 2008 ) , and wilkinson & uttley ( 2009 ) . _ a typical z / atoll spectrum _ :
unlike most black hole x - ray binaries with low - mass companions , neutron stars with low magnetic fields are typically persistent ( but variable ) sources . whereas black hole continua can often be well - characterized in terms of thermal disk emission and a power - law , neutron star spectra often require a third component ( lin , remillard , & homan 2007 ; cackett et al .
2008 , cackett et al .
2010 , da et al .
a simple blackbody likely corresponds to emission from the boundary layer between the accretion disk and stellar surface ( revnivtsev & gilfanov 2006 ) .
emission from this region may be comptonized , but blackbody emission that is compton - upscattered in a region with high @xmath21 and low @xmath22 can be modeled with a hotter , smaller blackbody ( london , taam , & howard 1986 ) .
we used the spectral continuum parameters reported by cackett et al .
( 2008 ) for serpens x-1 .
relativistic line components were added to all continuum spectra , apart from those for the `` high / soft '' state .
the `` laor '' relativistic emission line model ( laor 1991 ) was assumed to simulate emission lines from the inner disk . in all cases ,
the input inner radius was fixed at @xmath23 , the emissivity index was fixed at @xmath24 ( where @xmath25 ) , and the inclination was fixed at @xmath26 .
this radius is substantially larger than the innermost stable circular orbit for a maximally - spinning kerr black hole ( @xmath27 ) , and significantly larger than measured in a number of stellar - mass black holes ( miller et al .
however , one purpose of this investigation is to see if photon pile - up can artificially _ broaden _ relativistic lines , giving falsely small inner radii and falsely high spin values . setting the inner radius at @xmath23 , then , permits distortions due to photon pile - up ample opportunity to produce falsely broad lines . for all simulated black hole spectra ,
the input flux of the relativistic `` laor '' line was normalized to give an equivalent width of @xmath28 ev as per miller et al .
( 2004 ) .
similarly , @xmath23 is approximately 12.4 km for a 1.4 @xmath29 neutron star only slightly larger than the canonical stellar radius of 10 km employed in many circumstances . assuming the stellar surface ( if not the boundary layer ) must truncate the accretion disk , a relativistic line suggesting a smaller radius might be taken as evidence of a `` strange '' star . in the case of neutron stars , then , fixing the input radius at @xmath23 serves to assess the ability of photon pile - up to give false evidence of `` strange '' stars through falsely small disk radii . for all simulated neutron star spectra ,
the input flux of the `` laor '' line was normalized to given an equivalent width of @xmath30 ev as per cackett et al .
( 2008 ) .
iron emission lines in accreting black holes and neutron stars are merely the most prominent part of the disk reflection spectrum ( e.g. dovciak , karas , & yaqoob 2004 ; ross & fabian 2005 ) .
self - consistent modeling of real spectra requires modeling of the entire reflection spectrum , not just the broadened emission line . for simplicity , and because the line drives constraints on inner disk radii ( and therefore spin ) in fits to real data , disk reflection models are not treated in this analysis . in all spectra , thermal continuum emission from the accretion disk was simulated using the simple `` diskbb '' model ( mitsuda et al .
1984 ) within xspec . using this model ,
the inner disk radius can be derived via : @xmath31 , where @xmath32 is the model flux normalization , @xmath33 is the distance to the source , and @xmath34 is the inclination of the inner disk . if the mass of the source is known , this radius can be converted to gravitational units .
this is merely a `` color '' radius , however , and does not account for the effects of spectral hardening due to radiative transfer through the disk atmosphere ( shimura & takahara 1995 ; merloni , fabian , & ross 2000 ) .
it is well - known that this model is overly simple in other ways .
for instance , it does not include a zero - torque condition at the inner edge , whereas including this condition can reduce implied radii by a factor of @xmath352 ( zimmerman et al .
moreover , a number of new disk models have been developed in which black hole spin is an explicit parameter ( davis & hubeny 2006 ) .
deriving spins using these newer , more physical models requires extremely accurate knowledge of the absolute disk flux , however , which can be complicated by a number of effects ( e.g. the flux calibration of the detector , the accuracy to which line of sight absorption is known , the nature of the hard component , etc . ) . and
whereas new models can have as many as 10 free parameters , the `` diskbb '' model is able to accurately characterize the thermal continuum with only two ( temperature and flux normalization ) . owing to its simplicity and prior application to many neutron star and black hole spectra over numerous years and x - ray missions , then
, all thermal disk continua were simulated using the `` diskbb '' model , and our analysis is primarily concerned with changes in the flux normalization parameter . in the z / atoll spectra , a simple ( single - temperature )
blackbody function was used to describe emission from the stellar surface , independent of the thermal emission from the disk .
each input spectrum was modified by interstellar absorption using the `` tbabs '' model ( wilms , allen , & mccray 2000 ) . in all cases ,
the equivalent neutral hydrogen column density was fixed at @xmath36 .
this is a moderate value , consistent with columns that have facilitated studies of the thermal disk continuum in a number of sources and at different mass accretion rates . although a number observations made with the _ chandra_/hetgs ( e.g. di salvo et al .
2005 ) and with _ suzaku _ ( cackett et al.2008 , reis , fabian , & young 2009 , cackett et al .
2010 ) find broad , relativistic line shapes in the spectra of neutron stars , such lines are less established in neutron star spectra than in black hole spectra .
ng et al . (
2010 ) suggest that neutron star lines may actually be narrower ( @xmath37 kev , on average ) and symmetric .
therefore , we also simulated neutron star spectra using the same neutron star continuum described above and in table 1 , but with gaussian model with @xmath38 kev , @xmath37 kev , and an equivalent width of 150 ev , rather than a relativistic line .
the ccd spectrometers and modes considered in our simulations are detailed in table 2 .
the modes selected reflect those that users might realistically select when observing bright sources .
the `` burst '' mode of the epic - pn camera aboard _
xmm - newton _ is not considered , as it is capable of handling exceptionally high flux levels without pile - up distortions .
common imaging modes available on the acis and xrt instruments aboard _
chandra _ and _ swift _ , respectively , are also not considered ; these are known to readily suffer pile - up for even moderately bright sources and are typically avoided by users . finally , with _
chandra _ , pile - up can be largely mitigated while obtaining a high - resolution spectrum with the hetgs ; however , assessing pile - up in this circumstance is especially complex and beyond the scope of this investigation .
an `` effective '' frame time is given in table 2 .
this is the frame time that is important to consider when assessing the extent and impacts of photon pile - up .
it multiplies the time required to extract one element of charge ( typically a pixel , but sometimes a larger `` macropixel '' ) by the number of elements necessary to define an event box .
( some missions but not all redefine event boxes for different modes ; see the discussion below . ) if one could know a priori that a sequence of events were all `` singles '' wherein charge from photon events was contained in only one pixel , then the effective frame time would equal the time required to clock just one row of pixels . however , it is not possible to know this a priori , and _ in practice one only knows that an event was a `` single '' if the other rows in an event box are free of charge_. thus , for pile - up , the time scale of concern is the time required to transfer a full event box .
a significant simplification in our simulations is that the grade morphing parameter @xmath16 is assumed to be constant in each extraction region . in practice , this is not true ; @xmath16 would depend on radius as an actual telescope psf focuses successively less energy into successively larger annuli .
our simulations therefore capture the strongest impacts of photon pile - up distortions , and ignore how less distorted spectra from large annuli might affect results when summed with spectra from smaller annuli . in a partial effort to minimize the impact of the assumption of a constant @xmath16 , the extraction regions assumed in our simulations do not encircle the same fraction of the total incident energy , but rather attempt to sample the part of the psf where the encircled energy changes only linearly with radius .
this is not practical in the case of _ chandra _ , however , which has a very tight psf .
a secondary simplification is that the pile - up model and our simulations do not consider how charge from event boxes might bleed into adjacent boxes in the case of severe pile - up .
additional details concerning the individual spectrometers and modes considered , and how they are treated in our simulations , are given below : acis event boxes consist of a 3x3 box of pixels , each approximately 0.5 on a side .
in continuous clocking or `` cc '' mode , charge is continuously transferred from the ccd ( usually acis - s3 ) at a rate of 2.85 msec per row .
two - dimensional imaging is sacrificed in favor of fast read - out .
the definition of acis event boxes is unchanged between standard imaging modes and cc mode . in our simulations , then , we have assumed an effective frametime of 8.55 msec , since this is the time require to clock a full event box . as noted above ,
given the tight psf of _ chandra _ , it is impractical to extract photons from a region that encircles less than 90% of the incident energy . in our simulations , we therefore assumed an extraction radius of 2.5 , which corresponds to 90% of the incident energy .
in the one - dimensional regime of cc mode , the extraction region is then a bar 5 in length .
note that this means that only 3.3 event boxes tile the extraction region . where possible , generic , readily - available response matrices were used in order to facilitate checks on this work by other teams .
to simulate acis cc mode spectra , then , we used the acis response matrices made available by the mission for simulations supporting cycle 11 proposals : `` aciss_aimpt_cy11.rmf '' and `` aciss_aimpt_cy11.arf '' . for more information about the acis spectrometer and cc mode ,
please see garmire et al .
( 2003 ) and the _ chandra _ proper s observatory guide ( http://cxc.harvard.edu/proposer/pog/ ) . in `` windowed timing '' or `` wt '' mode , charge rows are grouped by 10 in the read - out direction ( effectively creating `` macropixels '' ) and rapidly tranferred from the ccd . in the standard photon counting mode ,
a 3x3 grouping of pixels ( each pixel is 2.36 on a side ) is used to define an event box . however , in windowed timing mode , the event box is altered to be 10x7 pixels : 1 macropixel in the clocking direction , and 7 pixels in the orthogonal direction . this change improves the ability of the detector to identify and screen events in windowed timing mode . note that because the new event box is only 1 macropixel in the clocking direction , the nominal and effective frame times are the same in windowed timing mode ( see table 2 ) . like _
`` continuous clocking '' mode , `` windowed timing '' mode has a livetime fraction of 1.0 . in our simulations
, we assumed an extraction radius of 9.0 , which roughly corresponds to the half power radius for the _ swift_/xrt .
the extraction region is tiled by only 1.1 event boxes . for simplicity
, we have also assumed an effective frame time of 1.77 msec ( see table 2 ) . in practice , depending on where a photon strikes , it may take up to two read - out times of 1.77 msec to clock the charge through the extraction region ( which roughly matches the half - power diameter of the psf ) . in this sense , our simulations slightly under - estimate the severity of pile - up distortions to _
swift_/xrt spectra obtained in `` windowed timing '' mode . to simulate `` windowed timing '' mode spectra , we used current , standardized response functions available through the heasarc calibration database , `` swxwt0to2s6_20070901v011.rmf '' and `` swxwt0to2s6_200101 01v011.arf '' .
these responses have been developed specifically for wt mode , and differ from the responses appropriate for normal imaging modes . for more information about the _ swift_/xrt
, please see hill et al .
( 2004 ) , burrows et al .
( 2005 ) , the _ swift _ technical handbook ( http://swift.gsfc.nasa.gov/docs/swift/proposals/appendix_f . html ) and the xrt user s guide ( http://swift.gsfc.nasa.gov/docs/ swift / analysis / xrt_swguide_v1.2.pdf ) .
the `` timing '' mode of the epic - pn camera is similar to the `` windowed timing '' mode of the _ swift_/xrt in some respects . to achieve a high time resolution ,
normal ccd pixels are grouped into `` macropixels '' .
each macropixel is 10 pixels in the read - out direction and 1 pixel in the orthogonal direction .
since each pixel is 4.1 on a side , this means that a single macropixel is actually 41 in the read - out direction . in its long dimension , then , a single macropixel is equivalent to the 70% encircled energy diameter of the telescope . in standard imaging modes ,
_ xmm - newton _ event boxes are the standard 3x3 pixel element , as per other x - ray ccd cameras . in timing mode , however , an event box is 3 macropixels by 3 macropixels .
this means that a single event box is 123 long ( equivalent to the 95% encircled energy diameter ) and 12.3 wide ( appoximately equal to the half - power diameter ) .
this event box is extremely large , and roughly equivalent to the extraction region that is often used by observers using this mode .
creating macropixels enables a short read - out time of 0.03 msec _ per macropixel_. this translates to an effective frame time of 0.09 msec _ per event box_. again , this is a simplification that may serve to under - estimate the severity of photon pile - up distortions to `` timing '' mode spectra . in practice , each macropixel takes at least three read - out times to be completely clocked , and it may take four timescales depending on where the photon strikes .
thus , each macropixel may contain more charge than anticipated in these simplified simulations .
the short frame time achieved in `` timing '' mode is impressive , especially considering that `` timing '' mode operates with a live - time fraction of 0.99 .
unlike its cousin , `` burst '' mode ( live - time fraction : 0.03 ) , `` timing '' does not handle high fluxes by making exceptionally short exposures while discarding the bulk of the incident flux .
these abilities come at the expense of tiling the psf with many event boxes , however , which is another important means of reducing the severity of photon pile - up distortions .
a distinctive feature of the _ xmm - newton_/epic - pn camera is that diagonally adjacent pixels are treated as two single pixel events .
differences between detectors at this level are assumed to be small .
generic and separate redistribution and ancillary matrix files are not readily available to _ xmm - newton _ users .
timing mode response matrices were generated using the `` rmfgen '' and `` arfgen '' tools in sas version 9.0.0 .
these matrices where then used to simulate all timing mode spectra . for more information about _ xmm - newton _ and the epic - pn camera ,
please see strder et al .
( 2001 ) , haberl et al .
( 2004 ) , and the user s handbook ( http://xmm.esac.esa.int/external/xmm_user_support/ documentation / index.shtml ) .
`` full frame '' mode is a standard imaging mode for the mos 1 and mos 2 cameras aboard _
xmm - newton _ , not a specialized timing mode like those described above .
there are two reasons for considering it in this analysis .
first , the small , 1.1 pixels of the mos cameras allow for a large number of event boxes to tile an extraction region , which is one viable strategy for reducing the effects of photon pile - up .
second , by extracting annuli that excise central regions within a piled - up source image , observers are often able to make use of data obtained in this and similar operating modes . whether or not pile - up is totally mitigated through this procedure is sometimes in doubt , and this analysis can help to clarify the nature of any distortions imposed on the extracted spectrum by residual pile - up . in simulating mos `` full frame '' spectra , circular
extraction regions with 20 radii were assumed . within this range , the encircled energy fraction of the psf changes almost linearly with radius . the regions assumed would encircle approximately 70% of the total incident energy . as with pn `` timing '' mode , responses were generated using the `` rmfgen '' and `` arfgen '' functions within sas version 9.0 .
these responses were then used to create all simulated `` full frame '' spectra . for additional information on the _ xmm - newton_/epic - mos cameras ,
please see turner et al .
( 2001 ) .
the xis spectrometer aboard _ suzaku _ provides many different operational modes that can be used to optimize science returns and to reduce or eliminate photon pile - up .
each xis ccd has a nominal frame time of 8.0 seconds , but this can be reduced by choosing a `` window '' option .
for instance , using a 1/4 window means that only 1/4 of the ccd will be exposed , reducing the frame time to 2.0 seconds .
( smaller windows are possible but can not be used owing to drift in the spacecraft pointing over the course of an observation . ) the frame time can be further reduced by selecting a `` burst '' option ; in this case , the ccd only exposes for a fraction of the nominal frame time .
the remainder of the frame time is then dead time , reducing the overall observing efficiency and requiring longer exposures . for this analysis , the second - most conservative combination of window and burst option has been considered . for all simulated spectra
, we assumed a 1/4 window and 0.3 second burst option .
it should be noted that the xis can be operated in `` psum '' mode , wherein the full height of the ccd is read - out in the time normally required to clock one row of charge . in some respects ,
this mode is similar to the `` timing '' and `` burst '' modes available on the _ xmm - newton_/epic - pn camera , and for similar reasons it is actually less effective at mitigating pile - up than the most conservative combinations of xis window and burst options .
each simulated spectrum assumed a circular extraction region with a radius of 60 , which encircles approximately half of the incident energy for the _ suzaku _ psf . within this radius , the encircled energy fraction changes in a roughly linear way with radius .
each xis pixel is 1 on a side , and standard 3 pixel by 3 pixel boxes are used to define event grades . within the extraction region ,
then , there are 1256.0 event boxes . tiling a broad psf with many event boxes
acts to reduce the flux per event box . for more information on the xis
, please see the _ suzaku _ abc guide ( http://heasarc.gsfc.nasa.gov/docs/suzaku/analysis/abc/ ) and the _ suzaku _ technical description ( http://heasarc.gsfc.nasa . gov / docs / suzaku / aehp_prop_tools.html ) .
all simulated spectra were fit using the same model components that were employed to generate the spectrum , but without using the pile - up kernel .
this procedure permits an examination of how spectral fitting results are distorted by photon pile - up effects , when such effects have not been mitigated entirely ( e.g. by extracting events in an annulus , by selecting a shorter frame time , etc . ) . for the purpose of illustrating how parameters are distorted even in the event of severe photon pile - up ,
fits were made even when the @xmath39 goodness - of - fit statistic was not acceptable .
all errors reported in this analysis are @xmath40 confidence errors , and were determined using the `` error '' command within xspec . in the cases where the fits were unacceptable , the errors are not true @xmath40 errors . a few simplifying assumptions were made in fitting the simulated spectra .
the equivalent neutral hydrogen column density was held fixed at the input value , @xmath41 . in practice
, this might be equivalent to assuming the column density measured via a sensitive _ chandra_/hetgs or _ xmm - newton_/rgs spectrum in which individual neutral photoelectric absorption edges were detected ( e.g. miller , cackett , & reis 2010 ) .
some of the relativistic iron line parameters were constrained to ranges and/or fractional uncertainties that are common in the literature . specifically , all fits to the simulated line spectra assumed @xmath42 , @xmath43 , and @xmath44 .
fits to the simulated spectra were otherwise unconstrained ; line and continuum model parameters were free to take on whatever value best described the distorted spectra .
the goal of this analysis is simply to understand how disk reflection and disk continua are affected by pile - up .
a detailed treatment of distortions to ( more physical ) continuum spectra ( e.g. comptonization spectra , and/or coupled disk plus corona models such as `` eqpair '' ) might also be timely , but it is beyond the scope of this paper .
the sections that follow are narrowly focused on the results of fits to relativistic lines and simple disk continua .
the results of these limited exercises are clear : when photon pile - up is severe , it causes relativistic emission lines in black hole spectra to be measured as falsely narrow . in fits to the simulated spectra with the laor line model , measured radii
were all consistent with or significantly larger than the true value of @xmath45 .
figures 1 and 2 trace the evolution of measured radius as a function of pile - up severity ( set by the grade migration parameter @xmath16 ) .
figure 3 shows how the entire spectrum is affected by pile - up .
figure 4 shows a sequence of line profiles from simulations with increasing photon pile - up .
the same results hold in the case of our simulated neutron star spectra , with one modest exception ( see below ) .
even the baseline flux level assumed in our `` very high '' state simulations ( corresponding to low values of @xmath16 ) is found to completely overwhelm _
chandra_/acis `` continuous clocking '' mode and _ xmm - newton_/epic - pn `` timing '' mode .
the emission line could not be detected nor reliably fit except at low values of the grade migration parameter @xmath16 . in any real observation at an even higher flux level ( corresponding to a higher value of @xmath16 ) , then , a line may not be clearly detected .
similarly , for the _ xmm - newton_/epic mos in `` full frame '' mode , event loss dominates over grade migration in the simulated spectra , and so the _ shape _ of the spectrum including the line is preserved though the line and continuum flux are a small fraction of their input values .
the results shown in figure 1 do not indicate that the mos `` full frame ''
mode is equipped to deal with typical `` very high '' state flux levels .
the _ swift_/xrt in `` wt '' mode and _ suzaku_/xis ( with 1/4 window and 0.3 second burst option ) are less overwhelmed by the high flux associated with our baseline `` very high '' state spectrum . in fits to those simulated spectra , the inner disk radius correlates with the severity of pile - up as traced by the grade migration parameter @xmath16 ( again , @xmath16 serves as a proxy for increasing the flux over the baseline levels given in table 1 ) .
the results shown in figure 1 indicate that the _
suzaku_/xis is actually best - equipped to deal with such high flux levels , although even in this case fits to the iron line fail to recover the full line width and give falsely large inner disk radii .
the apparent fluctuations in inner disk radius at high values of @xmath16 seen in the xrt trend in figure 1 are likely the result of errors that are under - estimated because the overall fit is poor . in the `` low / hard '' state , the degree of spectral distortions is generally reduced with respect to the `` very high '' state ( see figure 2 ) .
the most accurate inner disk radii were recovered in the simulated _
suzaku_/xis spectra .
a 1/4 window option and 0.3 seconds burst option appear to suffer minimal disortion due to photon pile - up , at least for the flux levels that would generate the values of @xmath16 that were simulated .
photon pile - up distortions are more pronounced in the other spectra : measured inner disk radii are found to depart more strongly from the true value with increasing @xmath16 .
the _ xmm - newton_/epic - pn `` timing '' mode and _ swift_/xrt `` windowed timing '' mode appear to only suffer modest distortions as pile - up becomes more severe : in each case , lines are measured to be falsely narrow and to give inner disk radii that are too large by 23 @xmath46 .
fits to the simulated _ xmm - newton_/epic mos spectra in `` full frame '' mode show the most marked trend : even for the lowest flux levels ( lowest values of @xmath16 ) , pile - up is so severe that measured inner radii depart from the input value by factors of a few .
chandra_/acis `` continuous clocking '' mode is still overwhelmed at the baseline flux level give in table 2 , and lines can not be required in spectral fits for @xmath47 . as with the `` low / hard '' state , the simulated _
chandra_/acis `` cc '' mode neutron star spectra were severely distorted by photon pile - up , and other modes were less affected .
it was only possible to detect and fit the iron line reliably for @xmath48 for _ chandra_/acis `` continuous clocking '' mode ( see figure 5 ) . in the case of the other simulated neutron star spectra ,
a limited number of measured radii were 10% smaller ( e.g. @xmath49 ) than the input value .
this effect was only seen in simulations of spectra obtained with the epic - pn camera in `` timing '' mode .
if these results are an accurate characterization of the systematic errors incurred when fitting mildly piled - up spectra with the epic - pn camera , it is a modest systematic error .
systematic uncertainties in the flux calibration between different cameras on the same observatory are approximately 7% ( _ xmm - newton _ calibration document xmm - soc - cal - tn-0083 ) and uncertainties between cameras on different missions can easily exceed 10% or more .
the neutron star spectra that were simulated assuming a narrower and symmetric gaussian emission line with [email protected] kev and @xmath51 kev ( as per ng et al .
2010 ) also give clear results .
the gaussian line width , line centroid energy , and the normalization of the disk blackbody component are plotted versus @xmath16 in figure 6 .
( the disk blackbody component is the lowest - energy flux component in the spectral model , and changes to its flux trace the degree of flux redistribution due to photon pile - up . ) when photon pile - up is modest , the line width and line energy are not strongly affected . at the flux levels simulated ,
the _ xmm - newton_/epic mos `` full frame '' mode suffers more severe pile - up .
the line becomes _ narrower _ as photon pile - up ( traced by the grade migration parameter @xmath16 ) becomes more severe ( see figure 6 ) . in no case
do fits to the simulated spectra measure a line that is falsely broad , asymmetric , nor displaced to a significanly lower centroid energy . in short , narrower and symmetric lines
are not observed to take on a relativistic shape due to photon pile - up distortions .
the result that pile - up generally tends to produce _ falsely narrow _ lines can largely be understood in simple terms . at high flux levels ,
photon pile - up will cause a ccd to register some low energy events as having a higher energy .
_ photon pile - up adds to the high energy continuum_. adding extra flux in the fe k band has the effect of hiding the true profile of a relativistic line : the red wing blends with the continuum and the ( relatively ) narrow blue wing is all that remains .
the details of the input spectrum and the nature of a given ccd camera ( its effective area curve , its effective frame time , how many event boxes tile an extraction region ) are likely important .
this may partially explain why slightly different results are obtained from simulated epic - pn `` timing mode '' spectra of neutron stars .
the trends seen in figure 4 are generally observed whenever pile - up is important . a broad range of incident spectra peak at approximately 1.01.5 kev , as do many detector effective area curves .
pile - up of photons to twice that energy , combined with a typical drop in effective area at about 2 kev due abrupt changes in the mirror reflectivity and also si absorption in the ccd , causes a false excess in the 23 kev band . depending on the specifics of the detector area curve
, this excess may be structured .
pile - up also adds appreciably through and above the fe k band , causing the power - law to be falsely hard and falsely high in flux .
this is seen as a steadily increasing excess above 7 kev in the data / model ratio shown in figure 4 .
the flux excess is greater when pile - up is more severe .
when fit with typical spectral models , the flux excess at 23 kev and above 7 kev act to create a false flux deficit just below the line energy , which is again more severe as pile - up becomes more severe . owing to the fact that more flux redistribution occurs when pile - up is more severe , the full width of relativistic lines becomes increasingly difficult to measure accurately . as shown in figure 7
, severe photon pile - up can have a strong impact on efforts to study thermal emission from the accretion disk .
depending on the specific detector and the input flux level ( again , traced by @xmath16 ) , pile - up can falsely reduce the flux in the disk continuum by a factor of a few .
the radius inferred from fits with the `` diskbb '' model depends on the square root of the flux normalization .
more recent and physical disk models have more parameters , but still measure the flux to determine an innermost emission radius , and can be expected to be affected in the same degree .
the results shown in figure 7 depict the results of fits to simulated `` high / soft '' state spectra , but consistent results are also obtained in fits to the continuum in simulated spectra of other states ( see the bottom panel in figure 6 ) .
this effect may also be understood relatively simply .
spectra of bright x - ray binaries tend to peak at or near to the peak of the effective area curve for x - ray telescopes with ccd spectrometers .
this is typically also the energy range in which the thermal disk continuum dominates .
when pile - up is important , then , low energy flux is preferentially lost from the accretion disk component due to grade migration and flux redistribution .
it should be noted that even the baseline `` high / soft '' state flux levels are actually quite high ( see table 1 ) .
simulations with successively higher values of the grade migration parameter @xmath16 trace successively higher flux levels
. only the _
suzaku_/xis with a 1/4 window and 0.3 second burst option , and the _ swift_/xrt in `` windowed timing '' mode , are even marginally able to cope with this flux level . even for these detectors and modes , strong mitigations ( e.g. annular extraction regions with large inner radii ) would be required to recover accurate spectral parameters .
the non - response of the other detectors and modes to increases in @xmath16 ( see figure 7 ) merely indicates that the limits of the simulations are reached .
the flux decrements indicated in figure 7 for the _ xmm - newton_/epic mos in `` full frame '' mode , epic - pn in `` timing mode '' , and the _ chandra_/acis in `` continuous clocking '' mode , should not be taken as realistic estimates of observed flux decrements .
the results of the numerous simulations and fitting exercises detailed above suggest that severe photon pile - up affects relativistic disk spectra in clear and predictable ways : redistributing the low energy continuum to the high energy portion of the spectrum causes relativistic emission lines to become _ falsely narrow _ , giving a _ falsely high _
value for the inner disk radius . in turn
, this means that estimates of black holes spin parameters based on the inner disk radius would be _ falsely low_. in spectra where pile - up may be important and the success of mitigation is uncertain , emission lines give _ upper limits _ on the radius of the accretion disk , and _ lower limits _ on the value of the black hole spin parameter .
the same trends are also observed in the case of neutron stars , both when relativistic and simple narrow gaussian line functions are assumed .
redistributing the low energy continuum to the high energy portion of the spectrum has exactly the opposite impact on disk continua .
the modified disk continuum gives _
falsely low _ values of the inner disk radius , equating to _ falsely high _ values of black hole spin based on that radius .
therefore , when pile - up may be important , and/or when efforts to mitigate pile - up may not have been entirely successful , radii inferred from continuum fits are _ lower limits _ and inferred spin parameters are _ upper limits_. when two diagnostics are skewed in the same sense , it is difficult to detect a bias , and a potential check on derived quantities is lost .
the results of the simple exercise undertaken in this paper are therefore fortuitous : photon pile - up distorts relativistic disk lines and the disk continuum in _ opposing _ ways .
fitting both lines and the disk continuum with relativistic spectral models and either checking that the radii agree , or explicitly requiring agreement in the fit ( see miller et al .
2009 ) , will help to derive results that are robust against pile - up distortions .
this procedure may be more difficult in the case of neutron stars , however , because the disk continuum is typically less distinct in neutron star spectra .
efforts to understand the inner accretion flow geometry through correlations between the flux in a disk line and the ionizing continuum are thus also impacted by photon pile - up . a particularly interesting explanation of the flux trends seen in seyfert-1 agn such as mcg-6 - 30 - 15 ( miniutti & fabian 2004 ) and ngc 4051 ( ponti et al.2006 ) is that gravitational light bending is altering the flux that impinges on the disk from a power - law source of ionizing radiation . in observations of such sources ,
photon pile up _ is not _ a concern , but it might be a concern if ccd spectometers were to monitor a stellar - mass accretor intensively .
it is worth noting that current evidence of gravitational light bending in stellar - mass black holes is drawn from gas spectrometer data that is unaffected by pile - up ( miniutti , fabian , & miller 2004 ; rossi et al .
2005 ) .
two papers have recently commented on the possibility that photon pile - up might create falsely broad emission lines in black hole spectra . in the first example , _
suzaku_/xis spectra of gx 339@xmath14 in an `` intermediate '' state were extracted from different annuli and compared ( yamada et al .
spectra from larger annuli ( presumably suffering from less pile - up ) are not found to strongly require black hole spin .
yamada et al .
( 2009 ) suggest that pile - up may influence the line shape , but fitting results are consistent with a very broad line and a spinning black hole at the 90% level of confidence ( see miller et al .
the relatively small number of counts in the wings of the psf can partially account for the lack of a strong spin requirement and lower statistical certainty .
a separate but equally important issue is that the disk reflection model used by yamada et al .
( 2009 ) was not convolved with the line element expected for emission from the inner disk ; this is physically inconsistent in that it implies a stationary , non - orbiting reflector .
the immediate effect is to falsely add to the continuum in the vicinity of the emission line , falsely narrowing the line .
recent work by done & diaz trigo ( 2010 ) examined _ xmm - newton _ spectra of gx 339@xmath14 in a `` low / hard '' spectral state .
spectra from the mos cameras were found to suffer from pile - up , even when extracting counts in annuli , whereas the epic - pn `` timing '' mode spectra were claimed to be free of distortions from photon pile - up .
the pn spectra were also found to deliver narrower line profiles than the mos spectra , implying a larger inner disk radius compared to values reported in prior work ( e.g. miller et al .
2004 ; reis et al.2008 ; also see wilkinson & uttley 2009 )
. repeating the mos extraction exactly as detailed in reis et al .
( 2008 ) , it is apparent that the disk line profile does not vary with the inner radius of annular extraction regions ( see figure 8) .
the line profile in the annulus identified by done & diaz trigo ( 2009 ) as being largely free of pile - up , closely matches the line profiles seen in spectra extracted from smaller annuli , when each spectrum is allowed to have its own continuum .
this means that the radii measured in the mos spectra are not distorted by photon pile - up , but that the radii derived in fits to the pn spectra are distorted .
this conclusion is echoed by our simulations .
the results detailed above strongly suggest that the fact of a narrower line profile in the pn means that it is actually piled - up and does not measure the true line width . indeed
, the departure of 23 @xmath46 shown in figure 2 is commensurate with the radius difference between fits to the mos spectra reported by miller et al .
( 2006 ) and reis et al .
( 2008 ) , and the pn spectrum as fit by wilkinson & uttley ( 2009 ) .
the data / model ratio in figure 7 of done & diaz trigo ( 2009 ) shows a flux excess at 2 kev and a flux decrement between 3 kev and the fe k line , very similar to the trends shown in figure 4 in this paper .
the observed spectra and our simulation results both show that photon pile - up makes lines artificially narrow .
following done & diaz trigo ( 2010 ) , ng et al .
( 2010 ) have analyzed a number of spectra of neutron star low - mass x - ray binaries , observed with the _ xmm - newton_/epic - pn camera in `` timing '' mode .
their analysis suggests that the spectra suffer from a degree of photon pile - up .
when the center of the psf is excluded , the line profiles are found to be more consistent with gaussian profiles .
ng et al.(2010 ) conclude that pile - up acted to falsely create skewed , relativistic line profiles .
this conclusion runs counter to the results of our simulations , which show that ( 1 ) severe pile - up acts to falsely narrow both relativistic and simple gaussian lines , not to falsely broaden lines ; and ( 2 ) pile - up distortions to typical neutron star line spectra are not expected to be extreme ( see figure 5 ) for sources with flux levels similar to serpens x-1 , except for _ chandra_/acis `` continuous clocking '' mode .
independent _ suzaku _ and _ xmm - newton _ observations of serpens x-1 reveal relativistic fe lines with a clear red wing ( see figure 9 ; also see bhattacharyya & strohmayer 2007 , cackett et al .
2008 , cackett et al .
it is already clear that _ suzaku _ line profiles in this source and similar sources do not depend on the inner radius of event extraction annuli ( cackett et al .
2010 ; see figure 11 ) the relativistic profiles do not result from photon pile - up .
the nature of the line seen with _ suzaku _ strongly suggests that the red wing seen in the _ xmm - newton _ spectrum is due to dynamical broadening and red - shifting , not photon pile - up .
we re - reduced the _ xmm - newton _ pn observation of serpens x-1 exactly as detailed in bhattacharyya & strohmayer ( 2007 ) , and extracted spectra including and excluding the center of the psf .
the primary effect of excluding the center of the psf is merely to reduce the number of photons in the resultant spectrum , thereby lowering the sensitivity to the point that the red wing can not be detected .
the line is neither narrower nor weaker when the center of the psf is excluded
it is simply less defined ( see figure 10 ) .
ng et al.(2010 ) do not present direct comparisons of line profiles in spectra obtained in dfferent extraction regions . given that : ( 1 ) _
suzaku _ spectra of neutron stars reveal relativistic lines that are clearly not due to photon pile - up distortions , ( 2 ) a number of _ suzaku _ and _ xmm - newton _ line profiles are remarkably similar ( e.g. sax j1808.6@xmath13808 , see figure 3 in cackett et al . 2010 ;
gx 349@xmath112 , see figure 9 in cackett et al.2010 ) , ( 3 ) the effect of excluding the center of the psf in the _ xmm - newton _ spectrum of serpens x-1 is merely to reduce definition in the fe line ( not its width nor its strength ) , and ( 4 ) our simulations show that severe pile - up acts to artificially _ narrow _ both relativistic and simple gaussian lines , it is likely that dramatic reductions in sensitivity drove the results obtained by ng et al.(2010 ) and led to faulty conclusions regarding the ability of pile - up to create false relativistic line profiles .
the fact that sensitive spectra are required to detect the red wings of relativistic line profiles is not a revelation .
this fact fueled the requirement of minimum sensitivity thresholds in recent surveys of relativistic lines in seyfert agn ( see nandra et al .
recent work on neutron star spectra makes the necessity of sensitive spectra even more clear .
deep observations of a source such as gx 349@xmath112 with _ suzaku _ reveal a relativistic line profile , whereas short observations with the _ chandra_/hetgs which has a lower collecting area only recovers the narrower blue wing of the line profile ( see figure 3 in cackett et al .
spectrum is fully consistent with the relativistic line profile found using _
, it simply does not have the sensitivity needed to detect the red wing against the continuum .
the above discussion is mostly focused on _ xmm - newton _ and _ suzaku _ spectroscopy , because the large effective area of these missions ensures that they are well - suited to relativistic spectroscopy .
_ swift _ is highly flexible , but its smaller collecting area and short observations mean that typical spectra lack the sensitivity to detect and measure relativistic disk lines well .
however , _
swift _ is very well - suited to black hole spin measurements using the disk continuum , and our results suggest that even in `` windowed timing '' mode , annular extraction regions are needed to avoid spectral distortions due to photon pile - up ( see figure 6 ) .
brief discussions of pile - up and excluding the central portion of the psf in `` windowed timing '' mode are given in rykoff et al .
( 2007 ) and rykoff , cackett , & miller ( 2010 ) .
two combinations of detectors and modes are not treated in this work _ because _ they represent effective means of preventing photon pile - up . as noted previously , the `` burst '' mode of the epic - pn camera aboard _
xmm - newton _ may provide the _
best _ means of preventing photon pile - up , as its short read - out time is suited to extremely bright sources .
this comes at a cost in sensitivity , however , as `` burst '' mode has a livetime fraction of just 0.03 .
pile - up can also be avoided by observing with the _ chandra_/hetgs , which disperses a spectrum onto the acis spectrometer .
pile - up can be further avoided by reading - out the dispersed spectrum in `` continuous clocking '' mode . _
chandra_/hetgs observations of the stellar - mass black hole gx 339@xmath14 in an `` intermediate '' state measured an inner disk radius of @xmath52 ( miller et al .
this analysis used the same relativistically - blurred reflection model applied to a later observation in the same state using _ suzaku _ ( miller et al .
analysis of the later observation found a spin parameter of @xmath53 , correspinding to @xmath54 ( bardeen , press , & tuekolsky 1972 ) .
the radius values measured in gx 339@xmath14 are consistent in the two separate observations , using very different detectors .
similarly , _
chandra_/hetgs observations of the neutron star x - ray binary 4u 1705@xmath144 find a very broad line , with a gaussian width ( fwhm ) of @xmath55 kev ; when fit with a relativistic diskline model , a radius of @xmath56 is measured ( di salvo et al .
2005 ) . here again
, the line properties found using grating spectra which avoid severe pile - up are consistent with those found using _
suzaku _ and _ xmm - newton _
( reis , fabian , & young 2009 , di salvo et al .
2009 , cackett et al .
2010 ) , and inconsistent with the much narrower lines found by ng et al .
( 2001 ) when extracting only the wings of _ xmm - newton_/epic - pn `` timing '' mode spectra .
an _ xmm - newton_/epic - pn `` burst '' mode observation of grs 1915@xmath11105 in a `` plateau '' state ( similar to the low / hard state in most black holes ) may also validate the results of our simulations and the discussion above .
martocchia et al .
( 2006 ) observed grs 1915@xmath11105 twice in the `` plateau '' state once in `` timing '' mode and once in `` burst mode '' at consistent flux levels .
the `` timing '' mode spectrum may suffer from modest distortions due to photon pile - up .
the power - law index was measured to be harder than in the `` burst '' mode observation ( @xmath57 versus @xmath58 ) , indicative of possible pile - up .
the iron line is found to be fairly narrow in the `` timing '' mode observation : the line is only visible above @xmath356.2 kev when the spectrum is fit with a power - law , the inner radius is constrained to be greater than 240 @xmath46 , and a reflection fraction of @xmath59 is measured .
in contrast , when observed using `` burst '' mode , the line is visible down to 5 kev ( or lower ) relative to a simple continuum , a radius of less than @xmath60 is required , and a reflection fraction of @xmath61 is measured ( martocchia et al .
( this _ xmm - newton_/epic - pn `` burst '' mode spectrum is broadly conistent with a _ suzaku _
observation of grs 1915@xmath11105 in the `` plateau '' state . in that spectrum ,
the line is also broad , and consistent with the isco ; see blum et al .
these results square with a central result of the simulations presented in this paper : photon pile - up acts to make relativistic lines appear to be falsely narrow .
though an investigation is beyond the scope of this paper , it should be noted that efforts to mitigate pile - up can also distort spectra .
the psf of an x - ray telescope is energy - dependent .
extracting events in annuli may avoid the piled - up core of a given psf , but the spectrum is then derived from a portion of the psf that is not calibrated as well as the core .
moreover , depending on the psf , annular regions may only extract a vanishing fraction of the incident photon flux , complicating the detection of weak spectral lines .
relativistic disk lines are often 1020% features above the continuum ; extracting spectra from regions of the psf from which the energy and flux calibration are not known to much better than 10% could easily make it difficult to recover the details of a given line profile .
depending on how much flux is excluded in the annulus , limited photon statistics will also serve to complicate the detection of line asymmetry .
finally , it is worth emphasizing that our results depend on a specific combination of mirror technology and detector technology .
presently , ccd spectrometers sit at the focus of gold foil or gold - coated mirors ; this partially accounts for an overall telescope efficiency curve that is highest below 2 kev .
new mirror technology , such as si pore optics ( e.g. beijersbergen , m. , et al . , 2004 ) , may produce different ccd photon pile - up effects , and distortions to disk continua and disk lines may not longer skew in the opposite sense .
extensive simulations of photon pile - up for a number of spectral forms and detector parameters suggest that severe pile - up can distort spectroscopic signatures of the inner accretion disk , in largely predictable ways .
the results of our work can be summarized as follows : + @xmath62 the degree of photon pile - up in a spectrum depends on the input flux level from a source and the effective area of the telesecope , but it is also depends on the number of event boxes tiling the psf and the frame time of the ccd .
@xmath62 tiling the psf with many event boxes and short frame times are two independent means of reducing photon pile - up .
however , when short frame times are achieved by changing the size of an event box and under - sampling the psf , pile - up mitigation is partially compromised .
@xmath62 flux redistribution due to photon pile - up causes emission lines whether relativistic or symmetric and intrinsically narrow to be observed as falsely narrow .
measured inner disk radii are then falsely large , and inferred black hole spin parameters are falsely low .
relevent observed spectra , though small in number , support this finding .
@xmath62 grade migration due to photon pile - up causes the low energy spectrum principally the disk component to have a falsely low flux .
measured inner disk radii are therefore falsely small and inferred spin parameters are falsely high .
@xmath62 deriving inner disk radii and/or black hole spin parameters by linking that parameter to a joint value in both disk continuum and disk reflection models may yield more robust results when photon pile - up can not be avoided or successfully mitigated .
we wish to acknowledge the referee , keith jahoda , for a thoughtful review of this work and for comments that improved the paper . we thank john davis , matthias ehle , lothar strder , jrn wilms , and maria diaz - trigo , mark reynolds , dipankar maitra , and tiziana di salvo , and giorgio matt for helpful discussions .
lllll parameter & very high / intm . & low / hard & high / soft & z / atoll + @xmath63 & 0.5 & 0.5 & 0.5 & 0.5 + @xmath64 ( kev ) & 0.76 & 0.18 & 0.97 & 1.21 + @xmath65 & 2300.0 & 160000.0 & 7050.0 & 103.0 + @xmath66 ( kev ) & & & & 2.28 + @xmath67 ( @xmath68 ) & & & & 4.9 + @xmath69 & 2.60 & 1.60 & 2.48 & 3.60 + @xmath70 & 2.20 & 0.50 & 3.38 & 0.18 + @xmath71 ( kev ) & 6.7 & 6.7 & & 6.7 + @xmath72 & 3.0 & 3.0 & & 3.0 + @xmath73 & 6.0 & 6.0 & & 6.0 + @xmath74 ( deg . ) & 30.0 & 30.0 & & 30.0 + @xmath75 & 7.7 & 8.5 & & 8.0 + llllllllll & @xmath77 & radius@xmath78 & pixel size@xmath79 & event box@xmath80 & @xmath81 & regions@xmath82 + & ( @xmath83 s ) & ( arcsec ) & ( arcsec ) & ( pixels ) & ( @xmath83 s ) & + & 2.85 & 2.5 & 0.5 & 3x3 & 8.55 & 3.3 + & 1.77 & 9.0 & 2.36 & 10x7 & 1.77 & 1.1 + & 0.03 & 120 & 4.1 & 30x3 & 0.09 & 1.0 + & 2600 & 20 & 1.1 & 3x3 & 2600 & 115.3 + & 300 & 60.0 & 1.0 & 3x3 & 300.0 & 1256.0 + | x - ray charge - coupled devices ( ccds ) are the workhorse detectors of modern x - ray astronomy . typically covering the 0.310.0 kev energy range , ccds
are able to detect photoelectric absorption edges and k shell lines from most abundant metals .
new ccds also offer resolutions of 3050 ( e/@xmath0e ) , which is sufficient to detect lines in hot plasmas and to resolve many lines shaped by dynamical processes in accretion flows .
the spectral capabilities of x - ray ccds have been particularly important in detecting relativistic emission lines from the inner disks around accreting neutron stars and black holes .
one drawback of x - ray ccds is that spectra can be distorted by photon `` pile - up '' , wherein two or more photons may be registered as a single event during one frame time .
we have conducted a large number of simulations using a statistical model of photon pile - up to assess its impacts on relativistic disk line and continuum spectra from stellar - mass black holes and neutron stars .
the simulations cover the range of current x - ray ccd spectrometers and operational modes typically used to observe neutron stars and black holes in x - ray binaries .
our results suggest that severe photon pile - up acts to falsely narrow emission lines , leading to falsely large disk radii and falsely low spin values .
in contrast , our simulations suggest that disk continua affected by severe pile - up are measured to have falsely low flux values , leading to falsely small radii and falsely high spin values . the results of these simulations and existing data appear to suggest that relativistic disk spectroscopy is generally robust against pile - up when this effect is modest . |
a polynomial map @xmath0 comes equipped with a rich topological structure .
the fibres @xmath8 form a family of complex analytic curves parametrised by points in the base @xmath9 .
the generic fibre is a smooth connected topological surface of genus @xmath2 .
a finite set @xmath10 of values in the base have exceptional fibres @xmath11 with euler characteristic strictly greater than the generic euler characteristic . on the complement of @xmath12
the map @xmath5 defines a surface fibration with structure group the mapping class group of the generic fibre , also known as its geometric monodromy group .
the action of this monodromy group on the homology of the generic fibre is a well - studied topological invariant .
the object of study in this paper is concerned with the geometric monodromy left over after passing to finite covers .
for example , if the homological monodromy is trivial , or finite so that a cover is trivial , there can be interesting geometric monodromy remaining . an effective way to study the geometric monodromy group uses the moduli space @xmath13 of genus @xmath2 curves
this is rationally the classifying space of the mapping class group so it classifies bundles on the complement of the exceptional set . a polynomial map @xmath0 gives a map from the complement of its exceptional set @xmath14 by sending @xmath15 to its fibre @xmath8 a smooth genus @xmath2 curve .
finite covers of fibrations arise when one tries to extend over the exceptional set to get a map @xmath16 to the compactification which is the moduli space of _ stable curves _ of arithmetic genus @xmath2 .
a stable curve satisfies @xmath17 , it has at worst isolated nodal singularities , and any genus zero irreducible component contains at least three singular points @xcite .
typically , there will be exceptional fibres of @xmath5 which are not stable .
the main technical issue of how to extend the map @xmath18 across the set @xmath12 of exceptional values is dealt with in section [ sec : stab ] using the process of _ stable reduction _ of a family of curves .
the stable reduction around a singular curve @xmath19 in a family @xmath20 canonically produces a stable curve to replace @xmath19 and induces a canonical topological decomposition along circles of the generic curve in the family .
the stable curve corresponds to pinching along these circles .
the stable reduction is obtained by pulling back the ( singular ) fibration @xmath0 over a new base @xmath21 so that the fibres are reasonably behaved . in this case reasonably behaved means that the local monodromy around exceptional values is given by dehn twists around disjoint non - parallel curves , which is equivalent to the homological monodromy being unipotent .
this condition on the monodromy is equivalent to the fibres having at worst isolated nodal singularities .
if we also blow down any genus zero component of a fibre containing fewer than three singular points , in which case the fibre becomes a stable curve , then we have produced a map from @xmath22 , and a corresponding rational homology class @xmath23/d\subset h_2(\overline{\mathcal{m}}_g)$ ] where @xmath24 is the degree of the cover @xmath21 .
it is clear that the homology class depends only upon the topological type of the stable reduction and that it is independent of the finite cover .
the process of stable reduction around an exceptional fibre @xmath19 in a family @xmath20 is topologically equivalent to a canonical topological decomposition of the generic curve in the family along circles , together with the degree of the cover .
this process , and hence the decomposition , depends only on a neighbourhood of @xmath19 in the family .
the neighbourhood is encoded in its 3-manifold link , well - defined for arbitrarily small @xmath25 , @xmath26 which fibres over @xmath27 .
the 3-manifold @xmath28 possesses a canonical decomposition known as the jsj decomposition @xcite .
the jsj decomposition canonically decomposes @xmath28 along two - dimensional tori .
the intersection of the tori with a fibre @xmath20 of @xmath28 induces a canonical decomposition of the fibre along circles .
[ th : decomps ] given a polynomial map @xmath0 , in a neighbourhood of an exceptional fibre the two decompositions of the generic fibre along circles , arising from stable reduction and the jsj decomposition , coincide .
we have stated the theorem in its simplest possible form although more is true .
the jsj decomposition also detects further information contained in the stable reduction - the finite cover required during the base change , and the singularities at nodal points of the stable curve .
the singularities at nodal points were only mentioned implicitly in the statement that we should blow down some genus zero components of a fibre .
see section [ sec : stab ] for precise details .
the theorem can also be extended to the situation of a map @xmath0 in the presence of a curve @xmath1 that intersects all fibres .
the stronger version of the theorem with proof appears as theorem [ th : decomps2 ] in section [ sec : stab ] .
this generalisation is required in our applications and is explained below .
the theorem is simply an interpretation of the methods of eisenbud and neumann in @xcite which applies the jsj decomposition of the 3-manifold boundary neighbourhood of an exceptional fibre to piece together the geometric monodromy .
this ends up being perfectly suited to stable reduction .
the main purpose of the theorem is not only to give a topological construction of the rational homology class associated to @xmath0 , but also an efficient way to calculate . in section [ sec : blowups ]
we show the long process involved in stable reduction and in section [ sec : splice ] we show how theorem [ th : decomps ] makes the calculations much more efficient .
+ a curve @xmath1 marks points on each fibre of @xmath0 .
the generic fibre must intersect @xmath6 transversally in @xmath3 points , and in particular the set of exceptional values may enlarge to @xmath29 . as in the case when @xmath6 is empty , we can characterise an exceptional fibre @xmath11 by the property that the euler characteristic @xmath30 is strictly greater than the generic euler characteristic @xmath31 .
the curve @xmath6 decomposes into irreducible components @xmath32 , and we may assume that no @xmath33 lies in a fibre of @xmath5 , since we can just remove such components , so @xmath34 . if @xmath35 then the @xmath33 are sections of the map @xmath0 and the generic fibre @xmath8 is an element of @xmath36 , the moduli space of genus @xmath2 curves with @xmath3 labeled points .
the geometric monodromy fixes the labeled points and @xmath36 is rationally the classifying space of the subgroup of the mapping class group that fixes the @xmath3 labeled points .
when @xmath37 there is not a well - defined labeling of the @xmath3 points of intersection on a generic fibre . in other words ,
the monodromy acts non - trivially on the intersection of @xmath6 with the generic fibre , and in particular its action on homology is not unipotent .
there is a finite cover @xmath21 with monodromy acting trivially on the intersection of the pull - back of @xmath6 with the generic fibre , and hence the map @xmath38 possesses @xmath3 sections
. thus stable reduction , which ensures unipotent monodromy around exceptional fibres , is also used to enable a labeling of the @xmath3 points of intersection . as before we get a homology class @xmath23/d\subset h_2({\overline{\mathcal{m}}_{g , n}})$ ] where @xmath24 is the degree of the cover @xmath21 .
the compactification @xmath4 consists of stable curves of arithmetic genus @xmath2 and @xmath3 labeled points .
generalising the case @xmath39 to @xmath40 , a stable curve has at worst isolated nodal singularities distinct from the labeled points , and any genus zero irreducible component contains at least three points made up of labeled points and singular points .
there is no restriction on genus .
theorem [ th : decomps ] generalises to this situation in which the three - manifold now contains a link @xmath41 .
since the main purpose of these ideas is to do be able to do calculations , this paper contains examples .
a huge source of examples is supplied by polynomials in two variables @xmath42 . in this case
the @xmath3 points on a generic fibre are its points at infinity .
the stable reduction of a family is obtained by pulling back @xmath42 over a new base @xmath43 so that the homological monodromy around exceptional fibres and around infinity is unipotent .
this fits in with the picture above if we compactify @xmath44 to @xmath45 and extend @xmath5 to a map from @xmath45 to @xmath46 .
note that we can not simply choose @xmath47 since @xmath5 does not extend there .
section [ sec : stab ] gives details .
thus , a polynomial @xmath7 gives a ( multi - valued ) map from the complement of its exceptional set in its base @xmath48 to @xmath36 and stable reduction allows us to extend over the exceptional set and at the same time choose a labeling of the @xmath3 points at infinity to get a rational homology class in @xmath49 .
the jsj decomposition has previously been applied to polynomials in two variables in @xcite to introduce a powerful topological invariant of a polynomial @xmath7 , its _ link at infinity _ which is a link in the 3-sphere given by the intersection of a generic fibre of @xmath5 with the boundary of a very large ball in @xmath44 .
this is analogous to the link of a complex plane curve singularity obtained by intersecting the curve with the boundary of a very small ball around the singularity .
the link at infinity yields topological information about the polynomial such as the genus of the generic curve in the family , the monodromy around exceptional curves in the family , and possible exceptional curves that can arise in the family .
part of this paper addresses the question of how much the link at infinity of a polynomial can say about its second rational homology class in @xmath4 .
an effective description of the homology class associated to @xmath0 uses rational numbers obtained by evaluating the homology class on rational cohomology classes @xmath50 these cohomology classes are defined in section [ sec : coh ] where explicit calculations of the rational numbers are given .
the technique of stable reduction here is taken from @xcite .
as described in the introduction , generic fibres of a polynomial map @xmath0 with @xmath1 are smooth genus @xmath2 curves containing @xmath3 distinct points .
although exceptional fibres are not necessarily stable , the existence of the compactification @xmath4 means that as fibres approach an exceptional fibre they converge to a stable curve , so we may replace the exceptional fibre with that stable curve .
this process , known as stable reduction , involves passing to a cover @xmath21 in which the local homological monodromy around each exceptional fibre is unipotent .
the pull - back map @xmath38 will have the two important properties that every fibre is a stable curve , and that @xmath51 is the image of @xmath3 sections @xmath52 .
such a cover exists by the local monodromy theorem which asserts that there exists a power of the local homological monodromy around an exceptional fibre which is unipotent with exponent 2 , i.e. @xmath53 . there are many such covers since any power of a unipotent element remains unipotent , nevertheless the way in which we define the rational second homology class in the moduli space of stable curves is independent of the cover .
we decompose the cover into cyclic covers of prime degree since they are most easily understood .
the process is best demonstrated via explicit calculations so this is what we will do next .
the reader can consult @xcite for further details . given a polynomial @xmath42 compactify to a morphism @xmath54 .
this is achieved by first compactifying to @xmath55 .
the map @xmath56 is not a morphism ( unless it was trivially a coordinate to begin with ) or in other words there are points in @xmath57 on which @xmath56 takes on every value and hence is ill - defined .
thus we repeatedly blow up @xmath58 until the transform @xmath59 of @xmath5 is well - defined , and in particular there exist rational curves @xmath60 so that @xmath61 is onto , sometimes referred to as horizontal curves or quasi - sections .
if the degree of @xmath61 is 1 for each @xmath62 then @xmath63 and the map @xmath54 has @xmath3 sections and the generic fibres become points of @xmath36 . if the degree of @xmath61 is @xmath64 for some @xmath62 then the inverse map @xmath65 is a multiply defined @xmath66 map .
we must take a finite cover consisting of different choices of labeling the points at infinity meeting @xmath33 to get @xmath67 sections .
we arrange this at the same time that we do stable reduction around a non - stable fibre .
note that the fibre @xmath68 is never stable because it has components of multiplicity greater than 1 so we need to replace it .
let @xmath69 .
the generic fibre has genus 4 and 1 point at infinity .
extend to @xmath55 and resolve the indeterminate points of @xmath56 ( explicitly , change coordinates to @xmath70 and blow up multiple times , or do a single toric blow - up @xmath71 , followed by a change of coordinates @xmath72 and another toric blow - up @xmath73 followed by another change of coordinates and multiple blow - up ) to get the curve over infinity represented in figure [ fig : res ] by its dual graph . over infinity ,
height=75 ] in figure [ fig : res ] the nodes denote irreducible components , which are rational curves unless a genus is specified , the edges denote intersections between components , the self - intersection numbers are given in brackets , and the other numbers denote the multiplicity of @xmath59 on each component .
the multiplicities and intersection numbers are related by the fact that a component has zero intersection with the fibre at infinity .
the zero multiplicity on the right - most component denotes the fact that it maps surjectively onto @xmath46 , and in this case the map has degree 1 so it is a section .
the dark node corresponds to the proper transform of the line at infinity in @xmath57 .
the euler characteristic of the generic fibre can be calculated from the fibre over infinity by removing the points of intersection of irreducible components and taking the euler characteristic .
thus a node of multiplicity @xmath74 and valency @xmath75 is replaced by @xmath74 copies of a @xmath75-punctured 2-sphere .
the valency 2 genus 0 components contribute 0 to this sum which is @xmath76 so the generic fibre is a genus 4 curve minus a point at infinity .
each fibre meets the section exactly once .
the fibre over infinity is not stable since most of its components have multiplicity greater than 1 .
all other fibres are stable in this example because the exceptional fibres , which are precisely the singular fibres , contain only nodal singularities .
we will demonstrate the process of stable reduction which will replace the fibre at infinity by a stable curve .
break the branched cover up into a sequence of covers of prime degree .
take a 3 to 1 branched cover of the base totally ramified at infinity .
the fibre over infinity in the pull - back fibration is not smooth since the original fibre over infinity is not smooth .
for example , the cover @xmath77 is singular above the singularity @xmath78 , and we can resolve the cover using two intersecting @xmath79 curves . more generally we will need to consider @xmath80 for a @xmath5-fold cover near the intersection of curves of multiplicities @xmath81 and @xmath82 .
it turns out that the normalisation of the pull - back fibration is the same as ramifying along curves over infinity of multiplicity not divisible by 3 .
for example , in figure [ fig:3fold ] ramify on the curves of multiplicities 8 , 10 , 8 and 10 .
the rightmost curve of multiplicity 6 in figure [ fig:3fold ] pulls back to a 3-fold cover branched at its two intersection points with the multiplicity 8 and 10 curves .
its multiplicity is divided by 3 . for the intersecting multiplicity 10 and 8 curves the 3-fold
cover ramified along these two is singular .
it is a hirzebruch - jung singularity which can be resolved . in this case
it can be resolved by blowing up between the two curves downstairs to get a multiplicity 18 curve which is divisible by 3 , so we are reduced to the previous case of a 3-fold cover branched over 2 points , leaving a multiplicity 6 curve .
the multiplicity 12 curve is branched over its two intersection points with multiplicity 8 and 10 and hence 3 copies of the multiplicity 6 curve live in the cover , sharing multiplicity 2 each .
figure [ fig : stablered ] shows the fibre over infinity of a sequence of base changes of prime degree combining to give a branched cover @xmath83 that pulls back the fibration so that the fibre over infinity is a stable curve .
often the resolution of a hirzebruch - jung singularity contains a string of curves that can be blown down .
we have not shown the full calculation . only once in figure
[ fig : stablered ] have we shown the blow - downs preferring to combine steps and hide many intermediate rational curves .
the process ends with the stable curve of arithmetic genus 4 given by two genus 1 curves attached to a genus 2 curve as in figure [ fig : stableinf ] .
this stable curve is canonically associated to the family in that any other cover leading to stable reduction will give the same stable curve over infinity .
in the final diagram of figure [ fig : stablered ] the cross with a 7 on an edge means that the intersection of the genus 1 and genus 2 curves is a @xmath84 quotient singularity obtained by blowing down a string of six @xmath79 curves . in the family of curves , parametrised by @xmath85 ,
say , with @xmath86 corresponding to infinity , the nodal singularity is given by @xmath87
. often we stop at the previous step containing strings of @xmath79 curves , and hence semi - stable fibres , to avoid singularities upstairs .
of course if we take a further branched cover of degree @xmath24 say , the nodal singularity would be @xmath88 .
it is clear that the factor of 7 is essential and the extra factor of @xmath24 is the result of taking a less efficient cover .
the essential number is really @xmath89 , an invariant of the process .
similarly , it is @xmath90th of the stable curve in figure [ fig : stableinf ] that strictly replaces the fibre over infinity .
this reflects the fact that the moduli space is an orbifold . in the example
, the curve @xmath1 is already a section . more generally , when there are irreducible components of @xmath6 that are not sections we need to arrange to have sections in the cover during the stable reduction process .
in each step the @xmath5-fold cover is totally ramified at infinity and we can freely specify all other ramification .
we choose some of the points of ramification away from infinity to coincide with the points of ramification of the map @xmath91 in such a way to remove the ramification .
note that when the base is @xmath46 as for polynomials of two variables , we can also choose the ramification away from infinity so that the cover of the base is again @xmath46 .
the result of stable reduction produces stable curves to replace non - stable exceptional fibres .
when these new stable curves are included into the family there are usually singularities at nodes of the stable curves .
the degree of the cover of the stable reduction must be remembered in order to get invariant quantities such as the homology class .
the example demonstrates an algorithm for stable reduction that works in general .
next we will prove theorem [ th : decomps ] and apply it to deduce the stable reduction more efficiently .
theorem [ th : decomps ] is a special case of a more general theorem which we will now describe .
consider a polynomial map @xmath0 and a curve @xmath1 .
for any exceptional fibre @xmath19 define the link of @xmath19 by the pair @xmath92 consisting of a 3-manifold @xmath28 containing a 1-manifold @xmath93 , where @xmath94 is a small enough neighbourhood of @xmath19 .
the jsj decomposition decomposes @xmath95 along finitely many tori .
the manifold @xmath95 fibres over the circle via @xmath5 , with fibre the complement of @xmath3 distinct points on the generic fibre of @xmath5 .
equipped with the more detailed description of stable reduction above , we are ready to restate theorem [ th : decomps ] in its full generality .
[ th : decomps2 ] the stable reduction of an exceptional fibre @xmath19 of a polynomial map @xmath0 with a curve @xmath1 is obtained as follows . denote the generic fibre of @xmath5 by @xmath96 and let @xmath97 be the link of @xmath19 .
then @xmath95 fibres over the circle with fibre @xmath98 and monodromy @xmath99 satisfying : \(i ) the jsj decomposition of @xmath95 along tori induces a decomposition @xmath100 of each fibre along circles .
\(ii ) the monodromy @xmath99 preserves the decomposition ( [ eq : decomp ] ) .
it has finite order on each piece @xmath101 and acts by a fractional twist along the decomposing circles .
\(iii ) the stable reduction of @xmath19 in the family uses the cover on which the lift of @xmath99 is trivial on each @xmath101 .
in particular , the degree of the cover is the lowest common multiple of the finite orders of @xmath99 on each @xmath101 .
\(iv ) the quotient singularities at nodes of the stable fibre are detected by the twists of @xmath99 along the decomposing circles .
_ consider the case of a fibre @xmath19 that is a nodal curve with isolated singularities and no ( -1 ) rational curves meeting fewer than three other components .
equivalently the multiplicity of each irreducible component is 1 and each rational curve contains at least two singular points
. more generally , there may be @xmath3 local sections around @xmath19 , and we blow up if necessary to separate intersection points of the sections on @xmath19
. then @xmath19 is semi - stable and the stable reduction process is unnecessary .
the nearby generic fibre decomposes @xmath102 along circles known as vanishing cycles , which correspond to the singular points when pinched .
the monodromy is given by one full positive twist around each vanishing cycle , so in particular it is trivial on each @xmath101 .
a string of @xmath103 rational curves , each containing exactly two singular points and hence of self - intersection ( -2 ) , has @xmath104 vanishing cycles , so the monodromy twists exactly @xmath104 times along the common homology cycle of each vanishing cycle .
if we blow down the string of ( -2 ) rational curves the fibre becomes a stable curve , and the node is a @xmath105 quotient singularity in the ambient surface .
the theorem states that the jsj decomposition of @xmath95 is the union of @xmath106 along the boundary tori , and the gluing of the boundary tori encodes the quotient singularities .
the tori correspond to intersections of neighbourhoods of irreducible components of @xmath19 .
the intersection of two components looks locally like @xmath107 and @xmath28 is locally @xmath108 which intersects along the torus @xmath109
. properties ( i ) and ( ii ) are proven in @xcite .
there it is shown that @xmath95 decomposes along tori into seifert - fibred pieces corresponding to the nodes of valency @xmath110 and of positive genus .
the decomposition coincides with the jsj decomposition so in particular it is canonical .
the fibration @xmath111 equips @xmath95 with an element of @xmath112 .
an element of @xmath112 is sometimes known as a _ multilink _ structure on @xmath93 since it assigns a multiplicity to each component by evaluating along a meridian circle .
any fibre @xmath96 of the map @xmath113 has boundary the multilink @xmath93 .
thus when the multiplicities of all components of @xmath93 are 1 , @xmath96 is a seifert surface for @xmath93 , and if a multiplicity of a component @xmath114 is @xmath64 then the boundary of @xmath96 near @xmath114 gives @xmath67 copies of @xmath114 .
if @xmath115 then one can attach a disk to the @xmath62th boundary component of @xmath96 which meets @xmath114 transversally .
the fibre @xmath96 decomposes as in ( [ eq : decomp ] ) and each @xmath101 is the fibre of a fibration of the corresponding seifert - fibred piece over the circle .
the monodromy decomposes and since the monodromy of any seifert - fibred piece is finite this gives ( ii ) .
using the definition of a stable curve in terms of its homological monodromy being unipotent , ( iii ) follows from ( i ) and ( ii ) , since on a cover where the lift of @xmath99 is trivial on each @xmath101 , only twisting along separating circles remains . in the cover ,
the twisting is integral , not merely fractional .
a precise description of how to calculate the fractional twisting of the monodromy along circles is given in @xcite .
we will show this in calculations below and thus verify ( iv ) .
the homology class of a polynomial map is trivial if and only if all generic fibres are isomorphic .
since the stable reduction of such a family is necessarily a trivial fibre bundle , a necessary condition for the generic fibres to be isomorphic is that the stable reduction of any exceptional the fibre is smooth .
theorem [ th : decomps2 ] shows that this occurs precisely when the dual graph of the exceptional fibre has only one node of valency @xmath110 since this is equivalent to the monodromy having finite order . applying this idea to the fibre over infinity of a polynomial @xmath42
gives a severe restriction on polynomials with isomorphic general fibres .
kaliman classified all polynomials @xmath42 with isomorphic generic fibres @xcite .
it seems likely that the classification can be reproven beginning from the restriction just described .
many calculations involving the dual graph require only some of the nodes .
examples are : \(i ) the euler characteristic of the generic fibre ; \(ii ) a description of branched covers of the dual graph ; \(iii ) determinants of intersection forms of subgraphs ; \(iv ) the canonical class supported on the dual graph .
+ an efficient dual graph hides valency 2 rational components , retaining components of valency @xmath116 or genus @xmath117 . an efficient dual graph for the fibre over infinity of @xmath69 is given in figure [ fig : splice ] .
genus 0 nodes.,height=75 ] we have labeled the multiplicity of the polynomial @xmath5 on each node .
the calculation of the euler characteristic of the generic fibre using the fibre at infinity gets a contribution of zero from the rational valency 2 nodes and hence one needs only the efficient dual graph . in figure
[ fig : splice ] a node of multiplicity @xmath74 and valency @xmath75 is replaced by @xmath74 copies of a @xmath75-punctured 2-sphere and the euler characteristic is @xmath76 so the generic fibre is a genus 4 curve minus a point at infinity . +
a cyclic branched cover along a singular fibre behaves quite generally as in the example in section [ sec : blowups ] .
a cyclic branched cover along a linear chain of genus zero components gives another linear chain of genus zero components .
the example in section [ sec : blowups ] shows this explicitly .
the most interesting behaviour occurs on components intersecting at least two other components , corresponding to nodes of valency @xmath110 in the dual graph , and on positive genus components . by simply understanding the behaviour of a cyclic cover on the nodes of valency @xmath110 and of positive genus
, we can much more efficiently take cyclic covers of splice diagrams , hiding all chains arising from hirzebruch - jung singularities . the long calculation shown in section [ sec : blowups ]
is substantially shortened .
+ an efficient dual graph that hides valency @xmath118 rational nodes , and equipped with enough information to reproduce the full dual graph , is known as a _ splice diagram . _ integer weights assigned to the graph supply the information required to reproduce the full dual graph and in particular the multiplicities of each node .
determinants of intersection matrices of subgraphs have the property that they are unchanged by blowing up the subgraph . in particular , to each edge of the efficient dual graph there is a well - defined edge determinant given by the determinant of the intersection matrix of nodes in the full dual graph along that edge , excluding the endpoints . the efficient dual graph is called a splice diagram when we use a set of weights on half edges .
the weight on a half edge is the determinant of the intersection matrix of the branch of the dual graph disconnected from the half edge if we cut along that edge .
figure [ fig : splice1 ] shows figure [ fig : splice ] with weights on half - edges replacing multiplicities on nodes .
the weights give back the whole dual graph .
in particular , to get the multiplicity of a node take a path from the node to an arrow and multiply all weights adjacent to , but not on , the path .
the black node has weights 3 and 2 adjacent to the path to the arrow so its multiplicity is @xmath119 .
the closest node to the arrow has weights 5 and 2 adjacent to its path so its multiplicity is @xmath120 .
when there are several arrows do this for each path joining a node to an arrow and add .
we will not go further into the calculation of the weights , nor will we emphasise the important fact that the splice diagram encodes a 3-manifold containing a link spliced together from seifert - fibred pieces .
see @xcite for full details . in this particular case the multiplicities in figure [ fig : splice ]
are determined by , and determine , the weights in figure [ fig : splice1 ] .
+ given a polynomial map @xmath0 , we may assume the canonical class @xmath121 is supported on fibres of @xmath5 and sections .
this is because any irreducible component of a divisor @xmath122 is either contained in a fibre of @xmath5 or it maps onto @xmath123 and in the latter case we can pass to a cover @xmath21 so that the pull - back of @xmath124 is a collection of sections , introducing new support only in fibres .
let @xmath125 denote the support of @xmath126 and put @xmath127 .
define @xmath128 where @xmath129 .
although @xmath130 resembles a divisor it is not really well - defined since we can add zero times any curve to @xmath126 and this will change @xmath130 .
nevertheless , @xmath130 is a useful tool to study the support of @xmath126 .
the important property of @xmath130 is that for each @xmath131 @xmath132 where the valency of @xmath131 is its valency in the dual graph of @xmath125 .
this expression vanishes on rational curves of valency 2 , precisely those curves hidden in the efficient dual graph . in particular @xmath133 uses only the multiplicities on the nodes of the efficient dual graph . for a polynomial @xmath42 the multiplicities of @xmath130
are easily calculated from the splice diagram .
the node corresponding to the line at infinity in @xmath57 , indicated by the black node in the diagrams , has multiplicity 2 since @xmath126 has multiplicity -3 there . from any other node @xmath75
take a path toward the black node to get to the previous node @xmath134 of valency @xmath110 .
then the multiplicity of @xmath130 is given recursively by @xmath135 of the edge @xmath136 joining @xmath75 and @xmath134 , and @xmath137 is the product of the edge weights adjacent to @xmath75 and not on @xmath136 .
see @xcite for a thorough description of the calculation of @xmath130 from weights in the splice diagram . in figure
[ fig : splice1 ] the multiplicities of @xmath130 on the two valency 3 nodes are 5 and 3 respectively .
another feature of @xmath130 is that it behaves well under branched covers . locally along a curve of ramification @xmath138
the canonical divisor is given by @xmath139 , and hence a @xmath24-fold cover @xmath140 transforms @xmath141 , and along unramified curves the multiplicity remains the same .
thus @xmath130 is given locally by @xmath142 transforming to @xmath143 so the multiplicity simply multiplies by @xmath24 .
figure [ fig : splice3 ] shows how to decompose figure [ fig : splice ] into pieces and how the euler characteristic decomposes .
each node of valency @xmath110 corresponds to a seifert - fibred piece of the jsj decomposition of @xmath95 .
the @xmath144 on the arrow indicates that a fibre of the left piece consists of two identical disjoint components .
the greatest common divisor of the @xmath145 on an arrow and the default value of @xmath146 on the right arrow is 1 which is the number of components of the fibre of the right piece .
the numbers @xmath144 and @xmath145 are calculated using the weights of the splice diagram in the same way as multiplicities are calculated , by taking a path from the edge to an arrow and multiply all weights adjacent to , but not on , the path . thus the left piece has fibre two genus 1 surfaces each with one boundary component and the right piece has fibre a genus 2 surface with three boundary components .
the genus 4 generic fibre with one point removed decomposes along two circles as shown in figure [ fig : stable ] . by pinching along these circles , or vanishing cycles ,
one gets the stable curve in figure [ fig : stableinf ] .
this demonstrates ( i ) of theorem [ th : decomps2 ] . the decomposition in figure [ fig : splice3 ] gives the jsj decomposiiton of @xmath95 into pieces @xmath147 each fibring over the circle and containing a multilink @xmath148 .
the numbers @xmath144 and @xmath145 in figure [ fig : splice3 ] are the multiplicities of the multilink components .
the monodromy on the fibre of @xmath95 decomposes to give the monodromy of the fibre of @xmath149 .
the latter has finite order given by the multiplicity of the valency @xmath110 component .
this uses the fact that the monodromy on seifert - fibred spaces is completely understood .
thus the multiplicity is 12 on the left component - this consists of an order 6 map on the once - punctured torus and a swapping of the two components - and the multiplicity is 10 on the right component .
a 60:1 cover will pull back the monodromy to be trivial on each component and thus this will give the stable reduction of the fibre over infinity as seen in the calculation in section [ sec : blowups ] .
thus ( iii ) and part of ( ii ) theorem [ th : decomps2 ] are seen .
it remains to understand the twisting of the monodromy along the annuli in figure [ fig : stable ] .
upstairs , in the 60:1 cover we expect some whole number of dehn twists along the annuli , so downstairs we expect fractional twists . this calculation is subtler and it is completely solved in terms of the jsj decomposition as theorem 13.1 in @xcite .
it is subtle only because we are choosing to emphasise the more intuitive multipicities of the polynomial @xmath59 on each component of the efficient dual graph , rather than the weights of the splice diagram .
the fractional twisting requires the weights of the splice diagram .
it is given along an annulus corresponding to the edge @xmath136 by @xmath150 where @xmath151 is the number of annuli corresponding to @xmath136 , @xmath152 and @xmath153 are the multiplicities of the neighbouring nodes , and @xmath154 is the determinant of the intersection matrix of the nodes from the full dual graph lying on the edge . in the example
above , @xmath155 , coresponding to the 2 annuli , @xmath156 and @xmath157 .
( we have given the calculation of the edge determinant twice , the second using the splice diagram weights in figure [ fig : splice1 ] , not explained here further . )
thus the fractional twisting is @xmath89 and upstairs in the 60:1 cover it is a dehn twist of order 7 .
this tells us that locally the node of the stable curve is given by @xmath87 and there is a @xmath84 quotient singularity in the ambient surface .
in this section we evaluate the homology class associated to @xmath0 on rational cohomology classes @xmath50 since we consider 1-dimensional bases @xmath123 , we will abuse notation and treat the cohomology classes as numbers @xmath158 . a non - trivial homology class in @xmath49 is detected by the non - vanishing of at least one of the cohomology classes above .
this follows from work of wolpert @xcite where he shows that @xmath159 is a positive multiple of the khler class @xmath160 .
the integral of @xmath160 over any non - trivial homology class is positive and hence at least one of @xmath161 or @xmath162 is non - zero on the homology class .
there are obstructions to a dual graph appearing as the fibre at infinity of a polynomial @xmath42 .
the obstructions are not complete so there are dual graphs where it is not known if they appear as the fibre at infinity of a polynomial @xmath42 .
we do not know if the positivity condition @xmath163 gives a new obstruction on dual graphs .
the class @xmath164 is represented by the boundary divisor @xmath165 .
thus it counts the number of singularities of a polynomial @xmath0 .
the contribution from a stable fibre is its number of nodes , where a node locally given by @xmath166 contributes @xmath103 to the count .
equivalently , if we insist that the total space @xmath45 be non - singular and thus allow semi - stable fibres then the contribution from a semi - stable fibre is simply its number of nodes . the contribution from an unstable fibre
is calculated by first passing to the stable reduction .
in particular , the contribution is not necessarily an integer
. let @xmath38 be the semi - stable reduction of @xmath0 equipped with @xmath3 sections , for @xmath167 a smooth surface that @xmath24-fold covers @xmath45 .
define @xmath168 where @xmath169 is the signature of @xmath167 and @xmath170 is the self - intersection of the @xmath62th section of @xmath171 .
it is not hard to see that if we take a further cover the number of singularities multiplies by the degree of the cover and hence @xmath172 is well - defined .
it is subtler that the signature behaves the same way .
the signature of a branched cover is in general not easily calculated .
the definition above is given for simplicity whereas for the purpose of calculations we give a better definition below in which the classes @xmath161 and @xmath173 in families @xmath174 and @xmath175 for @xmath176 .
the relation between the two definitions of @xmath173 is due to smith @xcite and the others easily follow .
the forgetful map @xmath177 is defined by forgetting the @xmath178st point and possibly blowing down rational components of some fibres .
it defines a universal bundle over @xmath4 equipped with @xmath3 sections @xmath179 . over @xmath180
define the vertical canonical bundle to be the complex line bundle @xmath181 the bundle @xmath182 is used to define the classes @xmath183 as follows .
define the hodge classes @xmath184 the push - forward sheaf @xmath185 is a rank @xmath2 vector bundle over @xmath4 best understood in terms of the fibres which are the @xmath2-dimensional vector spaces of holomorphic 1-forms on the curve associated to a point of @xmath4 . over a stable curve
, one uses 1-forms holomorphic outside the singular set that have at worst simple poles at the singular points with residues summing to zero .
the main purpose of this paper is to enable calculations of the homology class of a polynomial map @xmath0 with 1-dimensional base @xmath123 and in particular , calculations of the rational numbers @xmath192 .
we will calculate these rational numbers for infinite families of polynomial maps @xmath42 .
we first do the calculations for the specific example @xmath69 .
the contribution from the fibre over infinity to @xmath172 comes from the two nodes of the stable fibre over infinity , each counted with multiplicity 7 .
the stable reduction is obtained from a 60:1 cover hence the contribution to @xmath172 is @xmath193 .
this is the fractional twisting of the geometric monodromy and we have given it a cohomological interpretation .
there are 8 finite singularities so @xmath194 we deduce the canonical class by the fact that it can be supported on the infinite fibre plus 59 copies of the fibre of @xmath5 .
this is because the canonical class can be supported on the divisor at infinity @xmath195 since it can be supported on @xmath196 and blowing up introduces support only on the exceptional divisiors .
a branched cover introduces canonical class along the ramification set so the 60:1 cover means 59 copies of the fibre of @xmath5 contributes to the canonical class and the rest remains supported on the fibre over infinity .
apply the adjunction formula @xmath197 to a fibre to see @xmath198 so the section is given multiplicity 6 in the canonical class . to calculate the other multiplicities in figure [ fig : canonical ] ,
similarly apply the adjunction formula to irreducible components of the fibre over infinity . thus @xmath199 and
@xmath200 note that @xmath173 times the degree of the cover is an integer , so the relation @xmath201 puts a mod 12 condition on @xmath202 .
finally @xmath203 is given by @xmath204=-b'\cdot b'$ ] by the adjunction formula .
thus @xmath205 the self - intersection @xmath206 used to calculate @xmath203 is encoded in the fractional twisting of the geometric monodromy .
it gives the fractional twisting of the generic fibre on the annulus around the marked point . in terms of the efficient dual graph , or splice diagram , the annulus @xmath207 corresponds to the edge @xmath136 joining the fibre over infinity to the section .
thus @xmath203 is given by @xmath208 where @xmath209 is the number of annuli corresponding to @xmath136 , and @xmath152 is the multiplicities of the node of @xmath136 .
the formula for the twist is theorem 13.5 in @xcite .
the example @xmath69 is a particular case of the general class of polynomials with one point at infinity .
the splice diagram for the general polynomial with one point at infinity is given in figure [ fig : stable ] .
the stable reduction of the fibre over infinity is easily calculated from theorem [ th : decomps ] .
we refer to the node with weights @xmath211 as the @xmath62th node and the edge between the @xmath62th and @xmath212th nodes as the @xmath62th edge .
the @xmath62th node corresponds to @xmath213 components of euler characteristic @xmath214 , each meeting @xmath215 of the @xmath216 components corresponding to the @xmath217th node .
the monodromy permutes the @xmath213 components and acts internally with order @xmath218 , so its total order on the components corresponding to the @xmath62th node is @xmath219 which is the multiplicity of the polynomial on that node . thus the stable reduction uses a cover of order @xmath220 ( or more efficiently @xmath221 . ) from the fractional twisting along annuli we see that the quotient singularity at each annulus above the @xmath62th edge has order @xmath222 , and the self - intersection of the section in the stable reduction is @xmath223 .
we calculate inequalities for the rational numbers @xmath224 since their maximum value is taken when the fibre of @xmath5 over infinity is the only unstable curve , and the rational numbers @xmath224 are upper - semicontinuous .
we have no good interpretation of the messy expressions for @xmath172 and @xmath161 so we show them only for the cases @xmath225 and @xmath226 . for @xmath225
the euler characteristic of the generic fibre is @xmath227 the stable reduction of the fibre over infinity uses a degree @xmath228 cover and results in a smooth fibre over infinity .
to calculate @xmath172 we use a result of suzuki @xcite @xmath229 where @xmath230 which expresses the euler characteristic of the generic fibre as a sum of the number of vanishing cycles near exceptional fibres .
if all fibres over finite values @xmath231 are stable , which is the generic case that gives an upper bound on @xmath172 , then @xmath232 counts the number of nodes in the fibre @xmath233 .
thus @xmath234 the fractional twisting along the boundary annulus is given by @xmath235 so @xmath236 if all finite fibres are stable , the canonical class upstairs in the @xmath228-fold cover is @xmath237 where @xmath96 is the generic fibre and @xmath6 is the section .
the fractional twisting @xmath238 shows that @xmath239 in the cover .
thus @xmath240 and @xmath241 when @xmath226 , the stable reduction of the fibre over infinity uses a cover of order @xmath242 and consists of an irreducible component with euler characteristic @xmath243 plus @xmath244 irreducible components with euler characteristic @xmath245 , joined along @xmath244 nodes with quotient singularities of order @xmath246 .
now @xmath247 and @xmath248 uses ( [ eq : suz ] ) to get an upper bound for nodes away from infinity , plus @xmath249 times the @xmath244 nodes over infinity , each counted with multiplicity @xmath246 . to calculate @xmath161 we need a more efficient method than has been used so far .
we use the divisor @xmath130 introduced in section [ sec : effdual ] .
the expression @xmath250 consists of @xmath133 which is independent of rational valency 2 nodes in the dual graph , and @xmath251 which does depend on rational valency 2 nodes in the dual graph , however it behaves well on the semi - stable reduction of a family .
a semi - stable curve contains strings of @xmath79 rational curves in place of the quotient singularities at nodes of stable curves .
the expression @xmath251 vanishes on @xmath79 rational curves and hence ( [ eq : kdotk ] ) uses only the multiplicities of @xmath126 on the irreducible components of the stable fibres , and on sections . from this
we calculate @xmath252 when @xmath42 has rational fibres @xmath254 and in particular we get equalities instead of inequalities for @xmath255 and @xmath172 , i.e. the generic behaviour always occurs , since if the sum of two upper - semicontinuous functions is continuous then the two functions are both continuous . when the @xmath3 points at infinity are sections , no branched covering is required for stable reduction .
we use the classification in @xcite to see that either the homology class is trivial or @xmath256 , @xmath257 , @xmath258 , @xmath259 and @xmath260 .
the homology class depends only on @xmath3 .
there are many inequivalent polynomials for each @xmath3 showing that the homology class is a rather course invariant .
the homological monodromy is trivial in both cases when @xmath261 or @xmath256 so the homology class is useful to distinguish these two cases . | a topological invariant of a polynomial map @xmath0 from a complex surface containing a curve @xmath1 to a one - dimensional base is given by a rational second homology class in the compactification of the moduli space of genus @xmath2 curves with @xmath3 labeled points @xmath4 . here
the generic fibre of @xmath5 has genus @xmath2 and intersects @xmath6 in @xmath3 points . in this paper
we give an efficient method to calculate this homology class .
we apply this to any polynomial in two complex variables @xmath7 where the @xmath3 points on a fibre are its points at infinity . |
molecules are a ubiquitous component of the dense interstellar medium . to date
, about one hundred distinct species have been detected in the interstellar gas ( ohishi 1997 ) ; they range from simple diatomic molecules to complex species containing as many as thirteen atoms .
the wide variety of interstellar molecules demonstrates the thermodynamic tendency of most elements to form molecules under the conditions present in the dense interstellar medium .
although most of the interstellar molecules that have been detected previously contain elements of cosmic abundance greater than @xmath15 , a few molecules containing elements of lower abundance have also been detected .
for example , chlorine , with a solar system abundance of only 2@xmath16 relative to hydrogen ( anders & grevesse 1989 ) , was detected more than a decade ago in the form of hydrogen chloride ( blake , keene & phillips 1985 ) .
recent estimates ( neufeld & green 1994 ; schilke , phillips & wang 1995 ; zmuidzinas et al .
1995a ) of the hcl abundances implied by observations of the hcl @xmath17 line toward sgr b2 ( zmuidzinas et al .
1995a ) and the orion molecular cloud 1 ( blake et al .
1985 ; schilke et al .
1995 ) have yielded results for @xmath18 in the range 0.3 2 @xmath19 .
if the elemental abundance of chlorine in these sources is the same as that in the solar system , then hcl accounts for only @xmath20 of the total number of chlorine nuclei .
theoretical models for the chemistry of chlorine - bearing molecules ( schilke et al .
1995 ) predict that hcl will account for @xmath21 of gas - phase chlorine .
thus the observed abundance of hcl can only be understood if the chlorine depletion in the dense ism is large .
the required depletion factors of @xmath22 greatly exceed the values inferred for the _ diffuse _ ism from uv absorption line studies ( e.g. harris , gry , & bromage 1984 ) .
prior to the observations reported in this letter , hcl was the only known _
interstellar _ molecule containing a halogen element .
motivated by the earlier observations of hcl , by the fact that the solar system abundance of fluorine lies only a factor of 6 below that of chlorine ( anders & grevesse 1989 ) , and by the large h f bond strength which suggests that hydrogen fluoride is likely to be a major reservoir of gas - phase fluorine ,
we have undertaken a search for interstellar hydrogen fluoride toward the strong far - infrared continuum source sgr b2 .
such a search would provide a unique probe of the chemistry of interstellar fluorine and of the fluorine depletion in the dense interstellar medium . because of its large rotational constant the largest of any diatomic molecule other than molecular hydrogen or heh@xmath23 hf possesses a rotational spectrum that lies entirely shortward of the atmospheric windows within which most molecules show a rotational spectrum .
observations of hf rotational transitions are therefore possible only from airplane or space - based observatories .
we have made use of the long wavelength spectrometer ( lws ; clegg et al .
1996 ) on board the infrared space observatory ( iso ; kessler et al.1996 ) to search for hydrogen fluoride .
the @xmath17 transition of hf lies longward of the wavelength range to which the photoconductive detectors used in lws are sensitive , so we have carried out observations of the @xmath1 line at 121.6973 @xmath0 m .
these observations are described in 2 below .
our results are presented in 3 , and discussed in 4 .
the detection of hydrogen fluoride reported here marks the first discovery of an _ interstellar _ molecule containing the element fluorine and the first time that a new astrophysical molecule has been identified by means of observations in the far - infrared ( 30 300 @xmath0 m ) spectral region .
using the lws of iso in fabry - perot mode at a resolving power @xmath24 of 9600 , we observed the @xmath1 line of hf toward the source sgr b2 on 1997 march 28th .
the rest frequency of the line is @xmath25 ( nolt et al .
1987 ) , corresponding to a vacuum wavelength of 121.6973 @xmath0 m . unlike hcl
, hf shows no electric quadrupole hyperfine structure , and the magnetic hyperfine splitting is far smaller than the intrinsic width of any astrophysical absorption line .
the iso beam , of approximate diameter @xmath26 fwhm , was centered midway between sgr b2(m ) and sgr b2(n ) , at coordinates @xmath27 , @xmath28 ( j2000 ) .
the observation , which covered the wavelength range 121.63 121.80 @xmath0 m , was carried out in fast scanning mode with 8 spectral samples per resolution element .
the total integration time was 2549 seconds .
the initial data reduction was carried out with version 6.1 of the iso pipeline software .
the isap software package was then used to remove bad data points , to co - add the individual spectral scans , and to derive estimates of the likely statistical error for each flux measurement from the variance of the individual scans .
the resultant spectrum showed a continuum flux level that varied by @xmath29% over the narrow wavelength range covered .
this behavior , which has been widely observed in lws fabry - perot spectra , is caused by the grating that is used as an order sorter for the fabry - perot etalon .
due to errors in the grating setting , the grating wavelength is sometimes offset slightly from the central wavelength of the fabry - perot scan , thereby superimposing an asymmetric envelope on the measured spectrum and compromising the absolute flux calibration .
that envelope was fit with a second order polynomial , and the absolute flux calibration was obtained from a spectrum taken in the grating mode of lws immediately after the fabry - perot scan .
figure 1 shows the lws spectrum observed toward sgr b2 , normalized with respect to the measured continuum flux of @xmath30 jy .
we used a @xmath31 analysis to fit the observed spectrum and assess the statistical significance of lines that we detected .
our analysis makes the implicit assumption that the errors in the fluxes measured at each spectral point are gaussian and independent .
we approximated the spectral response function of the fabry - perot instrument as a lorentzian with a width corresponding to the nominal resolving power @xmath24 of 9600 ( trams , clegg & swinyard 1996 ) ; and we assumed that the astrophysical absorption feature was narrow compared to the instrumental profile , as suggested by the intrinsic line widths measured for hcl ( zmuidzinas et al .
1995a ) .
fitting a single absorption feature to the spectrum , and with a quadratic fit to continuum as described in 2 above , we obtained an absorption line equivalent width of @xmath3 nm and a central lsr velocity of 67 km s@xmath32 for the 121.6973 @xmath0 m hf line .
the measured lsr velocity is in excellent agreement with the centroid velocity observed by zmuidzinas et al .
( 1995a ) for the hcl @xmath17 line in sgr b2 .
our detection of an absorption feature is statistically significant at the @xmath33 level .
in addition to the absorption feature that we attribute to hf , we find evidence at a lower level of statistical significance ( @xmath7 ) for an emission feature near the expected position ( for @xmath34 67 km s@xmath32 ) of the @xmath8 line of water , which has a rest wavelength of 121.7219 @xmath0 m ( toth 1991 ) .
the emission line equivalent width is @xmath9 nm . with the water emission line included in the fit , the reduced @xmath31 is 0.98 ( with 108 degrees of freedom ) .
our identification of the absorption feature in figure 1 with hf @xmath1 is based upon ( 1 ) the excellent agreement of the observed line position with the expected position of the hf line ; and ( 2 ) the absence of any other astrophysically plausible candidate in available line catalogues .
notwithstanding these considerations , the detection of additional hf lines would be desirable as a confirmation of this identification .
unfortunately , according to the excitation model discussed below , we do not expect absorption lines higher than @xmath1 to be strong enough to be detected by the iso spectrometers .
furthermore , the @xmath17 transition near 243.2 @xmath0 m lies longward of the spectral region that is observable with the photoconductive detectors used in lws .
fortunately , the earth s atmosphere permits observations of hf @xmath17 from airplane altitude : that line , with a large predicted equivalent width @xmath35 nm , will therefore be a prime target for observations with instrumentation for longer wavelengths that has been proposed for the stratospheric observatory of infrared astronomy ( sofia ) .
observations of hf @xmath17 will also be possible with instrumentation proposed for the far infrared and submillimetre telescope ( first ) . in deriving a hydrogen fluoride abundance from the observed equivalent width of the @xmath1 line , we have made use of an excitation model to determine the fractional level populations as a function of position within the source .
our model is essentially identical to that used by zmuidzinas et al .
( 1995a ) to model the excitation of hcl in sgr b2 .
the temperature and h@xmath5 density in the source are assumed to vary with the radial coordinate @xmath36 pc according to @xmath37 @xmath38 for an assumed dust opacity that varies with wavelength as @xmath39 , these temperature and density profiles yield a predicted dust continuum spectrum that is in excellent agreement with the observations . because of the large dipole moment ( 1.826 debye ; muenter 1972 ) and large rotational constant ( @xmath40 ) of the hf molecule ,
the critical density at which collisional processes become important in the excitation of hf @xmath41 is very much larger ( @xmath42 ) than the density of the gas ( @xmath43 \rm \sim 10 ^ 5\,cm^{-3}$ ] ) that we sampled in our observations of sgr b2 .
thus the excitation of hf is dominated by the effects of radiative pumping by dust continuum radiation .
the radiative transfer was treated using the method of accelerated lambda iteration ( rybicki & hummer 1991 ) , and the line equivalent width obtained as a function of the hf abundance . in the absence of available collisional data for hf
, we adopted as the rate coefficients for collisional de - excitation those computed previously for the de - excitation of hcl by he ( neufeld & green 1994 ) ; fortunately , because radiative pumping dominates the excitation of hf in sgr b2 , our results are almost entirely independent of the assumed collisional rate coefficients .
on the basis of this model , we found that the observed equivalent width of @xmath3 nm corresponds to an abundance @xmath44 .
the predicted equivalent width shows a nearly linear dependence on the assumed hf abundance ( equivalent width @xmath45 abundance@xmath46 ) for abundances in the range @xmath47 .
figure 2 shows the fractional level populations in @xmath48 , 1 , 2 and 3 as a function of radial coordinate for a hydrogen fluoride abundance of @xmath49 .
observations of the @xmath1 absorption line sample the region exterior to an effective photosphere at @xmath50 .
we also modeled the excitation of the h@xmath5o @xmath8 line that is evident in the spectrum at a lower level of statistical significance .
for the case of water , an excitation model had previously been developed to account for the observed strengths ( zmuidzinas 1995b ; gensheimer , mauersberger & wilson 1996 ) of the @xmath51 and @xmath52 emission lines of h@xmath53o and the @xmath54 absorption line of h@xmath53o toward sgr b2 .
the temperature and h@xmath5 density profiles assumed in that model were identical to those adopted for the hf and hcl excitation models . to fit the measured fluxes of all three lines , a variable water abundance must be assumed , with a smaller water abundance @xmath55 in the cooler outer parts of the source ( @xmath56 k ; @xmath57 ) , and a larger water abundance of @xmath10 in the warmer inner regions ( i.e. the `` hot core '' with @xmath58 k ; @xmath59 ) .
a variation in the water abundance of this nature could be explained by the vaporization of water ice from grain mantles in the warmer parts of the source .
water abundances @xmath60 have previously inferred from observations of h@xmath53o and h@xmath61o in other warm dense molecular regions ( jacq et al .
1988 ; cernicharo et al . 1994 ; gensheimer et al . 1996 ; van dishoeck & helmich 1996 ) . without the adjustment of any model parameter
, our excitation model for water predicts that the @xmath8 line of h@xmath61o which is excited primarily within the hot core where the assumed water abundance is @xmath10 will be an emission line with an equivalent width of 0.4 nm , in good agreement with the observations . to interpret our detection of hf ,
we have constructed a simple model for the steady - state chemistry of interstellar fluorine .
fluorine is qualitatively different in at least one important respect from all the other elements of which hydrides have been detected in the interstellar medium : atomic fluorine reacts exothermically with h@xmath5 via the reaction @xmath62 fluorine also undergoes an exothermic reaction with water : @xmath63 both reactions have been well - studied in the laboratory , although only at temperatures greater than 200 k. for reaction ( r1 ) , measurements by three independent groups have yielded results for the rate coefficient at 298 k in the narrow range @xmath64 ( wurzberg & houston 1980 ; heidner et al . 1980 ; stevens , brune , & anderson 1989 ) .
the temperature dependence of the measured rate coefficient implied the presence of a relatively small activation energy barrier that was variously estimated as @xmath65 , @xmath66 and @xmath67 k in these three studies .
we adopted atkinson et al.s ( 1992 ) review value of @xmath68 for the rate coefficient for ( r1 ) .
for reaction ( r2 ) , stevens et al .
( 1989 ) obtained a rate coefficient @xmath69 with no detectable variation over the temperature range 240373 k ; the implied limit on any activation energy barrier was ( @xmath70 ) k. for a water abundance of @xmath71 , these rate coefficients imply that reaction ( r1 ) dominates the formation of hf at temperatures @xmath72 k. at the h@xmath5 densities present in sgr b2 , both reactions are expected to dominate ion - neutral reaction networks as a source of hf .
once formed , hf is destroyed only very slowly .
hf is more strongly bound than any of the species cf , ch , of ,
oh , nf , nh , sf , sh , sf@xmath23 , sif , sih , or sif@xmath23 : thus reactions of hf with c , o , n , s , s@xmath23 , si , and si@xmath23 are all endothermic .
hf has a smaller proton affinity than h@xmath5o , co or n@xmath5 : thus reactions with h@xmath73o@xmath23 , hco@xmath23 , and @xmath74 are likewise endothermic .
destruction of hf takes place slowly by means of cosmic ray induced photodissociation and as a result of reactions with species of low abundance such as ch , he@xmath23 , h@xmath75 , and c@xmath23 . considering these formation and destruction processes for hydrogen fluoride
, we find that over a wide range of densities and temperatures , including those probed by our observations of hf @xmath1 in sgr b2 , hf accounts for more than @xmath76 of all fluorine nuclei in the gas - phase . in comparison with the ion - neutral reaction routes responsible for the formation of other interstellar hydrides , the neutral - neutral reactions ( r1 ) and ( r2 ) are extremely effective in producing hf . given ( 1 ) our theoretical expectation that hf is the dominant reservoir of gas phase fluorine along the line - of - sight to sgr b2 , ( 2 ) the assumption that the elemental abundance of f is the same as that measured in the solar system ( @xmath77 relative to h ; anders & grevesse 1989 ) , and ( 3 ) the observed hf abundance of @xmath4 relative to h@xmath5 , we conclude that a depletion factor @xmath78 results from the depletion of fluorine onto dust grains in sgr b2 .
this is somewhat smaller than the depletion factors inferred by zmuidzinas et al .
( 1995a ) and schilke et al .
( 1995 ) for chlorine in sgr b2 and in omc-1 .
we gratefully acknowledge the outstanding support provided by the infrared processing and analysis center and particularly by steven lord .
the work of d.a.n . was supported by nasa grant nagw-3147 from the long term space astrophysics research program and by nasa grant nagw-3183 .
j.z . acknowledges support from nasa contract nag2 - 744 and t.g.p . from nsf contract
ast 96 - 15025 .
zmuidzinas , j. , blake , g.a . , carlstrom , j. , keene , j. , miller , d. , schilke , p. , and ugras , n.g .
1995b , in _ proceedings of the airborne astronomy symposium on the galactic ecosystem : from gas to stars to dust _
haas , j.a .
davidson & e.f .
erickson ( san francisco : asp ) , p. 33
-20pt -20pt fig .
1 iso lws fabry - perot spectrum of the hf @xmath1 line toward sgr b2 .
the iso beam , of approximate diameter @xmath26 fwhm , was centered midway between sgr b2(m ) and sgr b2(n ) , at coordinates @xmath27 , @xmath28 ( j2000 ) .
the flux has been normalized relative to the continuum flux of @xmath30 jy .
the lsr velocity plotted on the horizontal axis applies to the hf @xmath1 line at 121.6973 @xmath0 m ( rest wavelength ) .
the solid line shows the fit obtained from the fitting procedure described in the text .
the location of the @xmath79 @xmath8 line in that fit is shown for the case @xmath80 . | we report the first detection of interstellar hydrogen fluoride . using the long wavelength spectrometer ( lws ) of the infrared space observatory ( iso ) ,
we have detected the 121.6973 @xmath0 m @xmath1 line of hf in absorption toward the far - infrared continuum source sagittarius b2 .
the detection is statistically significant at the @xmath2 level . on the basis of our model for the excitation of hf in sgr b2
, the observed line equivalent width of @xmath3 nm implies a hydrogen fluoride abundance of @xmath4 relative to h@xmath5 . if the elemental abundance of fluorine in sgr b2 is the same as that in the solar system , then hf accounts for @xmath6 of the total number of fluorine nuclei .
we expect hydrogen fluoride to be the dominant reservoir of gas - phase fluorine in sgr b2 , because it is formed rapidly in exothermic reactions of atomic fluorine with either water or molecular hydrogen ; thus the measured hf abundance suggests a substantial depletion of fluorine onto dust grains .
similar conclusions regarding depletion have previously been reached for the case of chlorine in dense interstellar clouds .
we also find evidence at a lower level of statistical significance ( @xmath7 ) for an emission feature at the expected position of the @xmath8 121.7219 @xmath0 m line of water .
the emission line equivalent width of @xmath9 nm for the water feature is consistent with the water abundance of @xmath10 relative to h@xmath5 that has been inferred previously from observations of the hot core of sgr b2 .
0.5 true in @xmath11department of physics & astronomy , the johns hopkins university , 3400 north charles street , baltimore , md 21218 @xmath12division of mathematics , physics , & astronomy , 320 - 47 , california institute of technology , pasadena , ca 91125 @xmath13max - planck institute for radio astronomy , auf dem hgel 69 , d-53121 , bonn , germany 0.5 true in @xmath14 based on observations with iso , an esa project with instruments funded by esa member states ( especially the pi countries : france , germany , the netherlands and the united kingdom ) with the participation of isas and nasa . 0.5 true in |
this is part 3 in a sequence of papers @xcite on multiadaptive galerkin methods , @xmath0 and @xmath1 , for approximate ( numerical ) solution of odes of the form @xmath3 , \\
u(0 ) & = u_0 , \end{split } \label{eq : u'=f}\ ] ] where @xmath4 \rightarrow { \mathbb{r}}^n$ ] is the solution to be computed , @xmath5 a given initial condition , @xmath6 a given final time , and @xmath7 \rightarrow { \mathbb{r}}^n$ ] a given function that is lipschitz - continuous in @xmath8 and bounded . in the previous two parts of our series on multiadaptive galerkin methods
, we proved a posteriori error estimates , through which the time steps are adaptively determined from residual feedback and stability information , obtained by solving a dual linearized problem . in this paper , we prove a priori error estimates for @xmath0 and @xmath1 .
we also prove the stability estimates and interpolation estimates which are essential to the a priori error analysis .
standard methods for the time - discretization of ( [ eq : u=f ] ) require that the resolution is equal for all components @xmath9 of the computed approximate solution @xmath10 of ( [ eq : u=f ] ) .
this includes all standard galerkin or runge kutta methods ; see @xcite
. using the same time step sequence @xmath11 for all components could become very costly if the different components of the solution exhibit multiple time scales of different magnitudes .
we therefore propose a new representation of the solution in which the difference in time scales is reflected in the _ componentwise _ time - discretization of ( [ eq : u=f ] ) , that is , each component @xmath9 is computed using an individual time step sequence @xmath12 .
the multiadaptive galerkin methods @xmath0 and @xmath1 first presented in @xcite are formulated as extensions of the standard continuous and discontinuous galerkin methods @xmath13 and @xmath14 , studied earlier in detail by hulme @xcite , jamet @xcite , delfour , hager , and trochu @xcite , eriksson , johnson , and thome @xcite , and estep et al.@xcite .
as such , the analysis of the @xmath0 and @xmath1 methods can be carried out within the existing framework , but the extension to multiadaptive time - stepping leads to some technical challenges , in particular , proving the appropriate interpolation estimates .
local ( multiadaptive ) time - stepping has been explored before to some extent for specific applications , including specialized integrators for the @xmath15-body problem @xcite and low - order methods for conservation laws @xcite .
early attempts at local time - stepping include @xcite .
recently , a new class of related methods , known as asynchronous variational integrators ( avi ) with local time steps , has been proposed @xcite .
the main results of this paper are a priori error estimates for the @xmath0 and @xmath1 methods , respectively , of the form @xmath16,l_1)}\ ] ] and @xmath17,l_1)}\ ] ] for @xmath18 or @xmath19 , where @xmath20 is an interpolation constant , @xmath21 is a ( computable ) stability factor , and @xmath22 ( or @xmath23 ) combines local time steps @xmath12 with derivatives of the exact solution @xmath8 . the norm @xmath24 is defined by @xmath25 .
these estimates state that the @xmath0 method is of order @xmath26 and that the @xmath1 method is of order @xmath27 in the local time step .
we refer to section [ sec : aprioriestimates ] for the exact results .
it should be noted that superconvergence is obtained only at synchronized time levels , such as the end - point @xmath28 . for the general nonlinear problem
, we obtain exponential estimates for the stability factor @xmath21 . in @xcite
, we prove that for a parabolic model problem , the stability factor remains bounded and of unit size , independent of @xmath29 ( up to a logarithmic factor ) .
the following notation is used throughout this paper .
each component @xmath9 , @xmath30 , of the approximate @xmath31 solution @xmath10 of ( [ eq : u=f ] ) is a piecewise polynomial on a partition of @xmath32 $ ] into @xmath33 subintervals .
subinterval @xmath34 for component @xmath35 is denoted by @xmath36 $ ] , and the length of the subinterval is given by the
_ local time step _ @xmath37 .
this is illustrated in figure [ fig : intervals ] . on each subinterval @xmath38
, @xmath39 is a polynomial of degree @xmath40 and we refer to @xmath41 as an _ element_. furthermore , we shall assume that the interval @xmath32 $ ] is partitioned into blocks between certain synchronized time levels @xmath42 .
we refer to the set of intervals @xmath43 between two synchronized time levels @xmath44 and @xmath45 as a _ time slab _ :
@xmath46 we denote the length of a time slab by @xmath47 .
we also refer to the entire collection of intervals @xmath38 as the partition @xmath48 .
since different components use different time steps , a local interval @xmath38 may contain nodal points for other components , that is , some @xmath49 .
we denote the set of such internal nodes on a local interval @xmath38 by @xmath50 .
$ ] for different components .
elements between common synchronized time levels are organized in time slabs . in this example
, we have @xmath51 and @xmath52.,width=377 ] the outline of this paper is as follows . in section [ sec : definition ] , we give the full definition of the multiadaptive galerkin methods @xmath0 and @xmath1 .
we also introduce the dual methods @xmath53 and @xmath54 , which are of importance to the a priori error analysis . in sections [ sec : existence ] and [ sec : stability ] , respectively , we then prove existence and stability of the discrete solutions as defined in section [ sec : definition ] . in section [ sec :
interpolation ] , we prove the interpolation estimates that we later use to prove the a priori error estimates in section [ sec : apriori ] . proving the interpolation estimates
is technically challenging , since the function to be interpolated may be discontinuous within the interval of interpolation . to measure the regularity of the interpolated function , it is then necessary to take into consideration the size of the jump in function value and derivatives at each point of discontinuity .
finally , in section [ sec : numerical ] , we present some numerical evidence for the a priori error estimates by solving a simple model problem and showing that we obtain the predicted convergence rates , @xmath55 and @xmath56 respectively , for the @xmath0 and @xmath1 methods .
in this section , we give the definitions of the multiadaptive galerkin methods @xmath0 and @xmath1 .
the multiadaptive methods are obtained as extensions of the standard ( monoadaptive ) galerkin methods @xmath13 and @xmath14 by extending the trial and test spaces to allow individual time step sequences for different components . as an important tool for the a priori error analysis in section [ sec : apriori ]
, we also introduce the discrete dual problem and the discrete dual methods @xmath53 and @xmath54 .
to formulate the @xmath0 method , we define the _ trial space _ @xmath58 and the _ test space _ @xmath59 as @xmath60)]^n : v_i|_{i_{ij}}\in \mathcal{p}^{q_{ij}}(i_{ij } ) , \
j=1,{\dotsc},m_i , \
i=1,{\dotsc},n \big\},\\ \hat{v } & = \big\{v : v_i|_{i_{ij}}\in \mathcal{p}^{q_{ij}-1}(i_{ij } ) , \
j=1,{\dotsc},m_i , \
i=1,{\dotsc},n \big\},\\ \end{split } \label{eq : spaces , mcg}\ ] ] where @xmath61 denotes the linear space of polynomials of degree @xmath2 on an interval @xmath62 .
in other words , @xmath58 is the space of vector - valued continuous piecewise polynomials of degree @xmath63 with @xmath64 on the partition @xmath48 , and @xmath59 is the space of vector - valued ( possibly discontinuous ) piecewise polynomials of degree @xmath65 on the same partition .
we now define the @xmath0 method for ( [ eq : u=f ] ) as follows : find @xmath66 with @xmath67 such that @xmath68 where @xmath69 denotes the @xmath70 inner product . with a suitable choice of test function @xmath71
, it follows that the global problem ( [ eq : fem , mcg ] ) can be restated as a sequence of successive local problems for each component : for @xmath30 , @xmath72 , find @xmath73 with @xmath74 given such that @xmath75 where the initial condition is specified for @xmath30 by @xmath76 .
we define the _ residual _ @xmath77 of the approximate solution @xmath78 by @xmath79 . in terms of the residual ,
we can rewrite ( [ eq : fem , mcg , local ] ) in the form @xmath80 that is , the residual is orthogonal to the test space on each local interval .
we refer to ( [ eq : fem , mcg , local , orthogonality ] ) as the _ galerkin orthogonality _ of the @xmath0 method . for @xmath1
, we define the trial and test spaces by @xmath81 that is , both trial and test functions are vector - valued ( possibly discontinuous ) piecewise polynomials of degree @xmath63 with @xmath82 on the partition @xmath48 . by definition , the @xmath1 solution @xmath66 is left - continuous .
we now define the @xmath1 method for ( [ eq : u=f ] ) as follows : find @xmath66 with @xmath83 such that @xmath84_{i , j-1 } v_i\big(t_{i , j-1}^+\big ) + \int_{i_{ij } } \dot{u}_i v_i \ , dt \right ] = \int_0^t ( f(u,\cdot),v ) \ , dt \quad \forall v\in \hat{v } , \label{eq : fem , mdg}\ ] ] where denotes the jump in @xmath9 across the node , and where .
the @xmath1 method in local form , corresponding to ( [ eq : fem , mcg , local ] ) , reads as follows : for @xmath30 , @xmath72 , find @xmath73 such that @xmath85_{i , j-1 } v(t_{i , j-1 } ) + \int_{i_{ij } } \dot{u}_i v \ , dt = \int_{i_{ij } } f_i(u,\cdot ) v \ , dt \quad \forall v\in \mathcal{p}^{q_{ij}}(i_{ij } ) , \label{eq : fem , mdg , local}\ ] ] where the initial condition is specified for @xmath30 by @xmath86 .
the residual @xmath77 is defined on the inner of each local interval @xmath38 by @xmath87 . in terms of the residual , ( [ eq : fem , mdg , local ] )
can be restated in the form @xmath85_{i , j-1 } v\big(t_{i , j-1}^+\big ) + \int_{i_{ij } } r_i(u,\cdot ) v \ , dt = 0 \quad \forall v\in \mathcal{p}^{q_{ij}}(i_{ij } ) \label{eq : fem , mdg , local , orthogonality}\ ] ] for @xmath88 , @xmath89 .
we refer to ( [ eq : fem , mdg , local , orthogonality ] ) as the galerkin orthogonality of the @xmath1 method .
the dual problem is the standard tool for error analysis , a priori or a posteriori , of galerkin finite element methods for the numerical solution of differential equations ; see @xcite .
for the a posteriori error analysis of the multiadaptive galerkin methods @xmath0 and @xmath1 in @xcite , we formulate a continuous dual problem . for the a priori error analysis of this paper ,
we formulate instead a discrete dual problem .
the discrete dual problem was first introduced for the family of discontinuous galerkin methods @xmath14 in @xcite .
as we shall see , the discrete dual problem can be expressed as a galerkin method for a continuous problem .
the discrete dual solution @xmath90\rightarrow { \mathbb{r}}^n$ ] is a galerkin approximation of the exact solution @xmath91\rightarrow { \mathbb{r}}^n$ ] of the continuous dual problem @xmath92 where @xmath93 is an interpolant or a projection of the exact solution @xmath8 of ( [ eq : u=f ] ) , @xmath94\rightarrow { \mathbb{r}}^n$ ] is a given function , @xmath95 is a given initial condition , and @xmath96 that is , an appropriate mean value of the transpose of the jacobian of the right - hand side @xmath97 evaluated at @xmath98 and @xmath10 . note that by the chain rule , we have @xmath99
the data @xmath100 of the dual problem allow us to obtain error estimates for different functionals @xmath101 of the error @xmath102 .
we define below two new galerkin methods for the dual problem ( [ eq : dual ] ) : the dual methods @xmath53 and @xmath54 .
we will later use the @xmath53 method to express the error of the @xmath0 solution of ( [ eq : u=f ] ) in terms of the @xmath53 solution of ( [ eq : dual ] ) .
similarly , we will express the error of the @xmath1 solution of ( [ eq : u=f ] ) in terms of the @xmath54 solution of ( [ eq : dual ] ) . in the formulation of the dual method of @xmath0 , we interchange the trial and test spaces of @xmath0 . with the same definitions of @xmath58 and @xmath59 as in ( [ eq : spaces , mcg ] ) , we thus define the @xmath53 method for ( [ eq : dual ] ) as follows : find @xmath104 with @xmath105 such that @xmath106 for all @xmath107 with @xmath108 , where @xmath109 notice the extra condition that the test functions should vanish at @xmath110 , which is introduced to make the dimension of the test space equal to the dimension of the trial space . integrating by parts , ( [ eq : mcgqd ] )
can alternatively be expressed in the form @xmath111_{ij } v_i(t_{ij } ) - \int_{i_{ij } } \dot{\phi}_i v_i \ , dt \right ] = \int_0^t ( j^{\top}(\pi u , u,\cdot ) \phi + g , v ) \ , dt.\ ] ] @xmath112with the same definitions of @xmath58 and @xmath59 as in ( [ eq : spaces , mdg ] ) , we define the @xmath54 method for ( [ eq : dual ] ) as follows : find @xmath104 with @xmath105 such that @xmath113_{i , j-1 } \phi_i\big(t_{i , j-1}^+\big ) + \int_{i_{ij } } \dot{v}_i \phi_i \ , dt \right ] = \int_0^t ( j(\pi u , u,\cdot)v,\phi ) \ , dt + l_{\psi , g}(v)\ ] ] for all @xmath114 with @xmath115 . integrating by parts , ( [ eq : mdgqd ] ) can alternatively be expressed in the form @xmath116_{ij } v_i\big(t_{ij}^-\big ) - \int_{i_{ij } } \dot{\phi}_i v_i \ , dt \right ] = \int_0^t ( j^{\top}(\pi u , u,\cdot)\phi+g , v ) \ , dt.\ ] ]
to prove existence of the discrete @xmath0 , @xmath1 , @xmath53 , and @xmath54solutions defined in the previous section , we formulate fixed point iterations for the construction of solutions .
existence then follows from the banach fixed point theorem if the time steps are sufficiently small .
[ lem : explicit ] let @xmath43 be a time slab with synchronized time levels @xmath44 and @xmath45 . with time reversed for the dual methods ( to simplify the notation ) , the @xmath0 , @xmath1 , @xmath53 , and @xmath54 methods can all be expressed in the following form : for all @xmath117 , find @xmath118 ( the degrees of freedom for @xmath119 on @xmath38 ) such that @xmath120}(\tau_{ij}(t ) ) f_i(u,\cdot ) \ , dt,\ ] ] where @xmath121 and @xmath122}\}$ ] is a set of polynomial weight functions on @xmath123 $ ] .
the result follows from the definitions of the @xmath0 , @xmath1 , @xmath53 , and @xmath54 methods , using an appropriate basis for the trial and test spaces .
see @xcite for details .
let @xmath124 be the maximum time slab length and define the lipschitz constant @xmath125 by @xmath126 \ \forall x , y \in { \mathbb{r}}^n.\ ] ] if now @xmath127 where @xmath128 is a constant depending only on the order and method , the fixed point iteration _ ( [ eq : explicit ] ) _ converges to the unique solution of _ ( [ eq : fem , mcg ] ) _ , _ ( [ eq : fem , mdg ] ) _ , _ ( [ eq : mcgqd ] ) _ , and _ ( [ eq : mdgqd ] ) _ , respectively .
the result follows by lemma [ lem : explicit ] and an application of the banach fixed point theorem .
see @xcite for details .
write the dual problem ( [ eq : dual ] ) for @xmath129 in the form @xmath130 for simplicity , we consider only the case @xmath131 . with @xmath132 , we have @xmath133 , and so ( [ eq : linear , dual ] ) can be written as a forward problem for @xmath134 in the form @xmath135 , \\
w(0 ) & = w_0 , \end{split}\ ] ] where @xmath136 and @xmath137 .
below , @xmath134 represents either @xmath8 or @xmath138 and , correspondingly , @xmath139 represents either the discrete @xmath140 approximation @xmath78 of @xmath8 or the discrete @xmath141 approximation @xmath142 of @xmath143 .
the general exponential stability estimate is based on the following version of the discrete gronwall inequality .
[ lem : gronwall ] assume that @xmath144 are nonnegative , @xmath145 for all @xmath146 , and @xmath147 for all @xmath15 .
then @xmath148 for @xmath149 . by a standard discrete gronwall inequality @xcite , @xmath150 if @xmath151 @xmath152 for @xmath153 and @xmath154 . here , @xmath155 @xmath156 , and so @xmath157 , since @xmath158 .
the result now follows if we take @xmath159 .
[ th : estimate , exponential ] let @xmath139 be the @xmath0 , @xmath1 , @xmath53 , or @xmath54 solution of _ ( [ eq : linear , common])_. then there is a constant @xmath160 , depending only on the highest order @xmath161 , such that if @xmath162,l_p ) } \leq 1 $ ] for @xmath163 , then @xmath164,l_p ) } \leq c \|w_0\|_{l_p } \exp\left ( \sum_{m=1}^{n-1 } k_m c \|b\|_{l_{\infty}([t_{m-1},t_m],l_p ) } \right)\ ] ] for @xmath163 , @xmath165 . by lemma [ lem : explicit ]
, we can write the @xmath0 , @xmath1 , @xmath53 , and @xmath54 methods in the form . applied to the linear model problem ( [ eq : linear , common ] ) , we have , and so @xmath166}(\tau_{ij}(t ) ) ( bw)_i \ , dt \bigg| \\ & \leq & \vert w_i(0 ) \vert + c \int_0^{t_{ij } } \vert ( bw)_i \vert \ , dt \leq \vert w_i(0 ) \vert + c \int_0^{t_n } \vert ( bw)_i \vert \ , dt , \end{aligned}\ ] ] where @xmath45 is smallest synchronized time level for which @xmath167 .
it now follows that for all @xmath168 $ ] , we have @xmath169 , and so @xmath170 the result now follows by letting @xmath171,l_p)}$ ] .
see @xcite for an extension to multiadaptive time - stepping of the strong stability estimate lemma 6.1 for parabolic problems in @xcite .
in this section , we introduce a pair of carefully chosen interpolants , and , which are central to the a priori error analysis of the @xmath0 and @xmath1 methods .
the interpolants are defined in section [ sec : interpolation : interpolants ] . in section [ sec : interpolation : basic ] , we discuss the interpolation of piecewise smooth functions , that is , the interpolation of functions which may be discontinuous within the interval of interpolation , and then present the basic general interpolation estimates for the two interpolants and . for the a priori error analysis of the @xmath0 and @xmath1 methods , we will also need a special interpolation estimate for the function @xmath172 , where @xmath173 is the jacobian of the right - hand side @xmath174 of ( [ eq : u=f ] ) and @xmath142 is the discrete dual solution as defined in section [ sec : definition ] , including estimates for the size of the jump in function value and derivatives for the function @xmath175 at points of discontinuity .
these estimates are proved in section [ sec : interpolation : special ] , based on a representation formula for the @xmath0 and @xmath1 solutions of ( [ eq : u=f ] ) .
the interpolant @xmath176 } } : v \rightarrow \mathcal{p}^{q}([a , b])$ ] is defined by the following conditions : @xmath177}}v(a ) = v(a)\quad \text { and } \quad { \pi_{\mathrm{cg}}^{[q]}}v(b ) = v(b ) , \\ & \int_a^b \big(v - { \pi_{\mathrm{cg}}^{[q]}}v\big ) w \ , dx = 0 \quad\forall w\in\mathcal{p}^{q-2}([a , b ] ) , \end{split}\ ] ] where @xmath58 denotes the set of functions that are piecewise and bounded on @xmath178 $ ] . in other words , is the polynomial of degree @xmath2 that interpolates @xmath71 at the end - points of the interval @xmath178 $ ] and additionally satisfies @xmath179 projection conditions .
this is illustrated in figure [ fig : picg ] .
we also define the dual interpolant as the standard @xmath180-projection onto . } } v$ ] ( dashed ) of the function @xmath181 ( solid ) on @xmath123 $ ] for @xmath182 ( left ) and @xmath183 ( right).,title="fig:",width=226 ] } } v$ ] ( dashed ) of the function @xmath181 ( solid ) on @xmath123 $ ] for @xmath182 ( left ) and @xmath183 ( right).,title="fig:",width=226 ] } } v$ ] ( dashed ) of the function @xmath181 ( solid ) on @xmath123 $ ] for @xmath184 ( left ) and @xmath183 ( right).,title="fig:",width=226 ] } } v$ ] ( dashed ) of the function @xmath181 ( solid ) on @xmath123 $ ] for @xmath184 ( left ) and @xmath183 ( right).,title="fig:",width=226 ] the interpolant @xmath185 } } : v \rightarrow \mathcal{p}^{q}([a , b])$ ] is defined by the following conditions : @xmath186}}v(b ) = v(b ) , \\ &
\int_a^b \big(v - { \pi_{\mathrm{dg}}^{[q]}}v\big ) w \ , dx = 0 \quad\forall w\in\mathcal{p}^{q-1}([a , b ] ) , \end{split}\ ] ] that is , is the polynomial of degree @xmath2 that interpolates @xmath71 at the right end - point of the interval @xmath178 $ ] and additionally satisfies @xmath2 projection conditions .
this is illustrated in figure [ fig : pidg ] .
the dual interpolant is defined similarly , with the difference being that the left end - point @xmath187 is used for interpolation . to estimate the size of the interpolation error @xmath188 for a given function @xmath71
, we express the interpolation error in terms of the regularity of @xmath71 and the length of the interpolation interval , @xmath189 . specifically , when @xmath190 ) \subset v$ ] for some @xmath191 , we obtain estimates of the form @xmath192 where @xmath193)}$ ] denotes the maximum norm on @xmath178 $ ] .
this estimate is a simple consequence of the peano kernel theorem @xcite if one can show that the interpolant @xmath194 ) \subset v$ ] is linear and bounded on @xmath58 and that @xmath195 is exact on @xmath196 ) \subset v$ ] , that is , @xmath197 for all @xmath198)$ ] . in the general case , where the interpolated function @xmath71 is only piecewise smooth ( see figure [ fig : piecewise_smooth ] )
, we also need to include the size of the jump @xmath199_x$ ] in function value and derivatives at each point @xmath200 of discontinuity within @xmath201 to measure the regularity of the interpolated function @xmath71 . in @xcite , we prove the following extensions of the basic estimate ( [ eq : basic ] ) .
and its interpolant @xmath202.,width=377,height=151 ] [ lem : estimate , piecewise ] if @xmath195 is linear and bounded on @xmath58 and is exact on @xmath203 ) \subset v$ ] , then there is a constant @xmath204 such that for all @xmath71 piecewise @xmath205 on @xmath178 $ ] with discontinuities at @xmath206 , @xmath207_{x_j } \big|\ ] ] for @xmath208 , @xmath209 .
[ lem : derivatives , piecewise ] if @xmath195 is linear and bounded on @xmath58 and is exact on @xmath203 ) \subset v$ ] , then there is a constant @xmath204 such that for all @xmath71 piecewise @xmath205 on @xmath178 $ ] with discontinuities at @xmath206 , @xmath210_{x_j } \big|\ ] ] for @xmath211 .
lemmas [ lem : estimate , piecewise ] and [ lem : derivatives , piecewise ] apply to both the @xmath176}}$ ] interpolant ( for @xmath212 ) and the @xmath185}}$ ] interpolant ( for @xmath213 ) defined in section [ sec : interpolation : interpolants ] .
the linearity of both interpolants follows directly from the definition of the interpolants .
the proofs that both interpolants are bounded and exact on @xmath203)$ ] are given in detail in @xcite and @xcite . to prove a priori error estimates for @xmath0 and @xmath1 in section [ sec : apriori ] , we need to estimate the interpolation error @xmath214 for the function @xmath175 defined by @xmath215 we note that @xmath216 may be discontinuous within @xmath38 if @xmath38 contains a node for some other component , which is generally the case .
this is illustrated in figure [ fig : varphi ] .
note that on the right - hand side @xmath174 is linearized around a mean value of @xmath93 and @xmath8 .
has a node within @xmath38 , then @xmath217 may be discontinuous within @xmath38 , causing @xmath216 to be discontinuous within @xmath38.,width=377,height=226 ] an interpolation estimate for @xmath214 follows directly from lemma [ lem : estimate , piecewise ] . to use this estimate , we need to estimate the size of the jump in function value and derivatives at each internal node @xmath218 of the partition @xmath48 . to obtain this estimate
, we need to make a number of additional assumptions on the right - hand side @xmath174 of ( [ eq : u=f ] ) and the partition @xmath48 .
these assumptions are discussed in section [ sec : assumptions ] . based on the assumptions and the representation formula presented in section [ sec : representation ]
, we obtain the jump estimates in section [ sec : jumpestimates ] and , finally , in section [ sec : special , estimate ] , the interpolation estimate for @xmath175 .
the proof of jump estimates for the multiadaptive galerkin methods @xmath0 and @xmath1 is based on expressing the solutions as certain interpolants .
these representations are obtained as follows .
let @xmath78 be the @xmath0 or @xmath1 solution of ( [ eq : u=f ] ) and define , for @xmath30 , @xmath219 similarly , for @xmath142 the @xmath53 or @xmath54 solution of ( [ eq : dual ] ) , we define , for @xmath30 , @xmath220 where @xmath221 .
we note that @xmath222 and @xmath223 .
it now turns out that @xmath78 can be expressed as an interpolant of @xmath224 .
similarly , @xmath142 can be expressed as an interpolant of @xmath225 .
we present these representations in lemmas [ lem : representation , cg ] and [ lem : representation , dg ] .
we remind the reader about the interpolants @xmath176}}$ ] , @xmath226}}$ ] , @xmath185}}$ ] , and @xmath227}}$ ] , defined in section [ sec : interpolation : interpolants ] .
[ lem : representation , cg ] the @xmath0 solution @xmath78 of _ ( [ eq : u=f ] ) _ can expressed in the form . similarly , the @xmath53 solution @xmath142 of _ ( [ eq : dual ] ) _ can be expressed in the form @xmath228}}\tilde{\phi}$ ] , that is , @xmath229 } \tilde{u}_i$ ] and @xmath230 } \tilde{\phi}_i$ ] on each local interval @xmath38 .
the representation formulas follow by the definitions of the @xmath0 and @xmath53 methods and the interpolants @xmath176}}$ ] and @xmath226}}{}$ ] .
see @xcite for details .
[ lem : representation , dg ] the @xmath1 solution @xmath78 of _ ( [ eq : u=f ] ) _ can expressed in the form . similarly , the @xmath54 solution @xmath142 of _ ( [ eq : dual ] ) _ can be expressed in the form @xmath231}}\tilde{\phi}$ ] , that is , @xmath232 } \tilde{u}_i$ ] and @xmath233 } \tilde{\phi}_i$ ] on each local interval @xmath38 .
the representation formulas follow by the definitions of the @xmath1 and @xmath54 methods and the interpolants @xmath185}}$ ] and @xmath227}}{}$ ] .
see @xcite for details . to estimate the size of the jump in function value and derivatives for the function @xmath175 defined in ( [ eq : varphi ] ) , we make the following assumptions .
given a time slab @xmath48 , assume that for each pair of local intervals @xmath38 and @xmath234 within the time slab , we have @xmath235 and @xmath236 for some @xmath237 and some @xmath238 .
the dependence on @xmath239 in the error estimates is weak ( see remark [ rem : alpha ] ) , so assumption ( a2 ) does not prevent multiadaptivity .
we also assume that the problem ( [ eq : u=f ] ) is autonomous , @xmath240 noting that the dual problem nevertheless will be nonautonomous in general .
furthermore , we assume that @xmath241 where @xmath242 is defined for @xmath243 and @xmath244 by @xmath245 @xmath246 , with the norm @xmath246 defined by @xmath247 @xmath248 for all @xmath249 , and @xmath250 the @xmath15th - order tensor given by @xmath251 furthermore , we choose @xmath252 such that @xmath253 for @xmath254 , and @xmath255_t\big\|_{l_{\infty } } \leq c_f \sum_{n=0}^p
c_x^{p - n } \big\|\big[x^{(n)}\big]_t\big\|_{l_{\infty}}\ ] ] for @xmath256 and any given @xmath257 , where @xmath258 denotes a constant such that @xmath259 for @xmath260 .
note that @xmath261 defines a piecewise constant function on the partition @xmath42 .
note also that assumption ( a4 ) implies that each @xmath262 is bounded by @xmath263 .
we further assume that there is a constant @xmath264 such that .a5 @xmath265 for each local interval @xmath38 .
we summarize the list of assumptions as follows : \(a ) the local orders @xmath40 are equal within each time slab ; the local time steps @xmath266 are semiuniform within each time slab ; @xmath174 is autonomous ; @xmath174 and its derivatives are bounded ; the local time steps @xmath266 are small . to estimate higher - order derivatives , we face the problem of taking higher - order derivatives of @xmath267 with respect to @xmath268 . in lemmas [ lem :
estimate , derivative ] and [ lem : estimate , jump ] , we present basic estimates for composite functions @xmath269 with @xmath270 and @xmath271 . the proofs are based on a straightforward application of the chain rule and leibniz rule and are given in full detail in @xcite . [
lem : estimate , derivative ] let @xmath243 be @xmath244 times differentiable in all its variables , let @xmath271 be @xmath272 times differentiable , and let @xmath258 be a constant such that @xmath273 for @xmath260 . then there is a constant @xmath274 such that @xmath275 [ lem : estimate , jump ] let @xmath243 be @xmath276 times differentiable in all its variables , let @xmath271 be @xmath272 times differentiable , except possibly at some @xmath277 , and let @xmath258 be a constant such that @xmath278 for @xmath260 . then there is a constant @xmath274 such that @xmath279_{t } \right| \leq c \|v\|_{d^{p+1}({\mathbb{r } } ) } \sum_{n=0}^p
c_x^{p - n } \big\|\big[x^{(n)}\big]_{t}\big\|_{l_{\infty}}.\ ] ] we now prove estimates for derivatives and jumps of the @xmath0 or @xmath1 solution @xmath78 of the general nonlinear problem ( [ eq : u=f ] ) , under the assumptions listed in section [ sec : assumptions ] . similarly , one can obtain estimates for the discrete dual solution @xmath142 and the function @xmath175 defined in ( [ eq : varphi ] ) , from which the desired interpolation estimates follow . to obtain estimates for the multiadaptive solution @xmath78 , we first prove estimates for the function @xmath224 defined in section [ sec : representation ]
. the estimates for @xmath78 then follow by induction . to simplify the estimates , we introduce the following notation . for given @xmath280 ,
let @xmath281 be a constant such that @xmath282 for @xmath283 , we define @xmath284 .
temporarily , we assume that there is a constant @xmath285 such that for each @xmath272 , .a5@xmath286 @xmath287 this assumption will be removed in lemma [ lem : u , derivatives ] . in the following lemma ,
we use assumptions ( a1 ) , ( a3 ) , and ( a4 ) to derive estimates for @xmath224 in terms of @xmath288 and @xmath263 .
[ lem : utilde ] let @xmath78 be the @xmath0 or @xmath1 solution of _ ( [ eq : u=f ] ) _ and define @xmath224 as in _ ( [ eq : tildefunction])_. if assumptions _ ( a1 ) _ , _ ( a3 ) _ , and _
( a4 ) _ hold , then there is a constant @xmath289 such that @xmath290 and @xmath291_{t_{i , j-1}}\big\|_{l_{\infty } } \leq c \sum_{n=0}^{p-1 } c_{u ,
p-1}^{p - n } \big\|\big[u^{(n)}\big]_{t_{i , j-1}}\big\|_{l_{\infty } } , \quad p=1,{\dotsc},\bar{q}+1,\ ] ] for each local interval @xmath38 , where @xmath292 is an internal node of the time slab @xmath48 . by definition , @xmath293 , and
so the results follow directly by lemmas [ lem : estimate , derivative ] and [ lem : estimate , jump ] , noting that @xmath294 . by lemma [
lem : utilde ] , we now obtain the following estimate for the size of the jump in function value and derivatives for @xmath78 .
[ lem : u , jumps ] let @xmath78 be the @xmath0 or @xmath1 solution of _ ( [ eq : u=f])_. if assumptions _ ( a1)__(a5 ) _ and _ ( a5@xmath286 ) _ hold , then there is a constant @xmath295 such that @xmath296_{t_{i , j-1}}\big\|_{l_{\infty } } \leq c k_{ij}^{r+1-p } c_{u , r}^{r+1 } , \quad p=0,{\dotsc},r+1 , \quad r = 0,{\dotsc},\bar{q},\ ] ] for each local interval @xmath38 , where @xmath292 is an internal node of the time slab @xmath48 .
the proof is by induction .
we first note that at @xmath297 , we have @xmath298_{t } & = & u_i^{(p)}(t^+ ) - \tilde{u}_i^{(p)}(t^+ ) + \tilde{u}_i^{(p)}(t^+ ) - \tilde{u}_i^{(p)}(t^- ) + \tilde{u}_i^{(p)}(t^-)-u_i^{(p)}(t^- ) \\ & \equiv & e_+ + e_0 + e_- .
\end{aligned}\ ] ] by lemma [ lem : representation , cg ] ( or lemma [ lem : representation , dg ] ) , @xmath78 is an interpolant of @xmath224 and so , by lemma [ lem : estimate , piecewise ] , we have @xmath299_{x } \big\vert\ ] ] for @xmath208 and @xmath300 .
note that the second sum starts at @xmath301 rather than at @xmath302 , since @xmath224 is continuous .
similarly , we have @xmath303_{x } \big\vert.\ ] ] to estimate @xmath304 , we note that @xmath305 for @xmath283 , since @xmath224 is continuous . for @xmath306 , lemma [ lem : utilde ]
gives @xmath307_t \vert \leq c \sum_{n=0}^{p-1 } c_{u , p-1}^{p - n } \|[u^{(n)}]_t\|_{l_{\infty}}$ ] . by assumption ( a2 ) , it then follows that ( [ eq : u , jumps , first ] ) holds for @xmath308 .
assume now that ( [ eq : u , jumps , first ] ) holds for @xmath309 .
then , by lemma [ lem : utilde ] and assumption ( a5@xmath286 ) , it follows that @xmath310_{x}\big\|_{l_{\infty } } \\ & \leq c k_{ij}^{\bar{r}+1-p } c_{u,\bar{r}}^{\bar{r}+1 } + c \sum k_{ij}^{m - p } c_{u , m-1}^{m - n } k_{ij}^{(\bar{r}-1)+1-n } c_{u,\bar{r}-1}^{(\bar{r}-1)+1 } \\ & \leq c k_{ij}^{\bar{r}+1-p } c_{u,\bar{r}}^{\bar{r}+1 } \left ( 1 + \sum ( k_{ij } c_{u,\bar{r}-1})^{m-1-n } \right ) \leq c k_{ij}^{\bar{r}+1-p } c_{u,\bar{r}}^{\bar{r}+1}. \end{split}\ ] ] similarly , we obtain the estimate @xmath311 .
finally , we use lemma [ lem : utilde ] and assumption ( a5@xmath286 ) to obtain the estimate @xmath312_t\big\|_{l_{\infty } } \leq c \sum_{n=0}^{p-1 } c_{u , p-1}^{p - n } k_{ij}^{(\bar{r}-1)+1-n } c_{u,\bar{r}-1}^{(\bar{r}-1)+1 } \\ & = & c k_{ij}^{\bar{r}+1-p } c_{u,\bar{r}}^{\bar{r}+1 } \sum_{n=0}^{p-1 } ( k_{ij } c_{u,\bar{r}})^{p-1-n } \leq c k_{ij}^{\bar{r}+1-p } c_{u,\bar{r}}^{\bar{r}+1}. \end{aligned}\ ] ] summing up , we thus obtain @xmath313_t \vert \leq |e_+| + |e_0| + |e_-| \leq c k_{ij}^{\bar{r}+1-p } c_{u,\bar{r}}^{\bar{r}+1}$ ] , and so ( [ eq : u , jumps , first ] ) follows by induction . by lemmas [ lem : utilde ] and [ lem : u , jumps ] , we now obtain the following estimate for derivatives of the solution @xmath78 . [
lem : u , derivatives ] let @xmath78 be the @xmath0 or @xmath1 solution of _ ( [ eq : u=f])_. if assumptions _ ( a1)__(a5 ) _ hold , then there is a constant @xmath314 such that @xmath315 by lemma [ lem : representation , cg ] ( or lemma [ lem : representation , dg ] ) , @xmath78 is an interpolant of @xmath224 and so , by lemma [ lem : estimate , piecewise ] , we have @xmath316_x \big\vert\ ] ] for some constant @xmath317 . for @xmath318
, we thus obtain the estimate @xmath319 by assumption ( a4 ) , and so ( [ eq : u , derivatives ] ) holds for @xmath318 . for @xmath320 , assuming that ( a5@xmath286 ) holds for @xmath321 , we use lemmas [ lem : utilde ] and [ lem : u , jumps ] ( with @xmath322 ) and assumption ( a2 ) to obtain @xmath323_x\big\|_{l_{\infty } } \\ & \leq c c_{u , p-1}^{p } + c \sum k_{ij}^{m - p } c_{u , m-1}^{m - n } k_{ij}^{(p-1)+1-n } c_{u , p-1}^{(p-1)+1 } \\ & \leq c c_{u , p-1}^{p } \left ( 1 + \sum ( k_{ij } c_{u , m-1})^{m - n } \right ) \leq c c_{u , p-1}^{p } , \end{split}\ ] ] where @xmath324 .
this holds for all components @xmath35 and all local intervals @xmath38 within the time slab @xmath48 , and so @xmath325 where by definition @xmath321 is a constant such that for @xmath326 . starting at @xmath318
, we now define @xmath327 with @xmath328 .
it then follows that ( a5@xmath286 ) holds for @xmath329 with @xmath330 , and thus @xmath331 where @xmath332 .
we may thus define @xmath333 .
continuing , we note that ( a5@xmath286 ) holds for @xmath334 , and thus @xmath335 where @xmath336 . in this way
, we obtain a sequence of constants @xmath337 , depending only on @xmath338 , @xmath339 , and @xmath239 , such that @xmath340 for @xmath341 , and so ( [ eq : u , derivatives ] ) follows if we take @xmath342 .
having now removed the additional assumption ( a5@xmath286 ) , we obtain the following version of lemma [ lem : u , jumps ] . [
lem : u , jumps , new ] let @xmath78 be the @xmath0 or @xmath1 solution of _ ( [ eq : u=f])_. if assumptions _ ( a1)__(a5 ) _ hold , then there is a constant @xmath314 such that @xmath343_{t_{i , j-1}}\big\|_{l_{\infty } } \leq c k_{ij}^{\bar{q}+1-p } c_f^{\bar{q}+1 } , \quad p=0,{\dotsc},\bar{q},\ ] ] for each local interval @xmath38 , where @xmath292 is an internal node of the time slab @xmath48 .
similarly , we obtain estimates for the discrete dual solution @xmath142 and the function @xmath175 . in lemma
[ lem : varphi ] , we present the estimates for the function @xmath175 . [
lem : varphi ] let @xmath175 be defined as in _ ( [ eq : varphi])_. if assumptions _ ( a1)__(a5 ) _ hold , then there is a constant @xmath314 such that @xmath344 and @xmath345_x \big| \leq c k_{ij}^{r_{ij}-p } c_f^{r_{ij}+1 } \|\phi\|_{l_{\infty}(\mathcal{t},l_{\infty } ) } \quad \forall x\in \mathcal{n}_{ij } , \quad p=0,{\dotsc},q_{ij}-1,\ ] ] with @xmath346 for the @xmath0 method and @xmath347 for the @xmath1 method .
this holds for each local interval @xmath38 within the time slab @xmath48 . using the basic interpolation estimate of section [ sec : interpolation : basic ] , we now obtain the following important interpolation estimates for the function @xmath175 . [
lem : interpolationestimate , phi ] let @xmath175 be defined as in _
( [ eq : varphi])_. if assumptions _ ( a1)__(a5 ) _ hold , then there is a constant @xmath314 such that @xmath348 } \varphi_i - \varphi_i \big\|_{l_{\infty}(i_{ij } ) } \leq c k_{ij}^{q_{ij}-1 } c_f^{q_{ij } } \|\phi\|_{l_{\infty}(\mathcal{t},l_{\infty } ) } , \quad q_{ij } = \bar{q } \geq 2,\ ] ] and @xmath349 } \varphi_i - \varphi_i \big\|_{l_{\infty}(i_{ij } ) } \leq c k_{ij}^{q_{ij } } c_f^{q_{ij}+1 } \|\phi\|_{l_{\infty}(\mathcal{t},l_{\infty } ) } , \quad q_{ij } = \bar{q } \geq 1,\ ] ] for each local interval @xmath38 within the time slab @xmath48 . to prove ( [ eq : varphi , estimate , cg ] ) , we use lemma [ lem : estimate , piecewise ] , with @xmath350 and @xmath283 , together with lemma [ lem : varphi ] , to obtain @xmath351 } \varphi_i - \varphi_i \big\|_{l_{\infty}(i_{ij } ) } \leq c k_{ij}^{q_{ij}-1 } \big\| \varphi_i^{(q_{ij}-1 ) } \big\|_{l_{\infty}(i_{ij } ) } + c \sum_{x\in \mathcal{n}_{ij } } \sum_{m=0}^{q_{ij}-2 } k_{ij}^{m } \big\vert \big [ \varphi_i^{(m ) } \big]_x \big\vert \\ & \quad \leq c k_{ij}^{q_{ij}-1 } c_f^{q_{ij } } \|\phi\|_{l_{\infty}(\mathcal{t},l_{\infty } ) } + c \sum_{x\in \mathcal{n}_{ij } } \sum_{m=0}^{q_{ij}-2 } k_{ij}^{m } k_{ij}^{q_{ij}-m } c_f^{q_{ij}+1 } \|\phi\|_{l_{\infty}(\mathcal{t},l_{\infty } ) } \\ & \quad = c k_{ij}^{q_{ij}-1 } c_f^{q_{ij } } \|\phi\|_{l_{\infty}(\mathcal{t},l_{\infty } ) } + c k_{ij}^{q_{ij } } c_f^{q_{ij}+1 } \|\phi\|_{l_{\infty}(\mathcal{t},l_{\infty } ) } , \end{split}\ ] ] from which the estimate follows . the estimate for @xmath352 } \varphi_i - \varphi_i$ ]
is obtained similarly .
[ rem : alpha ] note that there is only a weak dependence on @xmath339 and @xmath239 , since the jump term contains an extra factor @xmath266 .
if higher - order terms can be ignored , then the dependence on @xmath339 and @xmath239 can be removed .
to prove a priori error estimates for the @xmath0 and @xmath1 methods , we derive error representations in section [ sec : errorrepresentation ] and then obtain the a priori error estimates in section [ sec : aprioriestimates ] for the general nonlinear case .
we refer to @xcite for a sharp a priori error estimate in the case of a parabolic model problem . for each of the two methods , @xmath0 and @xmath1
, we represent the error in terms of the discrete dual solution @xmath142 and an interpolant @xmath93 of the exact solution @xmath8 of ( [ eq : u=f ] ) , using the special interpolants or defined in section [ sec : interpolation ] .
we write the error @xmath102 in the form @xmath353 where @xmath354 is represented in terms of the discrete dual solution and the residual of the interpolant .
an estimate for the second part of the error , @xmath355 , follows directly from an interpolation estimate . in lemma [ lem : errorrepresentation , cg ] , we present the error representation for the @xmath0 method , and then present the corresponding representation for the @xmath1 method in lemma [ lem : errorrepresentation , dg ] .
the error representations are obtained directly by choosing @xmath356 as a test function for the discrete dual problems ( [ eq : mcgqd ] ) and ( [ eq : mdgqd ] ) .
[ lem : errorrepresentation , cg ] let @xmath78 be the @xmath0 solution of _ ( [ eq : u=f ] ) _ , let @xmath142 be the corresponding @xmath53 solution of the dual problem _ ( [ eq : dual ] ) _ , and let @xmath93 be any trial space approximation of the exact solution @xmath8 of _ ( [ eq : u=f ] ) _ that interpolates @xmath8 at the end - points of every local interval .
then @xmath357 where @xmath354 .
[ lem : errorrepresentation , dg ] let @xmath78 be the @xmath1 solution of _ ( [ eq : u=f ] ) _ , let @xmath142 be the corresponding @xmath54 solution of the dual problem _ ( [ eq : dual ] ) _ , and let @xmath93 be any trial space approximation of the exact solution @xmath8 of _ ( [ eq : u=f ] ) _ that interpolates @xmath8 at the right end - point of every local interval
. then @xmath358_{i , j-1 } \phi_i\big(t_{i , j-1}^+\big ) + \int_{i_{ij } } r_i(\pi u,\cdot ) \phi_i \ , dt \right],\ ] ] where @xmath354 . with a special choice of interpolant , and respectively , we obtain the following versions of the error representations .
[ cor : errorrepresentation , cg ] let @xmath78 be the @xmath0 solution of _ ( [ eq : u=f ] ) _ and let @xmath142 be the corresponding @xmath53 solution of the dual problem _
( [ eq : dual])_. then @xmath359}}u,\cdot\big ) - f(u,\cdot),\phi\big ) \ , dt.\ ] ] integrate by parts and use the definition of the interpolant @xmath176}}$ ] .
[ cor : errorrepresentation , dg ] let @xmath78 be the @xmath1 solution of _ ( [ eq : u=f ] ) _ and let @xmath142 be the corresponding @xmath54 solution of the dual problem _
( [ eq : dual])_. then @xmath360}}u,\cdot\big ) - f(u,\cdot),\phi\big ) \ , dt.\ ] ] integrate by parts and use the definition of the interpolant @xmath185}}$ ] . using the error representations of section
[ sec : errorrepresentation ] , the stability estimates of section [ sec : stability ] , and the interpolation estimates of section [ sec : interpolation ] , we now prove our main results : a priori error estimates for general order @xmath0 and @xmath1 .
[ th : apriori , cg ] let @xmath78 be the @xmath0 solution of _ ( [ eq : u=f ] ) _ and let @xmath142 be the corresponding @xmath53 solution of the dual problem _ ( [ eq : dual])_. then there is a constant @xmath204 such that @xmath361,l_2)},\ ] ] where @xmath362 for @xmath363 , and where the stability factor @xmath21 is given by @xmath364}}u , u , \cdot ) \phi\|_{l_2 } \ , dt$ ] . furthermore
, if assumptions _ ( a1)__(a5 ) _ hold , then there is a constant @xmath365 such that @xmath366,l_1)},\ ] ] where @xmath367 for @xmath363 , and where the stability factor @xmath368 is given by @xmath369 by corollary [ cor : errorrepresentation , cg ] , we obtain @xmath370}}u,\cdot\big ) - f(u,\cdot),\phi\big ) \ , dt = \int_0^t \big({\pi_{\mathrm{cg}}^{[q]}}u - u , j^{\top}\big({\pi_{\mathrm{cg}}^{[q]}}u , u,\cdot\big ) \phi\big ) \
, dt.\ ] ] by lemma [ lem : estimate , piecewise ] , it now follows that @xmath371,l_2 ) } \int_0^t \big\|j^{\top}\big({\pi_{\mathrm{cg}}^{[q]}}u , u , \cdot\big ) \phi\big\|_{l_2 } \
, dt,\ ] ] which proves ( [ eq : estimate , cg,1 ] ) . to prove ( [ eq : estimate , cg,2 ] ) , we note that by definition , @xmath372 } u_i - u_i$ ] is orthogonal to @xmath373 for each local interval @xmath38 , and so , recalling that @xmath374}}u ,
u , \cdot ) \phi$ ] , @xmath375 } u_i - u_i\big ) \varphi_i \ , dt = \sum_{i , j } \int_{i_{ij } } \big({\pi_{\mathrm{cg}}}^{[q_{ij } ] } u_i - u_i\big ) \big(\varphi_i - { \pi_{\mathrm{cg}}}^{[q_{ij}-2 ] } \varphi_i\big ) \ , dt , \end{aligned}\ ] ] where we take @xmath376 } \varphi_i \equiv 0 $ ] for @xmath377 . by lemmas [ lem : estimate , piecewise ] and [ lem : interpolationestimate , phi ]
, it now follows that @xmath378 } u - u,\varphi - { \pi_{\mathrm{cg}}}^{[q-2 ] } \varphi\big)\big| \ , dt \\ & = \int_0^t \big|\big(k^{q-1 } c_f^{q-1 } \big({\pi_{\mathrm{cg}}}^{[q ] } u - u\big ) , k^{-(q-1 ) } c_f^{-(q-1 ) } \big(\varphi - { \pi_{\mathrm{cg}}}^{[q-2 ] } \varphi\big)\big)\big| \ , dt \\ & \leq c \big\|k^{2q } \bar{\bar{u}}^{(2q)}\big\|_{l_{\infty}([0,t],l_1 ) } \int_0^t c_f \|\phi\|_{l_{\infty}(\mathcal{t},l_{\infty } ) } \ , dt \\ & = c \bar{s}(t ) \big\|k^{2q } \bar{\bar{u}}^{(2q)}\big\|_{l_{\infty}([0,t],l_1 ) } , \end{split}\ ] ] where @xmath379 .
similarly , we obtain the following a priori error estimate for the @xmath1 method .
[ th : apriori , dg ] let @xmath78 be the @xmath1 solution of _ ( [ eq : u=f ] ) _ and let @xmath142 be the corresponding @xmath54 solution of the dual problem _
( [ eq : dual])_. then there is a constant @xmath204 such that @xmath380,l_2)},\ ] ] where @xmath362 for @xmath363 , and where the stability factor @xmath21 is given by @xmath381}}u , u , \cdot ) \phi\|_{l_2 } \ , dt$ ] . furthermore ,
if assumptions _ ( a1)__(a5 ) _ hold , then there is a constant @xmath365 such that @xmath382,l_1)},\ ] ] where @xmath383 for @xmath363 , and where the stability factor @xmath368 is given by @xmath384 using the stability estimate proved in section [ sec : stability ] , we obtain the following bound for the stability factor @xmath368 . [ lem : stability ] assume that @xmath385 for all time slabs @xmath43 , with @xmath386 the constant in theorem [ th : estimate , exponential ] , and take @xmath387 in ( [ eq : dual ] ) .
then @xmath388 where @xmath389 } c_f$ ] .
_ by theorem [ th : estimate , exponential ] , we obtain @xmath390 and so @xmath391 finally , we rewrite the estimates of theorems [ th : apriori , cg ] and [ th : apriori , dg ] for special choices of data @xmath392 and @xmath393 .
we first take @xmath394 . with @xmath395 for @xmath396 , @xmath397 for @xmath398 , and
@xmath399 we obtain and so @xmath400 by an inverse estimate . by definition
, it follows that @xmath401 .
note that for this choice of @xmath393 , we have @xmath402,l_2)}=\|g\|_{l_1([0,t],l_{\infty})}=1 $ ] . we also make the choice @xmath387 .
noting that @xmath403 , since @xmath404 , we obtain @xmath405 for @xmath406 and @xmath407 for @xmath408 , and @xmath409 for @xmath410 .
note that for both choices of @xmath392 , we have @xmath411 . with these choices of data , we obtain the following versions of the a priori error estimates .
[ cor : apriori , cg ] let @xmath78 be the @xmath0 solution of _ ( [ eq : u=f])_. then there is a constant @xmath204 such that @xmath412,l_{\infty } ) } \leq c s(t ) \big\|k^{q+1 } \bar{u}^{(q+1)}\big\|_{l_{\infty}([0,t],l_2)},\ ] ] where the stability factor @xmath364}}u , u , \cdot ) \phi\|_{l_2 } \ , dt$ ] is taken as the maximum over @xmath394 and @xmath402,l_{\infty})}=1 $ ] .
furthermore , if assumptions _ ( a1)__(a5 ) _ and the assumptions of lemma [ lem : stability ] hold , then there is a constant @xmath413 such that @xmath414,l_1)}\ ] ] for @xmath415 , where the stability factor @xmath368 is given by @xmath416 .
[ cor : apriori , dg ] let @xmath78 be the @xmath1 solution of _ ( [ eq : u=f])_. then there is a constant @xmath204 such that @xmath412,l_{\infty } ) } \leq c s(t ) \big\|k^{q+1 } \bar{u}^{(q+1)}\big\|_{l_{\infty}([0,t],l_2)},\ ] ] where the stability factor @xmath381}}u , u , \cdot ) \phi\|_{l_2 } \ , dt$ ] is taken as the maximum over @xmath394 and @xmath402,l_{\infty})}=1 $ ] .
furthermore , if assumptions _ ( a1)__(a5 ) _ and the assumptions of lemma [ lem : stability ] hold , then there is a constant @xmath413 such that @xmath417,l_1)}\ ] ] for @xmath415 , where the stability factor @xmath368 is given by @xmath416 .
the stability factor @xmath21 that appears in the a priori error estimates is obtained from the discrete solution @xmath142 of the dual problem ( [ eq : linear , dual ] ) , and can thus be computed by solving the discrete dual problem .
numerical computation of the stability factor reveals the exact nature of the problem , in particular , whether or not the problem is parabolic ; if the stability factor is of unit size and does not grow , then the problem is parabolic by definition ; see @xcite .
the error representations presented in section [ sec : errorrepresentation ] are based on the galerkin orthogonalities of the @xmath0 and @xmath1 methods . in particular , for the @xmath0 method , we assume that @xmath418 in the presence of quadrature errors , this term is nonzero . as a result , we obtain an additional term of the form @xmath419 where @xmath420 is the interpolant of @xmath174 corresponding the quadrature rule that is used .
a convenient choice of quadrature for the @xmath0 method is lobatto quadrature with @xmath421 nodal points @xcite , which means that the quadrature error is of order @xmath422 and so ( super)convergence of order @xmath26 is obtained also in the presence of quadrature errors .
similarly for the @xmath1 method , we use radau quadrature with @xmath423 nodal points , which means that the quadrature error is of order @xmath424 , and so the @xmath27 convergence order of @xmath1 is also maintained under quadrature .
we conclude by demonstrating the convergence of the multiadaptive methods in the case of a simple test problem .
consider the problem @xmath425 on @xmath123 $ ] with initial condition @xmath426 .
the solution is given by @xmath427 .
for given @xmath428 , we take @xmath429 for @xmath430 , @xmath431 for @xmath432 , and @xmath433 for @xmath434 , and study the convergence of the error @xmath435 with decreasing @xmath436 . from the results presented in figure [ fig : convergence ] and tables [ tab : convergence , cg ] and [ tab : convergence , dg ] ,
it is clear that the predicted order of convergence is obtained . | the multiadaptive continuous / discontinuous galerkin methods @xmath0 and @xmath1 for the numerical solution of initial value problems for ordinary differential equations are based on piecewise polynomial approximation of degree @xmath2 on partitions in time with time steps which may vary for different components of the computed solution . in this paper , we prove general order a priori error estimates for the @xmath0 and @xmath1 methods . to prove the error estimates
, we represent the error in terms of a discrete dual solution and the residual of an interpolant of the exact solution .
the estimates then follow from interpolation estimates , together with stability estimates for the discrete dual solution .
multiadaptivity , individual time steps , local time steps , ode , continuous galerkin , discontinuous galerkin , mcg(@xmath2 ) , mdg(@xmath2 ) , a priori error estimates , existence , stability , peano kernel theorem , interpolation estimates , piecewise smooth 65l05 , 65l07 , 65l20 , 65l50 , 65l60 , 65l70 10.1137/040604133 |
power quality control is one of the major concerns for power delivery systems to function reliably , and it requires measurements of voltage characteristics , among which the frequency measurement is a non - trivial task due to the presence of voltage sags and voltage harmonics mostly caused by nonlinear loads @xcite . in the particular case of three - phase power systems , the clarke s @xmath0 , @xmath1 transformation is widely used as the preprocessing method to create a complex - valued single - phase signal from the real - valued three - phase signals @xcite , so that traditional complex - valued spectrum estimation methods can be applied , such as the mvdr method or the recently proposed iterative mvdr ( i - mvdr ) method @xcite . to improve the resolution
, we can further apply the subspace methods and one representative example is the music method @xcite .
however , all the zero sequence voltages will be cancelled out in the complex - valued signal and hence can not be detected .
although these harmonic voltages would simply be blocked by a delta transformer , they will add up in the neutral , leading to overheating in the transformer and potential fire hazards @xcite . to detect these harmonics , as well as harmonics of other orders ,
we propose a quaternion - valued model and all the traditional spectrum estimation methods can be extended to this domain , such as mvdr and music .
we will show that harmonics of all orders will be reserved in the resulting quaternion - valued signal and will be detected by relevant estimation methods .
the rest of this paper is organised as follows .
a brief review of the complex - valued model is presented in section ii .
our quaternion - valued model is proposed in section iii , together with the fourier analysis and the mvdr and music - like estimation algorithms .
simulation results are provided in section iv and conclusions are drawn in section v.
we consider the following discrete - time balanced three - phase power system in the presence of @xmath2 harmonic distortions : @xmath3 where @xmath4 are the amplitudes of the harmonic signals , @xmath5 is the fundamental ( angular ) frequency , @xmath6 is the sampling interval , @xmath7 is the signal phase , and @xmath8 are the measurement noise . traditionally , the three - phase signals will be converted to a complex - valued single - phase signal via the clarke s @xmath9 transformation .
firstly , the three - phase signals are mixed into two parts , namely @xmath10 and @xmath11 , where @xmath12 @xmath13 then these two parts will be merged as a complex - valued signal @xmath14 . with this complex - valued signal
, we can exploit the mvdr spectrum to locate the frequencies , and it is given by : @xmath15 where @xmath16 is the hermitian - transpose operation , @xmath17 is the covariance matrix of dimension @xmath18 , and @xmath19^\mathrm t\ ] ] is the frequency sweeping vector .
we can also use the music spectrum which is expressed as : @xmath20 where @xmath21 denotes the euclidean norm , @xmath22 represents the noise subspace and comprises the eigenvectors of the covariance matrix @xmath17 which are corresponding to the @xmath23 smallest eigenvalues , where @xmath23 is assumed to be known or can be estimated using the information theory methods @xcite . in practice
, the covariance @xmath17 needs to be updated and estimated from the average of samples , @xmath24 where @xmath25 where @xmath26 is the number of observations . in detail , @xmath27 is composed of complex - domain harmonic signals that can be divided into two categories plus noise , @xmath28 where @xmath29 is the summation of all positive sequence voltages , @xmath30 and @xmath31 is the summation of all negative sequence voltages , @xmath32 and @xmath33 denotes the largest integer not greater than @xmath34 .
all the zero sequence voltages have been cancelled out .
zero sequence voltages of the same order are cophasial in the three voltage channels and will be eliminated since both rows of the transformation matrix @xmath35 are zero - mean vectors . to solve this problem
, we propose our quaternion - valued approach in the next section .
we construct a quaternion - valued signal from the three - phase signals as @xmath36 where @xmath37 are the three imaginary units of the quaternion algebra which are constrained by @xcite @xmath38 this quaternion - valued signal contains quaternion - domain harmonic signals that belong to three categories @xmath39 , @xmath40 , @xmath41 plus noise , @xmath42 where @xmath39 is the summation of all the positive sequence voltages , @xmath43\\ & \quad-\frac{\imath+\jmath+k}{\sqrt{3}}\sin[(3p-2)(\omega nt_s+\phi)]\big\}\\ & = \sum_{p=1}^{\lfloor\frac{h+2}{3}\rfloor}\frac{2\imath-\jmath - k}{2}v_{3p-2}e^{-\frac{\imath+\jmath+k}{\sqrt{3}}(3p-2)(\omega nt_s+\phi)}\ ; , \end{split } \label{3p-2}\ ] ] @xmath40 is the summation of all the negative sequence voltages , @xmath44\\ & \quad+\frac{\imath+\jmath+k}{\sqrt{3}}\sin[(3p-1)(\omega nt_s+\phi)]\big\}\\ & = \sum_{p=1}^{\lfloor\frac{h+1}{3}\rfloor}\frac{2\imath-\jmath - k}{2}v_{3p-1}e^{\frac{\imath+\jmath+k}{\sqrt{3}}(3p-1)(\omega nt_s+\phi)}\ ; , \end{split } \label{3p-1}\ ] ] @xmath41 is the summation of all the zero sequence voltages , @xmath45\\ & = \sum_{p=1}^{\lfloor\frac{h}{3}\rfloor}v_{3p}\frac{\imath+\jmath+k}{2}\big(e^{\frac{\imath+\jmath+k}{\sqrt{3}}3p(\omega nt_s+\phi)}\\ & + e^{-\frac{\imath+\jmath+k}{\sqrt{3}}3p(\omega nt_s+\phi)}\big)\;. \end{split } \label{3p}\ ] ] hence all the harmonic signals will be reserved in the quaternion - valued signal
. we may observe from ( [ 3p-2])([3p ] ) that the frequencies of the harmonic signals have been mapped into the frequencies of the quaternion - valued signal associated with the @xmath46 axis ( see fig .
axis , title="fig:",width=211 ] + + then we can adopt the mvdr spectrum in ( [ mvdr ] ) and the music spectrum in ( [ music ] ) by substituting the frequency sweeping vector as @xmath47^\mathrm t\;.\ ] ] the frequencies detected in the spectrum are either the original real - domain angular frequencies or their additive inverses , namely 1 .
if a peak is detected in the spectrum in the absence of its additive inverse , it corresponds to a positive or negative sequence voltage signal and this spectrum peak indicates its angular frequency or its additive inverse .
if two `` mirrored '' peaks are detected in the spectrum , they correspond to a zero sequence voltage signal and they indicate the signal s angular frequency and its additive inverse , respectively .
in this section , we provide some numerical examples to illustrate the performance of the proposed quaternion model . in all experiments ,
the fundamental frequency is 50 hz , the sampling frequency is @xmath48 khz , the initial phase is @xmath49 , and @xmath50 , @xmath51 .
there exist a second - order and a third - order harmonic signals , both set to be 6% in amplitude . in the first experiment
, we test the capability of the two modelings .
the mvdr and music spectra of the quaternion- and complex - valued models are plotted in fig .
2 , where snr @xmath52 db .
it can be observed that the proposed model is able to detect all the harmonic signals , namely 50 hz ( the fundamental frequency ) , @xmath53 hz ( the second - order harmonic ) , and @xmath54150 hz ( the third - order harmonic ) , while the complex - valued model fails at the third - order harmonic frequency . in the second experiment ,
we test the accuracy of relevant algorithms
. the estimation errors ( averaged via 300 monte carlo simulation runs ) of the quaternion- and complex - valued mvdr and music algorithms are plotted in fig .
3 , where the snr value varies from 25 to 45 db . we can see that all the algorithms have similar estimation accuracy .
we have presented a quaternion - valued model as an alternative preprocessing approach to convert the three - phase signals into a single - phase system . compared with the clarke s transformation
, the proposed model can additionally detect the zero sequence voltages .
simulated results show that the proposed model can detect all - order voltage harmonics effectively , while sharing similar estimation accuracy with the complex - valued model .
m. bollen , _ understanding power quality problems : voltage sags and interruptions_. wiley - ieee press , 2000 .
m. akke , `` frequency estimation by demodulation of two complex signals , '' _ ieee transactions on power delivery _ , vol .
1 , pp . 157163 , jan .
h. j. jeon and t. g. chang , `` iterative frequency estimation based on mvdr spectrum , '' _ ieee transactions on power delivery _
2 , pp . 621630 , apr .
y. xia and d. p. mandic , `` augmented mvdr spectrum - based frequency estimation for unbalanced power systems , '' _ ieee transactions on instrumentation and measurement _ , vol .
62 , no . 7 , pp
. 19171926 , jul . 2013 .
r. o. schmidt , `` multiple emitter location and signal parameter estimation , '' _ ieee transactions on antennas and propagation _ , vol .
3 , pp . 276280 , mar .
. w. m. grady and s. santoso , `` understanding power system harmonics , '' _ ieee power engineering review _ ,
811 , nov .
j. p. ward , _ quaternions and cayley numbers : algebra and applications _ , kluwer , normwell , ma , 1997 .
m. wax and t. kailath , `` detection of signals by information theoretic criteria , '' _ ieee transactions on acoustics , speech , and signal processing _
assp-33 , no .
2 , pp . 387392 , apr . 1985
. s. j. sangwine , `` fourier transforms of colour images using quaternion or hypercomplex numbers , '' _ electronics letters _ ,
19791980 , oct .
j. w. tao , `` performance analysis for interference and noise canceller based on hypercomplex and spatio - temporal - polarisation processes , '' _ iet radar sonar and navigation _ , vol . 7 , no .
3 , pp . 277286 , mar .
s. miron , n. le bihan , j. i. mars , `` quaternion - music for vector - sensor array processing , '' _ ieee transactions on signal processing _ ,
4 , pp . 12181229 , apr . | in this work , a quaternion - valued model is proposed in lieu of the clarke s @xmath0 , @xmath1 transformation to convert three - phase quantities to a hypercomplex single - phase signal .
the concatenated signal can be used for harmonic distortion detection in three - phase power systems . in particular , the proposed model maps all the harmonic frequencies into frequencies in the quaternion domain , while the clarke s transformation - based methods will fail to detect the zero sequence voltages .
based on the quaternion - valued model , the fourier transform , the minimum variance distortionless response ( mvdr ) algorithm and the multiple signal classification ( music ) algorithm are presented as examples to detect harmonic distortion .
simulations are provided to demonstrate the potentials of this new modeling method .
shell : bare demo of ieeetran.cls for ieee journals harmonics detection , fourier transform , minimum variance distortionless response , multiple signal classification , quaternion , three - phase power system |
realistic quantum systems interact and exchange information with their surroundings . the engineering of the decoherence and the flow of information between an open quantum system and its environment has recently allowed , e.g. , to drive quantum computation by dissipation @xcite , to control entanglement and quantum phases in many - body systems @xcite , to create an open system quantum simulator @xcite , and to control the transition from markovian dynamics to a regime with quantum memory effects @xcite .
moreover , there has been recently rapid progress in understanding various fundamental issues on quantum non - markovianity @xcite . in general , the power of quantum information processing is based on the presence of quantum correlations in multipartite systems .
however , as any realistic quantum system , those systems are vulnerable to the detrimental effects of the environment such as the loss of coherence and information . depending on its physical composition , an open multipartite system can interact either with a single common environment or each of the sub - parties can interact locally with their own reservoirs .
the latter case , which is in the focus of our paper , is a common scenario , e.g. , in quantum information processing and entanglement dynamics @xcite , and in energy transport in biological organisms @xcite .
the central question to be studied here is , can a composite open quantum system exploit during its time evolution the information carried initially by the environmental degrees of freedom , i.e. , the information to which it does not have access initially ? on the basis of a recent theoretical prediction @xcite , we demonstrate experimentally that initial correlations between local parts of the environment lead to nonlocal memory effects and allow to protect and maintain genuine quantum properties of an open system .
we will show further that the non - markovian dynamics of photonic open systems provides a diagnostic tool for characteristic features of their environments .
our experimental open system consists of the polarization degrees of freedom of a pair of entangled photons .
the photon pair is created by spontaneous parametric down - conversion after which the photons travel along different arms @xmath0 and move through different quartz plates of variable thickness ( see fig . 1 ) .
the quartz plates act as birefringent media which lead to a local coupling between the polarization degrees of freedom of the photon in arm @xmath1 and its frequency ( mode ) degrees of freedom forming its local environment .
this local interaction can be described by a local unitary operator which is defined by @xmath2 denotes the state of the photon in arm @xmath1 with polarization @xmath3 ( horizontal or vertical ) and frequency @xmath4 .
the refraction index for photons with polarization @xmath5 is denoted by @xmath6 , and @xmath7 represents the interaction time . experimental setup .
maximally entangled photon pairs are generated by parametric down conversion ( pdc ) with an ultraviolet ( uv ) pump pulse whose spectral width can be controlled .
then a series of quartz plates are added in arm @xmath8 and arm @xmath9 to realize the local dephasing channels .
the final two - photon polarization state is analyzed by state tomography .
key to the components : nc nonlinear crystal , fp fabry - prot cavity , dm dispersion medium , hwp half wave plate , qwp quarter wave plate , if interference filter , qp quartz plate , pbs polarizing beamsplitter , and spd single photon detector.,scaledwidth=40.0% ] any pure initial state for the polarization degrees of freedom of the photon pair , which forms our bipartite open system , can be written as @xmath10 .
the total system initial states are assumed to be product states @xmath11 where @xmath12 is the probability amplitude for the photon in arm @xmath8 to have frequency @xmath13 and for the photon in arm @xmath9 to have frequency @xmath14 , with the corresponding joint probability distribution @xmath15 .
the time evolution of the open system can be represented by means of the dynamical map @xmath16 @xcite which maps the initial polarization state @xmath17 to the polarization state state @xmath18 at time @xmath19 . calculating the total system time evolution and taking the trace over the environment yields
@xcite @xmath20 the dynamics represents a pure dephasing process whose decoherence functions can be expressed completely in terms of the fourier transform of the joint probability distribution @xmath21 as @xmath22 , @xmath23 , @xmath24 and @xmath25 , where @xmath26 denotes the birefringence .
the function @xmath27 describes the decoherence of the polarization state of photon @xmath1 which is induced by the local coupling to its frequency environment caused by the quartz plate in arm @xmath1 .
this can be seen by taking the trace over the polarization state of photon 2 to obtain the polarization state of photon 1 : @xmath28 or by tracing over photon 1 to find the state of photon 2 : @xmath29 the nonlocal character of the dynamical map @xmath30 is controlled by the decoherence functions @xmath31 and @xmath32 in eq .
( [ rho12 ] ) . as a matter of fact
, the map @xmath30 is a product of local dynamical maps , i.e. , @xmath33 if and only if @xmath34 and @xmath35 . in particular , this implies that frequency correlations in the initial state do not influence the local dynamics of each photon and can not be detected by observing the local polarization states .
however , the evolution of the composite two - photon polarization state is determined by the frequency correlations in the environment which lead to memory effects and global non - markovian dynamics .
this also means that , counterintuitively , by adding degrees of freedom to an open system one can change its dynamics from the markovian to the non - markovian regime .
to quantify non - markovianity we use the trace distance based measure @xcite which is defined by @xmath36 . here , @xmath37 and the trace distance @xmath38 between two states @xmath39 and @xmath40 is given by @xmath41 .
the amount of non - markovianity is therefore equal to the total increase of the trace distance during the time evolution and quantifies the total information flow from the environment to the system .
this measure has been used recently in several theoretical and experimental contexts , see , e.g. refs .
@xcite .
trace distance dynamics for different values of the spectral width of the pump pulse .
the widths and the amounts of non - markovianity are : ( a ) @xmath42 , @xmath43 , ( b ) @xmath44 , @xmath45 , ( c ) @xmath46 , @xmath47 , ( d ) @xmath48 , @xmath49 .
the @xmath50-axis represents time measured by the total effective path difference between the horizontal and vertical photons caused by quartz plates in arms @xmath8 and @xmath9 ( @xmath51 ) .
the @xmath52-axis is the trace distance between the time - evolved initial open system states @xmath53 and @xmath54 .
we first add quartz plates in arm 1 .
when the total effective path difference between horizontal and vertical photons equals @xmath55 , we add quartz plates in arm 2 .
the solid line shows the fit using the theoretical result of eq .
( [ d ] ) .
the error bars are due to the counting statistics .
see the supplementary information for more details on the counting statistics and the fitting of the theoretical curves.,scaledwidth=50.0% ] for anticorrelated frequencies @xmath4 the pair of the bell - states @xmath56 maximizes the increase of the trace distance and thus determines the non - markovianity measure @xmath57 @xcite .
these states are created by using spontaneous parametric down conversion as is illustrated in fig . 1 .
a femtosecond pulse ( the duration is about @xmath58 and the operation wavelength is at @xmath59 , with a repetition rate of about @xmath60 ) generated from a ti : sapphire laser is frequency doubled to pump two @xmath8mm - thick beamlike cut beta barium borate ( bbo ) crystals @xcite creating the two - photon entangled state . with the help of
qwp1 we can easily tune the entangled state between @xmath61 ( qwp1 is set at @xmath62 ) and @xmath63 ( qwp1 is set at @xmath64 ) .
the size of the anticorrelations between the photon frequencies is controlled in the experiment by the spectrum of the uv pump pulse .
since energy is conserved in the down conversion process , decreasing the spectral width of the pump pulse decreases the uncertainty of the sum of the frequencies of the photons , and hence increases the anticorrelation between the frequencies .
we use four different pulse widths as described in more detail in the methods section .
it is also worth noting that even though our experimental setup is based on downconversion , which is a commonly used tool to prepare specific two - photon states in quantum optical experiments , the physical phenomena that we describe and realize experimentally was theoretically predicted only very recently @xcite .
trace distance dynamics of the open system with different dispersions of the pump pulse .
the quartz plate on the path of the pump pulse has thickness ( a ) @xmath65 ( b ) @xmath66 ( c ) @xmath67 . in all cases @xmath68 and the measured values of non - markovianity are @xmath69 . the trace distance dynamics remains the same irrespective of the character of the correlations .
see the methods section for further details.,scaledwidth=50.0% ] assuming a gaussian joint frequency distribution @xmath70 with identical single frequency variance @xmath71 and correlation coefficient @xmath72 , the time evolution of the trace distance corresponding to the optimal bell - state pair is found to be @xcite @xmath73.\ ] ] for uncorrelated photon frequencies , i.e. for uncorrelated local environments we have @xmath74 and the trace distance decreases monotonically , corresponding to markovian dynamics . however , as soon as the frequencies are anticorrelated , @xmath75 , the trace distance is non - monotonic which signifies quantum memory effects and non - markovian behavior . in our first experiment
the local decoherence induced by the quartz plates in arm @xmath8 and arm @xmath9 act consecutively , and we change the size of the initial correlations between the local reservoirs by tuning the spectral width of the pump pulse .
the results for the open system s trace distance dynamics are displayed in fig . 2 , which shows how the initial environmental correlations induce and influence the quantum non - markovianity .
first , when only the quartz plate in arm @xmath8 is active , the trace distance decreases monotonically demonstrating that information flows continuously from the system to the environment .
when subsequently the quartz plate in arm @xmath8 becomes inactive and the quartz plate in arm @xmath9 active , the trace distance increases again highlighting non - markovian behavior and a reversed flow of information from the environment back to the open system .
it is important to note that this increase of the trace distance implies a revival of the entanglement between the photon polarizations . whilst there exists earlier literature , e.g. , on how the entanglement is transferred from the open system to the environment @xcite , or on entanglement dynamics when the qubits interact with local non - markovian reservoirs @xcite ,
our results are fundamentally different .
we demonstrate that the entanglement can revive due to the nonlocal character of the dynamical map despite of the fact that the local interaction of the qubits with their environments is completely markovian .
a further important aspect of the experimental scheme is that it enables to measure the frequency correlation coefficient @xmath72 of the photon pairs by fitting the theoretical prediction of eq .
( [ d ] ) to the experimental data .
the fits shown in fig . 2 corresponding to panels ( a)-(d ) yield the values @xmath76 , @xmath77 , @xmath78 , and @xmath79 .
thus we see that the open system ( polarization degrees of freedom ) functions as a quantum probe which allows us to gain nontrivial information on the correlations in the environment ( frequency degrees of freedom ) . to the best of our knowledge ,
our results present the first quantitative measurement of the correlation coefficient without reconstructing the joint frequency distribution .
the open system dynamics of eq .
( [ rho12 ] ) depends solely on the decoherence functions @xmath80 , @xmath81 , @xmath82 and @xmath83 which in turn are completely determined by the fourier transform of the joint frequency distribution @xmath70 .
therefore , also the dynamical map of the open system depends only on @xmath70 and not on the probability amplitudes @xmath12 .
this fact leads to the conclusion that the nonlocal memory effects , and the revival of entanglement observed above , do not depend on the specific phase relations between different frequency components , and consequently , whether the initial frequency correlations between the photons are of quantum or classical character . to test
experimentally whether the nature of the correlations plays a role in inducing the memory effects , we insert quartz plates also on the path of the pump pulse , thereby changing the dispersion relations of the pump pulse while keeping fixed the correlation coefficient @xmath72 . the results in fig .
3 show that the trace distance dynamics remains the same in this case and only depends on the correlation coefficient @xmath72 of the photon frequencies .
thus we see that nonlocal memory effects and the revival of the entanglement between the polarization states of the photons can be induced by purely classical correlations between local environments of the open system .
trace distance dynamics for five different quartz plate configurations and a fixed correlation coefficient of @xmath76 . for all curves the maximal length of the quartz plates in both arms is @xmath55 , giving a maximal total effective path difference of @xmath84 . from top to bottom , the curves show the results for the cases , when we start to add quartz plates to arm @xmath9 when the length of the plate in arm @xmath8 has reached the length of @xmath85 respectively , see the methods section for details .
hence , in the top curve the quartz plates are added simultaneously to both arms while in the bottom curve they act consecutively .
the results demonstrate that the final state is the same while in terms of the information flow the behaviour is completely different . for
the top ( red ) curve information flows slowly out of the system with a small rate and , in fact , the decoherence would completely halt in the ideal ideal case of fully anticorrelated frequencies ( @xmath86 ) . in the other extreme , for the bottom
( violet ) curve information first flows out of the system with a high rate and later on returns back nearly completely indicating strong memory effects .
the solid lines are fits on the basis of eq .
( [ d ] ) , see the methods section for details , scaledwidth=50.0% ] in the previous experiments the local dephasing environments acted consecutively . in our final experimental series
we reposition the quartz plates in arm 2 so that eventually the dephasing is active simultaneously in both arms ( see the methods section ) .
the results are shown in fig .
we observe that the final state of the open system is always the same when the total effective path difference is equal to @xmath84 .
however , the amount of non - markovianity and the flow of information from the environment to the open system decreases with increasing overlap of the action of the quartz plates .
this demonstrates that initial correlations between the local environments and the reordering of the quartz plates allow to engineer the information flow between local systems while keeping fixed the final state .
it is also important to note that when the local dephasing channels act simultaneously on our open bipartite system ( the top red curve in fig . 4 )
, the initial correlations between the composite environments protect the entanglement between the qubits .
one can see this also from eq .
( [ d ] ) which shows that for @xmath87 and @xmath88 the trace distance remains constant and that there is thus no outflow of information from the system for the considered bell - states .
we have realized nonlocal quantum dynamical maps of a photonic system and demonstrated experimentally how initial quantum or even classical correlations between the local environments of a composite open system induce nonlocal memory effects . while the global dynamics of the open system is non - markovian , its local subsystems behave perfectly markovian .
the observation of the open system dynamics yields also important information about the environment .
in fact , we have seen that a measurement of the non - markovian evolution of the two - photon polarization state leads to a novel method for the determination of the correlation coefficient of the photon frequencies .
thus , the open system dynamics can be viewed as a non - markovian quantum probe which enables to extract nontrivial information on characteristic features of the environment .
this fact as well as our demonstration that initial correlations in the environment can protect the entanglement within an open system , could find important applications in quantum technologies and control strategies , and in the development of novel diagnostics tools for complex quantum systems .
the width of the pump spectrum controls the frequency anticorrelations between the pair of photons created in the down conversion process . by using filters and
fused silica plates , we can control the width of the pump pulses which are depicted in fig .
the laser source is filtered to @xmath89 ( fwhf ) and then passes through the frequency doubler ( @xmath90 thick bib@xmath91o@xmath92 ) .
hence , the bandwidth of the uv pulse is about @xmath93 as shown in fig . 5 ( b ) .
to obtain a sharper spectrum , as shown in fig . 5 ( a ) , we insert a thin fused silica plate , which is @xmath94 thick and coated with a partial reflecting coating on each side , with approximately @xmath95% at @xmath96 . for the spectra displayed in fig . 5 ( c ) and ( d )
, we have no filter before the doubler and no fused silica plate after the doubler , and for fig . 5 ( d ) we also replace the doubler by a @xmath97-thick bbo .
the details of the counting statistics and pump power for the experimental results of fig .
2 are as follows . for fig .
2 ( a ) the pump power is about @xmath98 and the total coincident count rate on the basis vectors hh , hv , vh , vv is about @xmath99 in @xmath100 seconds , ( b ) @xmath101 and the rate is about @xmath102 in @xmath103 seconds , ( c ) @xmath104 and the rate is about @xmath102 in @xmath9 seconds , and ( d ) @xmath104 and rate is about @xmath105 in @xmath103 seconds . please note that the coincident count rate always depends also on the bandwidth of uv pump source . employing eq .
( 3 ) , the theoretical fits of the experimental data presented in figs . 2 and 3 have been carried out as follows .
we first use the function @xmath106 to make a fit for the process when the quartz plates are added to arm 1 . from this process , we can find @xmath107 and @xmath108 which agree with the experimental data . then in the second part , when the quartz plate in arm 2 is active , we use the function @xmath109 $ ] . here , @xmath110 corresponds to the effective path difference after the first quartz plate .
since @xmath107 and @xmath108 are determined by the first fit , we can easily determine @xmath72 from the second fit . for fig . 4 , again keeping in mind eq .
( 3 ) , we can do the fitting in the following way .
first , we use the functions @xmath111 and @xmath112 mentioned above to make a fit of the lowest curve and determine @xmath113 , @xmath114 , and @xmath115 .
then we use a function @xmath116 $ ] to fit the top curve of the figure . from here , we determine @xmath117 , @xmath118 , and @xmath119 .
then we determine @xmath120 , @xmath121 , and @xmath122 which are then used to plot the functions @xmath123 $ ] in the figure . for the experimental results of fig . 3 ,
the fwhm of the pump pulse is about @xmath124 , the pump power is about @xmath104 , and the total coincident count rate on the basis vectors hh , hv , vh , vv is about @xmath102 in @xmath9 seconds .
the frequency spectra of the ultra - violet ( uv ) pump pulses used for the down conversion process .
full widths at half maximum ( fwhm ) of the pulse spectra are ( a ) @xmath42 , ( c ) @xmath44 , ( c ) @xmath46 , and ( d ) @xmath48 .
the solid lines represent gaussian fits of the experimental data.,scaledwidth=50.0% ] figure 6 shows the five different quartz plate configurations used for fig .
4 . to obtain the data of fig . 4 ,
the fwhm of the uv pulse is about @xmath125 , the pump power is about @xmath98 , and the total coincident count rate on the basis vectors hh , hv , vh , vv is about @xmath99 in @xmath100 seconds . the schematic view of the different quartz plate configurations .
the configurations from ( a ) to ( e ) correspond to the curves of fig . 4 from top to bottom .
for all cases the maximal length of the quartz plates in both arms is @xmath55 , giving a maximal total effective path difference of @xmath84 .
we start to add quartz plates to arm @xmath9 when the length of the plate in arm @xmath8 has reached the length of ( a ) @xmath65 , ( b ) @xmath126 , ( c ) @xmath127 , ( d ) @xmath128 , and ( e ) @xmath55 . ]
this work was supported by the national natural science foundation of china ( grant nos . 10874162 , 11074242 , 11104261 and 60921091 ) , the national basic research program of china , the foundation for the author of national excellent doctoral dissertation of pr china ( grant no .
200729 ) , the fundamental research funds for the central universities ( grant nos . 2030020004 and 2030020007 ) , the anhui provincial natural science foundation ( grant no .
11040606q47 ) , the academy of finland ( mobility from finland 259827 ) , the magnus ehrnrooth foundation ( finland ) , the jenny and antti wihuri foundation ( finland ) , the graduate school of modern optics and photonics ( finland ) , and the german academic exchange service ( daad ) .
b .- h.l . , d .-
y.c , y .- f.h . ,
c .- f.l . , and g .- c.g . planned , designed and implemented the experiments .
c .- f.l , g .- c .
g. , e .- m.l .
, h .- p.b . , and
j.p . carried out the theoretical analysis and developed the interpretation .
, c .- f.l . , e .- m.l , h .- p.b . , and j.p . wrote the paper and all authors discussed the contents .
the authors declare that they have no competing financial interests .
correspondence and requests for materials should be addressed to c .- f.l . or j.p .
tony j. g. apollaro , t. j. g. , di franco , c. , plastina , f. & paternostro , m. memory - keeping effects and forgetfulness in the dynamics of a qubit coupled to a spin chain . _
a _ * 83 * , 032103 ( 2011 ) . | the study of open quantum systems is important for fundamental issues of quantum physics as well as for technological applications such as quantum information processing .
recent developments in this field have increased our basic understanding on how non - markovian effects influence the dynamics of an open quantum system , paving the way to exploit memory effects for various quantum control tasks .
most often , the environment of an open system is thought to act as a sink for the system information .
however , here we demonstrate experimentally that a photonic open system can exploit the information initially held by its environment .
correlations in the environmental degrees of freedom induce nonlocal memory effects where the bipartite open system displays , counterintuitively , local markovian and global non - markovian character .
our results also provide novel methods to protect and distribute entanglement , and to experimentally quantify correlations in photonic environments . |
three - dimensional ( 3d ) topological insulators ( tis ) such as hgte , bi@xmath0se@xmath1 , bi@xmath0te@xmath1 , sb@xmath0te@xmath1 , and their alloys , are materials that are characterized by a narrow band gap in the electronic band structure while their surfaces host non - gapped spin - momentum locked states@xcite .
these helical surface states are topologically protected owing to the strong spin - orbit coupling in the material , and they mimic relativistic dirac electrons @xcite .
tis have generated much interest , not only as fundamentally new electronic states of matter , but also for their potential future technological applications in the fields of spintronics and quantum computation @xcite .
bismuth telluride ( bi@xmath0te@xmath1 ) is a 3d ti material , which has been investigated extensively @xcite .
the conducting surface states of bi@xmath0te@xmath1 have been observed first using surface - sensitive techniques like angle - resolved photoemission spectroscopy ( arpes ) @xcite and scanning tunneling microscopy ( stm ) , complemented by the scanning tunneling spectroscopy ( sts ) method @xcite .
nevertheless , at present it remains challenging to find unambiguous evidence of surface state transport in experiments on bi@xmath0te@xmath1 single crystals or thin films .
this is mainly because the charge transport is dominated by bulk conductivity due to residual carriers , which complicates the direct exploitation of the remarkable properties of the surface states .
recently , progress has been made in synthesizing a new generation of high - quality , bulk - insulating bi@xmath0te@xmath1 thin films using molecular beam epitaxy@xcite .
we have been able to grow bi@xmath0te@xmath1 thin films on different insulating substrates ( sto [ 111 ] and al@xmath0o@xmath1 [ 0001 ] ) with the fermi level placed in the bulk band gap@xcite .
combining different surface - sensitive in - situ characterization techniques , we demonstrated that the surface morphology and electronic band structure of bi@xmath0te@xmath1 are not affected by in - vacuo storage and exposure to oxygen , whereas major changes are observed when films are exposed to ambient conditions .
independently , combined in - situ four - point probe conductivity and angle resolved photoemission spectroscopy experiments revealed the bulk insulating properties of bi@xmath0te@xmath1 thin films with metallic surface states @xcite . in the work of hoefer _ _ et al.__@xcite , the authors also demonstrated that short exposure to air causes the bulk conduction band to be filled with electrons and the conductivity is no longer determined by the surface states alone .
these observations highlight the need to cap bi@xmath0te@xmath1 thin films in - situ before exposing them to air and other ex - situ contaminations in order to protect the surface states from degradation and unintentional doping .
different studies employed epitaxially - grown amorphous layers of se / te@xcite or an evaporated layer of al @xcite as a protective capping layer for bi@xmath0te@xmath1 thin films whereas others used al@xmath0o@xmath1 grown with atomic layer deposition ( ald ) @xcite . however , it was noticed that the stoichiometry of the films was altered after removal of the te / se capping layer@xcite and that the ald - grown al@xmath0o@xmath1 layer can cause damage to the bi@xmath0te@xmath1 surface states@xcite .
such damage decreases the surface mobility and generates a large number of impurity states which pin the surface chemical potential , a situation that will complicate an efficient electrostatic tuning of the fermi level @xcite . only recently , elemental te was identified as a suitable capping material for the protection of the topological surface of intrinsically insulating bi@xmath0te@xmath1 thin films@xcite , but systematic magnetotransport investigations of such capped films are needed .
further , an insulating capping layer is preferable because it allows for the fabrication of nearly damage - free top gate structures @xcite .
thus , it would be much more advantageous to develop a method for growing insulating capping layers with minimal damage to the surface states of intrinsically insulating bi@xmath0te@xmath1 thin films .
on the other hand , as the materials issues are being solved , the focus should now turn to finding an efficient way of controlling the fermi level inside the bulk band gap .
an effective gate control of topological surface states is highly desired for investigations of electronic transport properties in bi@xmath0te@xmath1 thin films and related device applications @xcite . to - date
, uniform gate - modulation of the carrier density and chemical potential over large areas and wide ranges has not been demonstrated in ( capped ) bi@xmath0te@xmath1 samples .
our first demonstration of this capability in this work is an important step toward the implementation of large scale hybrid devices for quantum information or spintronics applications with bi@xmath0te@xmath1 thin films .
we report bottom - gating results on in - situ capped , intrinsically insulating bi@xmath0te@xmath1 topological insulator thin films .
we follow the growth recipe presented in ref .
@xcite in order to obtain ti thin films with low densities of intrinsic carriers .
first , we focus on determining the proper in - situ capping procedure of our bi@xmath0te@xmath1 thin films grown on different insulating substrates , namely al@xmath0o@xmath1 and srtio@xmath1 .
we use two different capping materials : an al@xmath0o@xmath1 layer , grown by pulsed laser deposition ( pld ) , and an epitaxially grown te layer .
secondly , we report a systematic magnetotransport characterization of the in - situ capped films by studying the bottom - gate dependence of the sheet resistance , 2d carrier density ( @xmath4 ) and weak antilocalization behavior . here
, we used the srtio@xmath1 substrate as the back - gate dielectric .
lastly , we compare the transport characteristics of capped films with uncapped films of the same thickness and grown in the same conditions .
dual evidence from both , extracted carrier densities and weak antilocalization ( wal ) phenomena , suggests that the fermi level is in the bulk band gap for the in - situ capped films since capping prevents the formation of native oxide and isolates the film from different extrinsic defects @xcite .
the transport data on in - situ capped bi@xmath0te@xmath1 thin films suggest that we can use the back - gate to tune the fermi level all the way from the bulk conduction band ( bcb ) into the topological surface transport regime , and further down into the bulk valence band ( bvb ) .
these data reveal that capping the sample before taking it ex - situ for further electronic transport studies helps in preserving the surface states and avoiding unintentional doping induced by ambient conditions .
the high - quality bi@xmath0te@xmath1 thin films used in this work were grown by molecular - beam epitaxy ( mbe ) on insulating substrates , sto[111 ] and al@xmath0o@xmath1[0001 ] .
the growth procedure was similar to ref.@xcite . here , we compare magnetotransport properties of two samples of 15 nm thickness grown on sto .
one sample was capped with amorphous al@xmath0o@xmath1 and the other with an epitaxially - grown te layer .
the thickness of the protective capping layers of either material was @xmath5 6 nm .
for the purpose of understanding the role of capping , control samples without the protective layer were grown under the same conditions for comparison . before and after capping , the surface condition of films were checked using x - ray photoemission spectroscopy ( xps ) and reflection high energy electron diffraction ( rheed ) . for in - situ surface elemental characterization and detection of any te / bi excess in our films
, we used xps . figure [ fig : surfacecheck ] shows the xps spectra of the pristine ( blue ) film . only the bi and te peaks are resolved with no appearance of extra peaks , confirming that the film is clean without contaminations .
the analysis of the high resolution scans around the te @xmath6 and bi @xmath7 peaks for the pristine bi@xmath0te@xmath1 film ( see fig . [
fig : surfacecheck ] and ) reveal narrow and symmetric core level lines , signaling the absence of te or bi excess in our film .
the surface chemical stoichiometry was determined following the procedure presented in ref .
@xcite and the te : bi ratio was determined to be 1.497@xmath8 .
figure [ fig : surfacecheck ] shows a typical rheed pattern of the pristine bi@xmath0te@xmath1 film before being capped with an al@xmath0o@xmath1 layer .
the sharp rheed streaks are indicative of a high - quality film with smooth single crystalline domains . after depositing the al@xmath0o@xmath1 layer by pulsed laser deposition ( pld ) , the rheed pattern of the same sample is shown in fig .
[ fig : surfacecheck ] for comparison .
for the deposition , the films had been transferred to the pld chamber without breaking high vacuum ( hv ) conditions as the pld system is connected to the growth chamber via an hv distribution chamber .
the capping layer was deposited at room temperature in argon flow ( 40 sccm ) at a background pressure of 1@xmath9 mbar using a laser energy density of 2.75 j/@xmath10 . in order to minimize the degradation of the bi@xmath0te@xmath1 surface states due to high energy ionic bombardment of the sample surface during the pld process
, the sample holder was placed at a relatively far distance ( @xmath11 mm ) from the target material .
after the al@xmath0o@xmath1 deposition , only the oxygen and aluminum peaks were resolved in the xps spectrum ( red curve in fig .
[ fig : surfacecheck ] ) . for samples capped with te ,
the capping was also performed at room temperature , immediately after the growth of bi@xmath0te@xmath1 thin films in the mbe growth chamber without breaking the ultra - high vacuum conditions .
the surface elemental composition of these films was also investigated using xps . only bi and te peaks are resolved in the xps spectra before and after the deposition of the 6 nm te capping layer ( see fig .
[ fig : surfacecheck ] ) .
the intensities of the bi core levels in bi@xmath0te@xmath1 are observed to decrease after the capping step . from the analysis of the high resolution xps spectra ( see fig . [
fig : surfacecheck ] and ) , the bi @xmath7 spectral intensity is determined to have reduced by @xmath5 70% without any changes in the lineshape or energy position .
this indicates that the bi in bi@xmath0te@xmath1 is not chemically affected by the deposited 6 nm capping layer of te@xcite . on the other hand , the te @xmath12 and te @xmath13 peaks
are slightly influenced by this capping procedure since their lineshape peak positions are shifted by @xmath5 0.8 ev toward higher binding energies upon the deposition of the te overlayer .
this shoulder has been observed previously and was attributed to elemental te @xcite .
the xps data on te capped film are in agreement with recent in - situ arpes and four - point probe conductivity data on bi@xmath0te@xmath1 films @xcite , which revealed that the surface states of the pristine bi@xmath0te@xmath1 samples are not affected by the te capping , thus suggesting the te capping method does not cause a doping of the sample .
in order to investigate the electronic transport properties of capped films and to study the effect of the protective capping layer in magnetransport properties of in - situ capped and uncapped bi@xmath0te@xmath1 thin films , we fabricated hall bar devices using standard photolithography methods and subsequent wet etching . for the etching step , we used a solution of nitric acid ( hno@xmath1 ) diluted with de - ionized water ( 1 hno@xmath1 :3 h@xmath0o ) .
gold contacts were added in a subsequent lithography step . to achieve proper electric contact in the al@xmath0o@xmath1 capped samples , we removed the insulating layer in the contact area by etching the aluminum oxide with the same developer solution ( a base ) that we use for photolithography .
figure [ fig : transportcharacteristics ] shows an optical image of a typical hall bar device .
we used the sto [ 111 ] substrate as the back - gate dielectric since it has a high dielectric constant at low temperature ( @xmath14 ) .
the same approach has already been used for bi@xmath0se@xmath1 and ( bi@xmath15sb@xmath15)@xmath0te@xmath1 thin films , and bi@xmath0te@xmath1 nanoribbons also using sto [ 111 ] as the back - gate dielectric@xcite .
figure [ fig : transportcharacteristics ] depicts a schematic illustration of the bottom - gated thin film .
electrical transport properties of the films were first characterized in the temperature range from @xmath16k down to @xmath17k using a quantum design ppms measurement system .
later , we conducted magnetotransport measurements at different back - gate voltages in a dilution refrigerator at 100 mk .
the measurements were performed using a standard lock - in technique with a small excitation current of 5 na .
both , the sheet resistance @xmath18 and the hall resistance @xmath19 , were measured as a function of magnetic field at different back - gate voltages ( @xmath20 ) , in the range of @xmath21v .
sweeping the gate voltage back and forth , we observed hysteretic behavior , probably due to the presence of impurity states that trap charges as has been suggested by previous studies@xcite . in order to keep consistency in the data ,
all graphs presented here were taken by sweeping @xmath20 from positive to negative values .
figure [ fig : transportcharacteristics ] presents the temperature - dependent zero magnetic field sheet resistance for two different capped bi@xmath0te@xmath1 films ( black : capped with al@xmath0o@xmath1 , blue : capped with te ) , and for uncapped film ( red curve ) .
all three samples show metallic behaviour over the entire temperature range ; but the sheet resistance is lower in the uncapped film by about @xmath22 and @xmath23 at 2 k compared to the te and al@xmath0o@xmath1-capped films , respectively .
this observation is consistent with recent in - situ arpes and four probe stm transport measurements on pristine , capped and exposed to air bi@xmath0te@xmath1 thin films @xcite .
the decrease in @xmath18 was attributed to originate from exposing films to ambient conditions ; and corresponding arpes spectra on the same pristine and capped films proved that the fermi level is inside the bulk band gap prior and after in - situ transport measurements , which always showed metallic behaviour of the bi@xmath0te@xmath1 thin films@xcite .
we first discuss the magnetotransport properties of the al@xmath0o@xmath1-capped bi@xmath0te@xmath1 thin film .
the slope of @xmath19 indicates @xmath24-type carriers .
it increases with decreasing @xmath20 .
the @xmath19 curves start to show nonlinearity below @xmath25v ( see fig . [
fig : transportcharacteristics ] ) .
similar behavior has been reported in previous work on bi@xmath0se@xmath1 and ( bi@xmath26sb@xmath27)@xmath0te@xmath1 thin films grown on srtio@xmath1 [ 111 ] substrate @xcite .
we do not observe a reversal of the sign of @xmath19 in the entire magnetic field range in contrast to , e.g. , ( bi@xmath26sb@xmath27)@xmath0te@xmath1 samples in ref . @xcite . in the case of bi@xmath0se@xmath1 thin films grown on sto [ 111 ] , the change of the sign in the hall signal
has only been achieved at much higher magnetic fields and for large negative back - gate voltages @xcite .
we conjecture that the ambipolar effect is not observed within the accessible back - gate field range because of the `` m - like '' shape of the bulk valence band ( bvb ) structure of bi@xmath0te@xmath1 @xcite ( see fig . [
fig : walbehaviour ] and related discussion below ) .
the dirac point ( dp ) is buried inside the bvb , thus only @xmath24-type topological surface states carriers are accessible inside the bulk band gap , as it has also been recently suggested from gate - tunable transport studies of bi@xmath0te@xmath1 nanoribbon @xcite .
the nonlinearity in the hall resistivity curves below 50v suggests the coexistence of multiple charge carrier types .
we extracted the low - field hall coefficients ( @xmath28 ) which are calculated from the slope of the @xmath19 curves for @xmath29 t ( see inset fig . [
fig : transportcharacteristics ] ) . it should be emphasized here that , in the case where @xmath19(b ) is nonlinear , a multiple band model approach should be taken .
however , the discrepancies in carrier density and mobility of the high - mobility carriers ( corresponding to the low - field hall coefficient ) are small between one- and two - band fits to @xmath19 .
we were not able to obtain a simultaneous fit to both components of the magnetoresistance data , @xmath30 and @xmath19 , within the framework of the standard two - band model . a full theory of band structure and quantum corrections to the conduction of 3d ti surface and bulk bands is still missing .
thus , the 2d carrier density ( @xmath31 ) is extracted from @xmath32 , with the low - field hall coefficient @xmath28 as defined above , and @xmath33 the electronic charge .
the mobility is calculated using @xmath34 at zero magnetic field .
the response of sheet carrier density and sheet resistance of the al@xmath0o@xmath1-capped bi@xmath0te@xmath1 thin film to the applied back - gate voltage is summarized in fig .
[ fig : transportcharacteristics ] .
the first notable observation is the appearance of a maximum in the sheet resistance at v@xmath35v ( region ii ) , which is the first time , to our knowledge , that this behavior is seen in bi@xmath0te@xmath1 .
previously , it has only been observed in bi@xmath0se@xmath1 and ( bi@xmath26sb@xmath27)@xmath0te@xmath1 single crystals and thin films @xcite , and bi@xmath0te@xmath1 nanoribbons placed on a srtio@xmath1 substrate @xcite .
however , we would like to point out that the resistance maximum is not correlated with the depletion of the surface states at the dirac point ( dp ) , cp . refs .
@xcite , as the dp is buried below the bvb maximum .
we suggest that the downturn and saturation of the sheet resistance at negative gate voltages ( region iii ) is similar to what was observed in the measurements by lee _
@xcite on bi@xmath36sb@xmath15te@xmath37se@xmath38 single crystals .
the results were interpreted as an increase in interband scattering when the fermi level touches the bulk valence band ( see region iii in fig . [
fig : walbehaviour ] ) . in total
, we modulate the carrier density @xmath4 by a factor of @xmath2 in the accessible back - gate voltage range ( @xmath39 ) .
the biggest drop in carrier density is seen in region ii of fig .
[ fig : transportcharacteristics ] .
the lowest extracted carrier density is @xmath40 at @xmath41v with a corresponding mobility of @xmath42 .
the low carrier concentration in the intermediate region strongly suggests that the fermi level is located in the bulk band gap , crossing only topological surface states .
this is further corroborated by weak antilocalization measurements presented later .
furthermore , @xmath4 versus @xmath20 is observed to be linear as one would expect from a parallel plate capacitor model ( dashed line fit in fig .
[ fig : transportcharacteristics ] ) .
all added carriers participate in electric conduction in the intermediate @xmath20 regime , a behavior which is expected for surface states . at large , positive and at negative back - gate voltages , the linear relationship does not hold . in region
iii , the carrier density starts to increase slowly again as the bulk valence band contributes mobile carriers .
the dirac point is buried deeper in the bulk valence band in bi@xmath0te@xmath1 samples which makes it unaccessible by electrostatic gating in the gate voltage range of our experimental setup@xcite . in region
i , we add fewer high - mobility carriers by increasing the gate voltage .
we attribute this to the onset of bulk conduction when the fermi level crosses the bottom of the bulk conduction band ( bcb ) .
additional information about the carrier distribution and the scattering between transport channels is obtained from a measurement of the weak ( anti-)localization effect@xcite .
the wal effect in tis has been studied intensively both theoretically @xcite and experimentally @xcite .
wal is a quantum effect that is commonly observed in tis due to the spin - momentum locking resulting from strong spin - orbit coupling@xcite .
the @xmath43-berry s phase of the topological surface states leads to destructive interference between time - reversed charge carrier paths , thus increasing the conductivity at zero magnetic field .
the cusp - like wal maximum can be destroyed by applying an external magnetic field to the sample , which breaks the time reversal symmetry and @xmath44-berry s phase@xcite .
figure [ fig : walbehaviour ] gives the sheet conductance @xmath45 as a function of magnetic field for various @xmath20 measured at 100 mk .
the @xmath46 data show sharp cusp - like maxima at zero magnetic field , indicating the existence of the wal . at high positive @xmath20 ,
the wal effect is suppressed , containing parabolic background ( see inset fig . [
fig : walbehaviour ] ) .
the suppression of wal at high @xmath20 can be due to either enhanced bulk conductivity when @xmath47 is located in the bulk or disorder induced electron - electron interactions in the system @xcite .
the former scenario is much more likely in our samples since it is consistent with the @xmath31 analysis presented in fig .
[ fig : transportcharacteristics ] and is similar to the behaviors reported previously in gating experiments @xcite .
the wal behaviour is described quantitatively using the simplified hikami - larkin - nagaoka ( hln ) equation@xcite . in the limit of strong spin - orbit coupling ( @xmath48 ; where @xmath49 is the dephasing time , @xmath50 the spin - orbit scattering time , and @xmath51 the elastic scattering time ) , with a negligible zeeman term
, the magnetoconductance correction is given by : @xmath52;\ ] ] where @xmath53 is a parameter whose value reflects the number of conduction channels .
@xmath54 is the digamma function , @xmath55 is the characteristic magnetic field with @xmath56 , which is the phase coherent length , and @xmath57 the diffusion constant .
the relationship between the prefactor @xmath53 and the number of conducting channels in equation is essential in differentiating the nature of transport channels of tis@xcite .
for each two dimensional conducting channel , a value of @xmath58 is expected , and if there are two independent 2d conducting channels , the prefactors @xmath53 add up . an effective dephasing length @xmath59 will replace the phase relaxation length of the individual channel@xcite . fitting magnetoconductance data at different @xmath20 to equation , we observe a modulation of the prefactor @xmath53 and of the dephasing length @xmath60 as presented in fig .
[ fig : walbehaviour ] .
the values of @xmath53 drop from @xmath61 at high , positive @xmath20 ( region i ) , to a value of @xmath62 in the intermediate regime at @xmath63v ( region ii ) , and saturate at a value slightly smaller than @xmath64 for negative back - gate voltages ( region iii ) .
the same trend is observed in the @xmath20 dependence of @xmath60 .
for large gate voltage , @xmath65 , the phase coherence is long , @xmath66 m , but it reduces to @xmath67 nm when @xmath68v and flattens out for negative gate voltages ( region iii ) . the extracted @xmath60 is larger than the thickness of the sample ( 15 nm ) , indicating that the bulk channels , if present , are 2d with regard to wal and will be contributing to the measured wal signal when @xmath47 is located in the bulk band as pointed out previously for magnetotransport data of bi@xmath0te@xmath1 samples @xcite .
schematic band diagrams are depicted in fig .
[ fig : walbehaviour ] illustrating the evolution of the shift of the fermi level for the three different back - gate voltage regimes .
our data suggest that at low temperature , we can use the back - gate to tune the fermi level all the way from the bcb to the topological surface states ; and then to the bvb . in most cases ,
wal data in ti materials are governed by the competition between phase coherence time @xmath49 and the surface - to - bulk scattering time @xmath69 @xcite .
when the carrier density is high , and the fermi level ( @xmath70 ) is located inside the bulk conduction band ( bcb , fig .
[ fig : walbehaviour ] , right - hand side ) .
if substantial surface - to - bulk scattering is present , the effective phase coherence time will be much larger than the surface - to - bulk scattering time , @xmath71 , resulting in a single effective 2d transport channel with a prefactor of 0.5 to wal . in the opposite limit , @xmath72 , surface states and bulk
are decoupled .
each region constitutes a separate coherent channel leading to a correction to the sample conduction of the order of @xmath73 .
whereas the topological surface states contribute with a negative sign , a positive correction with @xmath74 is expected from the bulk conduction band when the fermi level is close to the bottom of the band@xcite .
the individual contributions cancel partially , and we are left with a conductance correction @xmath75 . after depleting the bulk carriers , region
ii ( fig .
[ fig : walbehaviour ] , center ) , the fermi level @xmath70 crosses only the topological surface states .
the two surfaces are decoupled , and we expect wal with @xmath76 , which is consistent with the observed value .
similar bottom - gating wal results with @xmath53 tuned from -0.5 to -1 has been reported in bi@xmath0se@xmath1 thin films grown on sto [ 111 ] @xcite and bi@xmath0te@xmath1 nanoribbons placed on a sto-100 substrate@xcite .
it should be pointed out that this scenario demands that both conduction channels have nearly identical dephasing fields , which requires good control of the fermi level and surface morphology @xcite . at negative @xmath20
, @xmath53 saturates at a value close to unity @xmath77 indicating that the surfaces stay decoupled as we lower the fermi level ( region iii in fig .
[ fig : walbehaviour ] left - hand side ) .
te@xmath1 film capped in - situ with a @xmath78 nm te layer .
( a ) back - gate voltage ( @xmath20 ) dependence of the 2d carrier density extracted from the zero - field slope of @xmath79 and of the sheet resistance .
( b ) back - gate dependence of the wal effect of an in - situ te - capped bi@xmath0te@xmath1 film .
the prefactor @xmath53 and the dephasing length @xmath59 as function of @xmath20 , extracted from best fits to equation as illustrated in the inset.,title="fig:",scaledwidth=45.0% ] in figure [ fig : transporttecappedfilms ] , we plot the carrier density @xmath80 and the sheet resistance of the te - capped sample .
the sheet carrier density values were extracted from the linear part of the low - field hall resistance @xmath81 .
the prefactor @xmath53 and the dephasing length @xmath59 obtained from fit to equ .
[ hln ] are shown in figure [ fig : transporttecappedfilms ] for different back - gate voltages .
we observe a similar characteristic as in the al@xmath0o@xmath1-capped bi@xmath0te@xmath1 film .
the change in the total number of ( mobile ) charge carriers is @xmath82 @xmath83 ( see fig . [
fig : transporttecappedfilms ] ) .
this variation in charge carriers is comparable to the values reported recently for bottom - gated se - capped bi@xmath0se@xmath1 films grown on an amorphous sio@xmath0/si substrate@xcite .
sweeping the back - gate , both the sheet resistance and the carrier concentration saturate at around @xmath84v , the lowest carrier density being @xmath85 @xmath83 at @xmath86v .
although we do not observe a maximum in the sheet resistance or the ambipolar effect in the @xmath81 data within the gate voltage range , the sharp increase in the sheet resistance together with the corresponding sharp drop in 2d carrier density suggest that we move the fermi level through the bulk band gap region of the bi@xmath0te@xmath1 band structure diagram . at these energies
the transport properties are dominated by the surface states as the bulk conductivity is significantly reduced @xcite . &
@xmath87 & + & uncapped & 4.3@xmath8810@xmath89 & 964 + & uncapped & 6.2@xmath8810@xmath89 & 915 + & & 9.5@xmath8810@xmath90&1206 + & & 3.8@xmath8810@xmath90&1594 + & uncapped & 3.8@xmath8810@xmath89&826 + & uncapped & 2.8@xmath8810@xmath89&1023 + & uncapped & 1.2@xmath8810@xmath91 & 41 + & & 1.8@xmath8810@xmath89&612 + fits of the magnetoconductance curves to equation ( see inset fig .
[ fig : transporttecappedfilms ] ) , were carried out for all gate voltages .
we observe a distinct modulation of the prefactor @xmath53 and the dephasing length @xmath59 with @xmath20 shown in fig .
[ fig : transporttecappedfilms ] .
the results agree qualitatively with the al@xmath0o@xmath1-capped bi@xmath0te@xmath1 sample in that both , @xmath53 and @xmath59 , decrease as @xmath20 is reduced .
again , @xmath53 approaches @xmath92 for @xmath93 v suggesting a single conduction channel coupled by a conducting bulk or , alternatively , a weak localization correction of @xmath94 to the conductance from a decoupled bulk conduction band @xcite . as the bulk carriers are depleted , @xmath95 , suggesting that each surface acts as an independent coherent transport channel with comparable dephasing fields .
to asses the role of the capping layers in protecting the topological surface states of bi@xmath0te@xmath1 thin films , we compare transport characteristics from capped films to uncapped ones .
we choose films of the same thickness ( 15 nm ) , which were grown under the same conditions .
table [ table1_2 ] summarizes the extracted low field carrier densities and the corresponding zero magnetic field mobilities for the two capped samples ( sample iv and viii ) , we presented in this work , and of a capped bi@xmath0te@xmath1 film grown on sapphire ( sample - iii ) together with five other uncapped bi@xmath0te@xmath1 films grown on sto .
clearly , we observe an improvement in the transport characteristics ( i.e. , lower intrinsic carrier densities while maintaining similar or high mobilities ) for the in - situ al@xmath0o@xmath1- and te - capped films .
we demonstrated the effectiveness of in - situ capping for preserving topological surface states in bulk insulating bi@xmath0te@xmath1 thin films .
for this purpose , we tested capping with al@xmath0o@xmath1 and te layers and found that both methods lead to a significant decrease in the measured carrier density by protecting the ti surfaces from adsorbates and other contaminations .
we back - gated the samples using the srtio@xmath1 substrate as a dielectric , and we found that we were able to vary the carrier densities by @xmath96 carriers in the gate voltage range of @xmath97 . in samples with low intrinsic doping , this modulation is sufficient to move from a bulk - dominated transport regime to a bulk insulating sample with decoupled surface states , which is evidenced by a doubling of the weak antilocalization correction to the sample conductance .
the ability to access ti surface states is an important step toward device applications in the field of quantum information and spintronics .
we showed that thin film deposition techniques of ti materials with low intrinsic doping paired with protective capping of the topological surface and effective depletion of bulk carriers by back - gating provide a technological platform for large - scale integration of such electronic devices .
this work is supported financially by the dutch foundation for fundamental research on matter ( fom ) , the netherlands organization for scientific research ( nwo ) and by the european research council ( erc ) .
1 l. fu , c. l. kane , and e. j. mele , phys .
lett . * 98 * , 106803 ( 2007 ) . c. l. kane and e. j. mele , phys .
* 95 * , 146802 ( 2005 ) .
h. zhang , c. -x .
liu , x. -l qi , x. dai , z. fang , s. -c .
zhang , nature phys . * 5 * , 438 ( 2009 ) .
j. e. moore , nature * 464 * , 194 ( 2010 ) .
d. hsieh , y. xia , d. qian , l. wray , j. h. dil , f. meier , j. osterwalder , l. patthey , j. g. checkelsky , n. p. ong , a. v. fedorov , h. lin , a. bansil , d. grauer , y. s. hor , r. j. cava and m. z. hasan , nature * 460 * , 1101 ( 2009 ) . c. jozwiak ,
park , k. gotlieb , c. hwang , d .- h .
lee , s. g. louie , j. d. denlinger , c. r. rotundu , r. j. birgeneau , z. hussain and a. lanzara , nature phys .
* 9 * , 293 ( 2013 ) .
y. l. chen , j. g. analytis , j .- h .
chu , z. k. liu , s .- k .
mo , x. l. qi , h. j. zhang , d. h. lu , x. dai , z. fang , s. c. zhang , i. r. fisher , z. hussain and z .- x .
shen , science * 325 * , 5937 ( 2009 ) .
z. alpichshev , j. g. analytis , j. -h .
chu , i. r. fisher , y. l. chen , z. x. shen , a. fang , and a. kapitulnik , phys .
104 * , 016401 ( 2010 ) .
p. ngabonziza , r. heimbuch , n. de jong , r. a. klaassen , m. p. stehno , m. snelder , a. solmaz , s. v. ramankutty , e. frantzeskakis , e. van heumen , g. koster , m. s. golden , h. j. w. zandvliet , and a. brinkman , phys .
b * 92 * , 035405 ( 2015 ) .
k. hoefer , c. becker , d. rata , j. swanson , p. thalmeier , and l. h. tjeng , proc .
usa 111 , 14979 ( 2014 ) .
k. hoefer , c. becker , s. wirth , and l. h. tjeng , aip adv . * 5 * , 097139 ( 2015 ) .
s. e. harrison , b. zhou , y. huo , a. pushp , a. j. kellock , s. s. p. parkin , j. s. harris , y. chen , and t. hesjedal , appl .
lett . * 105 * , 121608 ( 2014 ) .
k. virwani , s. e. harrison , a. pushp , t. topuria , e. delenia , p. rice , a. kellock , l. collins - mcintyre , j. harris , t. hesjedal , and s. parkin , appl .
. lett . * 105 * , 241605 ( 2014 ) .
s. e. harrison , s. li , y. huo , b. zhou , y. l. chen , and j. s. harris , appl .
phys . lett . * 102 * , 171906 ( 2013 ) .
yu , l. he , m. lang , w .
jiang , f. xiu , z. liao , y. wang , x. kou , p. zhang , j. tang , g. huang , j. zou and k. l. wang , nanotechnology * 24 * , 015705 ( 2013 ) .
h. liu and p. d. ye , appl .
. lett . * 99 * , 052108 ( 2011 ) .
f. yang , a. a. taskin , s. sasaki , k. segawa , y. ohno , k. matsumoto , and y. ando , appl .
. lett . * 104 * , 161614 ( 2014 ) .
j. wiedenmann , e. bocquillon , r.s .
deacon s. hartinger1 , o. herrmann , t.m .
klapwijk , l. maier , c. ames1 , c. brne , c. gould , a. oiwa , k. ishibashi , s. tarucha , h. buhmann and l.w .
molenkamp , nat . commun .
* 7 * , 10303 ( 2016 ) . c. kurter
, a. d. k. finck , p. ghaemi , y. s. hor , and d. j. van harlingen , phys . rev .
b * 90 * , 014501 ( 2014 ) .
x. he , t. guan , x. wang , b. feng , p. cheng , l. chen , y. li , and k. wu , appl .
* 101 * , 123111 ( 2012 ) .
h. yuan , h. liu , h. shimotani , h. guo , m. chen , q. xue , and y. iwasa , nano lett . * 11 * , 2601 ( 2011 ) . j. xiong , y. khoo , s. jia , r. j. cava , and n. p. ong , phys .
b * 88 * , 035128 ( 2013 ) .
m. lang , l .
xiu , x. yu , j. tang , y. wang , x. kou , w .
jiang , a. v. fedorov , and k. l. wang , nano lett .
* 6 * , 295 ( 2012 ) .
j. moulder and j. chastain , _ handbook of x - ray photoelectron spectroscopy : a reference book of standard spectra for identification and interpretation of xps data _ ( physical electronics division , perkin - elmer corporation , 1992 ) .
g. zhang , h. qin , j. chen , x. he , l. lu , y. li and k. wu , adv .
mater . * 21 * , 2351 ( 2011 ) .
j. chen , x. y. he , k. h.wu , z. q. ji , l. lu , j. r. shi , j. h. smet , and y. q. li , phys .
b * 83 * , 241304(r ) ( 2011 ) .
j. chen , h. j. qin , f. yang , j. liu , t. guan , f.m .
qu , g. h. zhang , j. r. shi , x. c. xie , c. l. yang , k. h.wu , y. q. li , and l. lu , phys . rev .
lett . * 105 * , 176602 ( 2010 ) .
l. a. jauregui , m. t. pettes , l. p. rokhinson , l. shi and y. p. chen , sci . rep . * 5 * , 8452 ( 2015 ) .
f. xiu , l. he , y. wang , l. cheng , l .- t .
chang , m. lang , g. huang , x kou , y. zhou , x. jiang , z. chen , j .
zou , a. shailos and k. l.wang , nat . nano .
* 6 * , 216 ( 2011 ) .
j. tian , c. chang , h. cao , k. he , x. ma , q. xue , and y. p. chen , sci .
rep . * 4 * , 4859 ( 2014 ) . j. lee , j. park , j. lee , j. kim , and h. lee , phys . rev .
b * 86 * , 245321 ( 2012 ) . h. z. lu and s. q. shen , phys
b * 84 * , 125138 ( 2011 ) .
i. garate and l. glazman , phys .
b * 86 * , 035422 ( 2012 ) . s
chiu and j .- j .
lin , phys .
b * 87 * , 035122 ( 2013 ) .
lu and s .- q .
shen , phys .
* 112 * , 146601 ( 2014 ) .
s. maekawa , and h. fukuyama , j. of the physical society of japan * 50*(8 ) , 2516 ( 1981 ) .
j. wang , a. m. dasilva , c .- z .
chang , k. he , j. k. jain , n. samarth , x .- c .
ma , q .- k .
xue , and m. h. w. chan , phys .
b * 83 * , 245438 ( 2011 ) .
h. steinberg , j .- b .
lalo , v. fatemi , j. s. moodera , and p. jarillo - herrero , phys .
b * 84 * , 233101 ( 2011 ) .
s. hikami , a. i. larkin and y. nagaoka , prog .
. phys . * 63 * , 707 ( 1980 ) .
he , g. wang , t. zhang , i .- k .
sou , g. k. l. wong , and j .-
wang , phys .
106 * , 166805 ( 2011 ) .
d. kim , p. syers , n. p. butch , j. paglione , and m. s. fuhrer , nat .
* 4 * , 2040 ( 2013 )
. m. lang , l. he , x. kou , p. upadhyaya , y. fan , h. chu , y. jiang , j. h. bardarson , w. jiang , e. s. choi , y. wang , n. c. yeh , j. moore , and k. l. wang , nano lett .
* 13 * , 48 ( 2013 ) .
m. brahlek , n. koirala , m. salehi , n. bansal , and s. oh , phys .
lett . * 113 * , 026801 ( 2014 ) . y. h. liu , c.w .
chong , j. l. jheng , s. y. huang , j. c. a. huang , z. li , h. qiu , s. m. huang and v. v. marchenkov , appl .
. lett . * 107 * , 012106 ( 2015 ) .
s. s. hong , j. j. cha , d. kong , and y. cui , nat .
* 3 * , 757 ( 2012 ) .
y. s. kim , m. brahlek , n. bansal , e. edrey , g. a. kapilevich , k. iida , m. tanimura , y. horibe , s. w. cheong , and s. oh , phys .
b * 84 * , 073109 ( 2011 ) .
d. kim , p. syers , n. p. butch , j. paglione , and m. s. fuhrer , nat .
* 4 * , 2040 ( 2013 ) . | combining the ability to prepare high - quality , intrinsic bi@xmath0te@xmath1 topological insulator thin films of low carrier density with in - situ protective capping , we demonstrate a pronounced , gate - tunable change in transport properties of bi@xmath0te@xmath1 thin films . using a back - gate , the carrier density is tuned by a factor of @xmath2 in al@xmath0o@xmath1 capped bi@xmath0te@xmath1 sample and by a factor of @xmath3 in te capped bi@xmath0te@xmath1 films .
we achieve full depletion of bulk carriers , which allows us to access the topological transport regime dominated by surface state conduction .
when the fermi level is placed in the bulk band gap , we observe the presence of two coherent conduction channels associated with the two decoupled surfaces .
our magnetotransport results show that the combination of capping layers and electrostatic tuning of the fermi level provide a technological platform to investigate the topological properties of surface states in transport experiments and pave the way towards the implementation of a variety of topological quantum devices . |
while most of the interstellar gas mass in elliptical galaxies , typically @xmath0 @xmath1 , is at the galactic virial temperature @xmath2 k , the cores of @xmath3 percent of all large e galaxies contain dusty clouds of cold gas of mass @xmath4 @xmath1 ( e.g. van dokkum & franx 1995 ) .
in addition , diffuse optical line emission from gas at @xmath5 k is observed in most or all elliptical galaxies within a few hundred parsecs from the center ( e.g. caon et al .
2000 ) , particularly in those galaxies formerly thought to contain `` cooling flows . ''
the notion that the hot gas cools to low temperatures is not supported by xmm spectra ( e.g. peterson et al .
2001 ; xu et al .
2001 ) , i.e. , the observed cooling rate @xmath6 is lower than that expected in traditional cooling flows ,
@xmath7 @xmath1 yr@xmath8 . in any case , cooling of essentially dust - free hot gas can not produce dust clouds . in the following
we propose that the dusty clouds are formed from gas expelled from evolving metal - rich red giant stars in e galaxy cores .
even if these dusty stellar envelopes are heated to @xmath9 after leaving the stars , the gas can cool rapidly by collisions between thermal electrons and dust grains .
direct evidence that mass is being lost from evolving stars in x - ray luminous ellipticals is provided by near - ir ( @xmath10 m ) observations of hot circumstellar dust ( knapp et al . 1989 , 1992 ; athey et al .
the 9.7 @xmath11 m peak from ( oxygen - rich ) silicate grains apparent in iso observations places an upper limit on the grain radius , @xmath12 @xmath11 m ( figure 5 of laor & draine 1993 ) . according to athey et al .
, the total luminosity of excess near - ir radiation is consistent with expected total stellar mass loss rates in giant e galaxies , @xmath7 @xmath1 yr@xmath8 . however , as dusty gas moves away from orbiting , mass - losing red giant stars , it encounters the hot interstellar gas at @xmath13 kev and is violently disrupted by hydrodynamic instabilities .
the interaction of the stellar wind / planetary nebula with the hot gas is extremely complicated , but this ejected gas can not remain at @xmath5 k for more than @xmath14 years without exceeding the typical h@xmath15 + [ nii ] luminosities observed ( mathews & brighenti 1999 ) .
xmm spectra show no evidence for slowly heating gas at @xmath16 .
this empirical evidence supports rapid thermal mixing of the stellar ejecta , which is probably a consequence of gasdynamical instabilities that enormously increase the surface area between the stellar ejecta and the ambient hot gas . in view of the complexity of this interaction ,
the thermal evolution @xmath17 of gas ejected from stars is difficult to estimate , but limits can be set . in the following discussion
we adopt a useful limiting assumption for @xmath17 : the dusty gas is rapidly heated to the temperature of the local hot gas with the dust intact .
this assumption maximizes the likelihood that the stellar gas will remain thermally merged with the hot gas phase .
grain sputtering during the rapid heating process is unlikely since the heating time must be @xmath18 , which is very much less than the sputtering time ( see below ) .
as soon as the dusty gas is heated ( by assumption ) to the local gas temperature , it will begin to cool .
cooling occurs by thermal x - ray emission from the gas and by inelastic impacts of thermal electrons with the grains .
grain - electron cooling can greatly exceed that from thermal x - ray emission ( e.g. dwek 1987 ; dwek & arendt 1992 ) provided the grains are not sputtered away . if the cooling time is shorter than the dynamical time in the galactic potential , the gas will tend to cool approximately _ in situ _ , the case we consider here . during isobaric cooling
, the gas temperature decreases according to @xmath32 here @xmath33 is the usual thermal radiation coefficient for gas with metallicity @xmath34 in solar units ( sutherland & dopita 1993 ) and @xmath35 is the coefficient for dust - induced cooling .
the cooling of the hot plasma by grains is due mostly to inelastic impacts of thermal electrons with the grains .
for high temperature plasmas , such as we consider , the rate of electron - grain collisions is unaffected by the grain charge ( draine & salpeter 1979 ) which we assume is regulated by emission of low energy electrons from the grain surface .
an individual grain cools the local gas at a rate @xmath36 where @xmath37 and @xmath38 are the mean energy and velocity of electrons , @xmath39 and @xmath40 is the grain radius in microns . at pressures of interest , @xmath41 , the plasma mean free path , @xmath42 cm , is small ;
only electrons in the cooling regions collide with the grains ; electrons from the hot ambient gas thermalize in the cooling plasma before colliding with a grain . in ngc 4472 , a giant elliptical in the virgo cluster which we use for reference , @xmath43 @xmath44 so @xmath45 and @xmath46 at @xmath47 kpc and the effective radius @xmath48 kpc respectively .
we adopt a distance of 17 mpc to ngc 4472 .
grains can internally absorb the energy of thermal electrons provided @xmath49 kev ( dwek & werner 1981 ) , which we assume is generally satisfied .
cooling by electron impact on grains during the sputtering time , @xmath50 yrs , greatly exceeds the additional gas cooling required to fully ionize the grain atoms ejected in the sputtering process or to bring these sputtered ions up to local thermal velocities . at any moment
the space density of interstellar grains @xmath51 in luminous elliptical galaxies is very inhomogeneous , being concentrated near mass - losing stars .
if @xmath52 is the typical total mass of the stellar envelope ejected by evolving stars and @xmath53 gm @xmath44 is the density of ( silicate ) grains , then the total number of grains expelled by a single star is @xmath54 where @xmath55 is the ratio of the dust mass to the gas mass when ejected from the stars .
we assume that @xmath56 decreases with stellar metallicity as in ngc 4472 where @xmath57 in solar units ( brighenti & mathews 1999 ) .
the initial space density of grains is proportional to the local gas density @xmath58 . as the gas cools isobarically from some initial temperature @xmath59 , the grain space density increases , @xmath60 ( \mu p / k)/t$ ] . the dust cooling coefficient in equation 2 , @xmath61 ,
is then @xmath62 to explore the influence of grain cooling we integrate equations 1 and 2 beginning with gas at temperature @xmath63 and pressure @xmath64 at galactic radius @xmath65 in the hot gaseous atmosphere of ngc 4472 .
the pressure @xmath66 is constant during cooling . for simplicity , we assume the gas is filled with grains all having the same initial radius @xmath67 . as the gas cools the sputtered debris from the grains increases the metallicity of the gas phase . to evaluate the plasma cooling rate coefficient @xmath33 during the cooling
, we assume that the gas phase metal abundance in solar units increases with decreasing grain mass ( @xmath68 ) , @xmath69 , where @xmath70 is the maximum ratio of dust to gas mass if all elements heaver than he were in grains .
we describe solutions of equations 1 and 2 for the elliptical galaxy ngc 4472 with initial conditions at radii @xmath47 and @xmath48 kpc . according to our assumption ,
the dusty gas expelled from a star is rather quickly heated to the local temperature of the hot gas at these two galactic radii , @xmath71 and @xmath72 k , where the gas density is @xmath73 and @xmath74 @xmath44 respectively .
figures 1a and 1b show the combined evolution of the gas temperature @xmath17 and grain radius @xmath75 for grains of initial radius @xmath76 and 0.1 micron , assuming @xmath77 and @xmath78 at @xmath47 and @xmath79 kpc respectively . for comparison
we also plot the cooling curve @xmath17 of dust - free gas ( @xmath80 ) with the same @xmath59 and @xmath81 .
dust - enhanced cooling can reduce the cooling time of recently ejected gas by over an order of magnitude compared to the dust - free case .
for fixed @xmath55 , the ratio @xmath82 increases rapidly with decreasing grain radius @xmath67 because of the larger total surface area of smaller grains .
a defining characteristic of cooling flows is that the isobaric radiative cooling time @xmath83 yrs ( in ngc 4472 ) exceeds the dynamical time @xmath84^{1/2}$ ] in the galactic potential .
dust - free , moderately positive density perturbations do not cool appreciably faster than unperturbed regions . if overdense regions remain coherent , they oscillate radially in the hot gas atmosphere , becoming underdense relative to the ambient gas during half their cycle ( e.g. loewenstein 1989 ) .
the dust - rich regions we describe here are not initially in an overdense state , but become overdense as the cooling proceeds .
the dust - rich gas cools locally if the cooling time is less than the dynamical time at the two selected galactic radii , @xmath85 at @xmath47 kpc and @xmath86 yrs at 8.57 kpc .
figures 1a and 1b show that @xmath87 is possible at @xmath47 kpc , so gas at small galactic radii can cool to low temperature and also preserve a small amount of its original dust since @xmath88 .
the cooling curves in figures 1a and 1b are independent of @xmath52 and are therefore unchanged if the dusty envelope is disrupted or fragmented during its interaction with the hot gas .
we show one case in figure 1b in which dusty gas with @xmath89 @xmath11 m begins cooling at @xmath90 , before fully thermalizing with the hot gas . the cooling time is significantly reduced , confirming our assumption that @xmath91 maximizes the cooling time .
as dust - rich stellar winds flow away from orbiting red giant stars , the winds are decelerated by interactions with the local hot gas , but the grains , impelled by their higher momentum , can move into the ambient gas .
it is possible in principle that the ejected dust cloud occupies and cools a gas mass larger than that ejected from the star , @xmath52 .
if the cooled mass exceeds @xmath52 by a factor @xmath92 , then the corresponding cooling evolution is found simply by replacing @xmath55 in equation 4 with @xmath93 . in figure 1c
we show cooling curves for @xmath94 at the same two galactic radii for grains of initial radius @xmath89 micron .
both cooling times exceed the dynamical time at the galactic radii considered , so the cooling gas may have a more complex dynamical evolution that may inhibit the cooling .
stellar mass loss has been regarded as an important internal source of hot gas within group or cluster - centered e galaxies .
a @xmath95 gyr old stellar population of total mass @xmath96 typically expels @xmath97 @xmath1 yr@xmath8 . in cluster - centered galaxies , such as m87 in virgo and ngc 4874 in the coma cluster ,
the gas temperature rapidly decreases in the central @xmath98 kpc from the virial temperature of the cluster ( 3 - 8 kev ) to the stellar virial temperature @xmath7 kev at @xmath99 kpc ( molendi 2002 ; vikhliin et al .
detailed gasdynamical models of traditional cooling inflows in these cluster - centered galaxies indicate that this steep temperature gradient can only be understood if gas is being cooled by thermalization of stellar ejecta ( brighenti & mathews 2002a ) .
the increase in the hot gas oxygen abundance toward the center of m87 ( gastaldello & molendi 2002 ) is expected if mass lost from the snii - enriched stars has thermally mixed into the hot ism and if the stars are more o - rich than the virgo cluster gas .
nevertheless , we propose here that the dust component can cool the stellar ejecta soon after it thermally merges with the ambient hot gas , particularly at @xmath100 kpc .
\(1 ) _ what is the source of interstellar dust observed in the cores of @xmath3 percent of all large e galaxies ? _
van dokkum & franx ( 1995 ) and others describe small central dust disks , lanes or clouds typically a few 100 pc in size .
the small masses of dust @xmath4 @xmath1 in these cores can easily be produced in @xmath101 yrs as dust - rich gas cools near the galactic center with incomplete sputtering .
it is difficult to accurately estimate the rate that dust accumulates near the center without knowledge of the dust size distribution , the exact thermal history @xmath17 of the stellar ejecta , or the star formation rate in the cold , dusty clouds . since type ia supernova remnants occupy a very small fraction of the interstellar volume at any time ( mathews 1990 ) , heating by type ia supernovae is unlikely to interfere with the cooling we describe here .
heating by active galactic nuclei has often been suggested to explain why the hot gas in e galaxies fails to cool to low temperatures ( e.g. , rosner & tucker 1989 ; binney & tabor 1995 ; brighenti & mathews 2002b ) .
if such heating occurs , it must be gentle enough not to destroy the observed central dust clouds or this dust must be rapidly regenerated .
\(2 ) _ what is the source of the diffuse optical line emission ( h@xmath15 + [ nii ] ) observed in most or all cooling flow galaxies ? _
the velocities of this diffuse emission are unrelated to stellar velocities ( caon et al . 2000 ) . in some e galaxies
the optical line emission spatially correlates with x - ray features ( e.g. trinchieri & goudfrooij 2002 ) or traces the perimeters but not the centers of x - ray cavities ( mcnamara , oconnell & sarazin 1996 ; blanton et al .
2001 ) . while these observations suggest that some gas cools to @xmath102 k from the hot phase , there is no evidence for this in _ xmm _ x - ray spectra
however , even if dusty gas ejected from metal - rich stars is heated to @xmath9 , it can quickly cool back to @xmath102 k where the cooling may be temporarily arrested by absorption of galactic uv starlight ( binette et al .
exactly what happens next is unclear , but the dust can help cool the gas to much lower temperatures if the clouds become optically thick to uv radiation .
it has often been suggested that dusty gas at @xmath5 k derives from mergers with gas - rich dwarf galaxies ( e.g. caon et al .
2000 ; trinchieri & goudfrooij 2002 ) ; this can be verified if the @xmath103 k gas is counter - rotating .
nevertheless , cold , dusty gas can arise naturally from stars in e galaxy cores .
\(3 ) _ why is the observed cooling rate @xmath6 from x - ray observations so low ? _ xu et al . ( 2001 ) found from rgs xmm observations of elliptical galaxy ngc 4636 that x - ray lines expected from gas cooling near @xmath104 k are unusually weak , indicating a total cooling rate of @xmath105 @xmath1 yr@xmath8 in @xmath106 kpc ( for @xmath107 mpc ) .
fuse observations of ovi lines in ngc 4636 ( emitted at @xmath108 k ) indicate @xmath109 @xmath1 yr@xmath8 within 1.2 kpc ( bregman et al .
both cooling rates are less than the @xmath110 @xmath1 yr@xmath8 predicted in traditional cooling flow models ( e.g. bertin & toniazzo 1995 )
. rapid dust - assisted cooling may be relevant to this discrepancy in two ways : ( 1 ) by reducing the total rate that gas enters the hot phase , and ( 2 ) by reducing the x - ray emission from cooling thermal gas at subvirial temperatures . in models of traditional cooling flows , gas ejected from stars
is usually assumed to enter the hot interstellar gas .
the stellar mass loss rate in a giant e galaxy , @xmath111 yr@xmath8 , is an important source of gas since it is comparable to the expected cooling rate of the hot gas . in gas dynamical models
the cooling rate @xmath6 near the center of the flow varies inversely with the specific rate of stellar mass loss , @xmath112 .
this is true even when there is an extended reservoir of circumgalactic hot gas due to cosmic accretion onto the surrounding galaxy group .
therefore , the apparent cooling rate @xmath6 observed with xmm should be reduced approximately in proportion to the fraction of stellar gas that fails to enter the hot gas because of dust - assisted cooling .
however , if more than @xmath113 percent of the stellar ejecta fails to enter the hot phase , our gasdynamical models indicate that the density of hot interstellar gas is lowered sufficiently to initiate a strong galactic wind driven by type ia energy .
the transition to wind flows is rather sudden . since the very low @xmath114 characteristic of strong winds
are not observed , this sets a limit on the efficiency of dust - assisted cooling described here .
if dusty gas from stars is indeed heated to @xmath9 , xmm observations require that it not emit thermal x - rays at intermediate temperatures either as it is heated or as it cools afterward .
we argue above that the heating phase is likely to be rapid , @xmath115 yrs .
when heated to @xmath9 , the stellar gas may undergo rapid dust - assisted cooling , during which the x - ray emission from intermediate temperatures is greatly reduced .
the radiation expected from an x - ray line of emissivity @xmath116 ergs s@xmath8 gm@xmath8 , @xmath117 ergs gm@xmath8 , varies inversely with the cooling rate @xmath118 , which can be @xmath119 times larger than normal plasma cooling .
the x - ray line emission is suppressed by a similar factor . in dust - assisted cooling ,
much of the thermal energy in the hot gas is radiated by dust in the far - ir .
the maximum fir luminosity expected , @xmath120 erg s@xmath8 , is consistent with iso fir luminosities observed in ngc 4472 and ngc 4636 , @xmath121 erg s@xmath8 ( temi et al .
2003 ) .
nevertheless , it seems unlikely that dust - enhanced cooling can explain the weakness of x - ray cooling lines from gas at subvirial temperatures in clusters ( e.g. peterson et al .
2001 ) where the expected cooling rates , @xmath122 @xmath1 yr@xmath8 , greatly exceed the stellar mass loss rate in the central e or cd galaxy .
.4 in studies of the evolution of hot gas in elliptical galaxies at uc santa cruz are supported by nasa grants nag 5 - 8049 & atp02 - 0122 - 0079 and nsf grants ast-9802994 & ast-0098351 for which we are very grateful .
fb is supported in part by grants murst - cofin 00 and asi - ars99 - 74 .
athey , a. , bregman , j. , bregman , j. , temi , p. & sauvage , m. 2002 , apj 571 , 272 bertin , g. & toniazzo , t. 1995 , apj , 451 , 111 binette , l. et al .
1994 , a&a , 292 , 13 binney , j. & tabor , g. 1995 , mnras 276 , 663 blanton , e. l , sarazin , c. l. & irwin , j. a. 2001 , apj , 552 , 106 bregman , j. n. , miller , e. d. & irwin , j. a. 2110 , apj 553 , l125 brighenti , f. & mathews , w. g. 2002a , apj 573 , 542 brighenti , f. & mathews , w. g. 2002b , apj 567 , 130 brighenti , f. & mathews , w. g. 1999 , apj 515 , 542 caon , n. , macchetto , d. & pastoriza , m. 2000 , apjs 127 , 39 draine , b. t. & salpeter , e. e. 1979 , apj 231 , 77 dwek , e. 1987 , apj , 322 , 812 dwek , e. & arendt , r. g. 1992 , ara&a 30 , 11 dwek , e. & werner , m. w. 1981 , apj 248 , 138 gastaldello , f. & molendi , s. 2002 , apj 572 , 160 knapp , g. r. , guhathakurta , p. , kim , d - w . ,
jura , m. a. 1989 , apjs 70 , 329 knapp , g. r. , gunn , j. e. , wynn - williams , c. g. 1992 , apj 399 , 76 laor , a. & draine , b. t. 1993 , apj , 402 , 441 loewenstein , m. 1989 , mnras 238 , 15 mathews , w. g. 1990 , apj 354 , 468 mathews , w. g. & brighenti , f. 1999 , apj , 526 , 114 mcnamara , b. r. , oconnell , r. w. & sarazin , c. l. 1996 , aj , 112 , 91 molendi , s. 2002 apj ( submitted ) ( astro - ph/0207545 ) peterson , j. r. et al .
2001 , a&a 365 , 104 rosner , r. & tucker , w. h. 1989 , apj 338 , 761 sutherland , r. s. & doptia , m. a. 1993 , apjs , 88 , 25 temi , p. et al .
2003 , apj 585 , l121 trinchieri , g. & goudfrooij , p. 2002 , a&a 386 , 472 tsai , j. c. & mathews , w. g. 1995 , apj , 448 , 84 van dokkum , p. g. & franx , m. 1995 , aj 110 , 2027 vikhlinin , a. , et al .
2001 , apj 2001 , apj 555 , l87 xu , h. et al .
2001 , apj ( submitted ) ( astro - ph/0110013 ) | we propose a stellar origin for the central dust clouds observed in most giant elliptical galaxies .
dusty gas ejected from evolving red giant stars in e or cd galaxies can cool rapidly even after entering the hot , x - ray emitting gas . cooling by thermal collisions with dust grains can be faster than either the dynamical time in the galactic potential or the grain sputtering time .
some grains survive in the cooled gas .
dusty stellar outflows cool more efficiently in the central regions where the stellar metallicity is higher .
mergers with gas and dust - rich dwarf galaxies may occasionally occur but are not required to explain the observed dust clouds .
.2 in |
if @xmath0 are continuous operators on a hilbert space , we may write @xmath1 where @xmath2 is a universal , non commutative , homogeneous polynomial of degree @xmath3 with rational coefficients . the product @xmath4 is called , after @xcite , baker - campbell - hausdorff ( bch ) product : it is associative and the bch theorem asserts that every polynomial @xmath2 is a lie element , i.e. , is a linear combination of nested commutators .
however , the proof of the bch theorem does not give directly an explicit description of @xmath2 as a lie element ; moreover , such description is not unique in view of the jacobi identity .
the most famous explicit expression of @xmath5 , in terms of nested commutators , is probably the one due to e. dynkin ( see ( * ? ? ?
* equation 1 ) or ( * ? ? ? * equation 1.7.3 ) ) : @xmath6 where @xmath7 $ ] is the adjoint operator , the second sum is over all possible combinations of @xmath8 such that @xmath9 , for @xmath10 , and @xmath11 .
the literature about baker - campbell - hausdorff formula is huge .
for instance : in 1998 , v. kathotia @xcite derived a trees summation expression for the bch product over the real numbers using m. kontsevich s universal formula for deformation quantization of poisson manifolds ; the coefficientf of this formula are certain integrals on configuration spaces and it is still unknown if they are rational numbers . in the papers @xcite and @xcite , the authors recognize the equation @xmath12 as the maurer - cartan equation of the canonical @xmath13 structure on the conormalized complex of singular cochains , on the standard two dimensional simplex with values in a lie algebra .
therefore , the possibility of an explicit description of @xmath14 , again as a trees summation formula , by using the standard tools of homological perturbation and homotopy transfer theory @xcite .
the reader may also consult @xcite for a list of explicit and recursive formulas .
the aim of this paper is to give a simple and elementary combinatorial description of the polynomial @xmath2 that uses some notions about planar rooted trees .
the necessary combinatorial background is summarized in sections [ sec.subroot ] and [ sec.posetted ] . in particular , for every finite planar rooted tree @xmath15 , the set of its leaves admits a total ordering ( from left to right ) and also a partial ordering @xmath16 , which takes care of the position of leaves with respect to the subroots .
then , we define a posetted tree as a finite planar rooted tree , whose leaves are labelled by elements on a partially ordered set ( a poset ) , monotonically with respect to @xmath16 .
our main result ( theorem [ thm.bchtrees ] ) gives an explicit description of every @xmath2 as a linear combination with rational coefficients of nested commutators , indexed by a certain set of posetted trees with @xmath3 leaves .
the formula of the coefficients involves the bernoulli numbers and is completely described in terms of the combinatorial data of posetted trees .
this section is devoted to introduce the notion , already known in the parallel logic programming community @xcite , of subroots of a planar rooted trees .
recall that a tree is called a _ rooted tree _
if one vertex has been designated the _ root_. every rooted tree has a natural structure of directed tree such that , for every vertex @xmath17 , there exists a unique directed path from @xmath17 to the root
. we shall write @xmath18 if the vertex @xmath19 belongs to the directed path from @xmath17 to the root .
a _ leaf _ is a vertex without incoming edges : equivalently , a vertex @xmath17 is a leaf if the relation @xmath20 implies @xmath21 .
a vertex is called _ internal _ if it is not a leaf ; notice that , if a rooted tree has at least two vertices , then the root is an internal vertex .
( 200,70 ) = 0.20 mm a=(200,125)b=(110,70 ) c=(95,40)d=(80,10 ) e=(110,10 ) f=(140,10 ) j=(185,10 ) k=(215,10 ) h=(245,10)l=(275,10)m=(300,10 ) n=(330,10 ) o=(315,35 ) p=(260,35 ) q=(290,70 ) z=(260,10 ) t=(360,10 ) ( t , q ) ( z , p ) ( j , a ) ( k , a ) ( b , a ) ( e , c ) ( f , b ) ( q , a ) ( c , b ) ( d , c ) ( e , c ) ( f , b ) ( p , q ) ( h , p)(l , p ) ( m , o)(o , q ) ( n , o ) ( a)@xmath22(b)@xmath22 ( c)@xmath22 ( p)@xmath22 ( d)@xmath22 ( e)@xmath22(f)@xmath22 ( n)@xmath22 ( l)@xmath22 ( m)@xmath22 ( h)@xmath22(o)@xmath22 ( p)@xmath22 ( q)@xmath22 ( z)@xmath22 ( j)@xmath22 ( k)@xmath22 ( t)@xmath22 ( 277,0)@xmath23 ( 295,70)@xmath24 ( 195,135)@xmath25 from now on , we consider only planar rooted trees ; following @xcite , we denote by @xmath26 the set of finite planar rooted trees with the root at the top and the leaves at the bottom ( i.e. , every directed path moves upward ) , and such that every internal vertex has at least two incoming edges .
we also write @xmath27 where @xmath28 is the set of planar rooted trees with @xmath3 leaves and , for every @xmath29 , we denote by @xmath30 the set of leaves of @xmath15 .
the planarity of the tree gives , for every internal vertex @xmath19 , a total ordering of the edges ending in @xmath19 , from the leftmost to the rightmost ( see figure [ fig.orientazione ] ) .
( 250,75 ) = 0.20 mm a=(240,125)b=(110,70 ) c=(95,40)d=(80,10 ) e=(110,10 ) f=(140,10 ) g=(176,10 ) h=(200,10)l=(230,10)m=(245,10 ) n=(275,10 ) o=(260,35 ) p=(215,35 ) q=(237,71 ) r=(240,95 ) a=(390,70 ) b=(360,10 ) c=(420,10 ) z=(320,10 ) j=(95,10 ) ( 222,127)@xmath23 ( 78,-5)@xmath31 ( 93,-5)@xmath32 ( 108,-5)@xmath33 ( 138,-5)@xmath34 ( 174,-5)@xmath35 ( 198,-5)@xmath36 ( 228,-5)@xmath37 ( 243,-5)@xmath38 ( 273,-5)@xmath39 ( 318,-5)@xmath40 ( 358,-5)@xmath41 ( 418,-5)@xmath42 ( a , a ) ( a , b ) ( a , c ) ( z , r ) ( c , j ) ( a , b ) ( e , c ) ( b , f ) ( r , g ) ( e , c ) ( b , f ) ( r , g ) ( p , l ) ( o , m)(a , r ) ( q , r ) ( q , o ) ( o , n ) ( p , h ) ( q , p ) ( c , d ) ( b , c ) ( a)@xmath22(b)@xmath22 ( c)@xmath22 ( p)@xmath22 ( d)@xmath22 ( a)@xmath22 ( b)@xmath22 ( c)@xmath22 ( e)@xmath22(f)@xmath22 ( n)@xmath22 ( l)@xmath22 ( g)@xmath22(m)@xmath22 ( h)@xmath22(o)@xmath22 ( p)@xmath22(r)@xmath22 ( q)@xmath22 ( z)@xmath22 ( j)@xmath22 a _ rightmost branch _ of a planar rooted tree @xmath29 is a maximal connected subgraph @xmath43 , with the property that every edge of @xmath44 is a rightmost edge of @xmath15 .
a rightmost branch is called non trivial if it has at least two vertices .
( 200,70 ) = 0.20 mm
a=(200,125 ) b=(110,70 ) c=(95,40 ) d=(80,10 ) e=(110,10 ) f=(140,10 ) j=(185,10 ) k=(215,10 ) h=(245,10)l=(275,10)m=(300,10 ) n=(330,10 ) o=(315,35 ) p=(260,35 ) q=(290,70 ) z=(260,10 ) t=(360,10 ) ( z , p ) ( m , o ) ( 274,13)(262,34 ) ( 329,13)(316,34 ) ( j , a ) ( 358,12)(291,69 ) ( 202,123)(287,72 ) ( k , a ) ( b , a ) ( 109,13)(96,39 ) ( 139,13)(111,69 ) ( c , b ) ( d , c ) ( p , q ) ( h , p ) ( o , q ) ( a)@xmath22(b)@xmath22 ( c)@xmath22 ( p)@xmath22 ( d)@xmath22 ( e)@xmath22(f)@xmath22 ( n)@xmath22 ( l)@xmath22 ( m)@xmath22 ( h)@xmath22(o)@xmath22 ( p)@xmath22 ( q)@xmath22 ( z)@xmath22 ( j)@xmath22 ( k)@xmath22 ( t)@xmath22 [ def.rigthmost m(v ) d(v ) ] a _ local rightmost leaf _ is a leaf lying on a non trivial rightmost branch . given an internal vertex @xmath19 , we call @xmath45 the leaf lying on the rightmost branch containing @xmath19 .
we also denote by @xmath46 the distance between @xmath19 and @xmath45 , as defined in @xcite .
a _ subroot _ is the vertex of a non trivial rightmost branch which is nearest to the root .
the set of subroots of a finite planar rooted tree @xmath15 will be denoted by @xmath47 .
therefore , we have the natural bijections @xmath48 in the tree of figure [ fig.esempiomassimilocali ] , the subroots are the vertices @xmath49 and @xmath50 ; the rightmost leaves are the leaves @xmath51 and @xmath52 .
moreover , @xmath53 and @xmath54 ; and @xmath55 , @xmath56 and @xmath57 .
( 200,70 ) = 0.20 mm a=(200,125)b=(110,70 ) c=(95,40)d=(80,10 ) e=(110,10 ) f=(140,10 ) h=(246,10)l=(275,10)m=(300,10 ) n=(330,10 ) o=(315,35 ) p=(262,35 ) q=(290,70 ) ( 190,126)@xmath58 ( 98,70)@xmath59 ( 297,70)@xmath60 ( 82,40)@xmath61 ( 320,35)@xmath62 ( 252,35)@xmath63 ( 73,-5)@xmath31 ( 103,-5)@xmath32 ( 133,-5)@xmath33 ( 240,-5)@xmath34 ( 270,-5)@xmath35 ( 295,-5)@xmath36 ( 325,-5)@xmath37 ( b , a ) ( e , c ) ( f , b ) ( q , a ) ( c , b ) ( d , c ) ( e , c ) ( f , b ) ( p , q ) ( h , p ) ( l , p ) ( m ,
o)(o , q ) ( n , o ) ( a)@xmath4(b)@xmath4 ( c)@xmath4 ( p)@xmath4 ( d)@xmath22 ( e)@xmath64(f)@xmath64 ( n)@xmath64 ( l)@xmath64 ( m)@xmath22 ( h)@xmath22(o)@xmath22 ( p)@xmath22 ( q)@xmath22 a planar rooted tree @xmath65 is a _ binary tree _ if every internal vertex has exactly two incoming edges .
we use the notation @xmath66 where @xmath67 is the set of planar binary rooted trees with @xmath3 leaves . using the notion introduced above , it is very easy to see that a tree @xmath68 is a binary tree if and only if it satisfies the equality : @xmath69 let @xmath70 be a ( non associative ) algebra over a field @xmath71 and @xmath72 a planar rooted tree .
labelling the leaves of @xmath15 with elements of @xmath70 , we can associate the product element in @xmath70 obtained by the usual operadic rules @xcite , i.e. , we perform the product of @xmath70 at every internal vertex in the order arising from the planar structure of the directed tree . for instance , the following labelled tree ( 200,65 ) = 0.20 mm a=(200,125)b=(110,70 ) c=(95,40)d=(80,10 ) e=(110,10 ) f=(140,10 ) h=(245,10)l=(275,10)m=(300,10 ) n=(330,10 ) o=(315,35 ) p=(262,35 ) q=(290,70 ) ( 73,-5)@xmath73 ( 103,-5)@xmath74 ( 133,-5)@xmath75 ( 240,-5)@xmath76 ( 270,-5)@xmath77 ( 295,-5)@xmath78 ( 325,-5)@xmath79 ( b , a ) ( e , c ) ( f , b ) ( q , a ) ( c , b ) ( d , c ) ( e , c ) ( f , b ) ( p , q ) ( h , p)(l , p ) ( m , o)(o , q ) ( n , o ) ( a)@xmath4(b)@xmath4 ( c)@xmath4 ( p)@xmath4 ( d)@xmath22 ( e)@xmath64(f)@xmath64 ( n)@xmath64 ( l)@xmath64 ( m)@xmath22 ( h)@xmath22(o)@xmath22 ( p)@xmath22 ( q)@xmath22 gives the product @xmath80 . given any map @xmath81 ( the labelling ) , we denote by @xmath82 the corresponding product element . if @xmath83 , then the elements @xmath84 , with @xmath85 and @xmath86 , are a set of generators of the subalgebra generated by @xmath87 .
if @xmath70 is either symmetric or skewsymmetric ( e.g. , a lie algebra ) , then we may reduce the set of generators by a suitable choice of the labelling .
keeping in mind our main application ( the bch product ) , a possible way of doing that is by introducing the combinatorial notion of posetted trees .
using the notion of subroot , we can define a partial order @xmath88 on the set of leaves @xmath30 .
given two leaves @xmath89 and @xmath90 in a tree @xmath65 , we say @xmath91 if @xmath92 or there exists a subroot @xmath93 such that @xmath94 and @xmath95 .
( 200,60 ) = 0.20 mm b=(200,110 ) f=(115,10 ) t=(285,10 ) a=(200,65 ) c=(170,30 ) e=(230,30 ) d=(155,10 ) g=(245,10 ) n=(185,10 ) l=(215,10 ) ( a , b ) ( t , b ) ( c , d ) ( e , g ) ( l , e ) ( b , f ) ( c , n ) ( a , c ) ( a , e ) ( t)@xmath64 ( a)@xmath4(b)@xmath4 ( c)@xmath4 ( d)@xmath22 ( e)@xmath22(f)@xmath22 ( g)@xmath64 ( n)@xmath64(l)@xmath22 ( 103,-5)@xmath96 ( 153,-5)@xmath97 ( 183,-5)@xmath98 ( 213,-5)@xmath99 ( 243,-5)@xmath100 ( 288,-5)@xmath101 for every poset @xmath102 , we denote @xmath103 in a similar way , we define @xmath104 , and , for every @xmath105 , @xmath106 and @xmath107 .
we call _ posetted trees _ the elements of @xmath108 .
[ example .
b ) ] the sets @xmath109 , @xmath110 and @xmath111 contain 2 , 3 and 8 posetted trees , respectively ( see figures [ fig.b2ab ] and [ fig.b3ab ] ) . (
240,25 ) = 0.30 mm a=(120,35)b=(105,10 ) f=(135,10 ) p=(200,35)q=(185,10 ) r=(215,10 ) a=(280,35)b=(265,10 ) f=(295,10 ) ( a , b)(f , a ) ( 10,0)@xmath59 ( 50,0)@xmath60 ( 10,10)@xmath22 ( 50,10)@xmath22 ( 103,0)@xmath59 ( 135,0)@xmath59 ( 183,0)@xmath60 ( 215,0)@xmath60 ( 263,0)@xmath60 ( 295,0)@xmath59 ( f , a)(b , a ) ( r , p)(q , p ) ( a)@xmath4 ( b)@xmath22 ( f)@xmath64 ( p)@xmath4
( q)@xmath22 ( r)@xmath64 ( a)@xmath4(b)@xmath22 ( f)@xmath64 ( 200,30 ) = 0.23 mm aa=(-20,60)ba=(-35,35 ) ca=(-50,10)da=(-20,10 ) ea=(10,10)fa=(-5,35 ) a=(90,60)b=(75,35 ) c=(60,10)d=(90,10 ) e=(120,10)f=(105,35 ) a=(200,60)b=(185,35 ) c=(170,10)d=(200,10 ) e=(230,10)f=(215,35 ) aa=(310,60)bb=(295,35 ) cc=(280,10)dd=(310,10 ) ee=(340,10)ff=(325,35 ) aa=(420,60)ab=(405,35 ) ac=(390,10)ad=(420,10 ) ae=(450,10)af=(435,35 ) ( a , f ) ( f , e ) ( f , d ) ( a , c ) ( aa , fa ) ( fa , ea ) ( fa , da ) ( aa , ca ) ( a , f)(f , e ) ( a , c)(f , d ) ( aa , cc)(aa , ee ) ( bb , dd ) ( aa , ac)(aa , ae ) ( ab , ad ) ( a)@xmath4 ( c)@xmath22(d)@xmath22 ( e)@xmath64(f)@xmath22 ( a)@xmath4 ( c)@xmath22(d)@xmath22 ( e)@xmath64(f)@xmath22 ( aa)@xmath4 ( ca)@xmath22(da)@xmath22 ( ea)@xmath64(fa)@xmath22 ( aa)@xmath4(bb)@xmath4 ( cc)@xmath22(dd)@xmath64 ( ee)@xmath64 ( aa)@xmath4(ab)@xmath4 ( ac)@xmath22(ad)@xmath64 ( ae)@xmath64 ( 118,0)@xmath59(58,0)@xmath59 ( 88,0)@xmath59 ( -52,0)@xmath60(8,0)@xmath60 ( -22,0)@xmath60 ( 388,0)@xmath60(448,0)@xmath60 ( 418,0)@xmath60 ( 168,0)@xmath60(198,0)@xmath59 ( 228,0)@xmath59 ( 308,0)@xmath59(338,0)@xmath59 ( 278,0)@xmath59 ( 200,50 ) = 0.23 mm a=(60,60)b=(45,35 )
c=(30,10)d=(60,10 ) e=(90,10)f=(75,35 ) a=(200,60)b=(185,35 ) c=(170,10)d=(200,10 ) e=(230,10)f=(215,35 ) aa=(340,60)bb=(325,35 ) cc=(310,10)dd=(340,10 ) ee=(370,10)ff=(355,35 ) ( a , f ) ( f , e ) ( f , d ) ( a , c ) ( a , f)(f , e ) ( a , c)(f , d ) ( aa , cc)(aa , ee ) ( bb , dd ) ( a)@xmath4 ( c)@xmath22(d)@xmath22 ( e)@xmath64(f)@xmath22 ( a)@xmath4 ( c)@xmath22(d)@xmath22 ( e)@xmath64(f)@xmath22 ( aa)@xmath4(bb)@xmath4 ( cc)@xmath22(dd)@xmath64 ( ee)@xmath64 ( 28,0)@xmath59(58,0)@xmath60 ( 88,0)@xmath59 ( 168,0)@xmath60(198,0)@xmath60 ( 228,0)@xmath59 ( 308,0)@xmath60(338,0)@xmath59 ( 368,0)@xmath59 if @xmath112 with the usual order , then there exists a natural inclusion of @xmath106 into the set of admissible graphs with @xmath3 vertices of the first kind and @xmath113 vertices of the second kind considered in @xcite . assume that @xmath114 is a subset of a ( skew)commutative algebra @xmath70 and choose a total ordering on @xmath114 . then , it is easy to see that the elements @xmath84 , with @xmath115 generate , as a @xmath71 vector space , the subalgebra generated by @xmath114 .
let @xmath116 be a lie algebra over a field @xmath71 of characteristic @xmath117 , which is complete with respect to its lower descending series @xmath118 , @xmath119 $ ] . denote by @xmath120 the baker - campbell - hausdorff ( bch ) product , obtained formally by the formula @xmath121 .
it is well known that @xmath122+\frac{1}{12 } [ a,[a , b]]-\frac{1}{12 } [ b,[b , a]]+ \cdots,\ ] ] is an element of the lie subalgebra generated by @xmath123 and @xmath124 and , then , it can be expressed as an infinite sum @xmath125 for a sequence @xmath126 . clearly , in view of the alternating properties of the product and jacobi identity , such a sequence is not unique .
the dynkin formula provides a sequence as above where @xmath127 , whenever @xmath15 has at least 2 subroots : on the other hand , the explicit expression of the nonvanishing @xmath128 is rather complicated . here
, we describe another sequence @xmath129 of rational numbers with the above properties .
first of all , define the sequence of rational numbers @xmath130 , for every @xmath131 , by their ordinary generating function @xmath132 notice that @xmath133 where the @xmath134 are the bernoulli numbers .
in particular , the only non trivial odd term of the sequence is @xmath135 and we have : @xmath136 given a poset @xmath114 and a posetted tree @xmath137 , let us define @xmath138 where the @xmath139 s are the rational numbers above and , for every subroot @xmath140 , we have @xmath141 we remind that @xmath45 is the leaf lying on the rightmost branch containing @xmath19 ( definition [ def.rigthmost m(v ) d(v ) ] ) .
since the bracket is skewsymmetric , we have that @xmath154 , for every @xmath155 ; therefore , @xmath156 in ( * ? ? ?
* theorem .
1.6.1 ) and @xcite , the following recursive formula for the baker - campbell - hausdorff product is proved : @xmath157 where @xmath158 for every @xmath159 , let @xmath160 be the subset of posetted trees with exactly @xmath161 leaves labelled with @xmath123 ; we prove that , for every @xmath162 , we have @xmath163 this is clear for @xmath164 ; for @xmath165 , we have @xmath166 whereas @xmath167 , @xmath168 , is the set of posetted trees of bernoulli type @xcite , i.e. , moreover , every element of @xmath174 is obtained in a unique way starting from a tree @xmath175 and grafting , at each of the @xmath113 leaves labelled with @xmath124 , the roots of elements of @xmath176 , with @xmath177 ( for the definition of the grafting see ( * ? ? ?
* definition 1.37 ) ) .
therefore , the proof of follows easily by induction on @xmath161 .
next , since @xmath4 is associative , we have @xmath178 and formula follows immediately from .
finally , setting @xmath179 , we have that every posetted tree of @xmath180 can be described in a unique way as a posetted tree in @xmath181 , where at every leaf labelled with @xmath124 is grafted the root of a posetted tree of @xmath182 .
in view of the associativity relation @xmath183 we obtain that is a consequence of @xmath184 .
choose @xmath185 and @xmath186 in equation , and @xmath187 and @xmath188 in equation . comparing the coefficient of the product @xmath189 in both equations , we obtain the following relations @xmath190 indeed , the coefficient of @xmath189 in equation comes from the bernoulli tree @xmath191 and so it is exactly @xmath139 . on the other side ,
we need to consider the subset @xmath192 of trees @xmath193 with only one subroot , @xmath3 leaves labelled @xmath124 and one leaf labelled @xmath123 . for any @xmath194 , we have @xmath195 , and we can define @xmath196 as @xmath197 comparing the coefficients , we have @xmath198 next , let us compute @xmath196 recursively .
there are two different types of contributions to @xmath196 due to the following graphs .
the first contribution is due to the graph with only one subroot ; in this case , the coefficient is @xmath199)=-\dfrac{b_n}{n } ad(b)^{n}(a).$ ] the other contribution is due to the graphs obtained from a graph in @xmath200 , for every @xmath201 , and grafting , at the leaves labelled with @xmath123 , a graph of @xmath202 .
therefore , for every fixed @xmath203 , the coefficient is @xmath204)=-\frac{b_i}{n}c_{n - i } ad(b)^{n}(a).\ ] ] summing up , we have @xmath205 and , since @xmath206 , we get the relation @xmath207 note that , in the previous computation , we have just used the fact that the product is associative and therefore apply for every associative product defined by equation .
more precisely , let @xmath208 be any sequence in @xmath71 , and for any @xmath209 , define @xmath210 and the product @xmath211 one implication is clear , if @xmath214 , then @xmath215 this implies that the product @xmath212 is associative ( in the last equality we use that @xmath216 ) .
as regards the other implication , assume that the product @xmath212 is associative ; then , equation holds for the product @xmath212 instead of @xmath4 . arguing as in the above remark , we conclude that the numbers @xmath208 must satisfy equation , and this easily implies that @xmath217 , for every @xmath105 .
e. dynkin : _ calculation of the coefficients of the campbell - hausdorff formula .
, * 57 * , ( 1947 ) , 323 - 326 . an english translation may be found in : e.b .
dynkin , a.a .
yushkevich , g.m .
seitz , a.l .
onishchik ( eds . ) , selected papers of e.b .
dynkin with commentary , american mathematical society / international press , providence , r.i./cambridge , mass , ( 2000 ) . | we introduce the combinatorial notion of posetted trees and we use it in order to write an explicit expression of the baker - campbell - hausdorff formula . |
the x - rays from active galactic nuclei ( agn ) are thought to originate from the innermost regions of an accretion disk around a central supermassive black hole .
thus , in principle , the study of these x - rays should allow one to probe the immediate environment of the accreting black hole as well as the exotic physics , including strong - field general relativity , that operates in this environment . in the past decade x - ray astronomy
has begun to fulfill that promise . both _
exosat _ and _ ginga _ discovered iron k - shell features ( including the k@xmath0 fluorescent line of cold iron at 6.4kev ) in the x - ray spectra of seyfert galaxies which were interpreted as ` reflection ' of the primary x - ray continuum by cold , optically - thick material in the immediate vicinity of the black hole ( guilbert & rees 1988 ; lightman & white 1988 ; nandra et al .
1989 ; nandra , pounds & stewart 1990 ; matsuoka et al .
it was suggested that this cold reflecting material was the putative accretion disk of agn models . with the launch of _ asca _ and the advent of medium resolution spectroscopy
, the iron line in several objects was shown to be broad ( @xmath4 fwzi ) and skewed ( tanaka et al .
1995 ; nandra et al .
the overall line profiles are in good agreement with models for fluorescent line emission from the innermost regions of geometrically - thin black hole accretion disks ( fabian et al .
such data allow us to address issues such as the location of the radius of marginal stability , the spin of the black hole , and the inclination distribution of various classes of agn ( see reynolds 1999 and references therein for a review of these studies ) . in the current , _
rxte _ era , we can now probe the iron line and compton reflection hump in individual objects in some detail ( e.g. , mcg-5 - 23 - 16 , weaver et al .
1998 ; mcg6 - 30 - 15 , lee et al .
1998 , 1999a ; ngc 5548 , chiang et al . 1999 ) . while these spectral studies have been successful , a complete picture of the agn phenomenon
is not possible without addressing the timing properties .
timing studies are important for two intertwined reasons .
firstly , agn are _ inherently _ variable systems . in general , the variability timescale in a given object is seen to shorten as one considers higher frequency radiation . in the x - ray and @xmath5-ray bands , dramatic variability has been seen in many seyfert galaxies with doubling timescales of only a few minutes ( e.g. see reynolds et al .
although it is poorly understood to date , the nature of this violent variability is a vital component of any final agn model .
careful characterization of the timing properties , as well as determining the observed spectral evolution during dramatic temporal events , is required if we are to understand this phenomenon .
secondly , timing studies are needed to break certain degeneracies that exist in models which , to date , have only been constrained by purely spectral data .
the spin of the black hole in mcg6 - 30 - 15 provides an excellent example of such a degeneracy by fitting the ` very - broad ' state ( iwasawa et al .
1996 ) of the iron line in this object with models consisting of a thin , disk - hugging corona , dabrowski et al .
( 1997 ) inferred that the black hole in this agn must be almost maximally rotating , with a dimensionless spin parameter of @xmath6 .
however , by including line emission from within the radius of marginal stability , reynolds & begelman ( 1997 ) showed that a geometry in which the x - ray source is at some height above the disk plane can produce the same line profile even if the black hole is completely non - rotating .
while there are subtle spectral differences between the two scenarios ( young , ross & fabian 1998 ) the most obvious way of distinguishing these scenarios is through their timing properties .
the reynolds & begelman ( 1997 ) geometry predicts substantial time delays between fluctuations in the primary power - law continuum and the responding fluctuations in the iron line .
more generally , the reverberation characteristics of the iron line contain tremendous information on the mass and spin of the black hole as well as the geometry of the x - ray source ( stella 1990 ; reynolds et al . 1999 ) .
the observational situation is more complex .
lee et al .
( 1999b ) and chiang et al .
( 1999 ) have analyzed extensive _
datasets for mcg6 - 30 - 15 and ngc 5548 , respectively , in order to study the timing properties and spectral variability . in both of these objects , the same pattern of spectral variability is seen .
firstly , the x - ray photon index displays flux - correlated changes in the sense that the source is softer when it is brighter .
secondly , and more surprisingly , the iron line flux was found to be constant over the timescales probed by these direct spectral studies ( @xmath7ksec ) . as discussed by both sets of authors ,
these results are difficult to interpret in the framework of standard x - ray reflection models since the breadth of these lines indicate that they originate from a small region .
it appears that some feedback mechanism regulates the amount of iron line emission in order to produce approximately constant iron line flux .
flux - correlated changes in the ionization state of the disk represent one such mechanism ( we discuss this in more detail in section 5 of this paper ) .
unless this feedback mechanism operates instantaneously , we might still expect variability of the iron line flux on short timescales . driven by these motivations
, this paper addresses the problem of determining causal relationships between light curves in different x - ray bands , with particular emphasis on timescales shorter than those that can be probed by direct spectroscopy .
in particular , we use the long _ rxte _ observation of the bright seyfert 1 galaxy mcg6 - 30 - 15 reported by lee et al .
( 1999a , b ) and consider the relationship between the 24kev band ( hereafter called the continuum band ) and the 57kev band which contains most of the iron line photons ( and hereafter called the line band ) . an important special case is one in which there is a linear transfer function relating one band to the other : @xmath8 where @xmath9 and @xmath10 are continuum and line band fluxes respectively , and @xmath11 is the transfer function .
such relationships between bands contain much of the important physical information , such as the reverberation characteristics of the iron line .
mathematically , the linear transfer equation can be easily inverted using fourier methods to obtain , @xmath12 where @xmath13 represents the fourier transform of @xmath9 . however , in real situations , a large number of regularly sampled measurements are required to obtain an accurate deconvolution using this simple method .
more often , deconvolution is achieved using maximum entropy techniques or some other regularization method ( horne et al .
1991 ; krolik et al .
1991 ) .
another common approach ( and one that is often used with less well sampled data ) is to compute cross - correlation functions ( ccfs ) , or some variant thereof which accounts for the finite and irregular sampling often encountered in real data .
the discrete correlation function ( dcf ; edelson & krolik 1988 ) is one example of such a variant .
lee et al .
( 1999b ) apply such methods to the observation of mcg6 - 30 - 15 considered in this paper and detect both phase and time lags between _ rxte _ bands ( also see nowak & chiang 1999 ) .
while these methods are powerful , it can be difficult to separate subtle time leads / lags from the autocorrelation properties of the data . here
, we take an alternative approach which is heavily based on the method of press , rybicki & hewitt ( 1992 ; hereafter prh92 ) . in essence
, we use the correlation properties of the continuum band data to reconstruct an optimal continuum light curve in which the data gaps have been interpolated .
most importantly , we also compute the expected deviation of the continuum flux from the interpolated curve .
the reconstructed continuum band light curve is convolved with a trial transfer function and compared with the line band light curve in a @xmath14 sense .
we then examine changes in the @xmath14 statistic as a function of the parameters that define the trial transfer function .
section 2 recaps the prh92 method .
this is then applied to the _ rxte _ data for mcg6 - 30 - 15 in section 3 .
the robustness and validity of our approach is demonstrated by applying this method to simulated data ( section 4 ) .
section 5 draws together our results and discusses their implications for the nature of this source .
in particular , we argue that the black hole in this agn has a mass of only @xmath15 . in order to explain the spectral variability ,
it is suggested that there are flux correlated changes in the ionization state of the surface layers of the accretion disk .
section 6 presents a short summary of the results and relevant astrophysical implications .
the continuum band light curve is reconstructed from the data using the technique of prh92 . for completeness , this section summarizes their method .
the reader who is primarily interested in the application of this method may skip to section 3 .
suppose that the true flux of the source at time @xmath16 is @xmath17 , but we measure @xmath18 , where @xmath19 is the noise in the measurement . in our case , the noise is poisson in nature .
our knowledge of @xmath17 is further impeded by the fact that the measurement is only made at a finite number of times @xmath20 , where @xmath21 .
we denote @xmath22 as @xmath23 and refer to this as the continuum data vector .
we seek an optimal reconstruction of @xmath17 which is continuous in time , @xmath24 , such that @xmath25 ^ 2\rangle\ ] ] is minimized for all @xmath16 . as usual ,
angle brackets denote the expectation value .
we impose that @xmath24 is linear in the data vector in the sense that @xmath26 where @xmath27 are a set of inverse response functions that are also continuous in time . assuming that the noise is uncorrelated with both @xmath17 and itself
, prh92 showed that eqn ( 3 ) can be minimized to yield , @xmath28 here , @xmath29 is the total covariance matrix . to keep the notation concise , prh92 define the correlation statistics : @xmath30 these functions define what prh92 call the ` covariance model ' .
the expected variance of the real signal from the optimal reconstruct in eqn ( 5 ) is then given by @xmath31 once the covariance model is known , eqns ( 5 ) and ( 10 ) define the optimal reconstruction of the continuum light curve together with a statistic measuring the expected deviation of the real signal from the reconstruction . here
, again , we follow the method of prh92 to determine the covariance model for our continuum data . at this stage
, we make the assumption that the underlying process is statistically stationary so that @xmath32 where @xmath33 is the autocorrelation function that we have to determine .
this function is related to the first order structure function @xmath34 by @xmath35 and @xmath34 can be approximated by forming pair - wise estimates for all distinct pairs of data points in the continuum light curve , and then binning by the time lag of the pairs .
we find that the analytic form @xmath36 fits the structure functions of this paper well .
in fact , the reconstruction is fairly insensitive to the exact analytic form used to approximate the structure function . making the reasonable assumption that @xmath37 as @xmath38 , our final expression for the autocorrelation function is @xmath39
in this section , we apply the method outlined above to a long _ rxte _ observation of the bright seyfert 1 galaxy mcg6 - 30 - 15 .
_ rxte _ observed mcg6 - 30 - 15 for approximately @xmath40 starting on 4-aug-1997 .
we retrieved these data from the nasa - heasarc public archive situated at the goddard space flight center .
our data reduction closely parallels that of lee et al .
( 1999a ) who has studied the spectral characteristics of this observation .
since , as mentioned in the introduction , we are interested in the soft x - ray continuum and the iron line band , the proportional counter array ( pca ) is the appropriate instrument for us to consider .
examining the housekeeping files for this observation reveals that proportional counter units ( pcus ) 3 and 4 suffer occasional breakdown and shut off .
hence , we do not consider data from these units and , instead , extracted standard-2 data from pcus 02 .
we applied fairly standard faint - source screening criteria to these data : the source must be at least @xmath41 above the earth s limb ( elv@xmath4210 ) , the source must be located within @xmath43 of the nominal pointing position ( [email protected] ) , there must be at least three pcus on ( num_pcu_on@xmath422 ) , it has been at least 30 minutes since a passage of the south atlantic anomaly ( time_since_saa@xmath4230 ) , and the electron background is not too high ( [email protected] ) . after application of these screening criteria ,
approximately @xmath45 of ` good ' data remain . from these data , 24kev and 57kev light curves
were extracted using 64s bins .
we also extracted the 815kev light curve which we will use in section 3.3 .
the background was estimated using the l7 - 240 background models which are appropriate for faint sources such as agn .
background light curves were computed and subtracted from the measured light curves in order to form the final background subtracted light curves that we shall use in our study .
figure 1 shows the continuum band ( 24kev ) light curves that results from this procedure . for clarity ,
the light curve shown in this figure has been binned with 256s bins .
we now apply the procedure outlined in section 2 to these lightcurves . to begin with
, we must estimate the structure function for these data .
figure 2a shows a pair - wise estimate of the continuum band structure function obtained following the method of prh92 .
this figure also shows our analytic approximation which is given by eqns ( 13 ) and ( 14 ) with @xmath46 using this covariance model , the prh92 reconstruction was applied to the continuum light curve using @xmath47 data points .
a portion of the resulting reconstructed light curve is shown in fig .
the next stage in the procedure is to convolve the reconstructed continuum band light curve with a trial transfer function and compare the result with the line band light curve in a @xmath14 sense .
we can then minimize the @xmath14 statistic in order to constrain free parameters in the trial transfer function .
we also minimize @xmath14 over multiplicative and additive offsets between the continuum and line band light curves , i.e. we set @xmath48 and minimize over @xmath49 and @xmath50 as well as the parameters describing the trial transfer function @xmath51 . in this work , we choose two trial transfer functions .
the first represents the case where some fraction @xmath52 of the line band flux is a delayed copy of the continuum band with a time delay @xmath53 : @xmath54 the second represents the case where some fraction @xmath52 of the line band flux is a delayed and smeared copy of the continuum band flux , where a gaussian kernel is used : @xmath55 no extrapolations were performed during this procedure . in order to avoid extrapolating , the @xmath14 statistic was calculated using a subset of data points . for the trial transfer function @xmath56 ,
only data during times @xmath57 were used to compute @xmath14 , where @xmath58 and @xmath59 are the times of the start and end of the reconstructed continuum light curve . for @xmath60 , @xmath14 is computed based upon data from times @xmath61 .
figure 3 shows the @xmath14 surfaces and confidence contours once this procedure has been performed . when displaying the @xmath14 surfaces , we plot @xmath62 in order to highlight the topography of the surface near the global minimum in the surface .
it can be seen that the minimum of the @xmath632 surface corresponds to the two lines @xmath64 and @xmath65 , i.e. no time delayed component of the line band light curve is detected .
here we only show the results for @xmath56 the @xmath60 results are trivial ( i.e. @xmath14 surface is completely flat ) since the preferred solution always have @xmath64 .
the best fit values of the multiplicative and additive constants are @xmath66 and @xmath67 . by considering the @xmath68 slice through the @xmath14 surface produced with trial transfer function @xmath56 , we can examine _ overall _ lags and leads between energy bands . examining the 24kev and 57kev light curves for mcg6 - 30 - 15 in this way , we find that the @xmath14 slice possesses a minimum at zero lag i.e. we find no evidence for overall time lags or leads between the continuum and line bands down to 64s , the bin size of the data . performing the same procedure for the 24kev and 815kev light curves reveals a one bin offset in the position of the minimum ( fig . 4 ) , indicating that the 815kev light curve is delayed by @xmath6950100s as compared with the 24kev light curve .
lee et al .
( 1999b ) have applied ccf methods to this _ rxte _ dataset . by carefully comparing with simulations
, they find evidence that the 7.510kev band lags the lower energy bands with a _ phase delay _ of @xmath70 .
they also find evidence that the hard band ( 1020kev ) lags the softer bands with a _ time delay _ similar to that found in this work .
figure 5 shows the dcf for our 24kev and 815kev lightcurves ( this is very similar to fig .
17 of lee et al .
a small time lag of 50100s between these two bands is evident .
thus , ccf methods and the optimal reconstruction method both suggest a time lag of 50100s between the 24kev and 815kev bands . our fitting in section 3.2 clearly reveals the need for an additive offset ( i.e. non - zero @xmath50 value ) between the line and continuum band light curves .
in other words , the fractional variability about the mean level is less in the line band that it is in the continuum band . the spectral analysis of lee et al .
( 1999b ) allows this behaviour to be understood in terms of the spectral phenomenology .
firstly , lee et al . found that on timescales of @xmath71ksec , the iron line flux does not track the continuum flux and , instead , remains approximately constant .
secondly , it was found that there are flux correlated changes in the photon index by as much as @xmath72 in the sense that higher flux states are softer .
both of these spectral changes will tend to reduce variability in the line band as compared with the ( softer ) continuum band .
in order to assess the significance and robustness of the above results , this section describes the application of this method to simulations .
we tailor our simulation to match the _ rxte _ observation mcg6 - 30 - 15 as much as possible .
_ exosat _ showed that the high frequency fluctuations of mcg6 - 30 - 15 possess a power spectrum of the form @xmath73 .
we use this power spectrum with an additional low - frequency cutoff at @xmath74 : @xmath75 we then make a ( noiseless ) simulated continuum light curve , @xmath76 , by summing fourier components of random phase between @xmath77 and @xmath78 , i.e. @xmath79\ ] ] where @xmath80 is a uniformly randomly distributed in the range @xmath81 to @xmath82 for each distinct value of @xmath83 . without loss of generality , we assume that the line band flux possesses the same mean normalization as the continuum band light curve .
however , in order to mimic the situation found in section 3 as closely as possible , we assume that there is an additive offset between the continuum band and line band light curves as well as the convolution a transfer function .
in other words we compute a ( noiseless ) line - band light curve using the expression : @xmath84 here , @xmath85 is the non - zero - lag component of our imposed simulated transfer function for which we use a gaussian : @xmath86 where @xmath83 is the fraction of the continuum flux that is delayed , @xmath87 is the mean time delay , and @xmath88 is the temporal standard - width of the smearing . for concreteness , we set @xmath89 this value of @xmath83 is approximately the fraction of the line - band flux which originates from the iron line , and hence this simulation crudely mimics the effect of iron line reverberation with a @xmath90 time delay .
our value of @xmath50 is set to be similar that found for mcg6 - 30 - 15 above . from these ` perfect ' noiseless light curves , we formed possion sampled noisy light curves assuming a mean count rate of 4cps in both the continuum and line bands , and using 64s bins .
the datagap structure of the real mcg6 - 30 - 15 dataset was then imposed on the simulated light curves .
we now examine our realistic , simulated , data in order to assess how well we can detect the existence of the imposed lag and recover the properties of @xmath85 using our method .
we use the method of section 2.1 and 2.2 to form an optimally reconstructed , evenly - sampled continuum lightcurve . the covariance model used
is given by eqn ( 13 ) and ( 14 ) with @xmath91 a total of @xmath92 simulated data points were used to form the reconstruction which spans a simulated observation time of 400000s .
a portion of the simulated dataset and its reconstruction are presented in fig .
note how well the reconstruction algorithm recovers the real signal during the times with data , and brackets the real signal during other times .
figure 7 presents the @xmath14 surfaces and confidence contours that result from passing the simulated light curves through the trial transfer functions @xmath56 and @xmath60 , including minimization over any additive offset between the continuum and line band light curves .
both trial transfer functions clearly detect the imposed lag in so far as a a deep and isolated hole is present in the @xmath14 surface at approximately the right time delay , delay fraction and delay width .
note that the @xmath52 dimension , which has been suppressed in the @xmath60 plots , has a value of @xmath93 at the global minimum .
this demonstrates the power of this technique for finding and characterizing subtle time lags or leads that are present in such data .
in order to bring structure to the discussion that follows , we will summarize the pertinent results from this paper . 1 .
we clearly see reduced fractional variability in the iron line band ( 57kev ) as compared with the continuum band ( 24kev ) .
this is the origin of the additive offset , @xmath50 , that was introduced in section 3 .
the spectral fitting results of lee et al .
( 1999b ) suggests that this is due to a combination of a constant iron line flux and flux correlated changes in the photon index .
2 . our analysis finds no evidence for iron line reverberation effects . by running a number of simulations ,
we find that any reverberation time delays must be less than @xmath94 or greater than @xmath95ks . together with the above result
, this suggests an approximately constant iron line flux over these timescales .
thus , we can extend the work of lee et al .
( 1999b ) and infer a constant iron line on timescales down to @xmath96ksec .
3 . any overall time lag between the 24kev and 57kev band is less than @xmath95s .
however , we do find that the 815kev band is delayed with respect to the 24kev band by 50100s .
we can use this time delay to obtain a rough size scale for the comptonizing cloud that is producing the hard x - rays . assuming a coronal temperature of @xmath97 , it takes approximately 3 inverse compton scatterings for a photon to be boosted between the 24kev and 815kev bands .
thus , the mean free path of a photon is approximately 1530 light seconds .
this is a lower limit on the size of the comptonizing region .
as we will see , this combination of facts presents problems for current models .
initially , let us discuss these results in the light of simple x - ray reflection models ( e.g. george & fabian 1991 ) .
assuming that variations in the primary flux are not accompanied by gross changes in geometry , we expect to observe one of two cases .
firstly , if the light crossing time of the fluorescing part of the disk is shorter than the timescale being probed by the observation , an iron line with constant _ equivalent width _ will result ( i.e. the iron line flux will track the flux of the illuminating primary x - ray source ) . on the other hand ,
if the light crossing time of the fluorescing disk is greater than the timescale being probed , a constant _ flux _ line will result . within the context of these simple reflection models ,
we are forced to conclude that the light crossing time of the fluorescing region is larger than @xmath95ks .
since the line is relativistically broad , most of the fluorescence occurs in the central @xmath98 of the disk .
setting the light crossing time of this region to be greater than 50ks gives a black hole mass of @xmath99 .
given such a large black hole mass , the accretion rate must be less than 1% of the eddington rate in order to produce the observed luminosity of @xmath100 ( reynolds et al .
furthermore , the size of the x - ray emitting blobs must be small , @xmath101 , in order to produce the very small time delays seen between different bands . despite being so small
, these blobs must be at large distances above and below the accretion disk plane , or else one would still see iron line variability as a flaring blob illuminated the patch of disk directly beneath . to date , there are no dynamical measurements of the black hole mass in mcg6 - 30 - 15 , and hence such a model does not explicitly contradict any data .
however , there are several indirect arguments that lead us to reject the inference of a large black hole mass in mcg6 - 30 - 15 .
an independent indicator of the black hole mass is possible by estimating the bulge mass of the host galaxy and then applying the bulge / hole mass relationship of magorrian et al .
the b - band luminosity of the s0-galaxy which hosts this seyfert nucleus is approximately @xmath102 ( rc3 catalogue ) , and this is likely to be completely dominated by the bulge since the nucleus is heavily reddened in the b - band and the galactic disk is very weak . using a hubble constant of @xmath103 , the absolute b - band magnitude of the bulge
is then @xmath104 . using the standard relations ( faber et al .
1997 ) , the bulge mass is then @xmath105 . finally , applying the magorrian et al .
( 1998 ) scaling factor between bulge mass and black hole mass gives @xmath106 , an order of magnitude smaller than the black hole mass estimate in the previous paragraph .
there are also x - ray constraints that suggest a black hole mass significantly smaller than @xmath107 .
mcg6 - 30 - 15 has exhibited large amplitude x - ray variability on timescales as short as 100s .
however , the dynamical timescale of the accretion disk where the bulk of the energy is released is @xmath108 .
thus , if the black hole really is as massive as @xmath109 , large amplitude variability would be occurring on timescales as short as @xmath110 .
it is difficult to conceive of processes which would give such variability .
the final x - ray argument against a @xmath111 black hole in mcg6 - 30 - 15 comes from the power spectrum derived by lee at al .
( 1999b ) and nowak & chiang ( 1999 ) . by comparing the power spectral density ( psd ) of mcg6 - 30 - 15 with that of ngc 5548 and cygnus x-1
, they estimate that the black hole in mcg6 - 30 - 15 has a mass of @xmath112 .
since the application of simple x - ray reflection arguments led us to deduce an unacceptably large black hole mass , we must examine alternative avenues .
indeed , the spectral fitting of lee et al . ( 1999b ) forces us to consider complications beyond the simple reflection picture in their spectral fitting , they found that the compton reflection continuum fails to show the expected correlation with the iron line equivalent width ( in fact , they are anti - correlated ; lee et al .
very similar behaviour is also seen in ngc 5548 ( chiang et al . 1999 ) ionization of the disk surface is one of the few physical phenomenon that can ( partially ) decouple the strength of the compton reflection continuum from the strength of the iron line .
matt , fabian & ross ( 1993 ) demonstrated that the iron emission line is more sensitive to ionization effects than the general form of the compton reflection continuum .
in other words , patches of the disk with certain ( surface ) ionization parameters can produce a compton reflection continuum without producing appreciable iron fluorescence .
we use this fact to construct the following simple model .
let the x - ray flux illuminating the surface layers of the accretion disk be @xmath113 a variety of x - ray source geometries give @xmath114 at large radii , and @xmath115 as one approaches the innermost parts of the disk .
now , the ionization parameter of at the surface of the disk is given by @xmath116 where @xmath117 is the density of the surface layers of the disk .
we suppose that @xmath118 . hence
, we have @xmath119 standard disk models ( shakura & sunyeav 1973 ) give @xmath120 at large radii , and @xmath121 near the inner part of the disk . now , suppose that there exists a critical ionization parameter @xmath122 above which there is no iron line produced . for reasonable values of @xmath123 and @xmath5 , this gives a critical radius @xmath124 within which no iron line is produced .
the total iron line flux expected from the object is then given by @xmath125 which is readily manipulated to give @xmath126 for our canonical values of @xmath123 and @xmath5 , this gives @xmath127 .
thus , this simple model produces an iron line flux which is anti - correlated with the flux of the illuminating source . provided a strong compton reflection continuum can still originate from the ionized portions of the disk , this type of picture may explain the spectral behavior that we observe .
one simple prediction of this model is that the velocity width of the line profile gets smaller as the continuum flux increases ( due to an outward migration in the inner radius of the line emitting region ) .
of course , the toy model presented above only captures the crudest aspects of the problem .
fully self - consistent ionized reflection models must be calculated ( taking into account the vertical structure of the disk ; e.g. see nayakshin , kazanas & kallman 1999 ) and compared with the data in order to test whether the picture sketched here is reasonable or not .
even if global , flux - correlated changes in the ionization of the disk surface are responsible for the observed spectral changes , we would still expect reverberation signatures on short timescales .
we have set upper limits of @xmath128s on the timescale of any reverberation delay .
if the black hole mass is @xmath129 , the light crossing time of the iron line producing region is @xmath130s , and hence we need to infer a disk - hugging corona ( with @xmath131 ) in order to be compatible with the reverberation limits . if , instead , the black hole is @xmath132 , the light crossing time of the entire line producing region is only @xmath133 and so the x - ray source geometry is unconstrained by our reverberation limits .
the corresponding eddington ratios are @xmath134 and @xmath135 for black hole masses of @xmath136 and @xmath137 respectively .
in this paper , we have used an interpolation method based upon that of prh92 to search for temporal lags and leads between the 24kev , 57kev and 815kev bands in a long _ rxte _ observation of the bright seyfert 1 galaxy mcg6 - 30 - 15 . in essence , we use the prh92 method to compute an optimal reconstruction of the 24kev light curve in which the datagaps are interpolated across .
we then fold this reconstructed light curve through trial transfer functions and compare with data from the other bands in a @xmath14 sense .
our search for lags and leads was tailored to find reverberation effects in the iron line which is thought to originate from the innermost regions of the black hole accretion disk .
we find no evidence for any reverberation , and rule out reverberation delays in the range @xmath138ksec .
we can extend the conclusions of lee et al .
( 1999b ) , and infer that the iron line possesses a constant flux on timescales on timescales as short as 500s .
we also find that the hard band ( 815kev ) is delayed by 50100s relative to the 24kev band .
we attempt to put these various results together into a coherent picture for this object .
the constancy of the iron line flux leads one to consider large black hole masses ( in excess of @xmath107 ) .
however , such a large mass is found to be unacceptable from the standpoint of both x - ray variability constraints , and constraints based on the mass of the galactic bulge . indeed , using the bulge / hole scaling factor of magorrian et al .
( 1998 ) , we estimate that the hole has a mass of @xmath106 .
given that this is a more reasonable mass estimate , some mechanism beyond the simple x - ray reflection model must be invoked to explain the temporal variability of the iron line and compton reflection continuum .
we suggest that flux correlated changes in the average ionization state of the surface layers of the accretion disk may be such a mechanism .
while we support this suggestion with a toy model , the plausibility if this suggestion can only be assessed once detailed modeling has been performed .
i thank james chiang , rick edelson , andrew hamilton , julia lee and mike nowak for insightful discussions throughout the course of this work .
csr thanks support from hubble fellowship grant hf-01113.01 - 98a .
this grant was awarded by the space telescope institute , which is operated by the association of universities for research in astronomy , inc .
, for nasa under contract nas 5 - 26555 .
csr also thanks partial support from nasa under ltsa grant nag5 - 6337 .
this work has made use of data obtained through the high energy astrophysics science archive research center ( heasarc ) online service , provided by the nasa goddard space flight center .
chiang j. et al .
, 1999 , apj , submitted .
dabrowski y. , fabian a. c. , iwasawa k. , lasenby a. n. , reynolds c. s. , 1997 , mnras , 288 , l11 edelson r. a. , krolik j. h. , 1988 , apj , 333 , 646 faber s. m. et al .
, 1997 , aj , 114 , 1771 fabian a. c. , rees m. j. , stellar l. , white n. e. , 1989 , mnras , 238 , 729 guilbert p. w. , rees m. j. , 1988 , mnras , 233 , 475 horne k. , welsh w. f. , peterson b. m. , 1991 , apj , 367 , l5 iwasawa i. et al.,1996 , mnras , 282 , 1038 krolik j. h. , horne k. , kallman t. r. , malkan m. a. , edelson r. a. , kriss g. a. , 1991 , apj , 371 , 541 lee j. c. , fabian a. c. , brandt w. n. , reynolds c. s. , iwasawa k. , 1999a , mnras , in press lee j. c. et al .
, 1999b , in preparation lee j. c. , fabian a. c. , reynolds c. s. , iwasawa k. , brandt w. n. , 1998 , mnras , 300 , 583 .
lightman a. p. , white t. r. , 1988 , apj , 335 , 57 magorrian j. et al .
, 1998 , aj , 115 , 2285 matsuoka m. , piro l. , yamauchi m. , murakami t. , 1990 , apj , 361 , 440 matt g. , fabian a. c. , ross r. r. , 1993 , mnras , 262 , 179 nandra k. , pounds k. a. , stewart g. c. , 1990 , mnras , 242 , 660 nandra k. , pounds k. a. , stewart g. c. , fabian a. c. , rees m. j. , 1989 , mnras , 236 , l39 nandra k. , george i. m. , mushotzky r. f. , turner t. j. , yaqoob t. , 1997 , apj , 477 , 602 nayakshin s.,kazanas d. , kallman t. r. , 1999 , apj submitted ( astro - ph/9909359 ) .
nowak m. a. , chiang j. , 1999 , apjl , subimtted nowak m. a. , vaughan b. a. , wilms j. , dove j. , begelman m. c. , 1999 , apj , 510 , 874 press w. h. , rybicki g. b. , hewitt j. n. , 1992 , apj , 385 , 404 ( prh92 ) reynolds c. s. , 1999 , in high energy processes in accreting black holes , eds j. poutanen , r. svennson , p.178 , asp conf .
vol . 161 .
reynolds c. s. , begelman m. c. , 1997 , apj , 488 , 109 reynolds c. s. , young a. j. , begelman m. c. , fabian a. c. , 1999 , apj , 514 , 164 reynolds c. s. , ward m. j. , fabian a. c. , celotti a. , 1997 , mnras , 291 , 403 reynolds c. s. , fabian a. c. , nandra k. , inoue h. , kunieda h. , iwasawa k. , 1995 , mnras , 277 , 901 shakura n. i. , sunyaev r. a. , 1973 , a&a , 24 , 337 stella l. , 1990 , nat . , 375 , 659 tanaka y. et al . , 1995 , nat , 375 , 659 young a. j. , ross r. r. , fabian a. c. , 1998 , mnras , 300 , l11 | we use the method of press , rybicki & hewitt ( 1992 ) to search for time lags and time leads between different energy bands of the rxte data for mcg6 - 30 - 15 .
we tailor our search in order to probe any reverberation signatures of the fluorescent iron k@xmath0 line that is thought to arise from the inner regions of the black hole accretion disk .
in essence , an optimal reconstruction algorithm is applied to the continuum band ( 24kev ) light curve which smoothes out noise and interpolates across the data gaps .
the reconstructed continuum band light curve can then be folded through trial transfer functions in an attempt to find lags or leads between the continuum band and the iron line band ( 57kev ) .
we find reduced fractional variability in the line band .
the spectral analysis of lee et al .
( 1999 ) reveals this to be due to a combination of an apparently constant iron line flux ( at least on timescales of @xmath1 ) , and flux correlated changes in the photon index .
we also find no evidence for iron line reverberation and exclude reverberation delays in the range 0.550ksec .
this extends the conclusions of lee et al . and suggests that the iron line flux remains constant on timescales as short as 0.5ksec .
the large black hole mass ( @xmath2 ) naively suggested by the constancy of the iron line flux is rejected on other grounds .
we suggest that the black hole in mcg6 - 30 - 15 has a mass of @xmath3 and that changes in the ionization state of the disk may produce the puzzling spectral variability .
finally , it is found that the 815kev band lags the 24kev band by 50100s .
this result is used to place constraints on the size and geometry of the comptonizing medium responsible for the hard x - ray power - law in this agn . #
1to 0pt#1 # 1 # 1 to size # 1 to size # 1 = cmr10 scaled2 = cmbx10 scaled2 = cmti10 scaled2 ps .
galaxies : seyfert , galaxies : individual : mcg6 - 30 - 15 , line : formation , methods : statistical , x - ray : galaxies |
in recent years , it has been debated in the literature about possible alternatives to black holes , the black hole mimickers , which would look observationally almost like black holes but would have no horizon .
the existence of such objects can , in principle , put in doubt astrophysical data which otherwise are considered as observational confirmation in favor of black holes @xcite . on one hand , it is clear that the properties in the near - horizon region where gravity is strong can be quite different for both type of objects . on the other hand , the statements about the difficulties in discerning black holes from their mimickers are usually related to measurements at spatial infinity .
thus , one should insist on the question : can an observer at infinity catch the difference between both type of objects in some indirect way , or even rule out some possible mimicker ? in our view , the answer is positive and is connected with key properties , namely , regularity or singularity , of the corresponding geometries .
it turns out that the requirement of full regularity up to an arbitrary neighborhood of the gravitational radius of the object enables one to rule out the potential mimickers in most of the cases .
the goal of the present work is to examine the near - horizon properties , and their connection with far away asymptotic properties , of some candidates to black mimickers .
we study spherically symmetric configurations , and make , two major divisions , or classes , on those candidates .
first , uncharged or charged but non - extremal objects , and second extremal objects . within the uncharged or charged but non - extremal one can invoke as black hole mimickers , non - extremal @xmath0-wormholes on the threshold of the formation of an event horizon , some of which are called black foils @xcite ( see @xcite for the construction with other purposes of @xmath0-wormholes , which actually can also act as mimickers ) , and gravastars @xcite . within the extremal charged class one can invoke extremal @xmath2-wormholes on the threshold of the formation of an event horizon , quasi - black holes @xcite ( see also @xcite ) , and wormholes on the basis of quasi - black holes from bonnor stars , to name a few .
we want to elucidate , whether or not the objects belonging to these two classes remain regular in the near - horizon limit .
the arguments of @xcite which rule out non - extremal limiting configurations as becoming singular do not apply to the wormhole case @xcite . thus , we carry out the corresponding analysis anew for both classes of objects .
for our purposes we write a generic spherically symmetric metric as @xmath3 where @xmath4 is the radial coordinate , and the @xmath5 are generic parameters , that depend on each situation we are studying . for instance
when treating wormholes one has @xmath6 , and @xmath7 , such that @xmath8 and @xmath9 . here
@xmath10 is the radius of the would be horizon when @xmath11 is zero , @xmath12 is the radius of the other possible horizon , @xmath11 is in principle a small quantity , and @xmath13 is the radius of a possible matter shell , satisfying @xmath14 . in other situations , e.g. , for gravastars , one has @xmath15 and , possibly , the energy density and the pressure should be somehow included . for the metric ( [ 1 ] ) the components of the riemann tensor in an orthonormal frame , the hat frame , are equal to , @xmath16 @xmath17 @xmath18 @xmath19 where a prime denotes derivative with respect to @xmath4 . here
these components of the riemann tensor have a simple physical meaning . the @xmath20 component in eq .
( [ k(r ) ] ) yields the radial geodesic deviation , the @xmath21 component in eq .
( [ n(r ) ] ) yields the angular deviation , and analogously for the @xmath22 and @xmath23 components in eqs .
( [ f(r ) ] ) and ( [ h(r ) ] ) . in summary
, they describe the deviation of geodesics in the corresponding directions . in general , forcing a matching at @xmath13 , with @xmath24 , surface stresses @xmath25 appear , which , in a coordinate frame , are equal to @xcite @xmath26 \text { , } \label{00}\ ] ] @xmath27 + \left ( \frac { d\phi } { dl}\right )
_ { + } -\left ( \frac{d\phi } { dl}\right ) _ { -}\ , , \label{22}\ ] ] @xmath28 being equal to @xmath29 , and @xmath30 being the proper radial distance .
now , if metric ( [ 1 ] ) represents a wormhole , than the areal radius @xmath31 should have a local minimum at the throat .
thus , we have two branches emerging out of the minimum radius , one with @xmath32 and the other with @xmath33 . then , @xmath34 @xmath35 there are also the bulk stress - energy components , but those do not interest us here and do not need computation . in some situations one has to deal here with naked behavior .
this means there are cases in which the kretschmann scalar and other curvature quantities are finite on the horizon in a static coordinate system , but some of those quantities may blow up in a freely falling frame . such a kind of behavior is called naked behavior , and many instances of it have been found @xcite . another example is with quasi - black holes @xcite .
one of the features typical of quasi - black holes , consists in precisely showing naked behavior on and beyond the quasihorizon surface .
in addition , metrics obtained by gluing two spacetimes , can have a similar behavior , but now the surface stresses , which are finite in a static coordinate frame , blow up in a free - falling frame .
thus , since we have found in eqs .
( [ k(r)])-([22 ] ) the curvature and surface stresses in a static frame for the spacetime in question , we now examine the behavior of the same quantities for a free - falling frame .
consider then a radial local boost from a static frame with four - velocity @xmath36 to a free - falling frame with the velocity @xmath37 . under a boost
the four - velocity transforms according to @xmath38 , where the orthonormal vector @xmath39 is pointing in the radial outward direction , and @xmath40 is the velocity boost parameter . in relation to the tidal forces in the bulk , the curvature components ( [ k(r)])-([n(r ) ] ) in the orthonormal basis responsible for tidal forces transform according to @xmath41 @xmath42 where a bar means a quantity evaluated in the freely falling frame , @xmath43 ( see equation ( [ h(r ) ] ) for the definition of @xmath44 ) , @xmath45 , and @xmath46 is the energy of the particle frame ( see e.g. @xcite for more details ) .
the most interesting situation arises when @xmath47 is finite ( so , the kretschmann scalar is also finite ) but @xmath48 diverges .
the corresponding horizons can be called truly naked @xcite . in relation to the surface stresses , it is useful to define the quantity @xmath49 which represents the energy density of the shell as observed by the observer with the four - velocity @xmath50 . in a static frame @xmath51 .
then , considering a boosted motion along a radial geodesic with energy @xmath46 , one obtains @xmath52 this is a useful expression for analyzing naked behavior of wormholes and other objects . for the wormhole case it reduces to @xmath53 where @xmath13 is the radius at which the shell is located , and we took into account eq .
( [ s00 ] ) .
there are many ways of making wormholes @xcite . in this section
we are interested in making wormholes from charged metrics , more general than the reissner - nordstrm metric , but for certain choices the metrics can be reduced to the reissner - nordstrm metric . even for
this metric one probably can think of many manners of making wormholes .
we are interested in two different ways , that easily lead to the threshold of black hole formation , and the discussion of how they mimic black holes
. then we compound both ways into one single way .
the first way is the surgery approach , see ( see in particular section 15.2.1 of @xcite ) .
pick up a spherically symmetric metric of the form @xmath55 , where @xmath10 is the radius of the would be horizon , @xmath12 is the radius of the other possible horizon , and @xmath13 is the radius of a possible matter shell , satisfying @xmath56 .
take for instance the non - extremal reissner - nordstrm metric , where @xmath57 , with @xmath58 and @xmath59 , @xmath60 and @xmath61 being the mass and electrical charge of the object , respectively , and @xmath62 is newton s constant ( we use @xmath63 ) . cut the metric at some @xmath13 and join the resulting spacetime with a symmetric branch .
this is a non - extremal reissner - nordstrm surgery ( the schwarzschild surgery , with no charge and so @xmath64 , being a particular case of this ) , resulting in a non - extremal wormhole with a thin shell of matter at @xmath13 , the throat . in brief , one places at some radius , @xmath13 , a thin shell which separates two regions , with non - extremal geometries , with @xmath13 also defining the throat .
then , one introduces another radial coordinate @xmath30 , such that @xmath65 with @xmath66 , and which covers the whole of the manifold , @xmath67 .
the function @xmath68 is monotonically decreasing for the branch @xmath69 , which we call the @xmath70 branch , and monotonically increasing for @xmath71 , the + " branch .
in general , giving this construction , surface stresses @xmath25 appear .
a second way , i.e. , another approach , to build wormholes , is through metrics of the type @xmath72 , where @xmath0 is a small quantity , @xmath73 is the radius of the would be horizon when @xmath0 is zero , and @xmath74 is the radius of the other possible horizon @xcite .
metrics of this type , depending on the parameter @xmath0 , can be generically call @xmath0-spacetimes , which in special cases can become wormholes , i.e. , @xmath0-wormholes . in @xcite the model with metric @xmath75 , was considered , where thus @xmath76 , with @xmath77 being chosen appropriately . in turn , in @xcite the model with metric @xmath78 , was considered , where thus @xmath79 , with @xmath77 being chosen appropriately , and with @xmath80 being an additional parameter .
thus , a generic @xmath0-metric of the type given above yields a generic spacetime that comprehends the two cited models , one model studied in @xcite , the other in @xcite .
one can calculate the riemann tensor for the @xmath0-metric , and of course , since spacetime is not empty , there is a smooth energy - momentum tensor associated to the metric but we do not need to calculate it here .
all these @xmath0-spacetimes are smooth .
now , take the @xmath0-metric to construct a wormhole , i.e , an @xmath0-wormhole .
since for the construction we need to impose some more conditions , in particular on the the potentials @xmath81 and @xmath77 of the @xmath0-metric , let us adopt the following approach .
first , as above , one introduces the radial coordinate @xmath30 , such that @xmath65 and @xmath82 .
the function @xmath68 is monotonically decreasing for the @xmath83 branch , @xmath69 , and monotonically increasing for the @xmath84 branch , @xmath71 .
second , the dependence of the function @xmath81 on the parameter @xmath0 , @xmath85 , which can be of the type of the models considered above @xcite , is such that @xmath86 , and the dependence of the function @xmath77 on the parameter @xmath0 is also such that @xmath87 , so in the limit @xmath88 the original wormhole configuration indeed approaches a black hole .
third , if the first derivative @xmath89 is continuous at the throat , we have @xmath90 ( see @xcite ) .
when @xmath77 does not depend on @xmath0 at all , the throat is situated on the possible would be horizon .
this is the approach used in @xcite to build a wormhole .
this approach is smooth as long as @xmath91 . as a particular instance of this approach
, one can choose @xmath77 as being reissner - nordstrm , @xmath92 , as usual .
for @xmath93 one has a non - extremal choice for @xmath77 , the case @xmath94 , i.e. , @xmath95 , yielding the schwarzschild potential @xmath77 as a particular case , the one chosen in @xcite .
such @xmath2-wormholes have been called foils in @xcite . when @xmath96 we have the full non - extremal reissner - nordstrm metric .
so , let us compound both approaches , the surgery approach of @xcite , and the @xmath0 approach of @xcite .
write then a generic @xmath2-metric with surgery as @xmath97 i.e. , from ( [ 1 ] ) one chooses @xmath98 where @xmath0 is a small quantity , @xmath10 is the radius of the would - be horizon ( if @xmath99 ) , and @xmath13 is the radius of a possible matter shell , satisfying @xmath56 .
since we are studying here non - extremal metrics we have , when @xmath100 , that @xmath101 and @xmath102 .
essentially , what we have done is a surgery on @xmath11-metrics , non - extremal reissner - nordstrm ( with schwarzschild included ) metrics being particular @xmath100 instances .
this can also be thought of as a one - parametric deformation of the original @xmath11-wormhole metric by gluing two branches at the throat @xmath13 , with @xmath24 .
now , given the general @xmath11-metric with surgery , ( [ 1 ] ) and ( [ epsilonwormpot ] ) , and the correspondent wormhole construction , we are interested in getting a spacetime that mimics a black hole .
it is then not hard to understand that there are two distinct situations to obtain spacetimes on the threshold of being black holes .
if there is no shell , then wormholes approach black holes when @xmath103 .
if there is a shell but @xmath104 then the wormhole throat approaches the horizon when @xmath105 .
therefore , there is a play of two small parameters @xmath11 and @xmath106 and the limiting procedure should be considered with great care .
it gives rise to two distinct situations , depending on the order one takes the limiting procedures .
situation bt : this situation is achieved by , in the end turning the wormhole metric into the metric of a black hole ( b ) ( i.e. , taking @xmath103 as the last operation ) , after first having moved the shell towards the minimum throat ( t ) radius ( i.e. , taking @xmath105 as the initial operation ) .
formally , this means taking the limits in the following order @xmath107 .
situation tb : this situation is achieved by , in the end the location of the shell approaches the throat ( t ) ( i.e. , taking @xmath105 as the last operation ) , after first turning the wormhole metric into a black hole ( b ) ( i.e. , taking @xmath103 as the initial operation ) .
formally , this means taking the limits in the following order @xmath108 . note the case considered in @xcite for the metric ( [ 1 ] ) is a particular instance of the bt situation , since there @xmath109 always , and one only takes the @xmath103 limit .
so the situation bt is the one that yields black foils , following the nomenclature of ds .
one can calculate from equations ( [ n(r)])-([h(r ) ] ) that the components @xmath21 , @xmath22 and @xmath23 of the riemann tensor are always finite , and from equation ( [ 00 ] ) that both limits when applied to @xmath110 give zero , i.e. , @xmath111 .
but for the quantities @xmath47 , @xmath112 and @xmath113 ( see equations ( [ k(r ) ] ) , ( [ 22 ] ) and ( [ sigmacapital ] ) ) the situation may be different depending on the order one takes the limits .
now , as we have been seeing , in treating this problem there are many levels of distinction .
first , we can specify two models of @xmath0-metrics with surgery which depend on the parameter @xmath0 and @xmath13 , namely the model considered in @xcite , where @xmath114 , and @xmath77 is non - extremal , or some appropriate generalization of it , which we will call model 1 , and the model given in @xcite , where @xmath115 and @xmath77 non - extremal , or some appropriate generalization of it , which we will call model 2 .
second , within each of the two cases provided by model 1 and model 2 , we should study the situations bt and tb .
furthermore , as we want to examine the regularity of the system under discussion , the relevant quantities which we are going to calculate are the spacetime curvature components , and the surface stresses which appear on the glued boundary , i.e. , the shell .
we will also study the naked behavior of each case .
so , within each situation we have to study the behavior of the scalars , and in addition the naked behavior .
thus we have eight distinct cases to analyze .
we consider these eight cases , each in turn . here
we consider the one - parametric deformation , eqs .
( [ 1 ] ) and ( [ epsilonwormpot ] ) , such that for @xmath99 our metric represents the gluing of two non - extremal black holes .
note that in this non - extremal case the function @xmath77 of the metric ( [ 1 ] ) does not need to contain the parameter @xmath0 , so we put @xmath116 .
model 1 _ let the metric have the form ( [ 1 ] ) together with ( [ epsilonwormpot ] ) .
for model 1 choose the metric potentials as @xmath117 where @xmath118 can be any function that satisfies @xmath101 and @xmath119 .
for instance @xmath77 can be reissner - nordstrm , @xmath120 with @xmath121 and @xmath59 , with the case @xmath64 being the schwarzschild case , the one chosen in @xcite .
let us now work out generically the general behavior of this model , i.e. , how the curvature and stress - tensor quantities behave , and also work out generically the naked behavior .
then we apply these behaviors to the two situations bt and tb . in doing so
, we will display the properties of the quantities @xmath47 , @xmath122 , which characterize regular or singular general behavior , and the properties of the quantities @xmath48 , @xmath113 which characterize non - naked or naked behavior .
as for understanding the general behavior note that , from equations ( [ k(r ) ] ) and ( [ 22 ] ) , explicit calculations give @xmath123 and @xmath124 in relation to naked behavior , note that since from ( [ n(r ) ] ) one has @xmath125 , and from ( [ h(r ) ] ) one has @xmath126 , one finds @xmath127 .
thus , from ( [ barn ] ) @xmath128 \ , .
\label{newbarn}\ ] ] with this we can now study the situations bt and tb . 0.4 cm _
situation bt : _ as for the general behavior , one has that for a non - extremal system @xmath101 and @xmath119 .
then , it follows from equation ( [ e ] ) that @xmath129 and so , @xmath130 correspondingly , the kretschmann scalar @xmath131 also diverges .
it was shown in @xcite for the choice @xmath132 that for @xmath133 there exist geodesics which have no analogue for the schwarzschild black hole metric .
the timelike particles which move along them oscillate between turning points which are situated at different sides of the throat .
however , the problem is that in the limit @xmath103 these geodesics pass through a region of a strong gravitational field .
this gives rise to tidal forces in the radial direction which are of order @xmath134 .
if @xmath135is exponentially small @xcite , the tidal forces are exponentially large .
in addition , from equation ( [ 22e ] ) , it follows that @xmath136 now let us analyze the naked behavior .
here one has , @xmath137 , @xmath138 .
so , @xmath139 , @xmath140 .
thus , there is infinite contraction in the longitudinal direction and infinite transversal stretching .
moreover , since @xmath101 and @xmath141 , we obtain immediately from ( [ sigmacapital2 ] ) and ( [ efv ] ) that in the situation bt one has @xmath142 0.4 cm _
situation tb : _ as for the general behavior , now one finds , @xmath143 .
so , @xmath144 a result equal to that of a black hole .
also , @xmath145 , so @xmath146 note that equation ( [ inf ] ) is in agreement with the behavior of surface stresses of a wormhole obtained by gluing two copies of the schwarzschild metric ( see equation 15.46 of @xcite ) .
now let us analyze the naked behavior .
one has , @xmath147 , so @xmath48 is finite and negative .
thus , one obtains finite deformation in both directions .
moreover , it also follows from ( [ sigmacapital ] ) that in the situation tb @xmath148 i.e. , @xmath113 diverges .
thus , a free - falling observer encounters diverging surface energy density .
the same conclusion applies to the flux @xmath149 .
concluding here model 1 , we can say that there are two nonequivalent limits but each of them is bad " in that in the bt situation the kretschmann scalar diverges , whereas in the tb situation it is the surface stresses that diverge . 0.5 cm 7.4 cm _
_ model 2 _ _ 0.1 cm let the metric have the form ( [ 1 ] ) together with ( [ epsilonwormpot ] ) .
for model 2 choose the metric potentials as @xmath150 with @xmath151 and @xmath0 being parameters .
in @xcite the model with @xmath152 , was considered , in which case , when @xmath99 one has the schwarzschild metric , see also @xcite . here
, @xmath153 can be any function that satisfies @xmath101 and @xmath154 , a typical example being the reissner - nordstrm @xmath77 potential .
as for the general behavior , again after some calculations , we obtain @xmath155 and @xmath156 now let us analyze the naked behavior . also , again after some calculations , we obtain @xmath157 , so it follows from ( [ barn ] ) that @xmath158 \label{n2}\ ] ] with this we can now study the situations bt and tb
. 0.4 cm _
situation bt : _ as for the general behavior , in this situation one finds , @xmath159 and @xmath160 now let us analyze the naked behavior .
one has @xmath161 , and @xmath162 so one finds no longitudinal deformation , and an infinite transversal stretching . taking into account that @xmath101 , @xmath163\neq 0 $ ] , we obtain immediately from ( [ sigmacapital2 ] ) that in situation bt one finds @xmath164 . 0.4 cm _ situation tb : _ as for the general behavior in this situation , one finds , @xmath165 and @xmath166 now let us analyze the naked behavior .
one finds , finite transversal stretching and finite longitudinal contraction ( @xmath167 ) or stretching ( @xmath168 ) .
it follows from ( [ sigmacapital2 ] ) and ( [ efv ] ) that in the situation tb , @xmath169 i.e. , it diverges .
thus , a free - falling observer encounters diverging surface energy density in the situation tb .
the same conclusion applies to the flux @xmath170 .
concluding here model 2 , we can say that in both situations , bt and tb , the curvature components remain finite but the limiting values of @xmath47 do not coincide .
moreover , in both situations the surface stresses diverge . 0.5 cm _ overall conclusions for models 1 and 2 _ 0.1 cm as an overall conclusion for the situation tb in models 1 and 2
, we find the results agree for both models .
this can be explained and generalized as follows . since ,
in the situation tb , the limit @xmath88 is taken first , the dependence of the metric on @xmath0 drops out in the final expressions for the curvature and surface stresses , so any model gives the same result .
in addition , assuming that for @xmath99 one has @xmath171 , and taking into account that @xmath101 and @xmath172 , we obtain immediately from ( [ sigmacapital2 ] ) that in the situation bt , @xmath164 , and in the situation tb , @xmath113 diverges .
this holds independently of the kind of the model used .
thus , a free - falling observer encounters diverging surface energy density in the situation tb .
the same conclusion applies to the flux @xmath173 . in the non - extremal cases just studied , it turned out that each of the limits under discussion is singular : either the kretschmann scalar or surface stresses on the throat ( or both ) diverge .
thus , the limit is singular . in other words ,
a black hole mimicker made from a wormhole , and in particular a black hole foil , is not smooth .
it is convenient to summarize the results in a table shown in figure [ table ] .
3.7 cm [ width=16.0 cm ] table.eps 0.5 cm although the singular or almost singular behavior of these black hole mimickers based on @xmath0-wormholes casts doubts on their real existence , it is worth our while to study a little more on the effects of such mimickers on infalling sources and their detection by far away observers .
intuitively it is clear that there should be some observational effects if infalling sources are distorted by stronger than normal tidal fields .
first , we point out that indeed the fact that tidal forces grow unbound when @xmath103 , can be used , in principle , to distinguish a black hole from an @xmath0-wormhole which mimics it .
suppose a small mass falling freely into a massive body , such as an @xmath0-wormhole .
consider , for example , the situation bt .
one can compare two approaches . in the first approach ,
one neglects the size of the small mass and considers the geodesic along which such a point - like small mass moves @xcite . then
, if the throat is very close to the would - be horizon and subsequent pulses are emitted near the throat , the intervals of time measured at infinity , grow unbound as @xmath174 in the limit @xmath175 .
so an observer at infinity can not distinguish the fall of matter into a wormhole with vanishingly small @xmath11 from absorption of matter by a black hole , if the interval of observation time is less than @xmath176 , by construction a very long interval , see @xcite for details .
in the second approach , however , one takes the finiteness of the size of the free - falling small mass into account .
then , the overall picture changes since due to the growing tidal forces the small mass gets deformed .
if the small mass is a luminescent source , such a change can in principle be detected by an observer at infinity .
moreover , such changes happen much quicker than the typical times needed to penetrate the immediate vicinity of a horizon .
the intervals of time during which the size of a small mass changes can be estimated from eq .
( [ kinf ] ) and the geodesic deviation equation . for a process occurring near the throat it yields for the proper time the value @xmath177 , so that @xmath178 , as @xmath103 .
the corresponding interval of time @xmath179 at infinity is @xmath180 is finite .
therefore , in case of stretching , an observer at infinity would see an extended distorted image of an infalling small mass .
if the tidal forces are infinite , as they can be in some of the cases shown above for @xmath0-wormholes , then the stretching is correspondingly large .
of course , it is also possible that infinite contraction in some direction converts an extended small mass to zero size in this direction , see @xcite . for the usual black holes , tidal forces
are also present but such forces are much weaker than for @xmath0-wormholes , where near the throat they can be as large as one likes for sufficiently small @xmath0 .
thus , the key point is to look not to single geodesics as in @xcite , but to the separation between geodesics of the same congruence .
such a separation delivers , to an observer at infinity , meaningful information about the region of strong gravity where a black hole or a @xmath0-wormhole can be situated , and such information has very different properties , which depend on the massive body in question .
second , we indicate what further physical changes could be expected by performing some interesting estimates of the effects . here , one can take advantage of known results ( see @xcite , section 32.6 ) .
suppose then that a box of small proper mass @xmath181 and proper length @xmath30 is freelly falling towards an object of mass @xmath60 .
to simplify we consider an uncharged object , @xmath95 . from @xcite one
finds that the radial tidal force exerted on the box is equal to @xmath182 where @xmath47 has the meaning of a tidal radial acceleration .
indeed , @xmath47 is given precisely in eq .
( [ k(r ) ] ) . for the schwarzschild metric one
has @xmath183 , so that near the horizon of a schwarzschild black hole , where @xmath184 , one obtains the value @xmath185 .
however , for model 1 , if we correspondingly choose the metric potential as @xmath186 in the situation bt , it follows from ( [ kinf ] ) that near the would - be horizon one has @xmath187 .
let us suppose that a distant observer is able to recover from observational data the value of @xmath47 . if the observer thinks that the measured value of @xmath47 is due to a schwarzschild black hole , he should ascribe a mass @xmath188 to it .
however , if the object turns out to be a black hole mimicker , then , for the same @xmath47 , the value of the actual mass will be much greater , given by @xmath189 . if , in addition , the observer , knowing the mass @xmath60 of the object , in this case @xmath190 , insists in explaining it in terms of the usual black hole metric , he will find that @xmath191 , and will certainly start asking about the hidden mass " .
this example shows that hidden mass " in some situations may arise simply because the metrics of a black hole or of a black hole mimicker were not properly discerned .
as far as gravastars are concerned , they contain , by construction , a thin layer of normal matter with positive density @xmath192 and positive radial pressure @xmath193 on the border of the tension matter with vacuum . in the model
suggested in @xcite the stiff matter equation of state with @xmath194 was chosen .
then , it follows immediately from the field equations and the conservation laws that , as the border approaches the gravitational radius , the gradient of the pressure becomes infinite .
one can , in general , admit discontinuous radial pressure , giving rise to a surface pressure on the boundary between matter and vacuum . however , this surface pressure and other surface stresses also grow unbound in the horizon limit .
surely , the more the border approaches the horizon , the better a black hole mimicker it becomes , but , at the same time , the closer the system approaches the singular state .
this does not exclude in advance the astrophysical significance of gravastars as compact vacuum - like geometries but it shows that they can hardly pretend to be good black hole mimickers . in more detail , to see that the surface stresses go unbound
, we introduce the quantity @xmath195 , so that we can rewrite eq .
( [ 22 ] ) as @xmath196 + \frac { 1 } { b(r_{0})}\left[\left ( \frac{db}{dl}\right ) _ { + } -\left ( \frac{db } { dl}\right )
_ { -}\right]$ ] , where we have taken into account that the continuity of the first fundamental form demands @xmath197 . in the outer region
we have the schwarzschild metric , so @xmath198 .
let the radius of the shell approach that of the would - be horizon , i.e. , @xmath105 , and @xmath199 . by definition of a gravastar
, actually there is no horizon in the system , so the function @xmath200 can not cross @xmath4 at @xmath10 at all . in the inner region , either @xmath201 or @xmath202 .
thus , since there is a @xmath203 @xmath204 term in @xmath205 , and the other terms remain finite and non - zero , we find that @xmath206 .
this makes the gravastar unphysical in the near - horizon limit .
it is worth noting that this conclusion and its derivation is very close to the statement that quasi - black holes can not be non - extremal , if only finite surface stresses are allowed ( see sec .
iv of @xcite ) .
the only difference is that in the whole inner region @xmath207 everywhere for quasi - black holes whereas for gravastars @xmath208 can be non - zero there .
however , in the present context , only the vicinity of the would - be horizon is relevant , so the conclusions are similar .
in section [ nonextremalepsilonsurg ] on wormholes on the basis of nonextremal @xmath0-metrics with surgery , we have discussed wormhole configurations which mimic non - extremal black holes . here
we consider the one - parametric deformation ( [ 1 ] ) such that for @xmath99 our metric represents the gluing of two extremal reissner - nordstm black holes .
note that in the non - extremal case , when using metric ( [ 1 ] ) we have chosen a function @xmath77 which itself does not contain the parameter @xmath2 .
however , now such a simple construction can not be implemented .
our goal is to trace the relationship between a wormhole metric with a generic @xmath0 parameter ( @xmath209 ) , and a black hole metric ( @xmath99 ) .
if we start from equation ( [ 1 ] ) with @xmath210 , i.e. , @xmath211 , we encounter the immediate difficulty that spacetime is geodesically complete and represents an infinitely long horn , so one can not speak about a wormhole at all .
therefore , we should deform the extremal reissner - nordstrm metric in a somewhat different way and include the parameter @xmath0 not only into the function @xmath212 but into @xmath77 as well .
let us make the simplest choice for @xmath213 , namely , @xmath214 @xmath215 being such that @xmath216 , and @xmath217 . at some @xmath24 , we glue two copies of the spacetime , the `` @xmath218 '' branch to the `` @xmath219 '' branch .
then , the behavior of @xmath47 and @xmath122 follows from eqs .
( [ e ] ) and ( [ 22e ] ) .
note that if @xmath220 and @xmath221 we return to the deformed non - extremal reissner - nordstrm case and the results ( [ einf])-([0 ] ) are reproduced , @xmath222 being the schwarzschild case .
0.3 cm 7.4 cm _ model 1 _ 0.1 cm using ( [ efv ] ) , i.e. , @xmath223 , and ( [ vne ] ) , we can study model 1 in the extremal case . 0.4 cm _
situation bt : _ as for the general behavior one finds , @xmath224 and , @xmath225 let us analyze the naked behavior . in principle , the quantity @xmath226 may be finite or infinite depending on the model for @xmath227 . from eq .
( [ eqk ] ) , longitudinal contraction is finite if @xmath40 is finite , or infinite if @xmath40 is infinite
. the value of @xmath48 which determines transverse deformation can be found from eq .
( [ newbarn ] ) .
then , we obtain that in the limit under discussion @xmath228 where @xmath229
. then we have transverse stretching which is finite if @xmath230 is finite , or infinite if @xmath231 is infinite .
the stress component @xmath232 is finite , indeed zero , in the limit .
situation tb : _ as for the general behavior one finds , @xmath233 @xmath234 note that when @xmath235 in the situation bt , the quantities @xmath47 for two situations coincide .
let us analyze the naked behavior .
here there is finite longitudinal contraction , and no transverse deformation since , according to eq .
( [ newbarn ] ) , @xmath236 in the limit under consideration .
thus , the only manifestation of naked behavior is connected with the surface stresses . according to ( [ sigmacapital2 ] )
, the quantity @xmath237 behaves as @xmath238 i.e. , it diverges in situation tb . 0.3 cm 7.4 cm _
model 2 _ 0.1 cm using eq .
( [ phimodel2 ] ) , i.e. , @xmath239 , and ( [ vne ] ) , we can study model 2 in the extremal case . _
situation bt : _ as for the general behavior one finds , @xmath159 @xmath240 let us analyze the naked behavior .
there is no deformation in the radial direction .
the behavior of @xmath48 which is responsible for transverse deformation , can be obtained from ( [ n2 ] ) and coincides with ( [ nb ] ) in the limit under discussion .
thus , again transverse stretching is finite if @xmath231 is finite , or infinite if @xmath231 is infinite . note that it follows from the definitions of @xmath40 and @xmath231 that @xmath241 .
according to the definition ( [ vne ] ) of @xmath242 , @xmath243 .
therefore , if @xmath231 is finite , @xmath244 , then @xmath47 , @xmath122 , @xmath48 and @xmath113 are finite ( moreover , @xmath245 ) , so naked behavior is absent in this case . on the other hand , as @xmath246 , in case @xmath247 the quantity @xmath231 diverges and so does @xmath48 .
this means transverse stretching is infinite .
situation tb : _ as for the general behavior one finds , @xmath233 @xmath248 let us analyze the naked behavior .
one finds , finite contraction in the longitudinal direction and no transverse deformation .
@xmath238 i.e. , it diverges .
0.5 cm _ overall conclusions for models 1 and 2 _ 0.1 cm as an overall conclusion , assuming that for @xmath100 , one has @xmath249 and taking into account that @xmath250 , @xmath163\neq 0 $ ] , we obtain immediately from ( [ sigmacapital2 ] ) that in the situation bt , @xmath164 , and in the situation tb , @xmath113 diverges .
this holds independently of the kind of the model , and is valid for the non - extremal spacetimes , as well as for the extremal spacetimes .
thus , a free - falling observer encounters diverging surface energy density in the situation tb .
the same conclusion applies to the flux @xmath251 . 0.1 cm _ light - like shells _ 0.1 cm one essential feature of the almost extremal configurations under discussion consists in that we deal with shells which are timelike but become light - like in the limit when they are approaching the would - be horizon .
a natural question arises : what happens if we start out our analysis with two extremal black hole spacetimes , i.e. , putting @xmath252 and @xmath253 , and the shell in - between the two spacetimes lies on the event horizon and so is lightlike ?
more precisely , we are interested whether the stresses on the shell remain finite or become infinite
. one can expect from our previous results that they are finite but this is not so obvious in advance .
we have seen already that taking different limiting procedures in the near - horizon limit is a rather delicate issue .
moreover , the formalism for lightlike shells @xcite somewhat differs from that for timelike ones , so we perform our analysis anew .
we restrict ourselves to the case of taking two extremal reissner - nordstrm black holes .
the reason for this choice is that only extremal black holes are good candidates for gluing without severe singularities , i.e. , the gluing procedure maintains a finite kretschmann scalar throughout the spacetime as well as finite stresses on the glued surface .
this case is also a majumdar - papapetrou case , and could be analyzed in the next section as well , but since it is also a limit of what has been done above , namely the null limit of a timelike wormhole throat with two extremal external vacuum spacetimes , we discuss it now . to match two extremal reissner - nordstrm spacetimes at the light - like surface @xmath253 we follow the general formalism for light - like shells @xcite .
we write the metric in kruskal - type coordinates @xmath254 let the surface @xmath253 correspond , say , to @xmath255 . then , the effective energy density is @xmath256 , ( see , e.g. , equation ( 3.99 ) of @xcite ) .
here @xmath257n^{\alpha } $ ] , where the indexes @xmath84 and @xmath83 refer to the different sides of the shell , and the null vector @xmath258 is such that @xmath259 , with @xmath260 being a null tangent vector .
now , the only nonvanishing components of @xmath261 and @xmath262 , are by construction @xmath263 and @xmath264 .
since there is no rest frame in the null case , the measured energy density @xmath265 depends on the chosen observer . to check that it is finite and non - zero
, it is sufficient to check that @xmath266 $ ] is finite and non - zero . for this purpose , it is sufficient to exploit the result of @xcite where it is shown that @xmath267 on the horizon . in our case , as we deal with two black holes instead of a single one , the coordinate @xmath4 increases on both sides of the shell .
thus , @xmath268 has different signs and takes the values @xmath269 on each side .
so indeed , the difference is equal to @xmath270 and is finite and non - zero . in general , there are other contributions to the effective stress - energy tensor of the shell due to effective pressures and currents , but it is easy to show that in our case they are absent .
it is worth noting that the gluing in our case differs from the gluing between different extremal reissner - nordstrm black holes considered in @xcite , in that we replaced the usual metric inside the horizon by their @xmath83 branch .
it also differs from the wormhole construction used in @xcite where two spheres were cut out from a vacuum majumdar - papapetrou system , or more precisely , they were cut from a single spacetime containing two extremal reissner - nordstrm black holes .
0.3 cm _ classical electron models _ 0.1 cm as a by - product of this light - like shell construction , and a very interesting one , we have just found a configuration that represents a regular wormhole configuration which is also a black hole .
more important perhaps , in addition , it can serve as a classical model for an elementary particle in that ( i ) the system is characterized by a minimum number of fixed parameters like mass and charge and ( ii ) it is free of a central singularity inside .
we are aware that it is not entirely of electromagnetic nature because it has stresses on the horizon , a kind of poincar stresses .
but , anyway , such a surface stress , can be considered as a mild singularity when compared to the usual central singularity . in this model
a free - falling observer can penetrate to the inside but , of course , can not return back to the original asymptotic region , due to the existence of a horizon .
so the wormhole is an untraversable one .
thus , the body under discussion combines features of an untraversable wormhole and a regular black hole , and can be called a worm - black hole . since the proper distant to the extremal horizon is infinite , such a hybrid construction is similar to the null wormholes , or n - wormholes for short , see @xcite .
note that it differs from configurations representing quasi - black holes .
for example , in some quasi - black hole models , see @xcite , there is a region @xmath271 which becomes degenerate in the quasihorizon limit .
this region is missing in our worm - black hole model , since the black hole metric beyond the horizon is replaced by a branch with an areal radius that grows away from the horizon .
it also differs from the model considered in @xcite where the external reissner - nordstrm part was glued to the bertotti - robinson metric that leads to surface stresses that vanish in the horizon limit in the static frame but grow unbound in the free - falling one , and as a result , the inner region becomes impenetrable for a free - falling observer ( see @xcite for details ) .
another candidate for the role of a black hole mimicker is a quasi - black hole . roughly speaking
, it is an extremal object that appears when the system approaches the quasihorizon as nearly as one likes , along a family of quasi - static configuration .
it was argued in @xcite that such a limit can correspond to an extremal quasihorizon only if we restrict ourselves to static configurations which are regular in the strong sense .
the latter means that the kretschmann scalar should be finite everywhere in the system , and surface stresses at the quasihorizon should be finite as well .
there are subtleties in the non - trivial relation between regular and singular features of quasi - black holes @xcite .
for example , the whole region can look degenerate from the viewpoint of a distant observer and , nevertheless , the kretschmann scalar remains finite in that region . in more detail
, consider the static spherically symmetric metric ( [ 1 ] ) , and let it represent a spacetime in which there is an inner matter configuration , attached to an asymptotic flat exterior region .
the @xmath5 in ( [ 1 ] ) stand for the radius @xmath13 of the configurations and possibly some other parameters connected with the particular object one is analyzing .
for instance , the parameter @xmath0 also enters in the analysis , and here has a slightly altered meaning .
it means a small deviation from a quasi - black hole , rather than from a black hole solution as in section [ nonextremalepsilonsurg ] , see also below .
suppose the spacetime in question has the following properties : ( a ) the function @xmath118 in ( [ 1 ] ) attains a minimum at some @xmath272 , such that @xmath273 , with @xmath274 , this minimum being achieved either from both sides of @xmath275 or from @xmath276 alone , ( b ) for such a small but nonzero @xmath0 the configuration is regular everywhere with a nonvanishing metric function @xmath277 , at most the metric contains only delta - function like shells , and ( c ) in the limit @xmath278 the metric coefficient @xmath279 for all @xmath280 .
these three features define a quasi - black hole . in turn , these three features imply that , there are infinite redshift whole regions when @xmath281 , a free - falling observer finds in his own frame infinitely large tidal forces in the whole inner region , showing thus naked behavior , although the curvature scalars are finite .
moreover it has some form of degeneracy since , although the spacetime curvature invariants remain perfectly regular everywhere , in the limit , outer and inner regions become mutually impenetrable and disjoint . for a free - falling external nearby observer
it is as if a null singular horizon is being formed . for external far away observers
the spacetime may be said to be naively indistinguishable from that of extremal black holes .
however , if one makes experiments with infalling luminescent extended small masses , one might find differences , since as discussed previously @xcite , due to the naked behavior , quasi - black holes enlarge grossly the tidal forces on an infalling small mass when compared to the tiny effect of an extremal black hole on the same small mass .
thus , as with the extremal @xmath0-metrics studied before , the naked behavior shows that quasi - black holes are not so good mimickers , as was previously thought , but they are still better than black foils @xcite where the singularity is more severe .
a further important property is that quasi - black holes must be extremal . for a quasi - black hole the metric
is well defined and everywhere regular .
however , when @xmath99 , quasi - black hole spacetimes become degenerate , almost singular , see @xcite .
the quasi - black hole is on the verge of forming an event horizon , but it never forms one , instead , a quasihorizon appears . in summary quasi - black holes
have normal general behavior and singular naked behavior .
quasi - black holes may appear from bonnor stars , i.e. , systems composed of extremal charged dust and vacuum , from self - gravitating higgs magnetic monopole systems , and from composite spacetimes even in the case of pure electrovacuum , in which these vacuum systems are composed of an exterior reissner - nordstrm part glued to an inner bertotti - robinson spacetime or of an exterior reissner - nordstrm part glued to an an inner minkowski spacetime , see @xcite for a full discussion and references .
bonnor stars are majumdar - papapetrou matter systems with either a sharp or smooth boundary to an exterior vacuum . since bonnor stars are paradigmatic to understand the formation of quasi - black holes ( see @xcite )
, it is interesting to use those stars on the threshold of forming a quasi - black hole to understand whether wormholes on the basis of quasi - black holes which can be formed from bonnor stars , can mimic extremal black holes or not .
this is an interesting variant , although with similarities , to wormholes on the basis of extremal @xmath0-metrics .
we use bonnor stars , both in their compact version @xcite , as well as in their extended one @xcite .
now , we start from the configuration which contains majumdar - papapetrou matter inside and vacuum outside , a compact bonnor star @xcite .
let us have a compact object , a bonnor star , with extremal dust for @xmath282 , joined to an extremal reissner - nordstrm metric for @xmath283 .
the potential @xmath284 can be written as @xmath285 with the mass density @xmath286 and the function @xmath287 being connected through @xmath288 .
the function @xmath287 can be interpreted as the proper mass enclosed within a sphere of a radius @xmath4 .
in addition , for extremal dust , one has @xmath289 , where @xmath290 is the electric charge within this sphere . at the boundary @xmath291
, one has @xmath292 , and @xmath293 , such that @xmath294 , @xmath60 and @xmath61 being the total mass and charge , respectively . to construct a wormhole
, one can take the following procedure .
cut the solution somewhere in the interior at some radius @xmath295 and discard the region @xmath296 .
one obtains a majumdar - papapetrou matter region for @xmath297 , and a vacuum region for @xmath298 . for definiteness , let this spacetime be situated on the left .
a symmetric right branch , also containing interior and exterior , is again cut at the radius @xmath299 , and glued to the symmetric left branch , producing thus a boundary shell at @xmath300 .
then , it follows from the field equations that the left branch is given by @xmath301 \,,\quad r_{1}\leq r\leq r_{0}\ , , \notag \\
\exp ( \phi ) & = & 1-\frac{m}{r}\,,\quad r_{0}\leq r<\infty \ , , \label{phibon}\end{aligned}\ ] ] and that the proper radius @xmath30 and the coordinate radius @xmath4 are related by , @xmath302 as the bonnor star is here a compact object , the proper distance from @xmath300 to @xmath13 is finite , and tends to zero in the limit @xmath303 .
therefore , in this limit , the matter between right and left boundaries becomes negligible and the construction corresponds to gluing two extremal reissner - nordstrm black holes in the situation tb of section [ nonextremalepsilonsurg ] .
thus in the limit of our interest , @xmath304 , it is not surprising that the results coincide with ( [ 2bt1 ] ) and ( [ 2bt2 ] ) where @xmath305 .
indeed , one finds @xmath306 and @xmath307 being finite as well .
we have put @xmath294 , as is the case for these systems .
there is yet another procedure to produce a wormhole . in the above considerations
, we performed a symmetric construction , in the sense that the @xmath84 and @xmath83 branches differed by the sign of @xmath89 only .
now , we start again from a compact bonnor star configuration which contains matter inside and vacuum outside .
we want to preserve this feature , and so we have to make a non - symmetric deformation .
thus , we change the procedure and consider the following construction . again
, a bonnor star is made of majumdar - papapetrou matter for @xmath282 , which in turn is joined to an extremal reissner - nordstrm metric for @xmath283 .
now , in the region @xmath282 , the left @xmath83 branch , choose the distribution with @xmath308 .
the potential @xmath118 can be written again as in eq .
( [ vm ] ) . then , for the left branch , it follows from the field equations that @xmath309 \,,\quad 0\leq r\leq r_{0}\ , , \label{phibon2}\ ] ] and @xmath310 it is worth paying attention that
because of the property @xmath311 , the matter distribution which was originally compact , turned after deformation into a non - compact one since at left infinity @xmath312 , where @xmath313 , @xmath314 .
for the right @xmath84 branch we use the extremal reissner - nordstrm metric with the mass @xmath294 , which gives , @xmath315 and @xmath316 in the limit @xmath317 , @xmath318 one finds that @xmath319 and @xmath320 being finite as well .
we have put @xmath294 , as is the case for these systems .
thus , extremal quasi - black wormholes , made of majumdar - papapetrou matter are possible .
their distinctive feature is the presence of finite non - zero surface stresses on the horizon .
curvature components remain finite .
it is worth noting that , although in this subsection we did not introduce the parameter @xmath0 explicitly , actually its role is played , say , by the difference @xmath321 in the sense that this quantity is responsible for the deviation of the spacetime from its limiting state ( a quasi - black hole or two quasi - black holes glued together ) . here
we exploit the distribution of extremal charged dust given in @xcite . near the quasi - black hole limit ,
the first order corrected quasihorizon has radius @xmath275 given by @xmath322 \ , , \label{radiusquasi}\ ] ] where @xmath323 is the parameter that yields the deviation from the reissner - nordstrm solution , and @xmath324 is a quantity with units of electric charge , which is indeed the total charge @xmath61 when @xmath325 from the outset ( see @xcite for details ) .
the quasi - black hole limit is such that @xmath326 , with @xmath327 . in a sense
, the dimensionless parameter @xmath328 here corresponds to the @xmath0 parameter in the @xmath2-wormhole construction of a previous section .
then the solution has asymptotics near the quasihorizon @xmath275 given by @xmath329 and @xmath330 consider again the @xmath84 and @xmath83 branches , symmetric relative to the first order corrected quasihorizon radius in the region @xmath331 , but with different signs of @xmath89 .
each branch has the same dependence of the metric potential , eqs .
( [ as1])-([as2 ] ) , on @xmath4 .
thus , for simplicity , we restrict ourselves to the analog of the situation bt considered previously , which translated to here means taking the limits as follows , @xmath332 . then , simple calculations show that @xmath333 and @xmath334 being finite as well . in ( [ lim ] ) we have put @xmath335 in the limit @xmath336 and @xmath294 also in this limit , where @xmath60 is the mass of the configuration . in the limit under consideration the metric in the region @xmath337 is given by the extremal reissner - nordstrm metric , whereas in the immediate vicinity of the quasihorizon the metric is described by the bertotti - robinson metric . therefore , as in the preceding subsection , our construction gives two extremal reissner - nordstrm black holes glued along the quasihorizon with different signs of @xmath89 on opposite sides .
we have studied wormhole and other configurations as possible mimickers of black holes .
we have separated the configurations into non - extremal and extremal . for wormholes ,
we have examined separately two limiting procedures in which the wormhole throat approaches the black hole horizon . in the first procedure , we fix the location of an observer ( exactly on the throat ) .
then , we change the spacetime ( making a wormhole on the verge of being a black hole ) .
this is the situation bt . in the second , we change spacetime ( making a wormhole on the verge of a black hole ) , then place the shell outside the throat and move it toward the throat ( which coincides with the horizon )
this is the situation tb .
this procedure is carried out for non - extremal and extremal configurations separately . in the non - extremal case it turned out that each of the limits under discussion is singular : either the kretschmann scalar or surface stresses on the throat ( or both ) diverge .
thus , the limit is singular . in other words ,
a mimicker of a non - extremal black hole , made from a non - extremal wormhole , including the black foil of @xcite , is not smooth .
we have summarized the results for the non - extremal case in a table . for the extremal case both the kretschmann scalar and surface stresses remain finite .
this pronounced distinction between properties of the limiting configurations in the non - extremal and extremal cases is one counterpart of the conclusion made in @xcite that quasi - black holes can be only extremal .
however , one should not forget about some subtleties connected with the fact that singular behavior can , in general , manifest itself not only in the value of the kretschmann scalar . even if this scalar is finite , a naked behavior is possible or even inevitable as was shown in in the present paper and in @xcite ( see , e.g. , sec .
v of @xcite for a discussion about other subtleties in which the singular features of quasi - black holes are revealed ) .
there is also another candidate for the role of a non - extremal black hole mimicker , a gravastar @xcite .
however , the corresponding surface stresses grow unbound when the radius approaches the gravitational radius , as we have seen . from an astrophysical viewpoint , the situation bt in the non - extremal case ,
i.e. , a black foil , is more interesting since it implies no necessity of making a shell by hand .
it is the case considered in @xcite .
one of the questions raised in @xcite is whether it is possible or not to distinguish between a true non - extremal black hole , a schwarzschild black hole say , and a wormhole .
the main conclusion of @xcite is that it is impossible to distinguish for any finite time in the limit under discussion .
this conclusion is reached on the basis of considering properties of bodies moving along separate fixed geodesics and emitting signals detected at infinity . however , if from single geodesics we shift our attention to a congruence of geodesics , it turns out that the strong gravity forces , on the near horizon region , leave their imprint on the form of a moving body and , thus , on the properties of signals which an observer at infinity is detecting .
if the surface of the body is luminescent , an observer at infinity would see either a finite width instead of a point , a continuous detection instead of separate pulses , and so on .
it is essential that in the case discussed in @xcite the corresponding proper time of deformation tends to zero when the curvature grows unbound , with the time at infinity being finite .
thus , at least in principle , an observer can distinguish between a black hole and an almost singular wormhole .
the singular nature of the limit in the non - extremal case makes also questionable the applicability of the membrane paradigm used in @xcite . the key point in this paradigm
consists in boundary conditions according to which a free - falling observer sees a finite value of physical fields on the horizon ( see @xcite , sec .
ii ) . however , in the problem under discussion , typically , this observer as well as the geometry itself become ill - defined . only in some cases ( see model 2 , situation tb ) the curvature components remain finite in the free - falling frame . but even in such situations the infinite surface stresses on the horizon surface makes the physical meaning of the membrane paradigm unclear since this paradigm relies heavily on the concept of a regular surface .
objects based on nearly extremal wormholes , although of less interest astrophysically perhaps , have a much better behavior in the sense that both the geometry and surface stresses remain finite .
moreover , typically , there is no naked behavior . in this case , the effect of strong curvature is much less pronounced than in the case of quasi - black holes where a naked behavior is typical @xcite . in this sense ,
a wormhole composed on the basis of two extremal black holes seems to be the best mimicker of an extremal black hole .
as by - product , we have obtained a model of a regular black hole !
thus , if we try to arrange a ranking of black hole mimickers , both non - extremal and extremal , the list looks as follows from top to bottom : wormholes on the basis of extremal black holes or on the basis of quasi - black holes , quasi - black holes , wormholes on the basis of non - extremal black holes ( and within these the best are black foils ) , and gravastars .
bearing in mind that in observational astrophysics it is difficult to find extremal configurations ( the would be best mimickers ) , whereas non - extremal configurations are really bad mimickers , the task of distinguishing black holes from their mimickers seems to be less difficult than one could think of it .
in the present paper we have restricted ourselves to static spherically symmetric spacetimes .
meanwhile , in a recent work @xcite the status of black hole mimickers is undermined in the rapidly rotating case as well since it is argued that they are unstable . we have also circumscribed our discussion to particular wormholes , gravastars and quasi - black holes @xcite-@xcite , since these objects are well adapted to our goal of examining the near - horizon properties and their connection with far away asymptotic properties . however , there are many other objects with properties that make them also potential black hole mimickers ( see , e.g. , @xcite ) and which are worthy of study within our formalism .
o. z. thanks centro multidisciplinar de astrofsica
centra for hospitality and a stimulating working atmosphere .
this work was partially funded by fundao para a cincia e tecnologia ( fct ) - portugal , through project poci / fp/63943/2005 . | black hole mimickers are possible alternatives to black holes , they would look observationally almost like black holes but would have no horizon
. the properties in the near - horizon region where gravity is strong can be quite different for both type of objects , but at infinity it could be difficult to discern black holes from their mimickers . to disentangle this possible confusion , we examine the near - horizon properties , and their connection with far away asymptotic properties , of some candidates to black mimickers .
we study spherically symmetric uncharged or charged but non - extremal objects , as well as spherically symmetric charged extremal objects . within the uncharged or charged but non - extremal black hole mimickers , we study non - extremal @xmath0-wormholes on the threshold of the formation of an event horizon , of which a subclass are called black foils , and gravastars . within the charged extremal black hole mimickers we study extremal @xmath1-wormholes on the threshold of the formation of an event horizon , quasi - black holes , and wormholes on the basis of quasi - black holes from bonnor stars .
we elucidate , whether or not the objects belonging to these two classes remain regular in the near - horizon limit .
the requirement of full regularity , i.e. , finite curvature and absence of naked behavior , up to an arbitrary neighborhood of the gravitational radius of the object enables one to rule out potential mimickers in most of the cases . a list ranking the best black hole mimickers up to the worse , both non - extremal and extremal , is as follows : wormholes on the basis of extremal black holes or on the basis of quasi - black holes , quasi - black holes , wormholes on the basis of non - extremal black holes ( black foils ) , and gravastars .
since , in observational astrophysics it is difficult to find extremal configurations ( the best mimickers in the ranking ) , whereas non - extremal configurations are really bad mimickers , the task of distinguishing black holes from their mimickers seems to be less difficult than one could think of it . |
the cross section of the reaction @xmath0be(p,@xmath1b has recently become the subject of renewed intense scrutiny [ 1 ] owing to the pre - eminence of the @xmath3b reaction products as the main source of high energy neutrinos in the interior of the sun and , hence , the possible implications for the so - called solar neutrino problem " .
a number of new precision measurements of this cross section in various stages of preparation have been reported lately . we have set up an experiment for such a measurement at the weizmann institute , focusing on one of the major sources of uncertainty in previous experiments : the homogeneity of the areal density of the target material .
in general , when a nonuniform particle beam impinges on a nonuniform target , the reaction yield is given by : y = ds where @xmath6 are the respective total numbers of beam and target particles and @xmath7 are areal densities .
+ only when the target is known to be uniform and the beam is smaller than the target can eq.(1 ) be simplified to : y = ds = n_b in such a case , the evaluation of the cross section is independent of the areal distribution of the particle beam . on the other hand , in other cases - e.g. for radiochemically produced @xmath0be targets [ 2 ] -
the target can not be assumed to be strictly uniform and the full relation ( 1 ) has to be used in the evaluation .
the inherent uncertainties in the distributions @xmath8 and @xmath9 may thus lead to considerable uncertainties in the value of the integral and hence in the deduced cross section .
we have addressed this problem by inverting the arrangement : we use a homogeneous beam - produced by raster scanning - impinging on a target than the beam .
the relation ( 1 ) then reduces to : y = n_t independent of the potentially problematic determination of the target areal distribution . as a first step we measured the cross section of the reaction @xmath0li(d , p)@xmath3li at the top of the resonance at e@xmath4(lab.)=776 kev
the mirror nuclei @xmath3li and @xmath3b have very similar mean lives and similar @xmath10 decays ( @xmath11 and @xmath12 , respectively ) to @xmath3be @xmath13 2@xmath14 .
the main differences are ( i ) the much easier preparation of a @xmath0li target and ( ii ) the much higher yield from @xmath0li(d , p ) .
the reaction @xmath0li(d , p)@xmath3li can therefore provide a convenient check of the equipment and the method . beyond that , this reaction is of significance in its own right and as a stepping stone and calibration in some @xmath0be(p,@xmath5)@xmath3b experiments [ 1 - 4 ] .
+ in the course of the present investigation we have examined a new issue : reaction - product nuclei being backscattered from the target backing , leaving the target assembly and reducing the number of detected @xmath14 particles .
we present below experimental results which probe this issue and compare them to computer simulations .
the influence of this effect on past and future measurements is discussed .
the general scheme of the experiment is shown in fig .
1a . a d@xmath15 beam out of the weizmann institute 3 mv van de graaff accelerator is raster scanned over a rectangle of 7 mm@xmath16 6 mm .
the purpose of the scan is to obtain a beam of uniform areal density .
the d@xmath15 beam ( as well as the p beam , see below ) is collimated by a 3 mm diameter hole and impinges on a circular target of lif of 2 mm diameter , evaporated on al and pt foils ( see discussion below ) ; the target spot is aligned with a set of variable collimators downstream of the target . the target is mounted on an arm which is periodically moved out of the beam and in front of a 40 micron surface barrier detector registering the delayed @xmath14 s following the @xmath10 decay of @xmath17 .
the detector was surrounded by a shroud to prevent scattered beam particles from reaching the detector .
the time sequence of the whole cycle is : a.- 1.5 s beam - on - target ; b.- 100 ms rotation ; c.- 1.5 s of target in the counting position ; d.- 100 ms rotation . in the counting position a gate signal from the control unit opens the adc for @xmath14 counting and the gated scaler for faraday - cup beam monitoring . this sequence results in an efficiency factor for the @xmath14 count ( see below ) of @xmath18(cycle ) = 0.400(1 ) ( fig .
a liquid nitrogen cold cryofinger is placed close to the target area to protect the target surface from contamination .
the vacuum in the chamber was @xmath19 8@xmath2010@xmath21 torr .
the beam density @xmath8 was measured by collimating the beam by a series of holes of known areas downstream from the target position , integrating the collimated beam in an electron - suppressed faraday cup and counting the digitized counts in a gated scaler .
the current digitizer and the scaler were checked during the experiment with a a calibrated current source .
the beam homogeneity was virtually insured by the nature of the raster operation : a low frequency triangular @xmath22 scan and a high frequency triangular @xmath23 scan , in small , digitally controlled steps in clock - fixed time intervals .
the beam homogeneity was checked in two direct ways : 1.- by measuring the areal density of x - rays from a tin foil induced by the scanned proton beam in a phosphor image plate ( fig . 2a ,
5 ) , 2 ) by repeating the measurement with different downstream collimators of know apertures . for a homogeneous beam
the faraday - cup counts of the integrated beam , normalized to the @xmath14 yield , should be proportional to the area of the hole ( fig .
the collimator hole areas were measured to an accuracy better than 1@xmath24 by a microscope and by having an alpha source in front of the collimator - detector assembly .
the number @xmath25 of @xmath0li nuclei in the 2 mm target was determined in a similar measurement using the reaction @xmath0li(p , n)@xmath0be with a proton beam of 1.985 mev ( refs . 6 and 7 ) .
the same downstream beam collimator of 1.5 mm diameter was used in both measurements .
the number of @xmath0be s produced was measured by registering @xmath5 s from the beta decay @xmath0be@xmath26li@xmath27 ( 478 kev ) in a ge detector at the low - background counting laboratory of the nrc - soreq center .
we have also taken care to adjust the plate - voltage to ensure the same area for both the p and d@xmath15 beams , respectively .
for the ratio of the two yields we get from ( 1 ) : @xmath28 where @xmath29 refer , respectively , to the total number of deuterons and protons registered .
the number of @xmath3li products , y(@xmath3li ) , is evaluated from the measured @xmath14 counts : @xmath30 and y(@xmath0be ) is evaluated from the measured @xmath5 counts of the radioactive decay of @xmath0be : @xmath31 the efficiency factors @xmath32 and @xmath33 above are each products of individual and independent efficiency factors detailed below . inserting the expressions for n@xmath34 and n@xmath35 into ( 4 ) we obtain the @xmath0li(d , p)@xmath3li cross section in terms of the @xmath0li(p , n)@xmath0be cross section and measured quantities : @xmath36
1 . _ the solid angle_. + the detector solid angle was defined by a collimator of diameter d=8.14(3 ) mm which was placed at a distance @xmath37 = 81.92(10 ) mm from the target in the counting position .
the solid angle is given by : @xmath38 = 6.16(6 ) 10@xmath39 of @xmath40 .
2 . _ the time - sequence efficiency - @xmath18(cycle ) _ + this factor relates to the time fraction of the @xmath14 counting out of the entire irradiation - counting cycle . from the description above and following refs . 2,3 , this factor
is calculated to be : @xmath18(cycle ) = 0.400(1 ) .
the time intervals of the rotation cycle were measured by counting a time - reference signal in the scaler gated by the cycle time windows ( fig .
3 . _ the spectral efficiency . _ + the @xmath14 spectrum in fig .
was integrated from the point marked in the figure all the way to the top energy .
the cutoff entailed a loss of low energy @xmath14 s on one hand and on the other - the inclusion of the high energy tail of the noise spectrum .
these two quantities were estimated by rough extrapolations to be considerably less than 1@xmath24 and approximately equal and hence the integration error appears to be negligible .
the low energy tail of the @xmath14 spectrum was also evaluated directly from the measured @xmath14 spectrum in ref .
8 folded with the energy loss of the @xmath14 s in the target , yielding a lost fraction of 0.2@xmath24 .
1 . _ the @xmath5 branch . _
+ the @xmath5 branch of the decay of @xmath0be to the first excited state of @xmath0li at 478 kev is 0.1052(6 ) [ 9 ] .
_ absolute calibration of the @xmath0be activity . _ + the activity of the @xmath0be 478 kev @xmath5-ray source produced in the ( p , n ) reaction on the various targets was determined by comparison with calibrated @xmath41na , @xmath42cs and @xmath43ba @xmath5 sources at a fixed distance from a ge detector , shielded for low - background , at the @xmath5-counting laboratory of the nrc - soreq center .
the absolute intensities of the @xmath5 sources used are known to within 2.5 - 3@xmath24 each .
the number of 478 kev @xmath5 rays ( and hence the number of @xmath0li nuclei in the target ) is thus determined to an accuracy of 3@xmath24 .
3 . _ the @xmath10 decay of @xmath0be . _
the lif targets were bombarded by protons for several hours and @xmath5 s were counted for about a day a short time later .
the number of @xmath0be s was inferred from the number of @xmath5 counts through the relation : @xmath44 where @xmath45 is the @xmath5 counting time , @xmath46 is the time from the middle of the @xmath47 run to the middle of the counting time , @xmath48 is the @xmath0be mean life : @xmath49 d [ 9 ] and @xmath50 and @xmath51 are the efficiency factors referred to above . 1 . _ current integration . _
+ the quantity @xmath52 is the ratio of scaler counts from a current integrator monitoring the current in the faraday cup for the d@xmath15 and p beams , respectively , using the same collimator hole . the construction of the faraday cup , including the suppressor , ensures a reliable monitoring of the current .
a calibrated current source was used to check the accuracy of the current integrator . in any case , as only the @xmath53 of the counts enters the cross - section determination , any possible error from this source cancels out in first order .
2 . _ counting statistics - @xmath54 . _
+ the statistical errors of the @xmath14 and @xmath5 counts were typically 0.5@xmath24 and 0.7@xmath24 , respectively for all the targets used .
there is in addition a 1@xmath24 systematic error in the @xmath5 counting due to uncertainties in the background subtraction procedure .
3 . _ the @xmath0li(p , n)@xmath0be cross section . _
+ the cross section for this reaction exhibits a flat and constant region around the beam energy of e@xmath55 = 1.985 mev employed in this experiment .
the cross section at this energy was taken from refs .
6,7 , which are in very good agreement with each other .
each of these references quotes , separately , an absolute error of 5@xmath24 .
we therefore adopt the value of @xmath2li(p , n)@xmath0be ] = 269(9 ) mb at e@xmath55 = 1.985 mev .
+ table i presents a summary of the experimental errors discussed above . + another source of error which was considered is the production of @xmath0be by @xmath56li(d , n)@xmath0be , adding on to the @xmath0be produced by @xmath0li(p , n)@xmath0be .
the added fraction was evaluated from the known cross - sections [ 10 ] and the fraction of @xmath56li in the target to be less than 0.2@xmath24 and therefore insignificant . + a larger correction to
the extracted cross section can arise from the effect of @xmath3li lost from the target due to backscattering .
this issue is discussed in iii .
a d@xmath15 beam at the 776 kev resonance corresponds to an accelerator voltage of 1.552 mv .
the voltage was calibrated using the known thresholds of the @xmath57c(p , @xmath5 = 9.17 mev)@xmath58n reaction at 1.746 mev and the @xmath0li(p , n)@xmath0be reaction at 1.88 mev which are close to the d@xmath15 and p beam voltages .
in the course of this experiment it became apparent that under some conditions the backscattering of recoil nuclei out of the target can be significant and can affect the cross section measurements . in order to obtain a quantitative estimate of these effects , the propagation of the reaction products of the @xmath0li(d , p)@xmath3li and @xmath0be(p,@xmath5)@xmath3b reactions was simulated with the aid of the trim [ 11 ] code .
we have also determined these effects experimentally for the ( d , p ) reaction . for the @xmath0li(d , p)@xmath3li reaction one has to take into account the broad angular distribution of the protons [ 12 ] ( and
the @xmath3li recoils ) and also to evaluate the loss of the backscattered @xmath0be s from @xmath0li(p , n)@xmath0be reaction , used as a standard ( see section ii ) .
one can then infer the total effect of backscattering loss on the cross - section measurement .
calculations were carried out for lif targets on pt , ni and al backings at 776 kev deuteron energy .
the results presented here are an average of simulations of reactions occuring at three positions in the lif target . at a given depth in the lif target
the energy and angular distribution of @xmath3li ions were evaluated and used to obtain the angle and energy distributions of the backscattered @xmath3li ions from trim .
the backscattered nuclei were then traced in the direction opposite to the beam to obtain the number of ions leaving the target .
a similar procedure was carried out for the backscattered @xmath0be from the @xmath0li(p , n)@xmath0be reaction for the proton beam energy of 1.99 mev . the angle and energy distributions of the neutrons ( and the @xmath0be recoils )
were taken from ref 13 .
the results of the trim simulation are shown in table ii .
as we see from the table , the effect of backscattered particles can give rise to a sizeable correction for high z substrates .
similar backscattering effects also exist for the @xmath0be(p,@xmath5)@xmath3b reaction . in this case ,
the kinematics of the reaction is much simplified as the @xmath3b nuclei from the proton - capture reaction all proceed in the proton - beam direction with the same center - of - mass momentum . as an example
we have performed calculations which duplicate the conditions of ref .
2 . in that experiment
the target was electroplated on polished pt disks and contained 7 @xmath59 g of solid material with 11@xmath24 of @xmath0beo [ 14 ] . for low proton energy ,
the number of backscattered particles depends very strongly on the ( unknown ) target composition .
two hypothetical trial cases are shown in fig .
the first is for @xmath0beo dissolved in pure iron ; the second case is for a hypothetical composition of equal number of cuo , cu@xmath15n@xmath60 and c molecules .
the latter composition has been chosen following the statement in ref .
14 that the solid materials remaining after the plating and flaming procedure are most likely to be traces of light metal oxides , nitrates and carbon .
it is obvious from fig .
4 that heavy atoms in the target and target backing should be avoided as much as possible in precision measurements of the @xmath0b(p,@xmath5)@xmath3b cross section .
we have used several lif targets with al and pt backing materials in order to probe experimentally the backscattering issue .
the results of the cross section measurements for five targets , together with the backscattering fraction as calculated by trim , are presented in table iii .
the error of each measurement is about 5@xmath24 ( table i ) .
the experiments yield a 9.1(2.2 ) @xmath24 difference between the 400 @xmath61 lif on pt and the lif on al targets .
the relative error for the two types of targets is smaller than the individual errors since it does not include the common uncertainties of the @xmath0li(p , n)@xmath0be cross section and the @xmath5-detector efficiency .
the trim simulations are in good agreement with the experimental results .
.5cmas the final result of the @xmath0li(d , p)@xmath3li cross section we have adopted an average of the two aluminum - backed targets , believed to be free from the backscattering loss .
we get : @xmath62[@xmath0li(d , p)@xmath3li ] = 155(8 ) mbarn . + a recommended value for this cross section
is given in ref . 1 : @xmath62 = 147(11 ) mbarn .
this value is based on four measurements , two of which should be corrected for backscattering loss . based on the available target details as presented in the respective publications , we estimate the corrected value to be in the range 150 - 152 mbarn , very close to our value . previous measurements of the @xmath63be((p,@xmath1b cross section are also susceptible to backscattering losses . in particular , the results of ref .
2 should be corrected for both the @xmath0be(p,@xmath5)@xmath3b and the @xmath0li(d , p)@xmath3li reactions .
the @xmath0be(p,@xmath5)@xmath3b reaction was measured at various energies and , as can be seen from fig .
4 , the backscattering correction diminishes at low energies .
the @xmath0li(d , p)@xmath3li reaction at the resonance energy was then used by the authors to obtain the number of @xmath0li , and hence @xmath0be , atoms in the same platinum - backed target . since the correction due to this is a @xmath64 which is of the opposite sign and larger than the @xmath3b corrections ( table ii and fig .
4 ) , the overall effect is to reduce the derived s factor , especially at lower energies .
we would like to express our sincere thanks to ygal shahar for his expert help in operating the van de graaff accelerator and in constructing the beam line .
we wish to thank meir birk and moshe sidi for designing and constructing the raster scanner , and leo sapir and boris levine for the target preparation .
we acknowledge with thanks the expert help of ovadya even of the soreq center in the @xmath5-efficiency calibration .
we are grateful to prof .
claus rolfs for extensive and enlightening discussions .
filippone , a.j .
elwyn , c.n .
davids and d.d .
koetke , phys .
@xmath65 , 2222 ( 1983 ) .
filippone , a.j .
elwyn and w. ray , jr . , phys .
@xmath66 , 2174 ( 1982 ) .
f. strieder , l. gialanella , u. greife , c. rolfs , s. schmidt , w.h .
schulte , h.p .
trautvetter , d. zahamow , f. terrasi , l. campajolla , a.donorfrio , v. roca , m. romano and m. romoli , z. phys . * a 355 * , 209 ( 1996 ) . l. weissman , m. hass and v. popov , nucl . inst .
@xmath67 , 409 ( 1997 ) .
sekhran , h. laumer , b.d .
kern and f. gabbard , nucl .
@xmath68 , 253 ( 1976 ) .
r.l . macklin and j.h .
gibbons , phys .
@xmath69 , 571 ( 1959 ) .
warburton , phys .
@xmath70 , 303 ( 1986 ) .
table of isotopes , 8@xmath71 edition , j. wiley and sons , new york ( 1996 ) . c. r. mcclenahan and r. e. segel , phys .
@xmath72 , 370(1986 h.h .
andersen and j.f .
ziegler , _ the stopping and ranges of ions in matter _
( pergamon , new york , 1977 , vol .
sellschop , phys . rev . * 119 * , 251 ( 1960 ) .
h. liskien and a. paulsen , atomic and nuclear data sheets , * 15 * , 57 ( 1984 ) .
b. w. filippone and m. wahlgren , nim , * a243 * ( 1986 ) 41 - 44 . | a novel scheme for measuring the cross section of the @xmath0be(p,@xmath1b reaction , the major source of high energy neutrinos from the sun , is presented . the scheme involves a strictly uniform particle beam and overcomes some of the recognized experimental uncertainties of previous measurements .
a new measurement of @xmath2li(d , p)@xmath3li ] has been carried out using this setup , and the present value of @xmath2li(d , p)@xmath3li ] = 155(8 ) mbarn at the top of the e@xmath4(lab . )
= 776 kev resonance is compared to previous measurements .
a new issue regarding both the ( d , p ) and ( p,@xmath5 ) reactions has been examined : reaction - product nuclei which are backscattered out of the target .
measurements and simulations carried out in the course of this investigation are presented and discussed in the context of possible effects on the measured cross sections of these reactions .
pacs numbers : 95.30.cq ; 96.60.kx ; 25.45.hi |
the developments in spintronics and nanomagnetics are having enormous influence on the field of storage and memory devices and it has been shown that the write ( w ) and read ( r ) elements can also be integrated into units that implement boolean as well as non - boolean logic @xcite .
these applications , however , usually make use of stable magnets with energy barriers @xmath3 which require relatively large currents for their operation .
the critical spin current needed to switch a magnet with a thermal energy barrier of @xmath4 is given by @xcite @xmath5 @xmath6 where @xmath7 is the electronic charge , @xmath8 is the saturation magnetization , @xmath9 is the anisotropy field , @xmath10 is the demagnetization field , v is the volume , @xmath11 is the gilbert damping coefficient and the factor @xmath12 is equal to zero for perpendicular anisotropy magnets ( pma ) and one for inplane anisotropy magnets ( i m a ) . with @xmath13 and @xmath14 ,
the critical switching spin current for a pma magnet is @xmath15 magnets with lower barriers could operate with lower currents but their application in conventional memory or logic is severely limited due to their stochastic nature .
however , their possible use in unconventional applications has been discussed both theoretically and experimentally @xcite .
the implementation of logic operations based on an ensemble average over stable nanomagnets has been explored in @xcite while @xcite describes an approach to the traveling salesman problem based on a time average over unstable nanomagnets that cycle through millions of collective correlated states potentially at ghz rates .
the present paper describes the application of the latter approach to implement precise boolean logic operations like a 32-bit adder that provides the sum s for given inputs a and b. remarkably the adder also evaluates the inverse function , cycling through all combinations of a and b that add up to a given sum s. ) as a function of the bias spin current which is normalized to @xmath16 ( eq . [ ic0 ] ) .
average magnetization of pma magnets obtained from sllg which agrees well with the analytical solution from the fpe , eq .
inset shows a physical structure using a giant spin hall effect ( gshe ) material that could be used to convert a charge current into a spin current with the correct polarization to bias an i m a .
( b ) the magnetization m(t ) for i m a as a function of time for three different bias currents obtained from a numerical solution of sllg equation .
( c ) same plot for pma with the same barrier height .
note that the fluctuations are much faster and more telegraphic for i m a than for pma .
( d ) a connection scheme for two p - bits is shown where the magnetization of a p - bit is implicitly converted into the bias current / voltage for the next p - bit ( eq .
[ ik ] ) . ]
we have recently shown @xcite that a suitably interconnected network of tunable telegraphic noise generators or `` p - bits '' can be used to perform even precise arithmetic functions like a 32-bit adder . however , it is not clear whether such p - bits can be implemented with real physical systems , especially if the noise in these systems are not telegraphic but continuous .
the objective of this paper is to use simulations based on the stochastic landau- lifshitz - gilbert ( sllg ) equation to demonstrate that p - bits can be implemented using unstable nanomagnets with energy barriers as low as a fraction of a kt , _ even though their magnetization is not telegraphic _ and fluctuate randomly among all values from @xmath21 to + 1.we assume that the magnets are read with a thresholding device that translates all positive values to 1 and all negative values to zero . but
this thresholding is applied _ only _ when we need to read a magnet at the end of an operation . .
( b ) the decimal value of each configuration of the input - output nodes at each time step ( normalized by the factor @xmath17 ) is calculated according to @xmath18 where a , b and c are thresholded to obtain binary values ( 0,1 ) at the read out .
( c ) histograms of the different configurations of the p - bits are shown for a weaker @xmath19 and stronger @xmath20 correlation strength .
note the close match between the numerical values obtained from the sllg equation with the probabilities obtained analytically from the fpe result in eq .
[ ps ] which is related to the boltzmann law , especially for @xmath21=0.5 . for higher values of @xmath21
the numerical results tend to be stuck in metastable states requiring longer simulation times to converge to the steady - state fpe result . ] * _ outline of paper : _ * we start in _ section 2 _ by showing that low barrier magnets , both pma and i m a , exhibit the key property of p - bits , namely that they act as electrically tunable random number generators ( rng ) .
their magnetization @xmath22 fluctuates randomly in time , and the time - averaged @xmath23 can be tuned from @xmath21 to + 1 with a spin current .
it is well known that stable pma magnets ( @xmath24 , eq .
[ ic0 ] ) have a lower critical switching current compared to i m a ( @xmath25 ) .
interestingly for both pma and i m a magnets , the current needed to tune the average magnetization from say @xmath20.8 to + 0.8 is approximately the corresponding current needed to switch a magnet with a 1 kt barrier .
this means that i m a magnets require a larger current to tune , but this is offset by a more rapid fluctuation rate , allowing a faster evaluation of the time average , and hence faster operation ( fig .
[ fi : fig1 ] ) .
note also that the pma magnetization is relatively continuous compared to i m a magnetization which is more telegraphic in nature . both pma and i
m a stochastic nanomagnets provide a random magnetization whose average depends on the input current . to harness them for logic applications
, they have to be interconnected such that the spin current @xmath26 driving magnet ` k ' has to be derived from the magnetization of other magnets .
@xmath27 where @xmath16 is the critical current ( eq . [ ic0 ] ) for a magnet with a barrier @xmath28 and
@xmath21 determines the overall strength of the interconnections . the bias @xmath29 and interconnection @xmath30 $ ] matrices have to be designed appropriately in implementing specific operations .
we will not go into the implementation of these matrices since there are many options requiring careful discussion @xcite,@xcite,@xcite,@xcite .
we will assume that a network of stochastic nanomagnets ( pma and i m a ) has been interconnected according to eq .
[ ik ] and simulate their behavior using the stochastic landau - lifshitz - gilbert ( sllg ) equation to demonstrate useful functionalities .
all numerical examples are presented for i m a but similar results are obtained with pma as well .
are clamped , the adder gives the correct output @xmath31 .
( c ) unlike standard logic , these gates _ are invertible _ :
if the output nodes of the adder are clamped to fixed values , the adder gives all possible input combinations satisfying the output constraint . ] in _ section 3 _ we describe how simple logic gates can be implemented by suitably designing the @xmath29 and @xmath30 $ ] matrices so that the magnet configurations corresponding to the desired truth table represent ` low energy ' states where the network spends most of its time according to the boltzmann law of equilibrium probabilities @xmath32 although the use of spin currents does not in general permit us to write an energy functional @xcite , for symmetrically interconnected pma magnets we can use a functional of the form @xcite
@xmath33 to describe the network of interconnected magnets .
this can be seen by noting that from eqs .
[ p ] and [ e ] @xmath34 so that for a symmetric @xmath30 $ ] matrix , from eq .
[ ik ] @xmath35 which is exactly the steady - state condition for magnet ` k ' that we would obtain from the fokker - planck equation ( fpe ) ( @xcite eq .
( 3.9 ) ) for pma
. we will refer to eqs .
[ p ] and [ e ] as the fpe probability . ) with the carry out bit @xmath36 from each adder communicated in a directed fashion to the carry in bit @xmath37 of the next adder .
( b ) time evolution of the sum @xmath38 obtained from the sum bits \{s } as the coupling strength @xmath21 is ramped up starting from zero .
note that in a time @xmath39 ( @xmath40 is defined in fig .
2 ) , the sum converges ( with occasional jumps ) to the correct value which represents one out of @xmath41 billion possibilities .
( c ) although the individual adders are connected in a directed fashion through the carry bits , the overall 32-bit adder performs the inverse function as well .
if the sum bits \{s } are clamped along with one set of input ( b ) , the other input converges rapidly to the correct difference ( a ) . ]
the probability distributions obtained from the numerical solution of the sllg equation for both pma and i m a magnets follow the fpe result quite well ( fig . 2 ) .
the highest probabilities correspond to the lowest energy states , which correspond to the desired truth table relating the input magnets a and b to the output magnet c. if we force the inputs a and b to specific values by using appropriate values for @xmath42 and @xmath43 , c would take on the specific value required by the truth table , just like standard digital gates .
but unlike standard gates , these gates are invertible , similar to those discussed in the context of memcomputing @xcite .
they can be operated in reverse : if we clamp the output c to a specific value , the inputs a and b will spend most of its time in those configurations @xmath44 that produce that output .
we also illustrate this reversible operation with a more complex logic gate , namely a full adder treating it as a boltzmann machine ( bm ) and using the same principle of energy minimization to design the @xmath29 and @xmath30 $ ] matrices . finally in _ section 5 _
we demonstrate the operation of a 32-bit adder obtained from 32 full adders with the output carry from each bit connected to the carry in of the next higher bit through the appropriate element of the overall @xmath30 $ ] matrix .
note that these are unidirectional connections so that the overall @xmath30 $ ] matrix is not symmetric , though the @xmath30 $ ] matrix for each full adder is symmetric .
we show that this network of nearly five hundred nanomagnets exhibits a remarkably precise correlation that provides the exact sum s of any two given inputs , a and b ( fig.4 ) .
what is even more remarkable is that if we do not fix either the inputs or the outputs , the quantities a , b and s all fluctuate randomly and yet the quantity a+b@xmath2s is sharply peaked around zero , so that the network can be used to extract either a , b or s , if the other two are fixed , which is similar to the np - complete `` subset sum '' problem ( fig.5 ) @xcite . ) is that it works even when the inputs ( a , b ) and the output ( s ) are not unique and fluctuate in time amongst many allowed values as shown in ( a ) .
nevertheless , the quantity a+b - s is sharply peaked at zero ( b ) , demonstrating the correlation of hundreds of nanomagnets consistent with the addition function a+b@xmath2s=0 . ]
fig.1(b , c ) shows the time response of the magnetization @xmath45 along the easy axis calculated using the sllg equation with @xmath46 for i m a and @xmath47 for pma . @xmath48 where @xmath49 is the effective field including the uniaxial and shape anisotropy terms , as well as the thermally fluctuating magnetic field due to three dimensional uncorrelated thermal noise @xmath50 having gaussian distribution with mean @xmath51 and standard deviation @xmath52 along each direction
@xcite , @xmath53 is the gyromagnetic ratio and @xmath54 is the total number of bohr magnetons comprising the magnet . the time - averaged magnetization ( fig .
1a ) obtained from the sllg equation for pma magnets is in good agreement with that obtained analytically by averaging over the fpe result ( eq . [ pmk ] ) : @xmath55
in implementing any given truth table we need the @xmath29 and @xmath30 $ ] matrices that make the truth table correspond to the lowest energy states of the energy functional given in eq .
the choice of these matrices is not unique and @xcite provides a suitable set for and , or gates along with many other functions .
2a shows one possible implementation of an and gate using a network of three nanomagnets , representing a , b and c. the magnetization of the magnets a , b and c fluctuates continuously between @xmath21 and + 1 and are mapped into the binary values of 0 and 1 by a thresholding operation : all negative values map to zero , while positive values map to + 1 . the y - axis in fig .
2b shows the resulting binary number @xmath56 converted into a single number @xmath57 .
note how the values on the y - axis are clustered around 0 , 2 , 4 and 7 which correspond to the lines of the truth table shown in fig .
occasionally the system jumps to other values but it quickly returns to one of these preferred values .
this clustering is reflected in the histogram constructed from 678 normalized time steps ( fig .
2c ) which shows peaks around the preferred states defined by the truth table .
this agrees well with the probability plot constructed from the fpe result in eq .
[ p ] noting that we can label the thresholded states with a binary variable s , where @xmath58 where @xmath59 and @xmath60 so that from eq .
[ e ] @xmath61 @xmath62 note that the probabilities are strongly affected by the choice of @xmath21 as we might expect from the exponential dependence of the boltzmann function .
if we use a much smaller value of @xmath21 we obtain a uniform probability across all eight states as we would expect for three uncorrelated magnets .
if we use a much larger value of @xmath21 the boltzmann law predicts all states with equal energy to be equally occupied , but in a numerical simulation , the system tends to get stuck for long periods in one of the preferred states , instead of moving freely among them .
consider now a full adder having three inputs @xmath63 and two outputs @xmath64 , @xmath65 being the sum bit , and @xmath66 being the incoming carry and the outgoing carry bits .
3 shows a full adder constructed out of 14 p - bits treating it as a bm with a symmetric j - matrix which is obtained by a suitable extension of the principles developed in the context of hopfield networks ( @xcite , eq . 4.20 ) and extended in @xcite .
this design not only gives the correct output for a given input , _ but also the correct set of inputs for a given output_.
finally we demonstrate the operation of a 32-bit adder obtained from 32 full adders with a single directed connection from the @xmath67 of one bit to the @xmath37 of the next bit , in accordance with the standard design of ripple carry adders ( rca ) .
if we provide two input numbers a and b , and look at the sum s , which includes all the sum bits along with the carry - out from the last bit , @xmath68 we find numerically that the histogram is sharply peaked around the correct sum .
it is really quite surprising that a network of @xmath69 nanomagnets fluctuating continuously over the range @xmath70 get correlated precisely enough to point to the correct answer out of @xmath71 possibilities .
interestingly it also works as a subtractor : if we fix the sum and one of the inputs b , the remaining input shows a clear peak at s@xmath2b ( fig . 4 ) .
even more surprisingly , the overall system seems to act like a bm when all magnets are allowed to fluctuate .
each set of magnets a , b and s fluctuates randomly over a wide range of values .
but the quantity a+b@xmath2s shows a sharp peak around zero ( fig . 5 ) , showing that the interconnected network reflects the desired truth table .
the authors gratefully acknowledge many helpful discussions with behtash behin - aein ( globalfoundries ) , vinh quang diep ( sandisk ) , and with ernesto e. marinero ( purdue university ) . this work was supported in part by c
- spin , one of six centers of starnet , a semiconductor research corporation program , sponsored by marco and darpa , in part by the nanoelectronics research initiative through the institute for nanoelectronics discovery and exploration ( index ) center , and in part by the national science foundation through the ncn - needs program , contract 1227020-eec .
37ifxundefined [ 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 in @noop _ _ , vol .
( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) in @noop _ _ ( ) pp . @noop * * ( ) @noop * * ( ) @noop * * , ( ) @noop * * , ( ) in @noop _ _ ( , ) pp . @noop * * , ( ) in @noop _ _ ( , ) pp .
@noop ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) _ _ , @noop ph.d . thesis , ( ) @noop * * , ( ) @noop _ _ ( , ) | it has recently been shown that a suitably interconnected network of tunable telegraphic noise generators or `` p - bits '' can be used to perform even precise arithmetic functions like a 32-bit adder . in this paper
we use simulations based on the stochastic landau - lifshitz - gilbert ( sllg ) equation to demonstrate that similar impressive functions can be performed using unstable nanomagnets with energy barriers as low as a fraction of a @xmath0 .
this is surprising since the magnetization of low barrier nanomagnets is not telegraphic with discrete values of @xmath1 .
rather it fluctuates randomly among all values between @xmath21 and + 1 , and the output magnets are read with a thresholding device that translates all positive values to 1 and all negative values to zero .
we present sllg - based simulations demonstrating the operation of a 32-bit adder with a network of several hundred nanomagnets , exhibiting a remarkably precise correlation : the input magnets \{a } and \{b } as well as the output magnets \{s } all fluctuate randomly and yet the quantity a+b@xmath2s is sharply peaked around zero !
if we fix \{a } and \{b } , the sum magnets \{s } rapidly converge to a unique state with s = a+b so that the system acts as an adder .
but unlike standard adders , the operation is invertible . if we fix \{s } and \{b } , the remaining magnets \{a } converge to the difference a = s@xmath2b .
these examples suggest a new direction for the field of nanomagnetics away from stable high barrier magnets towards stochastic low barrier magnets which not only operate with lower currents , but are also more promising for continued downscaling . |
the governing system of equations for the compressible nematic liquid ( nlc ) crystal flows is the following system of scalar or vector fields @xmath2 , @xmath3 and @xmath4 for @xmath5 , for a bounded smooth domain @xmath6 : @xmath7 together with the initial value conditions : @xmath8 and the boundary value conditions : @xmath9 here we denote by @xmath10 , @xmath11 , @xmath12 the unknown density , velocity and orientation parameter of liquid crystal , respectively , and @xmath13 is the pressure function . besides , @xmath14 and @xmath15 are positive viscosity coefficients .
the non - standard term @xmath16 denotes the @xmath17 matrix , whose @xmath18-th element is given by @xmath19 .
@xmath20 is the unit matrix .
@xmath21 is a polynomial function of @xmath22 which satisfies @xmath23 , where @xmath24 is the bulk part of the elastic energy ; usually we choose @xmath24 to be the ginzburg
landan penalization , i.e. , @xmath25 and @xmath26 , where @xmath27 is a positive constant . in
what follows , we will assume @xmath28 since its specific value does not play a special role in our discussion .
besides , we assume that the pressure function @xmath29 satisfies @xmath30 the above system is a simplified version of ericksen lesile system modeling the flow of compressible nematic liquid crystals , and the hydrodynamic theory of liquid crystals was established by ericksen @xcite and leslie @xcite in the 1960 s .
when @xmath31 , the system becomes to the compressible navier stokes ( cns ) equations .
matsumura and nishida @xcite obtained global existence of smooth solutions for the initial data is a small perturbation of a non vacuum equilibrium . for the existence of solutions for arbitrary initial value , lions @xcite and feireisl @xcite established the global existence of weak solution to the cns equations .
cho et al.@xcite proved that the existence and uniqueness of local strong solutions of the cns equations in the case where initial density need not to be positive and may vanish in an open set .
xin in @xcite showed that there is no global smooth solution to the cauchy problem of the cns equations with a nontrivial compactly supported initial density .
hence , there are many works @xcite try to establish blow up criterion for the strong solution to the cns equations . in particular , it is proved in @xcite by huang , li and xin that the serrin s blow up criterion
( see @xcite ) for the incompressible navier
stokes equations still holds for the cns equations , i.e. , if @xmath32 is the maximal time of existence strong solution , then @xmath33 or @xmath34 where @xmath35 and @xmath36 satisfy @xmath37 , @xmath38 . in @xcite , huang et al .
established that the beale
majda criterion ( see @xcite ) for the ideal incompressible flows still hold for the cns equations , that is @xmath39 sun , wang and zhang in @xcite ( see also @xcite ) obtained another beale kato
majda criterion in terms of the density , i.e. , @xmath40 when @xmath10 is a positive constant , the system becomes to the incompressible nematic liquid crystal ( inlc ) equations , the global - in - time weak solutions and local - in - time strong solution have been studied by lin and liu @xcite . in @xcite , hu and wang established global existence of strong solutions and weak strong uniqueness for initial data belonging to the besov spaces of positive order under some smallness assumptions .
liu and cui in @xcite obtained that the blow up criterion or still holds for the solution of the inlc equations .
we also refer @xcite and the reference cited therein for other related work on the inlc equations .
inspired by the above mentioned works on blow up criterion of strong solution of cns and inlc equations , particularly the results of huang etal .
@xcite and sun et al .
@xcite , we want to investigate a similar problem for the compressible nematic liquid crystal flow . before stating the main result , we denote the following simplified notations of sobolev spaces @xmath41 when the initial vacuum is allowed , the well - posedness and blow up criterion for strong solutions to the compressible nematic liquid crystal flows have been investigated by liu et al . in @xcite . here
, we write down the main results of liu et al .
@xcite .
[ thm1.1 ] suppose that the initial value @xmath42 satisfies the following regularity conditions @xmath43 and the compatibility condition @xmath44 then there exist a small @xmath45 and a unique strong solution @xmath46 to the system with initial boundary condition such that @xmath47 moreover , let @xmath32 be the maximal existence time of the solution . if @xmath48 , then there holds @xmath49 where @xmath50 satisfying @xmath51 and @xmath52 .
* remark 1.1 * another similar system of partial differential equations modeling compressible nematic liquid crystal flows has been studied by huang , wang and wen in @xcite .
they obtained the existence of local in time strong solution and two blow up criteria under some suitable assumption condition @xmath1 and @xmath22 or @xmath10 and @xmath22 .
the purpose of this paper is to obtain the beale
majda blow up criterion only in terms of the gradient of the velocity still holds for the liquid crystal flows .
our main result is the following [ thm1.2 ] assume that @xmath53 is the strong solution constructed in theorem [ thm1.1 ] , and @xmath32 be the maximal existence time of the solution .
if @xmath48 , then we have @xmath54 the proof of this theorem will be given in the next section . as a standard practice
, we will show that if ( [ eq1.9 ] ) does not hold then the strong solution @xmath46 can be extended beyond the time @xmath55 . to this end
we will step - by - step establish a series of higher - order norm estimates for the strong solution @xmath46 . the key fact used in this deduction
is that the boundedness of the @xmath56-norm of @xmath57 implies both boundedness of the @xmath58-norm of the density @xmath10 and boundedeness of the @xmath59-norm of @xmath22 with @xmath60 .
let @xmath53 be the unique strong solution to the system with initial
boundary condition . we assume that the opposite to holds , i.e. , @xmath61 in what follows , we note that @xmath62 denotes a generic constant depending only on @xmath63 and the initial data . by using the mass conservation equation @xmath64 and the assumption ,
it is easy to obtain the @xmath65-norm bounds of the density , multiplying the mass conservation equation @xmath64 by @xmath68 with @xmath69 , it follows that @xmath70 integrating the above equality over @xmath71 yields @xmath72 i.e. , @xmath73 the condition and the estimate imply that @xmath74 where @xmath62 is a positive constant independent of @xmath75 , letting @xmath76 , we obtain , and this completes the proof of the lemma . multiplying the momentum equation @xmath79 by @xmath1 , integrating over @xmath71 and making use of the mass conversation equation @xmath64 , it follows that @xmath80 where we have used the fact that @xmath81 . multiplying the liquid crystal equation @xmath82 by @xmath83 and
integrating over @xmath71 , we obtain @xmath84 combining and together @xmath85\text{d}x+\mu\int_{\omega } |\nabla u|^{2}\text{d}x+\lambda\nu \int_{\omega } |\delta d - f(d)|^{2}\text{d}x\nonumber\\ = & -\int_{\omega } u\nabla p\text{d}x=\int_{\omega}p\operatorname{div}u\text{d}x\leq \varepsilon\int_{\omega } |\nabla u|^{2}\text{d}x+c \varepsilon^{-1},\end{aligned}\ ] ] where we have used the estimates , and the young inequality .
taking @xmath86 small enough and applying the gronwall s inequality , we can establish the estimate immediately .
we first multiplying the liquid crystal equation @xmath82 by @xmath89 with @xmath90 , and integrating over @xmath71 , then there holds @xmath91 by using the gronwall s inequality , one obtains the inequality @xmath92 by letting @xmath76 , we notice that the estimate still holds . multiplying the gradient of the liquid crystal equation @xmath82 by @xmath93 with @xmath90 , and integrating over @xmath71 , then there holds @xmath94 where we have used the fact that @xmath95 letting @xmath76 , estimate still holds , and the inequalities and imply that estimate holds . to prove the estimate , we multiplying the liquid crystal equation @xmath82 by @xmath96 and integrating over @xmath71 , then @xmath97 where we have used the estimates and . integrating the above inequality over
@xmath98 $ ] gives the estimate .
noticing that the momentum equation @xmath79 can be rewrote as @xmath103 multiplying the equation above by @xmath104 and integrating over @xmath71 , one obtains the equality @xmath105 combining the mass conservation equation @xmath64 and the assumption , it follows that the pressure @xmath29 satisfies the following equation @xmath106 hence , we have @xmath107 to estimate the term @xmath108 , we have @xmath109 where we have used estimate in the last inequality . inserting and into , and integrating over @xmath98 $ ]
give that @xmath110 where we have used the estimate and . to estimate the terms on the right side of , by using lemma [ lem2.1 ] , the estimates , and , we get @xmath111 by using the stand elliptic regularity result to
, we have @xmath112 combining estimates and taking @xmath86 small enough , we can get @xmath113 to estimate the orientation parameter @xmath22 , by the standard elliptic regularity result to the liquid crystal equation @xmath82 , one obtains that @xmath114 multiplying the liquid crystal equation @xmath82 by @xmath115 , and integrating over @xmath71 , then we have @xmath116 where we have used the hlder inequality and estimates . for the term @xmath117 ,
applying estimate , we have for @xmath118 @xmath119 hence , taking @xmath120 small enough @xmath121 inserting into , taking @xmath86 , @xmath120 small enough and integrating above inequality over @xmath122 $ ] ensure that @xmath123 now , we will estimate the density @xmath10 . applying the operator @xmath124 to the mass conservation equation @xmath64 , then multiplying it by @xmath125 and integrating over @xmath71 yield @xmath126 where we have used the estimate in the above inequality . integrating the above estimate over @xmath127 $ ]
gives that @xmath128 combining estimates , and , and taking @xmath86 small enough , one obtains that @xmath129 since the energy estimate implies that @xmath130 . by using the gronwall s inequality , the elliptic regularity result @xmath131 and noticing that the assumption
, we deduce that the inequality holds . to prove the estimate , by using the standard elliptic regularity result on @xmath82
, we have @xmath132 where we have used the estimate in the last inequality . then by using the estimates , , and the above inequality
, we have @xmath133 this completes the proof of lemma [ lem2.4 ] . differentiating the momentum equation @xmath79 with respect to time , multiplying the resulting equation by @xmath135 , integrating it over @xmath136 and making use of the mass conservation equation @xmath64 ,
one obtains that @xmath137 differentiating the liquid crystal equation @xmath82 with respect to time gives @xmath138 multiplying the above equality with @xmath139 and integrating over @xmath71 , one obtains the equality @xmath140 where we have used the fact @xmath141 in the last equality . from the equation
, we can derive @xmath142 inserting the equalities and into derives @xmath143 we will estimate @xmath144 term by term . in the following calculations , we will make extensive use of sobolev embedding , hlder inequality , lemmas [ lem2.1][lem2.4 ] and the estimate , @xmath145 and @xmath146 to estimate the terms @xmath147 and @xmath148 , by using the standard elliptic estimate on and making use of the liquid crystal equation @xmath82 yield that @xmath149 where we have used the estimates and in the last inequality .
taking @xmath27 small enough yields @xmath150 making use of estimates and , we can estimate @xmath147 and @xmath148 as @xmath151 substituting all the estimates of @xmath144 into , and taking @xmath86 small enough , we obtain @xmath152\nonumber\\ \leq & c(\|\rho^{\frac{1}{2}}u_{t}\|_{l^{2}}^{2}+\|\nabla d_{t}\|_{l^{2}}^{2})(\|\nabla u\|_{l^{\infty}}+\|\nabla d\|_{h^{2}}^{2}+c)+c(\|\nabla d_{t}\|_{l^{2}}^{2}+1).\end{aligned}\ ] ] applying the gronwall s inequality to estimate , we deduce @xmath153 where we have used estimate and the assumption in the last inequality .
this completes the proof of lemma [ lem2.5 ] . from estimates , and , we have @xmath155 by using estimates and , we have @xmath156 where we have used the estimates , and in the last inequality . combining the estimates and above gives the estimate .
applying the operator @xmath124 to the mass conservation equation @xmath64 , then multiplying the resulting equation by @xmath157 and integrating over @xmath71 give @xmath158 that is @xmath159 by using the gronwall s inequality to the above estimate gives @xmath160 applying the standard elliptic regularity result @xmath161 , hlder inequality , sobolev embedding , the estimates and , we have @xmath162 inserting the estimate into yields @xmath163 where we have used the estimate , then applying the gronwall s inequality gives @xmath164 from and , we have @xmath165 it is easy to known that the estimate follows and immediately .
this completes the proof of lemma [ lem2.6 ] .
* proof of theorem [ thm1.2 ] . * from the existence result of theorem [ thm1.1 ] , we know that @xmath166 , @xmath167 , @xmath168 and @xmath169 are all continuous on the time interval @xmath170 . from the above lemmas [ lem2.1][lem2.6 ] , we see that for all @xmath171 , @xmath172 furthermore , there hold @xmath173;l^{2}),\end{aligned}\ ] ] and for all @xmath171 , @xmath174 where @xmath175 . therefore , from and
, we can take @xmath176 with any @xmath177 as the initial data and apply theorem [ thm1.1 ] to extend the local strong solution to a time interval @xmath178 $ ] for a uniform @xmath179 which only depends on the bounds obtained in these lemmas , so that the solution can be extended to the time interval @xmath180 .
this contradicts with the maximality of @xmath32 .
hence , the assumption @xmath181 can not be true .
this completes the proof of theorem [ thm1.2].@xmath182 + + | in this paper , we prove a beale kato majda blow - up criterion in terms of the gradient of the velocity only for the strong solution to the 3-d compressible nematic liquid crystal flows with nonnegative initial densities .
more precisely , the strong solution exists globally if the @xmath0-norm of the gradient of the velocity @xmath1 is bounded .
our criterion improves the recent result of x. liu and l. liu ( @xcite , a blow - up criterion for the compressible liquid crystals system , arxiv:1011.4399v2 [ math - ph ] 23 nov . 2010 ) .
* keywords * : compressible nematic liquid crystal flows ; strong solution ; blow - up criterion ; compressible navier
stokes equations * 2010 ams subject classification * : 76a15 , 76n10 , 35b65 , 35q35 |
let @xmath0 be an interval and @xmath1\to[0,1]$ ] , @xmath2 , a one - parameter family of maps of the unit interval such that , for every @xmath2 , @xmath3 is piecewise @xmath10 and @xmath11}|\partial_x t_a(x)|\ge\la>1 $ ] , where @xmath12 is independent on @xmath6 .
assume that , for all @xmath2 , @xmath3 admits a unique ( hence ergodic ) absolutely continuous invariant probability measure ( a.c.i.p . ) @xmath4 . according to @xcite and @xcite , for lebesgue
almost every @xmath5 $ ] , some iteration of @xmath13 by @xmath3 is contained in the support of @xmath4 . from birkhoff
s ergodic theorem we derive that lebesgue almost every point @xmath5 $ ] is _ typical _ for @xmath4 , i.e. , @xmath14 in this paper we are interested in the question if the same kind of statement holds in the parameter space , i.e. , if a chosen point @xmath5 $ ] is typical for @xmath4 for lebesgue a.e . @xmath2 , or more general , if , for some given @xmath15 map @xmath16 $ ] , @xmath17 is typical for @xmath4 for lebesgue a.e .
@xmath6 in @xmath18 . in section [ s.uniform ]
we try to establish sufficient conditions on a one - parameter family such that the following statement is true .
_ for lebesgue a.e .
@xmath2 , @xmath17 is typical for @xmath4 . _
the method we use in this paper is a dynamical one .
it is essentially inspired by the result of benedicks and carleson @xcite where they prove that for the quadratic family @xmath19 on @xmath20 there is a set @xmath21 of @xmath6-values of positive lebesgue measure for which @xmath22 admits almost surely an a.c.i.p . and for which the critical point is typical with respect to this a.c.i.p .
the main tool in their work is to switch from the parameter space to the dynamical interval by showing that the @xmath6-derivative @xmath23 is comparable to the @xmath13-derivative @xmath24 .
this will also be the essence of the basic condition on our one - parameter family @xmath3 with an associated map @xmath25 , i.e. , we require that the @xmath6- and the @xmath13-derivatives of @xmath26 are comparable ( see condition ( i ) below ) . some typicality results related to this paper can be found in @xcite , @xcite , and @xcite . the one - parameter families @xmath3 , @xmath2 , considered in these papers have in common that their slopes are constant for a fixed parameter value , i.e. , for each @xmath2 there is a constant @xmath27 such that @xmath28 on @xmath29 $ ] .
the advantage of our method and the main novelty of this paper is that we can drop this restriction and , thus , we are able to consider much more general families .
this paper consists mainly of two parts . in the first part , which corresponds to sections [ s.uniform]-[s.keylemmas ]
, we establish a general criteria for typicality . in the second part , which corresponds to sections [ s.beta]-[s.markov ]
, we apply this criteria to several well - studied one - parameter families and derive various typicality results for these families .
some of the results are presented in the following sections of this introduction .
we will shortly give a motivation and an overview of our criteria for typicality .
let @xmath30 $ ] be a ( small ) interval and set @xmath31 , @xmath2 , i.e. , @xmath32 is the forward iteration by @xmath33 of the points we are interested in . for @xmath34
fixed , the main estimate to be established in the method we apply is roughly of the form : @xmath35 where @xmath36 ( @xmath37 large ) are @xmath38 integers with large ( @xmath39 ) gaps between each other and @xmath40 is some constant .
such an estimate is easier to establish on the phase space for a fixed map @xmath3 in the family , i.e. , it is easier to verify the estimate @xmath41 ) . )
hence , in order to prove ( [ eq.intromain1 ] ) , the main idea in the first part of this paper is to compare sets in the parameter space @xmath18 with sets in the phase space @xmath29 $ ] .
this will be possible if the following three conditions , conditions ( i)-(iii ) , are satisfied .
the first two conditions are rather natural while the last condition is a bit technical and restrictive .
( however , as we will see in sections [ s.beta]-[s.markov ] these conditions are satisfied for a broad class of important one - parameter families of piecewise expanding maps . )
condition ( i ) roughly says that @xmath33 and @xmath42 are comparable , i.e. , there exists a constant @xmath40 such that @xmath43 for all @xmath44 , and @xmath2 for which the derivatives are defined .
this is a well - known condition for one - parameter families of maps on the interval . in the case of piecewise expanding unimodal maps if one chooses @xmath25 to be equal to the turning point ( or some iteration of it to make the @xmath13-derivative well - defined ) , this condition is equivalent to saying that the family is transversal ( cf .
lemma [ l.nondegenerate ] ) .
in fact , in order that ( [ eq.intromain0 ] ) holds for all @xmath44 it is enough to require that the map @xmath45 $ ] has a sufficiently high initial expansion for some @xmath46 ( see lemma [ l.startcalc ] ) , which makes condition ( i ) easy to check .
condition ( ii ) requires that the density for @xmath4 is uniformly in @xmath6 bounded above and below away from @xmath47 .
this ensures that there is a constant @xmath40 such that for all @xmath2 we have the estimate latexmath:[\[\label{eq.intromain21 } @xmath44 ( cf .
inequality ( [ eq.perronfrobenius ] ) ) .
condition ( iii ) requires that there is a kind of order relation in the one - parameter family in the sense that for each two parameter values @xmath49 satisfying @xmath50 the following holds .
the symbolic dynamics of @xmath3 is contained in the symbolic dynamics of @xmath51 and , furthermore , if @xmath52 is a ( maximal ) interval of monotonicity for @xmath33 and @xmath53 is the corresponding ( maximal ) interval of monotonicity for @xmath54 ( i.e. , @xmath53 has the same combinatorics as @xmath52 up to the @xmath55-th iteration ) , then the following holds .
the minimal distance of points in the image @xmath56 to points in the image @xmath57 is bounded from above by @xmath58 , i.e. , when the parameter are close then the images are close .
further , the size of @xmath56 is bounded above by the size of @xmath57 .
we shortly point out how to get an estimate as in ( [ eq.intromain1 ] ) .
if conditions ( i)-(iii ) are satisfied we first divide the parameter interval into intervals @xmath59 of length @xmath60 ( @xmath37 large ) in order to have good distortion estimates when switching from maps on the parameter interval to maps on the dynamical interval . on each such interval
@xmath61 we will establish roughly the estimate @xmath62 where @xmath63 is the right boundary point of @xmath61 , @xmath64 $ ] is an interval of approximately size @xmath60 containing the image @xmath65 , and @xmath66 is the interval twice as long as @xmath67 and having the same midpoint as @xmath67 .
the first inequality in ( [ eq.intromain3 ] ) is essentially due to condition ( i ) and ( iii ) where the main ingredient is lemma [ l.distortion1 ] . using the estimate ( [ eq.intromain21 ] ) ,
the last inequality in ( [ eq.intromain3 ] ) ( or , similarly , inequality ( [ eq.intromain2 ] ) ) is straightforward to verify whenever the family has the property that the image by @xmath33 of each ( maximal ) monotonicity interval for @xmath33 has size close to @xmath68 ( this is , e.g. , the case when the family preserves a markov structure ; cf .
section [ s.markov ] ) .
however , in the general case there are many monotonicity intervals with very small images which makes the proof more technical . a sufficient upper bound for the measure of exceptionally small monotonicity intervals
is established in lemma [ l.remainderintervals ] where an important ingredient is the @xmath69 gap between the @xmath70 s which makes it for too small intervals possible to recover. the example in section [ s.beta ] is a @xmath8-version of the @xmath9-transformation . by saying that a function is @xmath8 , we mean that it is @xmath15 and its derivative is in @xmath72 , i.e. , its derivative is lipschitz continuous with lipschitz constant bounded above by @xmath73 . for a sequence @xmath74 of real numbers such that @xmath75 as @xmath76 and a constant @xmath77 , let @xmath78 $ ]
be a right continuous function which is @xmath8 on each interval @xmath79 , @xmath80 .
furthermore , we assume that : * @xmath81 for each @xmath80 . * for each @xmath82 , @xmath83}\partial_xt(ax)\quad\text{and}\quad\sup_{x\in[0,1]}\partial_xt(ax)<\infty.\ ] ] see figure [ f.beta ] .
$ ] .,width=302 ] given a map @xmath84 as above , we obtain a @xmath8-version of the @xmath9-transformation @xmath1\to[0,1]$ ] , @xmath82 , by defining @xmath85 , @xmath5 $ ] . as we will see in section [ s.beta ] , for each @xmath82
, @xmath3 admits a unique a.c.i.p .
@xmath4 , and there are many functions @xmath25 for which we have almost sure typicality .
[ t.intro2 ] if @xmath86 $ ] is @xmath15 and @xmath87 , then @xmath17 is typical for @xmath4 for lebesgue a.e .
@xmath82 .
if we choose @xmath88 then @xmath89 and @xmath90 for all @xmath44 .
hence , if the condition @xmath87 in theorem [ t.intro2 ] is not satisfied , we can not any longer guarantee almost sure typicality for the a.c.i.p .
for an illustration of some curves on which we have a.s .
typicality see figure [ f.birkhoff ] ( when @xmath6 is fixed , we can apply birkhoff s ergodic theorem and get a.s .
typicality on the associated vertical line ) . observe that if we choose @xmath78 $ ] by @xmath91 , then @xmath92 is the usual @xmath9-transformation .
theorem [ t.intro2 ] generalizes a result due to schmeling @xcite where it is shown that for the usual @xmath71-transformation the point @xmath68 is typical for the associated a.c.i.p . for lebesgue a.e . @xmath82 .
-version of the @xmath71-transformation.,width=302 ] in section [ s.unimodal ] we investigate one - parameter families of piecewise expanding unimodal maps .
a map @xmath93\to[0,1]$ ] is a _ piecewise expanding unimodal map _ if it is continuous and if there exists a turning point @xmath94 such that @xmath95}$ ] and @xmath96}$ ] are @xmath8 , @xmath97}t'(x)>1 $ ] and @xmath98}t'(x)<-1 $ ] , and @xmath99 and @xmath100 .
( in section [ s.unimodal ] we will use a slightly different representation of piecewise unimodal maps which will be more convenient to work with . ) by @xcite ( or @xcite ) , since @xmath84 is only at one point not @xmath8 , there exists a unique a.c.i.p .
we call @xmath84 mixing , if it is topologically mixing on @xmath29 $ ] .
( this implies that @xmath102 $ ] . )
let @xmath0 be a finite closed interval .
for a one - parameter family of piecewise expanding unimodal maps @xmath3 , @xmath2 , we make some natural requirements on the parameter dependency as , e.g. , the following ( cf .
section [ ss.prel ] properties ( i)-(iii ) ) .
for each @xmath5 $ ] the map @xmath103 is @xmath72 on @xmath18 ( in particular , this implies that the turning point @xmath104 of @xmath3 is lipschitz continuous in @xmath6 ) , and if @xmath59 is an interval on which @xmath105 , then @xmath103 is @xmath106 and @xmath107 is @xmath72 on @xmath61 .
the main result in section [ s.unimodal ] can be stated as follows .
[ t.uniintro ] if @xmath3 , @xmath2 , is a one - parameter family of mixing piecewise expanding unimodal maps such that for some @xmath108 , @xmath109}|\partial_at_a(x)|}{\inf_{a\in\i}\inf_{x\in[0,1]}|\partial_xt_a(x)|-1},\ ] ] for all but finitely many @xmath2 ( i.e. , the set of @xmath2 for which @xmath110 is not differentiable is finite ) , then the turning point @xmath104 is typical for the a.c.i.p . for @xmath3 , for almost every @xmath2 .
other ways of stating theorem [ t.uniintro ] are : if there exists a @xmath108 such that the map @xmath111 satisfies condition ( i ) ( see section [ ss.main ] ) then the turning point is almost surely typical ; or if the family is non - degenerate ( or transversal ) in each point @xmath2 ( see section [ s.unimodal ] ) then the turning point is almost surely typical . in section [ s.unimodal ]
we will state a more local version of theorem [ t.uniintro ] . in section [ s.tent ] we will apply theorem [ t.uniintro ] to unimodal maps with slopes constant to the left and to the right of the turning point , the so called _ skew tent maps_. let these slopes be @xmath112 and @xmath113 where @xmath114 and denote the corresponding skew tent map by @xmath115 .
( in order that @xmath115 maps the unit interval into itself , we have also to assume that @xmath116 . )
fix two points @xmath117 and @xmath118 in the set @xmath119 such that @xmath120 , @xmath121 , and at least one of these two inequalities is sharp .
let @xmath122\to[\al_0,\al_1]\quad\text{and}\quad\be:[0,1]\to[\be_0,\be_1]\ ] ] be functions in @xmath123)$ ] such that @xmath124 , @xmath125 , and , for all @xmath126 $ ] , if @xmath127 , then @xmath128 and if @xmath129 , then @xmath130 .
consider the one - parameter family @xmath3 , @xmath126 $ ] , where @xmath1\to[0,1]$ ] is the skew tent map defined by @xmath131 .
the main result of section [ s.tent ] is the following .
[ t.intro3 ] the turning point for the skew tent map @xmath3 is typical for @xmath4 for lebesgue a.e .
@xmath126 $ ] .
theorem [ t.intro3 ] generalizes a result due to bruin @xcite where almost sure typicality is shown for the turning point of symmetric tent maps ( i.e. , when @xmath132 ) .
( see also @xcite . ) in section [ s.markov ] we consider one - parameter families preserving a certain markov structure .
a simple example for such a family are the maps @xmath1\to[0,1]$ ] defined by @xmath133 where the parameter @xmath134 .
see figure [ f.markov ] .
by @xcite , since this map has only one point of discontinuity , it admits a unique a.c.i.p .
@xmath4 ( which coincides in this case with the lebesgue measure on @xmath29 $ ] ) . in example
[ ex.markov ] in section [ s.markov ] we will show the following .
[ p.intro4 ] if @xmath135 is a @xmath15 map such that @xmath136 , then @xmath17 is typical for @xmath4 for a.e .
parameter @xmath134 .
observe that if @xmath137 where @xmath138 is the unique point of periodicity @xmath139 in the interval @xmath140 , then @xmath141 and @xmath17 is not typical for @xmath4 for any @xmath2 .
the very simple structure of the family in this section makes it a good candidate for serving the reader as a model along the paper .
example [ ex.markov ] in section [ s.markov ] is formulated slightly more generally by composing @xmath3 with a @xmath8 homeomorphism @xmath142\to[0,1]$ ] . where @xmath134.,width=302 ]
in this section we introduce the basic notation and put up a general model for one - parameter families of piecewise expanding maps of the unit interval . a map @xmath93\to\real$ ]
will be called _ piecewise @xmath8 _ if there exists a partition @xmath143 of the unit interval such that for each @xmath144 the restriction of @xmath84 to the open interval @xmath145 is a @xmath8 function .
observe that , by the lipschitz property , it follows that @xmath84 restricted to @xmath145 can be extended to the closed interval @xmath146 $ ] as a @xmath8 function .
let @xmath0 be an interval of finite length and @xmath1\to[0,1]$ ] , @xmath2 , a one - parameter family of piecewise @xmath8 maps where the lipschitz constant @xmath147 is independent on the choice of the parameter @xmath6 .
we assume that there are real numbers @xmath148 such that for every @xmath2 ,
@xmath149}|\partial_x t_a(x)|\quad\text{and}\quad\sup_{x\in[0,1]}|\partial_x t_a(x)|\leq\la.\ ] ] let @xmath150 be the partition of the unit interval associated to @xmath3 .
we make the following natural assumption on the parameter dependence . *
the number of monotonicity intervals for the @xmath3 s is constant , i.e. , @xmath151 , and the partition points @xmath152 , @xmath153 , are @xmath72 on @xmath18 .
furthermore , there is a constant @xmath154 such that @xmath155 for all @xmath156 and @xmath2 . *
if @xmath5 $ ] and @xmath59 is a parameter interval such that @xmath157 , for all @xmath158 and @xmath153 , then @xmath103 is @xmath106 and both maps @xmath103 and @xmath159 are @xmath72 on @xmath61 . in the sequel , instead of referring to @xcite and @xcite , we will refer to a paper by s. wong @xcite who extended the results in @xcite and @xcite on piecewise @xmath10 maps to a broader class of maps containing also piecewise @xmath8 maps . for a fixed @xmath2 , by @xcite , the number of ergodic a.c.i.p . for @xmath3
is finite and the support of an ergodic a.c.i.p . is a finite union of intervals . since we are always interested in only one ergodic a.c.i.p .
, we can without loss of generality assume that for each @xmath3 , @xmath2 , there is a unique ( hence ergodic ) a.c.i.p . which we denote by @xmath4 .
let @xmath160 .
by @xcite , for lebesgue a.e .
@xmath5 $ ] , the accumulation points of the forward orbit of @xmath13 is identical with @xmath161 , i.e. , @xmath162 for @xmath2 , let @xmath163 i.e. , the @xmath164 s are the monotonicity intervals for @xmath165 .
we assume the following . *
the number of @xmath164 s is constant in @xmath6 , i.e. , @xmath166 for all @xmath2 .
the boundary points of @xmath164 , @xmath167 , change continuously in @xmath2 . for a fixed parameter value @xmath2 ,
we denote by @xmath168 , @xmath44 , the partition on the dynamical interval consisting of the maximal open intervals of smooth monotonicity for the map @xmath169 .
more precisely , @xmath168 denotes the set of open intervals @xmath170 such that @xmath171 is @xmath8 and @xmath52 is maximal , i.e. , for every other open interval @xmath172 with @xmath173 , @xmath174 is no longer @xmath8 . clearly , the elements of @xmath175 are the interior of the intervals @xmath164 , @xmath167 . for an interval @xmath176 ,
we denote by @xmath177 the restriction of @xmath168 to the set @xmath178 . for a set @xmath179 ,
for which there exists , @xmath167 , such that @xmath180 , we denote by @xmath181 the index ( or symbol ) @xmath182 .
we will define similar partitions on the parameter interval @xmath18 .
let @xmath16 $ ] be a piecewise @xmath15 map from the parameter interval @xmath18 into the dynamical interval @xmath29 $ ] .
the points @xmath17 , @xmath2 , will be our candidates for typical points .
the forward orbit of a point @xmath17 under the map @xmath3 we denote as @xmath183 [ rem.critical ] since a lot of information for the dynamics of @xmath3 is contained in the forward orbits of the partition points @xmath152 , @xmath153 , it is of interest to know how the forward orbits of these points are distributed .
hence , an evident choice of the map @xmath25 would be @xmath184 where @xmath185 is an interval adjacent to @xmath152 .
let @xmath59 be an open interval ( or a finite union of open intervals ) and let @xmath186 denote the finite number of @xmath6 values in which @xmath25 is not differentiable .
by @xmath187 , @xmath44 , we denote the partition consisting of all open intervals @xmath52 in @xmath188 such that for each @xmath189 , @xmath190 , for all @xmath191 , and such that @xmath52 is maximal , i.e. , for every other open interval @xmath192 with @xmath173 , there exist @xmath193 and @xmath189 such that @xmath194 .
observe that this partition might be empty .
this is , e.g. , the case when @xmath195 for all @xmath2 or when @xmath3 is the usual @xmath71-transformation and the map @xmath25 is chosen to be equivalently equal to @xmath47 .
however , such trivial situations are excluded by condition ( i ) formulated in the next section [ ss.main ] . knowing that condition ( i ) is satisfied , then the partition @xmath187 , @xmath44 , can be thought of as the set of the ( maximal ) intervals of smooth monotonicity for @xmath196 $ ] ( in order that this is true one should also assume that @xmath197 for all @xmath6 contained in a partition element of @xmath187 ) .
we set @xmath198 ( and we will write @xmath199 instead of @xmath200 ) . if for an interval @xmath201 in the parameter space and for some integer @xmath202 the symbol @xmath203 exists for all @xmath204 , then it is constant and we denote this symbol by @xmath205 .
finally , in view of condition ( i ) below , observe that if a parameter @xmath2 is contained in an element of @xmath199 , @xmath44 , then also the point @xmath206 is contained in an element of @xmath168 which implies that @xmath33 is differentiable in @xmath17 . in this section we will state our main result , theorem [ t.main ] .
let @xmath37 be large .
in order to have good distortion estimates we will , in the proof of theorem [ t.main ] , split up the interval @xmath18 into smaller intervals @xmath59 of size @xmath60 .
the main idea in this paper is to switch from the map @xmath196 $ ] , @xmath207 , to the map @xmath208 $ ] where @xmath6 is the right boundary point of @xmath61 and @xmath209 is an interval of size @xmath210 oriented around @xmath65 ( assume @xmath211 $ ] has no discontinuities ) .
since the dynamics of the map @xmath3 is well - understood , we derive similar dynamical properties for the map @xmath42 , which then can be used to prove theorem [ t.main ] . to be able to switch from @xmath42 to @xmath33 , we put three conditions , conditions ( i)-(iii ) , on our one - parameter family and on the map @xmath25 associated to it . in condition ( i )
we require that the derivatives of @xmath42 and @xmath33 are comparable .
this is the very basic assumption in this paper . of course , the choice of the map @xmath16 $ ] plays here an important role .
if , e.g. , for every parameter @xmath2 , @xmath17 is a periodic point for the map @xmath3 , then @xmath42 will have bounded derivatives and the dynamics of @xmath42 is completely different from the dynamics of @xmath3 .
henceforth , we will use the notations @xmath212 * there is a constant @xmath213 such that for @xmath214 , @xmath44 , we have @xmath215 for all @xmath191 .
furthermore , the number of @xmath2 , which are not contained in any element @xmath216 is finite . given a @xmath15 map @xmath217 $ ] , as the following basic lemma asserts , to verify that there exists @xmath218 such that condition ( i ) is satisfied for the map @xmath219 it is sufficient to compute the @xmath6-derivative of @xmath220 for a finite number of @xmath55 s .
this makes it easy to check condition ( i ) numerically .
the proof of this lemma is given in section [ ss.lem22 ] .
[ l.startcalc ] assume that the parameter interval @xmath18 is closed , let @xmath217 $ ] be @xmath15 , and denote @xmath221 . if there exists @xmath218 such that @xmath222 , @xmath2 , and @xmath223 is piecewise @xmath15 ( with finitely many pieces ) such that @xmath224 then condition ( i ) is satisfied for the map @xmath225 .
we turn to condition ( ii ) . for @xmath2 ,
let @xmath226 denote the density for @xmath4 .
we assume that @xmath226 is uniformly in @xmath6 bounded away from @xmath47 .
* there exists a constant @xmath227 such that , for each @xmath2 , @xmath228 for @xmath4 almost every @xmath5 $ ] .
[ r.tomas ] in fact , instead of the uniformity of the constant @xmath229 one could @xmath229 allow to depend measurably on the parameter @xmath2 . in proposition [ p.main ] and its proof below , one could then consider closed sets @xmath230 with @xmath231 on which one has by lusin uniformity of @xmath229 .
the proof of proposition [ p.main ] restricted to such a set @xmath232 could then be adapted by observing that for the intervals @xmath61 in the proof it is only important that the right boundary point of @xmath61 lies in @xmath232 . however , in all the examples we are considering in this paper the uniformity of @xmath229 is easy to establish . while conditions ( i ) and ( ii ) are very natural requirements the following condition ( iii ) is more restrictive .
as we will see in sections [ s.beta]-[s.markov ] , condition ( iii ) is satisfied for important families of piecewise expanding maps . in particular ,
in section [ s.unimodal ] we will see that condition ( iii ) is satisfied for all non - degenerate families of piecewise expanding unimodal maps . for two non - empty sets @xmath233
, @xmath234 denotes the infimum of the distances @xmath235 over all possible points @xmath236 and @xmath237 .
* there is a constant @xmath238 such that the following holds . for all @xmath239 , @xmath240 , and
@xmath44 there is a mapping @xmath241 such that , for all @xmath242 , the elements @xmath52 and @xmath243 have the same symbolic dynamics : @xmath244 their images lie close together : @xmath245 and the size of the image of @xmath52 can be estimated above by the size of the image of @xmath243 : @xmath246 let @xmath1\to[0,1]$ ] , @xmath2 , be a piecewise expanding one - parameter family as described in section [ ss.prel ] , satisfying properties ( i)-(iii ) , and conditions ( ii ) and ( iii ) .
if for a piecewise @xmath15 map @xmath16 $ ] condition ( i ) is fulfilled , then @xmath17 is typical for @xmath4 for lebesgue almost every @xmath2 . in the considered examples below
, we will usually not verify conditions ( i)-(iii ) for the whole interval @xmath18 for which the corresponding family is defined .
instead we will cover @xmath18 by a countable number of smaller intervals and verify these conditions on these smaller intervals .
let @xmath44 and assume in the following formulas that , for the parameter values @xmath2 under consideration , @xmath247 and @xmath33 are differentiable in @xmath6 and @xmath248 , respectively .
for @xmath249 we have @xmath250 which implies @xmath251 for @xmath252 , choosing @xmath253 and @xmath254 , respectively , we get the following upper and lower bounds : @xmath255 where for the lower bound we used the assumption ( [ eq.startcomp ] ) . setting @xmath256 , this implies the existence of a constant @xmath257 as required in condition ( i ) .
it is only left to show that for each @xmath258 the number of @xmath2 which are not contained in any element @xmath214 is finite ( the partition @xmath199 is taken w.r.t . the map @xmath17 ) .
this is easily done by induction over @xmath55 . by the assumption in lemma [ l.startcalc ] ,
@xmath259 $ ] is not differentiable only in a finite number of points and , further , @xmath222 .
since @xmath260 and since by property ( i ) the boundary points @xmath152 are @xmath72 , we have that @xmath261 for all but finitely many @xmath2 .
thus , the number of @xmath2 which are not contained in any element @xmath262 is finite .
assume that @xmath44 and consider the partition @xmath263 .
from the lower bound in ( [ eq.startcalc2 ] ) , we derive that @xmath264 for all @xmath6 contained in an element of @xmath199 . as above
we derive that only a finite number of @xmath6 contained in an element of @xmath199 can be mapped by @xmath265 to a partition point @xmath152 , @xmath156 .
this concludes the proof of lemma [ l.startcalc ] .
the idea of the proof of theorem [ t.main ] is inspired by chapter iii in @xcite where benedicks and carleson prove the existence of an a.c.i.p . for a.e .
parameter in a certain parameter set ( the set @xmath266 ) .
their argument implies that the critical point is in fact typical for this a.c.i.p . let @xmath267\ ; \
q\in\rational , r\in\rational^+\right\}.\ ] ] we will show that there is a constant @xmath40 such that for each @xmath268 the function @xmath269 fulfills @xmath270 by standard measure theory ( see , e.g. , @xcite ) , ( [ eq.limsup ] ) implies that , for a.e .
@xmath2 , every weak-@xmath271 limit point @xmath272 of @xmath273 has a density which is bounded above by @xmath274 . in particular , @xmath272 is absolutely continuous . observe that , by the definition of @xmath32 , the measure @xmath272 is also invariant for @xmath3 and ,
hence , @xmath272 is an a.c.i.p . for @xmath3 . by the uniqueness of the a.c.i.p . for @xmath3
, we finally derive that , for a.e .
@xmath2 , the weak-@xmath271 limit of ( [ eq.weakstarlimit ] ) exists and coincides with the a.c.i.p .
@xmath4 . this concludes the proof of theorem [ t.main ] . in order to prove ( [ eq.limsup ] ) ,
it is sufficient to show that for all ( large ) integers @xmath275 there is an integer @xmath276 , growing for fixed @xmath67 at most exponentially in @xmath38 , such that @xmath277 for all @xmath278 ( see lemma a.1 in @xcite ) . in the remaining part of this section
, we assume that @xmath268 is fixed . for @xmath275 , we have @xmath279 observe that for a fixed parameter @xmath6 , there exist an integer @xmath182 and a set @xmath280 ( which is the union of finitely many intervals ) such that the system @xmath281 with the measure @xmath4 is exact and , hence , it is mixing of all degrees ( see @xcite and @xcite ) .
it follows that for sequences of non - negative integers @xmath282 , @xmath283 , with @xmath284 one has @xmath285 since the maps @xmath33 and @xmath42 are , by conditions ( i)-(iii ) , comparable it is natural to expect similar mixing properties for the maps @xmath42 . in fact , in the section [ ss.proofpropmain ] , we are going to prove the following statement .
[ p.main ] assume that conditions ( i)-(iii ) are satisfied . disregarding a finite number of parameter values in @xmath18
, we can cover @xmath18 by a countable number of intervals @xmath286 such that for each such interval @xmath287 there is a constant @xmath40 such that the following holds . for all @xmath275
, there is an integer @xmath276 growing at most exponentially in @xmath38 such that , for all @xmath278 and for all integer @xmath38-tuples @xmath288 with @xmath289 and @xmath290 , @xmath291 , @xmath292 seen from a more probabilistic point of view , proposition [ p.main ] says that whenever the distances between consecutive @xmath70 s are sufficiently large , the behavior of the @xmath293 s is comparable to that of independent random variables .
now , the number of @xmath38-tuples @xmath288 in the sum in ( [ eq.integral ] ) , for which @xmath294 or @xmath295 , is bounded by @xmath296
. hence , by proposition [ p.main ] , we obtain @xmath297 whenever @xmath298 since both terms in this lower bound for @xmath37 grow at most exponentially in @xmath38 , this implies ( [ eq.limsup ] ) , for a.e .
this concludes the proof of theorem [ t.main ] .
first we cover the parameter interval @xmath18 by smaller intervals @xmath287 which will be more convenient in the proof of lemma [ l.remainderintervals ] below . fix an integer @xmath300 so large that @xmath301 . for @xmath2 , let @xmath302 since condition ( iii ) holds , we can argue as in the proof of lemma [ l.app1 ] ( see inequality ( [ eq.kaap ] ) ) and , disregarding a finite number of parameter values in @xmath18 , we can cover @xmath18 by a countable number of closed intervals @xmath286 such that for each such interval @xmath287 there is a constant @xmath303 such that @xmath304 for all @xmath299 .
we assume also that @xmath287 is chosen such that @xmath305 $ ] is @xmath15 ( and not piecewise @xmath15 ) .
henceforth , we fix such an interval @xmath287 . instead of
@xmath287 we will write again @xmath18 . conditions ( i)-(iii ) enables us to switch from maps on the parameter space to maps on the dynamical interval . in order to have good distortion estimates , we split up the integral in proposition [ p.main ] and integrate over smaller intervals of length @xmath60 . in the following
, we are going to show that there exist a constant @xmath40 and an integer @xmath276 growing at most exponentially in @xmath38 such that , for @xmath278 , we have @xmath306 for all intervals @xmath59 of length @xmath60 and all @xmath38-tuples @xmath288 as described in proposition [ p.main ] .
if @xmath307 , this immediately implies that , for @xmath278 , @xmath308 which concludes the proof of proposition [ p.main ] ( where one has to adapt the constant @xmath274 in the statement of proposition [ p.main ] to @xmath309 ) . let @xmath37 be large and @xmath59 an open interval of size @xmath60 .
note that by condition ( i ) , for @xmath44 , there are only finitely many parameter values not contained in any element of @xmath187 .
hence , we can neglect such parameter values and focus on the partitions @xmath187 .
consider the set @xmath310 obviously if we show that @xmath311 , this implies ( [ eq.jjj ] ) .
our strategy of establishing this estimate for @xmath312 is to compare the elements in @xmath312 with elements in the partition @xmath313 where @xmath63 denotes the right boundary point of @xmath61 .
the following lemma will allow us to switch from partitions on the parameter space to partitions on the phase space .
its proof is given in section [ ss.distortion1 ] .
[ l.distortion1 ] under the assumption that conditions ( i ) and ( iii ) are satisfied , there is an integer @xmath314 and a constant @xmath315 such that for all open intervals @xmath59 of length @xmath60 the following holds .
there is an at most @xmath316-to - one map @xmath317 such that , for @xmath318 , the images of @xmath52 and @xmath319 are close : @xmath320 and the size of @xmath52 is controlled by the size of @xmath319 : latexmath:[\[\label{eq.sun } let @xmath64 $ ] be the interval obtained by the interval having the same midpoint as @xmath65 and having size @xmath322 intersected with the support @xmath323 of the a.c.i.p . for @xmath324 .
( observe that @xmath18 is chosen such that @xmath16 $ ] has no discontinuities . )
lemma [ l.distortion1 ] allows us to switch from considering @xmath312 to considering the set @xmath325 where @xmath66 denotes the interval twice as long as @xmath67 and having the same midpoint as @xmath67 .
indeed if @xmath326 then , by ( [ eq.sun2 ] ) , for all @xmath327 , there exists @xmath328 such that @xmath329 .
thus , by ( [ eq.sun ] ) , and since the map @xmath330 is at most @xmath316-to - one we obtain @xmath331 .
fix an integer @xmath332 such that @xmath333 , and assume that @xmath37 is so large that @xmath334 , which ensures that there are at least @xmath335 iterations between two consecutive @xmath70 s ( and between @xmath336 and @xmath37 ) .
let @xmath337 and , for @xmath338 , set @xmath339 clearly , @xmath340 .
notice that , by the assumption on @xmath341 , we have @xmath342 for all @xmath343 and , thus , @xmath344 note that , since the density @xmath345 is a fixed point of the perron - frobenius operator , we have , for @xmath346 , @xmath347 for a.e . @xmath348 . by condition ( ii ) , we get @xmath349 for a.e .
@xmath348 . for @xmath350 , @xmath44
, it follows that @xmath351 by lemma [ l.distortion ] ( where we set @xmath352 ) , which is stated and proven in the next section , we get the distortion estimate @xmath353 , for @xmath354 .
now if @xmath355 then we get @xmath356 in the remaining part of this section let @xmath357 . in order to apply ( [ eq.anteil ] )
, we will exclude in each set @xmath358 certain intervals with too short images . to this end
we define , for @xmath359 , the following exceptional sets ( let @xmath360 ) : @xmath361 the following lemma gives an estimate on the size of these exceptional sets .
it is proven in section [ ss.remainders ] .
[ l.remainderintervals ] there is a number @xmath276 growing at most exponentially in @xmath38 such that @xmath362 for all @xmath359 and @xmath278 . disregarding finitely many points , @xmath363 , @xmath359 , can be seen as a set of disjoint and open intervals @xmath364 such that each @xmath364 is an element of a partition @xmath365 , @xmath366 , and @xmath367 . by ( [ eq.anteil ] )
, we obtain @xmath368 which in turn implies that , for @xmath278 , @xmath369 hence , we have @xmath370 . since @xmath331
we conclude that @xmath371 where @xmath372 .
this implies ( [ eq.jjj ] ) which is the estimate we had to show .
in this section we will prove the key lemmas , lemma [ l.distortion1 ] and lemma [ l.remainderintervals ] , in the proof of proposition [ p.main ] .
as seen in section [ ss.proofpropmain ] , lemma [ l.distortion1 ] makes it possible to compare partition elements on the parameter space to partition elements on the phase space , and lemma [ l.remainderintervals ] provides us with a good estimate of exceptional partition elements in the phase space with too small images .
we establish first a distortion lemma .
[ l.distortion ] if the one - parameter family @xmath3 , @xmath2 , satisfies condition ( iii ) , then there exists a constant @xmath373 such that we have the following distortion estimate
. let @xmath374 and @xmath239 such that @xmath240 and @xmath375 . for @xmath242 , @xmath376 , we have @xmath377 for all @xmath378 and @xmath379 . if @xmath380 in lemma [ l.distortion ] , then we get a standard distortion estimate for piecewise expanding @xmath8 maps .
fix @xmath381 such that @xmath382 .
taking constant @xmath383 in lemma [ l.distortion ] greater than @xmath384 , for @xmath385 , the distortion estimate is trivially satisfied and we can assume that @xmath386 . observe that , by condition ( iii ) , there exist points @xmath387 and @xmath388 such that @xmath389 . for @xmath390 ,
let @xmath391 be the pre - images of @xmath392 and @xmath393 , i.e. , @xmath394 and @xmath395 .
note that , by ( iii ) , we have @xmath396 .
let @xmath397 , and denote by @xmath398 the ( maximal ) monotonicity interval @xmath399 , @xmath400 , for @xmath1\to[0,1]$ ] which contains the domain @xmath401 .
recall that , by property ( i ) , @xmath402 .
the distance between @xmath403 and @xmath404 , @xmath390 , satisfies @xmath405 ) , we will show inductively that @xmath406 .
this is obviously true for @xmath407 , so assume that ( [ eq.distance1 ] ) holds for @xmath408 where @xmath390 .
recall that @xmath375 .
hence , by property ( i ) , there is an interval @xmath409 such that @xmath410 and @xmath411 , for all @xmath412 $ ] .
( since @xmath413 and @xmath414 the interval @xmath67 is non - empty . ) take a point @xmath415 such that @xmath416 . since @xmath417
, we obtain @xmath418 by property ( ii ) the map @xmath419 is @xmath72 on @xmath420 $ ] which implies @xmath421 .
further , we have @xmath422 altogether , we obtain @xmath423 and we easily deduce that ( [ eq.distance1 ] ) holds for @xmath424 . for @xmath390 ,
let @xmath409 and @xmath415 as in the proof of ( [ eq.distance ] ) .
we obtain @xmath425 where we used property ( ii ) for estimating the term @xmath426 . using ( [ eq.distance ] )
, we get that @xmath427 where @xmath428 .
altogether , we obtain @xmath429 since @xmath207 , the product in the last term of inequality ( [ eq.distortion11 ] ) is clearly bounded above by a constant independent on @xmath37 .
hence , this shows the upper bound in the distortion estimate .
the lower bound is shown in the same way .
we define the map @xmath430 as follows .
let @xmath318 and @xmath191 . by the definition of the partitions associated to the parameter interval , we have @xmath431 , for all @xmath432 ( recall that @xmath433 ) . hence , there exists an element @xmath434 in the partition @xmath435 containing the point @xmath17 .
we set @xmath436 where @xmath437 is the map given by ( iii ) .
note that the element @xmath438 has the same combinatorics as @xmath52 , i.e. , @xmath439 @xmath432 . since there can not be two elements in @xmath313 with the same combinatorics , the element @xmath53 is independent on the choice of @xmath191 .
it follows that the map @xmath440 is well - defined . from the first claim in the proof of lemma [ l.app1 ]
, we get that the boundary points of @xmath441 change continuously in @xmath191 . hence ,
since @xmath442 is contained in @xmath441 and since @xmath443 , for all @xmath191 , we get @xmath444 where @xmath445 and @xmath446 denote the left and right endpoint of @xmath52 , respectively . by ( [ eq.drei2 ] ) , we get @xmath447 , for all @xmath191 , and , by ( [ eq.dreidist ] ) , we obtain @xmath448 from condition ( i ) , it follows that @xmath449 .
thus , we deduce that @xmath450 where in the last inequality we used once more ( [ eq.drei2 ] ) . by condition ( i ) and lemma [ l.distortion ] , we obtain that @xmath451 , for all @xmath191 and @xmath452 .
we conclude that @xmath453 which implies the estimate ( [ eq.sun ] ) in lemma [ l.distortion1 ] . in order to prove ( [ eq.sun2 ] ) ,
observe that @xmath454 , for all @xmath191 .
hence , by ( [ eq.distance ] ) , we have @xmath455 where @xmath274 is the constant in the righthand side of ( [ eq.distance ] ) . for @xmath37
sufficiently large we have @xmath456 , if @xmath457 .
thus , since @xmath32 is contained in @xmath458 , we conclude @xmath459 in order to conclude the proof of lemma [ l.distortion1 ] , it is only left to show that the map @xmath440 is at most @xmath316-to - one for some integer @xmath314 .
let @xmath460 be so large that @xmath461 for all @xmath462 and parameter values @xmath2 for which the derivative is defined ( @xmath73 is the lipschitz constant introduced in section [ ss.prel ] ) . if @xmath463 , using that the partition points @xmath464 are @xmath72 on @xmath18 , it is easy to show that the map @xmath465 is one - to - one .
hence , setting @xmath466 we derive that the map @xmath467 is at most @xmath316-to - one .
let @xmath44 and @xmath2 .
for each @xmath468 , we define the set @xmath469}(a)|\om'\ ; \ \nexists\;\tilde{\om}\in\p_{j+k}(a)|\om',\\ \tau\le k\le[\sqrt{n}],\ \text{s.t.}\ \tilde{\om}\supset\om\ \text{and}\ |t_a^{j+k}(\tilde{\om})|\ge\delta\}.\end{gathered}\ ] ] observe that the choice of @xmath470 in the beginning of section [ ss.proofpropmain ] implies that if @xmath471 , @xmath472 , then @xmath473 . from this
we deduce that if @xmath474 , @xmath475 , and @xmath472 then we have @xmath476 in other words only the elements in @xmath477 that are adjacent to a boundary point of @xmath478 can have a small image . by a repeated use of this fact and using that @xmath479 ( recall that @xmath480 is the number of elements in @xmath175 and , by property ( iii ) , @xmath480 does not depend on @xmath6 ) , we derive @xmath481}(a)|e_{\om'}\}\le p_1^\tau\cdot2\cdot2^{([\sqrt{n}]-\tau)/t_0}\le2p_1^\tau\sqrt{\la}^{[\sqrt{n}]},\ ] ] where in the last inequality we used the definition of @xmath300 .
it follows that @xmath482}}=:\gamma_n.\ ] ] observe that if we choose @xmath483 , @xmath359 , and @xmath484 then the exceptional set @xmath485 in lemma [ l.remainderintervals ] is contained in @xmath486 we have @xmath487 where @xmath488 .
set @xmath489 and , for @xmath490 , let @xmath491\cap t_a^j(\om)$ ] , where @xmath492 denotes the left boundary point of @xmath56 . by ( [ eq.gagaga ] ) and the distortion estimate in lemma [ l.distortion ] ( where @xmath493 ) , we get @xmath494 now , summing over all points in @xmath495 and moving the sum over the partition elements inside the integral , we derive that @xmath496}\sum_{\begin{subarray}{c } x\in k(a ) \\
t_a^j(x)=y \end{subarray } } \frac{1}{|t_a^j\,'(x)|}dy\le c_3c_1 ^ 2\gamma_n\#\o_j(a),\ ] ] where in the last inequality we used ( [ eq.perronfrobenius ] ) .
observe that for each @xmath497 there is a monotonicity domain @xmath498 for @xmath499 and a partition point @xmath500 such that @xmath501 for some @xmath390 .
thus , since @xmath207 , we have @xmath502 . recall that @xmath357 and @xmath503 . finally , in the case when @xmath483 , @xmath359 , and @xmath484 we deduce that @xmath504 for @xmath278 , where @xmath276 can obviously be chosen to grow only exponentially in @xmath38 .
this concludes the proof of lemma [ l.remainderintervals ] .
we apply theorem [ t.main ] to a @xmath8-version of @xmath71-transformations .
let the map @xmath78 $ ] be piecewise @xmath8 and @xmath74 be the associated partition , where @xmath75 as @xmath76 .
we assume that : * @xmath84 is right continuous and @xmath81 , for each @xmath80 .
* for each @xmath82 , @xmath83}\partial_xt(a x)\quad\text{and}\quad\sup_{x\in[0,1]}\partial_xt(ax)<\infty.\ ] ] see figure [ f.beta ] .
we define the one - parameter family @xmath1\to[0,1]$ ] , @xmath82 , by @xmath85 .
there exists a unique a.c.i.p .
@xmath4 for each @xmath3 as the following lemma asserts .
[ l.uniqueb ] for each @xmath82 there exists a unique a.c.i.p .
@xmath4 for @xmath3 .
the support @xmath161 is an interval adjacent to @xmath47 and the map @xmath505 , @xmath82 , is piecewise constant where the set of discontinuity points is countable and nowhere dense .
the proof of lemma [ l.uniqueb ] is given in section [ ss.eins1 ] .
henceforth , @xmath506 will always denote a closed interval on which @xmath507 is constant as well as the number of discontinuities of @xmath3 inside @xmath161 and @xmath29 $ ] , i.e. , the numbers @xmath508 and @xmath509 are constant on @xmath18 . for such an interval @xmath18
it is now straightforward to check that the one - parameter family @xmath3 , @xmath2 , fits into the model described in section [ ss.prel ] fulfilling properties ( i)-(iii ) .
now , we can state the main result of this section .
[ t.beta ] if for a @xmath15 map @xmath16 $ ] condition ( i ) is satisfied , then @xmath17 is typical for @xmath4 for a.e . @xmath2 . as the family @xmath3 we could also consider other models as , e.g. , @xmath510 where @xmath142\to[0,1]$ ] is a @xmath8 homeomorphism with a strict positive derivative .
even if this model is not included in the families described above , it would be easier to treat since , seen as a map from the circle into itself , it is non - continuous only in the point @xmath47 which , in particular , implies that @xmath511 $ ] . by theorem [ t.main ] , in order to proof theorem [ t.beta ] , it is sufficient to check conditions ( ii ) and ( iii ) .
we will show that there is a large class of maps @xmath512 for which we have almost sure typicality : [ c.beta ] if @xmath513 $ ] is @xmath15 such that @xmath514 , then @xmath248 is typical for @xmath4 for a.e . @xmath82 .
observe that the map @xmath515 @xmath82 , satisfies @xmath516 and @xmath514 , and , hence , corollary [ c.beta ] can be applied to these values which are important from a dynamical point of view .
we prove first lemma [ l.uniqueb ] .
for @xmath82 let @xmath4 be an a.c.i.p .
for @xmath3 with support @xmath161 and let @xmath179 be an interval .
since @xmath3 is expanding there exists an integer @xmath44 such that @xmath517 $ ] is not any longer continuous .
it follows that @xmath518 contains a neighborhood of @xmath47 . if @xmath3 had more than one a.c.i.p .
then , by @xcite , there would exist two a.c.i.p.s with disjoint supports ( disregarding a finite number of points ) .
this shows that the a.c.i.p .
@xmath4 is unique and its support @xmath161 contains an interval adjacent to @xmath47 .
since the image by @xmath3 of an interval adjacent to @xmath47 is again an interval adjacent to @xmath47 , using the ergodicity of @xmath4 , one deduces that @xmath161 is a single interval and @xmath519 where @xmath73 is a sufficiently small interval adjacent to @xmath47 .
observe that for an arbitrary interval @xmath520 $ ] adjacent to @xmath47 one has @xmath521 for all @xmath522 , which implies that the right end point @xmath523 of @xmath161 is non - decreasing .
now , we easily derive that there is @xmath346 such that @xmath524 and @xmath525 .
since @xmath526 is constant in @xmath6 , it follows that if @xmath527 then the support of the a.c.i.p . for @xmath51
is equal to @xmath161 for all @xmath528 $ ] where @xmath529 is maximal such that @xmath530 for all @xmath528 $ ] .
in the case when @xmath531 and @xmath532 , for some @xmath346 , then since @xmath533 for @xmath522 sufficiently close to @xmath6 , we derive that the support of the a.c.i.p . @xmath51 is constantly equal to @xmath161 for all @xmath522 close to @xmath6 .
thus , the only possible obstacle left is the set @xmath534 . but
using that the fixed points for @xmath3 are strictly decreasing in @xmath6 and @xmath523 is non - decreasing in @xmath6 , we deduce that this set is countable and nowhere dense .
this concludes the proof of lemma [ l.uniqueb ] .
we proceed with the proof of corollary [ c.beta ] which is an application of lemma [ l.startcalc ] . disregarding
countably many points we can cover @xmath535 by intervals @xmath18 as described in the beginning of section [ s.beta ] .
thus , in order to prove corollary [ c.beta ] , it is sufficient to verify the requirements of lemma [ l.startcalc ] for the family @xmath3 together with the map @xmath512 restricted to such a parameter interval @xmath18 .
recall that @xmath536 stands for a uniform lower bound for the expansion in the family . for @xmath5 $ ] ,
observe that by the definition of @xmath3 , we have @xmath537 , for all @xmath2 such that @xmath538 , @xmath80 .
recall the formula ( [ eq.startcalc1 ] ) for the derivative of @xmath221 , @xmath44 ( where we set @xmath253 ) . since @xmath514 and @xmath516 , for all @xmath2 , since the points of discontinuity @xmath539 for the map @xmath3 are strictly decreasing , and
since all the terms on the right - hand side of ( [ eq.startcalc1 ] ) are non - negative , we derive inductively that , for each @xmath44 , the maps @xmath247 and @xmath33 are differentiable in @xmath6 and @xmath248 , respectively , for all but finitely many @xmath2 . furthermore , since @xmath18
is closed we have that @xmath248 is uniformly bounded away from @xmath47 and , from the term in the sum in ( [ eq.startcalc1 ] ) when @xmath407 , we obtain @xmath540 thus , we find @xmath218 such that ( [ eq.startcomp ] ) is satisfied .
the only obstacle in applying lemma [ l.startcalc ] might be that @xmath541 .
however , by ( [ eq.betastartcalc ] ) and property ( [ eq.propertyk ] ) , we derive that , by possibly disregarding a countable number of points , we can cover @xmath18 by a countable number of intervals @xmath286 such that for each such interval @xmath287 there is an integer @xmath542 such that @xmath543 is @xmath15 satisfying ( [ eq.startcomp ] ) and @xmath544 for all @xmath299 . by lemma [ l.startcalc ] , it follows that condition ( i ) is satisfied for the map @xmath545 , @xmath299 . by theorem [ t.beta
] this concludes the proof of corollary [ c.beta ] .
the verification of condition ( iii ) in the section [ ss.betacdrei ] does not make use of condition ( ii ) .
hence , by lemma [ l.app1 ] , we can without loss of generality assume that there is a constant @xmath546 such that for each @xmath2 the density @xmath226 is bounded from above by @xmath274 and , further , there exists an interval @xmath547 of length @xmath548 such that @xmath226 restricted to @xmath547 is bounded from below by @xmath548 ( otherwise , disregarding a finite number of points , by lemma [ l.app1 ] , we can cover the interval @xmath18 by a countable number of closed subintervals on each of which this is true and then proceed with these subintervals instead of @xmath18 ) . to conclude the verification of condition ( ii ) it is left to show that there exists a lower bound for @xmath226 on the whole of @xmath161 . to make the definition of the intervals @xmath549 below work
, we assume that the interval @xmath547 is closed to the left and open to the right .
recall that , by property ( i ) in section [ ss.prel ] , we have @xmath550 , @xmath156 , for some constant @xmath551 .
let @xmath552 and take @xmath475 so large that @xmath553 .
we claim that @xmath554 .
let @xmath555 and assume that we have defined the interval @xmath556 , @xmath557 , where @xmath558 is a ( not necessarily maximal ) interval of monotonicity for @xmath559 . if @xmath560 , we stop and do not define @xmath549 .
if @xmath561 is not contained in @xmath562 then , since @xmath558 is a monotonicity interval for @xmath559 and by the definition of @xmath563 , it follows that there can lie at most one partition point @xmath539 in the image @xmath564 .
if there is no partition point in this image then we let @xmath565 , which is in this case also a monotonicity interval for @xmath566 .
if there is a partition point @xmath567 , then we define @xmath568 to be the interval of monotonicity for @xmath566 such that @xmath569 .
note that @xmath570|<\vep/\la$ ] , since otherwise we would have @xmath560 . assuming that @xmath571 is defined , we obtain @xmath572 where we used the definitions of @xmath563 and @xmath573 .
since @xmath574 $ ] , this is a contradiction and it follows that the maximal integer @xmath575 such that @xmath549 is defined is strictly smaller than @xmath573 .
hence , @xmath576 contains @xmath561 as claimed above .
let @xmath577 be the left boundary point of @xmath18 and denote by @xmath578 the right boundary point of @xmath579 ( which is constant in @xmath2 ) .
as observed in the proof of lemma [ l.uniqueb ] there exists a partition point @xmath580 , @xmath346 , such that @xmath581 and @xmath582 .
since @xmath51 is exact ( see @xcite ) we derive that there exists an integer @xmath583 such that @xmath584 .
since @xmath585 for all @xmath586 , it follows that @xmath587 for all @xmath2 .
altogether , we derive that for @xmath588 ( i.e. , @xmath55 is independent on @xmath2 ) we have @xmath589 for all @xmath2 .
now , by the perron - frobenius equality , it follows that @xmath590 for a.e .
@xmath591 , @xmath2 ( recall that @xmath592 is an upper bound for the maximal expansion in the family ) .
this concludes the proof of a lower bound for @xmath226 on the whole of @xmath161 .
we verify condition ( iii ) by induction over @xmath44 .
let @xmath239 such that @xmath240 .
note that @xmath175 consists of the elements @xmath593 , @xmath594 and the element @xmath595 where @xmath578 is the right boundary point of @xmath161 ( which does not depend on @xmath2 ) .
thus , if @xmath594 then we clearly have @xmath596 . since @xmath597 , for all @xmath2
, we derive that @xmath598 .
hence , ( iii ) holds for @xmath599 .
assume that ( iii ) holds for @xmath44 .
let @xmath600 and @xmath601 the corresponding element in @xmath602 .
note that the image by @xmath566 , @xmath557 , of an element in @xmath603 is always adjacent to @xmath47 .
since @xmath604 and the partition points @xmath539 s are decreasing , it follows immediately that for every element @xmath605 there is a unique element @xmath606 fulfilling @xmath607 , @xmath608 , and @xmath609 . defining @xmath610 shows that ( iii ) holds also for @xmath611 .
let @xmath84 be a piecewise expanding unimodal map as defined below and @xmath3 , @xmath612 , @xmath613 , a one - parameter family of piecewise expanding unimodal maps through @xmath84 , i.e. , @xmath614 . in this section
, we will show that if the family is non - degenerate ( or transversal ) then there exists @xmath615 such that for lebesgue almost every parameter values @xmath616 the turning point of @xmath3 is typical for the a.c.i.p . for @xmath3 . we will also state a condition for other points than the turning point in order to get almost sure typicality .
henceforth , let @xmath617 be a fixed real number .
we call a map @xmath618\to\real$ ] a _ piecewise expanding unimodal _ map if it satisfies the following properties : * @xmath619)$ ] and @xmath84 is @xmath8 on the intervals @xmath620 $ ] and @xmath29 $ ] . *
@xmath621}t'(x)>1\quad\text{and}\quad\sup_{x\in[0,1]}t'(x)<-1 $ ] .
* @xmath622,@xmath623 .
observe that the interval @xmath624 $ ] is forward invariant for @xmath84 . since @xmath84 is @xmath8 on @xmath625 $ ] except at the point @xmath47 , by @xcite , there exists a unique a.c.i.p .
@xmath101 for @xmath84 .
we denote the set of piecewise expanding unimodal maps by @xmath626 .
we say that a map @xmath627 is _ mixing _ if it is topologically mixing on @xmath624 $ ] .
observe that if @xmath84 is mixing then the support of @xmath101 is the whole interval @xmath624 $ ] .
in this section we will consider only maps @xmath627 which are mixing .
a _ one - parameter family through @xmath627 _ is a family @xmath628 , @xmath612 , @xmath613 , satisfying : * @xmath614 .
* for all @xmath629 $ ] the map @xmath103 is @xmath630 .
* the maps @xmath103 and @xmath107 are @xmath72 on @xmath631 .
let @xmath628 , @xmath612 , be a one - parameter family of mixing piecewise expanding maps through a map @xmath627 and @xmath632 the affine map from @xmath633 $ ] onto @xmath29 $ ] with , say , positive derivative . since @xmath634 is lipschitz in @xmath6 and since the length of the invariant interval @xmath633 $ ] is bounded from below by @xmath536 ( recall that @xmath12 is a lower bound for the expansion in the family ; cf .
( [ eq.la ] ) ) and from above by @xmath635 , it is obvious that the family @xmath636\to[0,1]$ ] , @xmath612 , satisfies properties ( i)-(iii ) in section [ ss.prel ] . henceforth , we will consider the family @xmath637\to[t_a(1),1]$ ] , @xmath612 , and not the family defined on the unit interval which is affinely conjugated to it .
the partitions in section [ ss.partitions ] are defined in an analogous way for the family @xmath3 .
to each one - parameter family @xmath628 , @xmath612 , through a map @xmath627 we associate a number @xmath638 given by @xmath639}\left|\partial_at_a(x)|_{a=0}\right|}{\la-1}.\ ] ] this number serves as a threshold in order to get almost sure typicality .
it stays in direct correspondence to the threshold given in lemma [ l.startcalc ] in order to verify condition ( i ) .
the main result of this section is the following .
[ t.main1 ] let @xmath628 , @xmath612 , @xmath613 , be a one - parameter family of mixing unimodal maps through a map @xmath627 .
if there exists @xmath108 such that @xmath640 , then there exists @xmath615 such that @xmath47 is typical for @xmath4 for almost every @xmath616 .
in order to prove theorem [ t.main1 ] , we will show that there exists @xmath615 such that the family @xmath3 , @xmath616 , together with the map @xmath641 , @xmath616 , satisfy conditions ( i)-(iii ) .
( to verify condition ( ii ) and ( iii ) we possibly have to divide @xmath642 into smaller intervals . )
knowing that conditions ( ii ) and ( iii ) hold for a one - parameter family , we can apply theorem [ t.main ] , and one directly gets almost sure typicality statements for other points than the turning point .
[ c.main2 ] let @xmath628 , @xmath612 , @xmath613 , be a one - parameter family of mixing unimodal maps through a map @xmath627 such that @xmath640 , for some @xmath108
. then there exists @xmath615 such that if @xmath643 $ ] is a @xmath15 map satisfying condition ( i ) then @xmath17 is typical for @xmath4 for almost every @xmath616 . before we start proving theorem [ t.main1 ] , we would like to point out the connection of the requirement @xmath640 in theorem [ t.main1 ] and corollary [ c.main2 ] to a standard non - degeneracy condition for one - parameter families of maps on the interval .
a one - parameter family through a map @xmath627 is _ non - degenerate _ or _ transversal _ if @xmath644 in the case where the turning point @xmath47 is not periodic for @xmath84 , or if @xmath645 in the case where @xmath47 has prime period @xmath646 for @xmath84 ( observe that @xmath647 ) .
this non - degeneracy condition ( or transversality condition ) appears , e.g. , in a generalization of jakobson s theorem in @xcite . in the context of piecewise expanding unimodal maps it appears , e.g. , in @xcite , @xcite , and @xcite .
we say that a piecewise expanding unimodal map @xmath84 is _ good _ if either @xmath47 is not periodic , or writing @xmath647 for the prime period of @xmath47 , if @xmath648 ( see remark [ r.good ] below . ) [ l.nondegenerate ] let @xmath628 , @xmath612 , @xmath613 , be a one - parameter family of unimodal maps through a good map @xmath627 . requiring that the family is non - degenerate is equivalent to requiring that there exists @xmath108 such that @xmath640 .
hence , if @xmath628 , @xmath612 , @xmath613 , is a non - degenerate one - parameter family of mixing unimodal maps through a good map @xmath627 then , by lemma [ l.nondegenerate ] and theorem [ t.main1 ] , we have almost sure typicality of the turning point for maps @xmath3 close to @xmath84 .
we prove lemma [ l.nondegenerate ] in section [ ss.unieins ] .
we verify first that there exists @xmath615 such that condition ( i ) holds for the map @xmath641 , @xmath616 .
we apply the criteria in lemma [ l.startcalc ] .
we consider first the case when @xmath649 for all @xmath650 .
it follows that @xmath651 for all @xmath650 and all @xmath6 sufficiently close to @xmath47 .
hence , @xmath641 is differentiable close to @xmath47 where @xmath652 is close to @xmath640 .
this implies that there exist @xmath615 and @xmath653 such that , for all @xmath616 , @xmath654 in ( [ eq.startcomp ] ) is to ensure that the partition points @xmath152 , @xmath153 , do not move faster than the points @xmath655 , @xmath542 , in lemma [ l.startcalc ] .
but in the setting here the turning point @xmath47 does not change in @xmath6 which implies that it is sufficient to have a constant @xmath653 instead of @xmath656 .
hence , by lemma [ l.startcalc ] , it follows that the one - parameter family @xmath3 , @xmath616 , together with the map @xmath641 satisfies condition ( i ) . regarding the verification of condition ( iii ) in section [ ss.cdreiuni ] we observe that the proof of lemma [ l.startcalc ] in fact implies that for either @xmath657 or @xmath658 ( @xmath659 is independent on the choice of @xmath55 and @xmath6 ) , @xmath660 for all @xmath542 and all @xmath616 for which the derivative @xmath661 exists .
concerning the verification of condition ( i ) , it is left to consider the case when @xmath47 is periodic for @xmath84 with prime period @xmath647 and @xmath662 .
we consider first the a - priori possible case when there is @xmath663 such that @xmath664 and @xmath665 .
let @xmath666 be minimal with this property .
since @xmath47 is periodic for @xmath84 with ( not necessarily prime ) period @xmath667 , we derive from ( [ eq.startcalc1 ] ) ( set @xmath668 ) the formula @xmath669 for @xmath670 . if @xmath671 is maximal such that @xmath672 , by equation ( [ eq.deg1 ] ) , we obtain that also @xmath673 .
thus , applying ( [ eq.startcalc1 ] ) with @xmath674 we derive that @xmath675 .
this in turn implies that we can assume that @xmath676 for all @xmath475 such that @xmath677 , which in turn implies that @xmath651 for all @xmath650 and all @xmath678 sufficiently close to @xmath47 .
hence , we are in a similar setting as in the first case above and we get that condition ( i ) is satisfied for the map
@xmath641 when @xmath6 is sufficiently close to @xmath47 .
this concludes the verification of condition ( i ) .
we turn to the proof of lemma [ l.nondegenerate ] . if @xmath47 is non - periodic for @xmath84 observe that @xmath679 if the family is non - degenerate then @xmath680 is growing in @xmath55 as @xmath681 and we find @xmath108 such that @xmath640 . in the other direction if @xmath640 for some @xmath108 , we have seen in the first part of this section that this implies that condition ( i ) is satisfied for @xmath6 close to @xmath47 . by ( [ eq.degstart ] )
, this shows that the family is non - degenerate . if @xmath47 has prime period @xmath647 then since @xmath682 it follows that @xmath683 .
since the map @xmath684 is good it follows that @xmath685 in equality ( [ eq.deg1 ] ) ( set @xmath686 ) is growing exponentially in @xmath573 , and we can find @xmath108 such that @xmath640 . on the other hand
if @xmath640 for some @xmath218 then we have seen above that this implies @xmath683 .
this concludes the proof of lemma [ l.nondegenerate ] .
[ r.good ] the fact that @xmath84 is good is only used to guarantee that @xmath685 ( when @xmath686 ) is growing to @xmath687 when @xmath573 increases .
we could replace the condition that @xmath84 is good by the requirement that @xmath688 as @xmath689 .
in section [ ss.cdreiuni ] it is shown that we can cover @xmath642 by a countable number of closed intervals @xmath18 such that condition ( iii ) is satisfied for the family @xmath3 , @xmath2 . since section [ ss.cdreiuni ] does not make use of condition ( ii ) , by lemma [ l.app1 ] , we can without loss of generality assume that there is a constant @xmath546 such that for each @xmath2 the density @xmath226 is bounded from above by @xmath274 and , further , there exists an interval @xmath547 of length @xmath548 such that @xmath226 restricted to @xmath547 is bounded from below by @xmath548 ( otherwise , disregarding a finite number of points , by lemma [ l.app1 ] , we can cover the interval @xmath18 by a countable number of closed subintervals on each of which this is true and then proceed with these subintervals instead of @xmath18 ) .
it follows that , there exists an integer @xmath690 such that @xmath691 . by @xcite and
since @xmath3 is mixing , we have that @xmath637\to[t_a(1),1]$ ] is exact , i.e. , for each set @xmath692 $ ] of positive lebesgue measure it follows that @xmath693\setminus t_a^j(s)|=0 $ ] .
observe that if @xmath627 and @xmath694 $ ] is an interval of length close to @xmath695 then we have @xmath696 $ ] .
thus , exactness implies that there is an integer @xmath182 such that @xmath697)=t_a^k([0,1/2c])=[t_a(1),1]$ ] .
since the image of an interval by @xmath33 , @xmath44 , changes continuously in @xmath6 and since the parameter interval @xmath18 is closed we can in fact choose the integer @xmath182 above independently on @xmath2 .
hence , for @xmath698 , we have @xmath699 $ ] , for all @xmath2 . by ( [ eq.perron22 ] )
, we get a uniform lower bound for the density .
this concludes the verification of condition ( ii ) .
as shown in section [ ss.unieins ] there exists @xmath700 such that condition ( i ) is satisfied for the family @xmath3 , @xmath701 $ ] , together with the map @xmath702 .
if @xmath47 is non - periodic for @xmath84 , we can choose @xmath703 so small that @xmath704 for all @xmath616 .
if @xmath47 is periodic for @xmath84 then we have shown in section [ ss.unieins ] that if @xmath705 such that @xmath706 then @xmath707 .
hence , there is @xmath703
such that , disregarding the point @xmath47 , we can cover the interval @xmath642 by a countable number of closed intervals @xmath18 such that for each such interval @xmath18 there exists a constant @xmath708 such that ( [ eq.j0 ] ) holds for all @xmath2 . henceforth , in the periodic case fix such an interval @xmath18 , and in the non - periodic case we use @xmath18 to denote the interval @xmath642 .
for @xmath2 , observe that if @xmath490 , @xmath44 ( note that @xmath168 is the partition on the phase space ) , then the image of @xmath52 by @xmath33 is of the form @xmath709 the choice of the integers @xmath182 and @xmath573 might not be unique due to the possibility that @xmath47 is periodic or pre - periodic for @xmath3 . if @xmath710 and @xmath711 , @xmath712 , we say that @xmath182 is _ minimal _ if the following holds .
if @xmath713 then @xmath182 is minimal if @xmath714 but @xmath715 for all @xmath716 . if @xmath717 or @xmath718 then @xmath182 is minimal if @xmath719 and @xmath720 , for all @xmath721
obviously for each element @xmath490 there exist unique minimal integers @xmath722 such that @xmath723 .
let @xmath199 , @xmath44 , be the partition in the parameter space associated to the map @xmath25 .
recall that @xmath724 . by ( [ eq.signconst ] )
, it follows that for all @xmath214 and @xmath191 we either have @xmath725 or @xmath726 , for all @xmath727 . without loss of generality
we assume that we are in the first case , i.e. , @xmath728 ( in the second case we would have to require @xmath729 instead of @xmath240 in the statement of condition ( iii ) . ) for @xmath2 , we claim that if @xmath490 , @xmath44 , and @xmath730 are chosen minimal such that @xmath723 then @xmath731 and @xmath732 are differentiable in @xmath6 and * if @xmath733 then @xmath734 ; and * if @xmath735 then @xmath736 .
we consider the case a ) .
since @xmath182 is chosen minimal we have that @xmath651 for all @xmath716 which ensures that the point @xmath731 is differentiable in @xmath6 . since @xmath731 is the left boundary point of the image @xmath56 , it is easy to verify that we must have @xmath737 ( for example use induction over @xmath55 ) . by ( [ eq.signchange ] )
, it follows that @xmath738 as claimed .
the case b ) is shown similarly . in order to verify condition ( iii ) we will first prove the following lemma .
[ l.helptwo ]
let @xmath239 and @xmath240 .
for all @xmath44 there is a mapping @xmath241 such that , for all @xmath242 , @xmath739 furthermore , for all @xmath242 we have the following .
let @xmath740 be the minimal integers such that @xmath741 then , we have @xmath742 and @xmath743 ; and , similarly , we have @xmath744 and @xmath745 .
we proof lemma [ l.helptwo ] by induction over @xmath55 .
observe that , by ( [ eq.j0 ] ) , it easily follows that for @xmath746 and for all @xmath239 , there is a bijection @xmath241 such that lemma [ l.helptwo ] is satisfied ( in this case the assumption @xmath240 is not necessary and one has always @xmath747 and @xmath748 ) .
thus , let @xmath749 and assume that the assertion of lemma [ l.helptwo ] holds for @xmath55 and all @xmath239 such that @xmath240 .
we will show that the assertion also holds for @xmath611 .
let @xmath239 , @xmath240 , and for @xmath750 and @xmath751 let @xmath752 be the corresponding minimally chosen integers . observe that by the induction assumption we have @xmath742 and @xmath744 .
we have @xmath753 $ and for all $ 1\le i<\max\{k_2,l_2\}$}.\ ] ] in particular this implies that the maps @xmath754 and @xmath755 are differentiable on the interval @xmath420 $ ] .
furthermore , we claim that if @xmath756 then @xmath757 , and if @xmath758 then @xmath759 . for @xmath412
$ ] , set @xmath760 , and let @xmath730 be the associated minimal integers for @xmath52 .
observe that since @xmath761 it follows from the induction assumption that @xmath762 and @xmath763 . by the minimality of @xmath182 and
@xmath573 it follows that @xmath651 for all @xmath764 .
thus , this implies ( [ eq.uuuni ] ) . if @xmath756 then ( [ eq.uuuni ] ) implies that @xmath420 $ ] is contained in an element of @xmath765 . by property
a ) and ( [ eq.signchange ] ) we have that @xmath766 . since condition ( i ) is satisfied we obtain that @xmath757 for all @xmath412 $ ] . by a similar argument using property b )
we obtain that @xmath759 for all @xmath412 $ ] , if @xmath758 .
we claim that @xmath767 $ , and } \\
\label{eq.b112 } \text{if $ t_{a_2}^{l_2}(0)<t_{a_1}^{l_1}(0)$},\quad & \text{then $ 0\not\in[t_{a_2}^{l_2}(0),t_{a_1}^{l_1}(0)]$}.\end{aligned}\ ] ] regarding ( [ eq.b111 ] ) , observe that if @xmath756 then the claim above implies that @xmath768 .
but by the induction assumption we have @xmath769 , and thus the case @xmath756 is not possible . for the case @xmath770 observe that , by ( [ eq.j0 ] ) , @xmath651 for all @xmath771 and @xmath2 . since the image of @xmath18 by the map @xmath754 contains @xmath772 $ ] and since @xmath743 , this shows ( [ eq.b111 ] ) . in a similar way one verifies ( [ eq.b112 ] ) .
it follows that if @xmath773 then @xmath774 and @xmath775 lie both either to the left or to the right of @xmath47 .
this implies that @xmath776 and @xmath777 are also elements of @xmath778 and @xmath779 , respectively .
we consider the case when @xmath774 and @xmath775 lie to the left of @xmath47 ( the other case is treated similarly ) .
let @xmath780 be the corresponding minimal integers for @xmath776 and @xmath777 ( seen as elements in the @xmath611-th partitions ) .
we immediately get that @xmath781 and @xmath782 . since by the induction assumption @xmath743 and since @xmath774 is to the left of @xmath47 where @xmath783 has positive slope , we have also @xmath784
the situation for the right boundary points is slightly more difficult .
first observe that if @xmath785 then @xmath786 if @xmath787 this follows immediately from ( [ eq.j0 ] ) .
if @xmath788 then we obtain from the claim above that @xmath789 .
since @xmath790 lies to the left of @xmath47 , this implies ( [ eq.uni111 ] ) . by the induction assumption
we get @xmath791 and it follows that @xmath792 .
the integer @xmath793 is equal to @xmath68 if @xmath794 , and @xmath795 otherwise . in both cases ,
since @xmath783 has positive slope on the left of @xmath47 , we obtain @xmath796 .
furthermore , @xmath797 holds .
hence , setting @xmath798 , we have shown that the assertion of lemma [ l.helptwo ] is satisfied in the case when @xmath773 . if @xmath799 , then each of the partitions @xmath800 and @xmath801 contain two elements @xmath802 and @xmath803 , respectively .
let @xmath804 and @xmath805 be the elements which are mapped after @xmath55 iterations , say , to the left of @xmath47 . looking at the to @xmath804 and @xmath805 corresponding minimal integers @xmath806 , we obviously have @xmath807 , and @xmath781 and @xmath782 . since by the induction assumption @xmath743 and since @xmath774 is to the left of @xmath47 where @xmath783 has positive slope , we have also @xmath808 .
an analogue situation appears for @xmath809 and @xmath810 .
thus , setting @xmath811 and @xmath812 , shows that the assertion of lemma [ l.helptwo ] in the case when @xmath799 .
the only case left is when @xmath813 but @xmath814 .
( observe that the case @xmath815 but @xmath816 is excluded by ( [ eq.b111 ] ) and ( [ eq.b112 ] ) . )
we have that @xmath776 is also an element of @xmath778 . without loss of generality
assume that @xmath774 lies to to the left of @xmath47 . if @xmath805 is the element in @xmath801 such that @xmath817 lies to the left of @xmath47 , then setting @xmath818 we derive with a similar reasoning as in the case when @xmath773 that the assertion of lemma [ l.helptwo ] is satisfied also in this last setting . to verify that condition ( iii )
is satisfied it is only left to show that there exists a constant @xmath238 such that properties ( [ eq.dreidist ] ) and ( [ eq.drei2 ] ) are fulfilled .
let @xmath239 , @xmath240 , take @xmath750 , @xmath44 , and set @xmath819 where @xmath820 is given by lemma [ l.helptwo ] .
let @xmath821 be the minimal integers corresponding to @xmath776 and @xmath777 , respectively .
regarding ( [ eq.dreidist ] ) , the only case to consider is when @xmath822 .
we consider the case when @xmath774 lies to the left of @xmath775 ( the other case is similar ) . by lemma [ l.helptwo ]
it follows that @xmath769 . by the claim in the proof of lemma [ l.helptwo ] the situation that @xmath823 is only possible
when @xmath770 .
it follows that @xmath824 which implies ( [ eq.dreidist ] ) .
regarding ( [ eq.drei2 ] ) , observe that if @xmath825 , then , since @xmath743 and @xmath745 , it follows immediately from the claim in the proof of lemma [ l.helptwo ] that @xmath774 is entirely contained in @xmath775 which implies ( [ eq.drei2 ] ) with the constant @xmath238 equal to @xmath68 .
furthermore , if @xmath826 then for all @xmath490 , @xmath2 , we have , by ( [ eq.j0 ] ) , that @xmath827 which implies ( [ eq.drei2 ] ) with a constant @xmath238 equal to @xmath828 .
so the only case left is when @xmath829 and @xmath542 .
assume that @xmath830 ( the case when @xmath831 is treated similarly ) .
consider the images @xmath832 and @xmath833 .
the boundary points of @xmath832 are @xmath834 and @xmath835 , and the boundary points of @xmath836 are @xmath68 and @xmath837
. obviously , @xmath835 and @xmath837 are the left boundary points . if @xmath838 then @xmath832 is contained in @xmath833 which , by the distortion estimate in lemma [ l.distortion ] , implies ( [ eq.drei2 ] ) with a constant @xmath238 equal to @xmath839 .
( note that lemma [ l.distortion ] requires that condition ( iii ) is satisfied .
but in its proof it is only needed that property ( [ eq.dreidist ] ) holds which we have already verified above . )
the case left is when @xmath840 . since @xmath841 it follows that @xmath842 and @xmath843 are not adjacent to @xmath47 for all @xmath844 .
hence , @xmath776 and @xmath777 are in fact also elements of @xmath845 and @xmath846 , respectively , and the minimal integers @xmath847 such that @xmath848 and @xmath849 are given by @xmath850 and @xmath851 .
thus , we can apply lemma [ l.helptwo ] and we obtain that @xmath852 . by the claim in the proof of lemma [ l.helptwo ] this implies that @xmath853 . by ( [ eq.j0 ] ) , it follows that @xmath854 and we obtain that @xmath855 .
hence , ( [ eq.drei2 ] ) is satisfied if we choose the constant @xmath238 equally to @xmath828 .
this concludes the verification of condition ( iii ) for the family @xmath3 , @xmath2 .
in this section we apply the results of section [ s.unimodal ] to an important family of piecewise expanding unimodal maps , the so - called skew tent maps .
we consider particular one - parameter families of skew tent maps for which it is easy to make general statements .
however , given a concrete one - parameter family of skew tent maps which does not fit into the families considered below , by theorem [ t.main1 ] , it is straightforward to check if one has almost sure typicality .
we will use the same representation as in @xcite , i.e. , we define the skew tent map with slopes @xmath112 and @xmath113 where @xmath114 , by the formula @xmath856 the turning point of @xmath115 is @xmath47 , @xmath857 and , by lemma 3.1 in @xcite , if @xmath116 then the interval @xmath858(=[1-\be,1])$ ] is invariant under @xmath115 ( if @xmath859 then there exists no invariant interval of finite positive length ) . for two parameter couples @xmath860 and @xmath861 we take the same order relation as the one which appears in @xcite , i.e. , we shall write @xmath862 if @xmath863 , @xmath864 , and at least one of these inequalities is sharp . fix @xmath117 and @xmath118 in the set @xmath119 such that @xmath865 . let @xmath122\to[\al_0,\al_1]\quad\text{and}\quad\be:[0,1]\to[\be_0,\be_1]\ ] ] be functions in @xmath123)$ ] such that @xmath124 , @xmath125 , and , for all @xmath126 $ ] , if @xmath127 then @xmath128 and if @xmath129 then @xmath130 . observe that @xmath866 , and @xmath867 , for all @xmath126 $ ] .
we define the one - parameter family @xmath3 as the family of skew tent maps given by @xmath868\to[t_{\al(a),\be(a)}(1),1],\quad\ a\in[0,1].\ ] ] observe that the family @xmath3 , @xmath126 $ ] , is a one - parameter family of piecewise expanding unimodal maps as described in section [ s.unimodal ] where we can set the constant @xmath869 equal to @xmath870}\be(a)-1 $ ] .
the main statement of this section is the following .
[ t.tent ] for a.e .
parameter @xmath126 $ ] the turning point @xmath47 is typical for the a.c.i.p . @xmath4 . for other points than the turning point , i.e. , given a @xmath15 function @xmath871\to\real$ ] ( such that @xmath872 $ ] ) , it is sufficient to check condition ( i ) in order to obtain a.s .
typicality for @xmath512 .
[ c.tent1 ] if the one - parameter family @xmath3 , @xmath126 $ ] , with the associated map @xmath873 satisfies condition ( i ) , then @xmath248 is typical for @xmath4 , for a.e . @xmath126 $ ] .
the following comments show that we can restrict ourself to non - renormalizable skew tent maps .
if @xmath874 then @xmath115 is renormalizable , see , e.g. , @xcite .
more precisely if ( [ eq.unirenorm ] ) holds , then @xmath875 is greater or equal than the unique fixed point in @xmath876 and @xmath877 restricted either to the interval @xmath878 $ ] or to the interval @xmath879 $ ] is affinely conjugated to @xmath880 restricted to the interval @xmath881 $ ] .
observe that the new slopes @xmath882 and @xmath883 still satisfy @xmath884 and @xmath885 ( the latter inequality follows by ( [ eq.unirenorm ] ) ) . since the function @xmath886 is decreasing for @xmath887 , we have that if @xmath684 is not renormalizable then @xmath3 is also not renormalizable for @xmath126 $ ] .
now , assume for the moment that @xmath3 is renormalizable for each @xmath126 $ ] and consider the one - parameter family defined by @xmath888 .
note that if we show typicality of the turning point for the family @xmath889 , for a.e .
@xmath126 $ ] , this implies a.s .
typicality of the turning point for the original family @xmath3 . furthermore ,
if condition ( i ) is satisfied for the family @xmath3 together with a map @xmath512 as in corollary [ c.tent1 ] , then it is easy to check that condition ( i ) is also satisfied for the family @xmath889 together with the map @xmath890 which is the map for the conjugated system corresponding to @xmath512 . since the @xmath6-derivative of @xmath891 is positive and the @xmath6-derivative of @xmath892 is non - negative , the new one - parameter family @xmath889 fits into the family of skew tent maps described in the beginning of this section .
furthermore , it is known that for each @xmath126 $ ] , @xmath3 is at most a finite number of times renormalizable where this number is bounded above by a constant only dependent on @xmath117 and not on the parameter @xmath6 ( this can easily be derived by looking , e.g. , at the topological entropy of @xmath3 , see @xcite page 137 ) .
altogether , we derive that in order to prove theorem [ t.tent ] ( and therewith also corollary [ c.tent1 ] ) we can without loss of generality restrict ourself to the case when @xmath3 , @xmath126 $ ] , is not renormalizable , i.e. , we assume that @xmath893 observe that it is only possible for the parameter @xmath894 to satisfy the equality @xmath895 .
thus , since we are only interested in lebesgue almost every parameter we can neglect skew tent maps whose slopes satisfy @xmath896 , i.e. , we assume that @xmath897 since we can restrict ourself to skew tent maps which are non - renormalizable , this immediately implies that these maps are mixing .
hence , we can apply theorem [ t.main1 ] and in order to prove theorem [ t.tent ] it is sufficient to show that there exists an iteration @xmath108 such that latexmath:[\[\label{eq.tentto } for almost all @xmath126 $ ] . the main computation needed for the verification of ( [ eq.tentto ] )
is already done in a paper by misiurewicz and visinescu @xcite ( see lemma 3.3 and 3.4 therein ; and proposition [ p.mv ] below ) , where they show monotonicity of the kneading sequence for skew tent maps .
( a reader not familiar with the basic notions and facts of kneading theory can find them in @xcite . ) the main result in @xcite ( see theorem a therein ) is the following .
[ t.monotonicity ] let @xmath860 and @xmath861 be in the set @xmath119 .
if @xmath862 then the kneading sequence of @xmath899 is strictly greater than the kneading sequence of @xmath115 . since the derivatives of @xmath900 and @xmath901 are non - negative and at least one of them is positive , we obtain strict monotonicity of the kneading sequence for our family @xmath3 , @xmath126 $ ] .
let @xmath44 .
note that if @xmath902 for some @xmath903 , then the kneading sequence of @xmath3 ends with @xmath274 and has length smaller than @xmath55 . by the strict monotonicity of the kneading sequence
, @xmath3 can have such a kneading sequence only for finitely many parameter values @xmath126 $ ] .
hence , for each @xmath44 , there are only a finite number of @xmath6 values such that @xmath904 is not differentiable in @xmath6 . to establish ( [ eq.tentto ] ) we will use some derivative estimates given in @xcite . to this end
we will look at the kneading sequences of the @xmath3 s .
every kneading sequence of a map @xmath3 in our family starts with @xmath905 and is smaller or equal than the sequence @xmath906 .
in fact , by ( [ eq.notlog2 ] ) and by the monotonicity of the kneading sequence , the kneading sequence of @xmath907 is strictly smaller than @xmath906 .
let @xmath908 be the integer such that the kneading sequence of @xmath909 starts with @xmath910 or is equal to @xmath911 .
from @xcite we derive the following result .
[ p.mv ] there exists a constant @xmath912 such that for all @xmath126 $ ] for which @xmath904 , @xmath913 , is differentiable in @xmath6 we have @xmath914},\ ] ] and , furthermore , @xmath915 the proof of proposition [ p.mv ] follows from lemma 3.3 and 3.4 in @xcite . for this note
that for each @xmath126 $ ] , the integer @xmath916 such that the kneading sequence of @xmath3 starts with @xmath917 or is equal to @xmath918 is smaller or equal than @xmath919 .
observe also that @xmath42 in @xcite corresponds to @xmath920 in our setting .
actually , lemma 3.4 in @xcite is only formulated for the case when @xmath921 . but considering lemma 3.4 i ) in @xcite it is easy to deduce that proposition [ p.mv ] also holds when @xmath922 . for @xmath913 , we have @xmath923 for all @xmath6 for which @xmath924 is differentiable .
since @xmath925 and at least one of the derivatives @xmath926 and @xmath927 is uniformly bounded away from @xmath47 , by ( [ eq.mv1 ] ) and ( [ eq.mv2 ] ) , @xmath928}|d_at_a^j(0)|$ ] is growing exponentially in @xmath55 .
thus , we can fix an integer @xmath108 such that ( [ eq.tentto ] ) is satisfied which concludes the proof of theorem [ t.tent ] .
assume that we have a one - parameter family @xmath1\to[0,1]$ ] , @xmath2 , as described in section [ ss.prel ] with a partition @xmath929 and satisfying properties ( i)-(iii ) .
we require additionally that the family @xmath3 fulfills the following markov property .
set @xmath930 , @xmath156 .
* for each @xmath156 the image @xmath931 , @xmath2 , is a union of monotonicity intervals @xmath932 , @xmath400 ( modulo a finite number of points ) . [ t.markov ] if the one - parameter family @xmath3 , @xmath2 , satisfies the markov property ( m ) and if for a @xmath15 map @xmath16 $ ] condition ( i ) is fulfilled , then @xmath17 is typical for @xmath4 , for a.e . @xmath2 .
[ ex.markov ] let @xmath933 and @xmath142\to[0,1]$ ] a @xmath8 homeomorphism such that @xmath934 and such that the set @xmath935 is non - empty .
clearly , @xmath18 is an ( open ) interval .
we define the one - parameter family @xmath1\to[0,1]$ ] as @xmath936 by @xcite , since @xmath3 has only one point of discontinuity , there exists a unique a.c.i.p .
@xmath4 . from the verification of condition ( ii ) in the proof of theorem [ t.markov ] , it will follow that @xmath937 $ ] .
[ p.markov ] if @xmath938 is a @xmath15 map such that @xmath939 , then @xmath248 is typical for @xmath4 , for a.e .
parameter @xmath2 .
le @xmath286 be a closed interval .
it follows that there exist constants @xmath148 which are uniform lower and upper bounds for the expansion in the family @xmath3 , @xmath299 ( cf .
( [ eq.la ] ) ) .
since the map @xmath940 is @xmath8 it follows that there exists @xmath941 such that @xmath3 is piecewise @xmath942 for all @xmath299 .
one easily checks that the one - parameter family @xmath3 , @xmath299 , fits into the model described in section [ ss.prel ] satisfying properties ( i)-(iii ) .
hence , we can apply theorem [ t.markov ] to this family .
clearly , @xmath3 satisfies the markov property ( m ) . in order to show a.s .
typicality , it is only left to verify condition ( i ) .
we will use the criteria in lemma [ l.startcalc ] .
we have @xmath943 which is non - positive for all @xmath299 .
since @xmath287 is a closed interval , the image of @xmath287 by @xmath512 has positive distance to @xmath47 and @xmath68 and , hence , there exists @xmath653 such that @xmath944 .
recall the formula ( [ eq.startcalc1 ] ) for the derivative of @xmath247 ( set @xmath253 ) .
all the terms in the right - hand side of ( [ eq.startcalc ] ) are non - positive and the term @xmath945 is decreasing faster than @xmath946 .
thus , we find @xmath218 such that ( [ eq.startcalc1 ] ) is satisfied for all parameter values @xmath299 in which @xmath223 is differentiable . in order to apply lemma [ l.startcalc ] ,
it is only left to show that the number of @xmath299 in which @xmath223 is not differentiable is finite .
since @xmath939 and since the point of discontinuity @xmath947 of @xmath3 is strictly increasing in @xmath6 , there can only be one point in the interval @xmath287 in which @xmath948 is not differentiable . since @xmath949 , @xmath44 , we can repeat this reasoning and it follows that there are only finitely many points in which @xmath223 is not differentiable . since we can cover @xmath18 by a countable number of such intervals @xmath287 , this concludes the proof of the proposition [ p.markov ] .
we turn to the proof of theorem [ t.markov ] . in order to proof theorem [ t.markov ] , it is sufficient to verify conditions ( ii ) and ( iii ) .
we consider first condition ( ii ) . as in the first paragraph in section [ ss.zweib ] , by lemma [ l.app1 ] , we can without loss of generality assume that there is a constant @xmath546 such that for each @xmath2 the density @xmath226 is bounded from above by @xmath274 and , further , there exists an interval @xmath547 of length @xmath548 such that @xmath226 restricted to @xmath547 is bounded from below by @xmath548 . since for each @xmath2
the expansion of @xmath3 is at least @xmath12 , we derive that there is an integer @xmath557 independent on @xmath6 such that the number of elements in @xmath950 is greater or equal than @xmath951 . by ( m ) , we derive that the image by @xmath559 of an element @xmath952 , which is not adjacent to a boundary point of @xmath547 , is a monotonicity interval @xmath932 , @xmath400 . by our assumption on the one - parameter family @xmath3 , the measure
@xmath4 is ergodic .
it follows that there is an integer @xmath953 such that @xmath954 .
( obviously @xmath955 can be chosen independently on @xmath2 and the monotonicity interval @xmath956 . ) thus , setting @xmath957 for almost every @xmath591 , there exists a point @xmath958 such that @xmath13 is mapped to @xmath959 after @xmath55 iterations , i.e. , @xmath960 .
now , inequality ( [ eq.perron22 ] ) provides us with a lower bound for the density .
note that from this argument follows that @xmath937 $ ] in example [ ex.markov ] . to verify ( iii ) we observe that , since in this markov setting @xmath161 , @xmath2 , is the union of monotonicity intervals @xmath932 , @xmath400 , there exists even a bijection @xmath241 for all @xmath239 and @xmath44 , satisfying ( [ eq.drei1 ] ) . since for each element @xmath490 the image
@xmath56 is a union of monotonicity intervals @xmath932 , @xmath400 , and since by property ( i ) in section [ ss.prel ] the boundary points of @xmath932 are lipschitz continuous in @xmath6 and @xmath961 , we get that also ( [ eq.dreidist ] ) and ( [ eq.drei2 ] ) are satisfied where we can take @xmath962 . [
[ a : appendix ] ] [ l.app1 ] let @xmath1\to[0,1]$ ] , @xmath2 , be a one - parameter family as described in section [ ss.prel ] , satisfying properties ( i)-(iii ) and condition ( iii )
. disregarding a finite number of parameters in @xmath18 , we can cover @xmath18 by a countable number of intervals @xmath286 such that on each interval @xmath287 the following holds .
there exists a constant @xmath963 such that for each @xmath299 the density @xmath226 of @xmath4 is bounded above by @xmath274 and , further , there exists an interval @xmath964 $ ] of size @xmath548 such that @xmath226 restricted to @xmath547 is bounded from below by @xmath548 . for each @xmath2 it follows from @xcite p.496 line 5 and @xcite p.484 line 6 , that the variation over the unit interval of the density @xmath226 is bounded above by a constant @xmath965 where the integer @xmath381 is chosen so large that @xmath966 and the number @xmath967 is given by @xmath968 ( in @xcite and @xcite @xmath967
is the minimal size of the monotonicity intervals for the map @xmath969\to[0,1]$ ] .
but since the elements of @xmath970 are monotonicity intervals for the map @xmath971 , the constant @xmath972 is greater or equal than the corresponding constant in @xcite and @xcite . ) for @xmath44 , let @xmath973 be a sequence of symbols @xmath974 , @xmath189 .
if @xmath975 is a parameter value such that there exists an element @xmath976 satisfying @xmath977 then there is a neighborhood @xmath978 of @xmath979 in @xmath18 such that for all @xmath980 there is an element @xmath981 having the same combinatorics as @xmath982 , i.e. , @xmath983 , @xmath189 .
furthermore , the boundary points of @xmath984 depend continuously on @xmath980 .
we prove the claim by induction over @xmath44 .
we do not make use of condition ( iii ) .
for @xmath599 the elements in @xmath175 corresponding to the symbols @xmath985 are the intervals @xmath164 , @xmath167 .
the boundary points of these intervals are , by property ( iii ) , continuous functions on @xmath18 . using properties ( i ) and ( ii )
, one can easily show that the boundary points of @xmath986 are continuous on @xmath18 .
now , assume that the statement holds for some @xmath44 .
fix a sequence @xmath987 of symbols in @xmath988 .
let @xmath975 be a parameter such that there exists an element @xmath989 satisfying @xmath990 , for all @xmath608 ( if there is no such a parameter @xmath979 for which the element @xmath982 exists then there is nothing to show ) .
let @xmath991 be the element containing @xmath982 .
by the induction assumption there exists a neighborhood @xmath992 of @xmath979 in @xmath18 such that for all @xmath993 there is an element @xmath994 having the same combinatorics as @xmath995 and the boundary points of @xmath996 and @xmath997 depend continuously on @xmath993 .
note that if @xmath998 is a boundary point of @xmath999 then it is either equal to a partition point @xmath1000 , @xmath153 , or it is a boundary point of @xmath1001 . by the continuity of the boundary points of @xmath997 on @xmath992 and the continuity of @xmath1002
, we deduce that there exists a neighborhood @xmath1003 of @xmath979 in @xmath18 such that for each @xmath980 there exists an element @xmath1004 having the same combinatorics as @xmath982 . since the boundary points of @xmath984 are continuous on @xmath978 , we can once more apply properties ( i ) and ( ii ) to deduce that also the boundary points of @xmath1005 are continuous on @xmath978 .
let @xmath973 be a sequence of symbols @xmath974 .
if for a parameter @xmath975 there exists an element @xmath976 which corresponds to this sequence of symbols then , by condition ( iii ) , for each @xmath1006 there exists an element @xmath981 corresponding to this sequence of symbols .
furthermore , @xmath1007 is a lower bound for the size of this element which is by the claim above continuous in @xmath6 .
this implies that there is a map @xmath1008 $ ] which is piecewise continuous with only a finite number of discontinuities and @xmath1009 .
hence , disregarding a finite number of parameter values in @xmath18 , we can cover @xmath18 by a countable number of intervals @xmath286 such that for each such interval @xmath287 there is a constant @xmath1010 such that @xmath1011 for all @xmath299 .
it follows that that there is a constant @xmath1012 such that the variation of @xmath226 is bounded from above by @xmath1013 for all @xmath299 . since @xmath1014
, this immediately implies that @xmath226 is bounded from above by @xmath1015 . to establish a lower bound on a subinterval of @xmath161 , we observe the following .
if the variation over @xmath29 $ ] of a function @xmath1016\to\real_+$ ] is bounded from above by a constant @xmath1017 , and if @xmath1018 , then there exists an interval @xmath61 of length @xmath1019 such that @xmath1020 for all @xmath1021 .
let @xmath1022 $ ] , divide the unit interval into @xmath1023 disjoint intervals @xmath1024 of length @xmath1025 , and , for @xmath1026 , set @xmath1027 and @xmath1028 . since @xmath1029 , it follows that @xmath1030 .
if @xmath1031 , for all @xmath1026 , it would follow that the variation of @xmath1032 is strictly greater than @xmath1033 , where the last inequality follows since @xmath1017 .
hence , at least for one @xmath1026 , @xmath1034 .
setting @xmath1035 this concludes the proof of lemma [ l.app1 ] .
i am grateful to m. benedicks and k. bjerklv for many fruitful discussions and for encouraging and supporting me during the writing of this paper .
a part of this paper was written at the institut mittag - leffler ( djursholm , sweden ) .
i thank tomas persson for pointing out that in condition ( ii ) it is sufficient to require that the constant @xmath238 depends measurably on the parameter @xmath6 ( cf .
remark [ r.tomas ] ) .
ds v. baladi , _ on the susceptibility function of piecewise expanding interval maps _ , comm . math .
* 275 * ( 2007 ) , 839859 . v.
baladi and d. smania , , nonlinearity * 21 * ( 2008 ) , no .
4 , 677711 . v. baladi and d. smania , , discrete contin .
* 23 * ( 2009 ) , no .
3 , 685703 m. benedicks and l. carleson , , ann . of math . ( 2 ) * 122 * ( 1985 ) , no .
1 , 125 . m. bjrklund and d. schnellmann , , indag . math .
( n.s . ) * 20 * ( 2009 ) , no .
2 , 167177 .
k. brucks and m. misiurewicz , , ergodic theory dynam .
systems * 16 * ( 1996 ) , no . 6 , 11731183 .
h. bruin , , fund .
* 155 * ( 1998 ) , no .
3 , 215235 . p.
collet and j .- p .
eckmann , `` iterated maps on the interval as dynamical systems '' , birkhuser , boston , 1980 .
b. faller and c .- e .
pfister , , ergodic theory dynam .
systems * 29 * ( 2009 ) , no . 5 , 15291547 .
a. lasota and j.a .
yorke , , trans .
* 186 * ( 1973 ) , 481488 .
li and j.a .
yorke , , trans .
* 235 * ( 1978 ) , 183192 .
p. mattila , `` geometry of sets and measures in euclidean spaces ( cambridge studies in advanced mathematics ) '' , cambridge university press , cambridge , 1995 .
m. misiurewicz and e. visinescu , , ann .
h. poincar probab .
statist . *
27 * ( 1991 ) , no .
1 , 125140 . v.a .
rohlin , , ( russian ) izv .
nauk sssr ser . mat . 25 ( 1961 ) , 499530 .
( english ) amer .
ser . 2 , * 39 * ( 1964 ) , 136
. j. schmeling , , ergodic theory dynam .
systems * 17 * ( 1997 ) , no .
3 , 675694 .
m. tsujii , , invent .
* 111 * ( 1993 ) , no .
1 , 112137
. g. wagner , , z. wahrsch
gebiete * 46 * ( 1979 ) , no .
3 , 317324 . s. wong , , trans .
* 246 * ( 1978 ) , 493500 . | let @xmath0 be an interval and @xmath1\to[0,1]$ ] , @xmath2 , a one - parameter family of piecewise expanding maps such that for each @xmath2 the map @xmath3 admits a unique absolutely continuous invariant probability measure @xmath4 .
we establish sufficient conditions on such a one - parameter family such that a given point @xmath5 $ ] is typical for @xmath4 for a full lebesgue measure set of parameters @xmath6 , i.e. , @xmath7 for lebesgue almost every @xmath2 .
in particular , we consider @xmath8-versions of @xmath9-transformations , piecewise expanding unimodal maps , and markov structure preserving one - parameter families . for families of piecewise expanding unimodal maps
we show that the turning point is almost surely typical whenever the family is transversal .
daniel schnellmann |
originally discovered by @xcite and @xcite , the magnetorotational instability ( mri ) was suggested by ( * ? ? ?
* bh91 hereafter ) to be the physical mechanism driving the redistribution of angular momentum required for the accretion process in keplerian discs orbiting compact objects ( see , e.g. , * ? ? ?
* for a review ) .
keplerian discs have a positive radial gradient in angular momentum , and therefore are linearly ( rayleigh-)stable .
purely hydrodynamic perturbations are unlikely to grow to amplitudes at which the associated stresses can account for efficient angular - momentum transport . in the presence of a weak magnetic field , however , a negative radial gradient in the angular velocity of the disc is mri - unstable , and seed perturbations can grow exponentially on time scales close to the rotational period . during this phase ,
_ channel modes _ develop .
channel modes are pairs of coherent radial up- and downflows stacked vertically and threaded by layers of magnetic field of alternating radial and azimuthal polarity . in these modes , the magnetic tension ( maxwell stress tensor )
transports angular momentum along the field lines from the inner parts of the disc outwards .
the criterion for the mri onset can be formulated in a rather simple manner , even if the thermal stratification ( gradients of entropy or molecular weight ) and non - ideal effects ( viscosity , resistivity ) are included @xcite .
this allows for its application beyond keplerian discs , in particular to proto - neutron stars ( pnss ) resulting from the core - collapse of rotating massive stars .
simplified simulations by @xcite showed that such pnss possess regions in which the mri can grow on shorter time - scales than the time between the bounce and the successful explosion .
this finding , later confirmed in multi - dimensional models ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , presents the possibility of generating strong magnetic fields that can tap the rotational energy of the core , power magnetohydrodynamics ( mhd ) turbulence @xcite , and become a potentially important ingredient in rapidly - rotating core - collapse supernovae ( ccsne ) .
how much these systems are affected by the mri crucially depends on both its growth rate and on the final amplitude of the seed perturbations .
we can give an upper limit by assuming that the mri ceases to grow once the magnetic field comes close ( within a factor of @xmath3 ) to equipartition with the energy of the differential rotation .
in ccsne , this would correspond to dynamically important field strengths up to @xmath4 .
similar energetic arguments can be used to express the mri - generated stresses in the framework of @xmath5disc models @xcite .
this estimate neglects possible effects quenching the mri before it reaches its maximally - allowed energy , and the effects of buoyancy as shown by @xcite , who performed mri simulations in the presence of buoyancy .
they showed that the termination amplitude ( at the end of the exponential growth ) is not necessarily correlated to the magnetic field strength in the turbulent state that followed .
the physics of the termination of the mri growth remains an active field of research with many studies devoted to finding the value of the @xmath3 parameter for the stress tensor .
we refer , among others , to the works of @xcite .
the model of parasitic instabilities by ( * ? ? ?
* gx94 hereafter ) , further studied and developed by @xcite , @xcite and @xcite provides a clear physical picture of the termination mechanism .
the mri channel modes are characterized by a shear layer and a current sheet in the vertical profiles of velocity and magnetic field , respectively .
hence , the ( laminar ) channel flows can be unstable against _ secondary _ ( or _ parasitic _ ) _ instabilities _ of kelvin - helmholtz ( kh ) or tearing - mode ( tm ) type .
initially , the role of the parasites is negligible , as they grow much more slowly than the mri .
however , since the growth rate of the secondary instabilities is proportional to the channel mode amplitude , it is clear that at some stage the parasites will grow faster than the mri channels whose growth rate is constant , whereas the parasites grow exponentially with time .
roughly at this point , the parasitic instabilities should disrupt the channel modes and terminate the mri growth , marking the transition to the turbulent saturation phase @xcite .
a further discussion of the mri saturated state is beyond the scope of this paper .
@xcite analytically studied the mri termination in resistive - viscous mhd by solving simplified model equations for the evolution of the parasitic instabilities .
he identified different parameter - space regimes where , depending on hydrodynamic and magnetic reynolds numbers , either the kh instability or the tm is the dominant ( i.e. faster developing ) secondary instability .
@xcite found that the magnetic field at the surface of the pns can be enhanced w.r.t . the interior regions . if this is also the case for rotating cores , mhd phenomena ( like , e.g. the mri ) should be most prominent at the pns surface . in this region ,
the reynolds numbers are large if the surface is located above the neutrinosphere , which however may not always be the case ( see fig.10 in * ? ? ?
* ) . in this paper
, we only investigate the regime of very high reynolds numbers in which , according to the parasitic model , the mri should be terminated by the kh instability .
the parasitic model has not been tested with direct global numerical simulations of the mri in ccsne .
@xcite found in their semi - global 2d ideal mhd simulations that , because of numerical resistivity , the mri was terminated by tms . in their 3d simulations the mri was terminated by non - axisymmetric parasitic instabilities , although a clear identification of their nature was not possible . to test the predictions of @xcite
, we performed a set of 2d and 3d resistive - viscous mhd simulations of the mri .
given initial conditions , we studied in particular the importance of non - ideal effects by varying the values of both the uniform viscosity and uniform resistivity , which influence the growth of kh modes and of tms , respectively . in the case of the kh instability , which is present already in ideal hydrodynamics , viscosity and resistivity
alter the properties of the unstable modes merely quantitatively .
on the other hand , tm are essentially driven by resistivity , i.e. they do not grow in ideal mhd , while viscosity plays a minor role by changing the properties of the unstable modes only quantitatively .
since we are mainly concerned in this work with identifying the type of the parasitic instability that limits the growth of the mri in particular models , we performed all our simulations ( except for one control model ) with a non - zero resistivity , but simulated models with both zero and non - zero viscosity .
we defer a more thorough quantitative analysis of the influence of non - ideal effects to a subsequent work . in sec .
[ sec : mri_theory ] we give the criterion for the onset of the mri , and we describe the initial stage of the instability during which channel modes develop .
next , we discuss possible scenarios of its termination , in particular , termination via the parasitic instabilities . in sec .
[ sec : numerics ] we describe the numerical code and the initial setup used in our 2d and 3d simulations .
we present the results of these simulations in sec .
[ sec : results ] , and summarize our findings in sec .
[ sec : summary ] .
we consider flows that can be described by the equations of resistive - viscous ( non - ideal ) mhd . in the presence of an external gravitational potential , @xmath6 ,
these equations read @xmath7 & = -\rho { { \bf{v } } } \cdot \nabla \phi , \\ \partial_t { { \bf{b } } } + { \nabla \cdot}\left({{\bf{v}}}\otimes{{\bf{b } } } - { { \bf{b}}}\otimes{{\bf{v } } } \right ) & = \eta \nabla^2 { { \bf{b } } } , \\ { \nabla \cdot}{{\bf{b } } } & = 0 , \label{eq : divb}\end{aligned}\ ] ] where @xmath8 , @xmath9 , @xmath10 , and @xmath11 are the fluid velocity , the density , a uniform resistivity , and the redefined magnetic field @xmath12 , respectively . the total energy density , @xmath13 , is composed of fluid and magnetic contributions , i.e. @xmath14 with the internal energy density @xmath15 and the gas pressure @xmath16 . the stress tensor @xmath17 is given by @xmath18 { { \bf{i } } } - { { \bf{b } } } \otimes { { \bf{b } } } - \rho \nu \left [ \nabla { { \bf{v } } } + ( \nabla { { \bf{v}}})^t \right],\ ] ] where @xmath19 is the unit tensor , and @xmath20 and @xmath21 are the kinematic shear and bulk viscosity , respectively .
we study the mri in a small portion of the rotating star at a given distance @xmath22 from the rotation axis , embedded in a magnetic field . for convenience
, we use cylindrical coordinates @xmath23 , hereafter .
we restrict our analysis to locations close to the equatorial plane ( @xmath24 ) and vertical perturbation wavevectors for which the mri is known to develop fastest ( see , e.g. * ? ? ?
* ) . in this case
, we can consider a differentially rotating fluid with angular velocity @xmath25 and linear velocity @xmath26 threaded by a uniform vertical magnetic field @xmath27 in the local perturbation analysis .
here , @xmath28 and @xmath29 are the unit vectors in @xmath30 and @xmath31 direction , respectively . with these assumptions and ignoring dissipative effects , the mri instability criterion is ( c.f . * ? ? ?
* ) @xmath32 where @xmath33 are the square of the brunt - visl frequency and the epicyclic frequency , respectively , and @xmath34 is the adiabatic index .
we consider an angular velocity with a radial dependence of the form @xmath35 where @xmath36 is the angular velocity at the characteristic radius @xmath37 , and @xmath38 is the local rotational shear given by @xmath39 the rotation profile is quite generic for astrophysical systems , e.g. for @xmath40 , one recovers the keplerian profile and for differentially rotating stars typical values of @xmath38 are in the range @xmath41 . the corresponding epicyclic frequency ( i.e. the radial oscillation frequency ) is @xmath42 which vanishes at @xmath43 ( rayleigh stability criterion limit for unmagnetised rotating fluids ) .
we also assume that the entropy and composition are constant within the simulated volume , and hence @xmath44 .
the influence of entropy gradients on the mri has been studied by @xcite and will not be discussed here . if condition ( [ eq : mri_instability ] ) , is fulfilled , i.e. @xmath45 , any perturbation of the form @xmath46 ( wkb ansatz ) is unstable for wavenumbers @xmath47 where @xmath48 is the alfvn speed in the vertical direction .
bh91 considered small perturbations of the velocity , @xmath49 , and the magnetic field , @xmath50 , in a system whose velocity and magnetic field are given by eqs.([eq : v_init ] ) and ( [ eq : b_init ] ) , respectively , i.e. @xmath51 we note that one must also introduce appropriate perturbations of the other thermodynamical quantities to fulfil the mhd eqs.([eq : cont ] ) - ( [ eq : divb ] ) , but we do not mention them here explicitly . for the linearized ideal mhd equations in the incompressible limit , bh91 found unstable solutions , which are usually referred to as _ mri channels _ , @xmath52 where the subscript _ c
_ stands for _ channel _ , @xmath53 is the unit vector in @xmath22 direction , @xmath54 and @xmath55 are the initial amplitudes , @xmath56 and @xmath57 are the angles between the @xmath22-axis and the direction of the velocity and magnetic field channels , respectively .
the wavenumbers @xmath58 correspond to values of @xmath59 fulfilling eq .
( [ eq : mri_kcrit ] ) . to simplify the notation
, we define @xmath60 and for brevity , we will often drop the explicit time dependence , i.e. @xmath61 and @xmath62 . gx94
generalized the results of bh91 , and they showed that the mri channels are an exact solution of the ideal incompressible mhd equations in the _ shearing sheet _ ( local ) approximation .
this approximation consists in transforming the equations to a frame corotating with a fiducial fluid element and linearising the rotational profile around a radius @xmath37 , i.e. @xmath63 . in this frame ,
the gravitational force and the centrifugal force balance each other for initially keplerian accretion discs , but the coriolis force has to be taken into account . in differentially rotating stars ,
additional pressure gradients are necessary to provide an equilibrium . in the ideal mhd limit ,
the mri growth rate and the channel angles , @xmath56 and @xmath57 , are given by @xmath64 and @xmath65 , \label{eq : v_b_pi1 } \\
\phi_b & = \phi_v + \frac{\pi}{2 } , \label{eq : v_b_pi } \end{aligned}\ ] ] respectively .
the amplitude ratio @xmath66 is a function of @xmath58 and @xmath38 ( cf .
* pc08 hereafter ) .
the mode with the wavenumber @xmath67 grows fastest at a rate @xmath68 note that we use capital letters in the subscripts to refer to the fastest - growing mode ( @xmath69 and @xmath70 ) to distinguish it from generic unstable modes with growth rates @xmath71 . for the fastest - growing mode ,
the magnetic field and the velocity amplitudes are related by @xmath72 where @xmath73 is the alfvn speed parallel to the mri channel , and the channel angles are @xmath74 and @xmath75 .
the properties of mri modes change when viscosity or resistivity are present in the system .
pc08 generalized the results of gx94 and showed that mri channels ( eqs .
( [ eq : channel_v ] ) and ( [ eq : channel_b ] ) ) are also exact solutions of the resistive - viscous incompressible mhd equations in the shearing sheet approximation .
they derived expressions for the growth rate @xmath76 , the amplitude ratio @xmath77 , and the channel angles @xmath78 and @xmath57 of mri - unstable modes for arbitrary hydrodynamic and magnetic reynolds numbers .
following pc08 , we define these two ( dimensionless ) numbers as @xmath79 we note that in the case of a zero entropy gradient the ( most general ) dispersion relation of @xcite reduces to the one analysed by pc08 . since we limit our studies to isentropic models , we can directly apply their results . in resistive - viscous mhd ,
the channel angles , @xmath78 and @xmath57 , are given by @xmath80 , \label{eq : phi_v } \\ \label{eq : phi_b } \phi_b & = \arctan \left [ \frac{- \omega_0 \ { 2 ( { \gamma_{\mathsmaller{\mathrm{mri}}}}+ { k_{\mathsmaller{\mathrm{mri}}}}^2 \eta ) + q ( \nu - \eta ) { k_{\mathsmaller{\mathrm{mri}}}}^2 \ } } { { c_{\mathrm{a } z}}^2 { k_{\mathsmaller{\mathrm{mri}}}}^2 + ( { \gamma_{\mathsmaller{\mathrm{mri}}}}+ { k_{\mathsmaller{\mathrm{mri}}}}^2 \eta ) ( { \gamma_{\mathsmaller{\mathrm{mri}}}}+ { k_{\mathsmaller{\mathrm{mri}}}}^2 \nu ) } \right]\end{aligned}\ ] ] when transforming eqs .
( 50 ) and ( 51 ) of pc08 into dimensional units . in general ,
velocity channels and magnetic channels are not orthogonal .
however , it is evident that if @xmath81 ( i.e. when the magnetic prandtl number @xmath82 ) , eq.([eq : phi_b ] ) reduces to eq.([eq : v_b_pi ] ) , so that the shift of @xmath83 between @xmath78 and @xmath57 may also hold in the non ideal mhd limit .
as the expressions for the mri growth rate @xmath76 and the amplitude ratio @xmath77 are quite complex , we do not give them here ( the interested reader will find the details in pc08 ) .
instead , we briefly discuss some of their key physical aspects .
both resistivity and viscosity reduce the mri growth rate .
they also shift both the critical wavenumber @xmath84 and the most unstable wavenumber @xmath70 to lower values than those given for ideal mhd by eq .
( [ eq : mri_kcrit ] ) and ( [ eq : kmri ] ) , respectively .
this behaviour can be readily understood , as dissipative effects are more pronounced in the resistive - viscous regime ( as they scale @xmath85 ) .
the smaller the reynolds numbers , the larger the growth rate reduction and the larger the shifts .
moreover , the magnetic reynolds number has a bigger influence than the hydrodynamic one , because the nature of the mri is more `` magnetic '' than `` hydrodynamic '' .
for any reynolds number , the mri growth rate @xmath86 tends to zero for sufficiently short wavevectors , i.e. @xmath87 for @xmath88 ( see eq.[eq : gr_ideal ] ) .
pc08 found analytic expressions for the growth rate @xmath69 , and the wavevector @xmath70 of the fastest - growing mode only in some limiting cases , usually for very large or very small reynolds numbers ( e.g. for @xmath89 and @xmath90 ) .
we note that for @xmath91 ( which holds for core collapse supernovae outside the neutrinosphere ) the ideal mhd expressions given in eqs .
( [ eq : gammamri ] ) and ( [ eq : kmri ] ) determine the values of the wavenumber and the growth rate of the fastest - growing mode with relative errors @xmath92 compared to the values obtained by the expressions of pc08 .
the high conductivity of the degenerate matter in a supernova core implies very high magnetic reynolds numbers .
using an order - of - magnitude estimate of @xmath93 @xcite , eq .
( [ eq : rm ] ) yields @xmath94 because the magnetic reynolds number is so large , we can safely neglect resistivity in the expressions for the growth rate and the wavelength of the mri .
the molecular viscosity of supernova matter ( @xmath95@xmath96s@xmath97 ) is only a few orders of magnitude larger than the resistivity , i.e. the respective hydrodynamic reynolds number @xmath98 @xcite . in the region below the neutrinosphere
, however , the tight coupling between neutrinos and matter changes the situation . according to @xcite the neutrino - matter interaction results in an effective viscosity that varies between a few @xmath99 deep inside the pns and @xmath100 near the neutrinosphere
, i.e. one finds @xmath101 the above equation implies that the expressions for the growth rate and for the angles @xmath57 and @xmath78 ( eqs .
[ eq : v_b_pi1]-[eq : gammamri ] ) , which are valid in the limit of ideal mhd , are not satisfied inside the pns . outside the neutrinosphere , the interaction between matter and neutrinos changes its character from diffusion to free streaming . in the free streaming regime neutrino drag
damps fluctuations not like an effective viscosity , but like a drag force that is independent of the wavelength .
its impact on the mri is therefore very different from the impact of a viscosity @xcite .
neutrino drag is relevant not only outside but also significantly below the neutrinosphere as long as the dynamics of interest is happening at a wavelength shorter than the neutrino mean free path , which varies from a metre inside the pns to several kilometre near its surface ( see fig.5 in * ? ? ?
because of these complications we will not investigate this effect here and refer the reader to the studies of @xcite and @xcite .
mri channels can not grow indefinitely , because their energy would constantly increase , whereas the energy of the system contained in differential rotation , which provides the reservoir for the mri , is finite . hence , there must be a physical mechanism terminating mri growth .
gx94 suggested that mri channels being exact solutions of the mhd equations ( in the shearing sheet and incompressible flow approximations ) may be unstable against parasitic instabilities , which could terminate the mri growth .
they found in their analytic calculations ( under the assumptions described in sec .
[ sec : termi_para ] ) that in ideal mhd ( shear driven ) kh modes can develop on top of mri channels .
gx94 also suggested that in resistive mhd , parasitic instabilities of the ( current driven ) tm type could develop , too .
we note that the importance of magnetic reconnection for mri termination was already discussed by bh91 .
analytic calculations by @xcite in resistive mhd confirmed this hypothesis . alternatively ,
if this scenario does not hold , the magnetic field of the mri channels could grow to a dynamically relevant strength ( when the alfvn speed becomes comparable to the sound speed ) , violating the approximation of incompressibility .
consequently , the magnetic field pressure would become important and it could push matter towards magnetic null surfaces of the mri channels , or mri channels could become buoyancy unstable ( gx94 ) . finally , small amplitude mri channels emerging in an already ( mri - driven ) turbulent state could be destroyed by non - linear mode - mode interactions or turbulent mixing @xcite . in this paper
, we only consider the first scenario , i.e. mri termination via parasitic instabilities . in sec .
[ sec : termi_para ] , we will briefly discuss assumptions and findings of the parasite model of @xcite and of its extension to resistive - viscous mhd by @xcite .
gx94 considered perturbations in a system with already well developed mri channels ( of the fastest - growing mode with @xmath70 ) given by @xmath102 where @xmath103 and @xmath104 are the velocity and the magnetic field of the parasitic instabilities , respectively . solving the equations governing
the evolution of the secondary ( parasitic ) instabilities is a very challenging task , because mri channels , which are treated as a background field for the perturbations , are non - stationary .
hence , standard techniques like an wkb ansatz can not be used . to make this task more tractable for analytic studies ,
gx94 considered a stage of mri growth when the amplitude of the mri channels is much larger than the initial weak magnetic field , i.e. @xmath105 .
the growth rate of the secondary instabilities @xmath106 ( which scales @xmath107 ) is then much larger than the mri growth rate , i.e. @xmath108 . under these conditions
the time evolution of the mri channels , the coriolis force ( which is of the order of @xmath69 ) , the background shear flow , and the initial background magnetic field @xmath109 can be neglected .
hence , instead of searching for solutions to perturbations according to eqs.([channel_full_p1 ] ) and ( [ channel_full_p2 ] ) , gx94 considered a more simplified system where the velocity and the magnetic field are given by @xmath110 with @xmath111 being the time at which the secondary perturbations are imposed .
similar assumptions were also made by @xcite and @xcite .
@xcite identified regions in parameter space , where depending on the values of the hydrodynamic and magnetic reynolds numbers either kh or tm is the dominant ( i.e. faster developing ) secondary instability that terminates mri growth . in particular , for the conditions prevailing in core collapse supernovae outside the pns ( @xmath112 ) the exponential growth phase of the mri should be terminated by kh instabilities .
tm should be dominant only in very resistive media , i.e. if @xmath113 .
@xcite also found that , in general , ( shear driven ) kh modes grow fastest along the mri velocity channels , @xmath114 , whereas ( current driven ) tm grow fastest along the magnetic field channels , @xmath115 .
one of the aims of this work is to test these predictions through numerical simulations .
we use the three - dimensional eulerian mhd code aenus @xcite to solve the mhd equations ( [ eq : cont])([eq : divb ] ) .
the code is based on a flux - conservative , finite - volume formulation of the mhd equations and the constrained - transport scheme to maintain a divergence - free magnetic field @xcite . using high - resolution shock - capturing methods ( e.g. * ? ? ?
* ) , the code employs various optional high - order reconstruction algorithms including a total - variation - diminishing piecewise - linear ( tvd - pl ) reconstruction of second - order accuracy , a third- , seventh- and ninth - order monotonicity - preserving ( mp3 , mp5 , mp7 and mp9 , respectively ) scheme @xcite , a fourth - order , weighted , essentially non - oscillatory ( weno4 ) scheme @xcite , and approximate riemann solvers based on the multi - stage ( musta ) method @xcite .
we add terms including viscosity and resistivity to the flux terms in the euler equations and to the electric field in the mhd induction equation .
we treat these terms similarly to the fluxes and electric fields of ideal mhd , except for using an arithmetic average instead of an approximate riemann solver to compute the interface fluxes .
the explicit time integration can be done with runge - kutta schemes of first , second , third , and fourth order ( rk1 , rk2 , rk3 , and rk4 ) , respectively .
choosing appropriate numerical schemes for our 3d simulations is an important issue , because we want to keep the numerical viscosity and resistivity as low as possible without unnecessarily increasing the computational cost .
therefore , before applying the code to the mri , we assessed its numerical viscosity and resistivity in an extended set of auxiliary simulations ( rembiasz et al .
, in preparation ) .
we performed test calculations of linear - wave propagation and the tm instability , comparing the numerical solutions with ( semi- ) analytic solutions .
our tests show that the very low numerical dissipation of the mp9 scheme well justifies its larger stencil ( requiring more ghost zones ) . in the tm simulations ,
the main contribution to the numerical dissipation comes from the spatial rather than the temporal discretisation errors .
we do not observe any gain in accuracy when using the rk4 instead of the rk3 scheme .
because we expect these findings to hold also for mri simulations , we performed the simulations reported here with the mp9 scheme , a musta solver based on the hlld riemann solver , and an rk3 time integrator @xcite .
we use the hybrid equation of state ( eos ) of @xcite , in which the gas pressure @xmath116 results from the addition of two contributions , namely a baryonic one @xmath117 and a thermal one @xmath118 .
these pressure contributions are given by @xmath119 where @xmath120 is the polytropic constant , and @xmath121 and @xmath122 are the barotropic index and the thermal adiabatic index , respectively .
the quantity @xmath123 is the thermal part of the internal energy @xmath124 , i.e. @xmath125 .
our study comprises a set of two - dimensional axisymmetric simulations and a set of three - dimensional simulations . for both kinds of simulations
we employed cylindrical coordinates @xmath126 and a computational domain centred around the equatorial plane at a radius @xmath127 . for this value of @xmath37 our computational box
is located in the middle of a nascent pns of radius @xmath128 .
however , @xcite have recently shown that the viscosity due to neutrinos can be much higher in collapsing cores than previously thought .
close to the neutrinosphere , i.e. whenever the neutrino mean free path is shorter than the length scale of interest , very low reynolds numbers are expected . the most favuorable place for mri amplification
is located close to the pns star surface , where differential rotation is stronger , and the reynolds numbers are larger if the surface is located above the neutrinosphere , which however may not always be the case ( see fig.10 in * ? ? ?
however this recent finding does not invalidate our studies as our models can be easily scaled for different initial conditions .
we may choose different values of the radius , @xmath22 , rotational velocity , @xmath25 , or density , @xmath9 , and scale the key physical quantities in following way : @xmath129 where @xmath130 , and @xmath131 , @xmath132 is the box length in the direction @xmath133 ( where @xmath134 ) , and the barred quantities are the values set up in or computed from our simulations .
we stress again that the most favorable place for the development of the mri is the region close to the surface of the pns , where the differential rotation gradient and the reynolds numbers are larger than deep inside the pns . since the choice of @xmath36 is ad hoc , it is evident that the ratio @xmath135 can be made as close as desired to one .
this means that neither the mri growth rate nor the typical growth time will change under a translation of the box to the surface of the pns .
such a translation will also imply that @xmath136 , while @xmath137 , hence , the typical magnetic fields and maxwell stresses would be 0.6 and 0.4 times smaller than in our models .
however , as we will show , none of these variations would change the foremost qualitative prediction of our models , namely , that the termination of the mri growth in collapsing stellar cores happens by the action of parasitic kh modes . in the 3d simulations , the typical box size is @xmath138 in @xmath139 , and @xmath31 direction , while the typical box size is @xmath140 in the @xmath141 plane in our 2d simulations .
we performed simulations involving up to @xmath142 zones ( see table[tab : main_results ] ) .
we assume periodic boundary conditions in both @xmath30 and @xmath31-direction .
this choice is natural for the @xmath30 direction , whereas in the vertical direction , @xmath31 , the core is obviously not periodic .
however , for the simulated region near the equatorial plane the vertical component of the gravitational force , @xmath143 , can be neglected , because it is much smaller than the radial one , i.e. @xmath144 .
unlike in local simulations of accretion discs , we can not use the shearing sheet boundary condition ( see * ? ? ? * ) in the radial direction , because it does not allow for global gradients of thermodynamic variables ( which are present in core collapse supernovae ) in the simulation domain .
therefore , like @xcite ( who followed * ? ? ?
* ) , we use the shearing disc boundary condition , which allows one to take these gradients into account .
we solve the full compressible mhd equations and do not perform a transformation to the frame corotating with the fluid , which discriminates our simulations from the common shearing - box approach . in radial direction , we apply periodic boundary conditions to the deviation of a variable ( here , density ) from its initial background state , e.g.@xmath145 i.e. we enforce periodicity of the perturbations .
we apply these boundary conditions to angular velocity , density , momentum , and entropy . because the initial magnetic field is homogeneous in all our simulations ,
we use periodic boundary conditions for this quantity too . like @xcite , we use equilibrium initial models based on the final stages of post - bounce cores from @xcite , in which ( several tens of milliseconds after core bounce ) the shock wave has reached distances of a few hundred kilometres and the post - shock region exhibits a series of damped oscillations as the pns relaxes into a _ nearly _ hydrostatic configuration .
the rotational profile ( given by eq.[eq : omega ] with @xmath146 ) that we used in our simulations , is similar to the one employed in the global mri simulations of @xcite . because the resulting centrifugal force is insufficient to balance gravity ,
the gas is kept in ( an initial hydrostatic ) equilibrium by an additional pressure gradient , so that @xmath147 the initial distributions of angular velocity , density , and gravitational potential are depicted in fig.[fig : mri_init ] .
( solid line , right ordinate ) , the density @xmath148 ( dashed line , left ordinate ) , and the angular velocity @xmath149 ( dash - dotted line , left ordinate ) . the entropy profile of this specific model is assumed to be flat . ] unless otherwise stated , we set the initial magnetic field strength to @xmath150 , the shear and bulk viscosity to @xmath151 , and the resistivity to @xmath152 , which implies a reynolds number @xmath153 and @xmath154 , respectively .
the reynolds numbers slightly vary in the box from @xmath155 , because both the alfvn speed and the rotational velocity are functions of radius ( see fig.[fig : mri_init ] , and eqs.[eq : re ] and [ eq : rm ] ) . for these default parameters ,
the wavelength of the most unstable mri mode ranges from @xmath156 , and the corresponding velocity channels and the magnetic field channels form ( for a rotational shear @xmath146 ) at an angle , @xmath157 and @xmath158 , respectively ( see eqs.([eq : phi_v ] ) and ( [ eq : phi_b ] ) ) .
these angles differ from those in ideal mhd ( @xmath159 and @xmath160 ) only very little . in all simulations presented below the initial magnetic field has only a uniform component in @xmath31 direction , as defined in eq.([eq : b_init ] ) .
this field geometry is a popular choice in mri simulations , ( see , e.g.bh91 ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , since the vertical component is the most important one for the development of the instability ( cf .
another common choice is a so - called _ zero net flux _ configuration ( see , e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , in which the magnetic field has a sinusoidal radial dependence , i.e. @xmath161 , where @xmath162 is chosen in such a way that an integer number of wavelengths fits the computational domain .
thus , @xmath163 with @xmath164 being a natural number .
the value of the initial magnetic field amplitude , @xmath109 , requires some further comments . according to state - of - the - art stellar evolution calculations
the pre - collapse magnetic field for the most strongly magnetised progenitors is less than about @xmath165 g @xcite . during the collapse phase the magnetic field
can be amplified by compression by two orders of magnitude to @xmath166 g @xcite . from eq.([eq : kmri ] ) , which is valid in ideal mhd , we estimate that for the pns the wavelength of the fastest - growing mri mode is @xmath167 for the typical reynolds numbers used in our simulations ( @xmath168 ) the wavelength of the fastest - growing mode is @xmath169 longer ( and the growth rate @xmath170 lower ) than the one in ideal mhd .
( for given @xmath171 , and @xmath20 ) using the expression from pc08 and determining the maximum @xmath69 and its location , @xmath70 , graphically . ]
assuming that @xmath172 zones per mri channel are needed to resolve it properly , a simulation with @xmath173 g would require a resolution of the order of @xmath174 zones per dimension , which is already unaffordable in 2d .
one could reduce the cost by using a smaller computational domain , but the high rotational and sound speeds ( @xmath175 ) would still limit the timestep to @xmath176 , i.e. almost @xmath177 iterations would be required to simulate the mri until its termination at @xmath178 . therefore ,
following @xcite , we use initial magnetic fields which are two orders of magnitude higher ( @xmath179 g ) to increase the wavelength of the fastest - growing mri mode to @xmath180 .
this reduces the minimum resolution to @xmath181 zones in a @xmath182 box , and the number of iterations to less than @xmath174 .
we pay attention to choose the values of the initial magnetic field strength such that an integer number of channels of the fastest - growing mri mode fits into the computational domain . in this way
this mode fulfils the periodic boundary conditions imposed in @xmath31 direction .
otherwise , the fastest - growing mode could be artificially suppressed by an unfavourable box size .
@xcite showed in test simulations that if , e.g. a box of size @xmath183 in @xmath31 direction is used , usually three mri channels form .
either all three of them have a wavelength smaller than @xmath184 or he finds a combination of larger and smaller channels . in any case , the mri developed at a rate which was lower than theoretically expected .
to trigger the mri we impose an initial velocity perturbation on the background velocity profile ( defined by eq.[eq : v_init ] ) of the form @xmath185 , \label{eq : v_with_sin}\end{aligned}\ ] ] where @xmath186 , @xmath187 , and @xmath188 are random numbers in the range @xmath189 $ ] , @xmath190 and @xmath191 are the perturbation amplitudes , @xmath59 is the radial perturbation wavenumber , and @xmath192 is the amplitude of the sinusoidal perturbation .
if not otherwise written , @xmath193 , @xmath194 , @xmath195 , and @xmath196 .
@xcite used a similar prescription for the initial perturbation , except that the sinusoidal part was not present in their case , i.e. @xmath197 .
we find that the sinusoidal term is more robust in exciting mri modes from small perturbations , which are sometimes suppressed by numerical effects if only random perturbations are imposed ( see * ? ? ?
imposing axisymmetry severely limits the number of modes that can grow in 2d simulations .
while the fastest - growing mri mode , which is an axisymmetric one , can freely develop in such simulations , the dominant parasitic instabilities , which for @xmath198 are non - axisymmetric kh modes , are suppressed @xcite .
hence , among axisymmetric secondary instabilities the fastest - growing modes are of tm rather than kh type .
this even holds for simulations with a very low or even a vanishing physical resistivity , as discussed by @xcite .
these authors performed extensive studies of the mri by means of local 2d ideal mhd simulations .
their simulations confirmed the instability criteria and the growth rates of the mri for the flow regimes relevant to core collapse supernovae .
they also found that the growth of the mri is terminated by a tearing mode ( tm ) instability developing because of the unavoidable presence of a numerical resistivity in ( even ideal ) finite - volume mhd codes .
figure[fig:2d_mri ] summarizes the evolution of an axisymmetric model simulated with a resolution of @xmath199 zones .
we performed appropriate convergence studies to ensure that the mri is properly resolved at this grid resolution ( see * ? ? ?
* for details ) .
panel displays the time evolution of the absolute value of the volume - averaged maxwell stress component @xmath200 where @xmath201 is the volume of the computational domain .
the other three panels display the colour encoded value of the radial component of the magnetic field in the @xmath202 plane at three different times . from the initial velocity perturbations ( see eq.[eq : v_with_sin ] ) , three mri channels have formed at @xmath203 ( _ top right _ panel ) that grow exponentially with time at a constant rate ( _ top left _ panel ) .
a linear fit of @xmath204 in the time interval @xmath205 \ , { \textrm{ms}}$ ] gives @xmath206 .
this value is consistent with the local linear analysis , eq.([eq : gammamri ] ) , which predicts values of @xmath207 varying from @xmath208 to @xmath209 within the box boundaries because of its dependence on @xmath210 . at @xmath211 ,
the maxwell stress reaches a value of @xmath212g@xmath213 and the mri growth is terminated by parasitic instabilities .
we note that in 2d simulations , the stress at termination is highly sensitive to the initial random perturbation imposed in the simulation . performing several realizations of the same simulation , with different seeds for the random number generator , we obtained a maxwell stress at the termination varying from @xmath214g@xmath213 to @xmath215g@xmath213 ( see * ? ? ? * for details ) . as we discuss in the next section , we do not observe this large scatter in our 3d simulations .
the colour map of the radial component of the magnetic field exhibits several x points at @xmath216 ( _ bottom left _
panel ) , where field lines reconnect .
these x points are located at @xmath217 , and @xmath218 , respectively .
there are also three o points recognisable in the centres of magnetic islands located at @xmath219 , and @xmath220 , respectively .
these patterns indicate mri termination by tm rather than by kh instabilities . shortly afterwards , at @xmath221 ( _ bottom right _ panel ) , the channel modes have been destroyed and mhd turbulence sets in . these results , which are qualitatively similar to those found by @xcite , confirm that in axisymmetric models the mri is artificially terminated by tms , which would grow slower than kh instabilities in 3d models . defined by eqs.([eq : hhat ] ) and ( [ eq : what ] ) and the cylindrical coordinates @xmath222 .
the angles @xmath78 and @xmath57 are given by eqs.([eq : phi_v ] ) and ( [ eq : phi_b ] ) , respectively .
compare with fig.2 of pc08 . ] according to predictions of @xcite , in general , the fastest - growing kh instabilities should develop along the velocity channels , i.e. @xmath114 , whereas the fastest tm should grow along the magnetic field channels , i.e. @xmath115 .
as these channels define two important directions in the horizontal ( @xmath139 ) plane , we will sometimes use in the following discussion another coordinate system @xmath223 , the axes @xmath224 and @xmath225 being aligned with the velocity and magnetic field channels , respectively .
the @xmath226 coordinates and the @xmath227 cylindrical coordinates are related through the following coordinate transformation ( fig.[fig : hw_coordinates ] ) @xmath228 where the angles @xmath78 and @xmath57 are given by eqs.([eq : phi_v ] ) and ( [ eq : phi_b ] ) , respectively . to denote vectors ( or their components ) in the @xmath227 plane . ] for @xmath229 , the axes @xmath224 and @xmath225 are orthogonal , i.e. during the phase of exponential growth of the mri @xmath230 and @xmath231 .
figure[fig : hw_coordinates ] depicts the horizontal plane in both coordinate systems .
the channel angles are set to @xmath159 and @xmath160 , which corresponds to the ideal mhd limit . in general , the vectors @xmath232 and @xmath233 are not orthogonal to each other , and their exact orientation depends on the reynolds numbers .
for example , in the limit @xmath234 and @xmath235 , one has @xmath236 and @xmath237 , i.e. @xmath238 and @xmath239 ( pc08 ) . for the values of the reynolds numbers used in our simulations , i.e. @xmath240 , and @xmath241
, we expect the channels to be oriented along @xmath242 , @xmath243 , @xmath244 , and @xmath245 , respectively .
these angles differ only little from those of the ideal mhd case .
hence , in either case , the fastest - growing kh mode should develop at an angle @xmath246 , where @xmath247 denotes the angle at which the dominant parasitic instability develops .
depending on the initial conditions , one either has @xmath248 or @xmath249 . in 2d axisymmetric simulations ,
the only allowed angle is @xmath250 .
km , which was used to simulate models # 5 to # 8 and # 10 ( see table[tab : main_results ] ) together with two cuts corresponding to the @xmath251 plane ( blue ) and the @xmath252 plane ( red ) used to display some results of model # 7 in fig.[fig : mri_cuts ] .
note that the computational volume is actually bent in @xmath30 direction and the two cuts are no planes but curved surfaces in cylindrical coordinates . ]
we studied the termination process in a number of 3d simulations , varying the size and aspect ratio of the computational domain , and the grid resolution ( see tab.[tab : main_results ] for the list of models ) .
the default simulation box has a size of @xmath253 , as shown in fig.[fig : mri_cut_geometry ] , which also depicts two cuts corresponding to the @xmath251 and @xmath252 planes .
we note that the computational volume is actually bent in @xmath30 direction and the two cuts are not planes but curved surfaces in cylindrical coordinates .
however , we will call them planes for simplicity . [ cols="^,<,^,^,^,^,^,^,<,^,^,^,^,^,^,^",options="header " , ] [ tab : main_results ] we begin the discussion of our results with model#7 ( see tab.[tab : main_results ] ) , which we simulated in the default box resolved with @xmath254 zones in @xmath139 , and @xmath31 direction , respectively .
except for the third space dimension , the initial conditions and the parameters of this 3d model are identical to those of the axisymmetric model discussed in sec.[ssec:2d ] . as illustrated by the time evolution of the maxwell stress ( fig.[fig : mri_time_evolution ] ) , the mri grows exponentially with time at the same rate as in the 2d model ( @xmath255 ) until saturation , which occurs a bit earlier than in the 2d model at @xmath256 . for the 3d models # 5 ( orange ) , # 6 ( red ) , # 7 ( black ) , # 10 ( violet ) , and # 11 ( blue ; run only until @xmath257 ) .
up to @xmath258 , the models give almost identical results , differences being only visible around termination ( see inset ) .
vertical lines indicate the times of the snapshots displayed for model#7 in fig.[fig:3d_mri ] .
for a better readability of the figure , the vertical lines around mri termination are only displayed in the inset . ]
figure[fig:3d_mri ] shows the 3d structure of the radial component of the magnetic field at six distinct times ( marked by vertical lines in fig.[fig : mri_time_evolution ] ) .
magnetic field perturbations grow in the form of three axisymmetric channels ( _ upper left _
panel ) , which are perturbed in turn by secondary instabilities that become recognisable only shortly before the mri is terminated at t = 11.2 . the sequence of snapshots around this time ( _ upper middle _ to _ bottom middle _ ) demonstrates that non - axisymmetric parasitic modes are responsible for mri termination .
after exponential growth of the magnetic field has ended , the simulation volume is dominated by a small - scale , turbulent magnetic field ( _ bottom right _ ) . comparing the structure of the radial magnetic field during mri termination ( fig.[fig : mri_2d_3d_term_comparison_local ] )
, we find profound differences between the 2d model ( _ left - hand _ panel ) and the 3d model ( _ right - hand _
the 2d model exhibits x and o points that are characteristic of the tm instability , whereas in the 3d simulation the channel modes bend strongly at the locations of vortex rolls which are a typical feature of the kh instability .
we also studied some geometrical aspects of the parasitic instabilities found in the 3d simulation . for this purpose
, we use the @xmath259 coordinates.velocity-shear driven kh instabilities should grow fastest in the @xmath224 direction and dominantly current driven tm along the @xmath225 axis . to simplify the expressions we performed the coordinate transformation @xmath260 with @xmath159 and @xmath160 instead of the theoretically expected angles @xmath261 and @xmath262 , which does not affect our qualitative analysis , however .
figure[fig : mri_cuts ] shows the distribution of the magnetic field components @xmath263 ( _ left - hand panels _ ) and @xmath264 ( _ right - hand panels _ ) in two - dimensional cuts through the computational domain of the 3d model#7 shortly before mri termination .
the component @xmath263 is considerably smaller than the component @xmath264 .
this is consistent with the theoretical expectation for an mri channel whose magnetic field should grow in the @xmath225 direction ( @xmath160 ) and vanish in the perpendicular @xmath224 direction ( @xmath265 ) .
close to termination , the mri channels are strongly perturbed and vortex rolls begin to form in the @xmath224 direction ( _ upper panels _ ) .
this indicates that kh instabilities developing along the velocity channels ( i.e. in the @xmath224 direction ) are responsible for the disruption of the mri channels . in the @xmath225 direction
( i.e. along the magnetic field channels separated by current sheets in which tm could develop ) , the channels remain almost unperturbed .
only box - size structures appear which are most likely a consequence of the radial boundary conditions imposed in our simulations ( see sec.[ssec : bc ] ) .
although we can not discard the presence of sub - dominant tm in this projection , it is clear that tm do not play a dominant role in the termination process , which can be understood completely in terms of parasitic kh instabilities . before we proceed
further we take another look at the spatial structure of @xmath266 during mri termination at @xmath267ms , which is illustrated in fig.[fig : deep_cuts ] . in the upper panel depicting @xmath266 in a @xmath268 cut at @xmath269 km ( this cut corresponds to the outer radial boundary of the computational domain shown in the upper right panel of fig.[fig:3d_mri ] ) , one can recognize vortex rolls developing along the mri channels . from the middle panel , which shows the distribution of @xmath266 in the @xmath270 plane at @xmath271 km , one could determine the horizontal wavelength @xmath272 and the angle at which parasitic instabilities develop ( @xmath273 ) .
however , we can obtain these quantities much more accurately using fourier transforms ( see next subsection ) .
the bottom panel provides another view of the structure of @xmath266 . by showing a part of both the @xmath251 cut and the @xmath252 cut
, it illustrates in which direction parasitic instabilities develop .
we note that the @xmath251 and @xmath252 cuts shown in figs .
[ fig : deep_cuts ] and [ fig : mri_cuts ] are different . in order to confront the theoretical expectations with our numerical results ,
we have calculated spatial discrete 3d fourier transforms of the magnetic field components @xmath274 with @xmath275 at a given time using a fast fourier transform ( fft ) algorithm .
we denote the complex fft coefficients as @xmath276 .
the power spectral density is proportional to @xmath277 , which is a measure of the average magnetic field energy density of the component @xmath274 in fourier space .
we chose the components @xmath266 and @xmath278 for the analysis , because they contain the information about both the mri channel flows and the parasitic instabilities .
we are particularly interested in determining the wavevectors , @xmath279 , of the dominant modes , which have the largest amplitudes in the fourier space .
we expect that mri channel flows appear as structures with a wavevector @xmath280 whose modulus is maximum for the fastest - growing mode , @xmath281 .
parasitic instabilities develop in the whole fourier space , whereas the mri does not contribute to modes with finite @xmath282 and @xmath283 , but @xmath284 .
hence , parasitic instabilities are expected to produce a characteristic signature with wavevectors @xmath285 which should be distinguishable from the mri signature .
the maximum fourier amplitude should be attained for the wavevector of the fastest - growing parasitic mode .
the analysis of model#7 reveals that the fourier amplitude along the line @xmath286 peaks at @xmath287km@xmath97 , which corresponds to the wavelength of the fastest - growing mri mode , @xmath288 km of this model .
the latter result holds for all our simulations during the phase of exponential growth of the mri .
the fourier amplitudes in the plane @xmath289 , displayed for three different times for model#7 in fig.[fig : fourier ] , show a power excess which peaks at @xmath290 km @xmath97 .
the location of this maximum barely changes with time , while the amplitudes of the fourier coefficients increase relative to those at @xmath291 .
this result is consistent with a super - exponential growth of parasitic instabilities .
in addition , as we will demonstrate below , the behaviour of the fourier amplitude in the @xmath289 plane is consistent with the development of parasitic kh instabilities feeding off the mri channels . for model#7 at @xmath292ms ( top ) , @xmath293ms ( middle ) , and @xmath294ms ( bottom ) ,
the colour coded quantity is the ratio @xmath295 of the fourier amplitude representing parasitic instabilities and the one representing mri channels .
the white square is the location of the energy - weighted barycenter computed with eq.([eq : barycenter]).,title="fig : " ] for model#7 at @xmath292ms ( top ) , @xmath293ms ( middle ) , and @xmath294ms ( bottom ) , respectively .
the colour coded quantity is the ratio @xmath295 of the fourier amplitude representing parasitic instabilities and the one representing mri channels .
the white square is the location of the energy - weighted barycenter computed with eq.([eq : barycenter]).,title="fig : " ] for model#7 at @xmath292ms ( top ) , @xmath293ms ( middle ) , and @xmath294ms ( bottom ) , respectively .
the colour coded quantity is the ratio @xmath295 of the fourier amplitude representing parasitic instabilities and the one representing mri channels .
the white square is the location of the energy - weighted barycenter computed with eq.([eq : barycenter]).,title="fig : " ] the average magnetic energy density of the field component @xmath274 can be computed from the fourier amplitudes as @xmath296 where @xmath297 similarly , we can estimate the average magnetic energy density of the mri channels or parasitic instabilities restricting the summation to locations in fourier space relevant for each kind of instability .
therefore , we define @xmath298 as proxies for the average magnetic field energy density associated with these instabilities . in case of the mri ,
the fourier amplitudes are distributed along the line @xmath299 , i.e. @xmath300 should be a good estimator .
however , in the case of the parasitic modes , @xmath301 contains not all contributions to the energy density and thus provides only a lower bound of the energy density associated with the parasitic instabilities .
fig.[fig : parasites_time ] shows the time evolution of the magnetic energy density for both the mri and the parasitic instabilities .
the behaviour of @xmath302 is very similar to the time evolution of the maxwell stress ( see fig.[fig : mri_time_evolution ] ) .
it is characterized by an exponential growth with the same growth rate as for the maxwell stress @xmath303 , and a termination point associated with the disruption of the mri channel flows .
the average magnetic energy density @xmath304 of the parasites starts to grow super - exponentially at @xmath305 from a value of about @xmath0 orders of magnitude smaller than that of the mri . at termination ,
the magnetic energy density of the parasites amounts to , at least , @xmath306 of the magnetic energy density of the mri . ) and the parasitic instabilities ( @xmath307 ) for different components @xmath274 of the magnetic field for model#7 . ] to determine the wavevector of the fastest - growing parasitic instability , we search for the maximum fourier amplitude with @xmath289 . for this purpose , it is sufficient to consider only positive components of the wavevector @xmath308 . numerically , we obtain the wavevector by computing the energy - weighted barycenter in the relevant part of fourier space , @xmath309 where we limited all summations to those fourier modes displayed in fig.[fig : fourier ] to avoid high frequency contaminations . substituting @xmath310 for
@xmath311 does not change the results significantly .
the values of @xmath312 obtained from eq.([eq : barycenter ] ) ( white squares in fig.[fig : fourier ] ) properly trace the location of the maximum fourier amplitude . with these values
one can compute both the wavelength and the angle of the parasitic instability according to @xmath313 figure[fig : parasitic_theta ] shows the time evolution of the wavelength , @xmath272 , and angle , @xmath314 , of the parasitic instabilities during the late stage of mri evolution . before mri termination ,
the evolution of the angle is compatible with its theoretically expected behaviour for kh instabilities , i.e. @xmath315 ( horizontal green line ) , within the accuracy of the angle determination . and
@xmath316 respectively , depends on the accuracy in the determination of @xmath317 , which is set by the size of the box . ]
the wavelength differs from its theoretically expected value @xmath318 ( horizontal blue line ) by a factor of @xmath319 .
whether the source of this disagreement is of a numerical origin or results from a limitation of the theoretical approach of @xcite is beyond the scope of this work . , and the wavelength , @xmath320 , of the parasites , respectively .
horizontal solid lines show the corresponding theoretical values for kh modes that one expects to develop in this simulation . ]
we have performed a similar analysis for all our other simulations .
table[tab : main_results ] gives the values of @xmath272 and @xmath314 at a representative time before termination ( i.e. at @xmath321ms for models # 1 to # 16 , and at @xmath322ms for models # 17 and # 18 ) . in the following three subsections ,
we discuss the influence of different numerical and physical parameters on the values of these two quantities .
next , we address whether the size and aspect ratio of the computational domain influences mri termination ( for a somewhat similar study of the post - termination phase , see * ? ? ?
following @xcite , we simulated models # 5 , # 6 , # 7 , and # 10 in a box of ( the default ) size ( @xmath323 ) .
we performed additional simulations reducing the size of the box in @xmath30 direction ( @xmath324 , model # 8) , and both in @xmath30 and @xmath31 direction ( @xmath325 , models # 2 , # 3 , # 4 , # 9 , and # 11 ) .
finally , we computed several models ( # 12 to # 18 ) where we varied the azimuthal size @xmath326 of the domain ( see table[tab : main_results ] ) .
the main motivation for using a smaller box in some of our simulations was computational cost reduction .
in accordance with theoretical predictions of @xcite for flows with @xmath112 , we found that in model#7 parasitic instabilities develop at an angle @xmath273 ( see fig.[fig : parasitic_theta ] , and middle panel of fig.[fig : deep_cuts ] ) .
this result suggests that it is unnecessary to use a box elongated in azimuthal direction ( @xmath327 ) .
instead , one could rather study the mri in a box with a horizontal aspect ratio @xmath328 to minimize the volume of the computational domain without affecting the development of parasitic instabilities . to test this hypothesis , we simulated model#8 in a box of size @xmath329 using the same spatial resolution as for reference model#7 .
in both simulations , the maxwell stress tensor at mri termination attained the same value , which confirms our expectations .
we can further reduce the computational domain in the vertical direction based on the following observation . in simulations performed with the default box , we chose the magnetic field strength in such a way that three of the fastest - growing mri modes fit in the computational domain .
thus , it should be possible to reduce the vertical extent of the box by a factor of @xmath330 , i.e. @xmath331 , without hindering the growth of the dominant mri mode .
however , in such a smaller domain parasitic instabilities are restricted to modes having a vertical wavelength equal to the width of an mri channel ( called type - i modes by gx94 ) , and modes with a longer vertical wavelength ( type - ii ) are suppressed .
nonetheless , according to the predictions of gx94 and @xcite , in the ideal mhd limit or for @xmath332 , the dominant parasitic modes should be ( kh modes ) of type - i .
hence , we expect the mri termination process to be unaltered by a box of smaller vertical size . to test this theoretical prediction , we simulated model#9 in a box of size @xmath333 using the same grid resolution as in models#7 and # 8 .
we found that the maxwell stress tensor reached the same value at mri termination in all three models .
this result not only confirms ( given our initial conditions ) the termination of the mri by `` type - i '' parasitic modes , which is an important result , but it also justifies the use of a computational box with a @xmath334 times smaller volume for studying the mri termination process .
we made use of this fact in our resolution studies ( see next subsection ) . in section [ ssec:2d ]
, we showed that the mri is terminated by tms in axisymmetric 2d simulations .
one can also impose axisymmetry in 3d simulations by choosing a box of vanishing azimuthal length , i.e. @xmath335 .
therefore , we expect that for a certain small , but non - zero @xmath326 , kh instabilities should be suppressed even in 3d simulations , and the mri should be terminated by tm instead . to determine this critical azimuthal box length , we performed a series of simulations varying @xmath326 from 1.0 to 0.1 km ( models # 12 to # 18 ) . moreover , to investigate the influence of viscosity , we ran these simulations with a non - zero shear viscosity corresponding to a hydrodynamic reynolds number @xmath336
. the shear viscosity should change the wavelength of the fastest growing mri mode , but only by a negligible amount from the numerical point of view ( i.e. less than @xmath337 ) .
as expected , the mri is terminated by kh instabilities in model#12 , which only differs from model#9 by the value of @xmath338 . in both simulations , the parasitic instabilities have the same horizontal wavelength , @xmath272 , and they develop at the same angle @xmath247 within the measurement error . in models#13#16 , the mri is also terminated by kh instabilities , but the value of the maxwell stress tensor at termination is somewhat larger .
this result can be easily understood : parasitic modes that would grow fastest in a box with @xmath339 are affected ( or even suppressed ) in narrower boxes , i.e. the mri can operate a bit longer before it is finally terminated .
the smaller the box , the stronger the suppression of the kh modes and the larger the deviation of such quantities as the maxwell stress at termination from model # 12 .
finally , in models#17 ( @xmath340 ) and # 18 ( @xmath341 ) , kh modes are suppressed and the mri is terminated by tms , showing similar features as in the 2d simulations described in section [ ssec:2d ] .
we note that the maxwell stress reaches considerably higher values at termination in these models , because tm are growing more slowly .
a similar behaviour was observed by @xcite in their mean - field shearing - box simulations of accretion discs . in simulations done in a box of size @xmath342 ,
kh modes were suppressed and the mri channels were disrupted by tm , whereas in simulations done in larger boxes in the horizontal directions ( i.e. @xmath343 and @xmath344 , or @xmath345 and @xmath346 ) , the mri was terminated by kh instabilities . to summarize , mri termination in models#12#18 is well explained within the parasite model , and the minimum azimuthal box length for which kh modes are not severely affected by boundary conditions is @xmath347 . to properly resolve mri termination , one needs to use a resolution that is high enough not only to resolve the mri channels , but also parasitic instabilities .
the latter criterion is more stringent , as parasitic instabilities develop finer structures in vertical direction than mri channels do ( gx94 , * ? ? ?
we explore this issue in a series of simulations performed in the small ( @xmath348 ; models # 2 , # 3 , # 4 , # 9 , and # 11 ) and in the default ( @xmath253 ; models # 5 , # 6 , # 7 , and # 10 ) computational box .
as we have demonstrated in the previous subsection , parasitic instabilities can develop equally well in both computational domains , i.e. the boxes are equally well suited for resolution studies . in models # 2#4 , which were simulated with 816 zones per mri channel in the vertical direction , we observe differences already during the phase of exponential growth .
first , channel modes emerge @xmath349 later than in the other better resolved simulations .
secondly , the channel modes somewhat differ from the analytic solution , i.e. some imperfections develop on channels which are of numerical and not physical origin .
thirdly , the mri growth rate measured with the help of @xmath303 slightly fluctuates during the phase of exponential growth .
these fluctuations of the order of a few percent can be explained by the fact that the above mentioned imperfections also contribute to @xmath303 . in models # 5#11 , which were simulated with at least 20 zones per mri channel , the channel modes appear roughly at the same time and grow at the same constant rate during the phase of exponential growth .
based on these observations , we conclude that 20 zones per mri channel are sufficient to resolve the phase of exponential growth with the mp9 scheme ( but more zones will be required for lower order reconstruction schemes ) .
this result is consistent with previous estimations of the numerical viscosity and resistivity of the code done by @xcite . with the help of tests involving alfn wave propagation and tm instabilities ,
he determined scaling laws for the numerical viscosity and resistivity of the code as a function of resolution and initial conditions . given our numerical setup , the numerical viscosity and resistivity of the code are @xmath350 , if the relevant length scale is covered by at least 20 zones ( see also section4.2.5.1 ) . to investigate the nature of the parasitic instabilities that are responsible for quenching the mri , we used both fourier analysis and the local magnetic field structure during mri termination . in models#2 to # 11 , we find parasitic instabilities of kh type developing at an angle consistent with @xmath351 independent of resolution .
the horizontal wavelength of the parasitic modes ranges from @xmath352 in the coarsest resolved model#2 to @xmath353 in the best resolved models#10 and # 11 .
remarkably kh instabilities develop even in model#2 , although it was simulated with the coarsest grid which noticeably affects the structure of the channel modes .
their identification was not straightforward , however . from the patterns recognizable in the magnetic field structure during mri termination
, one can exclude tm and other obvious numerical artefacts , e.g. , arising from imperfect boundary conditions .
we also find patterns resembling underresolved vortex rolls which we attribute to kh modes , although without the help of fourier analysis , our experience gained from observing similar patterns in better resolved simulations , and most importantly a theoretical model ( gx94 ; * ? ? ?
* ) , this identification would have been impossible .
the value of the maxwell stress at mri termination depends on the grid resolution ( see tab.[tab : main_results ] and green circle symbols in fig.[fig : reso ] ) . it is larger for coarse grids , the only exception being model#2 . using the mp9 reconstruction scheme , the maxwell stress converges to a value of @xmath354 in models # 10 and # 11 simulated with 67 and 134 grid cells per mri channel , respectively .
the dashed line in fig.[fig : reso ] shows that @xmath303 follows roughly a power law with index @xmath355 as a function of the number of zones per mri channel , except for the under - resolved model#2 with only eight zones per channel , in which mri growth terminates at a much smaller value of @xmath303 .
this behaviour of @xmath303 can be explained within the parasite model .
the kh modes responsible for mri termination grow from the shear layers created by the mri channels . during their growth ,
these secondary instabilities develop structures that are even smaller than the mri channels .
therefore , to resolve these structures a finer grid is necessary than for the mri channels themselves . as a result
, simulations may not be converged even if the numerically obtained mri growth rate is close to the theoretical value .
because badly resolved kh modes grow at a rate below the theoretical expectation , parasitic instabilities reach the necessary amplitude to disrupt the channels slightly later during the evolution , and the mri will terminate at a somewhat higher amplitude than in well - resolved simulations . the appearance of the outlier ( model#2 ) in fig.[fig : reso ] is not surprising .
we have already seen that mri channels are underresolved in this model .
therefore , one would not expect to find any reasonable termination amplitude for this model , i.e. it is probably a chance coincidence that this model gives roughly the same results as the other better resolved models . for completeness
, we included also models#12#18 in fig.[fig : reso ] ( red diamond symbols ) , for which the parasitic modes are restricted not only by the grid resolution but also by the box size and some physical viscosity .
the data of these models lie all above the dashed line indicating the influence of resolution alone .
finally , the asterisk symbol marks the result of the ideal mhd model#1 , which we discuss in the next subsection . .
] [ [ viscosity ] ] viscosity + + + + + + + + + we have performed the reference simulation ( model # 7 ; see tab.[tab : main_results ] ) , and the set of models described above without any explicit physical viscosity . to study whether a finite physical viscosity corresponding to @xmath356 ( for which the equations of ideal hydrodynamics should still approximately hold ) leads to more than mere quantitative changes , we compare the results of models # 9 and # 12 . as expected , in both of these two models , which differ only in the value of the hydrodynamic reynolds number ( @xmath357 for model # 9 , and @xmath356 for model # 12 ) the growth of the mri is terminated by kh modes .
even though no physical viscosity was used for model # 9 , it is affected by a non - zero numerical viscosity . to infer the quantitative importance of the numerical viscosity
, we first compare models which differ only in grid resolution in the inviscid limit , e.g. , model # 9 , where the evolution is still notably affected by numerical viscosity , and models # 10 and # 11 , which are already converged , as we have argued in the previous section .
we note that the numerical viscosity accounts for the @xmath358 difference in the volume - averaged maxwell stress at termination ( i.e. the parasitic instabilities are somewhat under - resolved ) , and has a negligible impact on the mri growth rate .
we now turn to the comparison of an inviscid model ( # 9 ) with one where a physical viscosity has been included ( # 12 ) . quantifying the exact contribution of the numerical viscosity in a model that includes also physical viscosity
is much more involved .
the reasons for this difficulty are multiple . on the one hand ,
the numerical viscosity is a scale dependent effect , affecting more the smallest scales .
hence , it is relatively more important for the development of parasitic modes than for the mri growth . on the other hand ,
the volume - averaged maxwell stress at termination depends chiefly on a delicate balance between two competing factors .
first , models computed with a finite physical viscosity have a smaller @xmath207 .
thus , for a similar growth time , @xmath359 should be smaller than for models computed without a physical viscosity .
secondly , this effect is offset , however , by the also smaller growth rate of the kh parasitic modes , which defers the saturation process and allows for a further growth of @xmath359 . thereby
, our comparison is of qualitative nature only . in view of the former reasoning
, we tentatively attribute the @xmath360 difference in @xmath361 between the inviscid model # 9 and the viscous model # 12 to the effect of a finite viscosity in the latter one , though we can not unambiguously say which part of this difference is caused by numerical viscosity . to support the argument that a finite physical viscosity and not a numerical viscosity is driving the dynamics , we note that the mri growth rate decreases significantly in model # 12 ( as also in all other viscous models from # 13 to # 18 ) .
this expected behaviour ( see section [ ssec : intability_criterion ] ) implies that including a physical viscosity causes measurable differences already in the mri growth phase .
[ [ resistivity ] ] resistivity + + + + + + + + + + + in a final set of simulations , we investigated the influence of a finite ( physical ) resistivity on mri termination .
model#1 , simulated in ideal mhd , represents the typical conditions prevailing in core collapse supernovae close to the surface of pns , whereas in models#19 and # 20 we explored the ( opposite ) limit of a very high resistivity . with the latter simulations we also address a prediction of @xcite , namely that for @xmath362 , the mri should be terminated by tm .
we performed the ideal mhd simulation of model#1 in the default size box , the results being very similar to those obtained for model#7 , which we simulated with the identical setup , but with @xmath154 .
as expected , the mri grew in the ideal mhd simulations at a somewhat higher rate ( @xmath363 ) , because its growth was not damped by resistivity .
we found that the mri is terminated by parasitic kh instabilities in this model , too .
the parasites develop at an angle @xmath273 in accordance with the parasite model of gx94 .
the maxwell stress at mri termination is @xmath364 , a value which is similar to those obtained for models#7 and # 12 ( the latter one being simulated with @xmath365 ) . from these findings
we conclude that for reynolds numbers @xmath366 viscosity and resistivity have no strong influence on the process of mri termination and the saturation value of the maxwell stress .
however , one should not forget that in all three simulations ( i.e. models#1 , # 19 , and # 20 ) , numerical resistivity and numerical viscosity do contribute to the total hydrodynamic and magnetic reynolds numbers ( by lowering them ) . in the limit of very high resistivity ,
i.e. for @xmath367 , the mri modes that grow fastest in ideal mhd are completely suppressed by magnetic dissipation ( pc08 ) . only modes with sufficiently long wavelengths survive the interplay between mri amplification and resistive damping in this limiting case , because magnetic dissipation is weaker for these modes .
the wavelength of the fastest - growing mode is @xmath368 ( pc08 ) . to keep @xmath369 in models#19 ( @xmath370 ) and # 20 ( @xmath371 )
, we had to reduce the initial magnetic field strength to values of @xmath372 and @xmath373 , respectively .
we find that the fastest mri mode grows at a considerably slower rate than in the ideal mhd limit in both models , the measured growth rates being @xmath374 ( model#19 ) and @xmath375 ( model#20 ) .
these values agree very well with those of pc08 , who predicted @xmath376 and @xmath377 , respectively .
pc08 also predicted that for @xmath378 and @xmath370 , we should find @xmath379 and @xmath380 , i.e. the velocity channels and the magnetic field channels should almost be aligned with the @xmath22 axis and the @xmath30 axis , respectively .
indeed , our simulations show that @xmath381 , i.e. @xmath382 , during the exponential growth phase ( see fig.[fig : parasites_tm_time ] ) . finally ,
according to @xcite , the mri should be terminated by tm and not kh instabilities for @xmath362 . for @xmath370 , the dominant parasitic mode should have a wavelength @xmath383 developing at an angle @xmath384 .
a fourier analysis of the mri modes and their parasites ( see fig.[fig : parasites_tm_time ] ) shows that shortly before mri termination ( @xmath385 ) the energy stored in horizontal fourier modes increases significantly growing super - exponentially with time .
however , because of numerical noise introduced by our imperfect boundary conditions , we could neither identify parasitic instabilities nor determine their angle with the help of the fourier analysis .
therefore , the type of the parasitic instability terminating the mri in model#19 ( as well as in model#20 ) remains unknown .
it is probably neither kh nor tm , because we could not recognise the characteristic features of none of these parasites .
the bottom line is that based on models#19 and # 20 , we can not confirm the predictions of @xcite that for @xmath362 the mri is terminated by tm . as the highly resistive limit is of no direct relevance for core collapse supernovae , we did not investigate this regime in more detail .
but for model#19 simulated with a high resistivity ( @xmath370 ) .
note the very different time - scale at which the mri evolves in this case . ]
@xcite first pointed out that the mri can develop in ccsne and could lead to a strong amplification of the magnetic field during the proto - neutron star phase .
the presence of a dynamically relevant magnetic field is of great interest because it may have implications for the supernova explosion mechanism and for explaining the origin of neutron stars magnetic fields .
however , for the typical magnetic field strength present in supernova progenitors , it is necessary to resolve numerically mri channels with a length scale @xmath386 m , which makes global simulations unfeasible even with present day supercomputers . to overcome the limitation of global simulations
@xcite performed semi - global simulations in a kilometre size box located near the surface of the proto - neutron star .
they confirmed the hypothesis of @xcite showing that the mri can strongly amplify the magnetic field and parasitic instabilities are able to destroy the mri channels , thereby quenching the amplification of the magnetic field .
however , they were unable to identify the agent terminating the mri growth .
hence , @xcite could not give an upper limit on the mri - driven magnetic field amplification .
this work tries to shed some light on the termination process .
we numerically studied the evolution of the mri in the context of a ccsn .
to this end , we performed a set of local 3d resistive - viscous mhd simulations in a box representing typical conditions in a proto - neutron star . in all of our simulations we observed the growth of the mri and its subsequent termination .
we identified that secondary kh instabilities are responsible for mri termination by acting as parasites on the mri channels .
our results are consistent with the predictions of the parasite model proposed by gx94 and further developed by @xcite and @xcite .
hence , the parasite model , which is based on a local stability analysis of the system , allows some insight into the termination process and provides a valuable guidance for interpreting numerical results .
this conclusion was obtained by an analysis of our numerical simulations based on mri theory and the parasite model in a number of tests : 1 .
we observe an exponential growth of mri channels of a size compatible with the predicted value , @xmath184 , for the fastest - growing mode ( bh91 ) .
the growth rates measured in the numerical simulations agree with those obtained from the local analysis within the simulation box .
the presence of viscosity or resistivity lowers the values of the growth rates as expected from theory . in the limit of large hydrodynamic and magnetic reynolds numbers , @xmath112 , the angle between velocity ( magnetic field ) in the mri channels and
the radial direction is close to the theoretically expected one in the ideal mhd limit by gx94 , i.e. @xmath387 ( @xmath388 ) .
2 . for simulations with large hydrodynamic and magnetic reynolds numbers , @xmath112 , the mri is terminated by kh instabilities , as predicted by gx94 and @xcite for this regime . to identify the kh instabilities we had to reduce the time interval between outputs of our simulations around termination to a value below the typical growth time of the parasites , in our case @xmath389 .
this explains why @xcite did not observe the development of kh instabilities , since they employed insufficiently frequent output . to help the analysis we project the magnetic field components into the directions @xmath78 and @xmath57 . shortly before termination
we observe the development of vortex rolls at the position of the shear layer located along an angle @xmath387 .
we identify the vortices as a consequence of kh instabilities , which eventually disrupt the channels causing turbulence .
this behaviour is in agreement with the predictions of gx94 and @xcite .
3 . analysing the numerical simulations by means of fourier transforms , we have been able to determine the properties of the parasitic instabilities , i.e. their horizontal wavelengths and the angles at which they develop .
our results confirm the theoretical predictions of gx94 and @xcite that for @xmath112 the dominant parasitic instabilities are kh modes developing parallel to the mri velocity channels ( i.e. at an angle @xmath273 in the ( @xmath139 ) plane ) .
the horizontal length of these kh modes is in a reasonable agreement ( within a factor of @xmath390 ) with the predictions of @xcite .
4 . motivated by the good agreement with the parasite model
, we explored the regime of very high resistivity ( i.e. @xmath362 ) , although it is not of direct relevance for ccsne . in this regime
@xcite predicts that the mri should be terminated by tms . to this end
we performed two simulations with @xmath370 and @xmath391 , but we were unable to identify the agent terminating the mri growth , because no clear signature of tm or kh instability could be identified in our simulations .
we conclude that possibly even higher resistivities are required for tm to become the dominant secondary instability .
a possible explanation of this discrepancy between theory and our simulations is the fact that the calculations of @xcite are based on somewhat simplified assumptions .
therefore , his results should be treated more like a guideline rather than exact predictions . to confirm the above conclusions , we have studied systematically the effect of resistivity , viscosity , box size , and grid resolution .
there are a number of numerical artefacts that can affect the simulations : 1 . in axisymmetric ( 2d ) simulations , the mri growth is terminated by tm .
this was already reported by @xcite , who observed tm in their 2d ideal mhd simulations , in accordance with theoretical results obtained by @xcite .
@xcite argued that this resistive mhd instability must have been triggered by numerical resistivity . in 2d simulations only axisymmetric parasitic modes can develop .
as axisymmetric kh modes are strongly suppressed by the magnetic field tension of the mri channels , tm become the dominant secondary instability .
they suffer less strongly from this constraint , develop faster than axisymmetric kh modes ( but slower than kh modes would grow in full 3d ) , and terminate the mri growth . as a result the mri in 2d is terminated at unrealistically large magnetic stresses and continues to grow after termination , a behaviour not observed in our 3d simulations .
we performed a set of simulations to study the dependence of the properties of the parasite modes . using a 9th - order spatial reconstruction scheme ( mp9 )
, we reached convergent results only in simulations with at least 60 zones per mri channel , i.e. the value of the magnetic stress at termination differed by less than @xmath392 between the two highest resolution runs .
this grid resolution is significantly higher than the one necessary to obtain convergence in the growth rate of the mri , for which 8 zones are sufficient to obtain a @xmath392 accuracy .
we note that the required number of zones will be significantly higher , if lower order reconstruction methods are used .
our result is not surprising when viewed in the light of the parasite model .
the kh instability is triggered by the shear layer between mri channels . at this
layer structures develop that are much finer than the width of the channel itself . failing to capture the kh instabilities properly because of a lack of resolution leads to artificially large magnetic stresses at mri termination .
however , the qualitative behaviour of the flow seems not to be affected by a lack of resolution . even for our lowest resolution simulation ( 8 zones per channel )
we are able to identify kh instabilities developing at @xmath273 as the main termination agent .
we studied how the box size can affect the development of parasitic instabilities . in 3d simulations with an azimuthal box length of at least @xmath393 the mri is terminated by kh modes , whereas simulations with an azimuthal box length @xmath394 gave very similar results as 2d simulations and the mri
was terminated by tm ( because kh modes are suppressed ; compare models # 16 and # 17 in tab.[tab : main_results ] ) .
taking into account that the parasitic kh instabilities develop at a @xmath395 angle , one should use a box size of at least @xmath396 is in the radial direction . however , we recommend using larger boxes in the horizontal directions to reduce the influence of the boundary conditions on the development of the parasitic instabilities . in the vertical direction
it is sufficient to use a box size of @xmath184 , i.e. it is sufficient to consider the evolution of one single mri channel to capture the termination process correctly . determining
the minimum box size has been critical to be able to perform the simulations with the highest resolution presented in this work ( model # 11 with 134 zones per channel ) .
our results have some important implications for the community modelling magnetorotational collapse of stellar cores . studying the termination of the exponential growth phase of the mri
, we learned that it involves parasitic kh instabilities , but does neither depend on the physical resistivity nor viscosity present in ccsne .
the effects of the interaction of neutrinos and matter , which at different locations in the pns can be described either as an effective viscosity or by a drag term , affect the properties of the mri , in particular its growth rate and wavelength @xcite .
consequently , this interaction should induce modifications of the termination amplitude . on the other hand
, we do not expect these effects to change the most important qualitative result of our study , namely that the mri is not terminated by tm , but by the kh instability .
the presence of background flows in ccsne ( like convection or turbulence ) can affect the growth and termination of the mri , too .
in 2d global simulations , @xcite observed the growth of coherent channel flows , albeit deformed , while @xcite observed the disruption of channel flows on timescales of milliseconds due to the presence of a dynamical background .
both , the parasitic instabilities and the effect of the background flow , seem to disrupt channel flows on similar timescales .
therefore , it is likely that both effects will be of relevance in studying the mri termination in global simulations .
furthermore , the presence of parasitic instabilities implies the existence of a maximum magnetic stress at termination .
this limits the ability of the mri to amplify the magnetic field to dynamically relevant values in ccsne .
the turbulence triggered by the mri may amplify the global magnetic field further , if conditions for dynamo action are encountered .
a study of the latter process is beyond the scope of this work , and will be addressed in a future publication .
the mri has been traditionally invoked in the ccsn community as a justification to use supernova progenitors with an artificially enhanced initial magnetic field strength , @xmath397@xmath398 times larger than the values expected in stellar evolution models .
this argumentation should be revisited if we want to move towards a more realistic modeling of magnetorotational core collapse .
unfortunately , given the minimum numerical requirements that we provide in this work , 3d simulations seem nowadays unfeasible . to resolve mri channels of @xmath399 m length scale in a proto - neutron star of about @xmath400 km radius , with a resolution of @xmath401zones per channel , it would require about @xmath402grid zones .
evolving this system for a typical dynamical timescale of the proto - neutron star evolution , @xmath403ms , would require about @xmath404 time steps , several zettabytes of memory and would consume @xmath405 cpu hours ( @xmath406iterations times @xmath402zones times @xmath407operations / zone at @xmath408gflops ) . with current petascale supercomputers
this would amount to @xmath409yr of uninterrupted computation with @xmath410cores .
a simulation which includes the whole iron core with a radius of @xmath411 km would be even more demanding .
reducing the dimensionality of the system by imposing axisymmetry reduces the computational time by a factor @xmath412 .
the use of adaptive mesh refinement ( amr ) techniques would alleviate , but not solve the problem . the ccsn community has been performing 2d and 3d simulations with resolutions of up to @xmath322 m ( 2d ; * ? ? ? * ) and @xmath413 m ( 3d ; * ? ? ? * ) .
we note that these authors could afford such high resolutions only by accepting other limitations in terms of geometry and input physics , while other simulations use resolutions of several 100 m @xcite , which even for the artificially enhanced initial magnetic field , fail to resolve the termination process of the mri . given the above computational requirements , the community should rethink the way to model this scenario .
the agreement that we have found here between local simulations and the parasitic instability analysis of gx94 and @xcite make us hope that there could be a way out . if we can understand how mri and turbulence work at sub - meter scales , this information could be incorporated in global simulations using appropriate subgrid models . as a first step along this path
we plan to gain a deeper understanding in the maximum magnetic stress achievable by the mri and in the properties of its turbulent saturated state .
tr acknowledges support from the international max planck research school on astrophysics at the ludwig maximilians university munich , em & tr acknowledge support from the max - planck - princeton center for plasma physics , and ma , pcd , tr and mo acknowledge support from the european research council ( grant camap-259276 ) .
we also acknowledge support from grants aya2013 - 40979-p and prometeoii/2014 - 069 .
the authors thank rob yates for proof reading the manuscript , and m. pessah , c. mcnally , and h. latter and j. guilet for helpful discussions .
the authors thank the anonymous referee whose useful remarks greatly improved the quality of this manuscript .
the computations have been performed at the leibniz supercomputing center of the bavarian academy of sciences and humanities ( lrz ) , the rechenzentrum garching of the max - planck - gesellschaft ( rzg ) , and at the servei dinformtica of the university of valencia .
, t. a. & stone , j. m. 2005 , in american institute of physics conference series , vol .
784 , magnetic fields in the universe : from laboratory and stars to primordial structures . ,
e. m. de gouveia dal pino , g. lugones , & a. lazarian , 475488 | the magnetorotational instability ( mri ) can be a powerful mechanism amplifying the magnetic field in core collapse supernovae . whether initially weak magnetic fields can be amplified by this instability to dynamically relevant strengths is still a matter of debate .
one of the main uncertainties concerns the process that terminates the growth of the instability .
parasitic instabilities of both kelvin - helmholtz and tearing - mode type have been suggested to play a crucial role in this process , disrupting mri channel flows and quenching magnetic field amplification .
we perform two - dimensional and three - dimensional sheering - disc simulations of a differentially rotating proto - neutron star layer in non - ideal magnetohydrodynamics with unprecedented high numerical accuracy , finding that kelvin - helmholtz parasitic modes dominate tearing modes in the regime of large hydrodynamic and magnetic reynolds numbers , as encountered close to the surface of proto - neutron stars .
they also determine the maximum magnetic field stress achievable during the exponential growth of the mri .
our results are consistent with the theory of parasitic instabilities based on a local stability analysis . to simulate the kelvin - helmholtz instabilities
properly a very high numerical resolution is necessary . using 9th order spatial reconstruction schemes
, we find that at least @xmath0 grid zones per mri channel are necessary to simulate the growth phase of the mri and reach an accuracy of @xmath1 in the growth rate , while more than @xmath2 zones per channel are required to achieve convergent results for the value of the magnetic stress at mri termination .
[ firstpage ] accretion , accretion discs - mhd - instabilities - stars : magnetic field - general |
an important goal of cosmology is to understand how galaxies form . in standard hierarchical models
dissipationless dark matter aggregates into larger and larger clumps as gravitational instability amplifies the weak density perturbations produced at early times .
gas associated with such dark haloes cools and condenses within them , eventually forming the galaxies we see today .
these two aspects of galaxy formation are currently understood at very different levels .
gravitational processes determine the abundance , the internal structure and kinematics , and the formation paths of the dark haloes , can be simulated in great detail using n - body methods .
well - tested analytic models are now available for describing these properties of the halo population .
the growth of the dark haloes is not much affected by the baryonic components , but determines how they are assembled into nonlinear units .
gas cools and collects at the centre of each dark halo until it produces an independent self - gravitating unit which can form stars , heating and enriching the rest of the gas , and perhaps even ejecting it from the halo .
these star - formation and feedback processes are not understood in detail and can be modeled or simulated only crudely .
furthermore , as haloes merge to form groups and clusters of galaxies , the cluster members interact and merge both with each other and with the intracluster gas .
it will not soon be possible to simulate galaxy formation _
ab initio_. at least the star - formation and feedback processes must be put in _ ad hoc _ based on observations of real galaxies . in a first exploration of the galaxy population expected in this hierarchical picture white & rees ( 1978 ) calculated the expected luminosity function .
the observed characteristic luminosity of galaxies could be reproduced provided ( i ) that feedback prevents efficient conversion of gas into stars in early objects , and ( ii ) that merging of galaxies is inefficient as groups and clusters of galaxies build up .
the models appeared to overproduce faint galaxies .
the formation of galaxy disks within this scheme was studied by fall & efstathiou ( 1980 ) .
they showed that if galactic spin is produced by tidal torques acting at early times , then extended massive haloes are necessary if disks as large as those observed are to form by the present day .
they also found that plausible initial angular momentum distributions for the gas could lead to near - exponential density profiles for the final disk .
more recent analytic work has refined the calculation of global properties of galaxies , for example the distributions of luminosity , colour , morphology and metallicity can all be calculated explicitly in currently popular cosmogonies ( e.g. white & frenk 1991 ; kauffmann , white & guiderdoni 1993 ; cole et al 1994 ; baugh et al 1997 ) .
there has , however , been relatively little work on the internal structure of galaxies .
kauffmann ( 1996b ) studied the accumulation , star - formation and chemical enrichment histories of individual disks , linking them to damped ly@xmath0 absorbers at high redshift , but she did not consider the scatter in angular momentum for disks of a given mass and the resulting scatter in their present - day properties .
dalcanton , spergel & summers ( 1997 , hereafter dss ) considered the latter issue in some detail , but did not tie their model explicitly to its cosmogonical context or consider its implications for disk evolution .
there have also been surprisingly few well - resolved simulations of galaxy formation from initial conditions which properly reflect the hierarchical model .
so far all have failed to produce viable models for observed spirals because excessive transfer of angular momentum from gas to dark matter results in overly small disks ( e.g. navarro & steinmetz 1997 ) . in the present paper
, we again address the problem of disk formation in hierarchical cosmogonies , formulating a simple model based on four key assumptions : 1 ) the mass of a disk is some fixed fraction of that of the halo in which it is embedded ; 2 ) the angular momentum of the disk is also a fixed fraction of that of its halo ; 3 ) the disk is a thin centrifugally supported structure with an exponential surface density profile ; and 4 ) only dynamically stable systems can correspond to real galaxy disks .
we derive the abundance of haloes in any specific cosmogony from the press - schechter formalism , we model their density profiles using the precepts of navarro , frenk & white ( 1996 , 1997 hereafter nfw ) , and we assume their angular momentum distribution to have the near - universal form found in n - body simulations .
( see lacey & cole 1996 for a summary of how well these analytic models fit simulation results . )
our four assumptions are then sufficient to predict the properties of the disk population , once values are assumed for the mass and angular momentum fractions and the redshift of disk formation . a mass - to - light ratio for the disk stars must also be assumed in order to compare with the luminosities of observed disks .
our assumptions are similar to those of dss but differ in detail .
we calculate many of the same properties as they do , but in addition we address broader issues such as the evolution of disks , the dependence of the disk population on cosmological parameters , and the link with qso absorption lines .
the outline of our paper is as follows .
section 2 presents our modelling assumptions in detail . to clarify the scaling of disk properties with model parameters , we first treat haloes as singular isothermal spheres and
neglect the gravitational effects of the disks .
we then consider the more realistic nfw profiles and include the gravitational effects of the disks . in sections 3 to 5
, we describe the predicted properties of present - day disks , we discuss tf relations , we compare high redshift disks with qso absorption line systems , and we study the effect of a nonrotating central bulge .
finally in section 6 , we discuss other issues related to our model .
in cold dark matter cosmogonies the evolution of the dark matter is determined by the parameters of the background cosmological model and by the power spectrum of the initial density fluctuations .
the relevant cosmological parameters are the hubble constant @xmath6 , the total matter density @xmath7 , the mean density of baryons @xmath8 , and the cosmological constant @xmath9 .
the last three are all expressed in units of the critical density for closure .
the power spectrum @xmath10 is specified in cdm models by an amplitude , conventionally quoted as @xmath11 , the _ rms _ linear overdensity at @xmath12 in spheres of radius @xmath13 mpc , and by a characteristic scale @xmath14 , where @xmath15 is the current radiation density in units of the standard value corresponding to a @xmath16k black - body photon field and equilibrium abundances of three massless neutrinos . for such cosmogonies
the abundance of haloes as a function of mass and redshift can be calculated quite accurately using the press - schechter formalism , while their radial density profiles can be calculated as a function of these same variables using the results of navarro et al ( 1997 ) .
we give some details of how we implement these formalisms in the appendix . in the main body of this paper
we will compare results for four different cdm cosmogonies .
these are specified by particular values of the parameter set ( @xmath17 ) .
we do not explicitly need to assume a value for the baryon density @xmath8 , although for consistency we should require @xmath18 where @xmath19 is the ratio of disk mass to halo mass adopted ( and assumed universal ) in our models .
this ensures that the disk mass is less than the total initial baryon content of the halo .
our four cosmogonies are : 1 . the standard cold dark matter model ( scdm ) , with @xmath20 , @xmath21 , @xmath22 , @xmath23 , and @xmath24 ; 2 . a cobe - normalized cold dark matter model ( ccdm ) , which has the same parameters as scdm , except that @xmath25 ; 3 . a @xmath26 model , with @xmath20 , @xmath21 , @xmath22 , @xmath27 , and @xmath24 ; and 4 . a @xmath28 model , with @xmath29 , @xmath30 , @xmath31 , @xmath27 , and @xmath32 .
the @xmath26 model could be realised if the @xmath33-neutrino were unstable to decay into other neutrinos and had a mass and lifetime satisfying @xmath34kev yr ( efstathiou et al 1992 ) . on the relevant scales
the power spectrum in this model is quite similar to that of a mixed dark matter ( mdm ) model with 20 per cent of the critical density in stable neutrinos .
the values of @xmath11 adopted for @xmath26 and @xmath28 are consistent both with the observed abundance of rich clusters ( see e.g. white , efstathiou & frenk 1993 ; mo , jing & white 1996 ) and with observed anisotropies in the cosmic microwave background ( see e.g. white & scott 1996 ) , whereas the value for scdm is consistent only with the cluster normalization and that for ccdm only with the cobe normalization .
many of the properties we predict for galaxy disks can be understood using a very simple model in which dark haloes are treated as singular isothermal spheres and the gravitational effects of the disks themselves are neglected .
we will examine the limitations of this rather crude treatment in section 2.3 .
for a singular isothermal sphere the density profile is just , ( r)=^24 g r^2 , where the circular velocity @xmath35 is independent of radius . based on the spherical collapse model ( gunn & gott 1972 ; bertschinger 1985 ; cole & lacey 1996 ) , we define the limiting radius of a dark halo to be the radius @xmath36 within which the mean mass density is @xmath37 , where @xmath38 is the critical density for closure at the redshift @xmath39 when the halo is identified .
thus , the radius and mass of a halo of circular velocity @xmath35 seen at redshift @xmath39 are [ rh_h ] = 10 h(z ) ; = ^2g=^3 10 g h(z ) , where [ hubblec ] h(z)=h_0 ^ 1/2 is the hubble constant at redshift @xmath39 .
the redshift dependence of halo properties , and so of the disks which form within them , is determined by @xmath40 . in order to understand this dependence better ,
figure 1 shows @xmath41 as a function of @xmath39 for open and flat cosmologies with various @xmath7 .
@xmath40 increases more rapidly with @xmath39 in universes with larger @xmath7 . for a given @xmath7 ,
the increase is more rapid for an open universe than for a flat universe . at @xmath42 ,
the ratio @xmath41 is about twice as large in an einstein - de sitter universe as in a flat universe with @xmath29 . as a result of this difference
, the properties of high redshift disks will depend significantly on @xmath7 and @xmath9 .
we assume that the mass which settles into the disk is a fixed fraction @xmath43 of the halo mass .
the disk mass is then [ md_sis ] = ^310 g h(z ) 1.710 ^ 11 h^-1(0.05)(250)^3 ^-1 . except in section 5 where we investigate the effects of including a central bulge
, we distinguish only halo and disk components , and we assume the disks to be thin , to be in centrifugal balance , and to have exponential surface density profiles , [ ex_disk ] ( r)=_0 ( -r/ ) . here
@xmath44 and @xmath45 are the disk scalelength and central surface density , and are related to the disk mass through [ srd ] = 2_0 ^2 .
if the gravitational effect of the disk is neglected , its rotation curve is flat at the level @xmath35 and its angular momentum is just [ jdtwo_sis ] = 2(r ) r^2 d r=4_0 ^3 = 2m_dr_d .
we assume this angular momentum to be a fraction @xmath46 of that of the halo , i.e. [ jdone_sis ] = , and we relate @xmath47 to the spin parameter @xmath48 of the halo through the definition [ jh_sis ] = ^1/2 g^-1 ^-5/2 , where @xmath49 is the total energy of the halo .
equations ( [ jdtwo_sis])-([jdone_sis ] ) then imply that [ rdone_sis ] = g^3/22 ^ 1/2 ( ) . the total energy of a truncated singular isothermal sphere is easily obtained from the virial theorem by assuming all particles to be on circular orbits : [ eh_sis ] = -g^22 = -m^22 .
on inserting this into equation ( [ rdone_sis ] ) and using equations ( [ rh_h ] ) and ( [ srd ] ) we get [ rd_sis ] = 1 ( ) 8.8 h^-1(0.05 ) ( 250 ) ^-1 ( ) , and [ sig_sis ] _
04.810 ^ 22h ^-2 ( 0.05 ) ( 0.05)^-2 ( 250 ) ( ) ^2 , where @xmath50 is the mass of a hydrogen atom . using equations ( [ ex_disk ] ) , ( [ rd_sis ] ) and ( [ sig_sis ] ) ,
we can also obtain the limiting radius , @xmath51 , at which the disk surface density drops below some critical hydrogen column @xmath52 ( which must be less than @xmath53 ) : [ rl ] r_l5.5 + .
this radius depends on @xmath36 , @xmath48 and @xmath54 primarily through the scalelength , @xmath44 .
equations ( [ md_sis ] ) and ( [ rd_sis])-([rl ] ) give the scalings of disk properties with respect to a variety of physical parameters .
these properties are completely determined by the values of @xmath35 , @xmath43 , @xmath46 , @xmath48 and @xmath40 ; other cosmological parameters , such as @xmath39 , @xmath7 and @xmath9 , affect disks only indirectly through @xmath40 .
since @xmath40 increases with @xmath39 , disks of given circular velocity are less massive , smaller and have higher surface densities at higher redshifts . at a given redshift ,
they are larger and less compact in haloes with larger @xmath48 , because they contract less before coming to centrifugal equilibrium .
we will approximate the distribution of @xmath48 by [ plambda ] p ( ) d=1 _ , where @xmath55 and @xmath56 .
this function is a good fit to the n - body results of warren et al ( 1992 ; see also cole & lacey 1996 ; steinmetz & bartelmann 1995 ) except possibly at very small @xmath48 ; note that it is narrower than the distribution used by dss .
since the dependence of @xmath57 on @xmath58 and @xmath10 is weak , we will not consider it further .
this model for @xmath57 is plotted as the dotted curve in figure 10 .
the distribution peaks around @xmath59 , and has a width of about 0.05 .
its 10 , 50 and 90 percent points are near @xmath60 0.05 , and 0.1 , respectively .
thus we see from equations ( [ rd_sis ] ) and ( [ sig_sis ] ) that about 10% of disks of given circular velocity will be more than twice the size of the typical disk , and so have less than a quarter the typical surface density , while about 10% will be less than half the typical size , and so have more than four times the typical surface density .
( the latter 10% may be unstable see section 3.2 . )
disk masses are , of course , directly proportional to @xmath61 , the fraction of the halo mass we assume them to contain .
for simplicity we will assume this fraction to be the same for all disks , but even in this case it is unclear what value of @xmath61 is appropriate .
a plausible upper limit is the baryon fraction of the universe as a whole , @xmath62 , but the efficiency of forming disks could be quite low , implying that @xmath61 could be substantially smaller than @xmath63 . taking a big bang nucleosynthesis value for the baryon density , @xmath64 ( e.g. walker et al 1991 , but see turner 1996 ) gives an upper limit on @xmath61 in the range 0.05 to 0.1 for the cosmological models we consider ( see section 2.1 ) . if the specific angular momentum of the material which forms the disk is the same as that of the halo , then @xmath65
this has been the standard assumption in modelling disk formation since the work of fall & efstathiou ( 1980 ) , but it is unclear if it is appropriate , particularly if the the efficiency of disk formation , @xmath66 , is significantly less than unity .
even when this efficiency is high , numerical simulations of spiral galaxy formation have tended to find @xmath54 values well below unity ( e.g. navarro & white 1994 ; navarro & steinmetz 1997 ) . equation ( [ rd_sis ] ) shows that disk sizes are directly proportional to @xmath54 , and we will find below that values close to unity are required to fit observed spirals .
this agrees with the fact that the simulations did indeed produce galaxies with too little angular momentum .
most of our discussion will assume @xmath67 , but we will comment on the effects of changing @xmath43 and @xmath46 independently .
the disk of our own galaxy has a mass of about @xmath68 .
its scalelength and circular velocity ( measured at the solar radius , @xmath69 ) are @xmath70 and @xmath71 ( see binney & tremaine 1987 , p17 ) . since the milky way appears to be a typical sbc galaxy , it is interesting to see whether a disk with these parameters is easily accommodated .
equation ( [ md_sis ] ) together with figure 1 implies that milky way - like disks can not form at @xmath72 in a universe with @xmath73 ; for our nucleosynthesis value of @xmath63 the disk mass for @xmath74 is too small even if disks are assumed to form with maximal efficiency .
the constraints are weaker for a low-@xmath7 universe , because @xmath63 is larger and @xmath40 increases less rapidly with @xmath39 .
a similar conclusion can be derived from @xmath44 . at high
redshift the disk scalelength for @xmath74 is too small to match the observed value unless @xmath75 .
such high values of @xmath48 are associated with only a small fraction of haloes , and so late formation is required for the bulk of the disk population . as we will see
later these constraints become stronger in more realistic models .
if we assume that all disks have the same stellar mass - to - light ratio , @xmath76 ( in solar units ) , we can cast equation ( [ md_sis ] ) into a form similar to the tf relation : [ tf_sis ] = a(250)^ , where @xmath77 is the slope , and [ zerop_sis ] a=1.710 ^ 11 h^-1 ^ -1 ( 0.05 ) ^-1 is the ` zero - point ' .
notice that @xmath48 does not appear in these equations .
as a result , all the disks assembled at a given time are predicted to lie on a scatter - free tf relation despite the fact that there is substantial scatter in their scale radii and surface densities .
the zero - point of this tf relation can , however , depend on the redshift of assembly , being lower at higher redshifts if the product @xmath78 is larger . for i -
band luminosities the observed value of @xmath0 is very close to 3 ( willick et al 1995 ; 1996 ; giovanelli et al 1997 ) .
since stellar population synthesis models suggest that @xmath79 should vary rather little in this band , we can take this as confirmation of the approximate validity of our assumption that @xmath43 is independent of @xmath35 .
the zero - point of the observed tf relation corresponds to a value of @xmath80 of about @xmath81 ( giovanelli et al 1997 ) . since the expected value of the disk mass - to - light ratio is @xmath82 ( bottema 1997 , see section 3.4 ) , comparison with equation ( [ zerop_sis ] )
again suggests that late assembly epochs are required ; for the appropriate values of @xmath61 the predicted zero - point would be too low if disks formed at high redshift .
the limiting radius @xmath51 which we defined in equation ( [ rl ] ) is chosen to be relevant to the damped ly@xmath0 absorption lines seen in qso spectra .
such systems have hi column densities of @xmath83 , comparable to those of present - day galactic disks ( see wolfe 1995 for a review ) . for given @xmath35 and @xmath48
disks are smaller at higher redshifts . as a result the cross - section for damped ly@xmath0 absorption decreases with @xmath39 .
as one can see from equation ( [ rl ] ) , the limiting radius for damped absorption is roughly @xmath84 . for @xmath85 and @xmath86
, we find @xmath51 between @xmath87 and @xmath88 at @xmath89 for @xmath7 between 1 and 0.3 .
given the distribution of dark haloes as a function of @xmath48 and @xmath35 , it is clearly possible to calculate the total cross - section for damped absorption along a random sightline , as well as the distributions of @xmath48 , @xmath35 , @xmath90 and impact parameter ( i.e. the distance between the centre of the absorbing disk and the sightline to the qso ) for the systems actually seen as damped absorbers .
we discuss this in detail in section 4 . the model presented in this subsection
is limited both because the dark haloes formed by dissipationless hierarchical clustering are only very roughly approximated by singular isothermal spheres , and because the gravitational effects of the disks are often not negligible . despite this , most of the characteristic properties of this model are retained with only rather minor modifications in the more realistic model which we discuss next . in a recent series of papers navarro , frenk & white ( 1995 ; 1996 ; 1997 , hereafter nfw ) used high - resolution numerical simulations to show that the equilibrium density profiles of dark matter haloes of all masses in all dissipationless hierarchical clustering cosmogonies can be well fit by the simple formula : @xmath91 where @xmath92 is a scale radius , and @xmath93 is a characteristic overdensity .
the logarithmic slope of this profile changes gradually from @xmath94 near the centre to @xmath95 at large radii . as before ( and following nfw )
we define the limiting radius of a virialised halo , @xmath36 , to be the radius within which the mean mass density is @xmath37 .
the mass within some smaller radius @xmath96 is then [ mh_nfw ] ( r ) = 4 r_s^3 , where @xmath97 , and [ concentration ] c is the halo concentration factor .
the total mass of the halo is given by equation ( [ mh_nfw ] ) with @xmath98 .
using equations ( [ nfw])-([concentration ] ) , we can obtain a relation between the characteristic overdensity and the halo concentration factor , [ delta_null ] = 200 3 c^3 ( 1+c)-c/(1+c ) .
we can obtain the total energy of the truncated halo as before by assuming all particles to be on circular orbits , calculating their total kinetic energy , and then using the virial theorem .
this yields [ eh_nfw ] e = - g ( 4 r_s^3)^22r_s = - g ^2 2 f_c , where f_c = c 2 1 - 1/(1+c)^2 - 2(1+c)/(1+c ) ^2 + ( c 21.5)^0.7 . the approximation for @xmath99 given above is accurate to within 1% for the relevant range @xmath100 ( to about 3% for @xmath101 ) .
comparing equation ( [ eh_nfw ] ) with equation ( [ eh_sis ] ) , we see that the total energy for an nfw profile differs from that for an isothermal sphere by the factor , @xmath99 , which depends only on the concentration factor .
if the gravitational effects of the disk were negligible , its rotation curve would simply follow the circular velocity curve of the unperturbed halo , @xmath102 .
this curve rises to a maximum at @xmath103 and then falls gently at larger radii ( see nfw ) .
in fact , disk formation alters the rotation curve not only through the direct gravitational effects of the disk , but also through the contraction it induces in the inner regions of the dark halo .
we analyse this effect by assuming that the halo responds adiabatically to the slow assembly of the disk , and that it remains spherical as it contracts ; the angular momentum of individual dark matter particles is then conserved and a particle which is initially at mean radius , @xmath104 , ends up at mean radius , @xmath96 , where [ action ] g _
f(r ) r = g ( r_i ) r_i . in this formula @xmath105
is given by the nfw profile ( [ mh_nfw ] ) and @xmath106 is the total final mass within @xmath96 .
[ see barnes & white ( 1984 ) for a test of this adiabatic model . ]
the final mass is the sum of the dark matter mass inside the initial radius @xmath104 and the mass contributed by the exponential disk .
hence [ adiabatic ] _
f(r ) = ( r ) + ( r_i ) ( 1- ) , where , as before , @xmath43 , is the fraction of the total has in the disk , and [ md_r ] ( r ) = . here
we implicitly assume that the baryons initially had the same density profile as the dark matter , and those which do not end up in the disk remain distributed in the same way as the dark matter . for a given rotation curve , @xmath107 , the total angular momentum of the disk
can be written as j_d = _ 0^ ( r ) r ( r ) 2 r dr = 200 _
0^/r_d e^-u u^2 ( u ) 200 du , where @xmath108 is unaffected by disk formation .
in practice we can set the upper limit of integration to infinity because the disk surface density drops exponentially and @xmath109 .
substituting equation ( [ eh_nfw ] ) into equation ( [ jh_sis ] ) and writing @xmath110 , we can use the argument of section 2.2 to obtain [ rd_nfw ] = 1 ( ) f_c^-1/2 ( , c , , ) where [ fr_nfw ] ( , c , , ) = 2 ^-1 . comparing equation ( [ rd_nfw ] ) with equation ( [ rd_sis ] )
we see two effects that cause the disk scalelength to differ from that in a singular isothermal sphere ; the factor @xmath111 comes from the change in total energy resulting from the different density profile , while the factor @xmath112 is due both to the different density profile and to the gravitational effects of the disk .
the rotation velocity @xmath113 is a sum in quadrature of contributions from the disk and from the dark matter : [ vc_nfw ]
^2(r ) = v_c , d^2(r ) + v_c , dm^2(r ) , where [ vc_bdm ] v_c , dm^2(r)=g / r , and @xmath114 is the rotation curve which would be produced by the exponential disk alone .
note that when calculating the latter quantity the flattened geometry of the disk has to be taken into account ( binney & tremaine 1987 , p77 ) . for a given set of parameters , @xmath115 , @xmath116 , @xmath48 , @xmath43 and @xmath46 , equations ( [ adiabatic ] ) , ( [ md_r ] ) , ( [ rd_nfw ] ) and ( [ vc_nfw ] ) must be solved by iteration to yield the scale length , @xmath44 , and the rotation curve , @xmath107 .
for example , we can start with a guess for @xmath44 by setting @xmath117 in equation ( [ rd_nfw ] ) .
we can then obtain @xmath118 from equation ( [ md_r ] ) . substituting this into equation ( [ adiabatic ] ) and using equation ( [ action ] )
, we can solve for @xmath104 as a function of @xmath96 and so obtain @xmath106 from equation ( [ adiabatic ] ) . with this and
with @xmath119 calculated for the assumed @xmath44 , we can get the disk rotation curve from equations ( [ vc_nfw ] ) and ( [ vc_bdm ] ) . inserting this into equation ( [ fr_nfw ] ) and using equation ( [ rd_nfw ] )
we then obtain a new value for @xmath44 . in practice
this iteration converges rapidly , so that accurate values for both @xmath44 and @xmath120 are easily obtained .
it is useful to have a fitting formula for @xmath121 which allows this iterative procedure to be avoided .
we find that the following expression has sufficient accuracy : [ fr_appr ] ( ^0.1)^-0.06 + 2.71 + 0.0047/^ ( 1 - 3 + 5.2 ^2 ) ( 1 - 0.019 c+0.00025 c^2 + 0.52/c ) , where @xmath122 .
we will see in section 3.1 that the rotation curves predicted by our procedure typically reach a maximum near @xmath123 and in much of the discussion which follows we will use the value at this point to characterise their amplitude . as a result , it is also useful to have a fitting formula for the dimensionless coefficient @xmath124 in [ vcc_nfw ] ( 3 ) = ( g ) ^1/2 = 200 .
we find [ fv_appr ] ( ^0.1)^-2.67 - 0.0038/^+0.2^ ( 1 + 4.35 - 3.76 ^2 ) 1 + 0.057 c-0.00034 c^2 - 1.54/c ^1/2
. the approximations in ( [ fr_appr ] ) and ( [ fv_appr ] ) are accurate to within 15% for @xmath100 , @xmath125 and @xmath126 .
according to the model set out in section 2.3 the shape of a disk s rotation curve depends on the concentration of its halo , @xmath116 , on the fraction of the halo mass which it contains , @xmath61 , and on its angular momentum as specified by the parameter combination @xmath127 . for given values of these three parameters
the amplitude of the rotation curve and its radial scale are set by any single scale parameter , either for the halo ( @xmath58 , @xmath36 , or @xmath115 ) or for the disk ( @xmath128 , @xmath129 , or @xmath45 ) .
figure 2 shows examples chosen to illustrate how changes in @xmath61 , @xmath130 and @xmath116 affect the shape of the rotation curve of a disk of fixed mass .
curves for other disk masses can be obtained simply by scaling both axes by @xmath131 . in all cases
the rotation curves are flat at radii larger than a few disk scalelengths . at smaller
radii the shape of the curves depends strongly on the spin parameter . for small values of @xmath130
the disk is more compact , and its self - gravity is more important .
the rotation velocity then increases rapidly near the centre to a peak near @xmath132 , thereafter dropping gradually towards a plateau at larger radii . for large @xmath130 ,
the disk is much more extended and its self - gravity is negligible .
the rotation velocity then increases slowly with radius in the inner regions .
since the surface density of a disk is roughly proportional to @xmath133 , these results imply a correlation between disk surface brightness and rotation curve shape which is very reminiscent of the observational trends pointed out by casertano & van gorkom ( 1991 ) .
our predicted rotation curves also become more peaked at small radii for larger values of @xmath61 , again because of the larger gravitational effect of the disk .
rotation curves as peaked as that shown in the lower left panel of figure 2 are not observed . as we will discuss below , this is probably because such disks are violently unstable .
halo concentration affects rotation curves in the obvious way ; more strongly peaked curves are found in more concentrated haloes . combining these trends
we conclude that slowly increasing rotation curves are expected for low - mass disks in haloes with large @xmath130 and small @xmath116 . since @xmath116 is smaller for more massive haloes ( see nfw ) , we would predict slowly rising rotation curves to occur preferentially in giant galaxies . in fact , the opposite is observed .
this is almost certainly a result of the greater influence of the bulge in giant systems .
a final trend which is clear from figure 2 is that at fixed disk mass more strongly peaked curves have larger maximum rotation velocities .
a similar trend can be seen in the observational data ( e.g. persic and salucci 1991 ) .
the modelling of the last two sections can lead to disks with a very wide range of properties .
not all of these disks , however , are guaranteed to be physically realisable .
in particular , those in which the self - gravity of the disk is dominant are likely to be dynamically unstable to the formation of a bar .
there is an extensive literature on such bar instabilities [ see christodoulou , shlosman & tohline ( 1995 ) and references therein ] . for our purposes
the most relevant study is that of efstathiou , lake & negroponte ( 1982 ) who used n - body techniques to investigate global instabilities of exponential disks embedded in a variety of haloes .
they found the onset of the bar instability for stellar disks to be characterised by the criterion , [ epsilon_m ] _
m1.1 , where @xmath134 is the maximum rotation velocity of the disk .
the appropriate instability threshold for gas disks is lower , @xmath135 , as discussed in christodoulou et al ( 1995 ) . in our model , _ m ( ^)^1/2 f_c^-1/4f_r^1/2 f_v .
thus , disks are stable if [ lambda_crit ] ^^_crit= _ m , crit^2
f_c^1/2f_r^-1 f_v^-2 , where @xmath136 is the critical value of @xmath137 for disk stability .
if the effect of disk self - gravity on @xmath44 and @xmath35 is weak , then @xmath138 .
figure 3 shows @xmath139 as a function of @xmath43 for our nfw models .
results are shown for @xmath140 , 1 , and 1.2 , in order to bracket the critical values of @xmath141 discussed above . in practice
we will normally adopt @xmath142 as a fiducial value .
the dependence of @xmath139 on halo concentration is weak , because for a given @xmath43 , the dependences of @xmath134 and @xmath44 on @xmath116 are opposite .
since halo mass and assembly time affect @xmath139 only through @xmath116 , we expect disk stability to be almost independent of these parameters .
as one can see from figure 3 , it is a useful rough approximation to take @xmath143 as the condition for stability .
thus in figure 2 the top two disks are marginally stable , the lower left disk is strongly unstable and the lower right disk is stable .
disk galaxies are common and appear to be stable . if we take account of the expected distribution of @xmath48 ( equation [ plambda ] ) and the likelihood that @xmath144 , then the results in figure 3 suggest that @xmath43 values of 0.05 or less will be needed to ensure that most haloes host stable disks rather than the descendants of unstable disks . for a given cosmogony
our assumptions determine the joint distribution of disk size and rotation speed once values are adopted for the mass and angular momentum fractions , @xmath61 and @xmath46 , and for the disk formation redshift . this distribution can be compared directly with observation provided our size measure , @xmath44 ,
can be identified with the exponential scalelength of observed luminosity profiles .
this seems reasonable since there is no evidence that the mass - to - light ratio of the stellar populations in real disks are a strong function of radius . in our simplified model the disk `` formation redshift '' is just the time at which the arguments of section 2 are applied .
we make no attempt to follow the actual formation and evolution of disks , and we assume that all disks form at the same time .
this formation redshift should be interpreted as the epoch when the material actually observed was assembled into a single virialised object .
figure 4 compares predicted distributions of @xmath44 as a function of @xmath35 with observational data for a large sample of nearby spirals .
the four panels give predictions for formation redshifts of 0 and 1 , and for two different cosmogonies , scdm and @xmath28 ; predictions for ccdm and @xmath26 are similar to those for scdm . as a characteristic measure of rotation speed for our model galaxies we use the value of @xmath35 at @xmath123 .
as discussed in 2.3 this is always close to the maximum of the rotation curve .
the solid line in each panel is the @xmath44-@xmath35 relation for _ critical _ disks ( i.e. those with @xmath145 ) for the case @xmath146 ; stable disks must lie above this line .
since @xmath147 in this case , and since , in general , we expect @xmath148 , implying @xmath149 , at most half of all dark haloes will contain stable disks for @xmath150 ( see the discussion around equation [ plambda ] ) .
the short - dashed line in each panel is the corresponding critical line for @xmath151 . in this case
about 90 percent of the dark haloes can host stable disks under the optimistic assumption that @xmath152 .
the critical lines would be even lower if @xmath153 , but values of @xmath48 below 0.025 are rare , and so such compact disks will not be abundant unless @xmath46 is often substantially smaller than @xmath61 .
finally , the long - dashed lines in figure 4 show @xmath44-@xmath35 relations for @xmath154 and @xmath155 . for these parameters
disk self - gravity is negligible and the line is almost unaltered for @xmath156 .
fewer than 10 percent of the dark haloes have @xmath157 in the @xmath48-distribution of equation ( [ plambda ] ) . thus for @xmath148 very few disks
are expected to lie above these long - dashed lines .
the slopes and relative amplitudes of all the theoretical lines in figure 4 can easily be understood from the scalings in equation ( [ rd_sis ] ) .
the data points in figure 4 are the observational results of courteau and coworkers ( courteau 1996 ; 1997 ; broeils & courteau 1996 ) for _ present - day _ normal spirals . for a formation redshift of zero ,
the observed distribution lies comfortably below the long dashed lines and above the short dashed lines in both cosmologies .
a significant fraction of the disks lie below the solid lines , particularly for @xmath28 , suggesting that these disks must have @xmath158 in order to be stable . for this formation
redshift it seems relatively easy to reproduce the observed distribution in both cosmogonies . for a formation redshift of 1 , in contrast ,
the observed points clearly lie too high relative to the predictions of the scdm model ( and similarly for ccdm and @xmath26 ) .
observed disks are too big to have formed at @xmath159 in a high density universe , unless there is some way for @xmath46 to be substantially larger than @xmath61 .
the observed points also lie rather high for @xmath160 in @xmath28 , but here the discrepancy is marginal provided that @xmath161 is indeed a realistic expectation .
it is clear that substantial transfer of angular momentum from baryons to dark matter , resulting in @xmath162 , would lead to unacceptably small disks for any assembly redshift in all of the cosmogonies we consider .
note that although the observed disks of figure 4 apparently formed late , this does not mean that disks were not present in large numbers at high redshift ; early disks should , however , be significantly smaller than present - day objects of the same @xmath35 . for the cosmogonies considered here , the comoving number density of galaxy - sized haloes is expected to peak at @xmath159 and then to decline gradually at lower redshift .
the peak density is comparable to the number density of galaxies observed today .
thus , as can be seen explicitly from more detailed semi - analytic modelling ( e.g. kauffmann et al 1993 ) these models do indeed predict enough haloes to house the observed population of disks .
late formation thus appears viable both in terms of the structural properties and the abundance of disks in cdm - like models . for a given cosmogony
the distributions of @xmath58 ( the halo mass ) and @xmath48 are known as a function of redshift [ see equations ( [ ps_mf ] ) and ( [ plambda ] ) ] , and furthermore the halo concentration factor , @xmath116 , can be calculated from @xmath58 and @xmath39 ( we neglect any scatter in @xmath116 , see the appendix ) . for any particular formation redshift
, we can therefore generate monte - carlo samples of the halo distribution in the @xmath58-@xmath48 plane , and , using specific values for @xmath61 and @xmath46 , transform these into monte - carlo samplings of the disk population .
figure 5 shows the distribution of disks in the @xmath35-@xmath45 plane for scdm with @xmath151 , 0.5 and 0.1 , @xmath67 and a formation redshift of zero .
unstable disks with @xmath163 are excluded .
the instability criterion leads to an upper cutoff in the central surface luminosity density @xmath45 . for the simple isothermal sphere model of 2.2
this upper cutoff can be written as _ crit h(z)h(z )
h(z ) , where we have used @xmath164 ( see equation [ lambda_crit ] ) . thus @xmath165 is larger for higher @xmath39 , for higher @xmath35 and for lower @xmath43 .
the last two dependences are clearly seen in figure 5 . to compare the distributions of figure 5 with real data we need to assume a stellar mass - to - light ratio for the disks . throughout this article
we will use the values found for high surface brightness disks by bottema ( 1997 ) . from disk dynamics , he inferred @xmath166 for @xmath167 . since @xmath168 for a typical disk galaxy ( mcgaugh & de blok 1996 ) , and @xmath169 , we obtain @xmath170 . solid squares in figure 5 correspond to median values for the courteau data in figure 4 .
we have converted from @xmath35 to @xmath171 using the tully - fisher relation of giovanelli et al ( 1997 ; equation [ zero - point ] ) , and from @xmath171 to @xmath45 using equation ( [ srd ] ) and @xmath172 .
error bars join the upper and lower 10% points of the distribution in each @xmath35 bin .
the typical central surface brightnesses found in this way are similar to the `` standard '' value advocated by freeman ( 1970 ) . for @xmath146 ,
the median central surface density for disks at @xmath12 in scdm appears to be about a factor of 2 too faint compared with the observed value
. however , this discrepancy should not be over - interpreted since low surface brightness galaxies are probably underrepresented in courteau s sample .
in addition , the model prediction has substantial uncertainties . for a given @xmath173 ,
the predicted disk surface density is proportional to @xmath40 , and so a formation redshift of 0.5 would increase the predictions by a factor of 1.8 . furthermore , the upper cutoff of the disk surface density distribution depends sensitively on the value of @xmath174 ( cf .
equation [ epsilon_m ] ) . for a given @xmath43 , @xmath175
; a 20% decrease in @xmath174 would thus increase the upper cutoff on @xmath45 by a factor of two . despite these uncertainties ,
figure 5 suggests that @xmath176 is preferred . notice that this conclusion is independent of the @xmath44-@xmath35 analysis of the last section which suggested similar values for @xmath43 .
it is interesting that the abundance of both high and low surface density galaxies is predicted to increase as @xmath61 decreases .
this is a consequence of the wider range of @xmath48 values which produce stable disks when @xmath43 is small .
there is currently considerable debate about the relative abundance of low - surface - brightness ( lsb ) galaxies .
our results confirm those of dalcanton et al ( 1997 ) ; hierarchical models can relatively easily produce a spread which appears broad enough to be consistent with the observations .
we note that an additional uncertainty in any detailed comparison comes from the fact that the star formation efficiency in disks may well vary strongly as a function of surface density .
this would further complicate the conversion between disk mass and disk light .
in one of the most complete recent studies of the tully - fisher relation for nearby galaxies giovanelli et al ( 1997 ) put together a homogeneous set of hi velocity profiles and i - band photometry for a set of 555 spiral galaxies in 24 clusters . by careful adjustment of the relative zero - points of the different clusters they derived a tf relation in the ( cousin ) i - band , @xmath177 where @xmath178 is the inclination - corrected width of the hi line profile , and , to a good approximation , is twice the maximum rotation velocity .
when comparing with our models we will assume @xmath179 ( see 2.3 ) .
giovanelli et al find a mean scatter around this relation of 0.35 magnitudes , but point out that the scatter is actually magnitude - dependent ; it increases significantly with decreasing velocity width .
a nearly identical tully - fisher relation was derived independently by shanks ( 1997 ) , but the zero - point of the tf relation in willick et al ( 1996 ) is fainter by about 0.5 mag . in order to derive luminosities for the disks in our models we need to specify their stellar mass - to - light ratio .
unfortunately , the appropriate value for this quantity is quite uncertain ; here we will again use the value @xmath180 which bottema ( 1997 ) derived for the disks of high surface brightness spirals . for a given mass - to - light ratio , the magnitude of a disk galaxy can be calculated as @xmath181 where we have used @xmath182 , and @xmath183 ( bessel 1979 , table ii ) .
note that in the observed tf relation @xmath184 is the total absolute magnitude of the galaxy , whereas for our models we are estimating the absolute magnitude of the disk alone ; we are therefore implicitly assuming that the bulge luminosity can be neglected for the galaxies under consideration ( but see 5 below ) . with these assumptions
, we can generate a monte - carlo disk catalogue , as in 3.4 , and plot @xmath185 versus @xmath186 ( @xmath187 ) .
figure 6 shows such plots for stable disks with a formation redshift of @xmath12 in scdm and @xmath28 . here
we assume @xmath154 .
for each cosmogony , the three panels show results for three choices of @xmath188 , the critical value of @xmath141 for disk instability ( equations [ epsilon_m ] to [ lambda_crit ] ) .
as one can see , the plots are similar in all cases .
the zero point for the scdm model is lower than that for the @xmath28 model for the reasons to be discussed in more detail in section 3.5.3 .
the scatter is slightly larger for smaller values of @xmath174 , because disks are then stable for smaller @xmath48 values where disk self - gravity is more important . by fitting plots such as these we can clearly derive tf relations for our models , and
so can compare their slope , scatter and zero - point with those of the observed tully - fisher relation .
we begin by considering the slopes of the model relations .
the solid lines in figure 6 are the linear regressions of @xmath185 on @xmath35 for each monte - carlo sample of simulated data .
it is clear that in all cases the slope is close to @xmath77 ( where @xmath0 is defined in equation [ tf_sis ] ) , and is consistent with the observed value of giovanelli et al which we indicate with a dashed line .
although there is clearly very good agreement , it is important to realise that this is a consequence of choosing to compare with the observed i - band data .
in fact , the observed slope is quite dependent on the photometric band used .
for example , in the b - band it is about @xmath189 ( strauss & willick 1995 and references therein ) .
this difference arises because the colours of disk galaxies vary systematically with their luminosity ; fainter galaxies show proportionately more star formation and are bluer .
stellar population models suggest that the stellar mass - to - light ratios of disks must vary quite strongly with luminosity in the b - band ( being smaller for the fainter , bluer galaxies ) but could plausibly be constant at i. for this reason the i - band tf relation does indeed appear to be the most appropriate one to compare with our models . in the models
the slight deviation of the slope from @xmath77 is caused by the fact that massive haloes are less concentrated than low - mass haloes . to see this ,
we recall nfw s discovery that the concentration factor of a halo is determined by its formation time @xmath190 ; massive haloes form later and so are less concentrated ( see the appendix ) .
since the value of @xmath191 in equation ( [ erfc ] ) is small , we have @xmath192 . for galactic haloes , @xmath193 in all the cosmogonies considered here , and so @xmath194 . writing the _ rms _ mass fluctuation defined in equation ( [ sigma ] ) as @xmath195
, we have @xmath196 . on galactic scales , the effective power index @xmath197 for our cdm models , so that more massive haloes have lower values of @xmath190 , and so also of @xmath116 .
the value of @xmath0 is slightly larger for scdm , because @xmath198 is slightly larger in this case . as shown by navarro
et al ( 1997 ) , if @xmath199 differs substantially from @xmath200 , then @xmath0 can be very different from 3 .
the effect of changing the _ amplitude _ of the fluctuation spectrum is negligible , and the slope is also insensitive to changes in @xmath43 , @xmath46 , and the formation redshift of disks , provided none of these parameters varies with @xmath35 .
the observed slope of the tf relation is thus a generic prediction of hierarchical models with cdm - like fluctuation spectra .
the observed scatter in the tf relation is quite small ; giovanelli et al ( 1997 ) quote 0.35 magnitudes as an overall measure of scatter , while willick et al ( 1995,1996 ) find @xmath201 magnitudes to be typical .
it is clear from figure 6 that we do indeed predict a small scatter , but it is important to understand why . in the model of section 2.2 the gravitational effect of disks
is neglected , all haloes of the same mass have identical singular isothermal density profiles , and all disks have the same @xmath61 and @xmath79 . with these assumptions
the tf relation has no scatter even though the broad @xmath48-distribution leads to a wide range of disk sizes and surface brightnesses . in the more realistic model of section 2.3 ,
scatter in the tf relation can arise from two effects . for given disk and halo masses , the maximum rotation is larger for more compact disks , i.e. for those residing in more slowly spinning haloes . as a result
, the scatter in the @xmath48-distribution of haloes translates into a scatter about the predicted tf relation .
this is the source of scatter in figure 6 ; this scatter remains small because the stability requirement eliminates the most compact systems .
in addition , even for given @xmath202 , @xmath58 and @xmath48 , the disk rotation velocity will be larger in more concentrated haloes .
thus any scatter in @xmath116 will produce scatter in the tf relation .
we now examine the consequences of these two effects in detail .
the solid curves in figure 7 show the scatter about the mean model tf relations as a function of @xmath203 , an assumed lower limit on @xmath48 for the haloes allowed to contribute observable disks . for given @xmath61 ,
this function is sensitive neither to cosmological parameters nor to the redshift of disk formation .
the scatter increases as haloes with lower values of @xmath48 are included , because disk self - gravity becomes important for such systems .
it increases with @xmath61 , because the contribution of a disk to the rotation velocity is then larger at fixed @xmath48 .
it is clear that the predicted scatter can be made sufficiently small , provided disks with large @xmath61 and small @xmath48 are excluded .
the arrows next to each curve in figure 7 show the smallest value of @xmath48 for which disks with the given value of @xmath61 are stable ( here we conservatively require @xmath204 ) .
it is clear that the predicted scatter is smaller than observed if @xmath205 . for even smaller @xmath43 , a larger range of @xmath48 is allowed for stable disks , leading to a larger scatter .
this is a consequence of the deviation between the nfw profile and an isothermal sphere .
since only about 10 percent of haloes have @xmath206 , the increased scatter in this case is due to a relatively small number of compact outliers .
the distribution of the halo concentration factor at fixed mass is poorly known . in the simulations of nfw ,
the _ rms _ scatter in @xmath116 is @xmath207 . to demonstrate the effect of such scatter on the predicted tf relation , the dashed curves in figure 7 show the predicted scatter
if we assume a 25% _ rms _ uncertainty in the concentration parameter of haloes of a given mass .
( note : nfw show that the scatter in @xmath116 is uncorrelated with @xmath48 . )
if this variability in halo profiles is indeed realistic , it is clear from figure 7 that it adds relatively little scatter to the tf relation . as nfw point out ,
this is in conflict with the conclusions of eisenstein & loeb ( 1996 ) from an analytic assessment of the expected variation in halo profiles . in summary
, the analysis of this section suggests that it may not be difficult for hierarchical models to produce a disk population with tully - fisher scatter within the observed limits .
we note , however , that we have not included the bulge contribution to the luminosity of the galaxies or possible variations in @xmath208 .
both presumably act to increase the tf scatter .
the zero - point of the predicted tf relation depends weakly on the distribution of @xmath48 and on our stability requirement . on the other hand
, it depends strongly on @xmath43 , the fraction of the halo mass in disks , on @xmath79 , the stellar mass - to - light ratio of the disks , and on @xmath39 , the redshift at which disk material is assembled within a single dark halo .
figure 8 shows the values of @xmath209 required to reproduce the observed zero - point of the @xmath210-band tf relation ( equation [ zero - point ] ) as a function of @xmath43 .
zero - points are calculated for stable disks with @xmath211 ( the results are very similar for @xmath204 or @xmath212 ) . in all cases
the required mass - to - light ratio is approximately proportional to @xmath43 .
the three arrows in each panel show the values of @xmath43 at which 10% , 50% and 90% of the haloes can host stable disks .
stable disk galaxies are common and there are no obvious observational counterparts for a large number of `` failed '' unstable disks .
the @xmath12 lines in this figure thus suggest @xmath213 in the @xmath210-band for the ccdm model , whereas for scdm , @xmath28 and @xmath26 the upper limit gets progressively weaker .
this trend is a consequence of the steadily decreasing concentration of haloes of given mass along this sequence , which results in a corresponding decrease in the maximum rotation velocity .
the same trend can be seen quite clearly in figure 6 .
the maximum allowed mass - to - light ratio is even smaller for higher assembly redshifts , because disks are then less massive for a given @xmath35 .
the horizontal dotted lines in figure 8 bracket the @xmath214 range derived by bottema ( 1997 ) from his analysis of the dynamics of observed disks .
this range takes account of the fact that some ( dim ) disks may have as much as half of their mass in gas .
as one can see , in order to reproduce the tully - fisher zero - point with @xmath215 in the allowed range , normal disks could form at high @xmath39 only if @xmath43 is large , while for low formation redshifts @xmath216 is possible . for @xmath217 most haloes _ can not _ host stable disks , so we infer that normal spiral disks must form late with an effective assembly redshift close to zero .
a reasonable zero - point then requires @xmath218 , in good agreement with the values obtained above from the observed sizes and surface brightnesses of disks .
finally we note that because disk self - gravity is less important for larger values of @xmath48 , somewhat higher mass - to - light ratios are needed if lsb galaxies are to fit on the same tully - fisher relation as normal spirals .
damped ly@xmath0 absorption systems ( dls ) , corresponding to absorbing clouds with a neutral column density , @xmath219 , are seen in the spectra of many high redshift qsos .
their abundance varies very roughly as @xmath220 over the redshift range @xmath221 , and their total hi content is a few tenths of a percent of the critical density , about the same the total mass in stars in the local universe ( see e.g. storrie - lombardi et al 1996 ) . over the last decade wolfe and
his collaborators have used a wide range of observational indicators , most recently the velocity structure in the associated metal lines , to argue that these systems correspond to large equilibrium disks with @xmath222 , the extended and gas - rich progenitors of present - day spirals ( wolfe et al 1986 ; lanzetta , wolfe & turnshek 1995 ; wolfe 1995 ; prochaska & wolfe 1997 , but see jedamzik & prochaska ) .
early theoretical studies showed that the total hi content of these systems can be reproduced in some but not all hierarchical cosmologies , provided systems with circular velocities well below 200 km / s are allowed to contribute ( mo & miralda - escud 1994 ; kauffmann & charlot 1994 ; ma & bertschinger 1994 ; klypin et al 1995 ) . a detailed study of the standard cdm cosmology by kauffmann ( 1996b ) concluded that disks in all haloes down to @xmath223 km / s must contribute , once a realistic model for star formation is included . numerical simulations by gardner
et al ( 1997 ) led these authors to a similar conclusion , while higher resolution simulations by haehnelt , steinmetz & rauch ( 1997 ) showed that the kinematic data of prochaska & wolfe ( 1997 ) can be reproduced by nonequilibrium disks in low mass haloes and so do not _ require _ large @xmath35 .
these simulations did not include star formation .
the cross - sections they imply should be viewed with caution since their resolution is quite poor and their underlying physical assumptions lead to a substantial underprediction of the sizes of present - day disks .
our models allow a more detailed study of the sizes and cross - sections of high redshift disks than does that of kauffmann ( 1996b ) .
figure 9 shows how the predicted abundance of absorbers at @xmath224 rises as disks within lower and lower mass haloes are allowed to contribute .
we parametrize the mass of the smallest contributing halo by @xmath225 , the rotation velocity of its central disk measured at @xmath123 and averaged over the @xmath48-distribution of stable disks .
this abundance refers to absorbers with @xmath226 and is calculated from @xmath227 where @xmath228 is the proper distance at redshift @xmath39 , @xmath229 is the comoving number density of galactic haloes given by equation ( [ ps_mf ] ) , @xmath230 is the central column density of the disk , @xmath231 is the minimum hi column density for a dls , and @xmath232 $ ] .
the step function , @xmath233 for @xmath234 and @xmath235 for @xmath236 , ensures that only stable disks are used . for this plot
we assume @xmath154 and @xmath142 .
we also assume that all baryons at the outer edge of the absorbing region are in the form of hi gas , and that the column density never drops below @xmath52 at smaller radii .
this may be unrealistic at low redshift because star formation may substantially reduce the mass of hi gas .
the horizontal dotted lines show a @xmath237 range for the observed dls abundance at @xmath224 , taken from storrie - lombardi et al ( 1996 )
. the lower limit on @xmath225 needed to reproduce the observations can be read off immediately .
this limit varies quite strongly between cosmogonies . for @xmath26 , which has the smallest power on galactic scales
, one has to include disks with rotation velocities smaller than @xmath238 .
this confirms the result found earlier for the mdm model which has a similar power spectrum ( mo & miralda - escud 1994 ; kauffmann & charlot 1994 ; ma & bertschinger 1994 ; klypin et al 1995 ) . for the other three models , about one third of the observed systems can arise in disks with @xmath239 , and two thirds or more in disks with @xmath240 .
it may not be difficult to accommodate disks with @xmath35 as large as suggested by wolfe and collaborators , although these will be predicted to be much smaller at @xmath224 than nearby disks with the same circular velocity .
it is interesting that the total cross - section of rotationally supported _ stable _ disks with @xmath240 can already explain the observations .
unless the formation of such disks is prevented , they should give rise to a large fraction of the observed dls .
some absorbers may correspond to _ unstable _ disks , but the abundance of this population should be relatively small ; such disks are compact and so have small cross - sections .
figure 10 shows the cross - section weighted distribution of the spin parameter for disks at @xmath224 . comparing this with the unweighted @xmath48-distribution ( shown as the dot - dashed curve )
, we see that absorption comes primarily from disks with large @xmath48 . for @xmath241 , @xmath242 , and less than 20% of dls
are produced by unstable systems ( assuming , of course , that instability does not change the cross - section ) .
the value of @xmath243 for stable disks is maximised for @xmath244 , but varies little for @xmath43 ( @xmath245 ) in the range from 0.025 to 0.07 .
the abundance is smaller both for larger and for smaller @xmath43 , in the first case because the number of stable disks is small , and in the second because the amount of gas is small .
this range is similar to the one we inferred in earlier sections from the properties of local disk galaxies .
as shown by haehnelt et al ( 1997 ) many dls may be produced by gas which has not fully settled into centrifugal equilibrium .
we are implicitly assuming that the cross - section for damped absorption does not vary strongly , at least on average , over this settling period .
it is also instructive to look at the predicted distribution of the impact parameter @xmath246 , defined as the distance between the line - of - sight and the centre of a disk which produces a dls .
this distribution can be written as [ pofb ] p(b ) dbdbdp()^_m_l dn_h(,z ) f[n_0(,,z)/n_l , b,(,,z ) ] ( _ m-1 ) , where @xmath247 is the lower limit on halo mass , and @xmath248 is the distribution of @xmath246 for a disk with scalelength @xmath44 and central surface density @xmath230 observed at a random inclination angle .
the function @xmath248 can be obtained from simple monte - carlo simulations .
the solid curves in figure 11 show @xmath249 at @xmath224 , with @xmath247 chosen to reproduce the observed abundance of dls . for @xmath26 ,
this distribution is narrow and peaks at quite small values of @xmath246 ; dls at large impact parameters are rare in this model . for the other three models , however , the upper quartile of the distribution is at @xmath250 , so systems with relatively large impact parameter do occur .
not surprisingly , they tend to be associated with high-@xmath35 disks , as can be seen from the dashed curves . at present
, there are very few measurements of the impact parameter for a high redshift dls .
djorgovski et al ( 1996 ) have identified a galaxy responsible for a dls at @xmath251 , and for this case the impact parameter is about @xmath252 ( for @xmath253 ) .
an observation by lu et al ( 1997 ) suggests that this is indeed a rotating disk , with @xmath85 . as one
can see from figure 11 , for this @xmath35 , the observed impact parameter is not exceptionally large .
one can not draw any statistical conclusions from a single event , but if systems like this turn out to be common the @xmath33cdm model would be disfavored .
it is worth noting that this consistency with the observational results requires that @xmath43 be in the range 0.025 to 0.07 , and that little angular momentum be transferred to the dark halo during disk formation .
significant angular momentum loss would produce smaller disks .
haloes with smaller @xmath35 would then have to be included in order to match the observed abundance . in this case ,
typical dls would be predicted to have smaller @xmath35 and smaller @xmath246 , but to have larger column densities .
substantial star formation would be required to reduce the column densities to the observed values , contradicting , perhaps , the low metal abundances measured in high redshift dls .
the galaxies responsible for dls are likely to be easier to identify at lower redshift .
it is therefore interesting to give predictions for such systems . as an example
, we show the impact parameter distribution for dls at @xmath160 in figure 12 .
results are shown only for scdm and @xmath28 ; those for ccdm and @xmath26 are similar .
the redshift is chosen to be typical for the low - redshift dls observed by , for example , steidel et al ( see steidel 1995 ) . in computing this distribution
we have used only disks with @xmath254 .
this choice is based on the fact that disks with smaller @xmath35 are difficult to identify , while those with larger @xmath35 are rare .
an upper limit on @xmath35 has to be imposed in our calculation , because massive haloes are likely to host elliptical galaxies or galaxy clusters which contain only a small amount of hi gas .
the median and the upper quartile of the distribution are at about 7 and @xmath88 , respectively ( see the solid curves ) .
larger impact parameters are expected for samples biased towards galaxies with larger @xmath35 ( see the dashed curves ) .
as before , of course , these predictions assume that star formation has not substantially reduced the gas column near the damped absorption boundary . some decrease in impact parameter may therefore be expected by @xmath160 .
however , the absorption cross - section of an exponential disk depends only logarithmically on @xmath43 , and so figure 12 will not change much provided star formation is only moderately efficient in the outer galaxy . on the other hand ,
the hi mass in a dls decreases in proportion to @xmath43 , so that systems observed at lower redshift may have systematically lower column densities and so give rise to a smaller cosmic density parameter in hi .
these trends are indeed observed ( wolfe 1995 ) , and arise naturally in models which take star formation into account ( kauffmann & charlot 1994 ; kauffmann 1996b ) . to see
what kind of disks produce dls at low redshifts , we can examine the cross - section weighted distribution of the spin parameter .
this distribution depends only weakly on @xmath39 , so that figure 10 is still relevant .
notice again the strong bias towards the large @xmath48 tail .
as discussed in 3 , the disks which form in haloes with large @xmath48 have large size and low surface densities .
we thus expect that galaxies selected as dls will be biased towards low surface densities .
as noted by mcgaugh & de blok ( 1996 ) and others , the star formation rates in such galaxies are low , giving rise to low surface brightnesses , and producing a relatively small reduction in their cross - section for damped absorption . the impact parameter distribution of figure 12
may then apply .
there are now some observations of galaxies responsible for low redshift dls .
steidel et al ( see steidel 1995 ) find the typical impact parameter for such systems to be 5-@xmath255 .
there are also indications that these galaxies tend to be blue , with low surface brightness ( steidel 1995 ) and low chemical abundance ( pettini et al 1995 ) .
all these points seem to agree well with our expectations .
obviously , more such data will provide stringent constraints on disk formation models of the kind we propose here .
so far we have considered galaxies to contain only disk and halo components . in reality
, most spirals also contain a bulge . for a galaxy of milky way type or
later , the bulge mass is less than 20 percent that of the disk , so its dynamical effects are small . for earlier - type spirals
, however , the bulge makes up a larger fraction of the stellar mass , and its effects may be significant . in this section
we use a simple model to assess how our results are affected by the presence of a central bulge .
we assume bulges to be point - like , to have a mass which is @xmath256 times that of the halo , and to have negligible angular momentum .
further , we assume either that the specific angular momentum of the disk is the same as that of its dark halo ( so that @xmath67 as usually assumed above ) , or that the total specific angular momentum of the stellar components is equal to that of the halo ( so that @xmath257 )
. the second case may be appropriate if angular momentum transfer to the halo is negligible , if low angular momentum gas forms the bulge , and if high angular momentum gas settles into the disk . under these assumptions it is straightforward to extend the model of section 2.2 to include the bulge .
equation ( [ adiabatic ] ) becomes _
f(r ) = ( r ) + ( r_i ) ( 1 - -)+ , and @xmath258 in equations ( [ vc_nfw ] ) and ( [ vc_bdm ] ) is replaced by @xmath259 , where @xmath260 is the contribution to the circular velocity from the bulge . with these changes ,
the procedure of section 2.2 can be applied as before to obtain the disk scalelength , @xmath44 , and the rotation curve , @xmath107 . in figure 13
we show the predicted rotation curves when disk and bulge masses are equal .
results are shown for the two cases mentioned above , namely @xmath67 and @xmath261 .
the same total halo mass and concentration are adopted as in the top left panel of figure 2 .
the rotation curves now diverge at small radius because of our unrealistic assumption of a point - like bulge . in the case
where @xmath262 the disk is substantially more extended , and the value of @xmath35 at given radius is slightly lower , than in the case where @xmath67 . to see the effect of the central bulge more clearly , we compare values of @xmath44 and @xmath186 for the two cases , @xmath263 and @xmath264 . figure 14 shows how the relative values vary with @xmath265 .
this figure assumes @xmath266 and @xmath267 , but in fact the ratios are quite insensitive to @xmath48 and @xmath116 .
as one can see , for @xmath67 the values of @xmath44 and @xmath35 do not change much even when half of the accreted gas is put into the central bulge . on the other hand , for @xmath262 , the disk scalelength increases by a factor of two , and the rotation velocity drops significantly below that for @xmath264 .
this case produces similar results to those for @xmath155 and @xmath264 .
the results in figure 14 have some interesting implications .
if disks have the same specific angular momentum as their dark haloes , and if the luminosity of a galaxy is proportional to its total stellar mass , then the zero - point of the tf relation is almost independent of the mass of the bulge . even in the extreme case where @xmath268 and disks have twice the specific angular momentum of their haloes , the disk rotation velocity for given @xmath269 is only reduced by about 20 percent
provided @xmath270 .
if disk and bulge had the same mass - to - light ratio , this would change the zero - point of the tf relation by about @xmath271 , but since the stellar mass - to - light ratio of bulges is undoubtedly higher than that of disks , the actual change would be smaller . in practice ,
most of the galaxies used in applications of the tf relation have bulge - to - disk ratios much smaller than one , so we expect the effects of the bulges to produce rather little scatter in the tf relation .
in this paper , we have formulated a simple model for the formation of disk galaxies .
although many of the observed properties of spirals and damped ly@xmath0 absorbers seem relatively easy to explain , it is important to keep in mind our underlying assumptions .
these leave open a number of difficult physical questions .
we have assumed that all galaxy haloes have the universal density profile proposed by nfw .
this appears relatively safe since the original nfw claims have been confirmed by a number of independent n - body simulations ( e.g. tormen et al 1996 ; cole & lacey 1996 ; huss et al 1997 ) .
more significantly we assume that all disks have masses and angular momenta which are fixed fractions of those of their haloes .
successful models require @xmath272 and @xmath273 .
these relations are not produced in a natural way by existing simulations of hierarchical galaxy formation .
while it is common to find all the gas associated with galactic haloes within condensed central objects ( so that @xmath274 ) , the gas usually loses most of its angular momentum to the dark matter during galaxy assembly .
this results in @xmath275 and produces disks which are too small ( navarro & benz 1991 ; navarro & white 1994 ; navarro & steinmetz 1997 ) .
the resolution of this problem may be related to the well - known need to introduce strong feedback in order to get a viable hierarchical model for galaxy formation ( e.g. white & rees 1978 ; white & frenk 1991 ; kauffmann et al 1993 ) .
such feedback is not included in the simulations , but is required in hierarchical models to suppress early star formation in small objects , thus ensuring that gas remains at late times to form large galaxies and the observed intergalactic medium . disks may then form from gas which remains diffuse and so loses little angular momentum during protogalactic collapse . unfortunately , while such effects may solve the angular momentum problem , they call into question the proportionality between disk mass and halo mass .
the disk / halo mass ratios required by our models are , in any case , much less than the observed baryon mass fractions in rich galaxy clusters , so substantial inefficiency in assembling the galaxies is necessary in any consistent model .
there are several interesting theoretical issues that are not addressed in this paper .
as is well known , disk galaxies , especially lsb galaxies , have lower correlation amplitudes than ellipticals ( see e.g. davis & geller 1976 ; jing , mo & brner 1993 ; mo , mcgaugh & bothun 1994 ) .
since hierarchical clustering predicts little correlation between the spin parameter of a halo and its large - scale environment , this observed clustering difference can not be considered a consequence of disk galaxies being associated with haloes with large @xmath48 .
however , as discussed above , disk galaxies , especially lsbs , form late in our model . on the other hand , elliptical galaxies , and in particular cluster ellipticals , form at relatively high redshifts in hierarchical models ( kauffmann 1996a ) .
since for given mass , haloes at low redshift are less correlated than those at high redshift ( mo & white 1996 ) , the morphology dependence of clustering can arise as a consequence of the morphology dependence of formation time .
this effect is seen quite clearly in the detailed semianalytic galaxy formation models of kauffmann et al ( 1993 ) and kauffmann , nusser & steinmetz ( 1997 ) .
local environmental effects may also lead to morphological segregation ; for example , extended disks may be destroyed in a high density environment by tidal effects and gas stripping ( moore et al 1996 ) .
detailed modelling is clearly needed for any quantitative comparison with observation .
in this paper we have considered the formation of disk galaxies only .
it is interesting to examine how other populations might fit into such a scenario .
earlier semianalytic modelling has shown that many of the global properties of giant ellipticals can be understood if they formed by mergers between disk systems which themselves formed at early times in overdense `` protocluster '' regions .
as we have seen , high redshift disks should be substantially denser and more compact than present - day disks of similar mass .
this may go a long way towards explaining the high densities of observed ellipticals .
some bulges and ellipticals may also form in dark haloes with low spin where any disk would be too dense to be stable .
such `` spheroidal '' systems may have systematically different properties from those which formed by mergers .
we intend to return to a number of these issues in future papers .
we are grateful to s. courteau for providing the data plotted in figure 4 .
we thank luiz da costa for helpful discussions .
this project is partly supported by the `` sonderforschungsbereich 375 - 95 fr astro - teilchenphysik '' der deutschen forschungsgemeinschaft .
bardeen j. , bond j.r .
, kaiser n. , szalay a.s . , 1986 , apj , 304 , 15 barnes j. , white s.d.m .
1984 , mnras , 211 , 753 baugh c.m . , cole s. , frenk c.s . , lacey c.g .
1997 , preprint ( astro - ph/9703111 ) bertschinger , e. 1985 , apjs , 58 , 39 bessel m.s . , 1979 , pasp , 91 , 589 binney j. , tremaine s. , 1987 , galactic dynamics .
princeton univ . press ,
princeton , nj bothun g. , impey c. , mcgaugh s. , 1997 , preprint ( astro - ph/9702156 ) bottema , r. 1997 , preprint ( astro - ph/9706230 ) broeils a.h . , courteau s. , 1996 , preprint ( astro - ph/9610264 ) carroll s.m . , press w.h .
, turner e.l .
, 1992 , araa , 30 , 499 casertano s. van gorkom j. , 1991 , aj , 101 , 1231 christodoulou d.m . , shlosman i. , tohline j.e . , 1995 , apj , 443 , 551 cole s. , aragon - salamanca a. , frenk c.s . ,
navarro j.f . , zepf s. e. , 1994 , mnras , 271 , 781 cole s. , lacey c. , 1996 , preprint ( astro - ph/9510147 v3 ) courteau s. , 1996 , apjs , 103 , 363 courteau s. , 1997 , apjs , 1 in press dalcanton j.j . ,
spergel d.n . , summers , f.j .
, 1997 , apj , 482 , 659 ( dss ) davis m. , geller m. , 1976 , apj , 208 , 13 djorgovski s.g .
, pahre m.a . , bechtold j. , elston r. , 1996 , nature , 382 , 234 efstathiou g. , bond j.r .
, white s.d.m . , 1992 , mnras , 258 , 1p efstathiou g. , lake , g. , negroponte , j. , 1982 , mnras , 199 , 1069 eisenstein , d.j . ,
loeb , a. , 1996 , apj , 459 , 432 fall s.m . , efstathiou g. , 1980 , mnras , 193 , 189 gardner j.p . , katz n. , weinberg d.h . , hernquist l. , 1997 , preprint ( astro - ph/9705118 ) giovanelli r. , haynes m.p . , da costa l.n .
, freudling w. , salzer j.j . , wegner g. , 1997 , preprint ( astro - ph/9612072 ) gunn j. , gott r. , 1972 , apj , 176 , 1 haehnelt m. g. , steinmetz m. , rauch , m. , 1997 , preprint ( astro - ph/9706201 ) huss a. , jain b. , steinmetz m. , 1997 , preprint ( astro - ph/9703014 ) jedamzik , k. , prochaska , j.x .
1997 , preprint ( astro - ph/9706290 ) jing y.p . , mo h.j . & brner g. , 1991 , a&a , 252 , 449 .
katz n. , gunn j. , 1991 , apj , 377 , 365 kauffmann g. , charlot s. , 1994 , apj , 430 , l97 kauffmann g. , white s.d.m . , guiderdoni b. , 1994 , mnras , 267 , 981 kauffmann g. 1996a , mnras , 281 , 487 kauffmann g. 1996b , mnras , 281 , 475 kauffmann g. , nusser a. , steinmetz m. , 1997 , mnras , 286 , 795 klypin a. , borgani s. , holtzman j. , primack . j. , 1995 , apj , 444 , 1 lacey c. , cole s. , 1993 , mnras , 262 , 627 lacey c. , cole s. , 1995 , preprint ( astro - ph/9510147 ) lanzetta k.m . , wolfe a. m. , turnshek d. a. , 1995 , apj , 440 , 435l lu l. , sargent w.l.w . , barlow t.a . , 1997 ,
preprint ( astro - ph/9701116 ) ma c.p . ,
bertschinger e. , 1994 , apj , 434 , l5 mcgaugh s.s .
, 1994 , nature , 367 , 538 mcgaugh s.s .
, de blok w.j.g . , 1996 ,
preprint ( astro - ph/9612070 ) mo h.j . , jing y.p . , white s.d.m . , 1996 , mnras , 282 , 1096 mo h.j . , mcgaugh s.s
, bothun g.d . , 1994 , mnras , 267 , 129 mo h.j . , miralda - escud j. , 1994 , apj , 430 , l25 mo h.j . , white s.d.m . , 1996 , mnras , 282 , 347 moore b. , katz n. , lake g. , dressler a. , oemler , a. jr . , 1996 , nature , 379 , 613 navarro j.f . , benz w. 1991 , apj , 380 , 320 navarro j.f . ,
white s.d.m . , 1994 , mnras , 267 , 401 navarro j.f . , frenk c. s. , white s.d.m . , 1995 , mnras , 275 , 720 navarro j.f . , frenk , c. s. , white , s.d.m . , 1996 , apj , 462 , 563 navarro j.f . , frenk , c. s. , white , s.d.m . , 1997 , in press ( preprint , astro - ph/9611107 ) ( nfw ) navarro j.f . , steinmetz m. , 1997 , 478 , 13 pettini m. , king d.l . , smith l.j . ,
hunstead r.w . , 1995 in qso absorption lines , ed .
g. meylan , springer : berlin , p.71 press w.h . , schechter p. , 1974 ,
apj , 187 , 425 ( ps ) prochaska j.x .
, wolfe a.m. , 1997 , apj , submitted percic , m. , salucci , p. , 1991 ,
apj , 285 , 307 shanks t. , 1997 , preprint ( astro - ph/9702148 ) sprayberry d. , bernstein g.m . , impey c.d . ,
bothun g.d .
, 1995 , apj , 438 , 72 steidel c.c . , 1995 , in qso absorption lines , ed
. g. meylan , springer : berlin , p.137 steinmetz m. , bartelmann m. , 1995 , mnras , 272 , 570 steinmetz m. , mller e. , 1995 , mnras , 276 , 549 storrie l.j . , mcmahon r.g .
, irwin m.j . , 1996 , mnras , 283 , l79 strauss , m. , willick , j.a . , 1995 , phys
. reports , 261 , 271 toomre a. , 1964 , apj , 139 , 1217 tormen g. , bouchet f.r . ,
white s.d.m . , 1997 , mnras , 286 , 865 turner , m.s . 1996 , preprint ( astro - ph/9610158 ) walker t. p. , steigman , g. , schramm , d. n. , olive , k. a. , kang , h .- s .
1991 , apj 376 , 51 warren m.s .
, quinn p.j . , salmon j.k . , zurek w.h . , 1992 ,
apj , 399 , 405 white m. , scott d. , 1996 , comments on astrophysics , 8 , no . 5 white s.d.m . , 1995 , les houches white s.d.m . , efstathiou g. , frenk c. , 1993 , mnras , 262 , 1023 white s.d.m .
, frenk c.s . , 1991 , apj , 379 , 52 white s.d.m .
, rees m.j . , 1978 , mnras , 183 , 341 williams r.e . , et al , 1996 , aj , 112 , 1335 willick j.a .
, courteau s. , faber s.m .
, burstein d. , dekel a. , kolatt t. , 1996 , apj , 457 , 460 willick j.a . , courteau s. , faber s.m . , burstein d. , dekel a. , 1995 , apj , 446 , 12 wolfe a. , turnshek , d.a . ,
smith , h.e . , cohen , r. d. , 1986 , apjs , 61 , 249 wolfe a. , 1995 , in qso absorption lines , ed .
g. meylan , springer : berlin , p.13 zwaan m.a . , van der hulst j.m .
, de blok w.j.g . ,
mcgaugh s.s . , 1995 , mnras , 273 , l35
we assume that the universe is dominated by cold dark matter ( cdm ) .
the initial power spectrum is p(k ) , with t(k)=ln(1 + 2.34q)2.34 q 1 + 3.89 q+(16.1q)^2+(5.46q)^3+(6.71q)^4 ^-1/4 and q ( see bardeen et al 1986 ) .
following efstathiou et al ( 1992 ) , we have introduced a shape parameter , @xmath276 , for the power spectrum . the rms mass fluctuation in top - hat windows with radius @xmath277 , @xmath278 ,
is defined by [ sigma ] ^2 ( r_0)=12 ^ 2_0^ k^3p(k)w^2(kr_0 ) , where @xmath279 is the fourier transform of the top - hat window function .
the power spectrum , @xmath10 , is normalized by specifying @xmath280 .
we use the press - schechter formalism ( press & schechter 1974 ) to calculate the mass function of dark haloes . in this formalism , the comoving number density of dark haloes , with mass in the range @xmath281 , at redshift @xmath39 is [ ps_mf ] n_h(,z)d= -(2)^1/2 _ 0 _ z ( r_0 ) d(r_0 ) d exp -_z^2 2 ^2(r_0 ) , where @xmath282 is the mean matter density of the universe at present time ; @xmath58 is the mass of the halo , related to the initial comoving radius @xmath277 of the region from which the halo formed ( measured in current units ) by @xmath283 , @xmath284 is the threshold linear overdensity for collapse at redshift @xmath39 . according to the spherical collapse model , [ deltaz ] _
z=(1+z ) , where @xmath285 ( see nfw ) ; @xmath286 is the linear growth factor at the expansion factor @xmath287 corresponding to the redshift @xmath39 ( @xmath288 , the present value of @xmath287 , is taken to be 1 ) : g(a)=52 ^ 4/7 -_+(1+/2)(1+_/70)^-1 ( see carroll , press & turner 1992 ) , with ( a)=a+(1-a)+(a^3-a ) , _ _ ( a ) = a^3a+(1-a)+(a^3-a ) . to put the nfw density profile in a cosmological context , we need to calculate the concentration factor @xmath116 defined in equation ( [ concentration ] ) . for this density profile
, @xmath116 is related to the characteristic overdensity , @xmath93 , by equation ( [ delta_null ] ) .
the appropriate value of @xmath116 depends on halo formation history and on cosmology .
nfw proposed a simple model for @xmath116 based on halo formation time .
the formation redshift @xmath190 of a halo identified at @xmath12 with mass @xmath58 is defined as the redshift by which half of its mass is in progenitors with mass exceeding @xmath289 , where @xmath290 is a constant . according to the formula given by lacey & cole ( 1993 ) based on the press - schechter formalism , the halo formation time
is then defined implicitly by [ erfc ] erfc=12 .
nfw found that the characteristic overdensity of a halo at @xmath12 is related to its formation redshift @xmath190 by [ delta_nullzf ] ( , f)=c(f)_0 1+z_f(,f)^3 , where the normalization @xmath291 depends on @xmath191 .
we will take @xmath292 as suggested by the n - body results of nfw . in this case @xmath293 .
thus , for a halo of given mass at @xmath12 , one can obtain the concentration factor @xmath116 from equations ( [ deltaz])-([delta_nullzf ] ) . in practice
, we first solve @xmath190 from equation ( [ erfc ] ) and insert the value of @xmath190 into equation ( [ delta_nullzf ] ) to get @xmath93 , we then use this value of @xmath93 in equation ( [ delta_null ] ) to solve for @xmath116 .
the distribution of @xmath116 is not well known . in the simulations of nfw ,
the scatter in @xmath116 is @xmath294 .
equations ( [ erfc])-([delta_nullzf ] ) can be easily extended to calculate concentrations for haloes identified at @xmath295 . in this case
, we just replace @xmath296 in equation ( [ erfc ] ) by @xmath284 , and @xmath7 in equation ( [ delta_nullzf ] ) by @xmath297 , where @xmath287 is the scale factor at @xmath39 . | we study the population of galactic disks expected in current hierarchical clustering models for structure formation . a rotationally supported disk with exponential surface density profile
is assumed to form with a mass and angular momentum which are fixed fractions of those of its surrounding dark halo .
we assume that haloes respond adiabatically to disk formation , and that only stable disks can correspond to real systems . with these assumptions
the predicted population can match both present - day disks and the damped ly@xmath0 absorbers in qso spectra .
good agreement is found provided : ( i ) the masses of disks are a few percent of those of their haloes ; ( ii ) the specific angular momenta of disks are similar to those of their haloes ; ( iii ) present - day disks were assembled recently ( at @xmath1 ) . in particular , the observed scatter in the size - rotation velocity plane is reproduced , as is the slope and scatter of the tully - fisher relation .
the zero - point of the tf relation is matched for a stellar mass - to - light ratio of 1 to 2 @xmath2 in the i - band , consistent with observational values derived from disk dynamics .
high redshift disks are predicted to be small and dense , and could plausibly merge together to form the observed population of elliptical galaxies . in many ( but not all ) currently popular cosmogonies , disks with rotation velocities exceeding 200 km
/ s can account for a third or more of the observed damped ly@xmath0 systems at @xmath3 .
half of the lines - of - sight to such systems are predicted to intersect the absorber at @xmath4 and about 10% at @xmath5 .
the cross - section for absorption is strongly weighted towards disks with large angular momentum and so large size for their mass .
the galaxy population associated with damped absorbers should thus be biased towards low surface brightness systems . |
striped patterns of spin and charge order have been observed in a wide range of antiferromagnetic oxides since the initial discovery in la@xmath7nd@xmath8sr@xmath9cuo@xmath10.@xcite interest has been sustained by continuing uncertainty about the true importance of stripe correlations for the mechanism of superconductivity in the layered cuprates . in some models ,
for example , stripe correlations assist in the formation of pairing instabilities that can lead to superconductivity,@xcite whereas according to experiment static charge stripes suppress superconductivity.@xcite .
investigations into the fundamental properties of stripes are therefore potentially important for an understanding of cuprate superconductivity , as well as providing insight into an interesting emergent phase of electronic matter .
experimental investigations on stripe phases have been made on various materials , but the layered nickelates la@xmath0sr@xmath11nio@xmath12 ( lsno ) have been a particularly informative model system to study.@xcite lsno , which is isostructural with the 214 " high temperature superconductor la@xmath0sr@xmath11cuo@xmath12 , exhibits spin and charge stripe order for @xmath13.@xcite the stripes form on the square nio@xmath14 layers and consist of diagonal bands of antiferromagnetically ( afm ) ordered ni@xmath15 spins separated by charged domain walls that act as antiphase boundaries to the magnetic order . at one - third doping ( e.g. @xmath16 )
the stripe order is particularly stable owing to a combination of two factors , firstly a commensurability effect that pins the charge stripes to the lattice , and second a stripe periodicity which is the same for the magnetic and a charge order.@xcite at this doping level lsno displays long range ( @xmath17 ) charge order,@xcite making it an ideal material in which to probe the charge ordered state . the magnetic excitation spectrum in the ordered stripe phase of lsno with @xmath18 has been investigated in some detail by neutron inelastic scattering and found to contain two distinct components , ( i ) quasi - two - dimensional spin - wave excitations of the afm - ordered regions , extending to @xmath1980 mev in energy,@xcite and ( ii ) a low energy ( @xmath20mev ) quasi - one - dimensional ( q-1d ) magnetic fluctuation consistent with short - range afm correlations along the charge stripes.@xcite a two - component spectrum with qualitatively similar characteristics has also been observed in la@xmath21sr@xmath22nio@xmath2 .
( ref . )
the observation of two types of magnetic dynamics is consistent with the existence two magnetic sub - systems associated with nominally ni@xmath15 and ni@xmath23 ions , the latter of which form the charge stripes . at this time
, however , there is no microscopic model of the magnetic interactions in lsno that provides a unified description of the complete magnetic spectrum .
the aim of the present study was to work towards a better understanding of the low - energy q-1d magnetic correlations by comparing neutron scattering data from two doping levels , @xmath24 and @xmath25 ( both with @xmath26 ) , and by examining in more detail the energy lineshape of the magnetic scattering .
we have found that the q-1d scattering is present at both doping levels , and our analysis of the energy lineshape reveals that the signal is gapped at the minimum of the q-1d dispersion .
single crystals of la@xmath0sr@xmath1nio@xmath10 were grown by the floating - zone method.@xcite the crystals were in the form of rods with typical dimensions 78 mm in diameter and @xmath1940 mm in length ( mass @xmath27 g ) .
neutron scattering measurements were performed on crystals with @xmath24 and @xmath25 on the triple - axis spectrometers ( tas ) in8 , in20 and in14 at the institut laue - langevin , and on rita - ii at sinq , paul scherrer institut .
the energies of the incident and scattered neutrons were selected by bragg reflection from crystal arrays of pyrolytic graphite ( pg ) crystals ( in8 , in14 , rita - ii ) , physically bent si crystals ( in8 ) or heusler arrays ( in20 ) .
the monochromators were vertically focused ( in8 , in14 ) and horizontally focused ( in8 , in20 , rita - ii ) to maximise neutron flux on the sample position .
the analyzers were horizontal focused on all instruments and vertically focused on in8 .
data were collected with fixed final neutron wavevectors of 2.662@xmath28 ( in8 , in20 ) , 1.5@xmath28 ( in14 , rita - ii ) and 1.2@xmath28 ( in14 ) .
a pyrolytic graphite filter ( in8 , in20 ) or be / beo filter operating at 77k ( in14 , rita - ii ) was placed between the sample and analyzer to suppress higher - order harmonic scattering .
on in20 polarized neutrons were employed , and the neutron spin polarization @xmath29 was maintained in a specified orientation with respect to the neutron scattering vector @xmath30 by an adjustable guide field of a few mt at the sample position .
a monitor is placed between the monochromator and sample position to determine the number of neutrons incident on the sample position .
the monitor count has an energy dependent contamination due to higher order neutrons in the incident beam , which we have corrected for when integrated intensities are shown .
the crystals we aligned so that the horizontal scattering plane was @xmath31 for @xmath5 and @xmath32 for @xmath3 ( we refer here to the tetragonal unit cell of the space group @xmath33 with unit cell parameters @xmath34 and @xmath35 ) .
the particular crystals used in the tas measurements were grown at oxford university and have been used in previous neutron scattering studies described elsewhere.@xcite to supplement the tas measurements we will also present some data on the @xmath24 composition collected on the maps time - of - flight spectrometer at the isis spallation neutron source . a report on the findings of this study at higher energy transfers ,
can be found elsewhere.@xcite the sample used on maps was an array of four crystals grown at kyoto university and co - aligned by x - ray diffraction to within about @xmath36 .
crystals with @xmath24 from the same source have been used in neutron diffraction studies of magnetic and charge order.@xcite the sample was mounted on maps in a closed - cycle refrigerator and aligned with the @xmath37 axis parallel to the incident beam direction .
a fermi chopper was used to select the incident neutron energy of 60mev .
the intensity was normalized and converted to units of scattering cross - section ( mbsr@xmath28mev@xmath28[f.u.]@xmath28 ) by comparison with measurements from a standard vanadium sample .
scattered neutrons were recorded in large banks of position - sensitive detectors .
the spin dispersion in la@xmath0sr@xmath1nio@xmath10 is highly two - dimensional,@xcite and so we project the data onto the @xmath38 two - dimensional reciprocal lattice plane .
the elastic energy resolution on maps was 2.7mev ( full - width at half - maximum ) .
to help visualise the experimental measurements we reproduce in fig .
[ fig : previous ] part of a figure from our previous publication on the q-1d magnetic fluctuations in la@xmath39sr@xmath40nio@xmath2 ( ref . ) . figure [ fig : previous](a ) is a simplified map of the low - energy scattering features in the @xmath31 plane of reciprocal space . the sharp peaks ( circles ) at @xmath41 and @xmath42 represent spin - wave scattering associated with the magnetic ordering wavevectors .
there are two pairs of peaks because there are two possible orientations of the stripes on the nio@xmath14 layers , either along the @xmath43 $ ] direction or the @xmath44 $ ] direction . the diagonal lines which run parallel to the stripe directions represent the approximate pattern of diffuse scattering observed in la@xmath39sr@xmath40nio@xmath2 , ref . .
figure [ fig : previous](b ) depicts an array of afm chains running parallel to the stripes which would give rise to the diagonal grid of diffuse scattering shown in fig .
[ fig : previous](a ) . in reality
, the the line of the observed diffuse scattering is not exactly straight but meanders slightly so as to follow approximately the magnetic zone boundaries , suggesting that the fluctuations on adjacent chains are weakly correlated .
the spins are shown pointing upwards to reflect the observation that the strength of the out - of - plane fluctuations is about twice the strength of the in - plane fluctuations.@xcite the diffuse scattering has a strong dynamic component on a thz frequency scale , and may even be entirely dynamic . whether the q-1d scattering is quasielastic or inelastic ( i.e. gapped ) is an open question which will be addressed later in this work . figure [ fig : x=033e=0 + 5mev ] presents maps of the magnetic scattering intensity measured in the @xmath45 plane of the @xmath46 crystal at a temperature of 2k .
the maps were collected at three fixed energies , ( a ) 0mev ( i.e. elastic scattering within the resolution of the spectrometer ) , ( b ) 2.5mev , and ( c ) 5mev . the 2.5mev data is reproduced from ref . .
figure [ fig : x=033e=0 + 5mev](d ) shows the areas covered in the three maps , which are not the same . in the elastic map , fig .
[ fig : x=033e=0 + 5mev](a ) , there is a magnetic bragg reflection at @xmath47 due to the pattern of afm order between the charge stripes , and a small spurious peak at @xmath48 .
no diffuse elastic scattering signal can be observed within the experimental precision . comparing the two inelastic maps , figs .
[ fig : x=033e=0 + 5mev](b ) and [ fig : x=033e=0 + 5mev](c ) , one can see that the q-1d diffuse scattering is broader at @xmath49mev than at at @xmath50mev .
this is consistent with our previous measurements@xcite which showed that the q-1d scattering disperses in the direction perpendicular to the scattering ridge , with a bandwidth of about 10mev .
the intensity of the q-1d scattering is modulated , with maxima adjacent to the spin - wave scattering from the afm order and at the positions where the diffuse ridges meet , i.e. @xmath51 and equivalent positions .
we now turn to the lsno crystal with @xmath3 . at this composition
the charge stripe order is incommensurate with the crystal lattice and it of interest to see whether this has any effect on the q-1d magnetic fluctuations .
figure [ fig : x=0_275e=3 - 5mev](a ) shows the distribution of scattering intensity from lsno ( @xmath24 ) measured on maps .
the scattering has been averaged over the energy range 35mev and plotted as a function of the in - plane components @xmath52 of reciprocal space .
there are four strong scattering signals at positions @xmath53 and @xmath54 with @xmath55 , from the steeply dispersing spin - wave excitations of the afm order . in fig
[ fig : x=0_275e=3 - 5mev](a ) the spin wave excitations from the afm order are too low in energy to be resolved into spin wave cones , and appear simply as spots .
in addition to these , there can also be seen ridges of diffuse scattering similar to the q-1d scattering observed from the @xmath5 crystal [ fig .
[ fig : x=033e=0 + 5mev](b , c ) ] .
figures [ fig : x=0_275e=3 - 5mev](b ) and [ fig : x=0_275e=3 - 5mev](c ) show cuts through this data along the paths marked a and b in fig.[fig : x=0_275e=3 - 5mev](a ) .
path b is chosen so that no intensity from the excitations of the ordered afm is observed in a scan along path b. the location for path b is determined experimentally , by performing scans perpendicular to path a through the excitations from the afm order .
the cut in fig .
[ fig : x=0_275e=3 - 5mev](b ) along path a shows the diffuse scattering as a shoulder to the sharper spin - wave peak , whereas the cut shown in fig .
[ fig : x=0_275e=3 - 5mev](c ) along path b shows just the diffuse scattering peak , which is centred on @xmath56 [ the @xmath57 coordinate measures the position along the scan projected perpendicularly onto line a , i.e. such that @xmath58 projects onto @xmath59 and @xmath60 projects onto @xmath61 . from these and
similar cuts we find that the diffuse scattering at @xmath62 follows the same slightly - meandering path as does the diffuse scattering at @xmath5 see figs . [
fig : x=033e=0 + 5mev](b ) and [ fig : x=033e=0 + 5mev](c ) . in fig .
[ fig : x=0_275e=3 - 5mev](b ) we also display the elastic scattering intensity along path a. this shows the magnetic bragg peak from the afm order but does not contain any elastic signal corresponding to the position of the diffuse inelastic signal . to confirm that the diffuse scattering signal is magnetic in origin we performed scattering measurements on in20 employing neutron polarization analysis . for these measurements we constrained the neutron polarization @xmath29 to be parallel to the scattering vector @xmath30 . in this
configuration scattering from electronic magnetic moments causes the neutron spin to flip , whereas scattering via non - magnetic processes does not .
figure [ fig : polx=0_275.eps ] displays the neutron spin - flip ( sf ) and non - spin - flip ( nsf ) scattering from the @xmath3 crystal along a line equivalent to path a. plots of elastic scattering data and inelastic scattering data with @xmath63mev are shown .
the elastic scan contains a strong peak in the sf channel centred on @xmath64 , the afm ordering wavevector for this composition .
a small peak in the nsf channel at the same position is due to imperfect spin polarization , which has not been corrected for . at @xmath65mev
there is a broad peak centred on @xmath66 in the sf channel but not in the nsf channel . due to their very steep dispersion ,
the spin - wave scattering from the afm order only accounts for the delta - shaped left - side of the sf peak.@xcite the remaining extent of the sf peak can be accounted for by the inelastic diffuse scattering .
the results of the unpolarized and polarized neutron scattering measurements presented here for @xmath3 are qualitatively very similar to our previous measurements@xcite on @xmath67 .
this shows that the q-1d diffuse scattering is magnetic in origin and has the same character in both the @xmath3 and @xmath68 samples .
next we turn to the question of whether the q-1d scattering is gapped or not . to answer this
we performed inelastic neutron scattering measurements on in14 on the @xmath5 crystal at energies @xmath69mev .
figure [ fig : x=0.333e=0.2,0.8 ] shows constant - energy scans at 0.2mev and 0.8mev along a direction equivalent to path b of fig . [ fig : previous](a ) .
the scan at 0.8mev reveals a peak centred on @xmath70 from the q-1d fluctuation , whereas the scan at 0.2mev shows no peak at @xmath71 within the statistical precision of the data .
the peak widths of both the q-1d diffuse scattering at 0.8mev and the spin - wave scattering at 0.6mev ( not shown ) are resolution - limited@xcite . by performing a series of constant - energy scans like those of fig.[fig : x=0.333e=0.2,0.8 ] and fitting these with gaussian peaks we determined the integrated intensity of the spin - wave and q-1d diffuse scattering peaks as a function of energy .
we remind the reader that in scans along path b only the q-1d excitations are observed , while in scans along path a the excitations from the afm order are clearly resolved from the q-1d.@xcite below 3mev we observed no energy variation of the integrated intensity of the spin excitations from the afm order along path a(not shown ) , placing an upper limit of 0.3mev on the in - plane anisotropy gap . in figure
[ fig : in14x=0.333 ] the energy variation of the integrated intensity of the diffuse scattering peak in @xmath5 is plotted . with increasing energy transfer the integrated intensity of the diffuse scattering increases monotonically up to @xmath72mev ,
then remains almost constant up to 3mev , the highest energy measured . in a separate experiment we investigated the q-1d diffuse scattering from the @xmath3 sample , performed on the rita - ii spectrometer . in the inset of figure [ fig : x=0_275lowe ]
we show constant - energy scans at 0.8mev and 0.5mev along a direction
equivalent to path a. for @xmath73mev the q-1d scattering is clearly observed and centred on @xmath74 . at @xmath75mev
there is still a small excess of scattering above the background at @xmath74 .
the centring is consistent with the meandering of the q-1d observed in fig .
[ fig : x=033e=0 + 5mev ] , and the energy dependence is consistent with our observations of the q-1d in the @xmath46 sample , see fig.[fig : x=0.333e=0.2,0.8 ] and fig.[fig : in14x=0.333 ] .
figure [ fig : x=0_275lowe ] shows a constant-@xmath30 scan at @xmath76 of the background - corrected amplitude of the q-1d in the @xmath3 , off - centred to avoid the excitations from the afm order .
the background was estimated from measurements at nearby wavevectors .
with decreasing energy transfer the signal remains roughly constant down to 0.8mev , then drops to a level close to zero at around 0.5mev . within the limitations of low counting statistics ,
this is consistent with the costant energy scans on the @xmath3 , and our observations of the q-1d in the @xmath46 sample , see fig.[fig : in14x=0.333 ] .
this work has been concerned with the nature of the low - energy q-1d diffuse scattering in la@xmath0sr@xmath1nio@xmath2 first found for @xmath68 ( ref . ) .
we have shown here that the diffuse scattering is also present at @xmath3 , and that it is the same for @xmath3 as for @xmath5 to within the experimental precision .
this implies that the q-1d scattering is not dependent on the periodicity of the spin - charge stripe order , and neither is it a consequence of the special circumstance found at @xmath25 in which the spin and charge order have the same periodicity and are commensurate with the crystal lattice .
instead , the diffuse scattering appears to be an intrinsic property of individual charge stripes embedded in the afm matrix formed by the ni@xmath15 spins , consistent with the interpretation in terms of q-1d correlations among spins in the charge stripes.@xcite in this work we have also carefully measured the low - energy lineshape of the q-1d diffuse scattering . to model the lineshape we recall that the dynamical part of the scattering is proportional to the linear response function @xmath77 where @xmath78 is the energy transferred to the system
, @xmath79 is the imaginary part of the dynamical susceptibility , and @xmath80 to describe the * q*-integrated intensities we used the phenomenological lorentzian lineshape constructed to satisfy detailed balance , @xmath81 \label{gap1}\ ] ] where @xmath82 is the lorentzian width ( half - width at half - maximum ) and @xmath83 is the energy of the undamped mode . as the q-1d has a large intrinsic energy width , @xmath83 is the characteristic energy for the q-1d , not a gap energy .
the best fit to the data achieved by this lineshape is shown as the solid line in fig .
[ fig : in14x=0.333 ] , and gives an energy @xmath84mev and @xmath85mev .
the fact that the fitted @xmath83 is non - zero provides a clear indication that the q-1d diffuse scattering is inelastic rather than quasielastic . to assess
how robust this result is we repeated the fit with @xmath83 fixed at zero , corresponding to a quasielastic lineshape .
the best fit thus obtained is shown with a broken line in fig.[fig : in14x=0.333 ] .
the width of the quasielastic fit was @xmath86mev .
the fit to the inelastic lineshape has a goodness - of - fit parameter @xmath87 = 1.19 compared with a @xmath87 = 2.25 for the quasielastic lineshape . on the strength of this evidence
we conclude that q-1d diffuse scattering corresponds to a gapped inelastic excitation . for an alternative representation of the inelastic lineshape we show in the inset to fig .
[ fig : in14x=0.333 ] the spectral weight function @xmath88 , which is related to the imaginary part of the dynamical susceptibility by @xmath89 .
the large energy width of the excitations indicates that the q-1d excitations are relatively short - lived , @xmath90s . in our original work on the q-1d spin correlations in lsno with @xmath68 we analyzed the magnetic spectrum with respect to that of an antiferromagnetic spin chain.@xcite assuming the doped holes reside in localized ni@xmath23 states with low - spin @xmath91 this implies that the fundamental excitations are spinons with a gapless dispersion . in the present study , however , we have presented evidence that the spectrum has a gap of 1.4mev .
further thought is needed , therefore , if we are to reconcile this information with our understanding of the q-1d magnetic diffuse scattering .
the model of an afm spin chain to describe the spin correlations along the ni@xmath23 charge stripes is reasonable providing the coupling between the spin chain and the afm order of the ni@xmath15 spins can be neglected .
the justification for neglect of this coupling is that the net heisenberg exchange acting on the ni@xmath23 sites from the afm - ordered ni@xmath15 spins cancels at the mean - field level , so the coupling between the spin chain and the afm order is frustrated .
however , if the individual ni@xmath23ni@xmath15 exchange interactions are strong enough then a weak ferromagnetic ( fm ) order could be induced on the ni@xmath23 stripes combined with a canting of the afm order of the ni@xmath15 spins .
such a possibility has been proposed by klingeler _
based on magnetization data.@xcite weak fm order would be difficult to detect in diffraction experiments because the associated magnetic bragg peaks would coincide with the bragg peaks from the crystal lattice . nevertheless , the competition between induced fm order and afm correlations from an effective afm exchange along the stripes could provide an explanation for the observed gapped dispersion .
such a model has been investigated recently with promising results.@xcite
this work has revealed that gapped , quasi-1d afm spin correlations are an intrinsic property of the charge - stripes in la@xmath0sr@xmath1nio@xmath2 .
the findings should inform theoretical models for the magnetic interactions in lsno , and hence contribute to a broader understanding of the formation and stability of spin
charge stripes in lsno and related systems .
the authors would like to acknowledge the help of h. j. woo in the maps experiment .
this work was performed in part at the swiss spallation neutron source sinq , at the paul scherrer institute ( psi ) , villigen , switzerland .
we are grateful for support from the engineering and physical sciences research council of great britain , the european commission under the 7th framework programme through the ` research infrastructures ' action of the ` capacities ' programme , contract no : cp - csa infra-2008 - 1.1.1 number 226507-nmi3 , and the support of a grant - in - aid for scientific research ( no .
22244039 ) from the mext , japan .
h. j. schulz , j. phys .
france * 50 * , 2833 ( 1989 ) ; h. j. schulz , phys .
lett . * 64 * , 1445 ( 1990 ) ; d. poilblanc , and t. m. rice , phys .
b * 39 * , 9749 ( 1989 ) ; j. zaanen , and o. gunnarsson , phys . rev .
b * 40 * , 7391 ( 1989 ) ; v. j. emery , and s. a. kivelson , physica c * 209 * , 597 ( 1993 ) ; h. johannesson and g. i. japaridze phys .
b 68 , 214507 ( 2003 ) .
s. m. hayden , g. h. lander , j. zarestky , p. j. brown , c. stassis , p. metcalf , and j. m. honig , phys .
lett . * 68 * , 1061 ( 1992 ) ; v. sachan , d. j. buttrey , j. m. tranquada , j. e. lorenzo , and g. shirane , phys .
b * 51 * , 12742 ( 1995 ) ; j. m. tranquada , d. j. buttrey , and v. sachan , phys . rev .
b * 54 * , 12318 ( 1996 ) ; e. d. isaacs , g. aeppli , p. zschack , s - w .
cheong , h. williams , and d. j. buttrey , phys .
lett . * 72 * , 3421 ( 1994 ) ; a. vigliante , m. von zimmermann , j. r. schneider , t. frello , n. h. andersen , j. madsen , d. j. buttrey , doon gibbs and j. m. tranquada , phys . rev .
b * 56 * , 8248 ( 1997 ) .
g. pashkevich , v. a. blinkin , v. p. gnezdilov , v. v. tsapenko , v. v. eremenko , p. lemmens , m. fischer , m. grove , g. gntherodt , l. degiorgi , p. wachter , j. m. tranquada , and d. j. buttrey , phys .
lett . * 84 * , 3919 ( 2000 ) .
h. yoshizawa , t. kakeshita , r. kajimoto , t.tanabe , t. katsufuji and y. tokura , phys .
b * 61 * , r854 ( 2000 ) ; h. yoshizawa , t. kakeshita , r. kajimoto , t. tanabe , t. katsufuji and y. tokura , physica b * 241 - 243 * , 880 ( 1998 ) . p. g. freeman , a. t. boothroyd , d. prabhakaran , m. enderle , and c. niedermayer , phys . rev .
b * 70 * , 024413 ( 2004 ) ; p. g. freeman , n. b. christensen , d. prabhakaran and a. t. boothroyd j. phys . :
* 200 * , 012037 ( 2010 ) .
the resolution functions for in14 were calculated using the restrax program@xcite .
for @xmath92@xmath28 the resolution ellipsoid was typically of dimensions 0.06mev in energy and 0.025@xmath28 parallel and perpendicular to the scan directions .
for @xmath93@xmath28 the resolution ellipsoid was typically of dimensions 0.14mev in energy , and 0.025@xmath28 and 0.035@xmath28 parallel and perpendicular to the scan direction , respectively . | the low energy spin excitations of la@xmath0sr@xmath1nio@xmath2 , @xmath3 and @xmath4 , have been investigated by unpolarized- and polarized - inelastic neutron scattering from single crystals . a pattern of magnetic diffuse scattering is observed in both compositions , and is consistent with quasi - one - dimensional afm spin correlations along the charge stripes .
analysis of the energy lineshape for @xmath5 indicates that the diffuse scattering is inelastic with a characteristic energy of @xmath6mev .
there is no discernible difference between the diffuse scattering from @xmath3 and @xmath5 , suggesting that it is an intrinsic property of the charge stripes . |
in recent years index type invariants of virtual knots have attracted a great deal of attention from researchers in virtual knot theory . roughly speaking , this kind of virtual knot invariants
are usually defined by counting the indices ( also called the weights ) of real crossing points in a virtual knot diagram .
the first invariant of this type was introduced by henrich in @xcite .
it is well known that there are no degree one vassiliev invariants in classical knot theory . in @xcite henrich constructed a sequence of vassiliev invariants of degree one for virtual knots .
by using the idea of parity discussed in @xcite , the author defined a polynomial invariant of virtual knots , say the odd writhe polynomial @xcite .
we choose this terminology because it generalizes the odd writhe , a numerical invariant of virtual knots which was first proposed by l. kauffman in @xcite .
later , these polynomial invariants were generalized independently by y. h. i m , s. kim and d. s. lee @xcite , lena c. folwaczny and l. kauffman @xcite , cheng and gao @xcite , s. satoh and k. taniguchi @xcite .
the key point of these invariants is one can assign an index to each real crossing point such that the signed sum of crossings with the same index is preserved under the generalized reidemeister moves .
several variations and applications of these invariants can be found in @xcite . an interesting feature of index type invariants is that one can use them to distinguish some virtual knot from its mirror image or inverse .
however there also exists one obvious drawback : if a real crossing has zero index then it has no contribution to the invariant .
recently this shortcoming was improved by myeong - ju jeong in @xcite . in @xcite
, jeong introduced the zero polynomial which focused on the real crossings with zero index .
some examples of virtual knots that have trivial writhe polynomial but nontrivial zero polynomial were given .
inspired by jeong s work , in this paper we will describe a new virtual knot invariant which generalizes several polynomial invariants mentioned above . for each oriented virtual knot diagram @xmath0
we will associate it with a transcendental function @xmath1 , which has the form @xmath2 . here
@xmath3 is a polynomial in @xmath4 . in section 4
it will be found that @xmath1 is invariant under generalized reidemeister moves , hence it is a virtual knot invariant .
we will discuss the relations between @xmath1 and other index type polynomial invariants mentioned above .
finally , some interesting properties of @xmath1 will be systematically studied in section 5 .
in this section we take a brief review of virtual knot theory and the definitions of several polynomial invariants of virtual knots .
virtual knot theory was first introduced by l. kauffman in @xcite . roughly speaking , classical knot theory studies the embeddings of circles in 3-dimensional euclidean space @xmath5 up to ambient isotopy .
evidently the ambient space @xmath5 can be replaced by @xmath6 . as an generalization of the classical knot theory , virtual knot theory studies the embeddings of circles in @xmath7 up to ambient isotopy , homeomorphisms of @xmath8 and the addition or substraction of empty handles , here @xmath8 denotes the closed orientable surface with genus @xmath9 . for simplicity
we say two embeddings are _ stably equivalent _ if one can be obtained from the other one by some operations above . when @xmath10 virtual knot theory recovers the classical knot theory @xcite
. another useful way to understand virtual knots is to regard them as the realizations of gauss diagrams . given a classical knot diagram , one can find a unique gauss diagram of it .
however for a given gauss diagram sometimes one can not realize it as a classical knot diagram . in order to settle this problem
, one needs to add some virtual crossings on the knot diagram , although these virtual crossings are not indicated on the given gauss diagram . since the way of adding virtual crossings is not unique , besides of the original reidemeister moves we need to add some virtual reidemeister moves to remove the arbitrariness @xcite .
see figure [ figure1 ] .
+ formally speaking , a virtual knot diagram can be obtained from a classical knot diagram by replacing some real crossings with virtual crossings .
usually a virtual crossing is represented by a small circle placed around the crossing point .
two virtual knot diagrams are _ equivalent _ if they can be connected by a finite sequence of generalized reidemeister moves .
_ virtual knots _ can be defined as the equivalence classes of virtual knot diagrams . for a given virtual knot diagram @xmath0 ,
consider the one point compactification of the plane where the diagram locates in , then one gets a 2-sphere of the diagram .
let us attach a 1-handle to a virtual crossing and regard this virtual crossing as an overpass locally .
after interpreting all virtual crossings like this and thickening the surface , we will obtain an embedding of @xmath11 in @xmath12 where @xmath13 is the number of virtual crossing points in @xmath0 .
an important fact is these two interpretations of virtual knots coincides .
@xcite two virtual knot diagrams are equivalent if and only if their corresponding surface embeddings are stably equivalent .
since one motivation of introducing virtual knots is to realize arbitrary gauss diagram , sometimes it is more convenient to study the corresponding gauss diagram than to study the virtual knot diagram .
let us give a short review of the definition of the gauss diagram .
let @xmath0 be a virtual knot diagram , the _ gauss diagram _ of @xmath0 , written @xmath14 , is an oriented ( counterclockwise ) circle where the preimages of each real crossing are indicated .
for the two preimages of a real crossing we add a chord connecting them , which is directed from the overcrossing to the undercrossing .
finally each chord is associated with a sign according to the writhe of the corresponding real crossing .
one simple example is given in figure [ figure2 ] .
we remark that virtual reidemeister moves @xmath15 have no effect on the gauss diagram .
therefore as an advantage , when we study virtual knots from the viewpoint of gauss diagrams we only need to consider the classical reidemeister moves @xmath16 .
+ in order to classify virtual knots , one needs to introduce some virtual knot invariants .
some classical knot invariants can be extended directly to virtual knots , such as the jones polynomial and the knot quandle .
later these invariants were generalized using the virtual structure " of virtual knots . for jones polynomial , based on kauffman s
approach to it , after smoothing all real crossings one can assign a weight to each circle .
the state - sum gives a generalization of jones polynomial for virtual knots .
this were independently finished by y. miyazawa @xcite and h. dye , l. kauffman @xcite , now known as the miyazawa polynomial and the arrow polynomial . on the other hand , the idea of quandle was generalized to some complicated algebraic structure , such as the biquandle @xcite and virtual biquandle @xcite .
we refer the reader to @xcite for more details on these developments .
now we recall the definition of the writhe polynomial introduced in @xcite .
later we will show how to obtain other index type virtual knot invariants from this polynomial invariant .
let @xmath0 be a virtual knot diagram and @xmath14 the corresponding gauss diagram .
since there is a bijection between the real crossings of @xmath0 and the chords of @xmath14 , we will use the same letter to refer a real crossing point and the corresponding chord . choose a chord @xmath17 in @xmath14 ,
now let us assign an index to it , which will play an important role in the definition of the writhe polynomial .
let @xmath18 @xmath19 denotes the number of positive ( negative ) chords crossing @xmath17 from left to right , let @xmath20 @xmath21 denotes the number of positive ( negative ) chords crossing @xmath17 from right to left ( see figure [ figure3 ] ) .
following @xcite , we define the _ index _ of @xmath17 as ind@xmath22 .
we mention some similar indices appeared in the literature . in @xcite henrich
defined an _ intersection index _ for each chord , which equals to the absolute value of the index here . in @xcite dye introduced a _ parity mapping _ from the chords of a gauss diagram to @xmath23 , which is exactly the inverse of the index used here .
+ here we list some useful properties of ind@xmath24 , which can be easily verified . 1 . if @xmath17 is isolated , i.e. no other chord has intersection with @xmath17 , then ind@xmath25 .
2 . the two crossings appeared in @xmath26 have the same index .
the indices of the three crossings appeared in @xmath27 are invariant under @xmath27 .
@xmath28 @xmath29 preserves the indices of chords that do not appear in @xmath28 @xmath29 .
5 . if @xmath0 contains no virtual crossings , then every crossing of @xmath0 has index zero .
ind@xmath24 is invariant under switching some other real crossings . due to the properties above
, one can define a numerical invariant of virtual knots as below @xmath30 @xmath31 , where @xmath32 denotes the writhe of @xmath33 and @xmath34 denotes the set of real crossings of @xmath0 .
note that @xmath35 @xmath31 , where @xmath36 denotes the writhe of @xmath0 .
the key observation is that @xmath37 can be generalized by counting chords with a fixed index . for the sake of simplicity
we write it in the form of a polynomial .
we define the _ writhe polynomial _
@xmath38 to be @xmath39 . note that @xmath40 . on the other hand
, we remark that this definition is slightly different from that given in @xcite , where the polynomial is equal to @xmath41 .
one can easily deduce that @xmath38 is a virtual knot invariant from the properties listed above . before discovering the indices of chords , it was first observed by l. kauffman @xcite that @xmath42 , which is named the _ odd writhe _ , is a virtual knot invariant .
later this numerical invariant was generalized to the _ odd writhe polynomial _ , which equals to @xmath43 .
again the definition we give here is also a bit different from the one we gave in @xcite . besides of the writhe polynomial
, the odd writhe polynomial was independently generalized by several groups .
we list the connections between them and the writhe polynomial below . 1 .
the _ parity writhe polynomial _
@xmath44 introduced in @xcite can be described as + @xmath45 .
+ note that replacing the variable @xmath46 with @xmath47 will not weaken the invariant . in this case , it coincides with @xmath48 .
the _ affine index polynomial _
@xmath49 defined in @xcite satisfies + @xmath50 .
+ it is worth mentioning that in @xcite the algebraic structure behind the affine index polynomial was discussed .
it was proved that this kind of index polynomial is essential the unique one derived from an affine linear flat biquandle with coefficients in a commutative ring without zero divisors .
the _ @xmath51th parity writhe _
@xmath52 @xcite is equal to the coefficient of @xmath53 in @xmath38 . before ending this section we will give an interesting application of using writhe polynomial to detect the non - classicality of virtual knots .
a long - standing open problem in knot theory is whether there exists a nontrivial knot with trivial jones polynomial .
however in virtual knot theory , it is well - known that there exist infinitely many nontrivial virtual knots with trivial jones polynomial .
more precisely , we have the following result .
@xcite let @xmath0 be a nontrivial classical knot , then there is a corresponding nontrivial virtual knot @xmath54 with trivial jones polynomial .
we give a sketch of the proof , see @xcite for more details .
choose a diagram of @xmath0 , let us still use @xmath0 to denote it .
we can find a set of crossings such that after switching all these crossings one will obtain a trivial knot . now for each crossing we take a local replacement ( virtualization ) , see figure [ figure4 ] .
denote the new virtual knot diagram by @xmath54 . note that switching a crossing and virtualizing a crossing have the same effect on the jones polynomial
, it follows that the jones polynomial of @xmath54 is trivial .
however @xmath54 has the same involutory quandle ( also called _ kei _ ) as @xmath0 , which is nontrivial since @xmath0 is nontrivial .
thus , @xmath54 is a nontrivial virtual knot with unit jones polynomial .
+ now we can ask the following two questions . 1 .
whether every nontrivial virtual knot that has trivial jones polynomial can be obtained by virtualization ?
is @xmath54 non - classical ?
if both the answers are yes , then we can conclude that jones polynomial detects the unknot .
for this reason , it was suggested by l. kauffman @xcite to characterize which virtualization can be detected by index type invariants .
here we give a partial answer to this question .
note that this result is in fact a special case of the main theorem in @xcite , which is obtained from an analysis of the knot group .
let @xmath0 be a positive ( or negative ) classical knot diagram with unknotting number 1 .
then the corresponding virtual knot @xmath54 is non - classical . without loss of generality , we assume that @xmath0 is positive . choose a crossing @xmath17 such that switching @xmath17 will unknot @xmath0 .
let us consider the gauss diagram @xmath14 .
first notice that @xmath17 can not be an isolated chord in @xmath14 , otherwise @xmath17 is a nugatory crossing and switching @xmath17 preserves @xmath0 . after taking virtualization on @xmath17
, we obtain the gauss diagram @xmath55 , which can be derived from @xmath14 by changing the sign of @xmath17 .
since @xmath0 is a classical knot diagram , it follows that ind@xmath25 in @xmath14 .
let @xmath56 @xmath57 denotes the number of positive chords crossing @xmath17 from left to right , then there are also @xmath56 positive chords crossing @xmath17 from right to left .
now it is an easy exercise to check that @xmath58 , which is nontrivial .
in this section we give the definition of @xmath1 . the main idea of
the construction is to replace the integer ind@xmath24 in @xmath38 by a polynomial @xmath59 .
let @xmath0 be a virtual knot diagram and @xmath17 a real crossing of it .
denote the chords crossing @xmath17 from left to right and the chords crossing @xmath17 from right to left by @xmath60 and @xmath61 respectively .
let @xmath62 denotes the canonical quotient map from @xmath23 to @xmath63 .
now we define the _ index function _ @xmath59 as below @xmath64 . more precisely , if ind@xmath25 , in this case @xmath65 , therefore @xmath66 . if ind@xmath67 , then @xmath62 sends every integer into zero , hence @xmath68 . in general
, @xmath59 takes values in @xmath69/(s^{|\text{ind}(c)|}-1)$ ] . for the writhe polynomial @xmath38 and the numerical invariant @xmath37
, we define a refined version for each of them by letting @xmath70 , and @xmath71 .
now we define the function @xmath1 as @xmath72 we remark that according to the definition of the index function , it is easy to see that @xmath73 .
hence @xmath74 in fact takes values in @xmath69/(s^{|g_{c_i}(1)|}-1)$ ] , which guarantees the definition above is well - defined .
@xmath1 is a virtual knot invariant . here
we make some remarks on this invariant .
first , in fact it is not essential to use two variables @xmath75 and @xmath4 .
one can easily find that replacing @xmath4 with @xmath75 will not weaken this invariant .
however it is more convenient to regard it as a generalization of some other polynomial invariants if we use @xmath1 rather than @xmath76 .
for example , we have @xmath77 , thus @xmath78 .
on the other hand , myeong - ju jeong introduced the zero polynomial in @xcite , which was defined by @xmath79 , where @xmath80 .
the zero polynomial can be used to detect the non - classicality of some virtual knots that have trivial writhe polynomials . according to the definition above one
can easily deduce that @xmath81 .
together with the discussion in section 2 , we see that @xmath1 indeed generalises all the polynomial invariants mentioned in section 1 .
another important thing we want to point out is , @xmath82 and @xmath83 are mutually independent .
let us make a bit more explanation about this .
at the beginning of this story , we have two numerical invariants of virtual knots , @xmath30 and @xmath84 .
obviously they are mutually the inverse of each other .
after the first generalization , we have two polynomial invariants , @xmath39 and @xmath85 .
we can regard the first part as the contributions from the chords with nontrivial indices , and the second part contains the contributions from the chords with index zero .
however , the second part is preserved hence it can be derived from the first part . after the second generalization , now we have two transcendental function invariants of virtual knots , @xmath70 and @xmath86 .
later we will give some examples to show that @xmath82 and @xmath83 are mutually independent , i.e. neither can be derived directly from the other one .
since @xmath87 , @xmath82 and @xmath83 can be viewed as two different generalizations of our original numerical invariant @xmath37 .
now we give one example to explain how to calculate @xmath1 .
let us consider the virtual knot @xmath0 , the gauss diagram @xmath14 is depicted in figure [ figure5 ] .
first we write down the index of each chord @xmath88 .
next we list the index function of each chord @xmath89 .
it follows that @xmath90 .
we remark that in this example the writhe polynomial and zero polynomial are both trivial . ,
title="fig : " ] + , title="fig : " ] + the second example concerns the three virtual knots described in figure [ figure6 ] .
direct calculation shows that & w_k_1(t , s)=t+t^-1 , q_k_1(t , s)=2 , + & w_k_2(t , s)=t+t^-1 , q_k_2(t , s)=-t^1-s^-1-t^s^-1 - 1 + 4 , + & w_k_3(t , s)=3t^-1-t^-3s , q_k_3(t , s)=2 .
note that @xmath91 , but @xmath92 .
this implies that in general @xmath82 and @xmath83 are mutually independent , as we claimed before .
theorem 2.3 tells us that writhe polynomial can be used to detect the non - classicality of the virtualization of some classical knots .
however there exists some classical knot whose virtualization has the trivial writhe polynomial .
the third example was suggested by l. kauffman in @xcite .
the classical knot @xmath0 illustrated in figure [ figure12 ] has unknotting number 3 .
for example , switching crossings @xmath93 yields the trivial knot .
straightforward calculation shows that @xmath94 but @xmath95 , which means @xmath54 is non - classical and hence nontrivial .
in this section we give the proof of theorem 3.1 . before doing this , we need a simple lemma .
let @xmath96 be the three crossings of the virtual knot diagram locally described on the left side of figure [ figure7 ] . then ind@xmath97-ind@xmath98+ind@xmath25 . according to the connecting ways outside of the local diagram ,
there are two possibilities of the gauss diagram of @xmath0 , say @xmath99 and @xmath100 ( see figure [ figure7 ] ) .
let us consider @xmath99 .
due to the sixth property of index we listed in section 2 , without loss of generality , we assume there are @xmath101 positive chords intersects @xmath56 and @xmath17 , @xmath102 and @xmath17 , @xmath56 and @xmath102 respectively . see the middle graph of figure [ figure7 ] for the directions of these chords .
it follows directly that ind@xmath103 , ind@xmath104 , ind@xmath105 , which implies that ind@xmath97-ind@xmath98+ind@xmath25 .
the proof for @xmath100 is analogous .
, title="fig : " ] + + now we give the proof of theorem 3.1 .
it suffices to check that @xmath1 is invariant under all reidemeister moves illustrated in figure [ figure8 ] @xcite .
the only chord involved in @xmath106 is isolated .
therefore it has index zero and the indices of other chords are invariant .
consequently the index function of the isolated chord is zero , and the index functions of other chords are invariant under @xmath106 .
it is obvious that @xmath1 is preserved under @xmath106 .
similarly one can prove that @xmath107 also preserves @xmath1 . for @xmath108
, first we notice that the two chords involved have the same index but different writhes .
secondly , if a chord intersects one of them then it also has nonempty intersection with the other one . for these reasons ,
the two chords involved in @xmath108 have the same index functions but different writhes .
it follows that the contributions from these two chords to @xmath1 will cancel out . on the other hand
, these two chords will not affect the index function of any other chord , which implies that the contribution from any chord which is not involved in @xmath108 is invariant under @xmath108 .
we conclude that @xmath1 is invariant under @xmath108 .
now let us consider the third reidemeister move @xmath109 . as we mentioned in lemma 4.1 , there exist two possibilities for the gauss diagram of the whole knot diagram .
we only consider the one corresponding to @xmath99 , the proof for @xmath100 is quite similar .
the variation of the gauss diagram @xmath99 under @xmath109 is depicted in figure [ figure9 ] .
recall the properties ( 3 ) and ( 4 ) mentioned in section 2 , the index of each chord is preserved under @xmath109 . for any chord @xmath110 which is not involved in @xmath109 , it is not difficult to observe that the index function @xmath111 is invariant
. therefore it suffices to consider the behaviors of @xmath112 under @xmath109 .
for the gauss diagram on the left side of figure [ figure9 ] , we have & g_a(s)=a(s)+s^_a(-(b))-s^_a(-(c ) ) , + & g_b(s)=b(s)+s^_b((a))-s^_b(-(c ) ) , + & g_c(s)=c(s)+s^_c((a))-s^_c((b ) ) . here
@xmath113 count the contributions from those chords not involved in @xmath109 to @xmath112 respectively , @xmath114 denote the projection from @xmath23 to @xmath115 respectively . for the gauss diagram on the right side of figure [ figure9 ] , one computes that g_a(s)=a(s ) , g_b(s)=b(s ) , g_c(s)=c(s ) .
recall that lemma 4.1 tells us ind@xmath97-ind@xmath98+ind@xmath24=0 .
thus & s^_a(-(b))-s^_a(-(c))=s^_a(-(a)-(c))-s^_a(-(c))=0 , + & s^_b((a))-s^_b(-(c))=s^_b((b)-(c))-s^_b(-(c))=0 , + & s^_c((a))-s^_c((b))=s^_c((b)-(c))-s^_c((b))=0 , + the proof is completed . under @xmath109,title="fig : " ] +
in this section we listed some basic properties of @xmath1 .
first of all , similar to the other index type invariants , @xmath1 is good at detecting the non - classicality of virtual knots .
choose a chord @xmath17 in @xmath14 , we denote the corresponding chords in @xmath123 and @xmath124 by @xmath125 and @xmath126 respectively . according to the definition of chord index , it is not difficult to observe that recall that the real crossing number of a virtual knot diagram is the number of real crossings of this diagram . in virtual knot theory ,
the _ real crossing number _
@xmath134 of a virtual knot @xmath0 is the smallest number of real crossings of any diagram of @xmath0 .
obviously @xmath135 if and only if @xmath0 is trivial , and there is no virtual knot with @xmath136 . according to the definition of @xmath1 , it is easy to see that @xmath1 gives a natural lower bounder for @xmath134 .
for example , let us consider the virtual knot @xmath0 described in figure [ figure11 ] .
direct calculation shows that @xmath141 and @xmath142 .
the writhe polynomial tells us that @xmath143 .
in contrast we have @xmath144 , which implies @xmath145 .
since the gauss diagram in figure [ figure11 ] has only 4 chords , we conclude that @xmath146 .
next we study the behavior of @xmath1 under a connected sum . because of the forbidden move , in general the connected sum
is not well - defined for virtual knots .
for example , the connected sum of two trivial knots maybe nontrivial @xcite . similar to other index type invariants of virtual knots , @xmath1 is additive under a connected sum .
this proposition reveals a shortcoming of @xmath1 .
if a nontrivial virtual knots is a connected sum of two trivial knots , then we can not detect it via calculating the @xmath1 of it .
for example , the kishino knot depicted in figure [ figure10 ] has trivial @xmath1 .
however it is well known that the kishino knot is nontrivial ( in fact even the corresponding flat virtual knot of it is nontrivial ) , which can be detected by , for example , the arrow polynomial @xcite .
we remark that the kishino knot also provides a counterexample to the first question we listed in the end of section 2 , since the involutory quandle of kishino knot is trivial .
further , if each crossing of @xmath0 has index zero , proposition 5.1 tells us that @xmath116 .
the kishino knot is a special case of this . for writhe
polynomial , if one chord has index zero then it has no contribution to the writhe polynomial . for @xmath1 ,
the contribution comes from index zero chord may not be zero .
however @xmath1 is powerless to distinguish a virtual knot from the unknot if each chord of it has index zero .
it is a challenging question to deal with this kind of knots by introducing some extended index invariants .
the three loop isotopy invariant @xcite defined by assigning a weight to a pair of non - intersecting chords may offer some hints to this problem . finally , let us take a moment to discuss the degree of the invariant we introduced in the present paper . in virtual knot theory , there are mainly two kinds of finite - type invariants .
one was introduced by l. kauffman in his introductory paper of virtual knot theory @xcite , the other one was proposed by goussarov , polyak , and viro in @xcite . here
we will focus on the kauffman finite - type invariants , for the gpv finite - type invariants we refer readers to @xcite .
we remark that if a virtual knot invariant is a gpv finite - type invariant then it must be a kauffman finite - type invariant . on the other hand , if a knot diagram is classical , in this case kauffman finite - type invariants are exactly the classical finite - type invariants @xcite .
recall that a virtual knot invariant @xmath154 taking values in an abelian group is called a _
finite - type invariant of degree @xmath155 _ , if for any virtual knot diagram @xmath0 with @xmath156 singular crossings we have here @xmath158 runs over all @xmath159-tuples of zeros and ones , @xmath160 denotes the number of ones in @xmath158 and @xmath161 is obtained from @xmath0 by replacing the @xmath162-th singular crossing with a positive ( negative ) crossing if the @xmath162-th position of @xmath158 is zero ( one ) .
the minimal @xmath51 is said to be the _ degree _ of @xmath154 .
let @xmath164 denote the two real crossings obtained from the singular crossings in @xmath161 @xmath165 .
first we assume that the chord corresponding to @xmath166 has no intersection with the chord corresponding to @xmath167 .
it is not difficult to observe that @xmath168 , @xmath169 , and the index of any other chord in @xmath161 @xmath165 is equivalent .
it follows that for any real crossing @xmath17 in @xmath0 , the corresponding chords in @xmath161 @xmath165 have the same index function .
therefore we only need to consider the contributions from @xmath164 to @xmath170 , which equals roughly speaking , the reason why @xmath1 is a finite - type invariant of degree one is that @xmath1 is a weighted sum of each chord in the gauss diagram . if one wants to define a finite - type invariant of higher degree , a sensible idea is to associate a weight to each combinatorial structure of several chords , then take the sum over all subgraphs with this structure in the gauss diagram .
recently , a family of finite - type invariants of degree two were defined in this way by chrisman and dye in @xcite .
it is an interesting question to find some explicit virtual knot invariants with higher degrees .
99 joan birman , xiaosong lin .
knot polynomials and vassiliev s invariants . invent .
, 1993 , 111 : 225 - 270 zhiyun cheng , hongzhu gao . a polynomial invariant of virtual links . journal of knot theory and its ramifications , 2013 , 22(12 ) : 1341002 ( 33 pages ) zhiyun cheng . a polynomial invariant of virtual knots .
proceedings of the american mathematical society , 2014 , 142(2 ) : 713 - 725 micah w. chrisman , heather a. dye . the three loop isotopy and framed isotopy invariants of virtual knots .
topology and its applications , 2014 , 173 : 107 - 134 heather a. dye , louis h. kauffman .
virtual crossing number and the arrow polynomial .
journal of knot theory and its ramifications , 2009 , 18(10 ) : 1335 - 1357 heather a. dye .
vassiliev invariants from parity mappings .
journal of knot theory and its ramifications , 2013 , 22(4 ) : 1340008 ( 21 pages ) r. fenn , m. jordan - santana , louis h. kauffman .
biquandles and virtual links .
topology and its applications , 2004 , 145 : 157 - 175 lena c. folwaczny , louis h. kauffman . a linking number definition of the affine index polynomial and applications .
journal of knot theory and its ramifications , 2013 , 22(12 ) : 1341004 ( 30 pages ) m. goussarov , m. polyak , o. viro .
finite - type invariants of classical and virtual knots .
topology , 2000 , 39 : 1045 - 1068 allison henrich . a sequence of degree one vassiliev invariants for virtual knots .
journal of knot theory and its ramifications , 2010 , 19(4 ) : 461 - 487 y. h. i m , s. kim , d. s. lee .
the parity writhe polynomials for virtual knots and flat virtual knots .
journal of knot theory and its ramifications , 2013 , 22(1 ) : 1250133 ( 20 pages ) myeong - ju jeong .
reidemeister moves and a polynomial of virtual knot diagrams .
journal of knot theory and its ramifications , 2015 , 24(2 ) : 1550010 ( 16 pages ) myeong - ju jeong .
a zero polynomial of virtual knots .
journal of knot theory and its ramifications , 2015 , 24 : 1550078 ( 19 pages ) louis h. kauffman .
virtual knot theory .
j. combinatorics , 1999 , 20 : 663691 louis h. kauffman .
a self - linking invariant of virtual knots .
, 2004 , 184 : 135158 louis h. kauffman , v. o. manturov . virtual biquandles . fund .
, 2005 , 188 : 103 - 146 louis h. kauffman .
introduction to virtual knot theory .
journal of knot theory and its ramifications , 2012 , 21(13 ) : 1240007 ( 37 pages ) louis h. kauffman . an affine index polynomial invariant of virtual knots .
journal of knot theory and its ramifications , 2013 , 22(4 ) : 1340007 ( 30 pages ) joonoh kim . the affine index polynomial invariant of flat virtual knots .
journal of knot theory and its ramifications , 2014 , 23(14 ) : 1450073 ( 10 pages ) t. kishino , s. satoh . a note on non - classical virtual knots .
journal of knot theory and its ramifications , 2004 , 13(7 ) : 845 - 856 greg kuperberg .
what is a virtual link ?
algebraic and geomertric topology , 2003 , 3 : 587 - 591 vassily manturov .
parity in knot theory .
sbornik : mathematics , 2010 , 201 : 693 - 733 vassily manturov , denis ilyutko .
virtual knots - the state of the art . world scientific , 2012 y. miyazawa .
magnetic graphs and an invariant for virtual links .
journal of knot theory and its ramifications , 2006 , 15(10 ) : 1319 - 1334 y. miyazawa .
a multivariable polynomial invariant for unoriented virtual knots and links .
journal of knot theory and its ramifications , 2008 , 17(11 ) : 1311 - 1326 michael polyak .
minimal generating sets of reidemeister moves .
quantum topology , 2010 , 1 : 399 - 411 s. satoh , k. taniguchi .
the writhes of a virtual knot . fund . math .
, 2014 , 225(1 ) : 327 - 342 d. silver , s. williams . on a class of virtual knots with unit jones
journal of knot theory and its ramifications , 2004 , 13(3):367 - 371 | in this work we describe a new invariant of virtual knots .
we show that this transcendental function invariant generalizes several polynomial invariants of virtual knots , such as the writhe polynomial @xcite , the affine index polynomial @xcite and the zero polynomial @xcite . |
results of molecular dynamics simulations of fission reactions @xmath0 and @xmath1 are presented .
dependence of the fission barriers on isomer structure of the parent cluster is analyzed .
it is demonstrated that the energy necessary for removing homothetic groups of atoms from the parent cluster is largely independent of the isomer form of the parent cluster .
importance of rearrangement of the cluster structure during the fission process is elucidated .
this rearrangement may include transition to another isomer state of the parent cluster before actual separation of the daughter fragments begins and/or forming a `` neck '' between the separating fragments .
fission of charged atomic clusters occurs when repulsive coulomb forces , arising due to the excessive charge , overcome the electronic binding energy of the cluster @xcite .
this mechanism of the cluster fission is in a great deal similar to the nuclear fission phenomena .
experimentally , multiply charged metal clusters can be observed in the mass spectra when their size exceeds the critical size of stability , which depends on the metal species and cluster charge @xcite .
we report the results of the _ ab initio _ molecular dynamics ( md ) simulations of the fission processes @xmath0 and @xmath2 .
both symmetric and asymmetric fission channels are considered .
we have investigated the parent cluster isomer dependence of the fission barrier for the reaction @xmath0 . to the best of our knowledge ,
a comparative study of fission barriers for various isomers by means of quantum chemistry methods has not been carried out before .
note that such a study is beyond the scope of simpler approaches which do not account for ionic structure of a cluster .
we found that the direct separation barrier for the reaction @xmath3 has a weak dependence on the isomeric structure of the parent cluster .
we note , however , that the groups of atoms to be removed from the parent cluster isomers must be chosen with care ; one has to identify homothetic groups of atoms in each fissioning isomer .
the weak dependence on the isomeric state of the parent @xmath4 cluster implies that the particular ionic structure of the cluster is largely insignificant for the shape and height of the fission barrier .
this is due to the fact that the maximum of fission barriers in considered cases are located at distances comparable or exceeding the sum of the resulting fragments radii . at such distances
the interaction between the fragments , apart from the coulombic repulsion , is mainly determined by the electronic properties rather than by the details of the ionic structure of the fragments .
this is an important argument for justification of the jellium model approach to the description of the fission process of multiply charged metal clusters .
we have demonstrated the importance of rearrangement of the cluster ionic structure during the fission process .
the possibility of rearrangement of the cluster structure leads to the fact that direct fission of a cluster isomer in some cases may not be the energetically optimum path for the fission reaction .
alternatively , the reaction can go through transition to another isomer state of the parent cluster .
this transition can occur in the first phase of the fission process , before separation of the fragments actually begins .
we show that this is the case for the fission of @xmath5 and @xmath6 isomers of @xmath4 cluster .
the rearrangement of ionic structure may be important also after the fragments began to separate . for @xmath1 reaction , two magic fragments @xmath7 form a metastable transitional state in which the fragments are connected by a `` neck '' .
this `` necking '' allows for significant reduction in the height of the fission barrier .
note that the similar necking phenomenon is known for the nuclear fission process @xcite . in our molecular dynamics simulations we utilize methods of density
functional theory ( dft ) . within the dft one has to solve the kohn - sham equations @xcite @xmath8 where the first term corresponds to the kinetic energy of an electron from the kohn - sham reference system
, @xmath9 describes the attraction of the @xmath10 electron to the nuclei in the cluster , @xmath11 is the electronic orbital , @xmath12 is the hartree part of the inter - electronic interaction , @xmath13 @xmath14 is the electron density , @xmath15 is the local exchange - correlation potential defined as the functional derivative of the exchange - correlation energy functional @xmath16}{\delta \rho({\bf r})},\ ] ] where the exchange - correlation energy is partitioned into two parts , referred to as _ exchange _ and _ correlation _ parts : @xmath17= e_{\rm x}(\rho ) + e_{\rm c } ( \rho).\ ] ] physically , these two terms correspond to same - spin and mixed - spin interactions , respectively . both parts are functionals of the electron density , which can be of two distinct types : either a local functional depending on the electron density @xmath18 only or a gradient - corrected functional depending on the electron density and its gradient @xmath19 . there is a variety of exchange - correlation functionals in the literature .
we have used the three - parameter becke - type gradient - corrected exchange functional with the gradient - corrected correlation functional of lee , yang , and parr ( b3lyp ) @xcite . for the explicit form of this functional we refer to the original papers @xcite .
the b3lyp functional has proved to be a reliable tool for studying the structure and properties of small metal clusters .
it provides high accuracy at comparatively low computational costs . for a discussion and a comparison with other approaches ,
see @xcite .
note that the density of the parent cluster and two daughter fragments ( including the overlapping region before scission point ) almost does not change during the fission process ( by analogy with the deformed jellium model , see @xcite for more details ) .
this means that the b3lyp method works adequately for any fragment separation distances , @xmath20 , during the fission process .
the calculations have been carried out with the use of the gaussian 98 software package @xcite .
the 6 - 311g(d ) and lanl2dz basis sets of primitive gaussian functions have been used to expand the cluster orbitals @xcite .
the 6 - 311g(d ) basis has been used for simulations involving @xmath4 cluster .
this basis set takes into account electrons from all atomic orbitals , so that the dynamics of all particles in the system is taken into account . for @xmath21 cluster
we have used more numerically efficient lanl2dz basis , for which valent atomic electrons move in an effective core potential ( see details in @xcite ) . to simulate the fission process
we start from the optimized geometry of a cluster ( for details of the geometry optimization procedure see @xcite ) and choose the atoms the resulting fragments would consist of .
the atoms chosen for a smaller fragment are shifted from their locations in the parent cluster to a certain distance .
then , the multidimensional potential energy surface , its gradient and forces with respect to the molecular coordinates are calculated .
these quantities specify the direction along the surface in which the energy decreases the most rapidly and provide information for the determination of the next step for the moving atoms . if the fragments are removed not far enough from each other then the attractive forces prevailed over the repulsive ones and the fragments stuck together forming the unified cluster again . in the opposite situation
the repulsive forces dominate and the fragments drift away from each other .
the dependence of the total energy of the system on the fragment separation distance forms the fission barrier .
the aim of our simulations is to find the fission pathway corresponding to the minimum of the fission barrier .
there are usually many stable isomers of a cluster with energies slightly exceeding the energy of the ground state isomer . in order to analyze the isomer dependence of the fission barrier in the reaction
@xmath22 we have picked two energetically low - lying isomers with the point symmetry groups @xmath5 and @xmath6 differing from the distorted @xmath23 point symmetry group of the ground state parent @xmath4 cluster .
three isomer states of the @xmath4 cluster are shown in figure [ isomers ] . cluster . from left to right : the ground state isomer of distorted @xmath24 point symmetry group ( total energy is -1622.7063 a.u . ) ; an isomer of @xmath5 point symmetry group ( total energy is -1622.6888 a.u .
, that exceeds the lowest energy state by 0.476 ev ) ; an isomer of @xmath6 point symmetry group ( total energy is -1622.6860 a.u .
, that exceeds the lowest energy state by 0.553 ev ) .
the homothetic group of three atoms marked by black color . ] in figure [ isobarriers ] we show fission barriers for separation three atoms from the @xmath5 , @xmath6 , and @xmath23 isomers of the @xmath4 cluster . in this figure zero level of energy
is chosen for each parent isomer separately and corresponds to the minimum of total energy of that isomer .
the initial distances between the centers of mass of two ( future ) fragments are finite so that the barriers do not start at the origin . ) from three isomers of @xmath4 cluster derived from molecular dynamics simulations ( direct @xmath0 fission channel ) .
the barriers plotted versus distance between the centers of mass of the fragments .
solid , dashed , and dashed - dotted lines correspond to distorted @xmath24 , @xmath5 , and @xmath6 point symmetry groups isomers of the parent cluster , respectively .
energies are measured from the energy of the ground state of the corresponding isomers , i.e. we plot @xmath25 , where @xmath26 is the total energy of the system and @xmath27 are the ground energies of the @xmath24 , @xmath5 and @xmath6 isomer states of the parent @xmath4 cluster , respectively . ]
the barriers for all three channels are close .
the weak sensitivity of the fission barrier on the isomeric states of the reactants can be explained if one notices that the barrier maxima are located at distances comparable to or exceeding the sum of the resulting fragments radii , that is not far from the scission point . at such distances
the interaction between the fragments , apart from coulombic repulsion , is mainly determined by the electronic properties rather than by the details of the ionic structure of the fragments .
this is an important argument for justification of the jellium model approach to the description of the fission process of multiply charged metal clusters .
it is important to note that the barriers presented in figure [ isobarriers ] are calculated in assumption that fission occurs for the fixed ( given ) isomers .
however , since @xmath5 and @xmath6 isomers are not the lowest energy states of @xmath4 system , there could be other processes competing with fission .
one of such processes is rearrangement of the cluster structure .
rearrangement of the cluster structure during the fission process may significantly reduce the fission barrier .
such rearrangement may occur before the actual separation of the daughter fragments begins or after that .
system and schematic barriers for transitions between these states .
energies are measured from the energy of the ground @xmath24 state of the @xmath4 cluster . ]
channel derived from molecular dynamics simulations as a function of distance between the centers of mass of the fragments .
energy is measured from the energy of the ground @xmath28 state of the @xmath21 cluster .
the arrow shows position of the meta - stable transitional state , see also figure .
this results is in a good agreement with the results of the jellium model @xcite . ] . from left to right : ground state of the parent cluster ; `` necking '' between the two fragments leads to a meta stable intermediate state and significantly reduces the fission barrier height ; two @xmath7 fragments drifting away from each other . ]
fission of @xmath5 and @xmath6 isomers of @xmath4 cluster is an example of situation where rearrangement of the cluster structure takes place before the fragments start to separate . in figure
[ levels ] we show schematically the total energies of the @xmath4 @xmath24 , @xmath5 and @xmath6 isomers and barriers for the transitions between those states .
it is seen from the figure that transition to the ground ( @xmath24 ) state with subsequent fission into the @xmath29 and @xmath30 fragments , @xmath31 , ( shown by solid lines ) is preferred path for fission of both @xmath5 and @xmath6 isomers of the @xmath4 cluster and requires only about 0.2 ev for @xmath5 isomer and 0.26 ev for @xmath6 isomer .
in contrast , the direct fission process , @xmath32 , ( shown by dashed lines ) requires about 0.5 ev .
we also show the barrier for the transition between the @xmath5 and @xmath6 isomers .
another example of cluster structure rearrangement in the fission process is the @xmath1 reaction .
the fission barrier for this reaction is shown in figure [ fission18 ] .
if two fragments of the parent cluster were not allowed to adjust their ionic structure the fission barrier would be about 1 ev .
rearrangement of the cluster structure allows to reduce the fission barrier down to 0.31 ev . during the fission process
the daughter fragments start to drift away from each other and a `` neck '' forms between the fragments .
formation of the `` neck '' results in a metastable transitional state .
the geometry of this state , as well as the geometry of the parent cluster are shown in figure [ rearrangement ] . in table
[ tab ] we have summarized our results for the fission barrier heights and compared them with the results of other molecular dynamics simulations and with the predictions of the jellium model .
@lll & @xmath0 & @xmath1 + md ( this work ) & 0.49 ( distorted @xmath24 ) & 0.31 + md@xcite & 0.67 & 0.52 + md@xcite & 0.54 & + jellium model@xcite & 0.16 & 0.48 + we have investigated two aspects of charged metal cluster fission process : dependence of the fission barrier on isomer state of the parent cluster and importance of rearrangement of the cluster ionic structure during the fission process .
we found that for a consistent choice of the atoms removed from the cluster the fission barrier for the reaction @xmath0 has a weak dependence on the initial isomer structure of the parent cluster .
this implies that the particular ionic structure of the cluster is largely insignificant for the height of the fission barrier .
this is an important argument for justification of the jellium model approach to the description of the fission process of multiply charged metal clusters .
we have shown importance of rearrangement of the cluster ionic structure during the fission process .
the fission reaction can go through transition to another isomer state of the parent cluster .
this transition can occur before actual separation of the fragments begins and/or `` neck '' between the separating fragments is formed . in any case
the resulting fission barrier can be significantly lower compared to the one for the direct fission path .
the authors acknowledge support of this work by the alexander von humboldt foundation , dfg , the studienstiftung des deutschen volkes , intas , russian foundation for basic research ( grant no 03 - 02 - 16415-a ) , russian academy of sciences ( grant 44 ) and the royal society of london . |
|
the suppression of @xmath5 production in relativistic heavy ion collisions maybe one of the important signatures to identify the possible phase transition to the quark - gluon plasma @xcite .
the dissociation of the @xmath5 in the quark - gluon plasma due to color screening can lead to a reduction of its production , on the other hand , the @xmath5 suppression maybe already present in the hadron - nucleus collisions .
it is necessary to know absorption of the @xmath5 by the co - mover light mesons before we can make a definitive conclusion .
the values of the @xmath5 absorption cross sections by the light hadrons are not known empirically , we have to resort to some theoretical approaches . among existing approaches , the one - meson exchange model and the effective @xmath2 theory
are typical @xcite .
the detailed knowledge about the strong coupling constants which are basic parameters in the effective lagrangian is of great importance .
furthermore , the strong coupling constants among the charmed mesons and light - mesons play an important role in understanding final - state interactions in the hadronic @xmath6 decays @xcite
. there have been many works dealing with the strong coupling constants concerning the charmed mesons , see e.g. refs.@xcite . in ref.@xcite , the authors calculate the strong coupling constant @xmath7 with the qcd sum rules , and obtain much larger value than what expected from the effective @xmath2 theory . in this article , we calculate the value of the @xmath8 with the light - cone qcd sum rules .
the light - cone qcd sum rules carry out the operator product expansion near the light - cone @xmath9 instead of the short distance @xmath10 while the non - perturbative matrix elements are parameterized by the light - cone distribution amplitudes ( which classified according to their twists ) instead of the vacuum condensates @xcite .
the article is arranged as : in section 2 , we derive the strong coupling constant @xmath1 with the light - cone qcd sum rules ; in section 3 , the numerical result and discussion ; and in section 4 , conclusion .
we study the strong coupling constant @xmath11 with the two - point correlation function @xmath12 , @xmath13 where the currents @xmath14 and @xmath15 interpolate the vector mesons @xmath16 and @xmath17 respectively , the external state @xmath18 has the four momentum @xmath19 with @xmath20 . according to the basic assumption of current - hadron duality in the qcd sum rules @xcite ,
we can insert a complete series of intermediate states with the same quantum numbers as the current operators into the correlation function to obtain the hadronic representation . after isolating the ground state contributions from the pole terms of the mesons @xmath21 , we get the following result , @xmath22 where the following weak decay constant and phenomenological lagrangian have been used , @xmath23 here @xmath24 , @xmath25 @xcite .
we carry out the operator product expansion for the correlation function @xmath26 in perturbative qcd theory , and obtain the analytical expressions at the level of quark - gluon degrees of freedom , then perform the double borel transformation and match the quark - hadron duality below the threshold @xmath27 , finally we obtain the sum rule for the strong coupling constant @xmath1 , @xmath28-\exp\left[- \frac{s^0_{\rho}}{m^2 } \right ] \right\ } \nonumber\\ & & + \exp\left[- \frac{m_c^2+u_0(1-u_0)m_{\rho}^2}{m^2 } \right]\left\ { \left[f_{\rho}^\perp - f_{\rho}\frac{m_u+m_d}{m_{\rho}}\right]m_c m_{\rho}^2 h_{||}^{(s)}(u_0 ) \right.\nonumber\\ & & \left.-\frac{f_{\rho } m_{\rho}^3a(u_0)}{4 } \left[1 + \frac{m_c^2}{m^2}\right ] - \frac{2f_{\rho}m_c^2 m_{\rho}^3}{m^2 } \int_0^{u_0 } d\tau \int_0^\tau dt c(t)\right\ } \ , , \end{aligned}\ ] ] where @xmath29 here we have neglected the contributions from the gluons , the contributions proportional to @xmath30 can give rise to three - particle ( and four - particle ) meson distribution amplitudes with a gluon ( and quark - antiquark pair ) , their corrections are always suppressed by the large borel parameter and will not exceed @xmath31 , for examples , one can consult the last two articles of ref.@xcite . in calculation , the two - particle @xmath18 meson light - cone distribution amplitudes up to twist-4 have been used , for explicit expressions , one can consult ref.@xcite . due to the special tensor structure @xmath32
, some two - particle twist-4 light - cone distribution amplitudes have no contributions .
the parameters in the light - cone distribution amplitudes are scale dependent , in this article , we take @xmath33 . as the dominating contribution ( about @xmath34 )
comes from the two - particle twist-2 term involving the @xmath35 , for other terms , the continuum subtractions will not affect the result remarkably , we can neglect the subtractions . in some cases ,
the contributions from the two - particle twist-3 light - cone distribution amplitudes are very large ( or dominating ) @xcite , they depend heavily on the currents we choose to interpolate the mesons .
the input parameters are taken as @xmath36 , @xmath37 , @xmath38 , @xmath39 , @xmath40 , @xmath41 , @xmath42 , @xmath43 , @xmath44 , @xmath45 , @xmath46 , @xmath47 , @xmath48 , @xmath49 , @xmath50 , @xmath51 , @xmath52 , @xmath53 , @xmath54 , @xmath55 and @xmath56 @xcite .
the central value of the decay constant @xmath57 from lattice simulation is about @xmath58 @xcite . in this article
, we take the value @xmath59 from the two - point qcd sum rules without perturbative @xmath60 corrections for consistency .
the duality threshold parameter @xmath27 is chosen to be @xmath61 , the numerical ( central ) value of @xmath27 is taken from the qcd sum rules for the mass of the @xmath62 @xcite .
the borel parameters are chosen as @xmath63 and @xmath64 , in those regions , the value of the @xmath1 is rather stable , the uncertainty from the borel parameter is very small , less than @xmath65 . in the limit of large borel parameter @xmath66 ,
the strong coupling constant @xmath1 takes up the following behavior , @xmath67 it is not unexpected , the contribution from the twist-2 light - cone distribution amplitude @xmath35 is greatly enhanced by the large borel parameter @xmath66 , its contribution is dominating , about @xmath34 , ( large ) uncertainties of the relevant parameters presented in above equation have significant impact on the numerical result .
the main uncertainties come from the two parameters @xmath57 and @xmath68 , variations of the two parameters can lead to large uncertainties , about @xmath69 . taking into account all the uncertainties ,
finally we obtain the numerical value for the strong coupling constant @xmath1 , which is shown in fig.1 , @xmath70 if we take the replacement @xmath71 \rightarrow \exp\left[- \frac{m_c^2+u_0(1-u_0)m_{\rho}^2}{m^2 } \right]-\exp\left[- \frac{s^0_{\rho}}{m^2 } \right ] $ ] to subtract the continuum contributions of the terms besides the twist-2 light - cone distribution amplitude , the central value @xmath72 will decrease about @xmath65 .
comparing with the values from the effective @xmath2 symmetry and vector meson dominance theory , @xmath73 @xcite , our numerical value @xmath74 is much smaller . in the vector meson dominance theory , @xmath75 , the contributions from the radial excited states @xmath76 , @xmath77 , @xmath78 , @xmath79
are neglected .
we can make a crude estimation with the simple replacement @xmath80 to take into account the contribution from the @xmath76 , @xmath81 , our numerical result is reasonable , the @xmath2 symmetry breaking effect for the strong coupling constants is small . in ref.@xcite ,
the authors introduce some functions to parameterize the form - factor for small spacelike @xmath82 , then extrapolate to the mass shell , much larger value is obtained , @xmath83 . although the model functions have solid theoretical foundation at large @xmath82 , extrapolation to the mass shell may have good or bad behaviors , which correspond to the systematic errors .
it is not unexpected that the values from the qcd sum rules and light - cone qcd sum rules are different . with the borel parameter @xmath66 .
in this article , we analyze the vertex @xmath0 with the light - cone qcd sum rules .
the strong coupling constant @xmath1 is an important parameter in evaluating the charmonium absorption cross sections in searching for the quark - gluon plasmas .
our numerical value for the @xmath1 is consistent with the prediction of the effective @xmath2 symmetry and vector meson dominance theory .
this work is supported by national natural science foundation , grant number 10405009 , 10775051 , and program for new century excellent talents in university , grant number ncet-07 - 0282 , and key program foundation of ncepu .
s. g. matinyan , b. muller , phys . rev .
* c58 * ( 1998 ) 2994 ; k. l. haglin , phys . rev .
* c61 * ( 2000 ) 031902 ; z. w. lin , c. m. ko , phys . rev .
* c62 * ( 2000 ) 034903 ; a. sibirtsev , k. tsushima , a. w. thomas , phys . rev .
* c63 * ( 2001 ) 044906 .
f. s. navarra , m. nielsen , m. e. bracco , phys .rev . *
d65 * ( 2002 ) 037502 ; z. h. li , t. huang , j. z. sun , z. h. dai , phys . rev .
* d65 * ( 2002 ) 076005 ; m. e. bracco , m. chiapparini , f. s. navarra , m. nielsen , phys . lett .
* b605 * ( 2005 ) 326 .
i. i. balitsky , v. m. braun and a. v. kolesnichenko , nucl . phys .
* b312 * ( 1989 ) 509 ; v. l. chernyak and i. r. zhitnitsky , nucl . phys . *
b345 * ( 1990 ) 137 ; v. m. braun and i. e. filyanov , z. phys .
* c44 * ( 1989 ) 157 ; v. m. braun and i. e. filyanov , z. phys .
* c48 * ( 1990 ) 239 .
p. ball , v. m. braun , nucl .
* b543 * ( 1999 ) 201 ; p. ball , v. m. braun , phys .
* d54 * ( 1996 ) 2182 ; p. ball , v. m. braun , y. koike , k. tanaka , nucl .
* b529 * ( 1998 ) 323 ; p. ball , g. w. jones , r. zwicky , phys . rev .
* d75 * ( 2007 ) 054004 ; p. ball , g. w. jones , jhep * 0703 * ( 2007 ) 069 . | in this article , we analyze the vertex @xmath0 with the light - cone qcd sum rules .
the strong coupling constant @xmath1 is an important parameter in evaluating the charmonium absorption cross sections in searching for the quark - gluon plasmas .
our numerical value for the @xmath1 is consistent with the prediction of the effective @xmath2 symmetry and vector meson dominance theory . + zhi - gang wang@xmath3 , zhi - bin wang@xmath4 + @xmath3 department of physics , north china electric power university , baoding 071003 , p. r. china + @xmath4 college of electrical engineering , yanshan university , qinhuangdao 066004 , p. r. china + pacs numbers : 12.38.lg ; 13.25.ft ; 14.40.lb * key words : * strong coupling constant , light - cone qcd sum rules |
applications of statistical mechanics to fermion systems with discrete spectrum , such as semiconductor quantum dots @xcite , naturally involve single - particle averages in statistical ensembles with a fixed number of particles .
in particular , the kinetic theory of tunneling @xcite through quantum dots with fast intra - dot electron relaxation involves average level occupation number , @xmath2 in a gibbs distribution of @xmath0 independent fermions populating a set of single - particle energy levels @xmath3 ( enumerated by @xmath4 ) . here
@xmath5 is the canonical partition function @xmath6 and @xmath7 is the inverse thermodynamic temperature . for fermions , occupation numbers
@xmath8 in the sum ( [ eq : zdef ] ) take values @xmath9 and @xmath10 .
the behavior of @xmath11 is simple in two extreme limits of the typical level @xmath12 . for @xmath13 and large @xmath0 ,
( [ eq : avdef])-([eq : zdef ] ) reduce to the standard fermi - dirac distribution , @xmath14 here @xmath15 is the chemical potential determined by the normalization condition @xmath16 . in the low temperature limit , @xmath17 , most of the statistical weight in eq .
( [ eq : zdef ] ) is in the ground state ( defined as @xmath18 for @xmath19 and @xmath20 otherwise ) . in this case
it is common @xcite to take only one excited state into account resulting in a two - state gibbs distribution which is equivalent to eq .
( [ eq : fd ] ) for @xmath21 with @xmath22 and @xmath23 @xcite .
for finite @xmath24 , the average occupation number @xmath25 deviates from eq .
( [ eq : fd ] ) in a non - universal way which depends on the details of the energy spectrum @xcite .
exact analytical investigation of this regime is complicated by the combinatorial explosion in the number of levels with comparable statistical weights .
particle number projection technique based on fourier extraction @xcite gives an exact closed - form formula for @xmath26 ( see eq .
( 140 ) of @xcite ) that scales quadratically with the number of levels ; however , its potential for analytical investigation appears to be limited due to the additional sum over the fourier variable . in the paper
we address the problem of exact evaluation of canonical occupation numbers @xmath26 by deriving a general formula that scales linearly in the number of levels and quadratically in the number of particles , eqs .
( [ eq : znmain ] ) and ( [ eq : occumain ] ) below .
the energy spectrum @xmath27 enters the formula only in terms of the single - particle partition function @xmath28 computed at @xmath0 different temperatures .
we apply this general result to equidistant spectrum , @xmath29 , and derive an exact formula for @xmath26 in terms of polynomials in @xmath30 , eq .
( [ eq : exactgfd ] ) . in the limit of degenerate fermi gas , @xmath31 ,
these polynomials converge to partial theta function @xcite which is involved in a number of combinatorial proofs @xcite of ramanujan s identities @xcite .
the exact result for the equidistant spectrum can be approximated well by tailoring _
two _ standard fermi - dirac distributions ( [ eq : fd ] ) with different chemical potentials for holes and for particles , @xmath32 and @xmath33 , respectively . in the high- and
the low - temperature limits , this approximation converges to the asymptotically exact fermi - dirac and two - state gibbs distributions , respectively . at intermediate temperatures ,
@xmath34 , particle - hole correlation effects due to fixed @xmath0 result in finite deviations from the exact solution .
grand canonical partition function @xmath35 serves as a generating function for the canonical partition functions @xmath5 if expanded power series of @xmath36 , @xmath37 @xmath38 is most conveniently calculated via its logarithm @xcite , @xmath39 where @xmath40 , and @xmath41 is the canonical partition function of a _ single _ particle .
relation between @xmath42 and @xmath43 is the same as between the raw moments and the cumulants of a univariate probability distribution and is given by the complete bell polynomials @xcite , @xmath44 the latter satisfy a recurrence relation @xcite , @xmath45 which translates into @xmath46 \ , . % \frac{\kappa_n}{n ! } + \sum_{m=0}^{n-1 } \left ( \begin{array}{c}\ ! n\!-\!1 \\ m\!-\!1 \end{array } \ !
\right ) \kappa_m \frac{(n - m)!}{n ! }
z_{n - m } = % \frac{\kappa_n}{n ! } + \sum_{m=0}^{n-1 } \frac{\kappa_m z_{n - m}}{n ( m\!-\!1)!}\ ] ] we set @xmath47 identically . combining eqs .
( [ eq : avdef ] ) , ( [ eq : plainy ] ) and ( [ eq : logy ] ) gives the generating function for the occupation numbers : @xmath48 expanding the r.h.s . in powers
series in @xmath49 gives @xmath50\ ] ] equations ( [ eq : znmain ] ) and ( [ eq : occumain ] ) constitute our main general result .
for @xmath51 , the grand canonical partition function ( [ eq : logy ] ) can be expressed by an infinite product , @xmath52 where @xmath30 and @xmath53 is the @xmath1-shifted factorial @xcite .
using @xmath1-analog binomial theorem of euler ( @xcite , formula 17.2.35 ) we can get the partition function directly from eq .
( [ eq : plainy ] ) , @xmath54 applying eq .
( [ eq : occumain ] ) , and transforming @xmath1-shifted factorials ( @xcite , formula 17.2.13 ) , one gets @xmath55 equation ( [ eq : gfdpoly1 ] ) defines occupation numbers for @xmath0 fermions populating equidistant levels at equilibrium .
it is clear from the explicit form ( [ eq : pdef ] ) that @xmath56 is a laurent polynomial ( the product contains negative powers of @xmath1 if @xmath57 ) .
however , since @xmath58 for @xmath59 by definition ( [ eq : avdef ] ) , the negative powers of @xmath1 must cancel , thus we conclude that @xmath56 is always an ordinary polynomial in @xmath1 for @xmath60 , @xmath61 .
this cancelation is not trivial and deems further mathematical investigation @xcite .
a number of recurrence formulas can be derived for @xmath56 @xcite , including a symmetry relation @xmath62 using ( [ eq : pdef ] ) in the r.h.s . of ( [ eq : symmetry ] )
gives a sum of products with no negative powers of @xmath1 at @xmath57 .
if @xmath63 then the fermi gas is degenerate @xcite and the limit of @xmath64 is appropriate . for @xmath65 , and @xmath66 the polynomial sum in eq .
( [ eq : pdef ] ) becomes a geometric series which gives @xmath67 where @xmath68 is known as partial theta function is an instance of _ false theta series _
@xcite in the sense of l .j .
rogrers @xcite . ]
the partial theta function is famous for a number of identities discovered by ramanujan in his lost notebook @xcite .
these identities have been studied extensively , including some recent proofs by combinatorial methods @xcite . using the symmetry relation ( [ eq : symmetry ] ) for @xmath57 ,
gives the particle - hole complementary result : @xmath69 . in terms of level energies
@xmath70 , the occupation numbers in a canonical degenerate fermi gas with constant levels spacing @xmath12 can be written as @xmath71 where @xmath72 and @xmath73.the function defined in eq .
( [ eq : piecwiseform ] ) can be considered a @xmath1-analog @xcite of the standard fermi - dirac distribution ( [ eq : fd ] ) since @xmath74 . and @xmath75 .
crosses mark the exact values , thick ( blue ) continuous line shows the @xmath1-analog of fermi - dirac distribution , eq .
( [ eq : piecwiseform ] ) , thin continuous line substitution @xmath76 in eq .
( [ eq : piecwiseform ] ) , and short - dashed line single fermi - dirac distribution , eq .
( [ eq : fd ] ) .
the long - dashed lines between @xmath77 and @xmath78 are extrapolations of the corresponding functions into the gap @xmath79 .
[ fig : fig1],scaledwidth=70.0% ] equation ( [ eq : piecwiseform ] ) expresses two essential deviations of canonical occupation numbers from fermi - dirac distribution .
firstly , the @xmath1-analog @xmath80 is different from @xmath81 .
an approximation of substituting @xmath76 in eq .
( [ eq : piecwiseform ] ) becomes exact both for @xmath82 and for @xmath83 .
numerically , we find maximal absolute deviation @xmath84 of @xmath85 reached for @xmath86 at @xmath87 .
for the two levels closest to the gap , @xmath21 approximating @xmath76 gives @xmath88 $ ] which is equivalent to two - state gibbs approximation @xcite .
a comparison between the exact result ( [ eq : exactgfd ] ) , the large-@xmath0 limit ( [ eq : piecwiseform ] ) and a single fermi - dirac distribution is shown in fig .
[ fig : fig1 ] for @xmath89 and @xmath75 . secondly , barring the difference between @xmath80 and @xmath81 , eq .
( [ eq : piecwiseform ] ) can be seen as a combination of two fermi - dirac distributions with different chemical potentials for particles , @xmath90 , and for holes , @xmath91 , respectively .
this is precisely what is to be expected if one considers particle and hole excitations from the ground state as two uncorrelated microcanonical ensembles .
in this view , the moderate difference in the functional dependence between @xmath92 and @xmath93 results from correlation between a particle and a hole created in a single pair - excitation act . note that the difference @xmath94 can not be ignored even as @xmath95 and @xmath96 , thus using a single fermi - dirac distribution , eq .
( [ eq : fd ] ) , necessarily fails for finite @xmath12 and low temperature .
the author is grateful to carlo beenakker , bruce c. berndt , jan mangaldan , juris smotrovs , and michael somos for discussions .
this research has been supported by esf project no .
2009/0216/1dp/1.1.1.2.0/09/apia / viaa/044 .
10 url # 1#1urlprefix[2][]#2 kouwenhoven l p , austing d g and tarucha s 2001 _ rep .
phys . _ * 64 * 701 | we consider equilibrium level occupation numbers in a fermi gas with a fixed number of particles , @xmath0 , and finite level spacing . using the method of generating functions and the cumulant expansion
we derive a recurrence relation for canonical partition function and an explicit formula for occupation numbers in terms of single - particle partition function at @xmath0 different temperatures .
we apply this result to a model with equidistant non - degenerate spectrum and obtain close - form expressions in terms of @xmath1-polynomials and rogers - ramanujan partial theta function .
deviations from the standard fermi - dirac distribution can be interpreted in terms of a gap in the chemical potential between the particle and the hole excitations with additional correlations at temperatures comparable to the level spacing . |
electrons have proven to be useful probes for the study of nuclear structure , primarily because the @xmath2 interaction is reasonably well - known and because the electromagnetic interaction is weak compared to the strong forces which dominate nuclear structure . the longitudinal contribution to the nuclear response
is of particular interest because of its relationship to two - body correlation functions through what is commonly known as the coulomb sum rule ( csr).@xcite with the assumption that the nucleus may be treated as a nonrelativistic system of interacting nucleons , the extraction of nucleon - nucleon correlations is formally direct : this is commonly referred to as the nonrelativistic coulomb sum rule ( nrcsr ) . in practice , however , one finds that to obtain useful information one must extend the experiments and analysis to energies and momentum transfers sufficiently large that the usual nonrelativistic assumptions fail
. this will be particularly the case for experiments at the continuous electron beam accelerator facility ( cebaf ) with @xmath3 .
therefore , it has become desirable to have a fully - relativistic formalism both for analyzing experimental data and for testing theoretical models of nuclear structure . in this paper , we present an extension of the nrcsr which successfully handles certain problems which arise in electron scattering from nuclei at large three - momentum transfers . before discussing further the goals of this paper
, it will be useful to review the form of the nrcsr , as well as its derivation .
we begin with the first - order born ( or one - photon exchange ) approximation for the scattering of ultrarelativistic electrons ( @xmath4 ) from nuclear targets , in which one can factor the @xmath0 differential cross - section into leptonic and nuclear parts . after performing a standard separation of this nuclear response function into longitudinal and transverse ( virtual ) photon contributions , the differential cross - section in the laboratory frame can be written @xmath5,\label{aa}\ ] ] where the mott cross - section @xmath6 describes the elastic scattering of ultrarelativistic electrons from a _ fixed _
point target with charge @xmath7 and no spin . here
@xmath8 is the four - momentum transfer to the target , @xmath9 is the electron scattering angle , @xmath10 is the electron beam energy and @xmath11 .
the first term in the square brackets in ( [ aa ] ) is the longitudinal contribution , and is usually expressed in terms of the coulomb response function : @xmath12 where @xmath13 and @xmath14 denote initial .
the results can easily be generalized to unpolarized targets with @xmath15 . ] and final nuclear states , respectively , and @xmath16 is the fourier transform of the nuclear charge density operator . in general , the operator @xmath16 should include contributions from both the nucleons and ( virtual ) charged mesons in the target , but it has been conventional in nonrelativistic approximations to ignore the latter .
then @xmath16 is given by the spatial fourier transform of the local nucleon charge density operator : @xmath17 where @xmath18 is the local ( nonrelativistic ) nucleon field - operator , @xmath19 is the charge operator in isospin space which projects out protons , and @xmath20 is the proton charge form factor .
( in a purely nonrelativistic treatment of the target , @xmath21 would be replaced by the fourier transform of the proton charge density @xmath22 . )
charge effects of neutrons are small and are usually neglected . to proceed to a sum rule
, we first define the nonrelativistic coulomb sum function for inelastic scattering at fixed three - momentum transfer * q * : @xmath23 where the integration is over all energies above the elastic peak at @xmath24 . expression ( [ ae ] ) then becomes a sum rule by subtracting the elastic term @xmath25 explicitly and using the closure relation @xmath26 on the resulting sum over nuclear states @xmath14 .
equation ( [ ae ] ) is then in the form of a fourier transform of the expectation value of four nucleon field operators . to separate out the two - body correlation part
, we put the operators in normal order ( creation operators to the left ) by anticommutation , using @xmath27 the sum rule then becomes @xmath28 the first term is simply the total nuclear charge @xmath29 , and the second term can be written in terms of a correlation function : @xmath30 and the elastic form factor of the nuclear target : @xmath31 where @xmath32 . combining ( [ ah])([aj ] )
, the nrcsr takes the compact form @xmath33 the function @xmath34 is simply the fourier transform of the spatial two - proton correlation function .
it has been defined so that @xmath35 for a system of @xmath29 uncorrelated protons .
since @xmath36 is measured by elastic @xmath0 scattering , the nrcsr ( [ ak ] ) provides a direct measure of proton - proton correlations . since
both @xmath36 and @xmath34 are expected to vanish in the limit of large @xmath37 , the approach of @xmath38 with increasing @xmath37 is a test of the adequacy of the theoretical assumptions .
experimental studies of @xmath0 carried out in the momentum transfer region up to @xmath39 over the last decade have left serious questions about the status of the nrcsr . for light targets : @xmath40 , the measured coulomb sum @xmath41 for the largest momentum values reached.@xcite for heavier targets , however , @xmath42 by 20 - 30% compared to theoretical predictions based on independent particle models,[38 ] with larger suppressions occurring for larger nuclei .
this is despite the fact that pauli correlations are expected to vanish for @xmath43 .
we shall not discuss all the possible effects which could be involved here , and which have been reviewed elsewhere.@xcite however , one problem which is always present in the application of the nrcsr to data is whether the sum defined in ( [ ae ] ) is in fact saturated by the experimental data : not only is there an absolute relativistic limit @xmath44 due to the spacelike nature of virtual photons in electron scattering , but other experimental difficulties prevent one from reaching even that limit in practice .
we expect that extending experiments to higher energies and momenta , such as will be done at cebaf , may improve this situation .
the main problem dealt with in this paper is the extension of the coulomb sum rule to include the relativity of nucleons in inelastic scattering .
this becomes unavoidable for momentum transfers @xmath45 , the nucleon mass , even if the nuclear target is essentially nonrelativistic .
this follows because even if the nucleons in a typical nucleus move at essentially nonrelativistic velocities , they become highly relativistic after the absorption of a virtual photon carrying large three - momentum .
thus to obtain an adequate description , one must describe nucleons with relativistic wavefunctions , e.g. , solutions to the dirac equation .
this in turn introduces the difficulty of treating antinucleon degrees of freedom , which we deal with approximately .
other issues also arise at higher energies , including the composite nature of nucleons themselves .
this includes both nucleon electromagnetic form factors , which we include , and inelastic excitations of nucleons , which we do not .
we also do not consider the effects of exchange currents among nucleons , e.g. , as carried by charged mesons .
these have been studied by schiavilla _
et al._,@xcite for example . for the purpose of this paper ,
the nucleus consists of interacting relativistic nucleons , considered to be elementary particles with form factors .
a number of authors have considered relativistic extensions of the nrcsr with the approximations just mentioned .
walecka@xcite made the most direct extension by integrating the relativistic coulomb response function over all energies to allow the use of closure .
this results in a sum rule which is nearly identical in form to the nrcsr .
matsui@xcite showed , however , that such a sum rule would never be saturated by electron scattering , for which @xmath44 .
he showed that , in the fermi gas model , excitations of the fermi sea ( @xmath46 scattering ) are found entirely in the spacelike response , while excitations of the dirac sea ( @xmath1 pair - production ) are found entirely in the timelike response . unlike the nrcsr , for a system of @xmath29 dirac protons the spacelike relativistic coulomb sum tends to @xmath47 in the high-@xmath37 limit as a result of this separation .
this work initiated interest in spacelike coulomb sum rules based on energy transfers of @xmath44 only .
several authors have studied the case of a fermi gas target.[1316 ] we will see that the exclusion of timelike contributions from the coulomb sum introduces kinematic factors related to relativistic nucleon recoil which must be accounted for in the formulation of _ sum rules _ for large @xmath37 .
deforest@xcite recognized the issue of relativistic nucleon recoil and proposed a correction in the form of a modified electric form factor : @xmath48 , where @xmath49 , which would be factored from the response function in the definition of the relativistic coulomb sum .
he conjectured that the resulting sum rule should have an interpretation similar to the nrcsr , but did not show this .
deforest s prescription has been used in the analysis of recent data.@xcite however , we will show ( in sec .
5 ) that a result similar to his can be derived ( rcsr - i ) and does lead to a sum rule which bears close resemblence similar to the nrcsr , but is only accurate for dirac nucleons , i.e. , nucleons without anomalous magnetic moments . donnelly _ et al._@xcite have also proposed a modification to the definition of the spacelike coulomb sum . by requiring that their modified coulomb sum tend to @xmath29 in the large-@xmath37 limit , an integral equation for the modified form factor can be derived , then solved via an expansion in moments of the nucleon momentum . in
lowest - order , this corresponds to the deforest prescription evaluated at zero nucleon momentum .
however , in higher order , this method introduces an energy dependence into the coulomb sum function which then is _ not _ equivalent to a _ non - energy - weighted _ sum rule of the nrcsr form , from which correlation information can be extracted .
we discuss these issues further in the concluding section . in this paper
we derive a spacelike coulomb sum rule valid for arbitrary three - momentum transfers , in which the relativity of the nucleons is taken into account .
we assume that nucleons are the only degrees of freedom contributing to the spacelike sum and ignore , for example , antinucleon degrees of freedom , meson exchange currents and internal excitations of nucleons .
in section 2 , we review the formalism for electron scattering from nuclear targets in the first - order born approximation , including nucleon form factors with an assumption about their off - shell continuation . in section 3
, we consider elastic scattering from a single , free nucleon at rest as the simplest example of relativistic nucleon recoil effects in the constant-*q * response ; this is shown to become a large effect for @xmath45 . in section 4 , we develop our `` nucleons - only '' approximation and derive a relativistic coulomb sum rule which is exact in its treatment of nucleon recoil and fermi motion .
we give an explicit form for the two - body correlation function . in section 5
, we derive an expansion of the one - body term in moments of the nucleon momentum .
this leads to practical sum rules for the analysis of experimental data . in section 6 ,
we present numerical results , including a test of our moment expansion in the fermi gas model .
section 7 contains a discussion and conclusions , including a recommendation for the application of these sum rules to data analysis .
we begin with the first - order born approximation for the inelastic scattering of ultrarelativistic @xmath50 electrons from nuclear targets .
initial and final electron four - momenta are denoted by @xmath51 and @xmath52 , and initial and final target four - momenta are denoted by @xmath53 and @xmath54 , respectively .
the differential cross - section can be written in factored form : @xmath55 where @xmath56 is the four - momentum transferred to the target in the laboratory frame by a single virtual photon , and @xmath57 is the fine - structure constant .
the lepton current tensor for an unpolarized electron beam is : @xmath58\bigl[{\bar u}_{s^\prime}({\bf k}^\prime ) \gamma^\nu u_s({\bf k})\bigr]^*,\label{bb}\ ] ] where the fermion spinors @xmath59 are given in ( [ aac ] ) .
all of the interesting target physics is contained in the nuclear response function : @xmath60 where @xmath13 and @xmath14 denote initial . ] and final nuclear many - body states , respectively .
the tensor @xmath61 may be decomposed into longitudinal and transverse response functions , with respect to the polarization of the virtual photon in a given frame : normally the target rest - frame .
( we will not discuss the transverse response further in this paper . )
if the z - axis is chosen along the three - momentum transfer * q * , i.e. , @xmath62 ( lorentz gauge ) and @xmath63 , can be written @xmath64 .
the longitudinal response function is then defined @xmath65 for theoretical reasons , it will be more convenient to work in the coulomb gauge , defining the coulomb response function @xmath66 where the second relation reflects this change of gauge .
it is the coulomb ( not longitudinal ) response function which appears in ( [ aa ] ) .
in general , the current operator @xmath67 includes contributions from both charged nucleons and charged mesons in the target ground state , as discussed in the introduction .
however , for the present work we ignore contributions to ( [ bc ] ) from meson currents .
this allows us to write the nucleon current operator in the one - body form : @xmath68 where @xmath18 is the dirac nucleon field operator in the schrdinger picture . in general , the field operators and therefore the current operator ( [ bf ] ) involve both nucleon and antinucleon degrees of freedom , as well as a sum over isospin projections . in the relativistic impulse approximation
, the field operators in ( [ bf ] ) refer to free particles .
it is then natural to expand these field operators in terms of free dirac wavefunctions , as in ( [ aab ] ) .
the @xmath69 vertex operator @xmath70 may then be specified in terms of its matrix elements between dirac plane - wave spinors : @xmath71 , therefore it is conventional to express @xmath70 in terms of dirac matrices : @xmath72 where @xmath73 is the nucleon anomalous magnetic moment and @xmath74 is the nucleon mass . in many - body calculations involving both protons and neutrons , ( [ bg ] )
includes an implicit sum over proton and neutron isospin projections . in general , the form factors @xmath75 and @xmath76 are scalar functions of @xmath77 , @xmath78 and @xmath79 . for scattering from free nucleons , the form factors @xmath75 and
@xmath76 can be shown to be functions only of the scalar variable @xmath80 .
these form factors are obtained from elastic @xmath2 scattering data , usually in terms of the more convenient electric and magnetic form factors : @xmath81 respectively , where @xmath49
. then @xmath75 and @xmath76 are given by @xmath82 with the normalization convention that @xmath83 and @xmath84 .
the continuation of these form factors for interacting nucleons is not unique , requiring dynamical information not contained in the impulse approximation .
a common assumption is to use the on - shell values @xmath85 and @xmath86 in constructing the vertex operator ( [ bg ] ) .
an equally reasonable assumption , which we adopt in this paper , is to evaluate _ only _ the electric and magnetic form factors at the off - shell variable @xmath87 , i.e. , the off - shell @xmath75 and @xmath76 are defined : @xmath88 where @xmath89 and @xmath90 . here
@xmath91 is the energy of a free nucleon with three - momentum @xmath92 .
the crucial feature of the off - shell prescription ( [ bj ] ) is that the photon energy @xmath93 enters @xmath70 in ( [ bg ] ) only through the form factors @xmath94 and @xmath95 .
consider the limit of point nucleons , for which the form factors @xmath85 and @xmath86 in ( [ bg ] ) become constants : then the plane - wave matrix elements @xmath96 depend only on the three - vectors * p * and * q*. for extended particles , both the @xmath97 s and the @xmath98 s depend on the scalar @xmath87 , and may also depend on another scalar , which we take to be @xmath99 .
( this choice is not unique , but is sensible given that * p * and * q * are the only other variables at the @xmath100 vertex in the impulse approximation . )
one may assume that _ either _ the @xmath97 s or the @xmath98 s , but not both sets , are functions of @xmath87 only ; we have made the second choice in ( [ bj ] ) .
this prescription is necessary to arrive at the particularly simple forms for the relativistic coulomb sum rule which follows .
we adopt the following parameterizations@xcite for the nucleon electric and magnetic form factors appearing in ( [ bj ] ) : @xmath101 where @xmath102 and @xmath103 are the proton and neutron anomalous magnetic moments , respectively . although the neutron electric form factor @xmath104 is not identically zero in ref .
@xcite , it is known to be small compared to @xmath21 . other nonzero forms for @xmath104 have also been given,@xcite and could be adopted for use in the present work . however , we shall make use of the fact that _ all _ form factors have @xmath21 as a common factor to simplify the form of the relativistic coulomb sum rule to follow .
we note , however , that this proportionality can not be exact for nonzero @xmath104 , since @xmath105 .
in this section , we consider the simplest nuclear target , one free nucleon , in the response function formalism of the previous section .
this allows us to illustrate directly some of the features of relativity which also occur in the general nuclear case , and suggest how they may be treated in the formulation of relativistic coulomb sum rules . in this case , the scattering is elastic on the target nucleon ; the differential cross - section in the laboratory frame , which may be obtained from ( [ ba ] ) , is the rosenbluth formula with form factors ( [ bh ] ) .
we are interested only in the coulomb scattering , which is expressed in terms of the coulomb response function @xmath106 . for a free nucleon , the coulomb response function ( [ be ] ) may be evaluated from ( [ bc ] ) and ( [ bf ] ) by specifying that initial and final nuclear states are simply dirac plane - wave states of momenta * p * and @xmath107 , respectively . using the nucleon field operator of ( [ aab ] )
, we find : @xmath108 where the 00-component of the nucleon current tensor is given by @xmath109\bigl[{\bar u}_{s^\prime}({\bf p}^\prime ) \gamma_0 u_s({\bf p})\bigr]^*,\label{cb}\ ] ] and the the energy factors in the denominator of ( [ ca ] ) reflect our convention that the fermion spinors ( [ aac ] ) are normalized to @xmath110 particles per unit volume .
the delta - function in ( [ ca ] ) guarantees energy conservation for the free nucleon states : @xmath111 .
notice that only nucleon , i.e. , not antinucleon , states appear in ( [ cb ] ) .
the energy required to excite a free particle of momentum * p * into a state with momentum @xmath112 satisfies @xmath113 , hence the process @xmath114 is restricted entirely to the spacelike region @xmath44 .
in contrast , the energy required to excite an @xmath1 pair from the vacuum satisfies @xmath115 , and hence is restricted entirely to the timelike region @xmath116 ) do not appear in the ( spacelike ) response function for a free nucleon
. we shall find it advantageous to use this separation of excitation spectra to explicitly remove the contribution of @xmath1 pairs from the relativistic sum rule .
( this issue was first raised by matsui.@xcite ) returning to the coulomb response function , we first give an explicit form for the 00-component of the nucleon current tensor ( [ cb ] ) . after substituting ( [ bg ] ) and ( [ bj ] ) into ( [ cb ] ) , a straightforward trace calculation yields @xmath117,\label{cc}\ ] ] where , of course , for a free nucleon @xmath118 and @xmath119 .
if we put the initial nucleon at rest , evaluation of ( [ cc ] ) at @xmath120 yields @xmath121 where we have used @xmath122 at @xmath120 .
thus stationary nucleons couple _ only _ through their electric form factor .
( a sensible feature of the off - shell prescription ( [ bj ] ) is that this reduction also occurs for interacting nucleons . )
the response function is obtained by substituting ( [ cd ] ) into ( [ ca ] ) .
the coulomb sum function is obtained by integrating @xmath123 over energy transfer @xmath93 , as in ( [ ae ] ) . for electron scattering from a free nucleon ,
it is natural to restrict the integral to include only spacelike ( @xmath44 ) photons , thus excluding @xmath1 pair production : @xmath124 where in the second relation we have made use of the nucleon energy - momentum relation at @xmath120 to express the result in a form which lends itself easily to interpretation .
the first factor @xmath125 has its origin in the energy denominator @xmath126 in ( [ ca ] ) , which in turn comes from the relativistic normalization of the fermion spinors ( [ aac ] ) .
the second factor @xmath127 arises from our choice to work in coulomb gauge , as can be seen by comparing with ( [ be ] ) . taken together ,
these factors represent the effect of relativistic recoil ( in coulomb gauge ) for a single nucleon . the first relation in ( [ ce ] )
shows that this recoil effect tends to unity in the nonrelativistic limit @xmath128 , and tends to the value @xmath129 in the ultrarelativistic limit @xmath130 .
( it is in this sense that coulomb gauge is a convenient choice , since the analog of ( [ ce ] ) in lorentz gauge tends to zero in the limit @xmath131 . )
matsui@xcite and dodang _
et al._@xcite have previously noticed this limiting value of @xmath129 for the spacelike coulomb sum , for a fermi gas of dirac ( @xmath132 ) nucleons .
the fermi gas result is easily obtained by inserting ( [ cc ] ) into ( [ ca ] ) and integrating over the fermi sphere @xmath133 with an appropriate factor to account for pauli blocking , as is done in ( [ fb ] ) .
it is easy to show that this quantity approaches @xmath47 ( for @xmath29 protons ) as @xmath134 . that this is a purely kinematic effect of relativistic nucleon recoil is clear from the single - nucleon example ( [ ce ] ) .
the high-@xmath37 behavior of ( [ ce ] ) is in contrast to that of the nrcsr ( [ ak ] ) , for which @xmath41 ( @xmath135 here ) as @xmath136 , suggesting that in the relativistic limit , the coulomb sum does _ not _ directly count the number of charged scatterers , as in the nonrelativistic case .
a simple modification of ( [ ce ] ) would , however , seem to remedy this for the case of a single nucleon : divide both sides by the kinematic recoil factor on the right - hand side , so that the large-@xmath37 limit is now unity , as would be expected for a sum rule . that this actually corresponds to a sum rule for the many - body case , with a suitable modification for the nucleon momentum distribution ( fermi motion ) in the target ,
is shown in the next two sections . a final note : for a free nucleon
, the delta - function of ( [ ca ] ) appears in all components of @xmath137 . to obtain the rosenbluth formula
, one must integrate ( [ ba ] ) over @xmath138 ( or @xmath139 ) at fixed electron scattering angle @xmath9 , rather than at fixed @xmath140 , as in the coulomb sum ( [ ce ] ) .
the constraint that @xmath93 and * q * are _ not _ independent variables , but are related by @xmath141 , then gives rise to the familiar recoil factor : @xmath142^{-1}\label{cf}\ ] ] which reduces to unity in the nonrelativistic limit @xmath143 , where @xmath10 is the electron beam energy .
thus it is well - known that not only form factors , but also nucleon recoil factors , modify the differential cross - section for electron scattering in the relativistic domain .
in this section , we derive a sum rule for a target of interacting nucleons , where the relativity of the nucleons is taken into account . following the example of a single nucleon in the previous section
, we find that if we restrict the sum to spacelike four - momentum transfers , which corresponds to @xmath0 scattering experiments , we are led to a `` nucleons - only '' approximation in which antinucleon degrees of freedom are ignored .
the new feature in the sum rule given here is the kinematic correction for relativistic recoil , which depends on the momentum of the nucleons in the target . in the following section ,
we show how this correction can be calculated in successive approximations in terms of moments of the nucleon momentum .
we begin with the coulomb response function in the rest - frame of a many - body nuclear system , which is given by @xmath144 of ( [ bc ] ) , evaluated at @xmath145 : @xmath146 where the coulomb transition matrix element is : @xmath147 these are the relativistic equivalents of ( [ ac ] ) and ( [ ad ] ) .
the @xmath100 vertex operator @xmath148 is given by ( [ bg ] ) and includes an implicit sum over isospin projections , and the field operator @xmath18 is given in ( [ aab ] ) and includes both nucleon and antinucleon degrees of freedom .
as discussed in section 2 , the nucleon form factors are assumed to be given by ( [ bj ] ) such that the only @xmath93 dependence in @xmath149 enters through the nucleon electric and magnetic form factors , @xmath94 and @xmath95 .
however , given the specific functional forms in ( [ bk])([bn ] ) , one can instead consider _ all _ the @xmath93 dependence in ( [ db ] ) to enter only through the common form factor @xmath21 .
dividing ( [ db ] ) by @xmath21 therefore removes _ all _ the @xmath93 dependence from the nuclear matrix element @xmath150 .
dividing ( [ da ] ) by @xmath151 then allows us to generate a _
non - energy - weighted _ sum rule , as follows .
we consider the coulomb sum function defined by @xmath152 where the integration is performed only over energies @xmath153 , corresponding to electron scattering experiments , as in ( [ ce ] ) .
this is unlike the nonrelativistic coulomb sum ( [ ae ] ) , in which the integration is over all @xmath154 in order to invoke closure . substituting ( [ da ] ) into ( [ dc ] ) and performing the integration
, we obtain : @xmath155 where the sum over all final states in ( [ da ] ) is now reduced to a sum over only those final states @xmath156 which are accessed by spacelike four - momentum transfers . for the case of a free nucleon target at rest , ( [ dd ] )
reduces to ( [ ce ] ) .
the nucleon field operator ( [ aab ] ) , and therefore the transition matrix elements ( [ db ] ) , involve both nucleon and antinucleon degrees of freedom .
however , because electron scattering experiments produce spacelike virtual photons exclusively , the resulting nuclear excitations are restricted to those whose excitation energies satisfy @xmath157 , as indicated in ( [ dd ] ) .
we have already seen in ( [ ce ] ) , for the case of a free nucleon , that restricting the energy transfer @xmath93 removes the contribution of @xmath1 pair production from the coulomb sum , which would enter if @xmath158 were included in ( [ dc ] ) .
similarly , for a noninteracting fermi gas , only excitations involving @xmath159 ( @xmath46 scattering ) enter into the spacelike sum .
this is because there are no antinucleons present in the renormalized nuclear ground state , thus the terms involving @xmath160 ( @xmath161 scattering ) and @xmath162 ( @xmath1 pair - annihilation ) vanish identically , and because the terms involving @xmath163 ( @xmath1 pair - creation ) are restricted entirely to the timelike region , since @xmath164 .
furthermore , the separation of nucleon and antinucleon degrees of freedom is complete in the fermi gas model : not only are antinucleon scattering and @xmath1 pair terms completely removed from the spacelike sum , but the @xmath114 contribution is guaranteed to lie completely in the spacelike region through the relation @xmath165 .
( this feature is unaffected by the inclusion of scalar and vector potentials like those in quantum hadrodynamics , for example , as noted by do dang _ et al._@xcite ) thus , in the fermi gas model , the spacelike nuclear response ( [ dd ] ) can be represented by effectively dropping all reference to antinucleon degrees of freedom at the level of the field operator ( [ aab ] ) . for an interacting nuclear system , however , the separation of nucleon and antinucleon degrees of freedom is no longer exact . in general , both @xmath46 and @xmath161 degrees of freedom
can contribute to the spacelike nuclear response , and some of the nucleons - only response @xmath166 , which is entirely spacelike in the fermi gas model , can be pushed into the timelike region by two - particle interactions . however
, we can assume that true @xmath1 pair production will still occur mostly for @xmath158 , and therefore can be neglected in the coulomb sum ( [ dd ] ) .
although tests of this assumption are outside the scope of this paper , we note that it is satisfied exactly for infinite nuclear matter in the rpa , i.e. , a spacelike photon excites only spacelike excitations even through the summation of an infinite number of rpa ring diagrams , and does not produce @xmath1 pairs .
antinucleon contributions , i.e. , those involving @xmath167 and @xmath162 , enter as relativistic effects in the nuclear target structure . in hartree calculations for finite nuclei ,
these terms make only small ( @xmath168 ) contributions to the relevant ( vector ) density,@xcite and shall also be neglected in this paper .
therefore , we adopt a `` nucleons - only '' approximation , in which antinucleon degrees of freedom are removed from discussion at the level of the field operator ( [ aab ] ) . in effect
, we are assuming that the _ dominant _ relativistic effect comes from single - nucleon recoil at high @xmath37 , and not from relativistic aspects of nuclear structure .
thus the initial and final nuclear states in ( [ dd ] ) will be treated as interacting many - nucleon states without @xmath161 components . substituting the field operators given by ( [ aab ] ) into the transition matrix element ( [ db ] )
, we have in our `` nucleons - only '' approximation : @xmath169 where the `` reduced '' spinor matrix element is defined : @xmath170 in ( [ de ] ) and ( [ df ] ) , nucleon spin states are represented by latin indices and nucleon isospin states are represented by greek indices .
the operator @xmath171 is diagonal in isospin space , therefore the transition matrix element ( [ de ] ) simply involves a sum over proton and neutron isospin projections .
the term `` reduced '' in ( [ df ] ) refers to the division by @xmath21 , following ( [ dc ] ) .
the matrix element ( [ df ] ) is then a function only of * p * and * q * , i.e. , not of @xmath93 , through our assumption of proportional form factors . substituting ( [ de ] ) into ( [ dd ] ) leads to : @xmath172 where we have subtracted the elastic term @xmath25 explicitly , in order to obtain a sum over the complete set of nuclear final states .
for notational simplicity , we assume a nondegenerate ground state @xmath13 .
equation ( [ dg ] ) is then the most general form for the coulomb sum function ( [ dc ] ) , given our nucleons - only approximation and our assumptions about nucleon form factors and their off - shell continuation . in order to obtain a sum rule from ( [ dg ] )
, we must sum over a complete set of final states @xmath14 and use closure .
the completeness of such a set necessarily includes excited nuclear states accessed by _ both _ spacelike ( @xmath44 ) and timelike ( @xmath158 ) photons , which can be expressed @xmath173 it is consistent with the nucleons - only approximation imposed in ( [ de ] ) to neglect the contribution of timelike states @xmath174 to the closure of the sum in ( [ dg ] ) , since these states predominantly involve antinucleons .
for a uniform fermi gas in the hartree - fock approximation , we have already seen that the separation between @xmath46 and @xmath161 degrees of freedom guarantees that the spacelike coulomb sum be completely exhausted in the nucleons - only approximation .
similarly , for an interacting fermi gas in the rpa , nucleons are excited _ only _ for @xmath44 , and @xmath1 pairs are produced _ only _ for @xmath158 , thus the same argument holds .
it is possible , however , that restricting the sum in ( [ dd ] ) to spacelike states @xmath175 does _ not _ in general exhaust the nucleons - only excitation spectrum , e.g. , finite size effects and two - body interactions beyond the rpa may push some of the nucleons - only response into the timelike region .
we are aware of no general argument which applies .
we will simply assume that the nucleons - only excitation spectrum _ is _ saturated by states @xmath176 , so that we can invoke closure in ( [ dg ] ) .
extension of the sum to include nucleonic excitations with @xmath158 could be done by theoretical means , as in schiavilla _
et al._,@xcite for example .
applying closure to the first term in ( [ dg ] ) removes the sum over final states , and the expression in curly brackets becomes : @xmath177 equation ( [ dg ] ) is now a _
sum rule _ , equating the sum ( [ dc ] ) to the ground state expectation values in ( [ di ] ) . to cast the sum rule in a more transparent form , we separate out the one- and two - body parts of the operator in the first term of ( [ di ] ) , using the momentum - space anticommutation relations for the nucleon creation and destruction operators : @xmath178 where the anticommutator of any two creation or destruction operators vanishes .
moving all creation operators to the left in the first term of ( [ di ] ) , we obtain : @xmath179 where we have extracted @xmath180 from the first matrix element , which is diagonal in spin projection . inserting ( [ dk ] ) into ( [ dg ] ) allows us to separate the sum rule expression into one - body , two - body and elastic amplitudes , in complete analogy to the nonrelativistic expression ( [ ah ] ) . here , however , we have used momentum operators , rather than local field operators , in order to eliminate the antinucleon components .
we separate the sum rule as follows : @xmath181 the first term gives the one - body contribution : @xmath182 where we have defined the momentum distribution function @xmath183 for isospin projection @xmath184 ( which we assume is independent of @xmath185 ) , and the relativistic recoil factor @xmath186 where @xmath187 is the nucleon tensor component ( [ cc ] ) . in the nonrelativistic limit @xmath188 , @xmath189 and @xmath190 , as in ( [ ah ] ) .
for relativistic momenta @xmath191 , however , we have @xmath192 , which represents the kinematic effect of relativistic nucleon recoil , as in ( [ ce ] ) .
if the fermi momentum in ( [ dm ] ) could be neglected entirely , i.e. , if we were to set @xmath120 in ( [ dn ] ) , then @xmath193 would be given by ( [ ce ] ) .
the second term of ( [ dl ] ) is given by @xmath194 this expression contains the full two - body probability in momentum space , in analogy with the second term of ( [ ah ] ) , but is weighted by the kinematic factor @xmath195 .
to separate out the correlation part of @xmath196 as in ( [ ai ] ) , we first evaluate ( [ do ] ) for an uncorrelated target ground state , for which we obtain @xmath197 where @xmath198 for protons and neutrons , respectively .
( in the uncorrelated case , there is no contribution in ( [ dp ] ) from @xmath199 . )
we may then define the correlation function @xmath200 to give a measure of true two - body correlations , and rewrite ( [ dl ] ) @xmath201 in analogy with ( [ ak ] ) .
this result will be referred to as the relativistic coulomb sum rule ( rcsr ) .
the uncorrelated term @xmath202 is related to the square of the elastic target form factor , with the difference that the proton and neutron contributions are added _ incoherently _ in ( [ dp ] ) .
if , however , the contribution of neutrons to ( [ dp ] ) can be neglected , then @xmath203 completing the analogy of ( [ dr ] ) to ( [ ak ] ) .
for example , if the contribution of the magnetic ( @xmath95 ) terms is zero , as discussed by friar@xcite for a spin - saturated target , then with @xmath204 , as assumed in ( [ bn ] ) , ( [ ds ] ) is exact . for @xmath205 , ( [ ds ]
) can be suitably modified under reasonable assumptions , e.g. , similar neutron and proton distributions in ( [ dp ] ) .
the rcsr given here is exact in its treatment of nucleon fermi motion and in principle allows the identification of nucleon - nucleon correlations to arbitrarily high @xmath37 , given our assumptions about antinucleon and meson degrees of freedom , nucleon excitations , and off - shell continuation of nucleon form factors . from ( [ dr ] )
, the extraction of the correlation function @xmath206 from the calculated coulomb sum ( [ dc ] ) requires the removal of the nuclear elastic form factor , as just discussed , as well as a reliable evaluation of the one - body term ( [ dm ] ) .
for the latter , this form of the rcsr presupposes that one has in hand the nucleon momentum distribution @xmath207 .
the momentum distribution @xmath207 is not generally available directly from experimental data . in order to use the rcsr as given in the previous section , therefore , one would have to rely on a theoretical model to calculate the one - body term @xmath208 .
an alternative approach , which we develop in this section , is to expand the one - body term ( [ dm ] ) in moments of the nucleon momentum , i.e. , in averages of powers of the nucleon momentum , weighted by the distribution @xmath209 . the one - body term ( [ dm ] ) is then replaced by a sum in terms of these moments , the first few of which may be known .
this procedure leads to a series of approximate sum rules , each of which depends on higher momentum moments . in each case , it is possible to identify appropriate modifications to the definition of the relativistic coulomb sum ( [ dc ] ) and arrive at a series of modified sum rules , which share certain features with the nrcsr .
of course , such a procedure is useful _ only _ if the moment expansion converges rapidly , and if an adequate number of moments can be obtained reliably . in the next section
, we will demonstrate that this expansion indeed converges quickly for a uniform fermi gas , and argue in the following discussion that this is likely to be the case in general .
the lowest - order momentum moment expansion is obtained by setting @xmath120 in ( [ dn ] ) : @xmath210 which is just the factor that appeared in ( [ ce ] ) for a single proton at rest .
this is equivalent to approximating the reduced matrix element ( [ df ] ) by its value at @xmath120 : @xmath211 since we evaluate at @xmath120 , ( [ ea ] ) and ( [ eb ] ) do not involve magnetic terms .
the absence of neutron electric contributions is a result of our approximation ( [ bn ] ) . inserting ( [ ea ] ) into ( [ dm ] ) ,
the one - body term becomes @xmath212.\label{ec}\ ] ] similarly , inserting ( [ eb ] ) into ( [ do ] ) and ( [ dp ] ) , the two - body correlation term ( [ dq ] ) becomes @xmath213 \sum_{{\bf pp}^\prime}\sum_{ss^\prime}\biggl \lbrace\langle i|a^\dagger_{{\bf p+q}s } a^\dagger_{{\bf p}^\prime s^\prime } a_{{\bf p}^\prime+{\bf q}s^\prime } a_{{\bf p}s}|i\rangle\nonumber\\ & & \nonumber\\ & & \qquad\qquad-\biggl({z-1\over z}\biggr ) \langle i|a^\dagger_{{\bf p}^\prime s^\prime } a_{{\bf p}^\prime+{\bf q}s^\prime}|i\rangle \langle i|a^\dagger_{{\bf p+q}s}a_{{\bf p}s } |i\rangle\biggr\rbrace , \label{ed}\end{aligned}\ ] ] where the proton isospin label has been suppressed for clarity .
since both @xmath214 and @xmath206 now appear with the _ same _ overall kinematic factor , i.e. , that which appeared in ( [ ce ] ) , which is a function only of * q * , it is possible to modify the definition of the coulomb sum ( [ dc ] ) and obtain a more conventional form for the sum rule .
we define the modified coulomb sum function : @xmath215 where the lowest - order recoil correction factor is @xmath216 and @xmath217 is to be extracted from experimental data , using ( [ aa ] ) .
we then find the _ approximate _ sum rule : @xmath218.\label{eg}\ ] ] this result would be exact for a target of free , stationary protons .
the two - proton correlation function in ( [ eg ] ) is given by @xmath219 in close analogy with ( [ ai ] ) .
expression ( [ eg ] ) will be referred to as rcsr - i .
like the nrcsr given in ( [ ak ] ) , rcsr - i involves a one - body term @xmath29 which simply counts the total number of charged scatterers , and a two - body term which contains information on nucleon - nucleon correlations .
the function @xmath220 , however , is _ not _ the fourier transform of the usual spatial correlation function , as defined in terms of local nucleon field operators , i.e. , simply inserting the dirac nucleon field operators of ( [ aab ] ) into ( [ ai ] ) will _ not _ yield ( [ eh ] ) , because of the antinucleon degrees of freedom which are eliminated from ( [ eh ] ) , and the kinematic factors related to relativistic nucleon recoil which result from that elimination .
however , if the _ nonrelativistic _ result ( [ ai ] ) is expressed in momentum space , it is exactly of the form ( [ eh ] ) .
thus this version of the rcsr , although based on a rather strong assumption about fermi motion in the target , is closest in form to the familiar nrcsr .
the procedure which led to rcsr - i can be extended to include higher moments of the nucleon momentum . in this subsection ,
we keep terms only through second order .
we first expand the relativistic recoil factor : @xmath221 for isospin projection @xmath184 .
since the recoil factor ( [ dn ] ) is a function only of * p * and * q * , i.e. , not @xmath93 , the expansion coefficients @xmath222 are functions only of * q*. explicit forms for the @xmath222 are given in appendix b , where we find @xmath223 as a consequence of approximation ( [ bn ] ) . substituting ( [ ei ] ) into the one - body term ( [ dm ] ) leads to @xmath224 + n_n({\bf p})\bigl[\bigl(r_2^n+{{\bf q}^2\over3 } r_3^n\bigr){\bf p}^2\bigr]\biggr\rbrace,\label{ej}\ ] ] where we have used the spherical symmetry of @xmath207 to eliminate the @xmath225 term and integrate the angles : @xmath226 .
we can now identify the average squared nucleon momentum , i.e. , the second momentum moment : @xmath227,\label{ek}\ ] ] where @xmath228 for protons and neutrons , respectively .
the one - body term ( [ ej ] ) can then be expressed @xmath229+n\biggl [ \bigl(r_2^n+{{\bf q}^2\over3}r_3^n\bigr ) \langle{\bf p}^2\rangle_n\biggr].\label{el}\ ] ] we thus have an approximate expression for the one - body term ( [ dm ] ) which depends only on the two parameters @xmath230 , and reduces to ( [ ec ] ) if @xmath231 .
we can now follow a procedure similar to that of ( [ ee ] ) , by defining the modified coulomb sum function : @xmath232 where the second - order relativistic recoil factor is @xmath233 .
\label{en}\ ] ] we thus find the second - order _ approximate _ sum rule : @xmath234,\label{eo}\ ] ] in analogy with ( [ eg ] ) .
the second - order nucleon - nucleon correlation function is given by @xmath235 where @xmath206 is defined in ( [ dq ] ) .
expression ( [ eo ] ) will be referred to as rcsr - ii .
explicit expressions for the two - body term @xmath208 and the correlation function @xmath236 could be given to the same order of approximation made in ( [ ei ] ) , but would require the expansion of @xmath237 to second - order in * p*. since we shall not use these explicit forms , we do not pursue this here . although rcsr - ii enjoys a less direct correspondence with the nrcsr than does rcsr - i , it is more accurate in its treatment of fermi motion . more importantly , as we will see in the next section ,
the higher - order terms which are included in rcsr - ii introduce effects from proton and neutron anomalous magnetic moments which are appreciable .
in this section , we present numerical results for our relativistic coulomb sum rules .
we first consider the recoil correction factors @xmath238 and @xmath239 , which appear in rcsr - i and rcsr - ii , respectively , and find that terms @xmath240 in our momentum expansion bring in anomalous magnetic moment effects which are appreciable .
we then illustrate in the fermi gas model the accuracy with which rcsr - i and rcsr - ii allow the evaluation of the one - body term .
ideally , one would like to extract @xmath241 directly from experimental data .
since that is not always possible , however , one may have to rely upon estimates based on theoretical models .
the simplest such model is a uniform fermi gas , where @xmath242 . in that case
, we have : @xmath243 where @xmath244 is the fermi momentum for isospin projection @xmath184 . in this example , we assume @xmath245 and take @xmath246 to match the average density in the interior of finite nuclei .
this gives @xmath247 .
for @xmath248 , it is reasonable to assume that the fermi momenta for protons and neutrons scale according to @xmath249 .
figure 1 shows the recoil correction factors @xmath238 and @xmath250 , given in ( [ ef ] ) and ( [ en ] ) , as functions of the three - momentum transfer * q*. the dotted curve is @xmath238 , and is indistinguishable ( on the scale of this plot ) from the factor given by deforest : @xmath251 , suitably averaged over angle . the dot - dashed curve is @xmath239 with @xmath252 , i.e. , for dirac nucleons .
both curves tend to the value @xmath129 as @xmath136 .
their difference is a measure of the importance of second - order terms in the moment expansion when anomalous magnetic moments are ignored , and is at most @xmath253 anywhere in the range @xmath254 .
the dashed curve is @xmath239 with @xmath255 and @xmath256 , and shows that the proton anomalous magnetic moment leads to a noticeable enhancement in the recoil factor : @xmath257 at @xmath258 , and reaching @xmath259 as @xmath134 , when compared to @xmath238 .
the solid curve is @xmath239 with @xmath255 and @xmath260 , and shows a similar enhancement due to the neutron anomalous magnetic moment when @xmath261 : @xmath262 at @xmath258 , and reaching @xmath263 as @xmath136 , when compared to @xmath238 .
it is clear from figure 1 that @xmath239 is significantly different from @xmath238 _ only _ when anomalous magnetic moments are included .
this can be understood by considering the coefficient of @xmath264 in ( [ en ] ) , i.e. , the quantity @xmath265 , where @xmath266 and @xmath267 are given by ( [ abc ] ) and ( [ abd ] ) , respectively .
we find that the coefficients of @xmath268 and @xmath269 in that quantity are of roughly the same magnitude and have opposite sign .
thus the second - order proton contribution to ( [ en ] ) is nearly zero when @xmath270 , and @xmath271 .
however , when @xmath255 , we have @xmath272 . in that case ,
the coefficients of @xmath268 and @xmath273 no longer cancel , and @xmath239 is significantly enhanced .
similarly , for neutrons , both @xmath274 and @xmath275 are positive and proportional to @xmath276 .
thus when @xmath103 , we have @xmath277 and @xmath239 is further enhanced . from the specific functional form of @xmath278 , as given in ( [ cd ] ) , we note that terms in ( [ ei ] ) which are of higher order in @xmath279 will _ not _ involve higher powers of the anomalous terms @xmath269 and @xmath276 .
therefore , we expect that the moment expansion will converge _ if _ the moments themselves converge
. we now present a numerical test of our moment expansion in the simple system which includes fermi motion : a uniform fermi gas .
the spacelike coulomb response function for this system has been studied by several authors.[1316 , 24 ] substituting ( [ cc ] ) into ( [ ca ] ) , and integrating over the fermi sphere @xmath133 with an appropriate factor to account for pauli blocking , we obtain : @xmath280 { l_{00}^\sigma({\bf p},{\bf q})\over4e_{{\bf p+q}}e_{\bf p } } \delta(\omega - e_{{\bf p+q}}+e_{\bf p}),\label{fb}\ ] ] where the occupancy function @xmath281 . inserting ( [ fb ] ) into ( [ dc ] )
gives the relativistic coulomb sum : @xmath282 r_\sigma({\bf p},{\bf q}),\label{fc}\ ] ] where we have used definition ( [ dn ] ) .
comparison with ( [ dm ] ) makes the identification of one- and two - body contributions trivial for this model .
furthermore , the two - body term involves only pauli correlations , which vanish for @xmath283 , and @xmath284 for @xmath285 .
we now demonstrate in several cases the accuracy to which @xmath208 of ( [ fc ] ) is represented by rcsr - i and rcsr - ii .
we take ( [ fc ] ) to represent the experimental coulomb sum obtained from ( [ dc ] ) and ( [ aa ] ) , and divide out the recoil factors @xmath238 and @xmath239 , as in ( [ ee ] ) and ( [ em ] ) . figure 2 shows several versions of the coulomb sum ( divided by @xmath29 ) for a free dirac proton gas , i.e. , @xmath252 , versus three - momentum * q * in units of the fermi momentum @xmath286 .
the dot - dashed curve is the unmodified coulomb sum @xmath287 , given by ( [ fc ] ) ; this is exactly the result given by matsui,@xcite and tends to the value @xmath129 as @xmath136 .
the dashed curve is @xmath288 , as defined in ( [ ee ] ) , and the solid curve is @xmath289 , as defined in ( [ em ] ) . here , as in figure 1
, the result given by deforest is indistinguishable on the scale of this plot from rcsr - i . both the dashed ( rcsr - i ) and solid ( rcsr - ii ) curves approach unity quickly as @xmath134 , although rcsr - i overshoots slightly ( @xmath290 ) in the @xmath291 gev region . the convergence to unity as @xmath292
is guaranteed in the dirac case , since both the unmodified coulomb sum @xmath287 and the recoil factors @xmath238 and @xmath250 approach the value @xmath129 , for @xmath252 .
figure 3 shows the same quantities for a free proton gas , with @xmath293 . in this case , @xmath294 as @xmath136 , due to the proton anomalous magnetic moment .
in contrast to the dirac case , rcsr - i _ fails _ to level off at @xmath295 , but rather @xmath296 as @xmath136 .
however , since @xmath239 accounts for the anomalous moment , rcsr - ii levels off immediately at @xmath295 and @xmath297 as @xmath136 .
the difference from unity is a measure of the importance of higher - order terms , e.g. , @xmath298 , which have been neglected in our expansion .
figure 4 shows the same results for a symmetric ( @xmath261 ) fermi gas . in this case
, neutron terms further enhance the unmodified sum : @xmath299 and @xmath300 as @xmath301 .
however , when the neutron anomalous moment is incorporated into the recoil factor @xmath239 , we again have @xmath302 as @xmath136 .
we have shown that it is possible to estimate reliably the one - body term in the coulomb sum if one accounts correctly for nucleon recoil . due to the simplicity of the fermi gas model
, however , the numerical study given here does _ not _ test the validity of our nucleons - only approximation , since the separation of @xmath46 and @xmath161 degrees of freedom is exact in this case , nor does it test our assumptions about the off - shell continuation of nucleon form factors , since @xmath303 in this case .
this study does , however , give an indication of the importance of anomalous magnetic moment effects , and how accurate these approximate sum rules _ can _ be for the extraction of @xmath304 correlation information from electron scattering data , when the momentum distribution of the nuclear target is like that of a fermi gas .
the main result of this paper is the derivation of a coulomb sum rule applicable to the analysis of inelastic electron scattering experiments on nuclei at high three - momentum transfers .
the restriction to spacelike four - momenta ( @xmath44 ) , appropriate for electron scattering , approximately eliminates @xmath1 pair - production from the response function .
we then ignore the small contributions which arise due to antinucleons in the target ground state ; this is our `` nucleons - only '' approximation .
the sum rule is given in three forms .
the most general form of the relativistic coulomb sum rule ( rcsr ) appears in ( [ dr ] ) . as for the well - known non - relativistic case ( nrcsr )
, the rcsr can be decomposed into a one - body term ( @xmath214 ) and a two - body correlation term ( @xmath206 ) after removing the elastic term .
the correlation term is of considerable interest , but is expected to be dominated by the one - body term for all but the smallest three - momentum transfers . in the nrcsr ,
the one - body term is simply @xmath29 , the number of protons . in the rcsr , however , this term is a function of * q * and includes a kinematic factor @xmath305 , as shown in ( [ dm ] ) .
this factor can be interpreted as the relativistic effect of nucleon recoil , and becomes appreciable for @xmath45 . in general
, the evaluation of ( [ dm ] ) requires knowledge of the full momentum distribution function @xmath207 , which is not accurately known from experiments .
this then limits the direct applicability of the rcsr , and in particular , the extraction of correlation information from the analysis .
we therefore develop an expansion of the one - body term in moments of the nucleon momentum .
this leads to approximate sum rules , which require only _ partial _ knowledge of the nucleon momentum distribution @xmath209 , i.e. , the first few moments , assuming that the moment expansion is rapidly convergent .
the simplest case is rcsr - i , given in ( [ eg ] ) , which involves only the zeroth moment .
this version of the sum rule requires no information about @xmath207 , and therefore has the most direct correspondence with the familiar nrcsr .
although rcsr - i is surprisingly accurate ( @xmath290 ) in approximating the one - body term for a uniform system of dirac protons , it neglects the contributions from the anomalous magnetic moments , which turn out to be non - negligible ( @xmath306 for @xmath307 and @xmath245 ) .
therefore , we keep the second moment contribution to obtain rcsr - ii , given in ( [ eo ] ) , which allows the evaluation of the one - body term to within @xmath168 for a uniform fermi gas .
this version of the sum rule requires knowledge of only two parameters , namely @xmath241 for protons and neutrons , which are approximately known from experimental and theoretical information ; we have used the fermi gas model for our estimate .
this study shows that the uncertainty in @xmath241 is likely to introduce much less error into the analysis than would ignoring the anomalous magnetic moments altogether . the excellent convergence of the moment expansion , shown here to second - order for a uniform fermi gas , is not expected to be very different for correlated systems ( see , for example , donnelly _
furthermore , the extension of this approach to include higher moments , if necessary , is straightforward .
we therfore conclude that rcsr - ii , which includes contributions from anomalous magnetic moments , is the most efficient and reliable method for evaluating the sum rule for @xmath0 data and separating the two - body information from the dominant one - body term .
we now discuss the relation of our work to that of earlier authors .
deforests@xcite approach to a relativistic sum rule is similar to that of rcsr - i , both formally and numerically .
the main difference is , in the language of section 5 , his replacement of @xmath308 in ( [ ea ] ) with the lorentz invariant factor @xmath309 .
however , a careful inspection of the recoil factor ( [ dn ] ) for a nucleon with arbitrary * p * shows that @xmath310 is _ not _ a lorentz scalar , and that anomalous magnetic moment effects _ must _ enter when @xmath311 . hence there is , in fact , no formal justification for writing the recoil correction at @xmath120 in scalar form .
it is interesting that the factor @xmath312 is numerically nearly identical ( for all @xmath313 ) to our recoil factor @xmath314 .
however , as we have seen in figure 1 , the correction due to fermi motion , which is large due to the effects of anomalous magnetic moments , must be included for a reliable evaluation of the one - body term .
for example , for @xmath315 , the factor given by deforest
_ underestimates _ relativistic recoil corrections by @xmath316 , compared to rcsr - ii ; therefore , division of the experimental data ( as done in ref .
@xcite , for example ) by deforest s factor in the calculation of the experimental coulomb sum will lead to an _ overenhancement _ of the sum by @xmath317 .
the goal of donnelly _
et al._@xcite is somewhat different : to find a factor @xmath318 which would replace @xmath319 in the definition of the spacelike coulomb sum function , such that the sum will reach @xmath29 in the high-@xmath37 limit .
although their approach can , in principle , achieve that goal to essentially any order of accuracy , it necessarily leads to a correction factor @xmath318 which depends explicitly on the excitation energy @xmath93 . however , unless the @xmath93-dependence can be extracted from the coulomb response function _ before
_ integration over @xmath93 , one will not arrive at a _ non - energy - weighted _ sum rule .
in particular , the approach taken by donnelly _
_ will necessarily involve energy - weighted sum rules , which differ formally from the nrcsr , and the rcsr we have derived , in that dynamical information becomes mixed with the correlation information .
a crucial step in the derivation of our sum rule is the extraction of the energy dependence from the response function _ before _
integration over @xmath93 . to do this , we needed two assumptions : first , that the off - shell form factors could be obtained from the on - shell form factors @xmath94 and @xmath95 as in ( [ bj ] ) .
although plausible , this requires theoretical justification , i.e. , a theory of the electromagnetic structure of nucleons .
the second assumption is that the form factors are all proportional , as in ( [ bk])([bn ] ) .
this is probably sufficiently accurate , except possibly for taking @xmath204 , as noted earlier .
a better approximation would be to take @xmath320 , using data from larger @xmath87 , although this is incorrect near @xmath321 .
the results of this work do have some bearing on the question of the saturation of the nrcsr , which we mentioned in section 1 .
the coulomb sum function usually extracted from @xmath0 experiments is that of ( [ ae ] ) , but with a finite upper limit on @xmath93 ; since the theoretical limit at fixed @xmath37 is @xmath44 , this is essentially identical to ( [ dc ] ) . for the relativistic sum rules given in section 5
, one must divide the sum by the relativistic recoil factor @xmath322 , i.e. , either @xmath238 or @xmath239 , to obtain a form which should `` saturate '' at @xmath29 in the limit that @xmath136 .
since @xmath323 , this correction will _ enhance _ the experimental sum , e.g. , as was shown in figures 24 . at the energies of interest ,
however , this is only a small effect , e.g. , at @xmath324 we have @xmath325 and @xmath326 .
it is interesting to note that these factors grow with increasing @xmath327 , and therefore do show _ in part _ the increased suppression seen in larger nuclei .
of course , many other effects may also contribute to the observed suppression .
although we have worked within a `` nucleons - only '' approximation , in principle , @xmath328 processes are easily included in this formalism ( since the matrix element of @xmath70 between @xmath161 states is the same as that between @xmath46 states ) .
however , we do not pursue this here because other @xmath161 contributions , e.g. , @xmath1 pair production , which may also enter the spacelike response , are likely to be of the same order . in order to understand fully the contribution of antinucleons to electron scattering
, it will be necessary to study the spacelike response function in interacting nuclear models , and devise an appropriate renormalization scheme to obtain finite results .
two issues which are not treated in this paper must also be addressed if the relativistic coulomb sum rule given here is to be used for the extraction of nucleon - nucleon correlation information : namely , the contribution of meson exchange currents , and internal excitations of the nucleon .
the first can be thought of as generating two - body ( or more ) terms in the nuclear current operator @xmath67 in ( [ bf ] ) , as done , for example in ref .
@xcite , where meson current contributions are actually calculated in a model . from the point of view of the sum rule , these many - body contributions can be grouped with the two - body term @xmath329 of ( [ do ] ) , to be determined _ experimentally _ as what is left over when @xmath214 is subtracted from the coulomb sum .
the second issue can be thought of , to a first approximation , as the problem of removing the `` background '' of @xmath330 excitations from the @xmath0 response , to leave the `` nucleons - only '' response , to which our relativistic coulomb sum rules apply .
this could be done theoretically by calculating the @xmath331 background , e.g. , in a noninteracting - nucleus model .
alternatively , some partially exclusive experimental information could help : for example , @xmath332 data in the quasi - free peak region can be used to normalize the knockout of protons ( as opposed to @xmath331 ) , although other process ( e.g. , @xmath333 ) will also contribute to this process .
these two problems , and the problem of antinucleons , are subjects for future investigation . 0.2 true in * acknowledgements * 0.2 true in this research was supported in part by the u. s. department of energy under grant no .
de - fg02 - 88er40425 with the university of rochester .
we would also like to thank the high energy physics group for the use of their vax computer .
the dirac equation for free point nucleons represented by the wavefunction @xmath334 can be written @xmath335 where @xmath336 . for a static , spatially uniform system , the solutions to ( [ aaa ] )
can be expressed as momentum eigenstates with momentum * p*. the energy eigenvalues for positive and negative energy solutions are @xmath337 , where @xmath338 . from the complete set of solutions to ( [ aaa ] ) , it is possible to construct a local field operator in terms of nucleon and antinucleon degrees of freedom : @xmath339,\label{aab}\ ] ] where @xmath340 represents the normalization volume .
the nucleon and antinucleon spinors are defined @xmath341\qquad v_s({\bf p})=\sqrt{e_{\bf p}+m } \biggl[{{{\bf \sigma\cdot p}\over e_{\bf p}+m}\chi_s\atop\chi_s}\biggr ] \label{aac}\ ] ] respectively , and have been normalized to @xmath342 particles per unit volume . here
@xmath343 is the usual two - dimensional pauli spinor , and @xmath344 and @xmath345 are destruction and creation operators for nucleon and antinucleons , respectively .
it is possible to generalize this formalism to include isospin degrees of freedom .
we simply define @xmath346 , where @xmath347 is a two dimensional spinor corresponding to isospin projection @xmath184 , and include a sum over @xmath184 in ( [ aab ] ) .
the expansion coefficients appearing in ( [ ei ] ) are : expressions ( [ abe ] ) and ( [ abf ] ) can be obtained from ( [ abc ] ) and ( [ abd ] ) , respectively , by keeping only terms proportional to @xmath273 and letting @xmath355 .
the corresponding expressions for dirac nucleons are obtained by setting @xmath252 .
99 k. w. mcvoy and l. van hove , phys .
rev . * 125 * , 1034 ( 1962 ) .
meziani _ et al . _ ,
* a446 * , 113 ( 1985 ) .
r. altemus _ et al .
_ , phys .
lett . * 44 * , 965 ( 1980 ) .
p. barreau _ et al .
_ , nucl .
* a358 * , 287 ( 1981 ) .
p. barreau _ et al .
_ , nucl .
* a402 * , 515 ( 1983 ) . m. deady _ et al . _ , phys . rev .
c * 28 * , 631 ( 1983 ) .
meziani _ et al .
_ , phys .
lett . * 52 * , 2130 ( 1984 ) .
m. deady _ et al .
_ , phys .
c * 33 * , 1897 ( 1986 ) . v. r. pandharipande , nucl .
phys . * a497 * , 43c ( 1989 ) .
o. benhar , v. r. pandharipande and s. c. pieper , rev .
* 65 * , 817 ( 1993 ) .
r. schiavilla , r. b. wiringa and j. carlson , phys .
lett.,*70 * , 3856 ( 1993 ) .
j. d. walecka , nucl .
phys . * a399 * , 387 ( 1983 ) . t. matsui , phys .
lett . b * 132 * , 260 ( 1983 ) .
g. do dang , m. lhuillier , nguyen van giai and j. w. van orden , phys .
c * 35 * , 1637 ( 1987 ) . c. j. horowitz , phys .
b * 208 * , 8 ( 1988 ) . t. w. donnelly , e.
l. kronenberg and j. w. van orden , nucl .
a494 * , 365 ( 1989 ) . t. de forest , jr .
* a414 * , 347 ( 1984 ) .
meziani _ et al .
_ , phys .
lett . * 69 * , 41 ( 1992 ) .
m. a. preston and r. k. bhaduri , _ the structure of the nucleus _ ( addison - wesley , reading , mass . , 1975 ) .
g. hohler _ et al .
_ , nucl .
* b114 * , 505 ( 1976 ) .
r. j. perry , phys .
b * 182 * , 269 ( 1986 ) .
r. schiavilla , v. r. pandharipande and a. fabrocini , phys .
c * 40 * , 1484 ( 1989 ) .
j. l. friar , ann .
phys . * 81 * , 332 ( 1973 ) .
g. do dang and pham van thieu , phys .
c * 28 * , 1845 ( 1983 ) . | a coulomb sum rule is derived for the response of nuclei to @xmath0 scattering with large three - momentum transfers . unlike the nonrelativistic formulation , the relativistic coulomb sum is restricted to spacelike four - momenta for the most direct connection with experiments ; an immediate consequence is that excitations involving antinucleons , e.g. , @xmath1 pair production , are approximately eliminated from the sum rule .
relativistic recoil and fermi motion of target nucleons are correctly incorporated .
the sum rule decomposes into one- and two - body parts , with correlation information in the second .
the one - body part requires information on the nucleon momentum distribution function , which is incorporated by a moment expansion method .
the sum rule given through the second moment ( rcsr - ii ) is tested in the fermi gas model , and is shown to be sufficiently accurate for applications to data . cut here -
@=11
@=12 0.5 true in pacs number(s ) : 25.30.-c , 25.30.fj , 24.10.-i , 11.50.li 0.5 true in submitted to * physical review c * |
with the recent discovery of several extra - solar planets ( and brown dwarfs ) around solar type stars ( e.g. mayor & queloz 1995 ; buter & marcy 1996 ; basri , marcy & graham 1996 ) the question of the potential existence of extraterrestrial `` intelligent '' civilizations has become more intriguing than ever .
while this topic has been the subject of extensive speculations and many ill - defined ( often by necessity ) probability estimates , at least one study ( carter 1983 ) has examined it from a more global , statistical perspective .
that study concluded , on the basis of the near equality between the timescale of biological evolution on earth , @xmath0 , and the lifetime of the sun , @xmath1 , that extraterrestrial civilizations are exceedingly rare , even if conditions favorable for the development of life are relatively common .
the conclusion on the rarity of extraterrestrial intelligent civilizations ( carter 1983 ; and see also barrow & tipler 1986 ) was based on one crucial _
assumption _ and one _ observation_. the _ assumption _ is that the timescale of biological evolution on a given planet , @xmath0 , and the lifetime of the central star , @xmath2 , are a priori entirely independent quantities . put differently
, this assumes that intelligent life forms at some random time with respect to the main sequence lifetime of the star .
observation _ is that in the earth - sun system @xmath3 ( to within a factor 2 ; for definiteness i will take from now on @xmath0 to represent the timescale for the appearance of land life ) . for completeness
, i will reproduce here the argument briefly . _
if _ @xmath0 and @xmath2 are indeed independent quantities , then most probably either @xmath4 or @xmath5 ( the set of @xmath3 is of very small measure for two independent quantities ) .
if , however , @xmath6 _ generally _ , then it is very difficult to understand why in the first system found to exhibit an intelligent civilization ( the earth - sun system ) , it was found that @xmath3 .
if , on the other hand , _ generally _ @xmath7 , then it is clear that the first system found to contain an intelligent civilization is likely to have @xmath8 ( since for @xmath7 a civilization would not have developed ) .
thus , according to this argument , one has to conclude that typically @xmath9 , namely , that _ generally _ intelligent civilizations will not develop , and that the earth is the extremely rare exception .
what i intend first to show in the present work is that this conclusion is at best premature , by demonstrating that not only that @xmath0 and @xmath2 may not be independent , but also that @xmath10 may in fact be the _ most probable _ value for this ratio .
this is done in 2 .
i use the recently determined cosmic star formation history to estimate the most likely time for intelligent civilizations to emerge in the universe .
superficially it appears that @xmath0 ( which is determined mainly by biochemical reactions and evolution of species ) and @xmath2 ( which is determined by nuclear burning reactions ) should be independent .
it suffices to note , however , that light energy ( from the central star ) exceeds by 23 orders of magnitude all other sources of energy that can drive chemical evolution in the prebiotic environment ( e.g. deamer 1997 ) , to realize that @xmath0 may in fact depend on @xmath2 ( note that the statement that intelligent life will not develop for @xmath7 also constitutes a qualitative dependence ) .
below i identify a specific physical process that can , in principle at least , relate the two timescales .
first i would like however to point out the following _ general _ property .
imagine that we find that the ratio @xmath11 can be described by some function of the form @xmath12 , and that we further find that the function @xmath13 is monotonically increasing ( at least for the narrow range of values of @xmath2 corresponding to stars which allow the development of life , see below ) .
this situation is shown schematically in fig . 1 .
in such a case , since for a salpeter initial mass function ( salpeter 1955 ) we have that the distribution of stellar lifetimes behaves like @xmath14 ( since for main sequence stars , @xmath15 e.g.allen 1973 ) , it is immediately clear from fig . 1 , that it is _ most probable _ that in the first place we encounter an intelligent civilization , we will find that @xmath10 ( since the number of stars increases as we move to the right in the diagram ) . therefore , if we can show that some processes are likely to produce a monotonically increasing ( @xmath2 ; @xmath11 ) relation , then the fact that @xmath3 in the earth - sun system will find a natural explanation , and it will not have any implications for the frequency of extraterrestrial civilizations .
i should note though that if the breadth of the ` band ' in fig .
1 becomes extremely large , this is essentially equivalent to no @xmath0@xmath2 relation and carter s argument is recovered .
i will now give a simple example of how a @xmath0@xmath2 relation may arise .
i should emphasize that this is not meant to be understood as a realistic model , but merely to demonstrate that such a relation _
could _ exist .
nucleic acid absorption of uv radiation peaks between @xmath16 and @xmath17 and that of proteins between @xmath17 and @xmath18(e.g .
davidson 1960 ; sagan 1961 ; caspersson 1950 ) .
absorption in these bands is highly lethal to all known forms of cell activity ( e.g. berkner 1952 ) .
of all the potential constituents of a planet s atmosphere only o@xmath19 absorbs efficiently in the 20003000 range ( e.g. watanabe , zelikoff & inn 1958 ) .
it has in fact been suggested that the appearance of land life has to await the build - up of a sufficient layer of protective ozone ( berkner & marshall 1965 ; hart 1978 ) .
thus , it is important to understand the origin and evolution of oxygen in planetary atmospheres . while clearly only a limited knowledge of all the processes involved exists ( and even that only from the earth s atmosphere ) , this will suffice for the purposes of the present example .
two main phases can be identified in the rise of oxygen in planetary atmospheres ( berkner & marshall 1965 ; hart 1978 ; levine , hays & walker 1979 ; canuto et al .
in the first ( which on earth lasted @xmath20 yr ) , oxygen is released from the photochemical dissociation of water vapor ( this led on earth probably to oxygen levels of @xmath21 of the present atmospheric level ( p.a.l . ) ) . in the second phase ( which on earth lasted @xmath22 yr ) , the amounts of o@xmath23 and o@xmath19 reach levels @xmath24 p.a.l .
, sufficient to shadow the land from lethal uv to allow the spread of life to dry land ( the uv extinction is normally expressed as @xmath25 , where @xmath26 is the impinging intensity , @xmath27 is the absorption coefficient and @xmath28 is the path length in the atmosphere ) . the important point to note ( berkner &
marshall 1965 ; hart 1978 ) is that the duration of the first phase is inversely proportional to the intensity of the radiation in the range 10002000 ( significant peaks in h@xmath23o absorption exist in the 11001300 and 16001800 ranges ) .
thus , for a given planetary size and orbit , the timescale for the development of shielding ( which we identify approximately with @xmath0 ) is dependent on the stellar spectral type , and therefore on @xmath2
. for typical main sequence star relations , @xmath29 ( with @xmath30 in the range 0.61 , for spectral types f5k5 , see below ) , and empirical fractions of the radiation emitted in the 10002000range ( stecker 1970 ; carruthers 1971 ) , a simple calculation ( e.g. livio & kopelman 1990 ) leads to an approximate relation of the form @xmath31 = 4.6 in clearly with the existence of a relation like ( 1 ) ( which is monotonically increasing ) , the highest probability is to find @xmath3 , and hence the near equality of these two timescales in the earth - sun system can not be taken to imply that extraterrestrial civilizations are rare .
i should note again that the detailed evolution of the atmosphere is surely more complicated than a simple dependence on the intensity of uv photons .
in particular , it may be that the different phases of the evolution have different dependences on the properties of the central star .
the important point , however , is that , as the above example shows , the _ existence _ of a @xmath0@xmath2 relation is not implausible , and that @xmath32 ) _ could _ increase faster than linearly .
given that extraterrestrial `` intelligent '' civilizations may not be exceedingly rare after all , one may ask what is a likely time in the history of the universe for such civilizations to emerge .
i will restrict the discussion now to carbon - based civilizations . assuming a principle of ` mediocrity '
, one would expect the emergence to coincide perhaps with the peak in the carbon production rate .
the main contributors of carbon to the interstellar medium are intermediate mass stars ( wood 1981 ; yungelson , tutukov & livio 1993 ) through the asymptotic giant branch ( agb ) and planetary nebulae ( pne ) phases .
recent progress has been made in the understanding of the cosmic history of the star formation rate ( sfr ) ( e.g. madau et al.1996 ; lilly et al .
1996 ; madau , pozzetti & dickinson 1997 ) .
assuming for simplicity that all the galaxies follow the same sfr history and stellar evolution processes , we can calculate the rate of formation of planetary nebulae ( and hence the rate of carbon production ) as a function of redshift . for this purpose a population synthesis code which follows the evolution of all the stars ( assumed to be mainly in binaries ) , including all episodes of mass exchange , common envelope phases etc .
, has been used ( see yungelson , tutukov & livio 1993 ; yungelson et al .
1996 for details of the code ; i am grateful to lev yungelson for carrying out the simulations ) . in fig . 2 , the assumed sfr as a function of redshift ( taken as an approximation to the results in madau et al . 1996 ) and the obtained pne formation rate as a function of redshift
are shown . as can be seen , the peak in the pn rate is somewhat delayed ( to @xmath33 ) with respect to the peak in the sfr , and is much more shallow at @xmath34 , due to the build - up of a reservoir during the previous epochs .
realizing that continuously habitable zones ( chzs ) exist only around stars in the spectral range of about f5 to mid k ( e.g. kasting , whitmore & reynolds 1993 ) , and that in general the biochemistry of life requires rather precise conditions , carbon - based life may be expected to start ( with the assumed sfr history ) around @xmath35 , corresponding to an age of the universe of @xmath36 yr ( for @xmath37 , as seems to be indicated by recent observations , e.g. garnavich et al . 1998 ; parlmutter et al.1998 ; reiss , press & kirshner 1996 ; carlberg et al .
1997 , and a present age of @xmath38 billion years ) .
given the fact that ( as i have shown in the previous section ) the time required to develop intelligent civilizations is @xmath3 , it is expected that civilizations will emerge when the age of the universe is @xmath39 billion years , or maybe even somewhat older , since the chzs around k stars are somewhat wider ( in log distance ) than around g stars . a younger emergence age will be obtained if the star formation rate does not decline at redshifts @xmath40 , but rather stays flat ( as is perhaps suggested by the recent cobe diffuse background experiment ; hauser et al .
1998 ; calzetti & heckman 1998 ) .
finally , i should note that the arguments presented in this paper should definitely not be taken as attempting to imply that extraterrestrial intelligent civilizations do exist .
rather , they show that the conclusion that they do nt , is at best premature ( see also rees 1997 for discussions of related issues ) . | it is shown that , contrary to an existing claim , the near equality between the lifetime of the sun and the timescale of biological evolution on earth does not necessarily imply that extraterrestrial civilizations are exceedingly rare .
furthermore , on the basis of simple assumptions it is demonstrated that a near equality between these two timescales may be the most probable relation .
a calculation of the cosmic history of carbon production which is based on the recently determined history of the star formation rate suggests that the most likely time for intelligent civilizations to emerge in the universe , was when the universe was already older then about 10 billion years ( for an assumed current age of about 13 billion years ) . |
one of the pillars of the big bang theory is its successful prediction of a primordial baryonic matter composition , made up of nearly one hundred percent hydrogen and helium with a trace amount of a few other light elements ( e.g. , * ? ? ?
* ; * ? ? ?
the metals , nucleosynthesized in stars later , are found almost everywhere in the observable igm , ranging from the metal - rich intracluster medium ( e.g. , * ? ? ?
* ) to moderately enriched damped lyman systems ( e.g. , * ? ? ?
* ; * ? ? ?
* ) to low metallicity lyman alpha clouds ( e.g. , * ? ? ?
* ) . when and where were the metals made and why are they distributed as observed
? we address this fundamental question in the context of the standard cold dark matter cosmological model @xcite using latest simulations .
our previous simulations @xcite provided some of the earlier attempts to address this question with measured successes . in this investigation
we use substantially better simulations to provide significantly more constrained treatment of the feedback processes from star formation ( sf ) that drive energy and metals from supernovae into the igm through galactic winds ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
metal - line absorption systems in qso spectra are the primary probes of the metal enrichment of the igm as well as in the vicinities of galaxies ( e.g. , * ? ? ?
the most widely used metal lines include @xmath222796 , 2803 doublet ( e.g. , * ? ? ?
* ) , @xmath221548 , 1550 doublet ( e.g. , * ? ? ?
* ) , and @xmath221032 , 1038 doublet ( e.g. , * ? ? ?
we here focus on the and absorption lines and the global evolution of metals in the igm .
we will limit our current investigation to the observationally accessible redshift range of @xmath0 , which in part is theoretically motivated simply because the theoretical uncertainties involving still earlier star formation are much larger . at @xmath12 the line ( together with and lines ) provide vital information on the missing baryons ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , predicted to exist in a warm - hot intergalactic medium ( whim ) @xcite . for a well understood sample of qso absorption lines
, one could derive the cosmological density contained in them ( e.g. , * ? ? ?
early investigations indicate that @xmath7 remains approximately constant in the redshift interval @xmath23 @xcite .
there have been recent efforts to extend the measurements of @xmath7 to @xmath24 @xcite and to @xmath25 @xcite .
observations in these redshift ranges have been difficult to carry out because transition moves to the uv at low redshift and to the ir band at high redshift .
@xcite find evidence of a rise in the mass density for @xmath26 .
@xcite and @xcite found evidence of density at @xmath27 being consistent with estimations at @xmath28 .
more recently , however , @xcite set upper limits for @xmath7 at @xmath29 and @xcite observe a decline in intergalactic approaching @xmath30 , which we will show are in good agreement with our simulations .
the ionization potential of and the relatively high oxygen abundance are very favorable for production of absorbers in the igm ( e.g. , * ? ? ?
* ; * ? ? ?
. the rest wavelength of ovi ( @xmath31 ) places it within the ly-@xmath32 forest , which makes the identifications of these lines more complicated , although being a doublet helps significantly . at @xmath33 , however , absorption can probe the metal content of the igm in ways complementary to what is provided by lines .
for example , the lines can probe igm that is hotter than that probed by the lines and can reach lower densities thank to higher abundance .
there are now several observational studies at redshifts @xmath14 that describe the properties of absorbers and attempt to estimate the mass density , @xmath8 @xcite . at @xmath34
there is a missing metals problem : only 10 - 20% of the metals produced by all stars formed earlier have been identified in stars of lyman break galaxies ( lbg ) , in damped lyman alpha systems ( dlas ) and @xmath35 forest .
the vast majority of the produced metals appear to be missing ( e.g. , * ? ? ?
the missing metals could be in hot gaseous halos of star - forming galaxies @xcite .
we will show that most of the missing metals are in a warm - hot ( @xmath15k ) but diffuse igm at @xmath14 of overdensities of @xmath36 that are outside of halos .
the outline of this paper is as follows . in
2 we detail our simulations and the procedure of normalizing the uncertain feedback processes from star formation . results on the metal enrichment of the igm are presented in 3 . in 3.1 we give a full description of the properties of the and lines at @xmath0 , followed 3.2 discussing and absorbers as metals reservoirs .
we devote 3.3 to a general discussion of global distribution of metals , addressing several specific topics , including the metallicity of the moderate overdense regions at moderate redshift , the missing metals at @xmath5 .
conclusions are given in 4 .
numerical methods of the cosmological hydrodynamic code and input physical ingredients have been described in detail in an earlier paper @xcite .
the simulation integrates five sets of equations simultaneously : the euler equations for gas dynamics in comoving coordinates , time dependent rate equations for hydrogen and helium species , the newtonian equations of motion for dynamics of collisionless ( dark matter ) particles , the poisson equation for the gravitational potential field and the equation governing the evolution of the intergalactic ionizing radiation field , all in cosmological comoving coordinates .
the gasdynamic equations are solved using a new , improved hydrodynamics code ,
cosmo " @xcite on a uniform mesh .
the rate equations are treated using sub - cycles within a hydrodynamic time step due to the much shorter ionization time - scales ( i.e. , the rate equations are very stiff " ) .
dark matter particles are advanced in time using the standard particle - mesh ( pm ) with a leapfrog integrator .
the poisson equation is solved using the fast fourier transform ( fft ) method on the uniform mesh .
the initial conditions adopted are those for gaussian processes with the phases of the different waves being random and uncorrelated .
the initial condition is generated by the cosmics software package kindly provided by e. bertschinger ( 2001 ) .
cooling and heating processes due to all the principal line and continuum atomic processes for a plasma of primordial composition with additional metals ejected from star formation .
compton cooling due to the microwave background radiation field and compton cooling / heating due to the x - ray and high energy background are computed .
the cooling / heating due to metals is computed using a code based on the raymond - smith code assuming ionization equilibrium that takes into account the presence of a time - dependent uv / x - ray radiation background , which we have included in our simulations since @xcite and has now been performed by other investigators ( e.g. , * ? ? ?
we follow star formation using a well defined , schmidt - kennicutt - law - like prescription used by us in our previous work and similar to that of other investigators ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
a stellar particle of mass @xmath37 is created ( the same amount is removed from the gas mass in the cell ) , if the gas in a cell at any time meets the following three conditions simultaneously : ( i ) contracting flow , ( ii ) cooling time less than dynamic time , and ( iii ) jeans unstable , where @xmath38 is the time step , @xmath39yrs ) , @xmath40 is the dynamical time of the cell , @xmath41 is the baryonic gas mass in the cell and @xmath42 is star formation efficiency ( e.g. , * ? ? ? * ) .
each stellar particle is given a number of other attributes at birth , including formation time @xmath43 , initial gas metallicity and the free - fall time in the birth cell @xmath44 .
the typical mass of a stellar particle in the simulation is about @xmath45 ; in other words , these stellar particles are like coeval globular clusters .
all variations of this commonly adopted star - formation algorithm essentially achieve the same goal : in any region where gas density exceeds the stellar density , gas is transformed to stars on a timescale longer than the local dynamical time and shorter than the hubble time .
since these two time scales are widely separated , the effects , on the longer time scale , of changing the dimensionless numbers ( here @xmath46 ) are minimal .
since nature does not provide us with examples of systems which violate this condition ( systems which persist over many dynamical and cooling time scales in having more gas than stars ) , this commonly adopted algorithm should be adequate even though our understanding of star formation remains crude .
stellar particles are treated dynamically as collisionless particles subsequent to their birth .
feedback from star formation , the effects of the cumulative sn explosions known as galactic superwinds ( gsw ) and metal - enriched gas , will be described in more detail in the next subsection .
while the code can self - consistently compute the ionizing uv - x - ray background using sources and sinks in the simulation , here we use the @xcite spectra for all runs such that we do not introduce additional variations due to otherwise varying uv backgrounds in the different runs .
however , a local optical depth approximation is adopted to crudely mimic the local shielding effects : each cubic cell is flagged with six hydrogen optical depths " on the six faces , each equal to the product of neutral hydrogen density , hydrogen ionization cross section and scale height , and the appropriate mean from the six values is then calculated ; analogous ones are computed for neutral helium and singly - ionized helium . in computing the local ionization and cooling / heating balance for each cell , self - shielding
is taken into account to attenuate the external hm ionizing radiation field .
both these two shielding effects are essential in order to obtain self - consistent radiation background evolution and neutral hydrogen evolution .
lccll n & 50 & 24 & @xmath47 & @xmath48 + l & 50 & 24 & @xmath47 & @xmath49 + * m * & * 50 * & * 24 * & * @xmath47 * & * @xmath50 * + h & 50 & 24 & @xmath47 & @xmath51 + * mr * & 50 & 48 & @xmath52 & @xmath50 we have run a set of four new simulations of a wmap5-normalized @xcite cold dark matter model with a cosmological constant : @xmath53 , @xmath54 , @xmath55 , @xmath56 , @xmath57 and @xmath58 .
the adopted box size is @xmath59mpc / h comoving and with @xmath60 cells of size @xmath61kpc / h comoving ; the dark matter particle mass and mean baryonic mass in a cell are equal to @xmath62 and @xmath63 , respectively .
some of the key parameters for the four simulations are summarized in table 1 .
the only difference among the four main runs is the strength of the gsw feedback : ( n ) no gsw , ( l ) low gsw feedback , ( m ) moderate gsw feedback and ( h ) high gsw feedback . in the next subsection
we will determine which feedback strength produces the star formation rate history that matches observations .
we run an additional lower resolution simulation with @xmath64 cells a side , each of size @xmath65kpc / h ( run `` mr '' ) to test convergence of results . when computing results using run mr " , we multiply the metallicity of each cell in run mr " by a constant factor such that its mean metallicity at any epoch match that of run m " .
we obtain an additional set of results by changing the amplitude of the uv background , run `` m2 '' , where it is reduced to one half of that in `` m '' .
it is well known that without impeding processes to counter the cooling and subsequent condensation of baryons , the stellar mass in the universe would be overproduced
the overcooling " problem ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
feedback from star formation is believed to play the essential role to prevent gas from overcooling .
the key question is : where does the feedback from sf throttle gas cooling and condensation ?
we consider three independent lines of evidence to address this question .
first , while metals from supernovae ejecta can be accelerated to velocities exceeding the escape velocity , the whole interstellar gas is very difficult to be blown away , even in starburst galaxies , based on simulations ( e.g. , * ? ? ?
* ) , although their adopted feedback strength may be on the low side .
second , observed normal galaxies in the local universe tend to be relatively gas poor ( e.g. , * ? ? ? * ) .
their progeniors or their building blocks were presumably gas rich in the past when most of the star formation occurred .
this implies that , once gas has collapsed , it would turn into stars on a time scale that is shorter than the hubble time .
finally , if gas were able to collapse inside halos without hinderance , the observed soft x - ray background would be overproduced by more than an order of magnitude @xcite .
these three lines of evidence together suggest that feedback from star formation likely exerts its effect outside normal stellar disks , probably in regions that are tens to hundreds of kiloparsecs from halo centers , before too much gas has either been collected inside the virial radius or cooled and condensed onto the disk .
it is currently difficult to fully model gsw in a cosmological simulation , although significant progress has been made to provide a better treatment of the multi - phase interstellar medium ( e.g. , * ? ? ?
* ; * ? ? ?
it is likely that a combination of both high resolution and detailed multi - phase medium treatment ( perhaps with the inclusion of magnetic fields and cosmic rays ) is a requisite for reproducing observations .
here we do not attempt to model the _ causes _ and generation of gsw , but , instead , to simply assume an input level of mass , energy and metals , and carefully compute the _ consequences _ of gsw on the surrounding medium and on subsequent galaxy formation .
our simulations have a resolution of @xmath61kpc / h comoving ( see table 1 ) , which may provide an adequate resolution for this purpose , given the aforementioned lines of evidence that feedback from star formation likely exerts most of its effects in regions on scales larger than tens of kiloparsecs . in our simulations , gsw energy and ejected metals
are distributed into 27 local gas cells centered at the stellar particle in question , weighted by the specific volume of each cell @xcite .
the temporal release of the feedback at time @xmath66 has the following form , all being proportional to the local star formation rate : @xmath67 @xmath68 $ ] @xmath69 $ ] . within a time step @xmath70 ,
the released gsw energy and mass to the igm from stars are @xmath71 and @xmath72 , respectively .
we fix @xmath73 , i.e. , 25% of the stellar mass is recycled with the ejecta metallicity of @xmath74 .
metals , collectively having the observed solar abundance pattern , are followed as a separate hydro variable ( analogous to the total gas density or nuetral hydrogen , hei density , heii density ) with the same hydrocode .
we do not introduce any additional diffusion " process for the metals .
we note that cooling process is never turned off , before or after the deposition of thermal energy , and hydrodynamic coupling between ejected baryons and surrounding gas is not turned off either , a departure from some of the previous simulations ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
this is physically made possible in part due to a deposition of energy at scales that are comparable or larger than the sedov radius in our current simulations , thanks to our limited spatial resolution .
-0.5 in -1.0 in the gsw strength is therefore controlled by one single adjustable parameter , @xmath75 .
we normalize @xmath75 by the requirement that the computed star formation rate ( sfr ) history matches , as closely as possible , the observations over the redshift range @xmath12 to @xmath30 where comparisons can be made .
figure [ fig : sfr ] shows the sfr history for the three runs with non - zero @xmath75 , ( l , m , h ) .
what is immediately evident is that the mechanical feedback strength from star formation has a dramatic effect on the overall sfr history , especially at low redshift ( @xmath76 ) . at the resolution of the simulation ,
run m " provides the best and excellent match to observations , where run
l " and h " , respectively , over- and under - estimate the sfr at @xmath77 . at the time of this writing
we prefer to avoid introducing additional ad hoc physics to remedy this and are instead content with the ballpark agreement at @xmath78 between simulations and observations , given the large uncertainties in the observational data as evidenced by the large dispersion among different observations .
at redshift zero we find that the stellar densities in the three models ( l , m , h ) are @xmath79 , which should be compared to the observed value of @xmath80 @xcite .
our experiments indicate that , had we set @xmath81 , the amount of stellar density @xmath82 at @xmath12 would exceed @xmath83 , in serious disagreement with observations . in this respect model
m " also agrees better with observations .
our findings are in agreement with @xcite and @xcite in that star formation rate history depends sensitively on the stellar feedback , but in disagreement with @xcite who find otherwise .
all the subsequent results presented are based on run
there is some indication that a model between m " and l " might provide a better match to the observations at low redshift ( @xmath84 ) if the compilation of @xcite is used .
but we note that such a model may run into a worse agreement with observations with respect to @xmath82 at @xmath12 .
currently , it is difficult to reconcile the observations of star formation rate history and @xmath82 at @xmath12 .
one might appeal to an evolving imf to provide an attractive reconcilation between the possible discrepancy @xcite .
this is well beyond the scope of this investigation . in any case , a slight varied simulation , say , using an @xmath75 value between the m " and l " would give qualitatively comparable results . in order to test for numerical convergence
we run one additional simulation , `` mr '' , which has the same parameters as run `` m '' but have half the resolution . to test the dependence of results on the extragalactic uv background we run our software pipeline through run m " but with halving the amplitude of the uv background , called run m2 " .
-0.6 in -2.7 in it is prudent to make a self - consistency check for the value of @xmath75 that is empirically determined . the total amount of explosion kinetic energy from type ii supernovae with a chabrier imf translates to @xmath85 .
observations of local sturburst galaxies indicate that nearly all of the star formation produced kinetic energy ( due to type ii supernovae ) is used to power gsw ( e.g. , * ? ? ?
given the uncertainties on the evolution of imf with redshift the fact that newly discovered prompt type i supernovae contribute a comparable amount of energy compared to type ii supernovae , we argue that our adopted best " value of @xmath86 is consistent with observations and entirely within physical plausibility .
the photoionization code cloudy @xcite is used post - simulation to compute the abundance of and , adopting the uv background calculated by @xcite . for @xmath35 absorption lines we use the computed neutral hydrogen density distribution directly from the simulation that was already using the @xcite uv background in the rate equations for hydrogen and helium species .
we have checked that the radiation field is consistent with observations by comparing the simulated mean transmitted flux as a function of redshift with observations .
figure [ fig : meanf ] shows the mean transmitted @xmath35 flux as a function of redshift from the simulation in comparison with observations .
we see the @xmath35 forest produced in the lcdm model using the adopted uv background provides an adequate match to observations over most of the redshift range compared , @xmath87 . at @xmath88
, our results do not seem to coincide with observations .
we attribute this to the uv background used : we have only considered a quasar background , while at these high redshifts the uv radiation coming from galaxies should have a significant effect on the @xmath35 forest .
nevertheless , we do not expect this to be an issue on the metal species considered in the following sections .
these correspond to much higher energies than 1 ryd that are not affected by the uv contribution from galaxies to the ionizing radiation .
we generate random synthetic absorption spectra for each of the three absorption lines by producing optical depth distribution along lines of sight parallel to one of the three axes of the simulation box , based on density , temperature and velocity distributions in the simulation ( i.e. , our calculations include redshift effects due to peculiar velocities and thermal broadening ) .
the code used is similar to that used in our earlier papers @xcite .
we identify each absorption line as a contiguous region in the flux spectrum between a down - crossing point and an up - crossing point , both at a flux equal to @xmath89 .
note that flux equal to @xmath90 corresponds to no absorption .
for each identified line we compute its equivalent width ( ew ) , doppler width ( @xmath91 ) , mean temperature ( @xmath92 ) , mean metallicity ( @xmath93 ) and mean gas overdensity ( @xmath94 ) , weighted by optically depth of each pixel .
we do not attempt to perform voigt profile fitting , a procedure often used to analyze observed spectra . because of this , we tend to not generate some of the very low column lines that are purely an effect of profile fitting process .
also , precise comparison between our mock absorbers and observed ones is not possible for some quantities , such as doppler width distributions .
-0.7 in [ cols="^,^ " , ] -2.5 in the second point , perhaps the most overlooked , is that the amount of metals contained in the and absorption line is a very small fraction of the overall metals .
the left panel of figure [ fig : omega_metals ] shows the ratios of mass density measured in the ( diamonds ) and lines ( triangles ) over the total amount of metals in the igm as a function of redshift .
we see that the amount of mass contained in the line remains at @xmath95 within a dispersion of @xmath96 , and at @xmath95 within a dispersion of @xmath97 for the line . in the right panel of figure
[ fig : omega_metals ] we show the amount of metals probed by each line as a function of redshift . here
is how we compute the metals probed by each line and use as an example . for each detected line
, a range of spatial locations ( i.e. , gas cells along the line of sight ) contributes to its column density ( see figures [ fig : spectrum1],[fig : spectrum2],[fig : spectrum3 ] ) .
roughly speaking , the amount of metals effectively probed by the line will be larger than the metals directly seen in the line by a factor of @xmath98 ( a similar relation for the line ) .
this ratio @xmath99 for the and line is shown in figure [ fig : fsw ] .
one point worth noting is that for there is an upturn of @xmath99 from @xmath100 at @xmath77 towards high redshift , reaching again the same value at @xmath30 .
this is caused from a transition from more collisionally dominated absorber population at @xmath101 to a more photoionization dominated one at intermediate redshifts .
the trend for the line is much less pronounced , indicative of a dominance of collisionally ionized absorbers over the entire redshift range @xmath0 , with a trend that it is more so at higher redshift . from the right panel of figure
[ fig : omega_metals ] we see that , within a factor of @xmath102 , the amount of metals probed by either or line is roughtly @xmath103 . combining the fact that the majority of -producing regions have not collapsed and virialized ( see figure [ fig : civdnddelta ] ) and a small fraction of all metals is probed by and lines at all redshifts , and absorbers are transients " ; in other words , only a small fraction of metals in the igm get lit up " as the or line at any given time . as we demonstrated earlier ,
these regions that produce absorption lines have a set of properties that seem to be created by a combination of physical processes including feedback shock heating and radiative cooling ( see figures [ fig : spectrum1],[fig : spectrum2],[fig : spectrum3 ] ) .
these close observations suggest that only a fraction of metals at any given time that has recently passed through shocks and cooled to an appropriate temperature shows up as absorption lines . in this sense , absorption lines trace the current feedback processes from star formation and how the current feedback energy and metal - enriched gas interact with the surrounding igm
. similar statements about the transient nature could be made for the line , except that the line corresponds to somewhat different physical states of the shocked regions : they are slightly hotter in temperature and dynamically hotter .
third , returning to figure [ fig : omegaciv ] , we would like to emphasize that , consistent with recent observations ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , there is indeed a sharp drop in @xmath7 from @xmath104 to @xmath105 by a factor of @xmath106 , respectively .
this has less to do with the evolution of the total amount of metals produced , rather it is tracing the phase of gas at any given time .
at redshift @xmath107 , lines at different redshifts appear to come from regions of comparable overdensity ( see figure [ fig : civdnddelta ] ) and comparable metallicity ( see figure [ fig : civdndz ] ) .
this allows us to test a very simple physical picture for the origin of lines .
they are produced by regions that were shock heated earlier by feedback shocks and have cooled to the temperature of @xmath108k when they are seen , and the duration of each line in this
phase " would then be inversely proportional to the cooling time of the gas in this phase , which is proportional to @xmath109 , where @xmath110 is cooling function at temperature @xmath92 and metallicity @xmath93 and the z - dependent term is due to density evolution with redshift .
then , the total amount of metals in lines , @xmath7 , will be proportional to @xmath111 , where @xmath112 is the star formation rate at @xmath113 .
taking @xmath94 , @xmath93 and @xmath92 as roughly being constant ( see figures [ fig : civdnddelta ] , [ fig : civdndz ] , [ fig : civdndt ] ) , we have @xmath114 , which is shown as the dashed curve on the left panel of figure [ fig : omegaciv ] .
it provides a reasonably good fit for the actual computed evolution of @xmath7 .
we now turn to present a global metal enrichment history of the igm to supplement what is captured by the and absorption lines . as in @xcite , in our analysis
we divide the igm into three components by temperature : ( 1 ) @xmath115k cold - warm gas , which is in low density regions or cooling , star forming gas , ( 2 ) whim at @xmath116k@xmath117k , ( 3 ) hot x - ray emitting gas at @xmath118k
. one additional component ( 4 ) is the baryons that have left the igm and been condensed into stellar objects , which we designate as stars " .
-1 in -1 in -1.6 in -1.4 in figure [ fig : compz ] shows the evolution of these four components .
the overall evolution of the four components are in good agreement with earlier findings @xcite and relevant observations ( e.g. , * ? ? ?
in particular , we see that @xmath119 of all baryons are in whim by the time @xmath12 , which is in excellent agreement with our previous findings @xcite .
it is also noted that @xmath120 of the baryons at @xmath12 reside in a relatively cool but diffuse component with @xmath121k ( the triangles in figure [ fig : compz ] ) .
it is likely that a significant portion of this cool component at @xmath12 , in the form of @xmath35 forest , is already seen by uv observations ( e.g. , * ? ? ?
as we noted earlier , the strength of feedback from star formation is chosen to match the observed overall star formation history .
each of the igm components is composed of different regions that have gone through distinct evolutionary paths and thus spans a wide range in density , shown in figure [ fig : hisrho ] .
the distribution of the cold - warm component ( triangles ) is always peaked at the mean density at all redshifts , reflecting the initial gaussian distribution of gas around the cosmic mean and indicating that the bulk of the igm at mean density or lower has never been shock heated by either strong gravitational shocks or feedback shocks .
the cold - warm gas extends to very high densities ( @xmath122 ) .
it is interesting to note that the amount of cold - warm gas that could potentially feed the star formation , i.e. , the cold - warm gas at density @xmath123 , remains constant , within a factor of @xmath124 , over the range redshift shown @xmath125 .
this is consistent with observations of the nearly non - evolving amount of gas probed by dlas ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the physical relation between this apparently non - evolving gas and the precipitous drop of star formation rate at @xmath84 is currently unclear .
the distribution of the whim also appears to peak at a constant overdensity of about 10 times the mean density .
this is rather intriguing . in order to properly interpret this interesting phenomenon
it is useful to understand the heating sources of whim .
there are two primary heating sources for whim : shocks due to the collapse of large - scale structure and gsw produced shocks .
earlier works have already shown that gravitational shock heating due to the formation of large - scale structure dominates the energy input for heating up and thus turning about @xmath126 of the igm into whim by @xmath12 @xcite .
it is , however , expected that heating due to hydrodynamic shocks emanating from galactic superwinds become increasingly more important at higher redshifts .
this is because the amount of energy from gravitational collapse of large - scale structure as well as the resulting shock velocity decreases steeply towards higher redshift .
the reason for this is simple : in the standard cosmological model the amount of power is peaked at a wavelength of @xmath127mpc / h and drops steeply towards small scales . to quantify the relative contribution of gsw in turning the igm into whim , we compare the simulation with gsw feedback to that without gsw feedback ( run n in table 1 ) .
then , we make the simple assertion that the difference in the amount of whim between the two simulations is due to gsw .
figure [ fig : gswz ] shows the fraction of whim that is produced ( cumulatively ) by gsw as a function of redshift .
consistent with previous results @xcite , the contribution from gsw to heating up whim by @xmath12 is subdominant at @xmath128 .
this relatively small contribution to whim from gsw can be understood based on simple energetics estimates .
but we see the gsw fraction increases rapidly with increasing redshift . at redshift
@xmath129 the gsw fraction is about @xmath126 , then reaching @xmath130 at @xmath104 and @xmath131 at @xmath132 .
thus , we see the primary heating source of whim at @xmath133 is gsw , whereas gravitational shocks due to structure formation are mostly responsible for heating the whim at @xmath24 . from figure
[ fig : hisrho ] it seems clear that whim does not distinguish between gravitational shocks and feedback shocks . in both cases shocks have largely stopped at overdensity of about @xmath134 .
let us try to understand why that happened .
first we note that the shocks originate approximately from the central regions of filaments , where pancakes collapse and shock for the case of gravitational shocks and galaxies are generally located for the case of gsw shocks . for gas shock heated to @xmath135k
the shock velocity is roughly @xmath136 . with that velocity
the shock will be able to travel roughly @xmath137kpc comoving over the hubble time at any redshift .
therefore , one should expect to see shocks have reached a few hundred kpc comoving at any redshift , which are about one to a few times the virial radius of typical large galaxies , which in turn correspond an overdensity in the vicinity of @xmath134 and are thus in good agreement with simulation results .
some shocks penetrate deeper into the igm , especially along directions with lower densities and steeper density gradients , as seen in figure [ fig : civslice ] ; but the amount of mass effected in these low density regions is small , corresponding to the sharp drop of whim mass at the low density end ( figure [ fig : hisrho ] ) .
this last point is best corroborated by the distribution of the hot gas at high redshift ( @xmath138 ) , in the bottom two panels of figure [ fig : hisrho ] .
there we see a small amount of hot gas heated up by gsw shocks is indeed produced in regions of density lower than the mean density and traces a larger amount of whim gas that is also produced there . at @xmath139 , some comparable
, small amount of hot gas is still produced at low density regions .
but the vast majority of hot x - ray emitting gas is now residing in the deep potential wells of x - ray clusters of galaxies , when the cluster scale turns nonlinear and collapses .
-1 in -1 in having obtained an overview of the thermal history of the igm , we now turn to the metal story
. we will first focus our attention on the whim here , because that is where most of the energy and metal exchanges between galaxies and the igm take place , as shown in figure [ fig : hisrho ] .
observationally , integrating the observed star formation rate history from high redshift down to @xmath140 suggests that the vast majority ( possibly @xmath141 ) of cosmic metals at @xmath142 appear to be missing ( e.g. , * ? ? ?
* ; * ? ? ?
note that this conclusion is insensitive to the choice of imf , since both uv light and metals are , to zeroth order , produced by the same massive stars .
metals that have been accounted for in the estimates include those in stars of lyman break galaxies ( lbg ) , damped lyman alpha systems ( dlas ) and @xmath35 forest , i.e. , cold - warm gas and stars . given the dominant heating of whim by gsw , one may immediately ask : could a significant fraction of metals that accompanies the gsw energy be heated up and in a phase like whim that is different from those where metals have been inventoried ? to better address this open question , we further break down the igm component ( 1 ) ( @xmath115k cold - warm gas ) into two sub - components with ( 1c ) ( @xmath143k cold gas ) and ( 1w ) ( @xmath144k warm gas ) .
the purpose of this finer division is to separate out the cold gas ( 1c ) , which can be more appropriately identified with @xmath35 forest clouds and dlas .
the results are shown in figure [ fig : mtlcompz ] .
we see that about one third of all metals produced by @xmath12 is locked up in stars , decreasing monotonically towards high redshift , dropping to about @xmath145 by @xmath132 .
the fraction of metals in the hot x - ray emitting component is at about @xmath145 level at @xmath12 , plummeting to about @xmath103 at @xmath20 and slowly rising back to about @xmath146 at @xmath30 .
it is likely that the metal fraction in the hot x - ray component at @xmath84 be somewhat underestimated given the relatively moderate simulation boxsize .
the remaining metals are in the general photoionized @xmath35 forest and the whim . at @xmath30
the @xmath35 forest ( @xmath143k , open triangles ) contains about @xmath147 of all metals , while whim ( @xmath148k , open circles ) , and warm igm ( @xmath144k , solid triangles ) contain @xmath149 and @xmath150 , respectively .
but the fraction of metals in the @xmath35 forest decreases steadily with time and becomes a minor component by @xmath12 at @xmath151 .
most of the metals is seen to be contained in the whim at all times below redshift five at @xmath152 , peaking at @xmath153 at redshift @xmath154 . in total
, the amount of metals contained in the igm with temperature @xmath155 constitutes about 2/3 of all metals produced by @xmath140 .
metals in this temperature range were not accounted for in the quoted observational inventory at @xmath140 .
thus , it seems probable that the missing metals problem at @xmath14 can be largely rectified , if one counts the metals in the igm at @xmath155k .
by now we have learned that a large amount of metals could be hidden in the whim of temperature @xmath156k spanning a wide range in density .
since the metallicity is a strong function of density , it is still unclear the location of the whim that dominates the missing metals .
figure [ fig : zrho ] shows the mean metallicity of the three igm components as a function of overdensity at four different redshifts .
it is evident that within each igm component there is a wide range in metallicity that is a non - trivial function of overdensity .
let us examine their behaviors in detail .
for all three igm components there is a strong correlation between the mean metallicity and overdensity at overdensity @xmath2 and they converge at the highest density . while the metallicity of the cold - warm gas at the high density end remains at about solar at high density , its mean metallicity at the low overdensity drops rapidly with increasing redshift .
for example , at @xmath157 , the mean metallicity is ( -2 , -2.5 , -3 , -4 ) in solar units at @xmath158 .
-1.3 in -1.4 in one may notice that all three distributions exhibit a minimum metallicity at some intermediate density range , @xmath159 for the cold warm - gas , @xmath157 for the whim and @xmath160 for hot gas ( only at @xmath161 ) .
this is entirely in agreement with the physical picture that we described earlier for the gsw shock propagation through the igm .
figure [ fig : zrho ] confirms that the transformation of cold gas to whim roughly stops at @xmath157 .
additional metal - enriched gas is further transported along some directions , such as those perpendicular to the filaments , to very low density regions and enrich these regions to higher metallicity ( due to a negligible amount of pre - existing gas there ) .
the behavior of cold and hot components at the low density end can be understood in the same way as the whim .
the metallicity of hot gas at the centers of clusters of galaxies ( at overdensity @xmath162 ) appear to stay in narrow range around @xmath4\sim -0.5 $ ] over the redshift range @xmath161 , consistent with observations ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
there is some indication of a still higher metallicity towards higher density regions , which may be in agreement with observations ( e.g. , * ? ? ?
the metallicity of the whim at the peak of its mass distribution ( @xmath163 ) at @xmath12 is @xmath4\sim -1 $ ] , in good agreement with observations ( e.g. , * ? ? ?
we find that the following formula fits well the metallicity of the whim as a function of overdensity @xmath164 at the redshift range @xmath165 : @xmath166_{whim}=-1.2 - 0.08z + ( 0.3 + 0.12z^{1/3})(\log\rho/\langle\rho\rangle -1),\ ] ] which are shown as the straight lines in the three panels of figure [ fig : zrho ] .
-1.6 in -1.6 in we now examine directly the distribution of metal mass as a function of density for each igm component , shown in figure [ fig : mtlrho ] .
a very interesting result is that at high redshift ( @xmath138 ) the metals mass in the whim tends to peak at a somewhat lower overdensity than that for the overall whim mass , thanks to the upturn of metallicity of the whim at low overdensity end .
specifically , at @xmath167 it appears that the metals mass peaks at @xmath168 , whereas the total whim mass peaks at @xmath169 .
this trend is reversed at lower redshift ; for example , at @xmath12 the metals in whim is now broadly peaked at @xmath170 , while the whim mass peaks at @xmath169 .
this reversal is likely due to accretion of metal - enriched gas onto high density regions during recent formation of large - scale structures .
quantitatively , we find that , at @xmath140 , only about @xmath171 of the metals in warm and whim gas is located within virialized regions . about @xmath172 of the metals in warm and whim gas resides in the igm with @xmath160 , with the remaining @xmath173 in underdense regions .
this confirms an earlier expectation that some of the missing metals may be in the hot halos of galaxies ( e.g. , * ? ? ?
* ; * ? ? ?
* ) ; but that accounts for only a small fraction of the total missing metals .
combining with our earlier statements on missing metals at @xmath14 , our finding on missing metals is that most of the missing metals are in the warm and whim gas with moderate overdensity broadly distributed between @xmath174 .
-1.0 in -1.4 in finally , our attention is turned to the cold - warm component , which displays a dramatic trough at the mean density .
physically , it suggests that gsw does not affect bulk of the igm .
comparisons with observations are useful here to shed light on this dramatic behavior .
we note that the mean metallicity at overdensity @xmath175 drops quickly below @xmath4=-3 $ ] .
the typical @xmath35 forest clouds of column density @xmath176 @xmath177 arise in these moderate density regions .
our simulations suggest that most of these clouds are not expected to be enriched to a level higher than @xmath4=-3 $ ] , which appear to be in agreement with direct metallicity measurements of @xmath35 forest clouds ( e.g. , * ? ? ?
* ; * ? ? ?
* ) . however , our results are at variance with recent measurements of metallicity in these moderate density regions using pod method in the sense that the observed metallicity seems to far exceed what we obtain in our simulations . to illustrate the disagreement we cast the information presented in figures [ fig : hisrho ] and [ fig : zrho ] into a different form in figure [ fig : zdelta ] , where we show the evolution of metallicity as a function of redshift at four fixed densities , @xmath178 , for the ease of comparison .
if one compares the middle solid square curve ( the median metallicity at overdensity @xmath134 from our simulations ) and the open squares curve ( the median metallicity at overdensity @xmath134 from observations , @xcite ) , the middle solid dots curve ( the median metallicity at overdensity @xmath90 from our simulations ) and the open circles curve ( the median metallicity at overdensity @xmath90 from observations , @xcite ) , the disagreement is clear and dramatic .
we predict that the metallicity in regions with overdensity less than about @xmath134 generally increases quite rapidly with decreasing redshift , whereas the observationally inferred trend goes in the opposite direction with a mild rate of change .
is our simulation incomplete or are the observations misinterpreted ?
recall from figure [ fig : civdnddelta ] that the typical overdensity for low column lines is about @xmath134 , comparable to that of @xmath35 forest clouds .
but that is a mere coincidence : the two types of absorbers are generally not co - located in physical space .
if we go back to figures [ fig : spectrum1],[fig : spectrum2],[fig : spectrum3 ] and study the temperature ( second rows ) and metallicity ( third row ) , we see there is a strong spatial correlation between temperature and metallicity ; regions where a significant amount of reside tend to have an elevated temperature that exceeds @xmath179k , whereas the metallicity in lower temperature regions , where hi reside in abundance to give rise @xmath35 forest clouds , seems extremely low . as we noted earlier , the regions with elevated temperature and civ lines have a width that corresponds to one to several hundred km / s .
interestingly , these regions also typically have peculiar velocities of several hundred km / s ( fourth row from top of figures [ fig : spectrum1],[fig : spectrum2],[fig : spectrum3 ] ) . as a result
, there should be some overlap in velocity space between some lines and @xmath35 forest lines , even when they are significantly displaced in physical space . this overlap may diffuse " , in velocity space , some of the metals in regions that producing lines into the @xmath35 forest lines , causing an apparent , moderate metallicity level in @xmath35 forest , as inferred by @xcite ) , when a method such as pod is employed . a closer look at the left panel of figure [ fig : civdndz ] indicates that typical absorbers show a decrease of metallicity with decreasing redshift in the range @xmath180 : roughly @xmath4=[-2.0,-1.5],[-2.3,1.4],[-2.6,-1.5]$ ] at @xmath181 .
this is in accord with the observed weak trend of increasing metallicity with increasing redshift , which otherwise is extremely difficult to understand in the context of the standard cosmological model .
needless to say , the lines located in regions that are spatially close to lines will also diffuse " into the @xmath35 forest in velocity space .
the fact that lines tend to have a higher metallicity , about @xmath4=0.2 $ ] to @xmath182 , than the lines over the redshift range of @xmath183 ( comparing the left and right panels of figure [ fig : civdndz ] ) and there are more lines than lines ( comparing the left and right panels of figure [ fig : civdndz ] ) would suggest that one may expect that the apparent oxygen abundance in the @xmath35 forest inferred from pod should be higher than that of lines .
this is indeed the case : @xcite found that @xmath184=0.66^{+0.06}_{-0.2}$ ] .
we argue that this provides independent , supporting evidence for our explanation that is self - consistent and physically plausible .
alternatively , the igm may be enriched to the observed level by first generation , pop iii galaxies that are not properly captured in our simulations . to further test our `` diffusion '' hypothesis ,
we have computed the cross - correlation between @xmath35 , and spectra and taken the mean along all lines of sight at @xmath185 for two cases : run `` m '' of our simulations with and without the effect of peculiar velocities taken into account .
we present in figure [ podcheck ] the following function : @xmath186 where @xmath187 is the cross - correlation function for the spectrum of hi and the corresponding ion averaged over all lines of sight and symmetrized for positive and negative velocity lags at @xmath185 .
@xmath188 is the same function computed in the case where there are no peculiar velocities .
figure [ podcheck ] shows that in the case of no peculiar velocity , the cross - correlation between @xmath35 and , and @xmath35 and , is weaker than in the case where peculiar velocities are considered .
this is compelling evidence that peculiar velocities effects could artificially diffuse metals into the @xmath35 forest .
-0.7 in -2.8 in
we have carried out the state - of - the - art cosmological hydrodynamic simulations of the standard cold dark matter model to investigate the process of metal enrichment of the intergalactic medium .
our simulations have substantially higher resolution than our previous simulations to address this problem . more importantly , we can now constrain the strength of the feedback process by matching the star formation history in our current simulation to the observed one in the range @xmath0 .
we find that our model reproduces the observed mean flux of the @xmath35 forest and the mass density of and absorbers .
it is also in general consistent with observed physical properties of absorbers .
this indicates that we can explain the metal enrichment of the igm by considering star formation to be the main feedback mechanism , with no apparent need of significant contribution from agn in terms of additional energy .
we conclude from our results that : \(1 ) the overall star formation history depends rather sensitively on the feedback strength .
this is likely due to gsw significantly reducing the concentration of cold gas around halos .
nevertheless , gsw do not significantly alter the overall large - scale filamentary baryonic structure that follows the cosmic web of dark matter distribution . while gsw could travel far into low density regions
sometimes , the amount of energy and metals that are deposited in underdense regions is very small .
most of the gsw energy and metals remain in regions of overdensity @xmath2 , with the distance of influence of gsw from galaxies limited to about @xmath1mpc .
metal bubbles blown by gsw coincide with temperatures bubbles , suggesting a tight coupling of energy and metal deposition , and they are terminated by shock fronts .
\(2 ) both and absorbers are located in regions that have been swept by feedback shocks , of elevated temperature ( @xmath3k ) , density ( @xmath2 ) and metallicity ( @xmath4=[-2.5,-0.5]$ ] ) , demarcated by a double shock propagating outwards , with absorbers typically having a higher metallicity than absorbers . within these shocked regions ,
most of absorbers tend to arise from moderate density peaks that are troughs in temperature and are thus relatively quiescent .
the absorbers are from regions that are dynamically hotter near shock fronts .
there is a trend for the population of and absorbers to be more collisionally ionized at higher redshift ; for collisional ionization dominates over the entire redshift range @xmath0 , whereas for the transition occurs at moderate redshift @xmath5 from collisionally dominated to photoionization dominated .
\(3 ) the evolution of the mass density contained in and lines , @xmath7 is in good agreement with observations , with both the latest observations and simulations of @xmath7 exhibiting an exponential drop beyond redshift @xmath21 ; @xmath7 drop exponentially beyond redshift @xmath104 ; the near constancy of @xmath7 at redshift @xmath189 does not reflect the evolution of the overall metal content in the igm .
in the case of @xmath8 , we find a less good agreement between observations and out results .
this might be in part due to cosmic variance .
\(4 ) most of and absorbers , while clustered around galaxies , are transient and intergalactic in origin , produced by galactic superwinds in the process of transporting both energy and metals from galaxies into the igm ; the metal mass densities contained in and lines in the range @xmath6 each constitutes @xmath190 of total metal density at all redshifts ; the amount of metals probed by and lines in the range @xmath6 is @xmath191 of the total metal density at all redshifts .
\(5 ) while gravitational shocks from large - scale structure formation dominate the energy budget ( @xmath11 ) for turning about @xmath126 of igm to the warm - hot intergalactic medium ( whim ) by @xmath12 , galactic superwind feedback shocks are energetically dominant over gravitational shocks at @xmath13 .
\(6 ) most of the so - called missing metals " at @xmath14 are hidden in a warm - hot gaseous phase ( @xmath155k ) that is heated up by star formation feedback shocks .
their mass distribution is broadly peaked at overdensity @xmath192 in the igm , outside virialized halos .
approximately @xmath17 of the total metals at @xmath12 are in ( stars , whim , x - ray gas , cold gas ) ; the distribution stands at @xmath18 and @xmath19 at @xmath20 and @xmath21 , respectively .
\(7 ) the metallicity of the igm with moderate overdensities ( @xmath192 ) that are probed by the @xmath35 forest shows a rapid increase with decreasing redshift .
we show that velocity `` diffusion '' effect that arises from the peculiar velocities could enhance the
apparent " metallicity of the @xmath35 forest clouds , as supported by our cross - correlation analysis .
tentatively , we suggest that this may reconcile , at least in part , the discrepancy between our simulations and observations at @xmath193 based on pixel optical depth ( pod ) method .
we are thankful to ben oppenheimer for useful conversations on the subject and an anonymous referee for a demanding but constructive report that helps significantly improve the paper .
computing resources were in part provided by the nasa high- end computing ( hec ) program through the nasa advanced supercomputing ( nas ) division at ames research center .
this work is supported in part by grants nnx08ah31 g and nas8 - 03060 . ,
s. , norberg , p. , baugh , c. m. , frenk , c. s. , bland - hawthorn , j. , bridges , t. , cannon , r. , colless , m. , collins , c. , couch , w. , cross , n. , dalton , g. , de propris , r. , driver , s. p. , efstathiou , g. , ellis , r. s. , glazebrook , k. , jackson , c. , lahav , o. , lewis , i. , lumsden , s. , maddox , s. , madgwick , d. , peacock , j. a. , peterson , b. a. , sutherland , w. , & taylor , k. 2001 , , 326 , 255 , e. , dunkley , j. , nolta , m. r. , bennett , c. l. , gold , b. , hinshaw , g. , jarosik , n. , larson , d. , limon , m. , page , l. , spergel , d. n. , halpern , m. , hill , r. s. , kogut , a. , meyer , s. s. , tucker , g. s. , weiland , j. l. , wollack , e. , & wright , e. l. 2009 , , 180 , 330 , m. , shimasaku , k. , okamura , s. , furusawa , h. , kashikawa , n. , ota , k. , doi , m. , hamabe , m. , kimura , m. , komiyama , y. , miyazaki , m. , miyazaki , s. , nakata , f. , sekiguchi , m. , yagi , m. , & yasuda , n. 2004 , , 611 , 660 | using the state - of - the - art cosmological hydrodynamic simulations of the standard cold dark matter model with star formation feedback strength normalized to match the observed star formation history of the universe at @xmath0 , we compute the metal enrichment history of the intergalactic medium ( igm ) .
overall we show that galactic superwind ( gsw ) feedback from star formation can transport metals to the igm and that the properties of simulated metal absorbers match current observations .
the distance of influence of gsw from galaxies is typically limited to about @xmath1mpc and within regions of overdensity @xmath2 .
most and absorbers are located within shocked regions of elevated temperature ( @xmath3k ) , overdensity ( @xmath2 ) , and metallicity ( @xmath4=[-2.5,-0.5]$ ] ) , enclosed by double shocks propagating outward .
absorbers have typically higher metallicity , lower density and higher temperature than absorbers . for absorbers collisional
ionization dominates over the entire redshift range @xmath0 , whereas for absorbers the transition occurs at moderate redshift @xmath5 from collisionally dominated to photoionization dominated .
we find that the observed column density distributions for and in the range @xmath6 are reasonably reproduced by the simulations .
the evolution of mass densities contained in and lines , @xmath7 and @xmath8 , is also in good agreement with observations , which shows a near constancy at low redshifts and an exponential drop beyond redshift @xmath9 . for both and most absorbers are transient and
the amount of metals probed by and lines of column @xmath6 is only @xmath10 of total metal density at any epoch .
while gravitational shocks from large - scale structure formation dominate the energy budget ( @xmath11 ) for turning about 50% of igm to the warm - hot intergalactic medium ( whim ) by @xmath12 , gsw feedback shocks are energetically dominant over gravitational shocks at @xmath13 .
most of the so - called missing metals " at @xmath14 are hidden in a warm - hot ( @xmath15k ) gaseous phase , heated up by gsw feedback shocks .
their mass distribution is broadly peaked at @xmath16 in the igm , outside virialized halos .
approximately @xmath17 of the total metals at @xmath12 are in ( stars , whim , x - ray gas , cold gas ) ; the distribution stands at @xmath18 and @xmath19 at @xmath20 and @xmath21 , respectively .
= 1 |
since its introduction by lutwak in @xcite , @xmath0 affine surface area ( defined below ) has become a fundamental concept in the @xmath0 brunn minkowski theory and has appeared in a growing number of works ( e.g. , ludwig @xcite , paouris & werner @xcite , werner & ye @xcite , and ye @xcite ) .
different approaches to @xmath0 affine surface area have been discussed , e.g. , meyer & werner @xcite , schtt & werner @xcite , and werner @xcite .
characterization theorems for @xmath0 affine surface area have been given , e.g. , haberl & parapatits @xcite , and ludwig & reitzner @xcite .
its relation to pde was explored , e.g. , lutwak & oliker @xcite .
connections between @xmath0 affine surface area and information theory were discovered in werner @xcite . for the @xmath1 case ,
@xmath0 affine surface area is an older notion usually referred to simply as affine surface area .
results here are even more numerous .
different approaches to this notion include leichtwei @xcite , lutwak @xcite , meyer & werner @xcite , schtt & werner @xcite , and werner @xcite
. a characterization of affine surface area was given in ludwig & reitzner @xcite .
connections between affine surface area and the affine plateau problem were discussed , e.g. , trudinger & wang @xcite .
it is also not surprising to see the appearance of affine surface area in polytopal approximation , e.g. , brny @xcite , brczky @xcite , and gruber @xcite .
perhaps , most importantly , it is the crucial ingredient in fundamental affine isoperimetric inequalities and gives rise to some affine analytic inequalities , e.g. , artstein - avidan , klartag , schtt & werner @xcite , caglar & werner @xcite , and lutwak @xcite .
let @xmath2 be a convex body ( compact convex set with non - empty interior ) .
the curvature measures of @xmath3 are a list of @xmath4 borel measures defined on the boundary of @xmath3 that can be defined via local parallel sets ( see e.g. , chapter 4 in @xcite ) . in the current paper
, the relationship between @xmath0 affine surface area and curvature measures will be explored . as a result
, a new representation of the existing notion of @xmath0 affine surface area using only curvature measures will be derived .
new proofs of the important properties of @xmath0 affine surface area , such as the upper semi - continuity and the @xmath0 affine isoperimetric inequality , will be given using the new representation .
the proofs given will be ones requiring no prior knowledge of the properties already established using other definitions of @xmath0 affine surface area .
it is also the aim of this paper to investigate the relationship between the new representation of @xmath0 affine surface area and the three existing ones .
this will be done by providing direct proofs of equivalence between the new representation and those previously given .
it will become apparent that the new form of @xmath0 affine surface area is , in a sense , `` polar '' to that of lutwak s and `` dual '' to schtt & werner s . in order to establish the equivalence ,
the lipschitz property of a sequence of restrictions of the inverse gauss map will be investigated , which may be of independent interest .
this will be discussed in section [ sec_lip ] .
it is important to note that it is the attempt of the current paper to give a new representation of the usual @xmath0 affine surface area ( not to define a new ( different ) @xmath0 affine surface area ) .
the notion of affine surface area traces back to affine differential geometry . in affine differential geometry , the affine surface area of a convex body @xmath3 with sufficiently smooth boundary ( at least @xmath5 ) and
everywhere positive gauss curvature is given by @xmath6 where @xmath7 is the gauss curvature of @xmath3 at @xmath8 ( the boundary of @xmath3 ) and @xmath9 is @xmath10 dimensional hausdorff measure .
when @xmath3 has @xmath5 boundary with positive gauss curvature , the gauss map @xmath11 is nice enough to allow the change of variable @xmath12 and we get : @xmath13 here @xmath14 is the curvature function of @xmath3 .
a very important result in affine differential geometry is the affine isoperimetric inequality which characterizes ellipsoids . for a convex body @xmath2 with @xmath5 boundary and positive gauss curvature , @xmath15 with equality if and only if @xmath3 is an ellipsoid . here
@xmath16 is the volume of @xmath3 and @xmath17 is the volume of the @xmath4-dimensional unit ball . extending the definition of affine surface area to one that works for general convex bodies ( without smoothness assumptions ) and still respects the basic properties of the classical definition was of huge interest during the late 80s and 90s ( in the previous century ) . in particular , is there a way to do the extension so that the affine isoperimetric inequality , with the same equality conditions , still holds ?
the first attempt was made by petty @xcite .
he observed that makes sense for convex bodies that possess curvature functions .
the affine isoperimetric inequality was also shown to hold under this extension .
although the gauss curvature and the curvature function do not necessarily exist for general convex bodies , the generalized gauss curvature and the generalized curvature function ( see @xcite or section [ section_preli ] ) exist almost everywhere on @xmath18 and @xmath19 with respect to @xmath10 dimensional hausdorff measure . thus and already suggest two possible extensions .
but , the two extensions are not trivial at all , since the two integrals might not make sense . with the notion of floating body ,
leichtwei was able to give a geometric meaning to when @xmath20 is the generalized curvature function .
let @xmath2 be a convex body .
the affine surface area of @xmath3 is given by @xmath21 where @xmath20 is the generalized curvature function .
one is tempted to use a strategy similar to how we arrived at to get to work for general convex bodies .
this turns out to be invalid since neither the gauss map nor the inverse gauss map , in this case , is smooth enough ( in fact , not even lipschitz ) to permit such a change of variable . in spite of this unfortunate fact
, there is a natural extension to .
schtt & werner @xcite , via the notion of convex floating body , were able to give a geometric meaning to the integral representation with @xmath22 being the generalized gauss curvature .
let @xmath2 be a convex body .
the affine surface area of @xmath3 is given by @xmath23 where @xmath22 is the generalized gauss curvature .
one of the major characteristics that distinguishes affine surface area ( look at either or ) and other geometric invariants is that affine surface area is not continuous with respect to the hausdorff metric .
for example , any convex body can be approximated by polytopes , but polytopes are always of zero affine surface area .
given this fact , only upper semi - continuity can be expected .
but , even establishing the upper semi - continuity of classical affine surface area ( in the smooth case ) was unsolved in the 80s .
one of the difficulties of establishing this lies in the lack of knowledge of the limit behaviors of the gauss curvature and the curvature function . since and take similar formulations , difficulty persists . note that the lack of continuity also adds to the difficulty of establishing the affine isoperimetric inequality , since we can not establish the inequality for a dense class of convex bodies and then take a limit .
the long conjectured upper semi - continuity of classical affine surface area was settled by lutwak . in @xcite , he found the following characterization of affine surface area for general convex bodies .
let @xmath2 be a convex body .
the affine surface area of @xmath3 is given by @xmath24 where the infimum is taken over all positive continuous functions @xmath25 , and @xmath26 is the surface area measure of @xmath3 .
note that lutwak s definition applies to all convex bodies ( even the ones without smoothness assumptions ) .
more importantly , the upper semi - continuity of follows directly from the weak continuity of the measures in the integrals and the fact that the infimum of a class of continuous functionals is upper semi - continuous . since agrees with classical affine surface area for smooth convex bodies with everywhere positive curvature function ( also shown in @xcite ) , this , in turn , proves the upper semi - continuity of classical affine surface area .
as shown in @xcite , the affine isoperimetric inequality can also be established in this case by using the blaschke santal inequality . a natural question to ask is : are definitions , , and
as was explained earlier , the equivalence of and is by no means a trivial problem for general convex bodies .
it was not until 1993 that schtt , in @xcite , proved the equivalence using a somewhat indirect method .
a direct proof was given in @xcite by hug .
that and are equivalent was shown by leitchwei in @xcite .
note that in lutwak s original definition , the infimum is only taken over all positive functions @xmath27 such that @xmath28 .
the removal of this restriction was proposed by leichtwei in @xcite and the equivalence between this formulation and lutwak s original definition was shown by dolzmann & hug in @xcite using a topological argument .
note that it is trivial to see that affine surface area is translation invariant .
schtt , in @xcite , proved that affine surface area is a valuation .
hence , affine surface area is an upper semi - continuous valuation that is invariant under volume preserving affine transformations . in a landmark work ,
ludwig & reitzner @xcite established the `` converse '' : if a real - valued upper semi - continuous valuation on the set of convex bodies is invariant under volume preserving affine transformations , then it must be of the form @xmath29 with @xmath30 and @xmath31 . here
@xmath32 is the euler characteristic , @xmath33 is volume , and @xmath34 is affine surface area . observe that and are `` polar '' to each other in the sense that one is defined as an integral over the boundary of the convex body ( domain of the gauss map ) , while the other is defined as an integral over the unit sphere ( image of the gauss map ) .
in fact , as the proof provided by hug in @xcite indicates , definitions and are linked by the gauss map .
also note that is `` dual '' to as one can see in the proof in @xcite .
recall that for a convex body @xmath2 , curvature measures are a list of @xmath4 borel measures defined on @xmath18 that can be defined via local parallel sets . for details ,
the reader should consult chapter 4 in @xcite . among these curvature measures ,
the @xmath35-th and @xmath10-th curvature measures have stronger geometric meanings .
more specifically , for a borel set @xmath36 , the @xmath35-th curvature measure @xmath37 and the @xmath10-th curvature measure @xmath38 of @xmath39 are given by @xmath40 where @xmath41 is the set of outer unit normals of @xmath3 at points in @xmath39 .
note that previous formulations of affine surface area involve @xmath10 dimensional hausdorff measure on @xmath18 ( definition ) , @xmath10 dimensional hausdorff measure on @xmath19 ( definitions and ) , and the surface area measure of @xmath3 ( definition ) .
since curvature measures are also a crucial type of measures associated to a convex body @xmath3 , it is natural to study the relationship between affine surface area and the curvature measures .
it is the purpose of this paper to investigate this missing element . to be precise
, the following theorem will be proved : [ thm_new_1 ] for a convex body @xmath2 , @xmath42 where the infimum is taken over all positive continuous functions @xmath43 . in light of theorem [ thm_new_1 ] , we may view the right side of as a new representation of the existing notion of affine surface area .
let @xmath2 be a convex body .
the affine surface area of @xmath3 can be defined by @xmath44 where the infimum is taken over all positive continuous functions @xmath43 .
note again that this form of affine surface area uses only curvature measures .
a list of properties of will be given in section [ sec_properties ] .
in particular , among other things , the upper semi - continuity ( theorem [ prop_new_1 ] ) and the affine isoperimetric inequality with equality condition ( theorem [ prop_new_2 ] ) will be demonstrated using the new representation .
the author would like to point out that although previously established , none of the results proved in section [ sec_properties ] require acknowledgement of properties proved under the existing forms . with the recent development of the @xmath0 brunn - minkowski theory , efforts were also made to generalize affine surface area to its @xmath0 analogue .
one of the key findings in the @xmath0 brunn - minkowski theory is the @xmath0 curvature function discovered by lutwak @xcite . given a convex body
@xmath3 that possesses a curvature function @xmath20 , the @xmath0 curvature function may be defined by @xmath45 with @xmath46 being the support function of @xmath3 .
since the generalized curvature function exists almost everywhere for an arbitrary convex body , the generalized @xmath0 curvature function also exists almost everywhere .
given this notion , @xmath0 affine surface area can be defined .
the notion of @xmath0 affine surface area was introduced by lutwak in @xcite .
he defined the @xmath0 affine surface area of a convex body @xmath2 that contains the origin in its interior to be , @xmath47 where the infimum is taken over all positive continuous functions @xmath25 .
although lutwak originally presented this definition for the case @xmath48 , it works perfectly fine for any @xmath49 as observed by hug in @xcite . as with the classical @xmath1 case ,
different forms of @xmath0 affine surface area exist . for each @xmath50 , the following form of @xmath0 affine surface area
was given by lutwak @xcite for convex bodies that possess a continuous curvature function and by hug @xcite for general convex bodies , @xmath51 analogously , for each @xmath50 , another form of @xmath0 affine surface area , was given by hug in @xcite , @xmath52 note that is the @xmath0 extension of while is the @xmath0 extension of .
the equivalence of and was proved by hug in @xcite .
that and are equivalent for convex bodies that possess a positive continuous curvature function was due to lutwak @xcite , and can in general be proved in a similar way as leitchwei did in @xcite for @xmath1 as pointed out in @xcite .
it is also possible to prove the analogue of theorem [ thm_new_1 ] in the @xmath0 setting .
namely , [ thm_new_2 ] let @xmath50 be a real number . for each convex body @xmath2 that contains the origin in its interior , @xmath53 where
the infimum is taken over all positive continuous functions @xmath43 .
we will be working mainly in @xmath54 with the canonical inner product @xmath55 .
the usual euclidean @xmath56-norm will be denoted by @xmath57 and the open ( resp
. closed ) ball of radius @xmath58 , which is centered at @xmath59 , will be denoted by @xmath60 ( resp .
we write @xmath17 for the volume of the @xmath4-dimensional unit ball . for a subset @xmath62 ,
we will write @xmath63 and @xmath64 for the closure of @xmath65 and the complement of @xmath65 , respectively . the characteristic function of @xmath66 , for any set @xmath66 , is written as @xmath67 .
a subset @xmath3 of @xmath54 is called a _ convex body _ if it is a compact convex set with non - empty interior .
the set of all convex bodies that contain the origin in the interior is denoted by @xmath68 .
the boundary of @xmath3 will be written as @xmath18 .
for an integer @xmath69 , we will write @xmath70 for @xmath71 dimensional hausdorff measure .
if @xmath72 is a measure on a topological space @xmath73 and @xmath74 is @xmath72 measurable , the restriction of @xmath72 to @xmath65 will be denoted by @xmath75 .
associated to each convex body @xmath3 is the support function @xmath76 given by @xmath77 for each @xmath78 .
the supporting hyperplane @xmath79 of @xmath3 for each @xmath80 is given by @xmath81 at each boundary point @xmath8 , a unit vector @xmath82 is said to be an _ outer unit normal _ of @xmath3 at @xmath59 if @xmath79 passes through @xmath59 . for a subset @xmath36 , the spherical image , @xmath83 , of @xmath3 at @xmath39 ,
is the set of all outer unit normal vectors of @xmath3 at points in @xmath39 .
a boundary point @xmath59 is _ regular _ if @xmath84 contains exactly one point in @xmath19 .
denote by @xmath85 the set of all regular boundary points of @xmath3 .
the _ gauss map _
, @xmath86 is the map that takes each regular boundary point to the unique outer unit normal of that point .
similarly , for each subset @xmath87 , the inverse spherical image , @xmath88 , of @xmath3 at @xmath89 , is the set of all boundary points of @xmath3 that have outer normal vectors in @xmath89 . a unit vector @xmath82 is _ regular _ if @xmath90 contains exactly one point in @xmath18 .
denote by @xmath91 the set of all regular normal vectors . the _ inverse gauss map _
, @xmath92 is the map that takes each regular normal vector to the unique point in @xmath90 .
both the gauss map and the inverse gauss map are continuous ( see lemma 2.2.12 in @xcite ) .
let @xmath93 be a convex function .
the set @xmath94 is called the _ subdifferential _ of @xmath95 at @xmath59 . if @xmath96 satisfies @xmath97 for each @xmath78
, then it is called a _ subgradient choice _ of @xmath95 .
moreover , @xmath95 is differentiable at @xmath59 if and only if @xmath98 contains only @xmath99 , the gradient of @xmath95 at @xmath59 .
the following notion of second order differentiability is useful .
we say @xmath95 is _ second order differentiable _ at @xmath100 in the generalized sense if @xmath95 is differentiable at @xmath100 in the classical sense and there exists a symmetric linear map @xmath101 such that @xmath102 for every @xmath103 .
it follows from @xcite that a convex function @xmath95 is _ second order differentiable _ at @xmath100 in the generalized sense if and only if there exists a neighborhood @xmath33 of @xmath100 and a symmetric linear map @xmath104 such that @xmath105 for all @xmath106 and all subgradient choices @xmath107 .
note that the generalized second order differentiability is a local property .
hence , the above notion extends naturally to the case where @xmath95 is only defined on an open subset of @xmath54 .
+ for a regular boundary point @xmath108 , suppose @xmath109 .
the _ tangent space _ , @xmath110 , of @xmath3 at @xmath100 is the linear subspace @xmath111 .
write @xmath112 for the orthogonal projection of @xmath113 to @xmath110 for each @xmath103 .
we can choose a number @xmath114 and a neighborhood @xmath115 of @xmath100 such that for each @xmath116 , @xmath117 where @xmath118 and @xmath119 is a convex function satisfying @xmath120 and @xmath121 .
we say a regular boundary point @xmath100 is _ normal _ if @xmath95 in is second order differentiable at @xmath122 in the generalized sense .
denote by @xmath123 the set of all normal boundary points of @xmath3 . with a proper choice of orthonormal basis @xmath124 satisfying @xmath125 and @xmath126 , it is possible to write @xmath95 as : @xmath127 where @xmath128 are the coordinates of @xmath59 under @xmath129 . here , @xmath130 is called a generalized principal curvature while @xmath131 is the associated generalized principal direction , for @xmath132 . in this case , the generalized gauss curvature @xmath133 of @xmath3 at @xmath100 is given by @xmath134 it follows from the alexandrov theorem @xcite that @xmath135 hence , @xmath7 is defined for @xmath9 almost all @xmath8 .
the set @xmath136 is given by @xmath137 the support function @xmath46 is differentiable at @xmath138 if and only if @xmath139 is a regular normal vector . in this case ,
( see corollary 1.7.3 in @xcite . )
it was the result of the alexandrov theorem @xcite that every convex function @xmath93 is @xmath141 almost everywhere second order differentiable in the generalized sense .
in particular , the support function @xmath46 is @xmath142 almost everywhere second order differentiable in the generalized sense .
denote by @xmath143 the set of all points at which @xmath46 is second order differentiable in the generalized sense .
the following properties can be easily seen from the homogeneity of @xmath46 : 1 .
if @xmath144 , then @xmath145 , for all @xmath146 . hence @xmath147 2 .
given @xmath144 , we have that @xmath139 is an eigenvector of @xmath148 with 0 being the associated eigenvalue . the fact that @xmath148 is symmetric tells us that @xmath149 is an invariant subspace of @xmath148 .
the generalized curvature function of @xmath3 at @xmath150 , denoted by @xmath151 , is defined to be the determinant of @xmath152 .
note that @xmath151 is defined for @xmath9 almost all @xmath80 .
the set @xmath153 is given by @xmath154 the surface area measure @xmath26 of a convex body @xmath3 is a borel measure on @xmath19 and is given by @xmath155 for each borel set @xmath87 .
recall that the @xmath35-th curvature measure @xmath156 and the @xmath10-th curvature measure @xmath157 are borel measures on the boundary of @xmath3 and are given by @xmath158 for each borel set @xmath36 .
it is obvious that @xmath156 and @xmath157 are finite measures .
the @xmath35-th curvature measure @xmath156 has the following decomposition ( see e.g. , hilfssatz 3.6 in @xcite or ( 2.7 ) in @xcite ) : for each borel set @xmath36 , @xmath159 where @xmath160 is a borel set and @xmath161 .
in particular , one has @xmath162 curvature measures are weakly continuous with respect to the hausdorff metric ( see @xcite ) .
the following definitions are needed for federer s coarea formula .
see @xcite for details .
a subset @xmath89 of @xmath54 is said to be _
rectifiable _ if @xmath164 and there exists @xmath165 such that @xmath166 is bounded , @xmath167 is lipschitz and @xmath168 .
let @xmath169 be a non - empty subset of @xmath54 .
the _ tangent cone _ of @xmath169 at a given point @xmath170 , denoted by @xmath171 , can be defined as the set of @xmath172 such that for every @xmath114 there exists @xmath173 and @xmath174 with @xmath175 and @xmath176 .
suppose @xmath72 is a measure on @xmath54 .
the _ @xmath10 dimensional density _
@xmath177 at @xmath170 is given by @xmath178 if the limit exists .
the _ @xmath179 approximate tangent cone _
@xmath180 at @xmath181 is given by @xmath182 suppose @xmath95 maps a subset of @xmath54 into @xmath54 .
we say that @xmath95 is _
@xmath183 approximately differentiable _ at @xmath181
if and only if there exists @xmath184 and a continuous linear map @xmath185 such that @xmath186 for every @xmath114 . in this case , the _ @xmath183 approximate differential _ of @xmath95 at @xmath181 , denoted by @xmath187 , is given by @xmath188 suppose @xmath189 are two @xmath10-dimensional hilbert spaces .
let @xmath190 and @xmath191 be the @xmath10th exterior power of @xmath33 and @xmath192 equipped with the induced inner products from @xmath33 and @xmath192 respectively .
every linear map @xmath193 induces a map @xmath194 . by @xmath195 , we mean the operator norm of @xmath196 .
note that @xmath195 is just the absolute value of the determinant of @xmath95 when @xmath197 .
see chapter 1 in @xcite for details . when @xmath72 is the restriction of @xmath9 to some @xmath9 measurable and @xmath163 rectifiable subset of @xmath54 , by theorem 3.2.19 in @xcite , the approximate tangent cone @xmath180 is an @xmath10 dimensional subspace of @xmath54 for @xmath9 almost all @xmath181 in that subset . in this case
( when @xmath198 is an @xmath10 dimensional subspace of @xmath54 ) , we call @xmath199 the _ @xmath183 approximate jacobian _ of @xmath95 at @xmath181 and denote it by @xmath200 .
the following is a special case of federer s coarea formula ( * ? ? ?
* theorem 3.2.22 ) .
note that the original theorem works for any non - negative measurable function by the obvious application of the monotone convergence theorem .
[ coarea formula ] suppose @xmath201 are @xmath9 measurable and @xmath163 rectifiable . if @xmath202 is lipschitz , then for each @xmath203 measurable non - negative function @xmath204 on @xmath192 , @xmath205 it is implied in theorem [ coarea formula ] that @xmath206 is @xmath9 measurable as a function in @xmath207 .
in this section , we will prove the promised theorem [ thm_new_2 ] , which will reveal the relationship between @xmath0 affine surface area and curvature measures .
notice that theorem [ thm_new_1 ] follows by setting @xmath1 in theorem [ thm_new_2 ] and the obvious fact that both sides of are translation invariant .
the following notations will be needed .
let @xmath208 note that when @xmath209 , the sets @xmath210 and @xmath211 have the following relationship : @xmath212 recall that @xmath136 is the set of normal boundary points with positive gauss curvature ( see ) .
for the sake of simplicity , let us introduce the following notations : @xmath213 note that @xmath214 is defined for @xmath215 , while @xmath216 is defined for any positive borel measurable function @xmath217 satisfying @xmath218 . for each @xmath215 , by hlder
s inequality , @xmath219 for each @xmath220 , by , @xmath221 this , the fact that @xmath222 for each borel set @xmath36 , , and imply that , @xmath223 for each @xmath220 .
equations , , and the fact that @xmath224 show @xmath225 to complete the proof , let us now show @xmath226 it suffices , for every @xmath114 , to find a @xmath220 such that @xmath227 let @xmath228 be defined by @xmath229 note that for each @xmath9 measurable subset @xmath230 , there exists a borel measurable set @xmath231 such that @xmath232 and @xmath233 .
this and the fact that @xmath234 is a borel set ensure that we can modify the value of @xmath235 on a subset @xmath236 with @xmath237 , such that the resulting function is a borel measurable function . denote the resulting function by @xmath238 .
clearly @xmath239 for any @xmath240 and @xmath241 for @xmath9 almost all @xmath8 .
define @xmath242 note that both @xmath238 and @xmath243 are borel measurable , and @xmath244 . by the fact that both @xmath156 and @xmath157 are finite measures , the assumption that @xmath3 contains the origin in its interior , ,
the fact that @xmath222 for each borel set @xmath36 , the choice of @xmath238 , and @xmath161 , we can compute the following limit , @xmath245 by , and the fact that @xmath246 , @xmath247 , @xmath248 equations and imply that there exists @xmath249 such that @xmath250 note that @xmath251 .
let @xmath252 .
note that @xmath72 is a finite positive borel measure on a compact metric space .
hence @xmath72 is regular . since @xmath253 , by lusin s theorem ( see theorem 2.24 in @xcite ) , there exists @xmath254 such that @xmath255 let @xmath256 it is easy to see that @xmath257 is still continuous and @xmath258 .
moreover , since whenever @xmath259 , it must be the case that @xmath260 , we have by , @xmath261 hence , @xmath262 this implies that @xmath263 . as a result , there exists @xmath264 such that @xmath265 choose @xmath266 . by and
, such a @xmath204 will satisfy .
it is immediate from and that @xmath267 theorem [ thm_new_2 ] suggests , in addition to , , and , there is a new representation of the existing notion of @xmath0 affine surface area : let @xmath50 be a real number and @xmath2 be a convex body that contains the origin in its interior .
the @xmath0 affine surface area of @xmath3 can be defined by @xmath268 where the infimum is taken over all positive continuous functions @xmath43 .
note that is the @xmath0 analogue of .
it is worthwhile to point out that , as can be seen from the proofs in this section , definition is `` dual '' to definition . since it has already been established that definitions , , are the same , it follows from theorem [ thm_new_2 ] that definition is equivalent to definition .
however , we wish to give a direct proof of the equivalence between definitions and as this will further reveal the relationship between the two formulations of @xmath0 affine surface area .
this will be carried out in section [ sec_equivalences ] .
in this section , some basic properties of @xmath0 affine surface area will be shown using the new representation ( if @xmath1 ) .
although the properties given in this section are not new , different proofs are given using definition that do not depend on any results about @xmath0 affine surface area established using the previously known forms .
the following proposition shows that @xmath269 is a homogeneous functional on the set of convex bodies containing the origin in their interiors .
[ prop_tmp 100 ] let @xmath50 be a real number .
suppose @xmath209 . for @xmath270 , @xmath271 .
for each positive continuous @xmath204 on @xmath18 , define @xmath272 on @xmath273 by @xmath274 notice that for each borel set @xmath275 , @xmath276 by and , @xmath277 by the homogeneity of the support function , @xmath278 by , , and , we get the desired result .
it is trivial to see that when @xmath1 , translation invariance is satisfied by @xmath279 .
let @xmath50 be a real number .
suppose @xmath3 is polytope that contains the origin in its interior .
then , @xmath280 note that the measure @xmath156 in this case is concentrated on the set of vertices of @xmath3 , which has only finitely many points .
this implies that we can let the second integral in be arbitrarily small while holding the value of the first integral constant .
hence @xmath281 .
another important property of @xmath0 affine surface area that historically took a long time to prove ( settled in @xcite ) is its upper semi - continuity .
[ prop_new_1 ] let @xmath50 be a real number .
then , @xmath269 is upper semi - continuous with respect to the hausdorff metric .
this is a direct result from the weak continuity of @xmath156 , @xmath157 and the fact that the infimum of continuous functionals is upper semi - continuous . by
, the following variant of will give us more flexibility in choosing the function @xmath204 : @xmath282 we will need the following two lemmas , which were established by schtt & werner in @xcite .
let @xmath209 and @xmath283 be a linear map with determinant either @xmath284 or @xmath285 .
for each integrable function @xmath286 , @xmath287 let @xmath209 and @xmath283 be a linear map with determinant either @xmath284 or @xmath285 .
suppose @xmath59 is a normal boundary point of @xmath3 .
then @xmath288 is a normal boundary point of @xmath289 and moreover , @xmath290 the next proposition shows that @xmath269 is invariant under volume preserving linear transformations .
[ prop_tmp 101 ] let @xmath50 be a real number
. suppose @xmath209 and @xmath291 is a linear map with determinant either @xmath284 or @xmath285 .
then @xmath292 note that @xmath293 and @xmath294 .
thus , we only need to show @xmath295 .
if @xmath59 is a regular boundary point of @xmath3 , then @xmath296 is a regular boundary point of @xmath289 and @xmath297 hence , @xmath298 the definitions of support function and outer unit normal , together with and , imply @xmath299 let @xmath215 be a function defined on @xmath18 such that @xmath300 . by , , and , @xmath301 by , , and , @xmath302 let @xmath303 be defined as @xmath304 the fact that @xmath9 almost all points on the boundary of a convex body are regular , the choice of @xmath204 , and show that @xmath272 is a positive , @xmath9 measurable function on @xmath305 and @xmath306 by , , and , we immediately have @xmath307 .
an immediate corollary of proposition [ prop_tmp 100 ] and proposition [ prop_tmp 101 ] is : let @xmath50 be a real number .
suppose @xmath209 and @xmath283 is an invertible linear map .
then @xmath308 the @xmath0 affine surface area functional is also a valuation .
that is , if @xmath309 are such that @xmath310 , then @xmath311 this property , however , is not immediate under the new form .
the reader is recommended to see e.g. , schtt @xcite , ludwig & reitzner @xcite ( and the references therein ) for a proof of the valuation property of @xmath0 affine surface area and the role of @xmath0 affine surface area in the theory of valuation .
for each @xmath215 , denote @xmath312 notice that for a convex body @xmath313 that contains the origin in its interior , @xmath314 here @xmath315 is the @xmath0 mixed volume of @xmath3 and @xmath313 and can be defined by @xmath316 built in is the following @xmath0 mixed volume type inequality , @xmath317 we will postpone the proof of the @xmath0 affine isoperimetric inequality ( the @xmath0 analogue of the celebrated affine isoperimetric inequality , which was given by lutwak in @xcite for @xmath318 , and by werner & ye in @xcite for all other @xmath319 ) to section [ sec_equivalences ] . the proof will utilize .
recall that @xmath321 is the gauss map defined on @xmath85 , the set of regular boundary points of @xmath3 , and @xmath320 is the inverse gauss map defined on @xmath91 , the set of regular normal vectors of @xmath3 .
see section [ section_preli ] .
one of the essential difficulties people encounter when trying to prove the equivalence between definitions and , as well as definitions and , is linking an integral over the unit sphere ( image of the gauss map or domain of the inverse gauss map ) with an integral over the boundary of a convex body ( domain of the gauss map or image of the inverse gauss map ) .
one of the direct bridges was built in @xcite by exploring the lipschitz property of restrictions of @xmath321 .
in particular , for each @xmath174 and each convex body @xmath2 , denote @xmath322 the following lemma was shown in @xcite : [ lemma 4.1 ] for each convex body @xmath2 , @xmath323 denote @xmath324 by @xmath325 . hug in ( * ? ? ?
* lemmas 2.1 , 2.3 ) showed that @xmath325 is a lipschitz map and calculated the approximate jacobian of @xmath325 : [ lemma 4.2 ] let @xmath2 be a convex body and @xmath174 .
the following results are true : 1 .
@xmath326 is a closed subset of @xmath18 , and @xmath325 is lipschitz ; 2 . for @xmath9 almost all @xmath327
, we have @xmath328 the following characterization of points at which the generalized curvature function is positive was established in ( * ? ? ? * lemma 2.7 ) : [ lemma 4.3 ] let @xmath329 be a convex body .
suppose @xmath330 .
then the following two conditions are equivalent : 1 .
there is some @xmath174 such that @xmath331 .
2 . @xmath332 . in this section
, the lipschitz property of restrictions of @xmath320 will be discussed , which will be useful in proving the equivalence between definitions and . in particular
, we will divide the unit sphere into a countable collection of subsets ( up to a set of measure 0 ) , such that @xmath320 restricted to each subset is lipschitz .
thus a change of variable will be made possible by using federer s coarea formula .
let @xmath2 be a convex body .
for each @xmath333 , we define @xmath334 by @xmath335\}.\ ] ] here @xmath90 is the inverse spherical image of @xmath336 ( see section [ section_preli ] ) .
note that it is easily seen that if @xmath337 , then @xmath82 must be a regular normal vector of @xmath3 .
hence , @xmath338 , \text{where } x=\tau_k(u)\}.\ ] ] we will denote the restriction of @xmath320 to @xmath339 by @xmath340 , for each @xmath333 . the following lemma was observed by hug in @xcite , which can be proved similarly to lemma 2.7 in @xcite .
[ lemma_a.e .
] let @xmath2 be a convex body . for each @xmath330
, there exists @xmath333 such that @xmath341,\ ] ] where @xmath342 .
the following corollary follows immediately from lemma [ lemma_a.e . ] and .
[ corollary_a.e .
] let @xmath2 be a convex body .
with respect to @xmath9 , almost every point in @xmath19 is contained in @xmath343 , i.e. , @xmath344 the following lemma will be needed .
[ lemma_closedness ] let @xmath2 be a convex body . for each @xmath333 , @xmath339 is closed .
suppose @xmath345 is a convergent sequence in @xmath339 .
denote by @xmath139 its limit .
let @xmath346 . since @xmath18 is compact , we can take a convergent subsequence of @xmath347 .
we denote the subsequence and the corresponding subsequence of @xmath345 again by @xmath347 and @xmath345 .
denote by @xmath100 the limit of @xmath347 .
let @xmath348 be the hyperplane that passes @xmath100 and has @xmath139 as its normal .
since @xmath346 , @xmath349 for each @xmath350 .
let @xmath351 , we have @xmath352 for each @xmath350 .
hence @xmath348 is a supporting hyperplane of @xmath3 at @xmath100 and @xmath353 . since @xmath354 , we have @xmath355 $ ] .
let @xmath351 .
we have @xmath356 $ ] , where @xmath353 .
this implies that @xmath357 .
hence @xmath339 is closed .
[ lemma_tau lip ] for each @xmath333 and each convex body @xmath2 , @xmath340 is lipschitz . let @xmath358 .
denote by @xmath359 $ ] the angle formed by @xmath360 and @xmath361 .
let @xmath362 , and @xmath363 .
we first assume @xmath364 .
suppose @xmath365 . otherwise , there is nothing to prove .
since @xmath366 are not parallel to each other , the points @xmath367 and @xmath368 can not both lie on the line passing @xmath369 .
suppose @xmath367 does not lie on the line passing @xmath369 .
denote @xmath367 by @xmath370 . without loss of generality ,
since this lemma is invariant under translation , we may assume that @xmath370 is the origin .
let @xmath371 be the two dimensional subspace spanned by @xmath372 and @xmath373 .
note that @xmath360 is parallel to @xmath372 .
write @xmath374 where @xmath375 and @xmath376 . since @xmath360 and @xmath361 are not perpendicular , we have @xmath377 .
let @xmath378 .
denote by @xmath379 $ ] the angle formed by @xmath360 and @xmath380 .
notice that by the definition of @xmath380 , , and the fact that @xmath381 , @xmath382 this implies that if @xmath364 , we have @xmath383 . in this case , by and that @xmath384 , we have @xmath385 . by the monotonicity of the cosine function on @xmath386 , we have @xmath387 when @xmath388 . by the definition of @xmath339 , the fact that @xmath389 implies that @xmath390 is a subset of the disc @xmath391 $ ] .
note that @xmath390 is non - empty and is either the line segment joining @xmath372 and @xmath373 or a convex body in @xmath371 . in either case , for any @xmath392 , by the definition of @xmath380 , , the fact that @xmath363 , , and the definition of @xmath380 once again , @xmath393 now , we show that @xmath394 as a result of @xmath365 . if otherwise , @xmath395 . since @xmath396 $ ]
, the line @xmath397 passing @xmath372 and perpendicular to @xmath360 can only intersect @xmath390 at @xmath372 .
note also that since @xmath362 , we have @xmath398 for each @xmath392 . but
implies that @xmath399 .
hence @xmath400 and as a result , we have @xmath401 .
this immediately implies @xmath402 , which contradicts with our assumption .
thus , it suffices to look at the case when @xmath403 , with @xmath365 . in this case , let @xmath404 be the line passing through @xmath373 and is perpendicular to @xmath380 .
extend @xmath405 so that it intersects @xmath406 at @xmath407 . starting from @xmath370 ,
make an ray in the direction of @xmath380 ( perpendicular to @xmath406 as a result ) , so that it crosses the boundary of @xmath408 $ ] at @xmath66 .
let @xmath409 be the angles indicated in figure [ fig:1 ] . by and that @xmath410 , the point @xmath407 does not belong to the interior of @xmath411 .
hence @xmath412 .
the fact that @xmath413 $ ] , together with , implies @xmath414 observe that @xmath415 . since @xmath416 we conclude that there exists @xmath417 , such that @xmath418 for any @xmath358 , satisfying @xmath419 .
( note that trivially holds if @xmath402 . ) for the case @xmath420 , since @xmath421 and @xmath3 is bounded , we have @xmath422 is bounded from above .
this and prove the existence of @xmath423 such that @xmath424 for any @xmath358 . hence @xmath340 is lipschitz .
the following characterization of normal boundary points with positive curvature was shown in @xcite as corollary 3.2 ( see also @xcite ) .
[ lemma_h>0 ] let @xmath329 be a convex body and @xmath100 be a normal boundary point .
the following two conditions are equivalent : 1 .
[ enumerate_1]@xmath425 ; 2 .
[ enumerate_2]there exists @xmath426 such that @xmath356 $ ] where @xmath109 .
the following lemma was proved in ( * ? ? ?
* lemma 2.5 ) .
[ lemma_reciporical of h and f ] let @xmath2 be a convex body .
suppose @xmath100 is a normal boundary point and @xmath427
. then @xmath428 .
there is still one piece missing that hinders us from applying federer s coarea formula to @xmath340 , namely , the jacobian of @xmath340 .
taking the jacobian of @xmath340 in the classical sense is impossible since the classical jacobian requires @xmath340 to be defined in an open set .
therefore we have to consider the approximate jacobian of @xmath340 instead .
[ lemma_jacobian ]
let @xmath2 be a convex body .
for each given @xmath333 , @xmath429 for @xmath9 almost all @xmath337 .
observe that @xmath339 is an @xmath163 rectifiable and @xmath9 measurable subset ( since @xmath339 is closed by lemma [ lemma_closedness ] ) of @xmath430 . by theorem 3.2.19 in @xcite , for @xmath9 almost all @xmath357 , @xmath431 is an @xmath10 dimensional subspace of @xmath54 . by definition , @xmath432
hence , @xmath433 for @xmath9 almost all @xmath357 .
this and imply that @xmath9 almost all vectors in @xmath339 are in @xmath143 and satisfy .
thus , to prove this lemma , we may assume that @xmath144 and @xmath139 satisfies .
let @xmath114 be a real number and @xmath434 be an arbitrary subgradient choice of @xmath46 .
the fact that @xmath144 implies that there exists @xmath435 and a symmetric linear map @xmath436 such that @xmath437 for @xmath438 .
since each @xmath337 is a regular normal vector , we have @xmath439 for each @xmath337 . hence by the definition of @xmath440
, we have @xmath441 by the definition of @xmath442 and , we have @xmath443 by this , the definition of @xmath444 , that @xmath149 is an invariant subspace of @xmath148 , and the definition of the generalized curvature function , we have @xmath445
in this section , the relationship between definitions and will be unveiled by providing a direct proof of the equivalence between the two .
the proof utilizes the tool we established in section [ sec_lip ] . at the end of this section , the @xmath0 affine isoperimetric inequality ( with equality condition )
is established using the new representation of @xmath0 affine surface area .
let us first introduce some notations .
let @xmath446 note that @xmath447 .
we claim that for each positive real number @xmath319 and each @xmath209 , @xmath448 indeed , for each @xmath215 , let @xmath449 notice that @xmath450 on @xmath451 . by , , the choice of @xmath452 , and the fact that @xmath453 is a finite number , @xmath454 this shows that the left side is less than or equal to the right side in . by the fact that @xmath455 and that all integrands in are non - negative
, the left side is greater than or equal to the right side in . by , we can further show that [ lemma_h+ ] let @xmath50 be a real number .
suppose @xmath209 .
we have @xmath456 recall that @xmath153 is the set of unit vectors at which the curvature function @xmath20 is positive ( see ) . by a similar argument from above
, @xmath457 this , together with hilfssatz 2 and 3 in @xcite ( see also @xcite ) , immediately implies [ lemma_f+ ] let @xmath50 be a real number .
suppose @xmath209 .
we have @xmath458 now , we shall prove the equivalence between definitions and .
[ thm_new_3 ] let @xmath50 be a real number .
suppose @xmath209 .
we have @xmath459 where the infimums are taken over all positive continuous functions @xmath460 and @xmath43 respectively . by lemmas [ lemma_h+ ] and [ lemma_f+ ]
, it suffices to show @xmath461 let us first show that the left side is greater than or equal to the right side in .
let @xmath215 be such that @xmath462 and @xmath174 .
we first observe that @xmath326 and @xmath19 are @xmath163 rectifiable and @xmath9 measurable ( since @xmath326 is closed by lemma [ lemma 4.2 ] ) . by lemma [ lemma 4.2 ] ,
@xmath325 , the restriction of @xmath321 to @xmath326 , is lipschitz and @xmath463 for @xmath9 almost all @xmath327 .
the fact that @xmath325 is lipschitz and give , @xmath464 by and federer s coarea formula , @xmath465 it is implied in federer s coarea formula that @xmath466 is @xmath9 measurable on @xmath19 in @xmath82 . by , , lemma [ lemma_reciporical of h and f ] , and , once again , the following holds for @xmath9 almost all @xmath80 , @xmath467 hence @xmath468 , as a function of @xmath82 , is @xmath9 measurable on @xmath19 . by and , @xmath469 note that @xmath326 is increasing as @xmath470 .
let @xmath470 in .
it follows from the monotone convergence theorem , lemma [ lemma 4.1 ] , and lemma [ lemma 4.3 ] that , @xmath471 with the same technique , one can prove that @xmath472 set @xmath473 note that @xmath474 and @xmath27 is @xmath9 measurable on @xmath19 .
by , @xmath475 is @xmath9 integrable on @xmath19 .
hence @xmath476 . by ,
, the left side is greater than or equal to the right side in .
now we show that the left side is less than or equal to the right side in .
+ let @xmath476 be such that @xmath477 and @xmath333 .
we first observe that @xmath339 and @xmath18 are @xmath478 rectifiable and @xmath9 measurable ( since @xmath339 is closed by lemma [ lemma_closedness ] ) . by lemma [ lemma_tau lip ] , @xmath340 ,
the restriction of @xmath320 to @xmath339 , is lipschitz . by lemma [ lemma_jacobian ] , for @xmath9 almost all @xmath337 ,
@xmath479 the fact that @xmath340 is lipschitz and give , @xmath480 by and federer s coarea formula , @xmath481 it is implied in federer s coarea formula that @xmath482 is @xmath9 measurable on @xmath18 in @xmath59 . by , , lemma [ lemma_reciporical of h and f ] and , once again , the following holds for @xmath9 almost all @xmath483 , @xmath484 hence @xmath485 , as a function of @xmath59 , is @xmath9 measurable on @xmath18 . by and , @xmath486 note that @xmath339 is increasing as @xmath487 .
let @xmath487 in .
it follows from the monotone convergence theorem , corollary [ corollary_a.e . ] , and lemma [ lemma_h>0 ] that @xmath488 with the same technique , one can prove @xmath489 set @xmath490 note that @xmath491 and @xmath204 is @xmath9 measurable . by ,
@xmath492 is @xmath9 integrable on @xmath18 .
hence @xmath215 . by ,
, the left side is less than or equal to the right side in .
the proof given above reveals that definition is `` polar '' to and they are linked by the gauss map and the inverse gauss map . the @xmath0 affine isoperimetric inequality , which is the extension of the affine isoperimetric inequality of affine differential geometry , was first established by lutwak in @xcite .
thanks to , we are ready to prove the @xmath0 affine isoperimetric inequality using the new representation .
[ prop_new_2 ] let @xmath50 be a real number .
suppose @xmath209 and has the origin as its centroid .
we have , @xmath493 with equality if and only if @xmath3 is an ellipsoid .
recall from that , for each convex body @xmath313 that has the origin as its centroid , @xmath494 based on the change of variable formula established in the proof of theorem [ thm_new_3 ] , @xmath495 where @xmath496 is the polar body of @xmath313 . combining and ,
we get , @xmath497 it was observed in @xcite that is stronger than .
indeed , just by replacing @xmath313 in by @xmath3 and using the blaschke - santal inequality , we get .
equality holds only if the equality holds for the blaschke - santal inequality , i.e. , @xmath3 is an ellipsoid .
that the equality does hold for ellipsoids follows from a direct calculation .
hug in @xcite proved the equivalence between definitions and by applying federer s coarea formula to the lipschitz map @xmath325 , the restriction of @xmath321 to @xmath326 . here
we provide a dual proof of the same result by applying federer s coarea formula to the lipschitz map @xmath340 , the restriction of @xmath320 to @xmath339 , which is made possible by the discussion in section [ sec_lip ] .
0 alexandrov , a. d. `` almost everywhere existence of the second differential of a convex function and some properties of convex surfaces connected with it . '' _ leningrad state university annals [ uchenye zapiski ] math .
_ 6 ( 1939 ) : 335 .
artstein - avidan , s. , b. klartag , c. schtt , and e. werner .
`` functional affine - isoperimetry and an inverse logarithmic sobolev inequality . ''
_ journal of functional analysis _ 262 , no . 9
( 2012 ) : 4181204 .
petty , c. m. `` affine isoperimetric problems . '' in _ discrete geometry and convexity ( new york 1982 ) _ , edited by j. e. goodman , e. lutwak , j. malkevitch , and r. pollack , 11327 .
annals of the new york academy of sciences 440 .
new york : new york acad .
sci . , 1985 . | the relationship between @xmath0 affine surface area and curvature measures is investigated . as a result
, a new representation of the existing notion of @xmath0 affine surface area depending only on curvature measures is derived .
direct proofs of the equivalence between this new representation and those previously known are provided .
the proofs show that the new representation is , in a sense , `` polar '' to that of lutwak s and `` dual '' to that of schtt & werner s . |
thermocapillary migration of a droplet or bubble at a uniform temperature gradient is a very interesting topic in both microgravity and microfluidics@xcite . on the one hand , due to the development of space exploration ,
the studies of the physical mechanism of the droplet / bubble migration phenomena in the microgravity environment become more and more important .
the first investigation of thermocapillary droplet migration was completed by young et al ( ygb ) in 1959 , who gave an analytical prediction for the migration speed at zero limit reynolds(re ) and marangoni(ma ) numbers@xcite .
subramanian@xcite proposed the quasi - steady state assumption for thermocapillary bubble migration and obtained analytical results with high order expansion at small ma numbers .
balasubramanian and chai@xcite derived an exact result for thermocapillary droplet migration at small ma numbers .
the experimental results for the droplet migration speed at small re numbers obtained by braun et al @xcite are in agreement with those of the ygb model .
on the other hand , the manipulation and actuation of droplets in microfluidic devices have an extensive application on the chemical industry and biological engineering .
many interest problems , such as static contact and dynamic control of the droplets on solid substrates , were theoretically analyzed and experimentally observed@xcite . although the thermocapillary droplet migration processes at small ma numbers under the microgravity environment are understood very well in the series of theoretical analyses and experimental investigations , the physical behaviors at large ma numbers are rather complicated due to the momentum and energy transfer though the interface of two - phase fluids . in large re and ma number area
, it was reported@xcite that the migration speed of a droplet increases with the increase of ma number , as is in qualitative agreement with the corresponding numerical simulation@xcite .
the above theoretical analysis and numerical simulation are based on assumptions of the quasi - steady state and non - deformation of the droplet . the experimental investigation carried out by hadland
et al@xcite and xie et al@xcite showed that the droplet migration speed non - denationalized by the ygb velocity decreases with the increase of ma number .
the results are not in qualitative agreement with the above theoretical and numerical ones .
moreover , a nonconservative integral thermal flux across the surface in the steady thermocapillary droplet migration at large ma(re ) numbers was identified@xcite .
it was indicated that the thermocapillary droplet migration at large ma(re ) numbers is an unsteady process .
therefore , the thermocapillary droplet migration at large ma(re ) numbers remains a topic to be studied with respect to its physical mechanism . in this paper , first , under the assumption of the quasi - steady process of thermocapillary droplet migration at large re and ma numbers , a thermal source is added in the droplet to change the integral thermal flux across the surface from nonconservative to conservative .
then , we make an analytical result for the steady thermocapillary migration of droplet with the thermal source at large re and ma numbers and elucidate effects of the adding thermal source in the droplet on thermocapillary migration process .
consider the thermocapillary migration of a spherical droplet of radius @xmath0 , density @xmath1 , dynamic viscosity @xmath2 , thermal conductivity @xmath3 , and thermal diffusivity @xmath4 in a continuous phase fluid of infinite extent with density @xmath5 , dynamic viscosity @xmath6 , thermal conductivity @xmath7 , and thermal diffusivity @xmath8 under a uniform temperature gradient @xmath9 .
the rate of change of the interfacial tension between the droplet and the continuous phase fluid with temperature is denoted by @xmath10 .
figure 1 displays a schematic diagram of thermocapillary migration of a droplet with a thermal source in a laboratory coordinate system denoted by a bar .
axisymmetric energy equations for the continuous phase and the fluid within the droplet are written as follows @xmath11 where @xmath12 and @xmath13 are velocity and temperature , a prime and @xmath14 denote quantities and a thermal source in the droplet . by going over an unstable migration process from beginning , the droplet migration can reach a steady state , i.e. , with the constant migration speed @xmath15 . using the coordinate transformation from the laboratory
coordinate system to a coordinate system moving with the droplet velocity @xmath15 @xmath16 the problem ( 1 ) can be formulated as @xmath17 by taking the radius of the droplet @xmath0 , the ygb model velocity @xmath18 and @xmath19 as reference quantities to make coordinates , velocity and temperature dimensionless , the energy equations ( 3 ) can be written in the following dimensionless form in the spherical coordinate system ( @xmath20 ) as follows @xmath21 where the small parameter @xmath22 and ma number are defined , respectively , as @xmath23 and @xmath24 in the space experiments@xcite , prandtl numbers ( @xmath25 ) are constants in terms of the fixed fluid media , but ma numbers are changed duo to the variable droplet sizes . it is common to adopt re numbers ( @xmath26 ) instead of pr numbers in order to capture velocity fields with the changes of re numbers . in general
, the deformation of droplet usually depends on the weber numbers [ @xmath27 , @xmath28 is the capillary number ] , it is not included in this study due to the small we numbers in the space experiments .
for example , @xmath29 is in @xmath30 when @xmath31@xcite .
thus , the solutions of eqs .
( 4 ) and ( 5 ) have to satisfy the following boundary conditions at the interface of two phase fluids ( @xmath32 ) @xmath33 and at places far away from the droplet @xmath34 besides the above temperature boundary conditions , the stress boundary conditions at the interface are also applied for the steady droplet migration .
however , the normal stress boundary condition is removed in terms of the non - deformation assumption .
the shear stress boundary condition is expressed by @xmath35 where @xmath36 for large re numbers , the inner and outer momentum boundary layers at @xmath37-direction are introduced near the surface of droplet . in the flow fields outside the momentum boundary layers ,
there are still the potential flows . the scaled inviscid velocity field in the continuous phase and hill s spherical vortex within the droplet can be written as , respectively@xcite @xmath38 and @xmath39 by introducing the perturbation velocity fields @xmath40 , @xmath41 and the depth of boundary layer @xmath42 @xmath43 into the boundary condition ( 11 ) and the momentum equation for the continuous fluid , they are derived and truncated in the leading orders @xmath44 and @xmath45 as , respectively@xcite @xmath46 and @xmath47 where @xmath48 , @xmath49 , @xmath50 , @xmath51 and @xmath52 are in the momentum boundary layers .
following the derivation for the steady migration speed of a bubble at large re numbers@xcite , the migration speed of the droplet is written as follows @xmath53 thus , to determine @xmath54 , the temperature @xmath55 on the surface of droplet is required . however , before the energy equations are solved , the self - consistency of temperature fields for the steady droplet migration system will be determined .
it means that the solutions of steady energy equations will satisfy the conservative integral thermal flux across the surface of droplet . in this case , the temperature field at infinity in eq .
( 10 ) was derived by using the asymptotic expansion method and expressed as@xcite @xmath56 integrating eqs .
( 4 ) and ( 5 ) in the continuous phase domain @xmath57,\theta\in[0,\pi])$ ] with boundary condition ( 18 ) and within the droplet region @xmath58,\theta\in[0,\pi])$ ] , respectively , we obtain @xmath59 and @xmath60 from eq . (
19 ) and eq .
( 20 ) , we have @xmath61 \sin \theta d \theta \\ = \frac{1}{3 \epsilon^2 } ( 1 + \frac{2 \beta}{\lambda})-\frac{2 \beta \omega}{3 \lambda g v_0 } ma . \end{array}\ ] ] for large ma numbers and finite @xmath54 , eqs .
( 19 ) and ( 20 ) should satisfy the thermal flux boundary condition ( 9 ) , i.e. , the right side of eq .
( 21 ) will be zero .
so , we have @xmath62 using eq . ( 22 ) , eq .
( 5 ) is rewritten as @xmath63 in following , we will focus on the steady thermocapillary migration of droplet with the thermal source under a uniform temperature gradient and determine the dependence of the migration speed on large ma numbers .
by using an outer expansion for the scaled temperature field in the continuous phase @xmath64 the energy equation for the outer temperature field in its leading order can be obtained from eq.(4 ) as follows @xmath65 by using the coordinate transformation from @xmath66 to @xmath67 to solve eq .
( 25 ) , its solution can be written as @xmath68 where @xmath69 is a function of @xmath70 ( the streamfunction in the continuous phase ) .
following@xcite , the solution near @xmath32 is simplified as @xmath71 by using the boundary layer approximation @xmath72 the temperature field near interface can be expressed as @xmath73 by using the following outer expansion for the scaled temperature field within the droplet in eq .
( 23 ) @xmath74 the equation in its leading order can be written as @xmath75 its solution is @xmath76 where @xmath77 is the streamfunction within the droplet .
the unknown function @xmath78 can be obtained from the following equation for the temperature field @xmath79 in its second order @xmath80 to solve eq .
( 33 ) , we use the coordinate transformation from @xmath66 to @xmath81 , with streamlines and their orthogonal lines as coordinate axes defined as @xmath82 eq.(33 ) can thus be written in the @xmath81 coordinate system as follows @xmath83 where @xmath84 , @xmath85 , @xmath86 are the metrical coefficients for the transformation . multiplying both sides of eq.(35 ) by @xmath87 and integrating with respect to @xmath88 for one circuit along a streamline , we obtain the following equation @xmath89 + \frac{\lambda}{2\beta } h(m),\ ] ] where @xmath90 and @xmath91 . from the equation of continuity
, it can be shown that @xmath92 is a constant .
we thus obtain a solution of eq .
( 36 ) as follows @xmath93.\ ] ] near @xmath94 , @xmath95 and @xmath96 can be expanded , respectively @xmath97 and @xmath98 using eqs . ( 38 ) and ( 39 ) , we thus obtain from eq .
( 37 ) @xmath99 + o(m^3\ln m).\ ] ] by using the boundary layer approximation @xmath100 the temperature field near interface can be expressed as follows @xmath101 from eqs .
( 19 ) and ( 20 ) , the leading orders of the inner temperature fields in the continuous phase and within the droplet can be obtained , respectively . a schematic diagram of the inner and outer regions with the leading orders near the surface of droplet is shown in fig .
2 . by using inner expansions for the continuous phase and the fluid in the droplet @xmath102 and the inner variables given in eqs . ( 28 ) and ( 41 ) , the scaled energy equations for the inner temperature fields in the leading orders can be written as follows @xmath103 the boundary conditions are @xmath104 where @xmath105 .
we transform the independent variables from @xmath106 $ ] to @xmath107 $ ] and the functions from @xmath108 to @xmath109 as @xmath110 and @xmath111 the corresponding energy equations for @xmath112 and boundary conditions can be written as follows @xmath113 and @xmath114 to solve eq .
( 50 ) , initial conditions are provided below @xmath115 following the methods given by harper and moore@xcite , the solution of eq.(50 ) can be obtained as @xmath116d \eta^ * \},\\ \end{array}\ ] ] @xmath117 + \frac{1-\delta}{1+\delta } \exp[-\frac{(\eta'+\eta^*)^2}{4\xi } ] \}d\eta^ * .\\ \end{array}\ ] ] from eqs.(49 ) and ( 53 ) , we obtain the inner temperature field in its leading order near the surface of droplet @xmath118\\ & = \frac{1}{1+\delta } [ b+ \frac{\delta \sqrt{\xi}}{3 \sqrt{\pi } } + \frac{2}{\sqrt{\pi } } \int_0^{\infty } g_0(2 \xi^{1/2 } \zeta ) \exp(-\zeta^2)d\zeta ] .
\end{array}\ ] ] when the inner expansion in the temperature field ( 43 ) is truncated at the zero order , we rewrite eq .
( 17 ) as @xmath119 d\theta.\ ] ] since @xmath120 , the migration speed of the droplet is evaluated as @xmath121 where @xmath122 and @xmath123 substituting eq .
( 55 ) into eq .
( 58 ) , we obtain @xmath124 d\theta .
\end{array}\ ] ] to determine the function @xmath125 in eq .
( 60 ) , we use the boundary condition within the droplet at the front and rear stagnation points in eq .
( 52 ) @xmath126 + \frac{1-\delta}{1+\delta } \exp[-\frac{(\eta'+\eta^*)^2}{8 } ] \}d\eta^*. \end{array}\ ] ] the integral of the third term on the right - hand side of eq .
( 61 ) is approximated as @xmath127 then , eq . ( 61 ) is evaluated in a linear system of equations @xmath128
+ \frac{1-\delta}{1+\delta } \exp[-\frac{(\eta'+\eta^*_1)^2}{8 } ] \ } \delta \eta^*\\ & - \frac{1}{4\sqrt{2\pi } } g_0(\eta^*_{n+1 } ) \{\exp[-\frac{(\eta'-\eta^*_{n+1})^2}{8 } ] + \frac{1-\delta}{1+\delta } \exp[-\frac{(\eta'+\eta^*_{n+1})^2}{8 } ] \ } \delta \eta^*\\ & - \frac{1}{2\sqrt{2\pi } } \sum_{j=2}^n g_0(\eta^*_j ) \{\exp[-\frac{(\eta'-\eta^*_j)^2}{8 } ] + \frac{1-\delta}{1+\delta } \exp[-\frac{(\eta'+\eta^*_j)^2}{8 } ] \ } \delta \eta^*\\ & - \frac{1}{2 } g_0(\eta^*_{n+1 } ) [ { \rm erfc } ( \frac{\eta^*_{n+1}+\eta'}{2\sqrt{2 } } ) + \frac{1-\delta}{1+\delta } { \rm erfc } ( \frac{\eta^*_{n+1}+\eta'}{2\sqrt{2 } } ) ] \\ & = \frac{\delta}{1+\delta } [ -(b + \frac{1}{6 } \eta'){\rm erfc}(\frac{\eta'}{2\sqrt{2 } } ) + \frac{\sqrt{2}}{3\sqrt{\pi } } \exp(-\frac{\eta'^2}{8 } ) ] , \\ \end{array}\ ] ] where @xmath129 and @xmath130 .
the physical coefficients used in the space experiment@xcite with the uniform temperature gradient @xmath131k / cm for the continuous phase of fluorinert fc-75 and the droplet of 5cst silicon oil at @xmath132k are adopted here to yield @xmath133 , @xmath134 and @xmath135 . a typical value for @xmath136
is chosen as 3 .
using the trial and error method to satisfy the above approximation , we determine the unknown constant @xmath137 and obtain the dependence of @xmath125 on @xmath138 as shown in fig .
3 . from eq . ( 60 ) , we can determine the root - mean - square of the leading order term of the migration speed as @xmath139 although equations and boundary conditions describing the second order term of the migration speed can be obtained , we are unable to find an analytical result for @xmath140 in eq .
( 56 ) . under the truncation after the leading order term in eq.(57 ) , we obtain the migration speed of the droplet @xmath141 by using the migration speed @xmath54 , the adding thermal source @xmath14 in eq .
( 22 ) can be determined as @xmath142
in this paper , the conservative integral thermal flux across the surface for a steady thermocapillary droplet migration in a uniform temperature gradient at large re and ma numbers has been guaranteed by adding a thermal source in the droplet . under the assumptions of quasi - steady state and non - deformation of the droplet
, we have made an analytical result for the steady thermocapillary migration of droplet with the thermal source at large re and ma numbers .
the result shows that the thermocapillary droplet migration speed slowly increases with the increase of ma number .
we emphasize that the modeling of the adding thermal source in the droplet is used to investigate the effect of integral thermal flux across the surface on the thermocapillary droplet migration process .
it is one of the physical means to satisfy the conservative integral thermal flux boundary condition in the steady migration process at large ma numbers@xcite . on the one hand , in general ,
when the droplet moves forward under the uniform temperature gradient , the thermal energy in the droplet is imported from the upper surface of the droplet and exported from the lower surface .
the adding thermal source in the droplet can decrease / increase the inner thermal flux from the interface / inside to the inside / interface at the upper / lower surface of the droplet and satisfy the thermal flux boundary condition at the surface of droplet for the steady migration process . on the other hand , at large ma numbers , the heat advection around the droplet is a more significant mechanism for heat transfer across / around the droplet . it is always keeping the weak transport of thermal energy from outside of the droplet to inside .
thus , the adding thermal source is conducive to providing the thermal energy in the droplet to meet the requirement put forward by the quasi - steady migration process .
meanwhile , the heat transport across the droplet surface becomes weaker than the heat convection as ma number increases . to reach the steady migration process ,
more thermal energy in the droplet will be provided .
the proportional relationship of the thermal source to ma number in eq .
( 66 ) qualitatively confirms the above requirement of more thermal energy in the droplet . in final ,
some suggestions are provided in order that the thermal source in the droplet can be implementend in a real experiment . as given in eqs .
( 1),(66 ) and shown in fig .
1 , the thermal source depending on ma number will be added in the droplet when it starts to move under the uniform temperature gradient
. the heat radiant technology may be one of the possible physical means to add the thermal source in the droplet . *
acknowledgments * we thank the national science foundation for partial support through the grant no .
11172310 and the imech / sccas shenteng 1800/7000 research computing facilities for assisting in the computation .
b. braun , c. ikier and h. klein , thermocapillary migration of dropletlets in a binary mixture with miscibility gap during liquid / liquid phase separation under reduced gravity , _ j. colloid interface sci .
_ * 159 * , ( 1993 ) 515 .
p. h. hadland , r. balasubramaniam , and g. wozniak , thermocapillary migration of bubbles and droplets at moderate to large marangoni number and moderate reynolds number in reduced gravity , _ exper .
* 26 * , ( 1999 ) 240 .
j. c. xie , h. lin , p. zhang , f. liu and w. r. hu , experimental investigation on thermocapillary droplet migration at large marangoni number in reduced gravity , _ j. colloid interface sci . _ * 285 * , ( 2005 ) 737 .
fig . 2 . a schematic diagram of the inner and outer temperature fields denoted by different colors near the surface of droplet in a coordinate system moving with the droplet velocity @xmath15 .
the light / dark gray and the white / black regions represent the inner [ @xmath143 and outer [ @xmath144 ) temperature fields in the continuous phase / within the droplet , respectively . | the _ unsteady _ process for thermocapillary droplet migration at large reynolds and marangoni numbers has been previously reported by identifying a nonconservative integral thermal flux across the surface in the _ steady _ thermocapillary droplet migration , [ wu and hu , j. math .
phys . * 54 * 023102 , ( 2013 ) ] . here
we add a thermal source in the droplet to preserve the integral thermal flux across the surface as conservative , so that thermocapillary droplet migration at large reynolds and marangoni numbers can reach a _ quasi - steady _ process . under assumptions of _ quasi - steady _ state and non - deformation of the droplet
, we make an analytical result for the _ steady _ thermocapillary migration of droplet with the thermal source at large reynolds and marangoni numbers .
the result shows that the thermocapillary droplet migration speed slowly increases with the increase of marangoni number . *
pacs * 47.20.dr ; 47.55.nb ; 47.55.d- ; 47.54.-r + * keywords * interfacial tension ; thermocapillary droplet migration ; large marangoni number ; microgravity + |
in this paper we consider a natural way of counting the size of @xmath2-terms , namely @xmath2-terms presented by de bruijn indices . ]
in which all the operators are counted with size @xmath0 .
this means that abstractions , applications , successors and zeros have all size @xmath0 .
formally @xmath3 for instance the term for which is written traditionally @xmath4 in the lambda calculus is written @xmath5 using de bruijn indices and we have : @xmath6 since there are two @xmath2 abstractions , one successor @xmath7 and one @xmath8 .
the term for ( which should not be confused with the successor symbol ) is written @xmath9 which is written @xmath10 using de bruijn indices and its size is : @xmath11 since there are three @xmath2 abstractions , three applications , three successors @xmath7 s , and four @xmath8 s .
the term @xmath12 which corresponds to the term @xmath13 has size @xmath14 and the term @xmath15 which corresponds to the term @xmath16 is written @xmath17 and has size @xmath18 .
the term @xmath19 which corresponds to the fixpoint @xmath20 is written @xmath21 and has size @xmath22 .
since the terms are either applications , abstractions or de bruijn indices , the set @xmath24 of lambda terms is solution of the combinatorial equation : @xmath25 where @xmath26 is the set of de bruijn indices which is solution of @xmath27 let us call @xmath23 the generating function for counting the numbers of the plain terms .
it is solution of the functional equation : @xmath28 which yields the equation : @xmath29 which has discriminant @xmath30 this gives the solution @xmath31 which has @xmath32 as pole closest to @xmath1 .
the @xmath33 first values of @xmath34{l_\infty}$ ] are : 0 , 1 , 2 , 4 , 9 , 22 , 57 , 154 , 429 , 1223 , 3550 , 10455 , 31160 , 93802 , 284789 , 871008 , 2681019 this sequence is * a105633 * in the _ online encyclopedia of integer sequences_. [ theorem : approx ] assume @xmath35 + that is @xmath36 @xmath37{l_\infty}\sim \left(\frac{1}{{\rho_{l_\infty}}}\right)^n \frac{c}{n^{\frac{3}{2}}}\ ] ] the proof mimics this of theorem 1 in @xcite .
let us write @xmath23 as @xmath38 where @xmath39 applying theorem vi.1 of @xcite , we get : @xmath37{l_\infty}\sim \left(\frac{1}{{\rho_{l_\infty}}}\right)^n \cdot \frac{n^{-3/2}}{`g(-\frac{1}{2})}~\widetilde{c}\ ] ] with @xmath40 notice that @xmath41 . from this
we get @xmath42 figure [ fig : approxloo ] shows approximations of @xmath43{l_\infty}$ ] .
@xmath44{l_\infty}}}\qquad\qquad\qquad \\\hline 10 & 3550 \\ 20 & 253106837 \\ 30 & 27328990723991 \\ 40 & 3503758934959966001 \\ 50 & 493839291745701673090756 \\ 60 & 73920774614279746859303111580 \\ 70 & 11535317831253359292868402823579507 \\ 80 & 1855899670106913269845444317474927546423 \\ 90 & 305649725186484753579669948042728038245882292 \\ 100 & 51274965000307280025396615989999357497440689837989\\ \hline\hline \textbf{n } & \mathit{\mathbf{\lfloor ( 1/{\rho_{l_\infty}})^n c / n^{3/2}\rfloor } } \qquad\qquad\qquad\\\hline 10 & 3767\\ 20 & 261489930\\ 30 & 27945182509468\\ 40 & 3563589864915927683\\ 50 & 500623883981281516056181\\ 60 & 74770204056757299054875868847\\ 70 & 11649230835743409545961872906078995\\ 80 & 1871967051054756263272240387385909197928\\ 90 & 308005368563187477433148735955649926279818246\\ 100 & 51631045600653143846184406311963448514677624135086 \end{array}\ ] ] using the package @xcite we were able to build a holonomic equation of which @xmath23 is the solution namely @xmath45 from this equation it is possible to derive the following recursive and linear definition for the coefficients : l_,0 = 0l_,1 = 1l_,2 = 2 l_,3 = 4 l_,n = the set @xmath48 of terms with free indices @xmath49 is described as @xmath50 the set @xmath51 is the set of closed lambda terms .
if we consider the @xmath2-terms with at most @xmath46 free indices , we get : @xmath52 which yields : @xmath53 let us state @xmath54 we have @xmath55 notice that @xmath56 is defined using @xmath57 .
if this definition is developed , then @xmath47 is defined by an infinite sequence of nested radicals .
the sequences @xmath58l_m)_{n`:{\ensuremath{\mathbb{n}}}}$ ] do not occur in the _ online encyclopedia of integer sequences_. assume we take another notion of size in which @xmath8 has size @xmath1 and applications have size @xmath59 , whereas abstraction and succession keep their size @xmath0 .
in other words : @xmath60 the generating function as a reference to the function @xmath61 described in sequence * a105632 * of the _ online encyclopedia of integer sequences_. ] @xmath62 fulfills the identity : @xmath63 the reader may check that @xmath64a_1 = [ z^{n+1}]{l_\infty}.\ ] ] hence both notions of size correspond to sequence * a105633*. in appendix [ sec : zeroforzerodot ] we consider the case where all the operators ( application , abstraction and succession ) have size @xmath0 and @xmath8 has size @xmath1 .
a difficult open problem is to count simply typable terms . in this section ,
we give empiric results we obtain by an implementation on counting closed typed terms .
@xmath65{r|r|r } \textbf{size } & \textbf{typables } & \textbf{all}\\\hline \mathbf{0 } & 0 & 0\\ \mathbf{1 } & 0 & 0\\ \mathbf{2 } & 1 & 1\\ \mathbf{3 } & 1 & 1\\ \mathbf{4 } & 2 & 3\\ \mathbf{5 } & 5 & 6\\ \mathbf{6 } & 13 & 17\\ \mathbf{7 } & 27 & 41\\ \mathbf{8 } & 74 & 116\\ \mathbf{9 } & 198 & 313\\ \mathbf{10 } & 508 & 895\\ \mathbf{11 } & 1371 & 2550\\ \mathbf{12 } & 3809 & 7450\\ \mathbf{13 } & 10477 & 21881\\ \mathbf{14 } & 29116 & 65168\\ \mathbf{15 } & 82419 & 195370\\ \mathbf{16 } & 233748 & 591007\\ \mathbf{17 } & 666201 & 1798718\\ \mathbf{18 } & 1914668 & 5510023 \\ \mathbf{19 } & 5528622 & 16966529 \\
\mathbf{20 } & 16019330 & 52506837 \\
\mathbf{21 } & 46642245 & 163200904 \\
\mathbf{22 } & 136326126 & 509323732 \\
\mathbf{23 } & 399652720 & 1595311747 \\
\mathbf{24 } & 1175422931 & 5013746254 \end{array } $ ]
a black - white binary tree is a binary tree with colored nodes using two colors , _ black _ @xmath67 and _ white _ @xmath68 . _
the root of a black - white binary tree is @xmath69 , by convention . _
a @xmath66-free black - white binary tree is a black - white binary tree in which edges from a set @xmath66 are forbidden .
for instance if the set of forbidden edges is @xmath70\\ { ` ( ! ) } } } } } , { { \raisebox{.8pc}{\tiny \xymatrix@=.05pc{{`(!)}\ar@{-}[dr]\\ & { ` ( ? ) } } } } } , { { \raisebox{.8pc}{\tiny \xymatrix@=.05pc{{`(!)}\ar@{-}[dr]\\ & { ` ( ! ) } } } } } , { { \raisebox{.8pc}{\tiny \xymatrix@=.05pc{{`(?)}\ar@{-}[dr]\\ & { ` ( ? ) } } } } } \}$ ] , this means that only edges in @xmath71\\ { ` ( ? ) } } } } } , { { \raisebox{.8pc}{\tiny \xymatrix@=.05pc{&{`(!)}\ar@{-}[dl]\\ { ` ( ! ) } } } } } , { { \raisebox{.8pc}{\tiny \xymatrix@=.05pc{&{`(?)}\ar@{-}[dl]\\ { ` ( ? ) } } } } } , { { \raisebox{.8pc}{\tiny \xymatrix@=.05pc{{`(?)}\ar@{-}[dr]\\ & { ` ( ! ) } } } } } \}$ ] are allowed .
the @xmath72-free black - white binary trees of size @xmath73 and @xmath14 are as many as lambda terms of size @xmath73 and @xmath14 .
they are listed in fig .
[ fig : bij3 ] and fig .
[ fig : bij4 ] second column . for @xmath70\\ { ` ( ! ) } } } } } , { { \raisebox{.8pc}{\tiny \xymatrix@=.05pc{{`(!)}\ar@{-}[dr]\\ & { ` ( ? ) } } } } } , { { \raisebox{.8pc}{\tiny \xymatrix@=.05pc{{`(!)}\ar@{-}[dr]\\ & { ` ( ! ) } } } } } , { { \raisebox{.8pc}{\tiny \xymatrix@=.05pc{{`(?)}\ar@{-}[dr]\\ & { ` ( ? ) } } } } } \}$ ] , like for @xmath74\\ { ` ( ? ) } } } } } , { { \raisebox{.8pc}{\tiny \xymatrix@=.05pc{&{`(!)}\ar@{-}[dl]\\ { ` ( ! ) } } } } } , { { \raisebox{.8pc}{\tiny \xymatrix@=.05pc{&{`(?)}\ar@{-}[dl]\\ { ` ( ? ) } } } } } , { { \raisebox{.8pc}{\tiny \xymatrix@=.05pc{{`(?)}\ar@{-}[dr]\\ & { ` ( ! ) } } } } } \}$ ] , which is obtained by left - right symmetry , the @xmath66-free black - white binary trees are counted by @xcite . in what follows we will consider @xmath72 and we will rather speak in terms of an allowed set of pattern namely @xmath62 . for simplicity , we will call in this paper _ black - white trees _ , the binary black - white trees with allowed pattern set @xmath62
. @xmath75\\ & { \bullet}\ar@{-}[dl]\\{\bullet } } & \xymatrix@=.2pc{{`*}\ar@{-}[dr]\\&{`*}\ar@{-}[dr]\\ & & { ` * } } % & s^3\zerodot \\
\hline % ` l s\zerodot & \xymatrix@=.2pc{&&{\bullet}\ar@{-}[dl]\\ & { \bullet}\ar@{-}[dl]\\ { ` ( ? ) } } & % \xymatrix@=.2pc{{`*}\ar@{-}[dr]\\&{`*}\ar@{-}[dl]\\ { ` * } } % & \zerodot\,(s\zerodot)\\ \hline % ` l ` l \zerodot & \xymatrix@=.2pc{&&{\bullet}\ar@{-}[dl]\\ & { ` ( ? ) } \ar@{-}[dl]\\ { ` ( ? ) } } & % \xymatrix@=.2pc{&&{`*}\ar@{-}[dl]\\ & { ` * } \ar@{-}[dl]\\{`*}}% & \zerodot\,(`l\zerodot ) \\
\hline % \zerodot\ , \zerodot & \xymatrix@=.2pc{&{\bullet}\ar@{-}[dl]\\ { ` ( ? ) } \ar@{-}[dr]\\ & { ` ( ! ) } } & % \xymatrix@=.2pc{&{`*}\ar@{-}[dl]\ar@{-}[dr]\\ { ` * } & & { ` * } } % & ( s \zerodot)\ , \zerodot \\
\hline \end{array } $ ] @xmath76\\ & & { \bullet}\ar@{-}[dl ] \\ & { \bullet}\ar@{-}[dl ] \\ { \bullet } } & % \xymatrix@=.2pc{{`*}\ar@{-}[dr]\\ & { ` * } \ar@{-}[dr ] \\ & & { ` * } \ar@{-}[dr ] \\ & & & { ` * } }
% & s^4\zerodot\\\hline % ` ls^2\zerodot & \xymatrix@=.2pc { & & & { \bullet}\ar@{-}[dl]\\ & & { \bullet}\ar@{-}[dl ] \\ & { \bullet}\ar@{-}[dl ] \\
{ \circ } } & \xymatrix@=.2pc { { ` * } \ar@{-}[dr]\\ & { ` * } \ar@{-}[dr]\\&&`*\ar@{-}[dl]\\&`*}% & \zerodot\,(s^2\zerodot)\\\hline % ` l`l s\zerodot & \xymatrix@=.2pc { & & & { \bullet}\ar@{-}[dl]\\ & & { \bullet}\ar@{-}[dl ] \\ & { \circ}\ar@{-}[dl ] \\ { \circ } } & % \xymatrix@=.2pc{&{`*}\ar@{-}[dr]\\ & & { ` * } \ar@{-}[dl ] \\ & { ` * } \ar@{-}[dl ] \\ { ` * } } % & \zerodot\,(`ls\zerodot)\\\hline % ` l`l`l \zerodot & \xymatrix@=.2pc { & & & { \bullet}\ar@{-}[dl]\\ & & { \circ}\ar@{-}[dl ] \\ & { \circ}\ar@{-}[dl ] \\ { \circ } } & % \xymatrix@=.2pc { & & & { ` * } \ar@{-}[dl]\\ & & { ` * } \ar@{-}[dl ] \\ & { ` * } \ar@{-}[dl ] \\ { ` * } } % & \zerodot\,(`l`l\zerodot)\\\hline % \zerodot\ , ( s \zerodot ) & \xymatrix@=.2pc { & & { \bullet}\ar@{-}[dl]\\ & { \bullet}\ar@{-}[dl ] \\ { \circ}\ar@{-}[dr ] \\ & { \bullet } } & \xymatrix@=.2pc{&{`*}\ar@{-}[dr]\\ & & { ` * } \ar@{-}[dl]\ar@{-}[dr]\\ & { ` * } & & { ` * } } % & ( s \zerodot)\ , ( s \zerodot ) \\\hline% % \end{array } % \qquad % \begin{array}{c|c } % `
l-\textrm{\bf terms}&\textrm{\bf black - white binary trees } % \\\hline \zerodot\ , ( ` l \zerodot ) & \xymatrix@=.2pc { & & { \bullet}\ar@{-}[dl]\\ & { \circ}\ar@{-}[dl ] \\
{ \circ}\ar@{-}[dr ] \\ & { \bullet } } & % \xymatrix@=.2pc{&&{`*}\ar@{-}[dl]\\ & { ` * } \ar@{-}[dr ] \ar@{-}[dl ] \\ { ` * } & & { ` * } } % & ( s\zerodot)\ , ( ` l \zerodot ) \\\hline % ( ` l\zerodot)\,\zerodot & \xymatrix@=.2pc { & { \bullet}\ar@{-}[dl]\\ { \circ}\ar@{-}[dr ] \\ & { \bullet}\ar@{-}[dl ] \\ { \circ } } & \xymatrix@=.2pc{&{`*}\ar@{-}[dl]\ar@{-}[dr]\\ { ` * } & & { ` * } \ar@{-}[dl]\\ & { ` * } } % & \zerodot\,\zerodot\,\zerodot\\\hline % ( s \zerodot)\,\zerodot & \xymatrix@=.2pc { & { \bullet}\ar@{-}[dl]\\ { \circ}\ar@{-}[dr ] \\ & { \bullet}\ar@{-}[dl ] \\ { \bullet}}&% \xymatrix@=.2pc{&{`*}\ar@{-}[dl]\ar@{-}[dr]\\{`*}&&{`*}\ar@{-}[dr]\\ & & & { ` * } }
% < - & ( s^2 \zerodot)\,\zerodot\\\hline % ` l(\zerodot\,\zerodot ) & \xymatrix@=.2pc { & & { \bullet}\ar@{-}[dl]\\ & { \circ}\ar@{-}[dl]\ar@{-}[dr ] \\
{ \circ } & & { \bullet}}&% \xymatrix@=.2pc{&&{`*}\ar@{-}[dl]\ar@{-}[dr]\\ & { ` * } \ar@{-}[dl]&&{`*}\\ { ` * } } % & \zerodot\,(\zerodot\,\zerodot ) \\\hline% \end{array } $ ] before giving the bijection , let us give the trees corresponding to @xmath77 , to @xmath78 , to @xmath79 , and to @xmath80 : @xmath81\\ & & ` ( ! ) \ar@{-}[dl]\\ & ` ( ? ) \ar@{-}[dl]\\ ` ( ? ) } & \textrm{for~ } \mathsf{s } \xymatrix@=.2pc { & & & & & ` ( ! ) \ar@{-}[dl]\\ & & & & ` ( ? ) \ar@{-}[dl]\ar@{-}[drr]\\ & & & ` ( ? ) \ar@{-}[dl]\ar@{-}[dr ] & & & ` ( ! ) \ar@{-}[dl]\\ & & ` ( ? ) \ar@{-}[dl]\ar@{-}[dl]&&`(!)\ar@{-}[dl ] & ` ( ! ) \\ & ` ( ? ) \ar@{-}[dl ] & & ` ( ? ) \ar@{-}[dr ] & \\ ` ( ? ) & & & & ` ( ! ) \ar@{-}[dl]\\ & & & ` ( ! ) \ar@{-}[dl]\\ & & ` ( ! ) } \end{array}\ ] ] @xmath82\\ & & & ` ( ? ) \ar@{-}[dl ] \ar@{-}[dr]\\ & & ` ( ? ) \ar@{-}[dl]&&`(!)\\ & ` ( ? ) \ar@{-}[dr]\\ & & ` ( ! ) \ar@{-}[dl]\\ & ` ( ? ) \ar@{-}[dl ] \ar@{-}[dr]\\ ` ( ? ) & & ` ( ! ) } & \textrm{for~ } \mathsf{y } \xymatrix@=.2pc { & & & & & & & ` ( ! ) \ar@{-}[dl]\\ & & & & & & ` ( ? ) \ar@{-}[dl]\ar@{-}[dr]\\ & & & & & ` ( ? ) \ar@{-}[dl]\ar@{-}[dr]&&`(!)&&\\ & & & & { ` ( ? ) } \ar@{-}[dl ] & & ` ( ! ) \ar@{-}[dl]\\ & & & ` ( ? ) \ar@{-}[dl]\ar@{-}[dr ] & & ` ( ! ) \\ & & ` ( ? ) & & ` ( ! ) \ar@{-}[dl]\\ & & & ` ( ? ) \ar@{-}[dl]\ar@{-}[dr]\\ & & ` ( ? ) \ar@{-}[dl]\ar@{-}[dr]&&`(!)&&\\ & { ` ( ? ) } & & ` ( ! ) \ar@{-}[dl]\\ & & ` ( ! ) } \end{array}\ ] ] assume @xmath83 is the empty tree which is usually not represented in drawing .
the @xmath72-free black - white binary trees are described by the following combinatorial equation : @xmath84\\ { \mathcal{b}}{\mathcal{w } } _ { ` ( ! ) } } } \ ` ( + ) \
\raisebox{.2cm}{\xymatrix@=.2pc{&{`(!)}\ar@{-}[dl]\\ { \mathcal{b}}{\mathcal{w } } _ { ` ( ? ) } } } \\ { \mathcal{b}}{\mathcal{w } } _ { ` ( ? ) } & = & \box \ ` ( + ) \
\raisebox{.2cm}{\xymatrix@=.2pc{&{`(?)}\ar@{-}[dl]\\ { \mathcal{b}}{\mathcal{w } } _ { ` ( ? ) } } } \ ` ( + ) \ \raisebox{.2cm}{\xymatrix@=.2pc{&{`(?)}\ar@{-}[dl]\ar@{-}[dr]\\ { \mathcal{b}}{\mathcal{w}}_{`(?)}&&{\mathcal{b}}{\mathcal{w}}_{`(!)}}}\end{aligned}\ ] ] which yields the following functional equations : @xmath85 hence @xmath86 and @xmath87 which is the same equation up to a multiplication by @xmath88 as ( [ eq : loo ] ) namely the equation defining @xmath23 let us define the function @xmath89 from @xmath2-terms to black - white trees : @xmath90\\{`(!)}}}}}\\ { \ensuremath{\mathsf{ltobw}}}(`l m ) & = & { { \raisebox{.8pc}{\tiny \xymatrix@=.05pc{&{{\ensuremath{\mathsf{ltobw}}}(m)}\ar@{-}[dl]\\{`(?)}}}}}\\ { \ensuremath{\mathsf{ltobw}}}(m_1\,m_2 ) & = & \raisebox{.5cm}{${\tiny \xymatrix@=.2pc{&{{\ensuremath{\mathsf{ltobw}}}(m_2)}\ar@{-}[dl]\\{`(?)}\ar@{-}[dr]\\ & { { \ensuremath{\mathsf{ltobw}}}(m_1)}}}$}\end{aligned}\ ] ] in other words a new node is added on the leftmost node of the tree . from black - white trees to @xmath2-terms
let us now define the function @xmath91 @xmath92\\{`(!)}}}}}\right ) & = & s({\ensuremath{\mathsf{bwtol}}}(t))\\ { \ensuremath{\mathsf{bwtol}}}\left({{\raisebox{.8pc}{\tiny \xymatrix@=.05pc{&{t}\ar@{-}[dl]\\{`(?)}}}}}\right ) & = & ` l\ , { \ensuremath{\mathsf{bwtol}}}(t)\\ { \ensuremath{\mathsf{bwtol}}}\left(\raisebox{.5cm}{${\tiny \xymatrix@=.2pc{&{t_2}\ar@{-}[dl]\\{`(?)}\ar@{-}[dr]\\ & { t_1}}}$}\right ) & = & { \ensuremath{\mathsf{bwtol}}}(t_1 ) \ , { \ensuremath{\mathsf{bwtol}}}(t_2)\end{aligned}\ ] ] in other words , to decompose a binary tree which is not the node @xmath69 , we look for the left most node .
* if the leftmost node is @xmath69 , then the @xmath2-term is a de bruijn index .
actually there are only @xmath69 s ( indeed @xmath93\\{`(!)}}}}}$ ] is forbidden ) and the tree is linear . if this linear tree has @xmath94
@xmath69 s it represents @xmath95 . *
if the leftmost node is @xmath96 and has no child , then the @xmath2-term is an abstraction of the bijection of the rest .
* if the leftmost node is @xmath96 and has a right child , then the @xmath2-term is an application of the bijection of the right subtree on the bijection of the above tree .
@xmath97 and @xmath98 in this section we describe haskell programs for the bijections .
first we define black - white trees .
we consider three kinds of trees : leafs ( of arity zero and size zero ) corresponding to @xmath83 and not represented in drawing .
+ + in order to translate @xmath2-terms to corresponding black - white trees we carry out a rather unusual induction , where after the recursive step we attach a new subtree to the leftmost node in one of the previously obtained black - white trees .
similarly , in the inverse translation from black - white trees to @xmath2-terms , we have to cut out the leftmost node of the current black - white tree and pattern match against the result .
this unusual recursion is a result of our natural top - down representation of black - white trees , where children are drawn below their parents .
note that if we change this convention so that children are drawn on the right to their parents , the previously leftmost node becomes the root of the black - white tree .
the data type for black - white trees does not change , but instead of top - down trees , we are working with left - right ones .
such a representation simplifies the overall implementation as the algorithm is no longer required to look for the leftmost node .
consider @xmath99 the set of binary trees with no zigzag i.e. , with no subtree like @xmath100\ar@{-}[dr]\\ ` * \ar@{-}[dr]&&\\ & { ` * } \ar@{-}[dl]\ar@{-}[dr ] \\ & & & & } \ ] ] @xmath99 is described by the combinatorial equations : @xmath101\\&{\mathcal{b}}{\mathcal{z}}_1 } } \ ` ( + ) \ { \mathcal{b}}{\mathcal{z}}_2\\ { \mathcal{b}}{\mathcal{z}}_2 & = & { ` * } \ \ ` ( + ) \ \
\raisebox{.4cm}{\xymatrix @=.05pc{&{`*}\ar@{-}[dl]\\ { \mathcal{b}}{\mathcal{z}}_2 } } \ ` ( + ) \ \raisebox{.4cm}{\xymatrix @=.05pc{&{`*}\ar@{-}[dl]\ar@{-}[dr]\\{\mathcal{b}}{\mathcal{z}}_2&&{\mathcal{b}}{\mathcal{z}}_1 } } \end{aligned}\ ] ] like @xmath23 and @xmath102 , @xmath103 is solution of the functional equation : @xmath104 sapounakis et al .
@xcite consider a similar sequence defined in term of avoiding dyck paths and give the formula : @xmath105bz_1 = [ z^{n}]{l_\infty}= \sum_{k=0}^{(n-1 ) \div 2 } \frac{(-1)^k}{n - k}\binom{n - k}{k}\binom{2n-3k}{n-2k-1}\ ] ]
let us call the bijection from black white trees to zigzag free trees .
notice that the fourth equation removes a @xmath69 and the last equation adds a @xmath106 , keeping a balance between @xmath69 nodes and @xmath106 nodes on the leftmost branch .
@xmath107\\ t}}\right ) & = & \raisebox{.4cm}{\xymatrix @=.05pc{{`*}\ar@{-}[dr]\\ & { \scriptsize { \ensuremath{\mathsf{bwtobz}}}}(t)}}\qquad \text{when~ } t = \raisebox{.4cm}{\xymatrix @=.05pc{&{`(!)}\ar@{-}[dl]\\ u}}\\ { \ensuremath{\mathsf{bwtobz}}}\left(\raisebox{.4cm}{\xymatrix @=.05pc{&{`(!)}\ar@{-}[dl]\\ t}}\right ) & = & { \ensuremath{\mathsf{bwtobz}}}(t)\qquad \text{when~ } t = \raisebox{.4cm}{\xymatrix @=.05pc{&{`(?)}\ar@{-}[dl]\\ u}}\\ { \ensuremath{\mathsf{bwtobz}}}\left(\raisebox{.4cm}{\xymatrix @=.05pc{&{`(?)}\ar@{-}[dl]\ar@{-}[dr]\\ t&&t'}}\right ) & = & \raisebox{.4cm}{\xymatrix@=.05pc{&{`*}\ar@{-}[dl]\ar@{-}[dr]\\ { \scriptsize { \ensuremath{\mathsf{bwtobz}}}}(t)&&{\scriptsize { \ensuremath{\mathsf{bwtobz}}}}(t ' ) } } \qquad \text{when~ } t = \raisebox{.4cm}{\xymatrix@=.05pc{&{`(?)}\ar@{-}[dl]\ar@{-}[dr]\\ u_1 & & u_2}}\\ { \ensuremath{\mathsf{bwtobz}}}\left(\raisebox{.4cm}{\xymatrix @=.05pc{{`(?)}\ar@{-}[dr]\\ & t}}\right ) & = & \raisebox{.4cm}{\xymatrix@=.05pc{&{`*}\ar@{-}[dl]\ar@{-}[dr]\\ ` * & & { \scriptsize { \ensuremath{\mathsf{bwtobz}}}}(t)}}\end{aligned}\ ] ] we use two functions @xmath108 and @xmath109 . notice also that on the leftmost branch a @xmath69 is added and a @xmath106 is removed ; @xmath110\\&t}}\right ) & = & \raisebox{.4cm}{\xymatrix@=.05pc{&{`(!)}\ar@{-}[dl]\\{\ensuremath{\mathsf{bztobw}}}_{`(!)}(t ) } } \qquad \text{when~ } t = \raisebox{.4cm}{\xymatrix@=.05pc{&{`*}\ar@{-}[dl]\ar@{-}[dr]\\ u_1 & & u_2}}\\ { \ensuremath{\mathsf{bztobw } } } _ { ` ( ! ) } \left(\raisebox{.4cm}{\xymatrix@=.05pc{&{`*}\ar@{-}[dl]\ar@{-}[dr]\\t&&t'}}\right ) & = & \raisebox{.4cm}{\xymatrix@=.05pc{&&{`(!)}\ar@{-}[dl]\\ & ` ( ? ) \ar@{-}[dl]\ar@{-}[dr]\\ { \ensuremath{\mathsf{bztobw}}}_{`(?)}(t)&&{\ensuremath{\mathsf{bztobw}}}_{`(!)}(t ' ) } } \qquad \text{when~ } t = \raisebox{.4cm}{\xymatrix@=.05pc{&{`*}\ar@{-}[dl]\ar@{-}[dr]\\ u_1 & & u_2}}\\\\ { \ensuremath{\mathsf{bztobw } } } _ { ` ( ? ) } ( ` * ) & = & \box\\ { \ensuremath{\mathsf{bztobw}}}_{`(?)}\left(\raisebox{.4cm}{\xymatrix@=.05pc{&{`*}\ar@{-}[dl]\ar@{-}[dr]\\t&&t'}}\right ) & = & \raisebox{.4cm}{\xymatrix@=.05pc{&{`(?)}\ar@{-}[dl]\ar@{-}[dr]\\ { \ensuremath{\mathsf{bztobw}}}_{`(?)}(t)&&{\ensuremath{\mathsf{bztobw}}}_{`(!)}(t ' ) } } \qquad \text{when~ } t = \raisebox{.4cm}{\xymatrix@=.05pc{&{`*}\ar@{-}[dl]\ar@{-}[dr]\\ u_1 & & u_2}}\\\end{aligned}\ ] ] @xmath111 and @xmath112 .
lest us call @xmath113 this bijection .
it is described in figure [ fig : ltobz ] @xmath114\\ & { ` * } } } \\\\ { \ensuremath{\mathsf{ltobz}}}(`l(m ) ) & = & \raisebox{.4cm}{\xymatrix @=.05pc{&{\ensuremath{\mathsf{ltobz}}}(m)\ar@{-}[dl]\\{`*}}}\\\\ { \ensuremath{\mathsf{ltobz}}}(m\,{\mbox{$0\hspace*{-4pt}\raisebox{.5pt}{$\cdot$}$\ , } } ) & = & \raisebox{.4cm}{\xymatrix @=.05pc{&{`*}\ar@{-}[dr]\ar@{-}[dl]\\ { ` * } & & { \ensuremath{\mathsf{ltobz}}}(m ) } } \\\\ { \ensuremath{\mathsf{ltobz}}}(m\,s(n ) ) & = & \raisebox{.4cm}{\xymatrix @=.05pc{{\ensuremath{\mathsf{ltobz}}}(n)\ar@{-}[dr]\\&{`*}\ar@{-}[dr]\ar@{-}[dl]\\ ` * & & { \ensuremath{\mathsf{ltobz}}}(m)}}\\\\ { \ensuremath{\mathsf{ltobz}}}(m_1\,m_2 ) & = & \raisebox{.4cm}{\xymatrix @=.05pc{&&t\ar@{-}[dl]\\&{`*}\ar@{-}[dr]\ar@{-}[dl]\\ ` * & & { \ensuremath{\mathsf{ltobz}}}(m_1)}}\qquad \text{when~ } { \ensuremath{\mathsf{ltobz}}}(m_2 ) = \raisebox{.4cm}{\xymatrix@=.05pc{&t\ar@{-}[dl]\\ ` * } } \end{aligned}\ ] ] the bijection called @xmath115 is defined in figure [ fig : bztol ] .
@xmath116\\&`*}}\right ) & = & s ( { \ensuremath{\mathsf{bztol}}}\left(n\right ) ) \\ { \ensuremath{\mathsf{bztol}}}\left(\raisebox{.4cm}{\xymatrix @=.1pc{&`*\ar@{-}[dl]\\`*}}\right ) & = & ` l{\mbox{$0\hspace*{-4pt}\raisebox{.5pt}{$\cdot$}$\,}}\\ { \ensuremath{\mathsf{bztol}}}\left(\raisebox{.4cm}{\xymatrix @=.1pc{&`*\ar@{-}[dl]\ar@{-}[dr]\\`*&&t}}\right ) & = & { \ensuremath{\mathsf{bztol}}}(t)\,{\mbox{$0\hspace*{-4pt}\raisebox{.5pt}{$\cdot$}$\,}}\\ { \ensuremath{\mathsf{bztol}}}\left(\raisebox{.4cm}{\xymatrix @=.1pc{n\ar@{-}[dr]\\&`*\ar@{-}[dl]\\`*}}\right ) & = & ` l\ , { \ensuremath{\mathsf{bztol}}}\left(\raisebox{.4cm}{\xymatrix @=.1pc{n\ar@{-}[dr]\\&`*}}\right ) \\ { \ensuremath{\mathsf{bztol}}}\left(\raisebox{.4cm}{\xymatrix @=.1pc{n\ar@{-}[dr]\\&`*\ar@{-}[dl]\ar@{-}[dr]\\ ` * & & t}}\right ) & = & { \ensuremath{\mathsf{bztol}}}(t)~`l\ , { \ensuremath{\mathsf{bztol}}}\left(\raisebox{.4cm}{\xymatrix @=.1pc{n\ar@{-}[dr]\\&`*}}\right ) \\ { \ensuremath{\mathsf{bztol}}}\left(\raisebox{.4cm}{\xymatrix @=.1pc{&&t\ar@{-}[dl]\\&`*\ar@{-}[dl]\\`*}}\right ) & = & ` l\ , { \ensuremath{\mathsf{bztol}}}\left(\raisebox{.4cm}{\xymatrix @=.1pc{&t\ar@{-}[dl]\\`*}}\right ) \\ { \ensuremath{\mathsf{bztol}}}\left(\raisebox{.4cm}{\xymatrix @=.1pc{&&t_2\ar@{-}[dl]\\&`*\ar@{-}[dl]\ar@{-}[dr]\\`*&&t_1}}\right ) & = & { \ensuremath{\mathsf{bztol}}}(t_1)~{\ensuremath{\mathsf{bztol}}}\left(\raisebox{.4cm}{\xymatrix @=.1pc{&t_2\ar@{-}[dl]\\`*}}\right ) \end{aligned}\ ] ] @xmath117 and @xmath118 .
let us look at the bijection on classical examples , namely , , @xmath16 and : @xmath119\\ & & ` * \ar@{-}[dl]\\ & ` * \ar@{-}[dl]\\ ` * } & \textrm{for~ } \mathsf{s } \xymatrix@=.2pc{&&&&&`*\ar@{-}[dl]\ar@{-}[dr]\\ & & & & ` * \ar@{-}[dl]\ar@{-}[dr]&&`*\ar@{-}[dr]\\ & & & ` * \ar@{-}[dl]&&`*\ar@{-}[dr]\ar@{-}[dl]&&`*\\ & & ` * \ar@{-}[dl]&&`*&&`*\ar@{-}[dr]\\ & ` * \ar@{-}[dl]&&&&&&`*\ar@{-}[dr]\\ ` * & & & & & & & & ` * } \end{array}\ ] ] @xmath120\ar@{-}[dr]\\ & & ` * \ar@{-}[dl]&&`*\\ & ` * \ar@{-}[dl]\ar@{-}[dr]\\ ` * & & ` * \ar@{-}[dl]\ar@{-}[dr]\\ & ` * \ar@{-}[dl]&&`*\\ ` * } & \textrm{for~ } \mathsf{y } \xymatrix@=.2pc { & & & ` * \ar@{-}[dl]\ar@{-}[dr]\\ & & ` * \ar@{-}[dl]\ar@{-}[dr]&&`*\\ & ` * \ar@{-}[dl]\ar@{-}[dr]&&`*\ar@{-}[dr]\\ ` * & & ` * \ar@{-}[dl]\ar@{-}[dr]&&`*\\ & ` * \ar@{-}[dl]\ar@{-}[dr]&&`*\\ ` * & & ` * \ar@{-}[dr]\\ & & & ` * } \end{array}\ ] ] + + we leave the straightforward implementation of from @xmath2-terms to zigzag free trees to the reader .
we are now interested in normal forms , that are terms irreducible by @xmath121 reduction that are also terms which do not have subterms of the form @xmath122 .
there are three associated classes : @xmath123 ( the normal forms ) , @xmath124 ( the neutral terms , which are the normal forms without head abstractions ) and @xmath26 ( the de bruijn indices ) : @xmath125 let us call @xmath126 the generating function of @xmath123 , @xmath127 the generating function for @xmath124 and @xmath128 the generating function for @xmath26 .
the above equations yield the equations for the generating functions : @xmath129 one shows that @xmath130 @xmath127 is the generating function of motzkin trees ( see @xcite p. 396 ) .
in this section we exhibit a bijection between motzkin trees and neutral normal forms as suggested by the identity between their genrating functions .
let @xmath131 denote the unary motzkin path of height @xmath94 .
we start with defining two auxiliary operations and , translating unary motzkin paths into @xmath132-paths and debruijn indices , respectively .
@xmath133\\ u_n } } } \right ) & = & \raisebox{.6 cm } { { \xymatrix @=.6pc { \lambda \ar[d]\\ { \ensuremath{\mathsf{untol}}}\left(u_n\right ) } } } \end{aligned}\ ] ] @xmath134\\ u_n } } } \right ) & = & \raisebox{.6 cm } { { \xymatrix @=.6pc { s \ar[d]\\ { \ensuremath{\mathsf{untod}}}\left(u_n\right ) } } } \end{aligned}\ ] ] using and we can now define ( figure [ fig : motone ] ) the translation from motzkin trees into corresponding neutral terms .
@xmath135 & \\ & \bullet \ar[dl ] \ar[dr ] & \\
t & & t ' } } } \right ) & = & \raisebox{.9 cm } { { \xymatrix @=.9pc { & @ \ar[dl ] \ar[dr ] & \\ { \ensuremath{\mathsf{motone}}}\left(t\right ) & & { \ensuremath{\mathsf{untol}}}\left(u_n\right ) \ar[d ] \\ & & { \ensuremath{\mathsf{motone}}}\left(t'\right ) } } } \\ { \ensuremath{\mathsf{motone}}}\left(\raisebox{.5 cm } { { \xymatrix @=.9pc { & \bullet \ar[dl ] \ar[dr ] & \\
t & & t ' } } } \right ) & = & \raisebox{.6 cm } { { \xymatrix @=.9pc { & @ \ar[dl ] \ar[dr ] & \\ { \ensuremath{\mathsf{motone}}}\left(t\right ) & & { \ensuremath{\mathsf{motone}}}\left(t'\right ) } } } \\ { \ensuremath{\mathsf{motone}}}\left(u_n\right ) & = & { \ensuremath{\mathsf{untod}}}\left(u_n\right ) \end{aligned}\ ] ] @xmath136 is injective .
the proposition is an easy consequence of the fact that @xmath136 preserves the exact number of unary and binary nodes .
what remains is to give the inverse translation from neutral terms to motzkin trees ( figure [ fig : netomo ] ) .
let and denote the inverse functions of and respectively .
let @xmath137 denote the unary @xmath132-path of height @xmath94 and @xmath138 denote the @xmath94-th debruijn index .
the translation is given by : @xmath139 \ar[dr ] & \\
t & & l_n \ar[d ] \\ & & t ' } } } \right ) & = & \raisebox{.9 cm } { { \xymatrix @=.9pc { & { \ensuremath{\mathsf{ltoun}}}\left(l_n\right ) \ar[d ] & \\ & \bullet \ar[dl ] \ar[dr ] & \\ { \ensuremath{\mathsf{netomo}}}\left(t\right ) & & { \ensuremath{\mathsf{netomo}}}\left(t'\right ) } } } \\ & & \qquad \text{where } t ' \text { does not start with a } \lambda\\ { \ensuremath{\mathsf{netomo}}}\left(\raisebox{.6 cm } { { \xymatrix @=.9pc { & @ \ar[dl ] \ar[dr ] & \\
t & & t ' } } } \right ) & = & \raisebox{.5 cm } { { \xymatrix @=.9pc { & \bullet \ar[dl ] \ar[dr ] & \\ { \ensuremath{\mathsf{netomo}}}\left(t\right ) & & { \ensuremath{\mathsf{netomo}}}\left(t'\right ) } } } \\ { \ensuremath{\mathsf{netomo}}}\left(d_n\right ) & = & { \ensuremath{\mathsf{dtoun}}}\left(d_n\right ) \end{aligned}\ ] ] @xmath140 and @xmath141 .
+ + in order to translate motzkin trees to corresponding neutral terms we have to consider two cases .
either we are given a motzkin tree starting with a unary node or a binary one .
the later case is straightforward due to the fact that binary nodes correspond to neutral term application .
assume we are given a motzkin tree starting with a unary path @xmath131 of size @xmath94 .
we have to decide whether the path corresponds to a debruijn index or a chain of @xmath2-abstractions .
this distinction is uniquely determined by the existence of the path s _ splitting node _ the binary node directly below @xmath131
. if @xmath131 has a splitting node then it corresponds to a chain of @xmath94 @xmath2-abstractions which will be placed on top of the corresponding right neutral term constructed recursively from @xmath131 s splitting node .
otherwise , @xmath131 corresponds to the @xmath94-th debruijn index .
we leave the straightforward implementation of from neutral terms to motzkin trees to the reader .
we are now interested in the set of head normal forms @xmath142 which yields the equations @xmath143 and @xmath144 from which we draw @xmath145 this can be explained by the following bijection ( see figure [ fig : bij3 ] and figure [ fig : bij4 ] ) : if @xmath146 is a neutral head normal form , it is of the form : * @xmath147 with @xmath148 ( of size @xmath149 ) then it is in bijection with @xmath150 ( of size @xmath151 ) , * @xmath152 ( of size @xmath149 ) then it is in bijection with @xmath153 ( of size @xmath151 ) , * @xmath154 ( of size @xmath0 ) , treated by the case @xmath155 .
from theorem [ theorem : approx ] we get : @xmath156k \sim \left(\frac{1}{{\rho_{l_\infty}}}\right)^n \frac{c}{n^{\frac{3}{2}}}\ ] ] with @xmath157 and @xmath158 .
the density of a set @xmath159 in a set @xmath160 is @xmath161 where @xmath162 ( respectively @xmath163 ) are the numbers of elements of @xmath159 ( respectively of @xmath160 ) of size @xmath94 .
for instance the density of @xmath164 in @xmath24 is @xmath165k}{[z^n]{l_\infty}};\ ] ] hence the proposition .
the density of @xmath166 in @xmath24 ( i.e. , the density of neutral head normal forms among plain terms ) is @xmath167 .
@xmath37h \sim \left(\frac{1}{{\rho_{l_\infty}}}\right)^n \frac{c_h}{n^{\frac{3}{2}}}\ ] ] with @xmath168 the proof is like this of theorem [ theorem : approx ] with @xmath169 figure [ fig : approxh ] compares the coefficients of @xmath170 with its approximation .
@xmath44h}}\qquad\qquad\qquad \\\hline 10 & 1902\\ 20 & 118768916 \\ 30 & 12338289374047\\ 40 & 1552505356757052270\\ 50 & 216408050593408223194666\\ 60 & 32156818736630052190010494575\\ 70 & 4992016749940033843389032870415375\\ 80 & 800041142163881275363093897487465240590\\ 90 & 131362728872240507612558556757894820073668254\\ 100 & 21984069003048322712483528437236630547685953755064\\ \hline\hline \textbf{n } & \mathit{\mathbf{\lfloor ( 1/{\rho_{l_\infty}})^n c_h / n^{3/2}\rfloor } } \qquad\qquad\qquad\\\hline 10 & 1581 \\ 20 & 109732518 \\ 30 & 11727010776119 \\ 40 & 1495436887319673848 \\ 50 & 210083497584679365571791 \\ 60 & 31376820974748144171493861802 \\ 70 & 4888522574435898663355075650509052 \\ 80 & 785558576073780985739070920824898277393 \\ 90 & 129252413184969184232722751628403772087829182 \\ 100 & 21666626365243195881127917362969390314273901016408 \end{array}\ ] ] the density of @xmath171 in @xmath24 ( i.e. , the density of head normal forms among plain terms ) is @xmath172 actually @xmath173
consider a term @xmath174 of size @xmath175 and the set @xmath176 of terms that contain @xmath174 as subterm . @xmath177 which yields @xmath178 and @xmath179 notice that @xmath180 then the discriminant is @xmath181 in the interval @xmath182 , @xmath183 is decreasing ( its derivative is negative ) and @xmath184 . hence the root @xmath185 of @xmath186 is larger than the root @xmath187 of @xmath183 , that is @xmath188 . beside : @xmath189 hence the number of terms that do not have @xmath174 as subterm is given by @xmath190 the density in @xmath24 of terms that do not have @xmath174 as subterm is @xmath1 . indeed
the smallest pole of @xmath191 is @xmath185 and the smallest pole of @xmath23 is @xmath187 .
therefore , @xmath192({l_\infty}- t ) & \bowtie & \left(\frac{1}{`r_t}\right)^n\\ { [ z^n ] } { l_\infty}&\bowtie&\left ( \frac{1}{`r_{{l_\infty}}}\right)^n \end{aligned}\ ] ] hence , since @xmath188 @xmath193({l_\infty}- t)}{{[z^n ] } { l_\infty } } = \left(\frac{`r_{{l_\infty}}}{`r_t}\right)^n = 0.\ ] ] for instance if @xmath194 , that is for instance if @xmath195 , then @xmath196 and @xmath197 the density in @xmath24 of terms that contain @xmath174 as subterm is @xmath0 .
asymptotically almost no @xmath2-term is strongly normalizing . in other words , _
the density of strongly normalizing terms is @xmath1 .
_ indeed , the density in @xmath24 of terms that contain @xmath198 is @xmath0 .
hence the density of non strongly normalizing terms is @xmath0 .
hence the density of strongly normalizing terms is @xmath1 .
figure [ fig : summary ] summarizes what we obtained on densities of terms . [ cols="^,^,^,^ " , ] = normal forms + = neutral head normal forms = head normal forms = terms containing subterm = strongly normalizing terms @xmath199 = non strongly normalizing terms moreover , this research opens many issues , among others about generating random terms and random normal forms using boltzmann samplers @xcite . 1 nicolas bourbaki . .
elements of mathematics .
springer - verlag , 2004 .
philippe flajolet and robert sedgewick .
. cambridge university press , 2008 .
katarzyna grygiel and pierre lescanne .
counting terms in the binary lambda calculus . , abs/1401.0379 , 2014 . published in the proceedings of _ 25th international conference on probabilistic , combinatorial and asymptotic methods for the analysis of algorithms _ 2014 , https://hal.inria.fr/hal-01077251 .
nancy s. s. gu , nelson y. li , and toufik mansour .
2-binary trees : bijections and related issues . , 308(7):12091221 , 2008 .
pierre lescanne .
boltzmann samplers for random generation of lambda terms .
technical report , ens de lyon , 2014 .
http://hal-ens-lyon.archives-ouvertes.fr/docs/00/97/90/74/pdf/boltzmann.pdf .
bruno salvy and paul zimmermann .
gfun : a maple package for the manipulation of generating and holonomic functions in one variable .
, 20(2):163177 , 1994 .
aristidis sapounakis , ioannis tasoulas , and panagiotis tsikouras .
ordered trees and the inorder traversal .
, 306(15):17321741 , 2006 .
de bruijn indices are a system of notations for bound variables due to nikolaas de bruijn and somewhat connected to those proposed by bourbaki @xcite .
the goal is to replace bound variables by placeholders and to link each bound variable to its binder . for instance ( see figure [ fig : bourbaki ] ) bourbaki ( @xcite p. 20 ) proposes to represent placeholders by boxes @xmath83 and to represent binds by drawn lines .
this requires a two dimensional notation .
for example , he considers the formula : @xmath200 notice that we use an infix notation whereas he uses a prefix notation which gives @xmath201 .
the formula contains the binder @xmath202 ( a binder that bourbaki introduces ) and two occurrences of the bound variable @xmath203 , this involves two @xmath83 s and two drawn lines from @xmath202 , namely to the first @xmath83 and to the second @xmath83
. de bruijn proposes to represent the placeholders ( in other words the variables ) by natural numbers which represent the length of the link , that is the number of binders crossed when reaching the actual binder of the variables . in our proposal
, we write natural numbers using the functions _ zero _
@xmath8 and _ successor _ @xmath7 . for instance , @xmath73 is written @xmath204 . with de bruijn notations , bourbaki s formula
is written : @xmath205 and the lambda terms @xmath9 is written @xmath10 which would correspond to the drawing of figure [ fig : bourbakis ] in bourbaki style . .
] @xmath206@/_.1pc/@{-}@(u , u ) [ rrr ] & ` l\ar@[green]@{-}@(u , u ) [ rrrr ] & ` l \ar@[red]@{-}@(u , u ) [ rr ] \ar@{-}@(u , u ) [ rrrr ] & ( \box & \box ) & ( \box & \box)}\ ] ]
another natural counting is a counting where : @xmath207 the generating function is solution of @xmath208 with discriminant @xmath209 and with root closest to @xmath1 : @xmath210 and @xmath211 .
the first values are : | we study the sequences of numbers corresponding to lambda terms of given sizes , where the size is this of lambda terms with de bruijn indices in a very natural model where all the operators have size @xmath0 . for plain lambda terms
, the sequence corresponds to two families of binary trees for which we exhibit bijections .
we study also the distribution of normal forms , head normal forms and strongly normalizing terms .
in particular we show that strongly normalizing terms are of density @xmath1 among plain terms . * * keywords:**lambda calculus , combinatorics , functional programming , test , random generator , bijection , binary tree , asymptotic |
sunyaev - zeldovich decrements in the cosmic microwave background are produced by the thomson scattering of the photons by hot electrons - typically by electrons that are confined in the dark matter potential well of a rich cluster of galaxies .
however , we demonstrate in this work that comparable temperature decrements can be produced by electrons that are swept up by the outflows powered by the mechanical luminosity of quasars . while the properties of the outflow are determined primarily by the energy output of the qso , the kinematics at late times are determined by the mean density and the density gradient in the vicinity of the qso .
since in variants of the standard cdm ( cold dark matter ) dominated cosmological models , quasars form from the collapse of high - significance peaks and tend to seed the high density regions which subsequently are the sites for the formation of clusters , the physical scale of the outflow depends on the ambient density and depth of the local potential well .
there are at present many new observations of the cmb in the millimeter wave - band planned or underway , and there is renewed interest in the radio morphology of high - redshift qsos and their hosts . in this paper
, we expand on the classical s z scenarios ( clusters , clusters with peculiar velocities , relativistic corrections to the scattering cross - sections and warm bubbles around qsos with finite peculiar velocity ) and include another class of objects that may produce such decrements . for this class of objects : quasar outflows , it is instructive to point out that the gas need not be very hot in order to produce an observable decrement .
in particular , when the cooling / expansion time - scales are interestingly long , i. e. if the cooling time is comparable to or exceeds the time - scale of the high luminosity phase of the qso , the effects of the gas can be observed well after the qso has faded .
this paper is organized as follows : in section 2 we motivate the role of quasar - driven outflows ; in section 3 the energetic constraints in terms of the implied physical scales out to which the outflows and their effects are likely to manifest themselves are outlined ; in section 4 we discuss the kinematic and the thermal s z decrement induced by outflows and their other observational implications ; and we present our conclusions and prospects for detection in the final section .
the ubiquity and high luminosity of quasars out to high redshifts suggests that they might have played an important role in the subsequent formation of structure ( efstathiou & rees 1988 ; rees 1993 ; silk & rees 1998 ) .
babul & white ( 1991 ) suggest that the ionization produced by uv photons from the quasar might inhibit galaxy formation in their vicinity due to the ` proximity ' effect .
thus quasar activity is likely to be one of the important non - gravitational processes to have influenced structure formation at redshifts @xmath6 .
we propose quasar - driven outflows as a possible physical mechanism for concentrating baryons .
such a scenario has been explored previously in the context of explosion induced structure formation models ( ostriker & cowie 1981 ; daly 1987 ) as well as in theoretical models of the inter - stellar medium ( mckee & ostriker 1977 ) .
mechanically powered outflows can push baryons around active quasars in shells cleaning out regions of @xmath7 .
we examine the effectiveness of this mechanism in the production of massive clumps of baryons that can in turn produce a signal in the cmb .
the observational evidence for these outflows are seen in a subset of radio - quiet qsos - the broad absorption line systems ( hazard et .
al 1984 ; weymann et .
al 1991 ) in which speeds v / c @xmath8 ( 0.01 0.1 ) are inferred during a wind - phase .
these ` hyper - active ' quasars can create both regions that are baryon - rich as well as voids via their energetic outflows .
this powered outflow model is compatible with both scenarios of quasar population evolution : the single generation of long - lived quasars as well as the multiple generation model wherein populations fade and flare ( haehnelt & rees 1994 ) .
superwinds of the kind detected by heckman et al .
( 1996 ) from galaxies can also in principle produce s z decrements , but they are typically an order of magnitude weaker than the qso outflows proposed here .
the total energy available to power a quasar can be directly estimated from the mass accreted by a black hole of mass @xmath9 ( modulo an efficiency factor , @xmath10 ) . in general , the total energy available is 10 - 20% of the rest mass of the central black hole ( @xmath11 ) .
we assume that roughly 50% ( @xmath12 ) of the total energy in turn might be available as mechanical power . by assumption
this sets up an outflow that is primarily mechanically powered , that is , a fraction of the energy emitted by the active quasar gets efficiently converted into kinetic energy of the intergalactic medium .
mechanical luminosities up to @xmath13 at redshifts @xmath14 for such qsos are postulated , this is not entirely implausible given inferred x ray luminosities of comparable magnitude for high @xmath15 qsos ( elvis et .
al 1994 ) .
clearly such energy coupled locally to the environment as mechanical luminosity would set up an outflow that sweeps out a region of scale @xmath16 : @xmath17 for our purposes typical swept up regions are of the order of 0.11 mpc or so ; @xmath18 is the local baryon fraction in units of 0.01 , and a density contrast of @xmath19 over the mean cosmological density at redshift @xmath20 is required for the numbers to be consistent , this is a plausible over - density for the highest significance peaks that will seed the massive clusters observed at lower redshifts . the total energy budget @xmath21 ; and the restrictions on the available parameter space in terms of the mass that needs to be swept up and the bulk velocity required to produce a kinematic s z decrement are dictated by the following relations , @xmath22 where @xmath23 is the quasar lifetime and @xmath24 the optical depth .
beam dilution imposes a minimum @xmath25 assuming the line of sight length through the gas is comparable to its transverse scale , which is likely for the resulting outflow geometry studied here . scaling to typical abell cluster parameters ,
the mean electron thermal speeds are @xmath26 . since the kinematic effect is first order in @xmath27 , we require @xmath28 .
this is consistent with speeds required for the mass of swept up gas which would produce the required optical depths over angular scales of a few arc minutes .
it is these mass and velocity requirements which drive us to postulate ultra massive central black holes to power the outflows .
note that @xmath29 whereas the mass swept up @xmath30 .
all the numbers quoted in this work have been computed for @xmath31 ; @xmath32 and @xmath33 . as illustrated by eqns . 2 and 3 , the best sources for large energy effects are the massive bhs that power the qsos
. we also note here that according to some models of unified schemes for agns , jets and efficient energy release occurs only if the qso is accreting close to or at super - eddington rates , possibly via the blandford - znajek mechanism for rotating bhs .
this might preferentially happen at high redshift , where a higher fraction of high luminosity qsos are observed , consistent with a picture wherein the most massive bhs form rapidly at high redshift and are fueled by the baryons that are present in these high - density regions .
we do not necessarily need fully assembled , virialized dark halos for this to happen , we only need an efficient way to funnel the baryons already there to the bottom of the deepest potential wells .
the outflow starts out being collimated close to the source and has some , small , opening angle so that it does not terminate or disrupt the accretion process which is its powerhouse . at a distance of @xmath34 away from the central engine ,
the flow opens out .
the mass @xmath25 ( of the igm ) swept out by the outflow is : @xmath35 where @xmath36 denotes the mechanical luminosity of the qso .
as was shown by phinney ( 1983 ) , jets from qsos start out mildly relativistic with a lorentz factor @xmath8 a few ; it is likely that the initial outflow is dominated by a pair plasma .
the total momentum that the outflow can deposit is determined primarily by the ratio of @xmath37 and therefore bulk speeds of the order of @xmath38 are expected , which are indeed observed at low redshifts even in low power radio galaxies like ngc 315 ( bicknell 1994 ) . the mass swept out by a single qso with an outflow assuming that 50% of its total bolometric luminosity is expelled mechanically is
, @xmath39 where @xmath40 is the duration of the outflow phase , @xmath41 the ejection velocity and @xmath42 the mechanical luminosity of the qso .
the total mass swept up is strictly bounded , @xmath43 while the initial outflow is likely to be a pair plasma , the number density of electrons in the bulk outflow is dominated by the swept up hydrogen from the igm and is estimated as , @xmath44 in good agreement with typical values obtained using the 151-mhz and 1.5-ghz data set of leahy , muxlow & stephens ( 1989 ) . the optical depth @xmath24 to thomson scattering is : @xmath45 with an approximate slab geometry as predicted in this outflow picture the gas mass required for a given optical depth can be several times smaller than that for an isothermal sphere whose virialized core provides the most of the optical depth for the clusters s z effect .
no dark matter is swept up , so the total mass involved is considerably less than the total cluster mass of a virialised cluster producing a comparable s z effect . as is evident from our approach , the micro - physics of the early stages of the jet propagation
have not been computed in detail .
we assume that the flow is initially collimated and breaks out on a scale of a few kpc and that approximately 50% of the energy released due to accretion is efficiently mechanically coupled to the environment .
the geometry of a typical region through which the outflow propagates is assumed to be a cosmologically overdense region with a strong external density gradient .
physically , this is required so that the initial outflow can remain collimated right out to the advancing surface where the bulk of the energy is deposited .
the advancing surface breaks out and expands once a critical ambient pressure is attained and the presence of the density gradient serves to anisotropize ( i. e. destroys the spherically symmetry of ) the post break - out expansion .
the s z decrement in the temperature of the cmb is produced by thomson scattering of the cmb photons by hot electrons encountered along the line - of - sight ( sunyaev & zeldovich 1970 ; 1972 ; 1980 ) .
the magnitude of the decrement is independent of the redshift of the hot plasma and is given by the integral of the pressure along the line - of - sight .
the effect has two components a * thermal * and a * kinematic * one ( sunyaev & zeldovich 1980 ) . the thermal s z effect causes fluctuations in the temperature and the intensity which are ( i ) frequency independent , of equal magnitude and the same sign in the rayleigh - jeans limit of the spectrum but ( ii ) are larger , differ in magnitude and depend on the frequency in the wien limit .
the overall sign of the contribution to the fluctuations ( in both the temperature and the intensity ) also differs in the rayleigh - jeans and the wien limit .
coherent streaming motion with a velocity @xmath46 with respect to the frame of reference in which the cmb is isotropic introduces a doppler shift in the scattered radiation producing the * kinematic effect * which additionally perturbs the temperature and intensity across the region .
this kinematic piece induces a frequency * independent * temperature gradient but a frequency * dependent * fluctuation in the intensity ; both of which have the same sign ( determined by the direction of the velocity vector ; note that if the outflow is spherical on large scales the kinematic s
z effect cancels as the line of sight passes through regions with opposite sign radial velocity ) .
sunyaev & zeldovich ( 1980 ) ; peebles ( 1993 ) , show that the decrement produced by the kinematic term is given by , @xmath47 where @xmath48 ; the kinematic decrement is first order in @xmath49 compared to the normal second order thermal effect produced by hot cluster gas . for typical values of the optical depth and the velocity of the outflow assumed in the above analysis , this yields a @xmath50 .
we stress here that * the ratio of the predicted thermal to the kinematic s z effects depends entirely on the bulk velocity and mean temperature of the plasma*. below we estimate the thermal component of the s z effect expected from a fiducial outflow . the material at the working surface may be shock - heated to temperatures as high as @xmath51 , but may cool down efficiently as we argue below .
the bulk of the mass ( which is the material in the igm ; note that only roughly a thousandth of the mass is expelled directly from the qso ) however is expected to be in a cooler phase .
the shock heated dense plasma will cool to a temperature @xmath52 such that its cooling time @xmath53 , where @xmath54 is its dynamical time ( rees & ostriker 1977 ) .
thermal brehmstrahlung is expected to be the primary process for energy loss in the early stages of the evolution of the bubble and x - ray fluxes @xmath55 are expected , which could be detectable for low - redshift bubbles ( see natarajan , sigurdsson & silk 1998 for a more detailed discussion ) .
x ray cooling is inefficient for these high power outflows .
the plasma can cool to lower temperatures rapidly via turbulent mixing which will tend to homogenize the temperature of the material on the dynamical time - scale .
synchrotron cooling is however , unlikely to be important , since for the typical estimated value of the igm magnetic field at high - redshifts @xmath56 g ( see kronberg 1994 ) the cooling time exceeds both the hubble time and the lifetime of the qso , therefore , unless there are significantly stronger b fields on very small scales energy losses via synchrotron processes will be negligible .
the thermal s z decrement produced by this warm plasma is : @xmath57 much lower than that expected for virialized gas in a core of a cluster of galaxies . for a range of frequencies , in fig . 1
, we plot the ratio of the expected kinematic and thermal components for the plasma in the outflow .
as shown by birkinshaw ( 1996 ) , the ratio of the thermal to the kinematic effect is independent of the optical depth and can be written as , @xmath58 therefore , the contribution from the kinematic effect could easily be one or even two orders of magnitude higher than the thermal effect , for warm ( @xmath59 ) plasma with significant bulk motion , as postulated here .
this mechanism might be relevant to a recent detection of a decrement in the region of the quasar pair pc1643 + 4631 where no cluster is detected .
the reported observation was obtained by jones et .
al ( 1997 ) using the ryle telescope at 15-ghz in the direction of the quasar pair pc1643 + 4631 a & b. the pair separation on the sky is @xmath8 200 and their @xmath60 magnitudes are 20.3 and 20.6 respectively .
the redshifts of the quasars are ( a ) @xmath61 and ( b ) @xmath62 ( schneider , schmidt & gunn 1994 ) . the central value of the measured decrement is 560 @xmath63k . interpreting the source as a cluster at @xmath8 5 kev , jones et .
al estimate the implied gas mass to be @xmath64 2 @xmath65 10@xmath66 @xmath67 and conclude that the result demonstrates the existence of a massive high redshift ( z @xmath64 1 ) system
. however , serendipitous and pointed rosat pspc observations ( kneissl et .
al 1997 ) as well as ground - based optical follow - ups ( saunders et .
al 1997 ) have failed to reveal a cluster of galaxies as the source of the hot plasma to produce this ` cold ' spot in the cmbr .
the hypothesis of lensing by an intervening massive cluster is therefore hard to justify .
another s - z decrement has been measured at the vla ( 3.6 cm ) by richards et .
al 1996 , also around a pair of radio - quiet qsos at a projected separation of 1 mpc . in this case
the pair have the same redshift ( z = 2.561 ) and although their spectral features are not identical their velocity separation being small , they are likely to be a lensed pair split by a high - redshift cluster with a large time delay of the order of 500 years .
the physical scales involved in both these observations are consistent with the role of quasar outflows and therefore our proposed model might provide a possible physical explanation . in a recent analysis aghanim et .
al ( 1996 ) , consider another scenario where the kinematic s
z effect is also much higher than the thermal s z effect .
they consider the bulk motion of highly ionized baryons in the proximity of qsos at speeds of @xmath68 and at temperatures of @xmath69 , essentially stationary bubbles around the qso moving with the qso at the local peculiar velocity .
although the physical picture that we present above is entirely distinct from that proposed by aghanim et .
al , we find the same scaling of the decrement with the relevant parameters : @xmath70 the magnitude of the kinematic s z effect from outflows scales primarily with the mechanical luminosity of the qso and the redshift , therefore , we expect stronger decrements preferentially from high luminosity and high redshift qsos .
bicknell ( 1994 ) finds for a local fri radio galaxy ngc 315 , @xmath71 ; @xmath72 and an associated jet that extends to scales of the order of @xmath73 ; our model considers qsos that are essentially the high - redshift , scaled - up ( scaled up in luminosity and hence the physical scale on which the outflow occurs ) versions of objects like ngc 315 .
we stress here that spherically symmetric outflows do not generate a kinematic s z signal , an anisotropy induced by the bubble expansion into a density gradient is required .
more importantly with every observed kinematic s z decrement a corresponding increment is produced due to the opposite sign bulk velocity on the other side of the qso .
several immediate predictions follow from our proposed scenario : while the temperature decrement is frequency independent , the intensity change is not ( shown explicitly in eqn . 8 above ) , so we expect a @xmath74 variation in @xmath75 relative to @xmath76 at say 15 ghz and 30 ghz .
we also expect a temperature @xmath77 of the same amplitude on the opposite side from the decrement ( see the marked ` hot ' spot hs and the cold spot as cs in fig .
2 ) , as we expect an outflow with the opposite sign radial velocity on the other side of the qso . since the optical depth @xmath24 and @xmath78 are of comparable order , both polarization effects ( see sunyaev & zeldovich 1980 ) arising due to ( i ) the finite angle subtended by the line - of - sight with respect to the velocity vector ( linear polarization - @xmath79 ) and ( ii ) the finite optical depth ( circular polarization @xmath80 ) are also roughly equal . therefore , we expect polarization at the @xmath81 level on 20 scales from flow transverse to the line of sight , along the line from the qso to the peak decrement . given the density and the temperature of the outflow material , the corresponding x - ray luminosity @xmath82 is expected to be @xmath83 , below the flux limit of current x - ray observations , and substantially redshifted to lower energies .
we also note that the presence of extended cool gas with @xmath84 as predicted by this outflow picture could be detected via absorption features seen blueward of the emission peak ( offset by @xmath85 km s@xmath86 ) in the qso spectrum .
given the available energy budget , diffuse lyman-@xmath87 emission from recombination in the warm gas is expected with approximate surface brightness within a narrow band filter ( width @xmath8 60 angstrom ) that is of the order of @xmath88 .
such diffuse emission may be detectable through narrow band imaging or long slit spectroscopy , it is interesting to note that somewhat smaller scale diffuse emission has been seen around both a low redshift , low luminosity radio loud qso 4c 03.24 ( van ojikpine et al .
1997 ) as well as in two high redshift radio loud qsos ( lehnert & becker 1998 ) .
the diffuse emission predicted for the qso pair ( pc1643 + 4631 ) is of comparable power to that seen around the quasar 4c 03.24 , but spread over a larger angular scale due to the greater spatial extent required of the outflow .
any fortuitous alignment of two qsos such that their outflows contribute additively to the s z decrement would make the effect unusually strong and hence detectable .
we would expect this effect to be rare in general , even in the proximity of bright , high redshift , radio - quiet qsos with the optimal opening angle ( @xmath89 ) .
it also strongly depends on the efficiency of the conversion of the bolometric qso luminosity to mechanical outflow with time and environment .
the luminosity of radio quiet qsos is seen to increase strongly with @xmath15 ( barvainis , lonsdale & antonucci 1996 ; taylor et .
al 1996 ) .
for instance , in the scheme of falcke et .
al ( 1996 ) the qsos in pc1643 + 4631 at their brightest would have been classified as friis with luminosities in excess of @xmath90 ( see falcke et .
al s figure 1 ) ; the radio jets have now faded and the accretion power decreased by many orders of magnitude , possibly due to depletion of the cooling flow in the inner kpc ( see eg .
fabian and crawford 1990 ) or decreased efficiency at lower mass accretion rates . since the outflow must not be spherical at large radii for a kinematic s z effect to be observable , substantial amounts of gas ought to remain in the vicinity of the qso in the regions outside the out - flowing cone , and would probably condense into the lyman absorbers or even a population of baryon - rich dwarf galaxies ( natarajan , sigurdsson & silk 1998 ) observed in the vicinity of qsos .
the presence of bright radio sources in the immediate vicinity could make this effect difficult to detect . in the context of unified agn models , the viewing angle implied for radio loud
qsos means looking straight down the jet , therefore detection of such extended bulk flows from them is impossible . hence one would only see such effects after the peak emission had diminished by several orders of magnitude and the extended radio lobes faded . in order to predict how
many such objects are expected to be detected , we need to estimate the number density of qsos which had strong outflows . if we believe the jet power scalings with luminosity and redshift predicted by models , the relevant number density is then of qsos with luminosities @xmath91 within the optimum redshift range @xmath92 . in order to not violate
( the admittedly weakly constrained ) local bh mass functions , we expect the ultramassive bhs needed for this scenario to have a mean space density of @xmath93 per mpc@xmath94 or less .
s z decrements measured in the vicinity of qsos where no clusters are detected either in the optical or the x - ray may be due to a concentration of baryons aggregated by the power of mechanical outflows from these quasars .
the above phenomenon leads to the following predictions ( note that the numbers below are for the * kinematic * s z decrement caused by the outflow from a * single * qso ) : * \(i ) it produces a frequency independent decrement @xmath0 ; * \(ii ) a frequency dependent fluctuation in the intensity @xmath1 a 20% variation relative to @xmath2 at say 15 ghz and 30 ghz ; * \(iii ) a temperature increment @xmath3 on the opposite side of the qso ; * \(iv ) a linear polarization signal over 20 at @xmath4 level ( pushing at the limits of the detection by the proposed new instruments on map and planck ) ; * \(v ) patches of signal @xmath3 on degree scales which will contribute to the power and hence affect the cmb multipole calculations in the rising region of the first doppler peak at @xmath5 for cdm - like models as well as defect models ; * \(vi ) we expect this to be a rare phenomenon , occurring only in 1 3 % of all high - redshift qsos ; and finally , * \(vii ) the thermal effect is one or two orders of magnitude smaller than the kinematic one . in principle
, the effect ought to be observable for qsos in cases with the appropriate opening angle and alignment of outflow to the line of sight , and preferentially for high mechanical luminosity , high redshift , radio - quiet qsos .
we note with interest the very large scale diffuse radio emission seen on two sides of a3667 ( rttgering et al .
if the source there is due to outflow from a now quiescent black hole in the central galaxy of the primary cluster , which is one of the possible explanations , then future observations should reveal a supermassive black hole in the cd , primarily through a sharply rising central dispersion in the inner regions of the galaxy . from the scale of the outflow ,
if it was powered by an agn outflow , a black hole mass of @xmath95 accreting at near eddington luminosity for at least a salpeter time is required , and the activity would have peaked @xmath96 years ago . since the physical model that we have presented here is speculative , we briefly summarize below the problems presented by the observations of s z decrements in qso fields with no detected clusters : \(i ) a wind - driven model composed of a pair plasma is somewhat analogous to the state of the bubble during the early phases of our proposition here but such material is too hot and would have a short lifetime , and , more importantly , the maximal scale out to which such plasma could be swept out is only of the order of @xmath97 100 kpc , insufficient to explain decrements on large scales @xmath97 1 mpc .
furthermore , such plasma should produce easily observable diffuse radio emission as it collides with the igm .
the observational implications of this variant have been discussed in natarajan , sigurdsson & silk ( 1998 ) .
\(ii ) it is possible that decrements can be caused by warm gas in the potential well of proto - clusters
. the gas would not be in virial equilibrium , such a scenario might fit with the essential outflow model we present here , but in the limit of a more modest spherical outflow .
any detected decrement would then be the signature of the hot gas formed by an asymmetric collision of the unconfined outflowing gas with cold infalling gas .
if the infall is strongly non spherical , as is likely in unvirialized proto clusters , this would produce a hot spot , therefore producing a one - sided thermal sz effect .
such a scenario is therefore easily distinguished from the pure kinematic effect discussed above , as the thermal decrement would show as a strong increment at shorter wavelengths and the predicted long wavelength increment on the opposite side of the qso would be absent .
although we have presented a somewhat speculative model , it makes several falsifiable predictions .
our model does require ultra - massive black holes , but that is not forbidden , in fact observations imply precisely that some of the high redshift qsos have total luminosities in excess of @xmath98 , implying central black hole masses of the order of @xmath99 .
we note that the estimated bh mass in the case of the nearby bl lac oj287 ( lehto & valtonen 1996 ) at z = 0.306 is @xmath100 , and this is unlikely to be the most massive black hole in the universe .
how ultramassive black holes could form at high redshift to produce such high luminosity sources ; and whether high luminosity qsos are particularly efficient at producing mechanical outflows are separate issues to be contemplated if further observations are consistent with the model proposed here .
we thank alastair edge , andy fabian , mike jones , richard mcmahon , martin rees and richard saunders for valuable discussions .
ss acknowledges the support of the european union , through a tmr cat .
30 marie curie personal fellowship .
we thank the referee sterl phinney for invaluable feedback that led to significant reorganizing of the paper .
aghanim , n. , desert , j. , puget , j. , & gispert , r. , 1996 , a&a , 311,1 babul , a. , & white s. d. m. , 1991 , mnras , 253 , 31 barvainis , r. , lonsdale , c. , & antonucci , r. , 1996 , aj , 111 , 1431 bicknell , g. , 1994 , apj , 422 , 542 birkinshaw , m. , 1996 , ` the sunyaev - zeldovich effect ' , phys .
reports , in press blandford , r. , phinney , e.s . , & narayan , r. , 1987 , apj , 313 , 28 daly , r.a . , 1987 , apj , 322 , 20 elvis , m. , fiore , f. , wilkes , b. , mcdowell , j. , 1994 , apj , 422 , 60 fabian , a.c . , crawford , c.s . , 1990 , mnras , 247 , 439 falcke , h. , gopal - krishna , & biermann , p. , 1996 , a&a , 298 , 395 goodrich , r. , 1997 , apj , 474 , 606 haehnelt , m.g . , rees , m.j . , 1993 , mnras , 263 , 168 hazard , c.,morton , d. , terlevich , r. , & mcmahon r. , 1984 , apj , 282 , 33 jones , m. , et .
al , 1997 , apj lett . , 479 , 1 kronberg , p. , 1994
, reports on prog . in physics , 57
, 325 kneissl , r. , et .
al , 1997 , preprint leahy , j. , muxlow , t. , & stephens , p. , 1989
, mnras , 239 , 401 lehnert , m. d. , & becker , r. h. , 1998 , a&a , 332 , 514 lehto , h. j. , & valtonen , m. j. , 1996 , apj lett . , 460 , 207 mckee , c. & ostriker , j. , 1977 , apj , 218 , 148 natarajan , p. , sigurdsson , s. , & silk , j. , 1998 , mnras , in press van ojik , r. , rttgering , h.j.a . , carilli , c.l . , miley , g.k . ,
bremer , m.n .
, macchetto , f. , 1997 , a & a , 321 , 389 ostriker , j. , & cowie , l. , 1981 , apj lett . , 243 , l127 peebles , p. , 1995 ,
principles of physical cosmology , princeton univ . press , princeton , pg . 587 .
phinney , e.s . , 1983 , phd thesis , univ . of cambridge phinney , e.s . ,
blandford , r.d . , 1986 ,
nature , 321 , 569 rees , m. j. , 1993 , proc .
usa , vol .
4840 rees , m. j. , & ostriker , j. , 1977 , mnras , 179 , 541 richards , e. , 1997 , aj , 113 , 1475 rttgering , h.j.a .
, wieringa , m.h . ,
hunstead , r.w . , ekers , r.d . , 1997 , mnras , 290 , 577 saunders , r. , et .
al , 1997 , apj lett . , 479 ,
5 schneider , d. , schmidt , m. , & gunn , j. , 1994 , astron .
j. , 107(4 ) , 1245 sunyaev , r. , & zeldovich , y. , 1970 , astrophys .
space sci .
, 7 , 3 sunyaev , r. , & zeldovich , y. , 1972 , comments astrophys .
space phys . , 4 , 173 sunyaev , r. , & zeldovich , y. , 1980 , mnras , 190 , 413 taylor , g. , dunlop , j. , hughes , d. , robson , e. , 1996 , mnras , 283 , 930 weymann , r. , morris , s. , foltz , c. & hewett , p. , 1991 ,
apj , 373 , 23 | sunyaev - zeldovich decrements in the temperature of the cosmic microwave background ( cmb ) are produced by thomson scattering of the cmb photons by the confined hot electrons in the intra - cluster medium encountered along the line - of - sight . in this paper , we propose an alternate physical process that can produce detectable thermal and kinematic sunyaev - zeldovich decrements around quasars or quasar pairs in regions where no assembled , virialized clusters are detected either in the optical or x - ray wave - band . invoking quasar outflows , we argue that both * thermal * and * kinematic * decrements can be produced by the baryons swept up by the mechanical luminosity of quasars .
in contrast to the case of decrements produced by the hot electrons confined within the cluster potential , the magnitude of these quasar - outflow induced effects depend primarily on the luminosity of the qso , and the physical scale of the outflow - which in turn depends both on the mean local density and the density gradient in the vicinity of the quasar .
quasar outflows produce ( i ) a frequency independent kinematic decrement @xmath0 ; ( ii ) a frequency dependent fluctuation in the intensity @xmath1 a 20% variation relative to @xmath2 , say at 15 ghz and 30 ghz ; ( iii ) a temperature increment @xmath3 on the opposite side of the qso ; ( iv ) a linear polarization signal over 20 at @xmath4 level ; and therefore ( v ) patches of signal @xmath3 on degree scales which will contribute to the power and hence affect the cmb multipole calculations in the rising region of the first doppler peak at @xmath5 for cdm - like models as well as defect models .
these effects are of interest since they are potentially observable in the context of the next - generation of cmb experiments - map & planck .
[ firstpage ] cosmology : cosmic microwave background diffuse radiation galaxies : quasars |
the connection between long - duration gamma - ray bursts ( grbs ) and type ic core - collapse supernovae ( sne ) has long been established , beginning with the association of grb980425 with sn1998bw @xcite .
subsequent associations between nearby grbs and spectroscopically - confirmed sne include grb030329/sn2003dh @xcite , grb031203/sn2003lw @xcite and most recently grb060218/sn2006aj @xcite .
these are spectroscopically confirmed examples of nearby grb - sn associations ( out to @xmath2 ) .
the characteristic ` bump ' of a supernova has been noted in the photometric data for a dozen or more grbs out to @xmath3 ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , while the majority of grbs lie at higher redshifts where such signatures would be impossible to detect ( @xmath4 * ? ? ?
* ; * ? ? ?
the grbs with sne are therefore rare , but provide a crucial insight into the grb
sn connection and the progenitors of long grbs .
the grbs with spectroscopically confirmed sne are generally underluminous and subenergetic in comparison to a typical long grb , though grb030329 is a notable exception to this .
the prompt emission of these bursts has a lower energy spectral peak than a typical grb @xcite and they are suggested to have less relativistic outflows , or be viewed more off - axis .
it has been suggested that the observed nearby ( @xmath5 ) , underluminous / subenergetic grbs , including grb100316d , may be drawn from a different population to the cosmological grb sample , motivated by estimates of the local grb rate several times greater than the rate of cosmological grbs ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
radio observations of grbs suggest that mildly relativistic ejecta and often high luminosities are the properties that distinguish the supernovae associated with grbs from the non - relativistic ordinary type ib / c supernovae @xcite .
grb060218 was a landmark long - duration , low - luminosity , soft spectral peak event , detected by the _ swift _
satellite @xcite with unprecedented multiwavelength coverage of the prompt emission .
the prompt spectrum was found to comprise both the non - thermal synchrotron emission ascribed to most grbs and thought to originate in the collision of fast - moving shells within the grb jet ( e.g. * ? ? ?
* ) , and a thermal component .
the presence of this thermal component , together with its evolution , led to the suggestion that we were observing the shock breakout of the supernova for the first time @xcite .
however , the non - thermal emission did not differ greatly from that of the x - ray flash class of soft grbs and an outflow speed close to the speed of light could be inferred @xcite , alternatively suggesting this to be an extension of the typical grb population and not requiring significantly slower ejecta or any special ( off - axis ) geometry @xcite .
@xcite speculate that the long - duration , low - luminosity events such as this one may point to a different central engine for these events compared with more typical grbs : a neutron star rather than a black hole ( see also * ? ? ?
* ; * ? ? ?
* ) . also of relevance to this discussion
are the supernovae for which no grb was detected , yet for which mildly relativistic ejecta provide a good explanation for the radio data : sn2007gr @xcite and sn2009bb @xcite . as grb jets
are thought to be highly collimated , it is expected that many will go undetected if the emission is directed away from our line of sight ( e.g. * ? ? ?
* ) , and a fraction will also lie below current detector sensitivity limits . @xcite
estimate the fraction of type ib / c sne with central engines at about one per cent , consistent with the inferred rate of nearby grbs .
in contrast , there are nearby grbs for which an accompanying supernova could be expected but none has been detected to extremely deep limits .
grbs 060505 and 060614 @xcite are the two best examples of this , highlighting the need for a greater understanding of the relationship between grbs and sne .
we present the discovery of a new grb - sn , grb100316d @xcite associated with sn2010bh @xcite .
this is an unusually long - duration , soft - spectrum grb positioned on a nearby host galaxy .
many of the properties of this grb appear unlike those of typical grbs , resembling rather grb060218 . in this paper
we report accurate astrometry for the grb - sn and the redshift of the underlying host galaxy .
we examine the grb prompt and afterglow emission observed with all 3 instruments on - board _ swift _ , and the broad characteristics of the host galaxy as observed with the gemini south telescope , the very large telescope ( vlt ) and the hubble space telescope ( hst ) .
we compare these properties with the grb sample as a whole and with a subset of grb - sne , to understand the origins of the grb emission and further our knowledge of the grb
sn connection .
on 2010 march 16 at 12:44:50 ut ( hereafter t@xmath6 ) , the _ swift _ burst alert telescope ( bat , @xcite ) triggered on and slewed immediately to grb100316d @xcite .
this was an image trigger , since no rapid , substantial rise in the count rate occurred , but a possible low - level peak about 100 s long was noted in @xcite .
the x - ray telescope ( xrt , @xcite ) on - board _ swift _ began observing the field at 12:47:08.1 ut , 137.7 s after the bat trigger , beginning with 10 s of windowed timing ( wt ) mode settling data and followed by 593 s of wt pointing data .
after a time gap due to observing constraints , data - taking recommenced at t@xmath6 + 33 ks with xrt in photon counting ( pc ) mode .
the best x - ray position for this source is the astrometrically corrected xrt position of ra , dec ( j2000 ) = 07h 10 m 30.63s , -56d 15 19.7@xmath7 , with an uncertainty of 3.7@xmath7 ( radius , at 90% confidence * ? ? ?
_ swift s _ ultraviolet - optical telescope ( uvot , @xcite ) began observing with the white filter 148 s after the bat trigger and continued observations in u band , thereafter cycling repeatedly through all its lenticular filters .
the _ swift _ data were processed and analysed with the _ swift _ tools version 3.5 , released as part of heasoft 6.8 , and the latest calibration data at 2010 march were used .
xrt wt spectra were grouped with a minimum of 20 counts in each bin , allowing the use of @xmath8 statistics , while for the pc spectrum 1 count per bin was required and cash statistics used .
we fit the x - ray spectra over the energy range 0.310 kev and the @xmath1-ray spectra over the range 1550 kev ( as this is a soft event very few counts are present above 50 kev ) using the spectral fitting package xspec @xcite .
we used the x - ray absorption models phabs and zphabs for which photoionization cross - sections are taken from @xcite and solar abundances from @xcite .
the galactic absorption in the x - ray band is fixed at 7@xmath910@xmath10 @xmath11 throughout @xcite .
we tested for any differences in the spectral fitting results from use of different grade selections or different energy ranges ( excluding 0.30.5 kev and/or including 50150 kev ) and confirm that these have no effect .
optical / uv photometry was performed with the uvot tool uvotsource . at the position of the sn reported by @xcite
, background subtracted count rates were extracted using a circular source aperture of 2.5@xmath7 .
source - free background was taken from a region situated outside of the host galaxy .
the count rate was aperture corrected to 5@xmath7 in order to be compatible with the uvot calibration @xcite .
with uvot we detect the host galaxy , but no new point sources .
the 3@xmath12 upper limit on any optical transient emission was taken to be the background subtracted count rate plus three times the count rate error and was converted into magnitude using the zero points provided in @xcite . at this time , galaxy subtraction can not be performed since there are no uvot pre- or post - explosion observations .
the host galaxy count rate was determined using an elliptical source region , which enclosed the whole galaxy , and a source - free background region .
photometry of the host was performed on a sum of all data from @xmath13 s after the trigger onwards for each filter .
since the ellipse is similar in area to a 5@xmath7 circular aperture ( semi - major and -minor axes of 7.5@xmath7 and 5.5@xmath7 respectively ) , we have taken the standard uvot co - incidence loss correction to be the best estimate of the co - incidence loss , and the standard uvot zero points have been used to convert the count rate into magnitudes . our first optical observations of grb100316d were obtained with gemini - south , starting on march 16 23:53 ut . in these images we identified two point sources ,
consistent with the initial xrt position @xcite .
we acquired spectroscopy of these with the 3-arm echelle spectrograph x - shooter @xcite , mounted on unit telescope 2 of the vlt .
these sources were later shown to be compact hii regions unrelated to the supernova . in this paper
we will make use of the spectra obtained of the brightest of these two regions , referred to as source ` a ' in @xcite ( see figure [ subtraction ] ) .
we obtained two spectra each of 1200 s exposure of source a , using a 5@xmath7 nod throw along the slit to achieve better sky subtraction .
we used slit widths of 1.0 , 1.0 and 0.9@xmath7 for the uvb , vis and nir arms , respectively .
data were reduced using version 0.9.4 of the eso x - shooter pipeline @xcite .
calibration data ( bias , dark , arc lamp , flatfield and flexure control frames ) were taken the same night .
extracted spectra were flux calibrated using spectrophotometric standard star exposures , also taken the same night , resulting in flux calibrated spectra spanning the ultraviolet to the near - infrared ( @xmath14 nm ) simultaneously .
we obtained further imaging of the localization with gemini - south on the subsequent three nights , using these we were able to identify a further bright point source superimposed on the host galaxy @xcite .
image subtraction performed between the different images showed a clear brightening in this source , which we subsequently identified as the associated rising supernova ( * ? ? ?
* see figure [ subtraction ] ) .
we obtained deep imaging of the position of grb100316d using the fors2 instrument on unit telescope 1 of the vlt .
seeing conditions during both epochs were good , averaging 0.8@xmath7 , but the second epoch observations suffer from fairly bright background caused by high airmass and a bright moon .
the grb - sn is clearly visible in both epochs .
following the announcement of a sn discovery @xcite , we triggered our cycle 17 hst programme .
here we report the first epoch , obtained on march 25 , @xmath159 days after the bat trigger .
the details of the ground - based and hst observations can be found in table [ table : opticallog ] .
the data were reduced using standard procedures in iraf . [ cols=">,<,^,^ " , ] in section [ sec : cfgrbsne ] we found that among the grb - sne , there is a clear commonality in the x - ray behaviour of 100316d and 060218 . comparing the prompt emission spectra in section [ sec : spec ] , we again find parallels in that both bursts seem to require a similar thermal component in addition to the typical grb synchrotron emission for which the synchrotron peak is observed to move to lower energies with time .
the time - evolving prompt spectral parameters are compared in figure [ epkevolcf060218 ] .
both events have low isotropic equivalent energies , of order 4@xmath910@xmath16 erg , and very little flux above 50 kev .
these two events lie in apparently quite different environments : 060218 in a faint , compact dwarf @xcite and 100316d in a luminous , disturbed possibly spiral host galaxy .
the metallicities are also different , with the host of 100316d likely being more metal - rich ( though the metallicity is below solar ) . a prompt optical component , as observed in 060218 ,
could not be detected in 100316d , but conditions for its detection were far less optimal in this case due to the superposition on a brighter host galaxy and the possibility of a higher extinction along the line - of - sight .
the higher intrinsic x - ray column density we measure for 100316d compared with similar modelling of 060218 does not necessarily imply a higher optical extinction @xcite . in summary , both 100316d and its predecessor 060218
are nearby ( @xmath17 ) , long - duration ( @xmath181300 s , 2100 s ) , initially relatively constant in x - ray flux ( figure [ xrtlcs_long ] ) , spectrally soft ( very few or no counts above 100 kev ; low @xmath19 , figure [ epkevolcf060218 ] ) , subenergetic ( both have @xmath20 erg ) grbs with an associated sn .
these two events show similar prompt emission properties and stand out among the grb - sne subsample considered here for their unusual x - ray evolution . however , their host galaxies are rather different in morphology and metallicity , with the host of 100316d more closely resembling the host of 980425 .
the thermal x - ray component , with a luminosity , temperature and radius similar to that observed in 060218 , dictates that the shock break - out of the supernova must be considered .
the optical / uv thermal component observed in 060218 is not seen here : if the extinction is similar we would expect a shock breakout similar to 060218 to be two or more times fainter in 100316d and with the relatively brighter host galaxy it is perhaps no surprise that early optical emission is not detected with uvot .
the presence of a thermal component in the prompt @xmath1-ray emission of batse ( burst and transient source experiment ) grbs , in addition to non - thermal emission , was proposed by @xcite .
the discovery of thermal components in the _ swift _ xrt x - ray spectra shown here for 100316d and in @xcite for 060218 will therefore be important to consider in future studies of grb prompt emission .
the dominant component of the high energy emission in 100316d remains the synchrotron - like non - thermal spectrum common to all types of grb ( with and without sne ) thought to originate in internal shocks in a relativistic jet .
the long duration of the early x - ray emission is curious , and exceedingly rare , perhaps suggesting a greater reservoir of material is available to feed the central engine and prolong its activity .
the discovery of grb100316d and its associated supernova sn2010bh and the analyses of its high energy emission and host galaxy properties presented here illustrates the diversity in grb - sne characteristics that must be understood if we are to fully appreciate the relationship between grbs and core - collapse sne .
this work made use of data supplied by the uk _ swift _ science data centre at the university of leicester .
based on observations made with eso telescopes at the la silla or paranal observatories under programme ids 084.d-0939 , 083.a-0644 and 084.a-0260(b ) . based on observations made with the nasa / esa hubble space telescope , obtained at the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas 5 - 26555 : these observations are associated with programme 11709 .
we thank stsci staff for their efforts in implementing the hst too observation , particularly alison vick .
we acknowledge the wider _ swift _ team for their many contributions .
rlcs , kw , ar , jpo , klp and pae acknowledge financial support from stfc
. financial support of the british council and platform beta techniek through the partnership programme in science ( pps ws 005 ) is gratefully acknowledged ( kw ) .
the dark cosmology centre is funded by the danish national research foundation .
dnb acknowledges nasa contract nas5 - 00136 .
ajvdh was supported by an appointment to the nasa postdoctoral program at the msfc , administered by oak ridge associated universities through a contract with nasa .
references alard c. , 2000 , a&as , 144 , 363 amati l. , 2006 , mnras , 372 , 233 amati l. , della valle m. , frontera f. , malesani d. , guidorzi c. , montanari e. , pian e. , 2007 , a&a , 463 , 913 arnaud k. , 1996 , aspc , 101 , 17 asplund m.m .
, grevesse n. , sauval a.j . ,
allende - prieto c. , kiselman d. , 2004 , a&a , 417 , 751 band d. et al . , 1993 , apj , 413 , 281 barthelmy s. et al .
, 2005 , space sci .
, 120 , 143 blustin a.j .
et al . , 2006 ,
apj , 637 , 901 bufano f. et al . , 2010 ,
10543 burrows d.n .
et al . , 2005 , space sci .
, 120 , 165 cantiello m. , yoon s .- c . , langer n. , livio m. , 2007 , a&a , 465 , l29 campana s. et al .
, 2006 , nature , 442 , 1008 campana s. , thoene c.c . , de ugarte postigo a. , tagliaferri g. , moretti a. , covino s. , 2010 , mnras , 402 , 2429 chapman r. , tanvir n.r .
, priddey r.s . ,
levan a.j . , 2007 , mnras , 382 , l21 chornock r. et al . , 2010 , apjl submitted , arxiv:1004.2262 christensen l. , vreeswijk p.m. , sollerman j. , thne c.c .
, l floch e. , wiersema k. , 2008 , a&a , 490 , 45 cobb b.e .
, bailyn c.d . , van dokkum p.g . , natarajan p. , 2006
, apj , 645 , l113 costa e. et al . , 1997 ,
nature , 387 , 783 de pasquale m. et al .
, 2006 , a&a , 455 , 813 della valle m. et al . , 2006 ,
apj , 642 , l103
dodorico s. et al . , 2006 ,
proceedings of the spie , volume 6269 , p.98 evans p.a .
et al . , 2007 ,
a&a , 469 , 379 evans p.a .
et al . , 2009 , mnras , 397 , 1177 fan y .- z . , zhang b .- b . , xu d. , liang e .- w . ,
zhang b. , 2010 , apj submitted , arxiv:1004.5267 ferrero p. et al . , 2006 , a&a , 457 , 857 fruchter a.s .
et al . , 2006 , nature , 441 , 463 fynbo j.p.u .
et al . , 2006 , nature , 444 , 1047 fynbo j.p.u .
et al . , 2009 ,
apjs , 185 , 526 galama t.j .
. , 1998 , nature , 395 , 670 gal - yam a. et al .
, 2006 , nature , 444 , 1053 gehrels n. et al .
, 2004 , apj , 611 , 1005 gehrels n. et al . , 2008 ,
apj , 689 , 1161 ghisellini g. , ghirlanda g. , tavecchio f. , 2007 , mnras , 382 , l77 goldoni p. et al . , 2006 ,
spie , 6269 , 80 gonzalez delgado r.m .
, leitherer c. , heckman t.m . , 1999 , apjs , 125 , 489 hammer f. , flores h. , schaerer d. , dessauges - zavadsky m. , le floch e. , puech m. , 2006 , a&a , 454 , 103 hjorth j. et al .
, 2003 , nature , 423 , 847 hurkett c.p .
et al . , 2008 ,
apj , 679 , 587 izotov y.i . , thuan , t.x . ,
lipovetsky v.a . , 1994 ,
apj , 435 , 647 jakobsson p. et al . , 2006 , a&a , 447 , 897 kalberla p.m.w . , burton w.b . , hartmann d. , arnal e.m . , bajaja e. , morras r. , poppel w.g.l . , 2005 , a&a , 440 , 775 kaneko y. et al . , 2007 , apj , 654 , 385 kennicutt r.c . , 1998 , ara&a , 36 , 189 kulkarni s.r .
et al . , 1998 ,
nature , 395 , 663 levan a.j .
, tanvir n.r . , davanzo p. , vergani s.d .
, malesani d. , 2010 , gcn circ .
10523 levesque e.m . ,
berger e. , kewley l.j . ,
bagley m.m .
, 2010 , aj , 139 , 694 liang e. , zhang b. , virgili f. , dai z.g .
, 2007 , apj , 662 , 1111 malesani d. et al . , 2004 ,
apj , 609 , l5 mazzali p.a .
et al . , 2005 , science , 308 , 1284 mazzali p.a .
et al . , 2006 , nature , 442 , 1018 mirabal n. , halpern j.p . , an d. , thorstensen j.r .
, terndrup d.m . , 2006 ,
apj , 643 , l99 modjaz m. et al . , 2006 ,
apj , 645 , l21 oates s.r .
, de pasquale m. , stamatikos m. , 2010 , gcn circ .
10520 osterbrock , 1989 , astrophysics of gaseous nebulae and active galactic nuclei , university science books paragi z. et al . , 2010 , nature , 463 , 516 pei y.c . , 1992 ,
apj , 395 , 130 perley d.a .
, foley r.j . , bloom j.s . , butler n.r . , 2006 ,
5387 pian e. , antonelli l.a . , piro l. , feroci m. , 1998 , gcn circ . 158 pian e. et al . ,
2000 , apj , 536 , 778 pian e. et al .
, 2006 , nature , 442 , 1011 piro l. et al .
, 2005 , apj , 623 , 314 poole t.s .
et al . , 2008 ,
mnras , 383 , 627 prochaska j.x .
et al . , 2004 , apj , 611 , 200 rees m.j .
, mszros p. , 1994 ,
apj , 430 , l93 roming p.w.a .
et al . , 2005 , space sci .
, 120 , 95 ryde f. , 2004 , apj , 614 , 827 sakamoto t. et al . , 2008 , apjs , 175 , 179 sari r. , piran t. , narayan r. , 1998 , apj , 497 , l17 schlegel d.j . , finkbeiner d.p . ,
davis m. , 1998 , apj , 500 , 525 soderberg a.m. et al . , 2006 , nature , 442 , 1014 soderberg a.m. et al .
, 2010 , nature , 463 , 513 sollerman j. , stlin , g. , fynbo j.p.u . , hjorth j. , fruchter a. , pedersen k. , 2005 , newa , 11 , 103 sollerman j. et al . , 2006 , a&a , 454 , 503 stanek k.z .
et al . , 2003 ,
apj , 591 , l17 stamatikos m. et al . , 2010 ,
10496 starling r.l.c .
et al . , 2008 ,
mnras , 384 , 504 starling r.l.c .
, evans p.a .
, stamatikos m. , 2010 , gcn circ .
10519 stratta g. , perri m. , siegel m.h .
, krimm h.a . ,
barthelmy s.d . ,
roming p. , burrows d.n . ,
gehrels n. , 2009 , gcn report 212.1 svensson k.m . , levan a.j . ,
tanvir n.r . ,
fruchter a.s .
, strolger l.g . , 2010 , mnras in press tanvir n.r .
, 2010 , apj in press , arxiv:0812.1217 toma k. , ioka k. , sakamoto t. , nakamura t. , 2007 , apj , 659 , 1420 vergani s. , levan a.j . ,
davanzo p. , covino s. , malesani d. , hjorth j. , tanvir n.r .
, antonelli l.a . , 2010 ,
10513 verner d.a . ,
ferland g.j . ,
korista k.t . ,
yakovlev d.g . , 1996 ,
apj , 465 , 487 watson d. et al .
, 2004 , apj , 605 , l101 watson d. , hjorth j. , fynbo j.p.u .
, jakobsson p. , foley s. , sollerman j. , wijers r.a.m.j .
, 2007 , apj , 660 , l101 waxman e. , mszros p. , campana s. , 2007 , apj , 667 , 351 wiersema k. et al . , 2007 , a&a , 464 , 529 wiersema k. , davanzo p. , levan a.j .
, tanvir n.r .
, malesani d. , covino s. , 2010 , gcn circ .
10525 willingale r. et al . , 2004 , mnras , 349 , 41 wilms j. , allen a. , mccray r. , 2000 , apj , 542 , 914 woosley s.e .
, bloom j.s . , 2006 , ara&a , 44 , 507 zeh a. , klose s. , hartmann d.h . , 2005 , in proc .
22nd texas symp . on relativistic astrophysics , ed .
p. chen et al .
( stanford : slac ) , arxiv : astro - ph/0503311 | we report the _ swift _ discovery of the nearby long , soft gamma - ray burst grb100316d , and the subsequent unveiling of its low redshift host galaxy and associated supernova .
we derive the redshift of the event to be @xmath0 and provide accurate astrometry for the grb - sn .
we study the extremely unusual prompt emission with time - resolved @xmath1-ray to x - ray spectroscopy , and find that the spectrum is best modelled with a thermal component in addition to a synchrotron emission component with a low peak energy .
the x - ray light curve has a remarkably shallow decay out to at least 800 s. the host is a bright , blue galaxy with a highly disturbed morphology and we use gemini south , vlt and hst observations to measure some of the basic host galaxy properties .
we compare and contrast the x - ray emission and host galaxy of grb100316d to a subsample of grb - sne .
grb100316d is unlike the majority of grb - sne in its x - ray evolution , but resembles rather grb060218 , and we find that these two events have remarkably similar high energy prompt emission properties .
comparison of the host galaxies of grb - sne demonstrates , however , that there is a great diversity in the environments in which grb - sne can be found .
grb100316d is an important addition to the currently sparse sample of spectroscopically confirmed grb - sne , from which a better understanding of long grb progenitors and the grb
sn connection can be gleaned .
[ firstpage ] gamma - ray burst : individual : grb100316d ; supernovae : individual : sn2010bh |
the research was supported by the russian scientific foundation grant ( n14 - 50 - 00043 ) and rfbr grants ( n15 - 42 - 02573 , 16 - 02 - 00567 , 16 - 32 - 00210 ) .
sf acknowledges support of the russian government program of competitive growth of kazan federal university . | the origin of ultraluminous x - ray sources ( ulxs ) in external galaxies whose x - ray luminosities exceed those of the brightest black holes in our galaxy by hundreds and thousands of times is mysterious .
the most popular models for the ulxs involve either intermediate mass black holes ( imbhs ) or stellar - mass black holes accreting at super - eddington rates .
here we review the ulx properties , their x - ray spectra indicate a presence of hot winds in their accretion disks supposing the supercritical accretion .
however , the strongest evidences come from optical spectroscopy .
the spectra of the ulx counterparts are very similar to that of ss433 , the only known supercritical accretor in our galaxy .
ultraluminous x - ray sources ( ulxs ) are x - ray sources with luminosities exceeding the eddington limit for a typical stellar - mass black hole @xmath0 erg s@xmath1 . despite their importance in understanding the origin of supermassive black holes that reside in most of present galaxies , the basic nature of ulxs remains unsolved.@xcite the most popular models for the ulxs involve either intermediate mass black holes ( imbh , @xmath2@xmath3m@xmath4)@xcite with standard accretion disks or stellar - mass black holes ( @xmath5m@xmath4 ) accreting at super - eddington rates .
the last idea has been suggested@xcite because of an analogy with ss433 , the only known super - accretor in the galaxy,@xcite and developed in ( ref .
13 , 18 ) .
it was proposed that ss433 supercritical disk s funnel being observed nearly face - on will appear as extremely bright x - ray source . both scenarios
, however , require a massive donor in a close binary .
most of the ulxs are associated with the star - forming regions and surrounded with nebulae of a complex shape , indicating a dynamical influence of the black hole.@xcite they are not distributed throughout galaxies as it would be expected for imbhs originating from low - metallicity population iii stars .
the imbhs may be produced in a runaway merging in a core of young clusters . in this case
, they should stay within the clusters .
it has been found@xcite that all brightest x - ray sources in antennae galaxies are located nearby to very young stellar clusters .
it was concluded that the sources were ejected in the process of formation of stellar clusters in the dynamical few - body encounters and that the majority of ulxs are massive x - ray binaries with the progenitor masses larger than 50m@xmath4 .
the x - ray spectra of the ulxs often show a high - energy curvature with a downturn between @xmath6 and @xmath7kev .
it was called `` ultraliminous state''.@xcite the curvature hints that the ulx accretion disks are not standard .
inner parts of the disks may be covered with hot outflow or optically thick corona , which comptonizes the inner disk photons .
spectra of almost all optical counterparts of studied ulxs ( with ss433 included ) are shown in fig.1 .
main features in all the spectra are the bright heii@xmath8 , hydrogen h@xmath9 , and h@xmath10 emission lines .
the lines are obviously broad ; the widths range from 500 to 1500 km s@xmath1 .
some hydrogen lines are contaminated by nebulae ( over - subsraction of narrow absorptions ) associated with the ulxs .
however the broad emission components in h@xmath9 are clearly seen . in the ss433 spectrum , besides moving lines which form in the relativistic jets , stationary emission lines , the bowen blend ciii / niii @xmath11 and three hei lines are observed .
some of these lines are seen in the ulx spectra.@xcite all the spectra of the ulxs ( the spectra were reduced by us ) are surprisingly similar to one another .
the optical spectra are also similar to that of ss 433 , although the ulx spectra indicate a higher wind temperature . it was suggested in ( ref .
6 ) that the ulxs must constitute a homogeneous class of objects , which most likely have supercritical accretion disks .
the total luminosity of a supercritical disk is proportional to the eddington luminosity with an additional logarithmical factor depending on the original mass accretion rate,@xcite because the excess gas is expelled as a disk wind and the accreted gas is advected with the photon trapping , contributing little to the photon luminosity .
however , the uv and optical luminosity in such disks may strongly depend on the original mass accretion rate , because these budgets are mainly produced by the reprocess of the strong irradiation from the wind ( the excess gas ) . optical spectra of ss433 and the ulx counterpart are nearly the same , but in x - rays they are drastically different because we can not observe the funnel in ss433 directly .
it was found@xcite that the mass accretion rates in the ulxs may be by a factor of 1.56 smaller and their wind temperatures are by 1.44 times higher than those in ss433 . in fig.2
we show absolute magnitude of all well studied ulxs together with ss433 .
thus , one may interpret that ss433 is intrinsically the same as ulxs but an extreme case with a particularly high mass accretion rate , which could explain the presence of its persistent jets.@xcite in fig.3 we present one more evidence of the super eddington accretion in ulxs .
the power density spectrum of ss433 exhibit a flat part in the @xmath12hz frequency range.@xcite the presence of such a part is related to the abrupt change in the disk structure and the viscous time at the spherization radius . in this place
the accretion disk becomes thick , which reduces drastically the time of passage of matter through the disk.@xcite the same picture is observed in the well - studied ulx ngc5408 x-1 .
we need longer observations of the ulxs to study their power density spectra in more details . |
theoretical models of the evolution of cataclysmic variables ( cvs ) make predictions about the distributions of their physical parameters such as orbital period , mass ratio and the individual masses of the two stars .
if these parameters can be measured , the ( many ) assumptions that go into the theory can be tested .
these assumptions are of interest in the broader context of binary star evolution .
the same parameters are needed to gain an accurate understanding of the accretion processes which dominate these stars .
unfortunately , the only reliable physical parameter that can be measured for the majority of cvs is their orbital period , with perhaps the spectral type of the donor star coming a distant second . in this ranking ,
the masses of the white dwarf and the donor star are so uncertain that they are often barely mentioned in the context of evolution .
this is partly because of the difficulty of measuring orbital inclinations common to all forms of close binary , but also because it is hard to measure the the radial velocities of the binary components .
the optical spectra of most cvs are dominated by emission from accreting material , rather than the white dwarf . in the case of non - magnetic systems , the emission lines are formed in discs , and often have a double - peaked shape resulting from the doppler shift of matter rotating in a keplerian disc .
the high - velocity wings of the emission lines are expected to form close to the white dwarf , where it is hoped the flow is axi - symmetric .
if so , these wings should trace the motion of the white dwarf .
this is the idea behind many studies of cvs .
however , this method fails in practice : in those systems in which one has independent knowledge of the orbital phase , the emission line radial velocities invariably fail to match expectations ( stover 1981 ) .
nevertheless , the white dwarf velocities measured in this way are still used for mass determination , perhaps because there has not been any direct evidence for the magnitude of the distortion .
this changed after long & gilliland ( 1999 ) measured directly the radial velocity semi - amplitude of the white dwarf in u gem from narrow photospheric lines visible in the ultraviolet . in this case , the phase agreed perfectly with the predicted conjunction phase .
long & gilliland ( 1999 ) measured @xmath5 .
the direct nature of this measurement makes it the most accurate to date for any cv , and gives us the chance to test what the nature of the distortion of the emission line measurements in one system at least .
one other cv for which we have a good estimate of @xmath6 is ae aqr ( welsh et al 1995 ) . in this case @xmath6
is measured from the pulsations of the white dwarf .
however , ae aqr is a very unusual cv with flaring emission lines that are not suited to testing the general method of measuring @xmath6 from line emission .
u gem is interesting from several other points of view : it is a double - lined , partially - eclipsing system , and is one of the brightest dwarf novae .
it has prominent emission from the gas stream / disc impact region which has a velocity in - between that of a keplerian disc and a ballistic stream ( marsh et al .
finally , during outburst it has shown spiral shocks ( groot 2001 ) .
we will see that our new data has something to tell us on all of these issues .
the observations were taken in january 2001 at la palma ( see table [ tab : data ] for details ) .
the 4.2-m william herschel telescope ( wht ) was used with the double - beamed isis spectrograph .
two datasets , covering wavelength ranges around h@xmath7 and h@xmath8 were acquired using the highest resolution gratings .
the spectrograph s slit was oriented at a position angle of @xmath9 to include a comparison star for slit loss corrections . the spectrophotometric standard hd19445 and
one featureless star were observed during the nights with best seeing ( @xmath100.7 " ) for flux calibration and to remove the effect of telluric lines on the red data ( bessell 1999 ) .
flat - fields and comparison arc spectra were taken approximately every 60 minutes for the red and the blue dataset .
the weather was clear throughout most of our run with seeing of order 1 arcsecond for most of it .
[ cols="<,<,^,^,^,^,^,^ , > , > , > " , ] [ tab : low_velocity_emission ] simulations of the stream in u gem by kunze et al .
( 2001 ) predict that more than half the stream overflows the outer edge of the disc . with high signal - to - noise ratio and smooth disc profiles ,
our data are very well suited to detection of such overflow which can be expected to cause emisison along the path of the stream up to the point of closest approach to the white dwarf ( livio et al .
1986 ; armitage & livio 1998 ; lubow 1989 ) . in u
gem the point of closest approach is @xmath11 from the white dwarf , and in velocity space corresponds to a region in the lower right quadrant of the doppler maps of figs [ fig : dopplermaps_blue ] and [ fig : dopplermaps_red ] . to search more rigourously , we subtract the smooth background from the disc by using an elliptical isophote fitting procedure described in appendix [ app : ellipses ] . after performing the subtraction we obtained the tomograms seen in fig .
[ fig : continuum_subtracted ] , which are scaled to only 3% of the peak height of the raw tomograms .
the bright spot is clear in the maps , but there is no evidence of stream - disc overflow along the path of the stream even at this low level . in order to quantify this result we measured averages and peak values over several samples of the region
along which the hypothetical overflow would travel , before and after the subtraction of the elliptical continuum . by comparing those values
we conclude that any possible overflow would contribute less than @xmath12 to the flux from the disc surface .
[ fig : continuum_subtracted ] does show an excess extending up and to the left of the main disc / stream impact region to well over @xmath13 .
a similar effect was seen in ex dra ( billington & marsh 1996 ) who suggested that non - kinematic broadening , such as stark broadening , could be responsible .
figure [ fig : continuum_subtracted ] reveals the presence of weak spiral structure near the level of the continuum ( we refer here to the structure within @xmath14 of the centre of the images ) . in figure
[ fig : spirals_quiescence_vs_outburst ] we show the maps with contours of spiral shocks observed during outburst ( steeghs , morales - rueda & groot in preparation ) which line up well with the asymmetries that we see in quiescence .
this spiral structure has been previously reported for u gem in outburst ( groot 2001 ) , making it one of a handful of systems in which the phenomenon has been detected .
we emphasize that the amplitude of the spiral structure we find here is _
far _ less than that suggested by neustroev and borisov ( 1998 ) for the same system .
their data shows pronounced asymmetries , but were taken with poor spectral and phase resolution , and most importantly , without complete orbital coverage .
in contrast our raw maps are very symmetric .
we believe their claim of strong spiral structure to be incorrect , and a result of the difficulty of interpreting the complex variations of u gem with inadequate orbital coverage . the presence of spiral structure in the disc has been regarded as evidence of `` spiral shocks '' ( steeghs , harlaftis & horne 1998 ) , which had been predicted in simulations of accretion discs ( see matsuda et al .
2000 for a review ) .
the spiral shock interpretation requires the presence of large or unusually hot discs .
this prompted smak ( 2001 ) and ogilvie ( 2002 ) to propose alternatives to the spiral shocks to explain the phenomenon .
they suggest that the spiral features revealed by doppler tomography could be explained as the consequence of 3-body effects ( paczynski 1977 ) .
ogilvie extends the model of paczynski from 2 to 3 dimensions , allowing for a complete set of physical conditions in his model .
he concludes that , due to tidal distortions , some regions of the disc would thicken , later being irradiated by the white dwarf .
the pattern of thickening would be slightly spiral , but no waves or shocks would be involved in the process . in doppler tomography
the pattern would show up as the kind of open spiral structure seen in published tomograms .
these two competing explanations predict a different evolution of the spiral structures from outburst to quiescence .
the spiral shock model requires hot material to have open spirals . in quiescence , the spirals , if present , are expected to be relatively tightly wound ( steeghs and stehle 1999 ) .
assuming that the asymmetries that we see are indeed related to the outburst spiral structure , our observations show little change in morphology from outburst to quiescence , i.e. the spirals do not appear tightly wound .
this argues against the spiral shock model as an explanation of the spiral structure detected in u gem and favours ogilvie s ( 2002 ) and smak s ( 2001 ) explanation .
although similar patterns are seen in several lines , we remain cautious about claiming a definitive detection of spiral structure in the quiescent u gem .
we emphasize again that the maps are really extremely symmetrical , and given the presence of strong bright - spot components which have strengths modulated in orbital phase , it is hard to rule out the possibility that we are seeing low - level artefacts
. it is also worth noting that the size of the disc we find from the bright - spot of @xmath100.3a ( section [ sec : bright_spot ] ) is close to the minimum observed radius of u gem s disc ( smak 1984 ) and well inside the tidal radius where one would expect ogilvie s and smak s model to operate , let alone tidally - driven spiral shocks .
a notable feature in most of the tomograms , in particular in , is the mass donor s irradiated face .
small sections of these maps , centered on the donor star , are presented in the top row of figure [ fig : disc_shielding ] .
the figure shows that the irradiation - induced emission avoids the equator of the mass donor , indicative of shielding by the disc ( harlaftis 1999 ; morales - rueda et al .
the irradiation in and seems to be located further from the l1 point than for the other lines .
this is suggestive of a changing effective thickness due to wavelength dependent opacity at the euv wavelengths needed for photoionisation of the elements involved .
this has been previously seen in other cvs ( morales - rueda et al .
2000 ) . in the bottom row of figure [ fig : disc_shielding ]
we show simulated tomograms for a system with the same parameters of u gem but with varying height - to - radius ratio ( h / r ) in the disc .
we included some broadening in the simulations by convolving the synthetic datasets with a gaussian with fwhm @xmath15 .
values of h / r from 0.10 to 0.25 were considered ( increasing by 0.05 each time ) . for @xmath16 ,
the height completely prevents the irradiation to take place .
we estimate from this that the h / r ratio for u gem is @xmath100.15 ( ) , @xmath100.20 ( ) and @xmath100.25 ( and ) .
these values could be a little smaller if @xmath17 is larger than we have assumed ( raising the roche lobes of figure [ fig : disc_shielding ] ) , as suggested by the sharp spot in heii ( section [ sec : bright_spot ] ) . and
@xmath18 respectively .
the central panels show data from the balmer lines .
the central panels were processed to improve visibility of the extension of the eclipse ( see text for details ) .
all the panels are saturated at 100 per cent .
note that the data are affected by emission components not included in our simulations.,width=384 ] zhang & robinson ( 1987 ) derived an orbital inclination angle @xmath19 for u gem from a detailed lightcurve fitting model .
other estimates of @xmath20 have displayed a trend to somewhat smaller values , but mainly in the region of @xmath21 .
given the radial velocity semi - amplitiudes , the inclination determines the component masses .
long ( 2000 ) and more recently naylor et al ( 2005 ) encountered a problem when combining a measurement of the gravitational redshift of the white dwarf , which fixes @xmath22 , with other constraints including inclination estimates .
the problem , summarised in fig . 8 of naylor et al .
is that the gravitational redshift gives a rather small radius for a given mass , which only crosses the hamada - salpeter mass - radius relationship for @xmath23 implying at the same time a low inclination of @xmath24 and a radius well below the value predicted from u gem s uv spectrum and parallax ( long 2000 ; harrison 1999 ) .
the hamada - salpeter relation and the radius measured from the uv flux are in agreement for a much higher inclination of @xmath25 to @xmath26 and a lower white dwarf mass , @xmath27 ( naylor et al . 2005 ) , but then one must suppose that the gravitational redshift , a combination of careful measurements from the uv ( long & gilliland 1999 ) and i - band ( naylor et al .
2005 ) to be seriously in error .
in this section we use the eclipse of the emission lines , and specifically the highest velocity in the lines that is eclipsed to provide a new independent constraint on the orbital inclination .
we find a high inclination favouring the low - mass solution for the white dwarf .
the key point is that in a partially eclipsing system , since the eclipse does not reach the white dwarf , there will be a velocity in the lines above which there is no eclipse .
if we can determine this point , we can measure the inclination .
we do so by simulating the eclipse and comparing the simulation with the data ( figure [ fig : eclipses ] ) .
the top panel shows a simulated trailed spectra when @xmath20 is set to @xmath28 , with @xmath6 and @xmath29 set to the values mentioned before .
the three central panels show real data after removal of the mean spectrum .
the extension of the eclipse in these panels easily reaches @xmath30 while in the simulated data of the top panel it does not go beyond @xmath31 .
the inclination must be larger than @xmath32 ; we favour a value of @xmath33 ( lowest panel , figure [ fig : eclipses ] ) .
the emission line eclipse strongly favours the low - mass , high inclination solution and is in conflict with the measurements of gravitational redshift .
our data allows us to check the systemic - velocity of the donor star which is one component of the gravitational redshift .
we did this by subtracting from the dataset the orbital motion corresponding to the centroid of the mass donor s irradiated region .
then we averaged the spectra in phase ranges for which the separation between the signal from the irradiated region and other contaminating signals ( bright spot , disc ) was maximum .
this produced well defined peaks to which we fitted gaussian solutions .
we calculated the systemic - velocity of the donor as the offset of the centre of the gaussians with respect to their wavelengths at rest .
the result is @xmath34 ( average of , and ) .
this is compatible with @xmath35 from naylor et al .
( 2005 ) but , being higher , it reduces a little the value of the redshift . however , this reduction is far from enough to resolve the discrepancy between the two mass solutions for u gem .
3400 km @xmath36 seems appropriate to avoid most of the jump in noise close to 4000 km @xmath36 in the @xmath37/@xmath0 plot .
this separation corresponds to a plateau in the @xmath0 plot , from which the value @xmath1095 km @xmath36 is read for @xmath6 .
the dashed horizontal line in the top panel marks the level 107.1 km @xmath36.,width=288 ] seems appropriate in this case .
this points to a value of 150 km @xmath36 for @xmath6 .
the dashed horizontal line in the top panel marks the level 107.1 km @xmath36.,width=288 ] is seen twice in the plot
. shows no clear trend towards ( 0,107.1 ) km @xmath36 .
, width=480 ] the value of @xmath6 for u gem is known accurately thanks to a direct observation of metal lines from the white dwarf s surface ( long & gilliland , 1999 ) .
the result , @xmath38 , can be used to test the effectiveness of the techniques already in use , which are mostly based on assuming that the emission lines originated on the accretion disc will provide information from which the motion of the white dwarf can be tracked .
two traditional techniques to measure @xmath6 are the diagnostic diagram ( shafter , szkody & thorstensen 1986 ) and the light - centre diagram ( marsh 1988 ) .
both use the double - gaussian method ( schneider & young 1980 ; hereafter sy80 ) to measure the radial velocities from a dataset .
this method consists in convolving the continuum subtracted line with two gaussians of identical width and varying separations . in the diagnostic diagram technique , the velocities measured
are then fitted with an orbital solution @xmath39)$ ] , and the parameters of the fit are plotted in a diagram like those presented in figs .
[ fig : rv_diag_halpha ] and [ fig : rv_diag_hbeta ] .
@xmath6 is considered to coincide with @xmath0 when @xmath0 is stable over a range of gaussian separations ( a in the plot ) or at the point just before @xmath40 increases sharply . in figs .
[ fig : rv_diag_halpha ] and [ fig : rv_diag_hbeta ] we see that this happens for a = @xmath41 giving values of @xmath6 = 95 and @xmath42 for and respectively .
the plots in the diagnostic diagrams display curves that behave only partially as expected .
in particular there is no simultaneous convergence towards final values in the panels ( compare , for instance , top and bottom panel of figure [ fig : rv_diag_halpha ] ) . in the light - centre technique ,
the radial velocities measured using the double - gaussian method are plotted in velocity space and the line of points obtained is extrapolated towards the @xmath43 axis .
the value for @xmath6 is the point at which the extrapolated line crosses @xmath43 .
the plot obtained , not reproduced here , is very similar to figure [ fig : rv_new_method ] , which is described below .
a problem with the double gaussian measurement is that each spectrum is treated one by one , before the individual radial velocities are fitted by a sinusoid .
the process breaks down as soon as the signal - to - noise in one of the spectra becomes too low for the measurements to be made .
we can do somewhat better by fitting all spectra simultaneously , still employing the double - gaussian method because of its isolation of particular velocities in the line profiles . denoting the velocity , flux and uncertainty of the @xmath20-th pixel of the @xmath44-th spectrum by @xmath45 , @xmath46 and @xmath47 respectively
, we therefore miminise the following function : @xmath48^{2}\ ] ] where @xmath49 is the velocity of a pixel relative to the velocity of the spectrum @xmath50 , @xmath51 @xmath52 is a weighting function given by @xmath53 and @xmath54 represents two gaussians of standard deviation @xmath55 separated by @xmath56 , @xmath57 the function of eq .
[ eq : chi2 ] is minimised by fitting the velocities of the spectra @xmath50 , @xmath58 to @xmath59 .
if these quantities were fitted independently , then we are back to the standard double gaussian method for one spectrum which finds the velocity @xmath60 for which @xmath61 .
the difference is that we do not allow the @xmath50 to be independent , but instead constrain them to represent a circular orbit as so : @xmath62 fitting instead the standard orbital parameters @xmath63 , @xmath64 and @xmath43 .
uncertainties were estimated by a bootstrap procedure ( diaconis & efron 1983 ) , with 1000 bootstrap samples being chosen from the original set of spectra .
the end products are @xmath64 and @xmath43 values as a function of gaussian separation , exactly as with the old procedure .
the gain is greater robustness in the presence of noise .
figure [ fig : rv_new_method ] shows the application of this new method for performing the analysis of radial velocities .
we use a small cross to mark the point with coordinates ( 0,-107.1 ) @xmath65 , the direct measurement for @xmath6 from long & gilliland ( 1999 ) .
the points in the diagrams show the variation with separation for two gaussians of fwhm @xmath66 .
the initial separation is @xmath67 and is marked with a large grey circle .
small circles mark steps of @xmath68 while large circles marks steps of @xmath69 . in both panels
an inset with a blow up of the region close to ( 0,-107.1 ) @xmath65 has been included . in the insets
we have ommited the error bars .
although the plot does on two occasions shows trends towards the known value , it would be impossible in the absence of long & gillilands s value to deduce it from our measurements , although they are at least in its vicinity .
this should not come as a suprise ; the diagnostic diagrams have a phase offset from zero , and whatever causes the distortion will distort the new method as well .
the chief conclusion is that even far in the wings , there is something that distorts the radial amplitudes by of the order @xmath2 to @xmath68 , and that we can not improve on stover s ( 1981 ) estimate of @xmath70 .
we carried out several simulations to see what could disturb the wing velocities using the model applied in section [ sec : inclination_angle ] .
we began the simulations considering just an accretion disc and an opaque secondary star . in this way we obtained double - peaked phase - shifted spectra
in which an eclipse by the secondary star was included but no flux was contributed by it to the spectra .
we found that both techniques were able to accurately recover the correct value for @xmath6 from this dataset .
we then added a nearby line so that the separation and relative fluxes of both lines would approximate the situation of having or close to or respectively .
again both methods correctly extracted @xmath6 from the data
. then we added a bright spot ( its parameters tuned to fit the feature observed in our data ) to the main line ( -like and -like ) and found that the measurements were seriously distorted by its presence .
however , after increasing the gaussian separation beyond @xmath71 the methods again reliably extracted @xmath6 from the data .
lastly , we added a bright spot to the companion lines ( -like and -like ) and found that the recovery of @xmath6 was no longer possible . in summary
, we can not recover u gem s @xmath6 to better than @xmath72 , and we suspect that the problem is caused by the need to go far into the line wings to avoid the bright - spot , which brings with it the problem of contamination by the bright - spot emission of nearby lines .
we have presented high spectral resolution and signal - to - noise optical spectra of u gem with the aim of testing the use of the broad emission lines in tracing the motion of the accreting white dwarf in cvs .
u gem is particularly well suited for this task because the true motion of its white dwarf has been measured from _
_ spectra by long et al ( 1999 ) .
despite observations on a @xmath73m - class telescope of one the brightest dwarf novae , we find values of the semi - amplitude of the white dwarf ranging from 80 to @xmath74 which contain not a hint that the correct value is the @xmath75 measured by long et al ( 1999 ) . neither the diagnostic diagram ( shafter et al 1986 ) nor the light centre technique ( marsh 1988 ) were of any help .
the root cause of the problem seems to be the bright - spot which causes an asymmetry that extends to large velocities in doppler maps , and which is present in the same way in pairs of close spectral lines ( and ; and ) , thus violating the usual assumption of a trend towards symmetry at high velocity .
our data reveal a number of new phenomena for which we can offer only tentative explanations .
these include low velocity absorption and emission that we suggest imply the presence of prominences on the mass donor star .
u gem shows a complex structure in the region of the gas stream / impact region .
extended emission lies between the velocity expected for the ballistic gas stream directly and the velocity of the keplerian disc along the gas stream .
this emission extends both to very high and very low velocities in the maps .
most surprising of all is the presence of a narrow component of the bright - spot in , which appears to be unresolved even at @xmath76,@xmath77 .
this narrow spot appears to come from the disc rather than the stream , but seems to require a slightly larger value of radial velocity amplitude for the secondary star than has been measured to date .
we find no evidence of stream overflow , but we find weak spiral asymmetry in the ( quiescent ) disc , similar in nature to those seen during outburst
. this may provide support for ogilvie s and smak s explanation for the presence of spiral structure based upon 3-body effects .
we find an inclination angle of @xmath78 .
this supports a solution for the mass of u gem s white dwarf in agreement with the hamada - salpeter mass - radius relationship and with the radius measured from the uv flux .
this reinforces the discrepancy first noted by naylor et al .
( 2005 ) regarding the mass estimated from gravitational redshift .
this constitutes an important open puzzle .
eu was supported by pparc ( uk ) and fundacin andes ( chile ) under the program gemini pparc - andes throughout most of this research .
lmr was supported under a pparc pdra grant and trm under a pparc srf during some of the period over which this work was undertaken .
danny steeghs comments about sections of this article greatly improved the final version .
we also wish to thank him for his doppler map of u gem in outburst .
the authors acknowledge the data analysis facilities provided by the starlink project which is run by the university of southampton on behalf of pparc . this research has made use of nasa s astrophysics data system bibliographic services , as well as of the wonderful resources provided by hartley library at the university of southampton . during the long time that took the writing of this article , soledad martnez labrn had to use many times her incredible ability to make starry the darkest nights .
eu is lovingly indebted to her .
99 armitage , p. , livio , m. , 1998 , apj , 493 , 898 bessell , m. , 1999 , pasp , 111 , 1426 burke , p. , 1987
, byte magazine ( july ) cabrera , j. , meer , p. , 1996
, ieee trans .
pami , 18 , 752 collier cameron , a. et al . , 1999 , mnras , 308 , 493 diaconis , p. , efron , b. , 1983 , sci .
am . , v. 248 , n. 5 , 96 eggleton , p. , 1983 ,
apj , 268 , 368 fitzgibbon , a. , pilu , m. , fischer , r. , 1999 , ieee trans .
pami , 21 , 476 friend , m. , martin , j. , smith , r. , jones , d. , 1990 , mnras , 246 , 637 greenstein , j. , kraft , r. , 1959 , apj , 130 , 99 groot , p. , 2001 ,
apj , 551 , l89 harlaftis , e. , 1999 , a&a , 346 , l73 harrison , t. et al . , 1999 ,
apj , 515 , 93 jedrzejewski , r. , 1987 , mnras , 226 , 747 kent , s. , 1983 , mnras , 266 , 562 kunze , s. , speith , r. , hessman , f. , 2001 , mnras , 322 , 499 livio , m. , soker , n. , dgani , r. , 1986 , apj , 305 , 267 long , k. , gilliland , r. , 1999 , apj , 511 , 916 lubow , s. , 1989 , apj , 340 , 1064 marsh , t. , 1988 , mnras , 231 , 1117 marsh , t. , 1989 , pasp , 101 , 1032 marsh , t. , horne , k. , 1988 , mnras , 235 , 269 marsh , t. , horne , k. , schlegel , e. , honeycutt , r. , kaitchuck , r. , 1990 , apj , 364 , 637 matsuda , t. , makita , m. , fujiwara , h. , nagae , t. , haraguchi , k. , hayashi , e. , boffin , h. , 2000 , ap&ss , 274 , 259 morales - rueda , l. et al . , 2000 ,
mnras , 313 , 454 naylor , t. , allan , a. , long , k. , 2005 , mnras , 361 , 1091 neustroev , v.v , borisov , n.v . , 1998 , a&a , 336 , l73 ogilvie , g. , 2002 , mnras , 330 , 937 osaki , y. , 1996 , pasp , 108 , 39 paczynski , b. , 1977 , apj , 216 , 822 schneider , d. , young , p. , 1980 ,
apj , 238 , 946 shafter , a. , szkody , p. , thorstensen , j. , 1986 , mnras , 308 , 765 shanno , d. , phua , k. , 1980 , acm tran . on mathematical software , 6 , 618 smak ,
j , 2001 , acta astron . , 51 , 279 smak , j , 2001 , acta astron . , 51 , 295 smith , m. et al . , 1998 , apj , 507 , 94 smith , m. , dhillon , v. , marsh , t. , 1998 , mnras , 296 , 465 steeghs , d. , harlaftis , e. , horne , k. , 1998 , mnras , 296 , 463 steeghs , d. , horne , k. , marsh , t. , donati , j. , 1996 , mnras , 281 , 626 steeghs , d. , stehle , r. , 1999 , mnras , 307 , 99 stover , r , 1981 , apj , 249 , 673 wade , r. et al . , 1981 , apj , 246 , 215 watson , c. , dhillon , v. , rutten , r. , shcwope , a. , 2003 , mnras , 341 , 129 welsh , r. , horne , k. , gomer , r. , 1995 , mnras , 275 , 649 zhang , e. , robinson , e. , 1987 , apj , 321 , 813
in a doppler tomogram is usually possible to find at least two closed contours at a particular flux level .
this happens because the tomogram is usually dominated by the disc continuum , in such a way that the flux increases from the centre of the map towards a ring of maximum intensity , the flux then decreasing from this ring outwards .
this scenario makes the traditional methods to fit ellipses from the field of galactic surface photometry unstable ( e.g. kent 1983 , jedrzejewski 1987 ) , because in them most levels occur only once . on the other hand ,
the techniques employed in computer vision , for the characterization of conics in an image , are very suitable for application here .
we devised a new strategy by dividing the problem in two stages : tracing the contours at a certain level and extracting the points pertaining to a single closed contour at the same level are two completely different problems .
the former is readily solved in a purely graphical manner ( e.g. burke 1987 ) , while the latter ( ours ) is non - trivial and , as far as we know , not certainly solvable without human assistance .
consequently the first stage admits several approaches varying in efficiency .
we devised one in which from many angular directions either zero or one pixel was chosen , according to its proximity to the pixel selected in the previous direction and to the general pattern of the contours ( e.g. no crossed contours were allowed ) .
the procedure could be carried out almost on its own in the case of a smooth continuum , but it required close supervision when dealing with finely detailed ( or noisy ) structure .
the main danger is for the algorithm not to close the contour or to take a wrong detour before coming back to the expected path .
once all the points for a particular level were chosen , we fitted them with ellipses according to an analytical ellipse - specific method proposed by fitzgibbon , pilu & fischer ( 1999 ) .
the task is accomplished via a very efficient matrix procedure . in a non - matrix form
we can explain the procedure as the fitting of a general conic of the form : @xmath79 but adding the constraint : @xmath80 so that the result will specifically be an ellipse .
the constraint adds an arbitrary factor of scaling to the solution , but the algorithm provides for the removal of the effect at a later stage .
the resulting general coefficients @xmath81 , @xmath82 , @xmath83 , @xmath84 , @xmath85 and @xmath86 were converted to the more meaningful values @xmath87,@xmath88 ( coordinates of the centre of the ellipse ) , @xmath89 , @xmath90 ( size of the two semiaxes ) , @xmath91 ( eccentricity ) and @xmath92 ( angle of rotation ) .
uncertainties were estimated by a bootstrap procedure over the points . the solution given by the method of fitzgibbon et al .
( 1999 ) has been proved to be unique , but it must be noted that it has a certain bias to low eccentricities , specially if only a few points are fed to the algorithm . this is fine for our purposes because we use many points for the fitting ( basically limited by the resolution of the map ) and also we do not want unrealistically high values for @xmath91 ( which for an accretion disc we expect to be close to 0 for most isophotes ) .
if this was a problem , the bias could be removed by another bootstrap procedure at the cost of computing time ( see , for instance , cabrera & meer 1996 ) . by subtracting the disk from the original doppler map we can greatly improve the visibility of asymmetrical features .
although we expect disc isophotes to have low values of @xmath91 , this value is generally not be exactly 0 .
we therefore used the results from the ellipse - fitting process described above to subtract an interpolated elliptical background . to do this
, we found for every point the two closest ellipses and then we linearly interpolated their fluxes according to the distance of the point to both ellipses .
special cases were : a point exactly in the path of an ellipse ( its flux was used ) , and points beyond the region fitted ( no flux was assigned ) . | the dwarf nova u gem is unique in having a direct measurement of the @xmath0-velocity of its white dwarf from _ hst _ spectra ( @xmath1 , long et al . 1999 ) .
we present high - resolution optical spectra of the dwarf nova u gem in quiescence taken to test the accuracy to which the _ hst _ value can be recovered from optical data .
we find that , even with data of very high signal - to - noise ratio on this archetypal system , we can not recover long et al.s value to better than about @xmath2% by any method .
contamination by neighbouring emission lines seems a likely culprit .
our data reveal a number of new features : doppler tomograms show emission at low velocity , close to the centre of mass , and a transient , narrow absorption feature is seen in the balmer lines near the line centres at the time of eclipse .
we suggest that stellar prominences , as previously invoked for the dwarf novae ip peg and ss cyg in outburst , may explain both of these features .
the 4686.75 line emission is dominated by the gas stream / disc impact region .
two distinct spots are seen in doppler maps , the first being very narrow and showing a velocity close to that of the accretion disc in the impact region , and the second much broader and located between the velocities of the ( ballistic ) stream and the ( keplerian ) disc .
we present tentative evidence of weak spiral structure , which may support explanations for `` spiral shocks '' based upon 3-body effects .
we find no evidence of stream - disc overflow in the system .
our data suggests an inclination angle @xmath3 , adding to the evidence supporting the existence of a puzzle in the mass of u gem s white dwarf .
the mass donor is clearly seen in the doppler maps , with emission concentrated towards its poles , and mainly on the side facing the white dwarf .
this suggests irradiation with shielding by the disc from which we estimate an @xmath4 ratio between 0.15 and 0.25 .
[ firstpage ] binaries : close binaries : eclipsing binaries : spectroscopic stars : individual : u gem |
in this paper we derived an analytic solution for the flavor transition probabilities of solar neutrinos in a general scheme with an arbitrary number of sterile neutrinos , without any constraint on the mixing .
hence , we improved the study presented in ref .
@xcite , where it was assumed that the electron neutrino has non - negligible mixing only with the two massive neutrinos which generate the solar squared - mass difference .
here we only assumed a realistic hierarchy of neutrino squared - mass differences in which the smallest squared - mass difference is effective in solar neutrino oscillations .
in section [ 091 ] we have illustrated the validity of the analytical approximation ( [ 068 ] ) of the flavor transition probability by comparing its value with that obtained with a numerical solution of the evolution equation in three examples of the possible mixing matrix in the simplest case of four - neutrino mixing .
hence , we are confident that the expression ( [ 068 ] ) of the flavor transition probability can be used in a reliable analysis of the solar neutrino data which could provide important information on active - sterile neutrino mixing @xcite .
we would like to thank the department of theoretical physics of the university of torino for hospitality and support .
in this paper we used four different bases of the neutrino states and amplitudes : 1 .
the flavor basis of amplitudes in eq .
( [ 017 ] ) , related to the flavor states on the left - hand side of eq .
( [ 005a ] ) .
the vacuum mass basis of amplitudes in eq .
( [ 027 ] ) , related to the vacuum mass basis of states in the right - hand side of eq .
( [ 005a ] ) .
3 . the effective mass basis in matter of amplitudes in the truncated 1 - 2 sector , @xmath28 and @xmath29 , introduced in eq . ( [ 042 ] ) .
4 . the effective interaction basis in matter of amplitudes in the truncated 1 - 2 sector , @xmath30 and @xmath31 , introduced in eq . ( [ 044 ] ) . since @xmath32 we have @xmath33 therefore , the effective interaction amplitudes @xmath30 and @xmath31 for which the matter potential in the truncated 1 - 2 sector is diagonal are linear combinations of the flavor amplitudes @xmath34 , @xmath35 , @xmath36 , @xmath37 , @xmath38 , @xmath39 .
furthermore , from eqs .
( [ 042 ] ) and ( [ 046 ] ) we derive immediately the relation between the amplitudes @xmath28 and @xmath29 of the effective energy eigenstates in matter and the flavor amplitudes @xmath40 : @xmath41 in the effective mass and interaction bases in matter the solar neutrino state in eq .
( [ 014 ] ) is given by @xmath42 with the effective massive states in matter @xmath43 and the effective interaction states @xmath44 thus , a flavor neutrino states can be expressed as @xmath45 for @xmath46 .
one can see that in general each flavor state contains both the effective massive states @xmath47 and @xmath48 as well as both the effective interaction states @xmath49 and @xmath50 . | the matter effects for solar neutrino oscillations are studied in a general scheme with an arbitrary number of sterile neutrinos , without any constraint on the mixing , assuming only a realistic hierarchy of neutrino squared - mass differences in which the smallest squared - mass difference is effective in solar neutrino oscillations .
the validity of the analytic results are illustrated with a numerical solution of the evolution equation in three examples of the possible mixing matrix in the simplest case of four - neutrino mixing .
r * matter effects in active - sterile solar neutrino oscillations + c. giunti@xmath0 , y.f .
li@xmath1 + _ _ * [ cols="^ , < " , ] we solved numerically the effective two - neutrino evolution equation ( [ 041 ] ) for the exponential electron number density profile @xmath2 which is a good approximation of the real electron number density in the solar radiative and convective zones ( see ref .
@xcite ) .
we considered two values of the electron fraction @xmath3 : the realistic value @xmath4 , which is a good approximation in the solar radiative and convective zones ( see fig . [ fig01 ] ) , and the unrealistic value @xmath5 , which is useful to test the effects of the neutral - current potential and the case @xmath6 .
the panels in figs .
[ fig02 ] and [ fig03 ] show , respectively , the averaged @xmath7 survival probability and the averaged probability of @xmath8 transitions as functions of the neutrino energy @xmath9 for the three examples of mixing matrices m1 , m2 , and m3 in tab . [ tab01 ] .
notice that in order to check beyond any doubt the validity of the analytic flavor transition probability ( [ 068 ] ) both in the adiabatic and non - adiabatic regimes , we considered a very wide range of energies , from @xmath10 to @xmath11 , which is much wider than the energy range of measurable solar neutrinos , from about @xmath12 to about @xmath13 .
moreover , the inequality ( [ 026 ] ) , which allows the approximation in eq .
( [ 029 ] ) , is not valid for energies not much smaller than about 100 mev , for which the value of @xmath14 in the center of the sun is of the same order as @xmath15 .
hence , we emphasize that the energy values larger than those of solar neutrinos in figs .
[ fig02 ] and [ fig03 ] are considered only for the purpose of checking the validity of the analytic solution of the evolution equation ( [ 041 ] ) . from figs .
[ fig02 ] and [ fig03 ] one can see that the values of the flavor transition probability obtained with the analytic expression ( [ 068 ] ) and the numerical solution of the evolution equation ( [ 041 ] ) agree perfectly in all the three examples considered .
also , the extreme non - adiabatic limit discussed in section [ non - adiabatic ] is perfectly described by the analytic approximation .
the figures illustrate also some different peculiar behavior of the averaged oscillation probabilities in the three examples . with the realistic electron fraction @xmath4 , @xmath16 has the same qualitative behavior in the three examples , whereas @xmath17 is enhanced by the matter effect in m1 and m3 and suppressed in m2 .
this is due to the different signs of @xmath18 ( see tab .
[ tab01 ] ) . in m1 and m3
@xmath18 is negative and the adiabatic part of the transition probability @xmath19 is positive , leading to an enhancement of @xmath17 . in m2 the opposite happens . in the case of the unrealistic electron fraction @xmath5
, one can see from figs .
[ fig02 ] and [ fig03 ] that @xmath20 because @xmath6 , as discussed in section [ non - adiabatic ] .
the three examples have the following interesting differences in the extreme non - adiabatic limit . from fig . [ fig02 ]
one can see that @xmath21 in m1 and m2 and the inequality is reversed in m3 .
this happens because @xmath22 in m1 and m2 , whereas @xmath23 in m3 . from fig .
[ fig03 ] one can see that @xmath24 in m1 and the inequality is reversed in m2 and m3 .
these behaviors are due to the fact that @xmath22 and @xmath25 in m1 , @xmath22 and @xmath26 in m2 , and @xmath23 and @xmath25 in m3 .
after the check of the analytic expression ( [ 068 ] ) for the flavor transition probability in the case of the perfect exponential density ( [ 155 ] ) and a constant electron number fraction , we also checked its validity for the realistic bp04 standard solar model density @xcite , taking into account the variation of the electron number fraction ( plotted in fig .
[ fig01 ] ) .
the results of the analytic expression ( [ 068 ] ) and of the numerical solution of the evolution equation ( [ 041 ] ) are presented in fig .
[ fig04 ] for the three examples of mixing matrices m1 , m2 , and m3 .
we considered only neutrino energies smaller than about 50 mev , for which the inequality ( [ 026 ] ) and the approximation ( [ 029 ] ) are valid . in this range of energies
the flavor transition is practically adiabatic and we expect that non - adiabatic effects due to the variation of the electron number fraction are negligible . indeed , from fig .
[ fig04 ] one can see that the analytic approximation is very accurate for solar neutrinos and the effects of the variation of the electron number fraction , which are neglected in the analytic approximation , are really negligible . from the left panel in fig .
[ fig04 ] one can see that active - sterile transitions as those in the examples m1 , m2 , and m3 can change significantly the @xmath7 survival probability with respect to the standard three - neutrino mixing value in the case of negligible @xmath27 ( dash - dotted line ) .
therefore , it is important to perform an analysis of solar neutrino data @xcite in the framework of active - sterile neutrino mixing that we have considered in order to obtain information on the possible existence of sterile neutrinos and their mixing with the active ones . |
since the pioneering discovery of superconductivity in cecu@xmath5si@xmath5 , @xcite elucidation of the mechanism of unconventional superconductivity in heavy fermion compounds has been one of the central issues both in the experimental and theoretical research fields of condensed matter physics . in particular
, further discoveries of superconductivity in uranium compounds such as ube@xmath6,@xcite uru@xmath5si@xmath5,@xcite upt@xmath7,@xcite upd@xmath5al@xmath7,@xcite and uni@xmath5al@xmath7@xcite have triggered a rapid increase of vigorous investigations on exotic properties of superconductivity in @xmath0-electron materials .
recently there has occurred a second rush of new discoveries of superconductivity in @xmath0-electron compounds .
one such is the coexistence of superconductivity and ferromagnetism in uge@xmath5@xcite and urhge.@xcite after this discovery , much attention has been attracted to the mechanism of unconventional triplet superconductivity in @xmath0-electron systems , since it is naively believed that triplet pairing can coexist in the ferromagnetic phase .
a second example is the family of ce - based heavy fermion superconductors cemin@xmath4 ( m = ir , rh , and co).@xcite surprisingly , cecoin@xmath4 exhibits a superconducting transition temperature @xmath8=2.3k , which is the highest among those yet observed for heavy fermion materials at ambient pressure . on the other hand
, ceirin@xmath4 shows @xmath8=0.4k , which is much less than that of cecoin@xmath4 .
note that cerhin@xmath4 is an antiferromagnet with a nel temperature @xmath9=3.8k at ambient pressure , while under high - pressure it becomes superconducting at @xmath8=2.1k .
it is interesting that the ground state properties of these compounds are easily changed by transition metal ions .
third is the superconductivity of the filled skutterudite compound pros@xmath10sb@xmath11 with @xmath8=1.85k and a mass enhancement factor as large as fifty.@xcite again , a relatively high @xmath8 is surprising , while an interesting point is that the ground state is suggested to belong to @xmath12 , leading to a potential role played by quadrupole fluctuations in the mechanism of superconductivity .
another amazing result is the discovery of superconductivity in pucoga@xmath4 with @xmath8=18.5k.@xcite although it is difficult to synthesize this compound and to measure various physical quantities , owing to the inclusion of plutonium , this is truly high @xmath8 and its mechanism is rather mysterious . among several remarkable features both of new and old @xmath0-electron superconductors ,
one interesting and important point is the common observation of anisotropic cooper - pairing , experimentally suggested by the power - law behavior of physical quantities in the low - temperature region .
as is well known , the value of the power sensitively depends on the node structure of the gap function on the fermi surface .
thus , by analysing carefully the temperature dependence of physical quantities in experiments , it is possible to deduce the symmetry of the cooper pairs under the group - theoretical restrictions . as a result of this phenomenological analysis
, it has been clarified that the superconducting gap in @xmath0-electron materials has nodes on the fermi surface , indicating that non @xmath13-wave cooper pairing occurs in heavy fermion superconductors .
the appearance of non @xmath13-wave pairing itself is easily understood as an effect of strong short - range correlation , if we recall the fact that the cooper pair is composed of a pair of @xmath0 electrons with heavy effective mass . since
superconductivity in heavy fermion materials is a second - order transition phenomenon , the ginzburg - landau theory is still applicable to @xmath0-electron superconductors if we take into account the symmetry of the cooper pairs as well as the crystal structure with the help of group - theoretical arguments .
such efforts have been quite useful to promote the understanding of the nature of unconventional superconductivity without considering the mechanism of the cooper - pair formation.@xcite as for the microscopic aspects of the mechanism of unconventional superconductivity , it has at least been confirmed that the strong short - range coulomb interaction plays a crucial role .
for instance , it brings about antiferromagnetic spin fluctuations , @xcite which induce singlet @xmath14-wave superconductivity , @xcite consistent with the node structure suggested by some experiments .
it is true that in some @xmath0-electron materials , anisotropic singlet pairing is stabilized , but in upt@xmath7 , it has been experimentally confirmed that triplet pairing occurs.@xcite quite recently , also in uni@xmath5al@xmath7 , a possibility of triplet pairing has been discussed.@xcite unfortunately , the triplet pair formation as well as the coexistence of superconductivity and ferromagnetism can not be simply explained as the effect of antiferromagnetic spin fluctuations .
one may naively consider that instead , we can include the effect of ferromagnetic spin fluctuations , which was proposed for the origin of superfluidity in helium 3 .
however , we then immediately face another problem , i.e. , how to obtain such fluctuations from the strong short - range coulomb interaction
. it may be possible to derive this based on a realistic electronic band structure , but we do not find that paramagnons are always dominant in the spin fluctuation spectrum when we survey the magnetic properties of those materials .
in fact , in ruthenate sr@xmath5ruo@xmath10,@xcite which is experimentally confirmed to be a triplet superconductor,@xcite instead we find that the incommensurate antiferromagnetic spin fluctuation is enhanced , as observed in neutron scattering experiments.@xcite namely , in contrast to our naive expectation , paramagnons do not seem to play a central role in the occurrence of triplet superconductivity in the solid state . thus , it is still a puzzling and challenging problem to clarify what is the key issue which determines the symmetry of cooper pairs , singlet or triplet , in @xmath0-electron superconductors . in order to understand this point ,
it is necessary to perform a microscopic analysis based on a model hamiltonian appropriate for @xmath0-electron systems , but unfortunately , we find such a model to be lacking . it may be possible to consider this problem based on the single - band hubbard model , which is believed to describe well the properties of strongly correlated electron systems . in fact , several basic features of high @xmath8 cuprates have been understood based on the hubbard model by analyzing it using various kinds of analytical and numerical techniques.@xcite however , this model is too simplified to include differences among materials characterized by @xmath15 degrees of freedom , although recently , the active role of orbitals has been widely recognized in several materials as a means for understanding novel magnetism and unconventional superconductivity .
for instance , in manganites , @xmath16 orbitals play a primary role for electronic properties,@xcite since orbital ordering has been successfully observed in manganites by means of the resonant x - ray scattering measurement .
@xcite in order to investigate systems with active orbital degrees of freedom , it is necessary to analyze the multi - orbital hubbard model .
such an extension of the model does not bring essential difficulty , at least in @xmath14-electron materials .
for instance , in @xmath17 electron systems such as ruthenate sr@xmath5ruo@xmath10 introduced above , it has been found that a three @xmath18-orbital tight - binding model can well reproduce the experimental results , as can band - structure calculations.@xcite however , if we try to construct such a model also for @xmath0-electron materials , we immediately find a serious problem .
since the spin - orbit interaction is fairly strong in @xmath0-electron systems , the meaning of a single @xmath0-electron state is unclear .
even if this point is overcome , we find another difficulty in the evaluation of hopping amplitudes as well as coulomb interactions based on a good representation for one @xmath0-electron state .
in addition , we need to include correctly the important effects of crystalline electric fields ( cef ) , which are very often essential for some @xmath0-electron systems . in this paper , we attempt to go beyond all those problems through use of the @xmath1-@xmath1 coupling scheme , in order to construct an appropriate microscopic hamiltonian for @xmath0-electron compounds , including the hopping amplitudes , coulomb interactions , and cef terms .
the hamiltonian obtained is a multi - orbital hubbard - like model .
after construction of the model , we exhibit a couple of practical models , simplified by discarding one orbital degree of freedom .
then , as an typical example , one simplified model is analyzed by using exact diagonalization techniques to evaluate several kinds of correlation functions .
especially , the superconducting pair correlation function is carefully analyzed based on the concept of off - diagonal long - range order .
finally , we will discuss the relation between the present theoretical results and some properties of cemin@xmath4 .
we also try to give a possible explanation for superconducting and antiferromagnetic tendencies in `` 115 '' materials including cemin@xmath4 , umga@xmath4 , and pucoga@xmath4 , based on our microscopic picture .
the organization of this paper is as follows . in sec . 2 ,
the construction of the microscopic model hamiltonian based on the @xmath1-@xmath1 coupling scheme is discussed in detail .
the coulomb interaction terms , cef terms , and the hopping amplitudes are evaluated by paying due attention to @xmath0-orbital symmetry .
two types of simplified models are also introduced . in sec . 3 ,
one of the simplified models will be analyzed by showing the results of spin , charge , orbital , and pair correlation functions .
a prescription to measure the pair correlation will be also discussed . finally in sec . 4 , the paper is summarized and properties of 115-materials such as cemin@xmath4 , umga@xmath4 , and pucoga@xmath4 are discussed . throughout the paper ,
we use units such that @xmath19=@xmath20=1 .
in this section we construct a microscopic model hamiltonian for @xmath0-electron systems in order to discuss unconventional superconductivity and novel magnetism from the microscopic viewpoint . for this purpose
, the model must include correctly the itinerant nature of @xmath0 electrons as well as strong electron correlation and the effect of cef .
namely , the hamiltonian @xmath21 should be composed of three parts as @xmath22 where @xmath23 , @xmath24 , and @xmath25 are @xmath0-electron hopping , coulomb interactions among @xmath0 electrons , and the cef term , respectively . in order to express the hamiltonian in the form of eq .
( [ eq : h ] ) , it is requested that the one @xmath0-electron state should be well - defined by using a label @xmath2 .
the physical meaning of @xmath2 will be discussed later in detail , but here it is convenient to define a second quantized operator to express the state @xmath26 as @xmath27 where @xmath28 denotes the vacuum and @xmath29 is the creation operator for @xmath0 electrons in the @xmath2-state at site @xmath30 . by using the second - quantized operator @xmath29 defined above
, the kinetic term can be written as @xmath31 where @xmath32 is the overlap integral between the @xmath2- and @xmath33-states connected by the vector @xmath34 .
the coulomb interaction term is given in the form of @xmath35 where @xmath36 is the matrix element for coulomb interactions . finally , the cef term is expressed by @xmath37 where @xmath38 indicates the potential energy for @xmath0 electrons which expresses the effect of the cef .
now our tasks in this section are as follows : ( i ) specification of the meaning of @xmath2 for one @xmath0-electron state .
( ii ) evaluation of the hopping amplitude , coulomb matrix elements , and cef coefficients .
( iii ) simplification of the model for practical purposes . in the following subsections ,
all of these points will be clarified in this order .
first let us discuss the meaning of @xmath2 in the hamiltonian eq .
( [ eq : h ] ) . to make it clear ,
it is instructive to start the discussion with the single ion problem using two typical approaches , the @xmath40 coupling and @xmath1-@xmath1 coupling schemes . in general
, we consider the @xmath41 configuration , where @xmath42 is the number of @xmath0 electrons included on a localized ion . in the @xmath40 coupling scheme ,
first the spin @xmath43 and angular momentum @xmath44 are formed by following hund s rules as @xmath43=@xmath45 and @xmath44=@xmath46 , where @xmath47 and @xmath48 are spin and angular momentum for @xmath49-th @xmath0-electron .
note here that the hund s rules are based on the pauli principle and coulomb interactions among @xmath0 electrons . after forming @xmath43 and @xmath44
, we include the effect of spin - orbit interaction @xmath50@xmath44@xmath51@xmath43 , where @xmath50 is the spin - orbit coupling .
note that @xmath50@xmath520 for @xmath42@xmath537 , while @xmath50@xmath530 for @xmath42@xmath527 .
note also that a good quantum number to label such a state is the total angular momentum @xmath54 , given by @xmath54=@xmath44+@xmath43 . following from simple algebra ,
the ground - state level is characterized by @xmath55=@xmath56@xmath57@xmath58 for @xmath42@xmath537 , while @xmath55=@xmath59+@xmath60 for @xmath42@xmath527 . as is easily understood from the above discussion
, the @xmath40 coupling scheme is valid under the assumption that the hund s rule coupling is much larger than the spin - orbit interaction , since @xmath43 and @xmath44 are formed by the hund s rule coupling prior to the inclusion of spin - orbit interaction .
it is considered that this assumption is valid for insulating compounds with localized @xmath0 electrons . however , when the spin - orbit interaction is not small compared with the hund s rule coupling , especially in actinide ions , the above assumption is not always satisfied .
in addition , if the @xmath0 electrons become itinerant owing to hybridization with the conduction electrons , the effect of coulomb interactions would thereby be effectively reduced . in rough estimation ,
the effective size of the coulomb interaction may be as large as the bandwidth of @xmath0 electrons , leading to a violation of the assumption required for the @xmath40 coupling scheme . for @xmath0-electron systems in which the spin - orbit interaction becomes larger than the effective coulomb interactions , it is useful to consider the problem in the @xmath1-@xmath1 coupling scheme .
here we emphasize that the @xmath1-@xmath1 coupling scheme is also quite convenient for including many - body effects using the standard quantum - field theoretical techniques , since individual @xmath0-electron states can be clearly defined , as explained below .
first , we include the spin - orbit coupling so as to define the state labelled by the total angular momentum @xmath61 for the @xmath49-th electron , given by @xmath61=@xmath47+@xmath48 . for @xmath0-orbitals with @xmath62=3
, we immediately obtain an octet with @xmath1=7/2(=3 + 1/2 ) and a sextet with @xmath1=5/2(=3@xmath571/2 ) , which are well separated by the spin - orbit interaction .
note here that the level for the octet is higher than that of the sextet .
then , we take into account the effect of coulomb interactions to accommodate @xmath42 electrons among the sextet and/or octet , leading to the ground - state level in the @xmath1-@xmath1 coupling scheme . here , the meaning of @xmath2 in our hamiltonian becomes clear . in the case of @xmath42@xmath537
, @xmath2 should be the label to specify the state in the @xmath1=5/2 sextet , namely , the @xmath3-component of the total angular momentum @xmath1=5/2 and takes the values of @xmath2=@xmath63 , @xmath64 , @xmath65 , @xmath66 .
note here that for 3@xmath53@xmath42@xmath537 , @xmath1=7/2 octet is @xmath67 occupied , since we presume that the effect of spin - orbit interaction is larger than that of the hund s rule coupling in the @xmath1-@xmath1 coupling scheme . on the other hand , for the case of @xmath42@xmath687 ,
@xmath2 should be considered to specify the state in the @xmath1=7/2 octet , since @xmath1=5/2 sextet is fully occupied .
note again that spin - orbit interaction is larger than that of the hund s rule coupling . in this paper , we concentrate only on the case of @xmath42@xmath537 , especially the situation with @xmath42=1 and 2 , corresponding to ce@xmath69 and u@xmath70 ions .
thus , in the following , @xmath2 indicates the @xmath3-component of the total angular momentum which specifies the state in the @xmath1=5/2 sextet .
now let us consider the ground - state multiplet appearing in the @xmath1-@xmath1 coupling scheme .
for this purpose , first it is necessary to define the coulomb interaction @xmath36 in @xmath24 , which is the sum of two contributions , written as @xmath71 where the former indicates the coulomb term , while the latter denotes the exchange one .
it should be noted that @xmath36 vanishes unless @xmath2+@xmath33=@xmath72+@xmath73 due to the conservation of @xmath3-component of total angular momentum .
the matrix element @xmath74 is explicitly given by @xmath75 where @xmath76 is the clebsch - gordan coefficient , given as @xmath77 with @xmath78=@xmath79 ( @xmath80 ) for up ( down ) real spin .
the coulomb matrix element among @xmath62-orbitals is given by @xmath81 where @xmath82=@xmath83 and @xmath84 is the wavefunction for the @xmath85 state in @xmath62=3 orbitals , expressed by the product of the radial function and spherical harmonic @xmath86 .
the matrix element @xmath87 can be written in the form of @xmath88 where the sum on @xmath89 includes only even values ( @xmath89=0 , 2 , 4 , and 6 ) , @xmath90 is the slater - condon parameter@xcite including the complex integral of the radial function , and @xmath91 is the gaunt coefficient,@xcite which is tabulated in the standard textbooks of quantum mechanics ( see , for instance , ref . ) . when two electrons are accommodated in the @xmath1=5/2 sextet , the allowed values for total angular momentum @xmath55 are 0 , 2 , and 4 due to the pauli principle .
thus , the coulomb interaction term should be written in a 15@xmath9215 matrix form .
note that `` 15 '' is the sum of the basis numbers for singlet ( @xmath55=0 ) , quintet ( @xmath55=2 ) , and nontet ( @xmath55=4 ) .
as is easily understood , this 15@xmath9215 matrix can be decomposed into a block - diagonalized form labelled by @xmath93 , including one 3@xmath923 matrix for @xmath93=0 , four 2@xmath922 matrices for @xmath93=@xmath942 and @xmath941 , and four 1@xmath921 for @xmath93=@xmath95 and @xmath96 .
we skip the details of tedious calculations for the matrix elements and here only summarize the results in the following by using the racah parameters @xmath97 ( @xmath89=0,1,2),@xcite which are related to the slater - condon parameters @xmath90 as @xmath98 for @xmath93=4 and 3 , we obtain @xmath99 and @xmath100 respectively . for @xmath93=2 and 1 , we obtain @xmath101 and @xmath102 finally , for @xmath93=0 , we obtain @xmath103 note here the following relations : @xmath104 and @xmath105 by using these two relations and eqs . ( [ eq : jz4]-[eq : jz0 ] ) , we can obtain all the coulomb matrix elements.@xcite as a typical example , let us consider the @xmath106 configuration . in the @xmath1-@xmath1 coupling schemes , two
electrons are accommodated in the @xmath1=5/2 sextet .
when we diagonalize the 15@xmath9215 matrix for coulomb interaction terms , we can easily obtain the eigen energies as @xmath107@xmath57@xmath108 for the @xmath55=4 nontet , @xmath107+@xmath109 for the @xmath55=2 quintet , and @xmath107+@xmath110 for the @xmath55=0 singlet.@xcite since the racah parameters are all positive , the ground state is specified by @xmath55=4 in the @xmath1-@xmath1 coupling scheme . in the @xmath40 coupling scheme ,
on the other hand , we obtain the ground - state level as @xmath111 with @xmath60=1 and @xmath59=5 from the hund s rules . on further inclusion of the spin - orbit interaction ,
the ground state becomes characterized by @xmath55=4 , expressed as @xmath112 in the traditional notation .
note that we are now considering a two - electron problem .
thus , if we correctly include the effects of coulomb interactions , it is concluded that the same quantum number as that in the @xmath40 coupling scheme is obtained in the @xmath1-@xmath1 coupling scheme for the ground - state multiplet . in order to understand the physical meaning of racah parameters ,
let us consider a simplified coulomb interaction term . in the above discussion
, the expressions using racah parameters are not convenient since they depend on the orbitals in a very complicated manner , although they keep the correct symmetry required by group theory . to clarify their meanings
, it may be useful to restart the discussion from the following interaction form among @xmath113 orbitals : @xmath114 where @xmath115 is the annihilation operator for an @xmath0 electron with spin @xmath78 in the @xmath85 state of angular momentum @xmath62 at site * i * and @xmath116= @xmath117 . in this equation , we include only three interactions ; intra - orbital coulomb interaction @xmath118 , inter - orbital coulomb interaction @xmath119 , and the exchange interaction @xmath55 .
note that the relation @xmath118=@xmath119+@xmath55 holds among coulomb interactions to ensure rotational invariance in the orbital space . by using clebsch - gordan coefficients , @xmath115 with real - spin @xmath78 can be related to @xmath120 as @xmath121 note here that we consider only the @xmath122 sextet . the coulomb interaction term for @xmath1=5/2 is given by @xmath123 where @xmath124=@xmath125 , @xmath126=@xmath127 , @xmath128=@xmath119@xmath57@xmath129 , @xmath130=@xmath131 , and @xmath132 is the operator for total angular momentum with @xmath55=5/2 .
explicitly , @xmath133 is written as @xmath134,\end{aligned}\ ] ] with @xmath135 = @xmath136 . for two electrons in the @xmath1=5/2 sextet , based upon the simplified coulomb interaction term
, we can easily obtain the energy levels as @xmath128@xmath57@xmath137 for the @xmath55=4 nontet , @xmath128+@xmath138 for the @xmath55=2 quintet , and @xmath128+@xmath139 for the @xmath55=0 singlet .
when we compare these energy levels with the results obtained using racah parameters , we understand the correspondence such as @xmath107@xmath140@xmath128 and @xmath141@xmath140@xmath130 .
namely , @xmath107 is the effective inter - orbital coulomb interaction , while @xmath141 denotes the hund s rule coupling .
note that @xmath142 does not appear , since it is related to the pair - hopping interaction which is not included here . finally , we note the smallness of @xmath130 , given as @xmath130=@xmath131 .
the origin of the large reduction factor 1/49 is , in one word , due to the neglect of @xmath1=7/2 octet . in the coulomb interaction term eq .
( [ eq : hint2 ] ) , the hund s rule term is simply written as @xmath143 .
note the relation @xmath43=@xmath144 with @xmath145 the land s @xmath146-factor . for @xmath1=5/2
, we easily obtain @xmath145=6/7 , indicating @xmath43=@xmath147 .
thus , the original hund s rule term is simply rewritten as @xmath148 .
as is well known , the ground - state multiplet is further split due to the effect of the cef . in this paper
we take into account the effect of the cef by simply accommodating several electrons in the one - electron potential .
namely , @xmath38 in eq . (
[ h : cef ] ) is expressed by the cef parameters for @xmath55=5/2 .
since the cef term depends on the crystal structure , it is convenient to write down the cef formula for @xmath55=5/2 in the typical lattice structure , expressed as @xmath149 for a cubic lattice , @xmath150 for a tetragonal lattice , and @xmath151 for a hexagonal lattice . in each equation
above , @xmath152 is the stevens equivalent operator,@xcite composed of @xmath93 , @xmath153 , and @xmath154 .
the matrix formulae have been tabulated in a paper by hutchings.@xcite the coefficients @xmath155 are conventionally called the cef parameters , which are , in actuality , determined by the fitting of experimental results for physical quantities such as magnetic susceptibility .
note here that the above formulae have been obtained from the case of @xmath55=5/2 .
in general , the cef term is expressed in matrix form , depending on the value of @xmath55 ; for @xmath55 larger than 5/2 , higher terms in @xmath155 should occur . however , as already mentioned above , since in this paper the effect of the cef is considered as a one - electron potential based on the @xmath1-@xmath1 coupling scheme , it is enough to use the cef term for @xmath55=5/2 . in order to check the validity of our viewpoint regarding the cef term ,
let us see , for instance , the case of cubic structure , in which we can easily obtain @xmath156 and zero for other @xmath2 and @xmath33 . in the following we consider the cases of both @xmath157 and @xmath106 configurations .
when one electron exists in @xmath0-orbitals , i.e. , in the case of the ce@xmath69 ion , the ground state is easily obtained by the diagonalization of eq .
( [ eq : cubic - cef ] )
. then we immediately obtain two eigen energies , @xmath158=@xmath159 for the @xmath160 doublet and @xmath161=@xmath162 for the @xmath163 quartet .
thus , the ground state is changed due to the sign of @xmath164 , as shown in fig .
qualitatively , the ground state is deduced from the positions of ligand anions and the shapes of wavefunctions . as in the case of the ce monopnictide , when anions are located along the direction of [ 1,0,0 ] from the ce@xmath69 ion , the @xmath163 level is energetically penalized , since the corresponding wavefunctions are expanding along the axis directions . on the other hand , as in the case of ce hexaborides with anions located along the [ 1,1,1 ] directions from the ce@xmath69 ions , the @xmath160 levels are higher due to the shape of @xmath160 wavefunction .
= 1 under the cubic cef .
( b ) energy levels as a function of @xmath164 for @xmath42=2 in the @xmath1-@xmath1 coupling scheme .
( c ) energy levels as a function of @xmath165(=@xmath166 ) for @xmath55=4 nontet under the cubic cef with @xmath167(=@xmath168)=0.1 .
( d ) energy levels as a function of @xmath165 for @xmath55=4 nontet with @xmath167=@xmath169 . ] in order to consider the @xmath106 configurations such as pr@xmath70 and u@xmath70 ions based on the @xmath1-@xmath1 coupling scheme , we put two electrons in the potential for the @xmath55=5/2 cef term . here
some readers may have a naive question regarding its validity . to clarify this point
, we analyze the cef level schemes for the @xmath106 configuration by diagonalizing @xmath24+@xmath25 . in fig .
1(b ) , we show the eigen energies as a function of @xmath164 for a set of typical values of racah parameters , @xmath107=5 , @xmath142=2 , and @xmath141=0.5 . for @xmath164@xmath530 ,
the ground state is the @xmath170 triplet , while the @xmath171 singlet appears for @xmath164@xmath520 . at @xmath164=0 ,
the ground - state is the @xmath55=4 nontet . here
it should be noted that the ground - state wavefunction consists of a mixture of the @xmath55=4 , @xmath55=2 , and @xmath55=0 states .
note also that @xmath93 , not @xmath55 , is a good quantum number to label the state , since the rotational invariance in @xmath55 space is broken by the effect of the cef for @xmath55=5/2 .
however , we confirm that the contribution of the @xmath55=4 component is the largest in the ground state wave function . in the @xmath40 coupling scheme , on the other hand , we need to consider a 9@xmath929 matrix form for the @xmath55=4 nontet .
the cef term for @xmath55=4 under cubic symmetry is given by @xmath172= @xmath173+@xmath174 .
note that for the case of @xmath55=4 , there appears a higher order term including @xmath175 .
referring to the paper by hutchings,@xcite we obtain the cef coefficients as @xmath176 for the @xmath93=@xmath942 sector , @xmath177 for the @xmath93=@xmath943 and @xmath1781 sectors , and @xmath179 for the @xmath93=@xmath944 and 0 sectors .
note that we obtain three 2@xmath922 and one 3@xmath923 matrices .
after diagonalization , we obtain four eigen energies as @xmath180 as shown in fig .
1(c ) , for @xmath175@xmath1810 , the ground state is a @xmath170 triplet for @xmath164@xmath520 , while it is a @xmath171 singlet for @xmath164@xmath530 . on the other hand , for @xmath175@xmath530 ,
the ground state is also a @xmath170 triplet for @xmath164@xmath520 and a @xmath171 singlet for @xmath164@xmath530 , respectively , except for the region around @xmath164@xmath1400 , in which case the doublet @xmath12 becomes the ground state , as shown in fig .
when we compare the @xmath40 coupling results with those for the @xmath1-@xmath1 coupling , the following two points are noted : ( i ) although we commonly observe conversion of the ground state between a @xmath170 triplet and a @xmath171 singlet , the correspondence with the sign of @xmath164 is reversed between the @xmath40 and the @xmath1-@xmath1 coupling schemes .
( ii ) the @xmath12 doublet does @xmath67 appear as the ground state in the @xmath1-@xmath1 coupling scheme . the point ( i )
may not be serious , since it is just a problem of parametrization .
however , point ( ii ) looks strange at first glance . in order to clarify this point ,
it is useful to change the basis in the @xmath1-@xmath1 coupling . as shown in fig .
1(a ) , in the one @xmath0-electron case , for @xmath164@xmath520 ( @xmath530 ) , the @xmath160 doublet is the ground state ( excited state ) , while the @xmath163 quartet is the excited state ( ground state ) .
if we introduce pseudo - spin to distinguish the kramers doublet , @xmath160 indicates a single orbital , while @xmath163 denotes doubly degenerate orbitals . when we put two electrons into the situation shown in fig .
1(a ) , owing to the hund s rule coupling we obtain three cases :
( i ) two electrons occupy @xmath160 for @xmath164@xmath520 .
( ii ) they occupy @xmath163 for @xmath164@xmath530 .
( iii ) one occupies @xmath160 and the other does @xmath163 around @xmath164@xmath1820 . after algebraic calculations , we confirm that ( i ) and ( ii ) correspond to a @xmath171 singlet and a @xmath170 triplet , respectively , in agreement with the above result .
however , in contrast to the naive expectation , ( iii ) results in a @xmath183 triplet due to the hund s rule coupling . note that this @xmath183 triplet does @xmath67 become the unique ground state , as shown in fig .
1(b ) , except at @xmath164=0 . in this case
, the @xmath12 doublet appears only as an excited state , since this @xmath12 doublet is found to be composed of two singlet states , each of which includes a couple of electrons , one in @xmath160 and the other in @xmath163 .
as is easily understood , a local singlet state in multi - orbital systems is energetically penalized due to the hund s rule coupling , and thus , the @xmath12 doublet is always the excited state , higher than the @xmath183 triplet .
as long as we consider the @xmath1-@xmath1 coupling scheme for the @xmath1=5/2 sextet , the @xmath12 doublet never appears as the ground state , which is attributed to ignoring the @xmath1=7/2 octet .
however , it is @xmath67 allowed to include the effect of this octet in the present @xmath1-@xmath1 coupling scheme , since it just contradicts the original assumption .
namely , if the hund s rule coupling is strong enough so that @xmath1=7/2 can not be neglected , we must @xmath67 consider the @xmath1-@xmath1 coupling schemes , but take the @xmath40 coupling scheme from the outset .
here we have an interesting conclusion : in the @xmath1-@xmath1 coupling scheme , @xmath12 never becomes the ground state . only in the @xmath40 coupling scheme with @xmath175@xmath530 and @xmath164@xmath1820 ,
the ground state becomes @xmath12 . except for the disappearance of the @xmath12 doublet
, it is concluded that the cef terms can be treated within the @xmath1-@xmath1 coupling schemes .
now we consider the hopping motion of @xmath0 electrons based on the @xmath1-@xmath1 coupling scheme .
as discussed in the previous subsections , the one @xmath0-electron state is labelled by @xmath2 . within the second - order perturbation ,
the hopping amplitude @xmath184 includes two contributions , as shown in fig . 2 : ( a ) the direct hopping of @xmath0 electrons between the @xmath2-state at @xmath30 site and the @xmath33-state at @xmath185 site and ( b ) the hopping via ligand anions located along the @xmath34-direction from the @xmath0-ion site .
in order to evaluate both contributions , basically we employ slater s two - center integral method , @xcite as has been done in @xmath14-electron systems .
in fact , the table for two - center integrals has been extended to the case including @xmath0 orbitals by takegahara _
@xcite however , as emphasized in the previous subsections , it is necessary to pay due attention to the fact that spin and orbital are mixed .
thus , in this case , it is @xmath67 allowed to simply use the slater - koster table .
-electron hopping @xmath32 between neighboring sites along @xmath34-direction and ( b ) hopping amplitude @xmath186 via ligand anions . ] to avoid any confusion , it is convenient to step back to @xmath0-electron operators in the @xmath62=3 multiplet , defined as @xmath115 . using clebsch - gordan coefficients , @xmath115 with real - spin @xmath78
is related to @xmath120 as @xmath187 note again here that @xmath78 is defined to be @xmath79 ( @xmath80 ) for up ( down ) spin .
first , let us consider the direct hopping of @xmath0 electrons along the @xmath34-direction , as shown in fig .
since the real spin should be conserved in the hopping process , @xmath184 is given as @xmath188 where @xmath189 is the hopping amplitude of electrons between @xmath190- and @xmath191-states along the @xmath34-direction .
now the problem is reduced to the evaluation of @xmath189 .
although we can simply consult the paper of slater and koster,@xcite a convenient formula has been obtained by sharma for the overlap integral between two orbitals , @xmath190 and @xmath191,@xcite connected by unit vector @xmath34 .
it is expressed as @xmath192 where @xmath193 denotes slater s two - center integral through the @xmath78 bond , for instance , it is @xmath194 for @xmath62=@xmath195=3 and @xmath196 for @xmath62=3 and @xmath195=1 .
@xmath197 is the spherical harmonic function with @xmath85=@xmath198 , @xmath65 , @xmath62 .
@xmath199 and @xmath200 are polar and azimuth angles , respectively , to specify the vector * a * as @xmath201 let us consider the direct hopping between @xmath0 orbitals in nearest neighbor sites by putting @xmath62=@xmath195=3 .
after some calculations , we obtain the hopping amplitudes as follows . for diagonal elements ,
we obtain @xmath202 where the energy unit @xmath203 is given by @xmath204 where @xmath194 is the slater - koster two - center integral between adjacent @xmath0 orbitals .
note that @xmath205 .
for off - diagonal elements , we obtain @xmath206 and @xmath207 note that @xmath208=@xmath209 .
next we consider the hopping of @xmath0 electrons @xmath210 through ligand anions , as shown in fig .
assume that @xmath62-orbitals are situated along the @xmath34-direction and @xmath0-electrons hop to the neighboring @xmath0-site through these @xmath62-orbitals .
again we note that the real spin is conserved in the hopping process . in second - order perturbation , we can easily obtain @xmath211 where @xmath212 and @xmath213 are the energy levels for @xmath62- and @xmath0-orbitals and @xmath85 runs between @xmath198 and @xmath62 .
due to the orthogonality of spherical harmonics , it is easily reduced to @xmath214 it is stressed that the @xmath34-dependence of @xmath210 is the same as that of @xmath215 , although the energy unit is changed from @xmath194 to @xmath216 .
note that in general , the sign is also changed .
in previous subsections , we have completed the construction of a model hamiltonian which is expected to be a basic model to investigate the microscopic properties of @xmath0-electron systems .
however , it includes six states per site , i.e. , three kramers doublets . for practical purposes ,
it is more convenient to simplify the model further by discarding one of the three kramers doublets .
although the hopping amplitudes are obtained in a general form applicable to any type of crystal structure , it is necessary to set the crystal structure for the purpose of actual calculations .
first let us consider a simple cubic lattice composed of ions with @xmath0 electrons . under the cubic cef ,
as discussed above , for @xmath164@xmath530 the @xmath163 quartet becomes the ground state , while the @xmath160 doublet is the excited state .
this situation is considered to be realized when ligand anions are situated in the face - centered or body - centered position , due to a naive discussion of the point - charge picture .
note that the effect of ligand anions is included only through the cef term , and we do not consider its effect on @xmath0-electron hopping .
then , it is possible to consider a model consisting of @xmath163 , neglecting @xmath160 under the condition of @xmath164@xmath530 . to consider the model including @xmath163 levels ,
it is useful to define new operators with `` orbital '' degrees of freedom to distinguish two kramers doublets included in @xmath163 as @xmath217 for @xmath218 orbital electrons and @xmath219 for @xmath220 orbital electrons , respectively .
note that the @xmath160 state is given as a @xmath221 orbital as @xmath222 for the standard time reversal operator @xmath223=@xmath224 , where @xmath225 denotes an operator to take the complex conjugate , we can easily show the relation @xmath226 note that this has the same definition for real spin . for hopping in the @xmath227 plane and along the @xmath3-axis
, we simply set ( @xmath199 , @xmath200 ) to be ( @xmath228,0 ) , ( @xmath228,@xmath228 ) , and ( 0,0 ) for @xmath34=@xmath229=[1,0,0 ] , @xmath230=[0,1,0 ] , and @xmath231=[0,0,1 ] , respectively .
then , by using the general results in the previous subsection , we easily obtain @xmath184 between neighboring @xmath0 orbitals in the @xmath227 plane and along the @xmath3 axis .
further we transform the basis by the above definitions for operators with orbital degrees of freedom .
the results are given as@xcite @xmath232 for the @xmath229-direction , @xmath233 for the @xmath230 direction , and @xmath234 for the @xmath231 direction .
note that @xmath235=@xmath236=(9/28)@xmath194 . here
we note two points : ( i ) hopping amplitudes among @xmath163 orbitals are just the same as those for the @xmath16 orbitals of @xmath237 electrons.@xcite this is quite natural if we recall the fact that @xmath163 is isomorphic to @xmath238 , where @xmath12 indicates @xmath239 representation for the orbital part and @xmath240 denotes the spin part .
( ii ) the @xmath160 orbital is localized , since the corresponding wavefunction has nodes along the axis direction . since @xmath160 is the excited state for @xmath164@xmath530 and it is also localized
, electrons do not occupy the @xmath160 orbital .
after lengthy and tedious calculations to transform the basis of the coulomb interaction terms , the hamiltonian for @xmath163 is given by@xcite @xmath241 where @xmath242= @xmath243 and @xmath244=@xmath245 .
note that in the second term , the level splitting @xmath246 is simply introduced to include partially the effect of a tetragonal cef . in the coulomb interaction terms , @xmath118 , @xmath119 , @xmath55 , and @xmath247 denote intra - orbital , inter - orbital , exchange , and pair - hopping interactions , respectively , expressed by the racah parameters @xmath97 as @xmath248 note that the relation @xmath118=@xmath119+@xmath55+@xmath247 holds , ensuring rotational invariance in pseudo - orbital space for the interaction part .
we believe that this @xmath163 hamiltonian provides a simple , but non - trivial model to consider superconductivity and magnetism in @xmath0-electron systems .
note again that it is essentially the same as the model for @xmath16 electron systems such as manganites,@xcite although the coupling with jahn - teller distortion is not included in the present model .
due to the complex interplay and competition among charge , spin , and orbital degrees of freedom , a rich phase diagram has been obtained for manganites .
@xcite thus , it is definitely expected that a similar richness will also be unveiled for @xmath0-electron systems based on the @xmath163 model hamiltonian .
the @xmath163 hamiltonian @xmath249 is believed to be a non - trivial model for @xmath0-electron systems , since we can immediately grasp some essential aspects of @xmath0-electron materials viewed as charge - spin - orbital complex systems .
in fact , based on @xmath249 , we can have insight into the effect on superconductivity of cooperation and competition among spin and orbital fluctuations in the new family of heavy fermion materials cemin@xmath4 ( m = ir , co , and rh ) , with the hocoga@xmath4-type tetragonal structure .
@xcite among cemin@xmath4 materials , the de haas - van alphen ( dhva ) effect has been successfully observed in ceirin@xmath4 and cecoin@xmath4,@xcite both of which have huge electronic specific heat coefficients of several hundreds of mj / k@xmath250mol .
the angular dependence of major experimental dhva frequency branches is well explained by a quasi two - dimensional fermi surface.@xcite this is a clear advantage when we construct a model hamiltonian , since it is enough to consider the two - dimensional case . in order to analyze further the electronic properties from a quantitative viewpoint , it is desirable to include more precisely the effect of the actual crystal structure .
for instance , in the previous subsection , we have introduced the level splitting @xmath246 between @xmath163 levels in a simple way , in order to take into account partially the effect of a tetragonal cef . however , in an actual tetragonal situation , there occur three kramers doublets , two @xmath160 and one @xmath240 .
thus , in this subsection , we attempt to incorporate the effect of a tetragonal cef more faithfully into the model .
first let us consider the tetragonal cef term .
after the diagonalization , we obtain three kramers doublets , two @xmath160 and one @xmath240 , expressed as @xmath251 the coefficients @xmath252 and @xmath253 are determined by the cef parameters as @xmath254 where @xmath199=@xmath255 and @xmath256 with @xmath257 the corresponding eigen energies are given by @xmath258 respectively . if the @xmath240 level is higher than the other two @xmath160 levels , it is possible to obtain another simplified model for the @xmath160 levels .
in fact , in some experimental papers on cemin@xmath4 , we have found that the position of the @xmath240 is the highest among the three kramers doublets.@xcite thus , a simplified and realistic model for cemin@xmath4 can be constructed by including only two @xmath160 levels .
( m = ir , rh , and co ) . ] in this crystal structure , the ce ions form a two - dimensional tetragonal lattice , while the in ions are located in the face - centered positions of this ce - ion network , as shown in fig . 3 .
we include the nearest - neighbor hopping between ce @xmath259 orbitals as well as next - nearest neighbor hopping via in @xmath260 orbitals .
due to the symmetry , the latter process contributes only along the diagonal direction , while hopping via @xmath252-orbitals along the bond direction vanishes . based on the same two - dimensional lattice
, we have further constructed a hamiltonian called the `` @xmath0-@xmath252 model '' by keeping the @xmath0-@xmath252 hybridization explicitly.@xcite in order to compare the tight - binding dispersion directly with the results of a relativistic band - structure calculation , it is important to keep the hybridization term .
however , here we consider an @xmath0-electron model wherein we include the effects of @xmath0-@xmath252 hybridization in the next - nearest neighbor hopping of @xmath0 electrons .
we believe it will be adequate for the treatment of superconductivity and magnetism if we further add the coulomb interaction terms to the model .
since two @xmath160 s are given by the linear combinations of @xmath2=@xmath943/2 and @xmath945/2 , we obtain the @xmath160 model just by discarding the @xmath2=@xmath941/2 states .
it is convenient to define additional new operators as @xmath261 again , we can easily show the relation @xmath262=@xmath263 . for the nearest neighbor hopping ,
we obtain @xmath264 note that @xmath235 has the same definition as for the @xmath163 model . on the other hand , for the next - nearest neighbor hopping via the @xmath252 orbitals , we obtain @xmath265 for both @xmath266 $ ] and @xmath267 $ ] directions .
the next - nearest neighbor hopping amplitude is given by @xmath268 where @xmath196 is the slater - koster two - center integral for @xmath0 and @xmath252 orbitals .
note that in general @xmath269@xmath530 .
the @xmath160 model is then written as @xmath270 where @xmath34 and @xmath271 , respectively , denote the vectors connecting nearest- and next - nearest neighbor sites . @xmath272 and @xmath273 are pauli matrices . the energy level shift has been neglected here for simplicity .
the intra - orbital coulomb interactions are given by @xmath274=@xmath107@xmath57@xmath275+@xmath142 and @xmath276=@xmath107+@xmath108+@xmath142 for the 1 and 2 orbitals , respectively , and the inter - orbital coulomb interaction is denoted by @xmath119=@xmath107@xmath57@xmath108 .
the hund s coupling and the pair - hopping interactions are given by @xmath55=@xmath277 and @xmath247=@xmath142 , respectively .
three comments on @xmath278 are in order .
( 1 ) there is no transverse component in the hund s coupling term .
( 2 ) @xmath274@xmath279@xmath276 since the 1 and 2 orbitals belong to different symmetries . ( 3 ) for the case of @xmath42=1 , both @xmath55 and @xmath247 are sometimes neglected for simplicity , since they become irrelevant in the large-@xmath119 limit . when we are interested in the strongly correlated region , it is possible to safely set @xmath142=@xmath141=0 for the case of @xmath42=1 , leading to @xmath274=@xmath276=@xmath119=@xmath107 .
in the previous section , we have discussed in detail the construction of a microscopic model for @xmath0-electron systems based on the @xmath1-@xmath1 coupling scheme . for practical purposes ,
we have also prepared a couple of simplified models corresponding to an actual crystal structure .
those models can be analyzed with several techniques from both an analytical and a numerical viewpoint .
it is impossible to show here the results for all models . in this section ,
we show exact diagonalization results for @xmath249 as a typical example of an orbital degenerate system . before proceeding to the exhibition of calculated results ,
let us discuss the quantities we would measure in a model including orbital degree of freedom .
first of all , in order to see the spin and charge structure , it is quite natural to measure the spin and charge correlation functions , defined as @xmath280 and @xmath281 respectively , where @xmath282@xmath57@xmath283 and @xmath284 . on the other hand , in order to see the orbital structure , it is quite useful to calculate the orbital correlation function , given by @xmath285 as is easily understood from the above equations , even if the orbital degree of freedom is further included in the model hamiltonian , there is no ambiguity in the definitions for spin and charge correlations , given by sums in terms of an orbital index . for the orbital correlations , we also consider as an orbital - dependent charge correlation , which can be defined without any difficulty .
however , when we attempt to discuss the superconducting pair correlation , we immediately encounter some serious problems .
namely , it is necessary to determine the form factors for the pairing states in both momentum and orbital space .
note that even in the single - orbital hubbard model , it is necessary to determine the momentum dependence , but in general , our task can be reduced through the help of a group - theoretical analysis . in the orbital degenerate model , the extra task of determining the form factor in orbital space
is also imposed upon us . especially in the present realistic models , in which the orbital degree of freedom does @xmath67
provide a good quantum number , the situation seems to be very severe at first glance , since we can not resort to a group - theoretical analysis for the orbital dependence of the form factor . in order to overcome this problem ,
we carefully reconsider it based on the concept of off - diagonal long - range order.@xcite let us define the pair correlation as @xmath286 where the subscript @xmath57 ( @xmath287 ) denotes spin singlet ( triplet ) pairing and @xmath288 is a pair operator , given in general by @xmath289 since a pair with zero total momentum is considered here , @xmath290 depends only on @xmath34 , which is the vector connecting the sites .
note also that the orbital dependence of @xmath290 is not solely determined by symmetry considerations .
we can easily show that the pair correlation is written in the form @xmath291 where the matrix element @xmath292 is written as @xmath293/2.\end{aligned}\ ] ] here it is necessary to recall the concept of off - diagonal long - range order.@xcite namely , the occurrence of superconductivity is detected when the largest eigenvalue of the pair correlation matrix becomes the order of @xmath294 , where @xmath294 is the number of electrons .
thus , in a small cluster calculation , the possible superconducting pair state should be defined by the eigenstate with the maximum eigenvalue @xmath295 of the matrix @xmath296 . of course , in order to prove the existence of off - diagonal long - range order , we need to show that @xmath295 becomes the order of @xmath294 with increasing the cluster size . however , at some cluster size , a possible pairing state is determined without ambiguity by the eigenstate with @xmath295 . a prescription to define the pair operator is briefly summarized .
first , we determine the pair correlation matrix @xmath292 .
then we diagonalize it , and the form factor @xmath290 will be defined by the eigenstate corresponding to the maximum eigenvalue of the pair correlation matrix . in the previous subsection , we have discussed a way to define the pair operator in orbital degenerate systems .
here we further consider the symmetry of the cooper pair in orbital space from a general point of view .
anomalous green s functions for spin triplet and singlet pairs are , respectively , defined by @xmath297,\ ] ] and @xmath298,\ ] ] where @xmath299=@xmath300 and @xmath301 denotes imaginary time . due to cyclic invariance of the trace
, we obtain the following relations : @xmath302 where @xmath220 is the inverse temperature .
further , after fourier transformation , we obtain @xmath303 where @xmath304 is the fermion matsubara frequency , given by @xmath304=@xmath305(2@xmath42 + 1 ) with @xmath42 an integer .
these are the basic relations representing the symmetry of cooper pairs in orbital degenerate systems .
except for differences in the spin states such as singlet and triplet , `` parity '' is the only extra good quantum number to specify the pairing state , since there is no conservation law regarding orbital degrees of freedom .
thus , we can produce even- and odd - parity states , satisfying @xmath306 where @xmath307 indicate the operator to change @xmath308 to @xmath309 .
it is easy to understand that @xmath310 is given by linear combinations of @xmath311 and @xmath312 .
although we can not define @xmath310 uniquely , it is convenient to consider combinations of odd and even functions both for orbital exchange and sign - change in frequency .
if we introduce two orbitals `` a '' and `` b '' , then after some simple algebra , we obtain @xmath313 where @xmath314 is an arbitrary number and @xmath315 is given by @xmath316 since the present hamiltonian is @xmath67 invariant with orbital exchange , @xcite the terms `` orbital symmetric '' and `` orbital antisymmetric '' can not be used as a label to specify the state .
however , it is convenient to categorize @xmath315 as follows : the @xmath317 ( @xmath49=1@xmath1403 ) are orbital - symmetric and even function of frequency , while the @xmath318 ( @xmath49=1@xmath1403 ) are orbital - symmetric and odd function of frequency . on the other hand
, @xmath319 is orbital - antisymmetric and an odd function of frequency , while @xmath320 is orbital - antisymmetric and an even function of frequency .
if we consider the zero - frequency limit , we can simply drop the odd functions of frequency .
namely , the even - parity pair is expressed as a combination of orbital - symmetric functions , while the odd - parity solution is just given by @xmath320 , the orbital - antisymmetric part .
note , however , that this is just an accident due to the vanishing of odd - functions of frequency .
we must @xmath67 consider that the orbital state provides a good label to specify the pairing symmetry , although it is a convenient way to classify the pair state .
again , it is stressed that spin state and parity are labels used to specify the pairing symmetry . for the case of spin triplet , it is easy to repeat the same discussion as that for spin singlet .
note that in the zero - frequency limit , the odd - parity solution is given by the combination of orbital - symmetric parts , while the even - parity function is expressed only by the orbital - antisymmetric part .
up to now we have discussed the general situation in which the system is @xmath67 invariant under the exchange of orbitals .
we believe that this is quite natural since in actual materials , the shapes of orbitals are different from one another .
however , it is possible to consider a fictitious hamiltonian which is invariant under orbital exchange . in such a situation ,
the words `` orbital symmetric '' and `` orbital antisymmetric '' recover physical meanings which distinguish the pairing states .
namely , we can further classify anomalous green s functions as @xmath321 where the subscripts @xmath287 and @xmath57 denotes parity even and odd , while `` s '' and `` a '' indicate the orbital - symmetric and orbital antisymmetric , respectively . in this case
, the frequency dependence is automatically classified : even - function for @xmath322 and @xmath323 and odd - function for @xmath324 and @xmath325 .
finally , it is again stressed that when we consider pair correlation in the static limit , at first glance , orbital symmetric and antisymmetric parts seem to be separated , but such a separation is spurious owing to the speciality of the static limit .
next we show results obtained using exact diagonalization techniques for an 8-site tilted cluster .
since we attempt to make some comparison with cemin@xmath4 , we consider a two - dimensional lattice .
first we consider the case of quarter filling , i.e. , @xmath42=1 .
note here that @xmath55 and @xmath247 are simply neglected , since they are considered to be irrelevant in the case of @xmath42=1 . as shown in fig .
4(a ) , for @[email protected] , there is no dominant component in @xmath326 , indicating a paramagnetic ( pm ) phase , while for @[email protected] , @xmath327 becomes abruptly dominant , strongly suggesting an antiferromagnetic ( af ) phase .
this abrupt change is due to level crossing , but it is instructive to see the charge and orbital correlations . in fig .
4(b ) , we show charge correlation , exhibiting again an abrupt change around @[email protected] .
a decrease of charge correlation is the signature of a metal - insulator transition , since the charge fluctuations are suppressed in the mott insulator . as observed in fig .
4(c ) , we can see an orbital disordered ( od ) phase for @[email protected] and a ferro - orbital ( fo ) phase for @[email protected] . by monitoring the change in spin correlations
, we can draw the ground - state phase diagram for @xmath42=1 , as shown in fig .
we see od / pm phase for small @xmath118 and fo / af phase for large @xmath118 region . for @xmath118=@xmath119=4 and @xmath55=@xmath247=0 .
( d ) ground - state phase diagram for the @xmath163 model for @xmath118=@xmath119 with @xmath55=@xmath247=0 .
( e ) pair correlations vs. @xmath246 both for spin singlet and triplet states .
( f ) pair amplitude in momentum space for the singlet state having the maximum eigenvalue of the pair correlation matrix .
note that it is shown separately for @xmath328 , @xmath329 , and @xmath330+@xmath331 pair states . ] in fig .
4(e ) , we show the eigenvalues of the pair correlation matrices for spin singlet and triplet states as a function of @xmath246 .
the state with the largest eignevalue is a singlet , while a triplet appears as the second largest eigenvalue state .
even if we introduce a finite value of @xmath55 , unfortunetely , this triplet state is @xmath67 the largest eigen - value state , at least within the present 8-site calculation .
however , in the previous analysis using the random phase approximation , the appearance of a triplet state has been suggested in the small @xmath246 region.@xcite it may be related to the present result having the triplet state as the second largest eigenvalue state . in fig .
4(f ) , the pair amplitudes for the largest eigen - value state is shown in momentum space .
since we are measuring the static pair correlation , we can observe here an `` orbital symmetric '' state with `` aa '' , `` bb '' , and `` ab+ba '' , while the `` orbital antisymmetric '' part such as `` ab@xmath57ba '' is not mixed .
as discussed in the previous subsection , it is just an accident due to the static limit , and the orbital symmetric state is @xmath67 separated in general .
next we focus on the momentum dependence of each component . as seen in fig .
4(f ) , the
`` aa '' and `` bb '' parts of the pair amplitude exhibit @xmath332 , while `` ab+ba '' seems to belong to @xmath333 . if we recall that `` a '' and `` b '' each correspond to @xmath16 orbitals , then they have different signs for @xmath334 rotation in space
thus , there exists a discrepancy in the symmetry between ( `` aa '' , `` bb '' ) and `` ab+ba '' components , although the total symmetry should be @xmath332 in this case . we have obtained a simultaneous onset of spin and orbital ordering , which can be interpreted as a transition between pm and afm phases controlled by the suppression of orbital fluctuations .
moreover , just around the transition regime between pm and afm phases , we can expect the appearance of @xmath14-wave superconductivity induced by afm spin fluctuations , which has been identified as the @xmath332 pairing state with the largest eigenvalue .
in fact , the present model has been analyzed using the random phase approximation , leading to a transition between pm and @xmath14-wave superconducting phases with an increase of @xmath246.@xcite if we further increase @xmath246 , eventually the system becomes an afm insulating phase , and this successive transition agrees with the present results .
here it is stressed that the orbital degree of freedom plays a key role to control the ground - state properties .
we believe that this viewpoint is applicable to the appearance of @xmath14-wave superconductivity in cemin@xmath4 , although the quantitative features should be discussed in detail using the @xmath160 model .
= 0 and @xmath42=1.5 . note that the region with @xmath335 is unphysical .
( b ) amplitude for triplet pairing in momentum space , evaluated from the eigenstate corresponding to the maximum eigenvalue of the pair correlation matrix . ]
next we consider the case of @xmath42=1.5 for the same @xmath163 model .
for @xmath246=0 , we depict the phase diagram in the ( @xmath119,@xmath55 ) plane , as shown in fig .
note that the region with @xmath55@xmath52@xmath119 is unphysical and here we just ignore it .
we see the appearance of the ferromagnetic ( fm ) phase , which is stabilized due to the strong hund s rule coupling .
the interesting point is the pair amplitude inside the fm phase , as shown in fig .
the momentum dependence of the largest eigenvalue state is characterized by even parity . as naively expected , due to the hund s rule coupling
, a local triplet is formed , but in order to gain kinetic energy , this pair should be spatially extended , as shown in fig .
note here that this is @xmath333-like , since it is an even - parity state .
since we are now considering the pairing state in the strong coupling region , there is no reason to exclude the triplet state with even parity , which is @xmath67 directly connected to the usual odd - parity triplet state .
this is not surprising , since local triplet pair must have even parity , provided that one ion with @xmath0-orbitals is included in the unit cell .
thus , quite generally , we can conclude that the odd - parity triplet pair is not stabilized in the local hund s rule coupling , at least in the bravais lattice , which is consistent with the previous discussion.@xcite as for the mechanism of the attractive interaction in the pair inside the fm phase , it should be considered to be orbital fluctuations , since spin fluctuations are dead in the fully spin - polarized state .
this point is intriguing due to a possible connection with the coexistence of superconductivity and ferromagnetism in uge@xmath5 .
of course , in order to confirm such an exotic type of superconductivity , it would be necessary to consider the problem in the bulk limit using , for instance , the green s function method .
this case will be discussed elsewhere .
finally , we note that due care should be paid to the correspondence between pseudo - spin and total angular momentum @xmath54 , when we intend to make a comparison with actual materials . in experiments on @xmath0-electron compounds , for instance
, the external magnetic field couples to @xmath54 , not to the pseudo - spin .
thus , when we attempt to calculate physical quantities observed in actuality , it is necessary to consider the response of @xmath54 .
the present analysis based on the pseudo - spin model is useful to see qualitiave tendencies , while a quantiative discussion is left for future development .
-electrons in ce , u , and pu ions for the pseudo spin representations .
each horizontal bar indicates the pseudo orbital . ]
we have constructed a microscopic model hamiltonian for @xmath0-electron systems based on the @xmath1-@xmath1 coupling scheme : the coulomb matrix elements have been expressed in terms of the racah parameters , and the cef terms for the two - electron case have been determined by those for @xmath55=5/2 .
the validity of this @xmath1-@xmath1 coupling picture has been carefully dicsussed in the present paper .
the @xmath0-electron hopping amplitudes are obtained using slater s two - center integral .
two types of simplified models have also been presented : one is a hamiltonian including @xmath163 and the other for @xmath160 levels .
these are believed to be useful for further investigations of superconductivity and magnetism in @xmath0-electron systems in this paper the @xmath163 model has been analyzed using the exact diagonalization method .
we have measured several kinds of correlation functions and developed a prescription to define the pairing state in multi - orbital systems .
furthermore , the symmetry of the orbital dependent cooper pair has been carefully discussed . in the analysis of the @xmath163 model , it has been suggested that unconventional superconductivity should appear around the phase boundary between metallic od - pm and insulating fo - afm phases .
in addition , we have found evidence for triplet superconductivity in the ferromagnetic phase , induced by strong hund s rule coupling . with the present method of defining the pairing state
, it has been found that the triplet pairing state is @xmath67 localized , but extended spatially to gain kinetic energy .
based on the @xmath1-@xmath1 coupling scheme , it is easy to consider the case of several @xmath0 electrons per site .
although we have not shown any results in this paper for the case of @xmath42=2 , corresponding to the @xmath336 ion ( see fig .
6 ) , we have obtained that the ground - state is an afm insulator , which is driven by afm superexchange coupling between neighboring @xmath60=1 spins formed by the local hund s rule coupling .
when we introduce a level splitting into this afm insulating state , we have found robustness for this phase , compared with the transition around @xmath246=0.5 for the case of @xmath42=1 . for umga@xmath4@xcite isostructural with cemin@xmath4 ,
if the uranium ions are considered to take u@xmath69 , again it is naively deduced to have an afm ground state , if the local spin is formed by hund s rule coupling , as shown in fig .
however , without any calculations , it is impossible to determine whether it is a metallic or insulating state , although in experiments , metallic afm phase has been suggested.@xcite this is left as a future problem . as introduced in sec . 1
, pucoga@xmath4 amazingly exhibits superconductivity with @xmath8=18.5k.@xcite in order to understand this high @xmath8 , first of all , we generally consider that the spatial extent of the @xmath337 wavefunction is large compared to that of @xmath259 electrons , leading to a larger hybridization among conduction and @xmath337 electrons .
namely , the basic energy scale ( or effective bandwidth of the quasiparticles ) in the pu - compound should be larger than that of ce - materials .
here we envision a scenario similar to the high-@xmath8 cuprates : if superconductivity in those `` 115 '' materials is considered to originate from an electronic mechanism , @xmath8 should be increased as the energy scale is enhanced . it may be useful to point out a possible analogy with high-@xmath8 materials , but we emphasize as a separate important issue the active role of orbitals , which discussed intensively in this paper .
it has been deduced from the local moment behavior that the plutonium valence is pu@xmath69,@xcite indicating that five @xmath0-electrons are accommodated . in our @xmath1-@xmath1 coupling picture with pseudo - spin representation ,
we can put five electrons in the sextet of @xmath1=5/2 , as shown in fig .
6 , which is just the hole version of cemin@xmath4 .
thus , the appearance of superconductivity itself may be understood without any calculations on the basis of our @xmath1-@xmath1 coupling picture .
as explained above , the drastic changes of @xmath8 in the `` 115 '' materials may be understood , based on the difference in the energy scale for @xmath259 and @xmath337 electrons , but we believe that another key issue related to this problem would be the shape of the orbitals . a difference between the electron and hole pictures should exist stemming from the shape of the active orbital , as easily deduced from fig .
if such differences among orbitals , i.e. , level schemes relating to the cef effect , will be clarified in those materials , it may be possible to obtain a unified view to understand unconventional superconductivity and novel magnetism in `` 115 '' materials including cemin@xmath4 , umga@xmath4 , and pucoga@xmath4 .
this point will be further investigated based on the present model . as repeatedly mentioned in this paper ,
there remain several open questions in the electronic properties of @xmath0-electron materials .
we believe that the microscopic models presented and discussed in this paper will be useful to advance our understanding of unconventional superconductivity and novel magnetism in @xmath0-electron systems .
the authors thank t. maehira , k. miyake , y. nuki , t. takimoto , and k. yamada for discussions .
we are grateful to r. e. walstedt for his valuable comments .
the authors are separately supported by the grant - in - aid for scientific research from japan society for the promotion of science .
w. schlabitz , j. baumann , b. pollit , u. rauchschwalbe , h. m. mayer , u. ahlheim , and c. d. bredle , abstracts of ivth int .
conf . on valence fluctuaions ( cologne , 1984 ) ; t. t. m. pastra , a. a. menovsky , j. van den berg , a. j. dirkmaat , p. h. kes , g. j. nieuwenhuys , and j. a. mydosh , phys . rev . lett . *
55 * , 2727 ( 1985 ) .
s. s. saxena , p. agarwal , k. ahilan , f. m. grosche , r. k. w. haselwimmer , m. j. steiner , e. pugh , i. r. walker , s. r. jullian , p. monthoux , g. g. lonzarich , a. huxley , i. sheikin , d. braithwaite , and j. flouquet , nature ( london ) * 406 * , 587 ( 2000 ) .
e. d. bauer , n. a. frederick , p. -c . ho , v. s. zapf , and m. b. maple , phys . rev .
b * 65 * , 100506(r ) ( 2002 ) ; m. b. maple , p. -c . ho , v. s. zapf , n. a. frederick , e. d. bauer , w. m. yuhasz , f. m. woodward , and j. w. lynn , j. phys . soc .
. suppl . * 71 * , 23 ( 2002 ) .
h. tou , y. kitaoka , k. asayama , n. kimura , y. nuki , e. yamamoto , and k. maezawa , phys .
77 * , 1374 ( 1996 ) ; h. tou , y. kitaoka , k. ishida , k. asayama , n. kimura , y. nuki , e. yamamoto , y. haga , and k. maezawa , phys .
lett . * 80 * , 3129 ( 1998 ) ; y. murakami , h. kawada , h. kawata , m. tanaka , t. arima , y. moritomo , and y. tokura , phys .
lett . * 80 * , 1932 ( 1998 ) ; y. murakami , j. p. hill , d. gibbs , m. blume , i. koyama , m. tanaka , h. kawata , t. arima , y. tokura , k. hirota , and y. endoh , phys .
* 81 * , 582 ( 1998 ) .
h. shishido , r. settai , d. aoki , s. ikeda , h. nakawaki , n. nakamura , t. iizuka , y. inada , k. sugiyama , t. takeuchi , k. kindo , t. c. kobatashi , y. haga , h. harima , y. aoki , t. namiki , h. sato . and y. nuki , j. phys . soc . jpn .
* 71 * , 162 ( 2001 ) . as mentioned here
, the @xmath163 hamiltonian @xmath249 is not invariant for the operation of orbital exchange , but if the operation of spatial rotational is also combined with orbital exchange , it is possible to make @xmath249 invariant . | we construct a microscopic model for @xmath0-electron systems , composed of @xmath0-electron hopping , coulomb interaction , and crystalline electric field ( cef ) terms . in order to clarify the meaning of one @xmath0-electron state , here
the @xmath1-@xmath1 coupling scheme is considered , since the spin - orbit interaction is generally large in @xmath0-electron systems .
thus , the @xmath0-electron state at each site is labelled by @xmath2 , namely , the @xmath3-component of total angular momentum @xmath1 . by paying due attention to @xmath0-orbital symmetry ,
the hopping amplitudes between @xmath0-electron states are expressed using slater s integrals .
the coulomb interaction terms among the @xmath2-states are written by slater - condon or racah parameters .
finally , the cef terms are obtained from the table of hutchings .
the constructed hamiltonian is regarded as an orbital degenerate hubbard model , since it includes two pseudo - spin and three pseudo - orbital degrees of freedom . for practical purposes
, it is further simplified into a couple of two - orbital models by discarding one of the three orbitals .
one of those simplified models is here analyzed using the exact diagonalization method to clarify ground - state properties by evaluating several kinds of correlation functions .
especially , the superconducting pair correlation function in orbital degenerate systems is carefully calculated based on the concept of off - diagonal long - range order .
we attempt to discuss a possible relation of the present results with experimental observations for recently discovered heavy fermion superconductors cemin@xmath4 ( m = ir , co , and rh ) , and a comprehensive scenario to understand superconducting and antiferromagnetic tendencies in the so - called `` 115 '' materials such as cemin@xmath4 , umga@xmath4 , and pucoga@xmath4 from the microscopic viewpoint . |
the quark and gluon substructure of the hadrons is ruled by quantum chromodynamics ( qcd ) , and it is reflected in the baryon sector by a set of bumps in the cross sections of different probing processes , taken as functions of the center of mass energy @xmath10 .
these bumps are identified as baryon resonances characterized by spin , isospin , orbital angular momentum , radial excitation and parity quantum numbers .
the lowest energy bump , the @xmath11 baryon , is clearly isolated from the background as a state of spin and isospin 3/2 and positive parity . heavier resonances are not so clearly isolated from the background .
this happens in the so called second resonance region , where the @xmath12 , @xmath13 and @xmath14 resonances show up .
although in quark models these resonances can be described as three - quark systems confined by a potential like the harmonic oscillator potential @xcite , some properties , like their decay width , can be better understood within a dynamical meson - baryon coupled - channel reaction model . also , in constituent
quark models the baryon spectrum is difficult to interpret since the negative parity partner of the nucleon , the @xmath2 state ( @xmath15 ) is lighter than the first radial excitation of the nucleon ( @xmath16 ) , the roper ( or @xmath17 state ) @xcite .
it was only recently that lattice qcd simulations with very small pion masses @xcite , reconstructed the natural order of the baryon spectrum ( where the @xmath2 state is heavier than the @xmath17 state ) , suggesting a fundamental role of the quark - antiquark polarization , or meson cloud dressing , in the baryon systems , as a correction to the valence quark effects .
in this work we will use the notation @xmath1 to represent the @xmath14 nucleon excitation ( @xmath18 ) , and we will focus on the electromagnetic structure of this resonance , in particular on the calculation of the @xmath0 transition form factors , within a covariant constituent quark model .
precise data for the @xmath0 amplitudes is available at present @xcite . besides being one of the lightest nucleon resonances ,
the @xmath1 baryon is particularly interesting for several reasons : it is very well isolated in the spin 1/2 and negative parity configuration ; it decays strongly to the @xmath19 channel ( with a branching ratio @xmath20 50% ) , allowing a very precise determination of the electromagnetic structure , and providing therefore an extra challenge for theoretical models . also , because of the strong coupling with the @xmath21 channel ( with a branching ratio @xmath22 50% ) , the @xmath1 is crucial for the analysis of meson photoproduction from the nucleon @xcite .
another interesting aspect of the @xmath1 is its vicinity to another @xmath2 resonance with higher mass , the @xmath23 also called as @xmath24 .
the two resonances differ in their decay modes , and the differences in their structure is yet to be explored .
several formalisms have been used to describe the @xmath1 system .
they are based either on quark models or on effective meson - baryon interaction models . in the first case , there are non - relativistic constituent quark models @xcite , relativistic quark models @xcite , quark models with explicit quark - antiquark contributions @xcite and qcd sum rules @xcite .
alternatively , in the second case , the @xmath1 is interpreted as a molecular - type state dynamically generated by the meson - nucleon interaction @xcite with a particular dominance of the @xmath25 quasi - bound state @xcite .
a particular class of effective meson - baryon interaction models are the dynamical coupled - channel reaction models @xcite , where the baryon bare core is parametrized phenomenologically and the meson dressing is included non perturbatively .
thus @xmath1 does not only provide a crucial test for the methods just mentioned , but it is also a crucial resonant structure for the analysis of nucleon excitation reactions @xcite . within a constituent quark model picture ,
the nucleon excitation @xmath1 can be represented as a mixture of two different configurations .
since the @xmath2 excitation has total angular momentum @xmath26 and orbital angular @xmath27 ( @xmath28 state excitation ) , its core spin may be either @xmath29 or @xmath30 .
then , in the usual spectroscopic notation @xcite , the @xmath2 channel of the nucleon excitation is a mixture of the @xmath31 and @xmath32 states , which have spin 1/2 and 3/2 respectively .
this mixture of the two core spin components is defined by a mixing angle @xmath33 determined by a color hyperfine interaction between the quarks , which may have distinct origins : one - gluon - exchange @xcite , one - pion - exchange @xcite or goldstone - boson - exchange @xcite . in the classical isgur - karl model
it turns out that the spin core spin 1/2 component dominates in the @xmath1 , with a mixing angle given by @xmath34 @xcite . in our work
, we apply the covariant spectator quark model , which is based on the covariant spectator theory @xcite , to the @xmath1 system .
the model describes the nucleon @xcite , the roper @xcite , the @xmath35 and the @xmath36 @xcite experimental form factors , as well as the lattice qcd simulations for the nucleon , the @xmath37 transition , and the baryon decuplet @xcite . in our framework
the baryons are represented as a quark - diquark system .
the quark couples to the electromagnetic field by means of a constituent quark current which is parametrized by vector meson dominance , and the diquark is a spectator during the electromagnetic interaction , and therefore is taken on - mass - shell @xcite .
the model is phenomenological since it does not derive the structure of the baryon from a dynamical wave function equation . instead
, the baryon systems are described effectively in terms of their intrinsic properties ( spin , flavor , angular orbital momentum and parity ) which dictate the form of their wave function and the experimental value of their mass @xmath38 . as in the previous applications of the model , in particular to the @xmath39 and the roper resonances , we are focused on the role of the valence quarks for the electromagnetic transition .
because of this and also as a consequence of the kinematics ( the difference of mass between the @xmath1 and the nucleon is 0.60 gev ) our model can only be applied to the high @xmath9 region . as we will show , the domain of validity of our calculations can even be established more precisely and quantitatively , as the region @xmath3 gev@xmath4 . in this region
the meson cloud effects are expected to be small and valence quark degrees to dominate .
we use two additional assumptions : i ) the @xmath1 is represented exclusively by the spin 1/2 core [ no mixture with the @xmath24 excitation ] ii ) the diquark is pointlike . with these assumptions , and
taking the momentum distribution of the diquark the same as for the nucleon , we relate the nucleon and the @xmath1 wave functions .
these assumptions allow us to reduce the number of degrees of freedom to a minimum , since no additional parameters to the ones taken for the nucleon case are needed to describe the spin 3/2 core contributions , or the diquark internal structure .
our results are then true predictions , with no new adjustable parameters .
all parameters were fixed in the previous applications by the quark current and nucleon wave function , represented as s - wave system .
both assumptions can be tested in the future , once the structure of the nucleon is extended to the inclusion of p- and d - states , which demand in turn a spin 3/2 core and/or diquark with internal p - state structure @xcite .
this work will be organized as follows : in sec .
[ secsqm ] we introduce the wave functions of the nucleon and the @xmath1 ( details in appendix [ apwf ] ) . in sec .
[ seccurrent ] we derive the transition current for the @xmath0 transition ( with details presented in appendix [ apcurrent ] ) . explicit formulae for the form factors and helicity amplitudes come in sec .
[ secff ] . in sec .
[ secwf ] we parametrize the momentum dependence of the wave functions . the results and discussion are presented in sec .
[ secresults ] and the conclusions in sec .
[ secconclusions ] .
when the momentum transfer exceeds the mass of the constituent quarks the electromagnetic excitation requires necessarily a relativistic treatment .
this is one of the reasons for us to use the framework provided by the covariant spectator quark model for baryons @xcite . in this formalism
the baryons are phenomenologically described as constituent quark systems , and the covariant wave function has a form compatible with their symmetry properties ( flavor , spin , orbital angular momentum and parity ) and a totally anti - symmetric color wave function @xcite . for the nucleon
the s - state approximation was made and the spin , flavor and spatial wave function and is represented by @xcite _ n(p , k)= _ n(p , k ) , [ eqpsin ] where the nucleon and diquark four momenta are @xmath28 and @xmath40 respectively , @xmath41 is a dirac spinor , @xmath42 the diquark polarization vector in the fixed - axis representation @xcite and u^(p)= _ 5 ( ^- ) u(p ) , [ equn ] the spin 1/2 vector spin state [ direct product of states 1 ( diquark ) and 1/2 ( quark ) for a total spin state of 1/2 ] .
@xmath43 is the nucleon mass .
the wave function ( [ eqpsin ] ) is written in terms of the states corresponding to a diquark composed by the quark pair ( 12 ) and the quark 3 .
the isospin functions @xmath44 depend on the isospin projection @xmath45 and are shown in table [ tabisospin ] .
note that the spin-0 ( isospin-0 ) and the spin-1 ( isospin-1 ) states are respectively anti - symmetric and symmetric in the exchange of quarks 1 and 2 .
.isospin states for the nucleon and @xmath2 systems . [ cols="<,^,^",options="header " , ] to write down the @xmath1 wave function we applied the @xmath46 constituent quark model representation , where the @xmath1 state is a member of the @xmath47 $ ] supermultiplet ( dimension 70 , with @xmath48 ) , and part of the @xmath49 subset ( octet with @xmath50 ) @xcite .
we have also followed very closely the notation established in refs .
we use @xmath51 to label the @xmath1 mass .
the @xmath1 is defined as the excitation of the nucleon to the state @xmath52 .
this state has the same flavor content and the same spin ( 1/2 ) of the nucleon , but has negative parity .
the negative parity defines a spatial symmetry implied by the excitation of internal relative angular momentum @xmath27 , and requires the presence of p waves at least in one quark pair .
consequently , the spin structure also changes relatively to the one of the nucleon , in order to accommodate a total symmetric form for the flavor - spin - momentum space wave function . to represent the wave function in a basis of momentum states ,
one decomposes , as usual , the system into a pair of quarks [ or diquark labeled ( 12 ) ] , and a spectator quark [ labeled quark ( 3 ) ] , and one defines the momentum variables corresponding to those diquark and spectator quark sub - systems ( the so - called jacobi momenta ) . if the individual quark momenta are @xmath53 ( @xmath54 ) , the jacobi momenta are @xmath55 , the relative momentum of the quarks in diquark ( 12 ) , and @xmath56 , the diquark center of mass momentum with with respect to quark ( 3 ) .
the center of mass momentum is @xmath57 .
the momentum states that define our basis to represent the wave function are the eigenvectors of the jacobi momenta @xmath58 and @xmath59 .
they are called @xmath60-type and @xmath61-type states , with mixed symmetry is anti - symmetric for the exchange of quarks 1 and 2 , while the jacobi momentum @xmath58 is symmetric for the same exchange .
the jacobi momenta @xmath59 and @xmath58 eigenvector basis states are therefore anti - symmetric and symmetric , respectively , under that exchange . for another particle exchange @xmath62 , with @xmath63 , those states are , however , states of mixed symmetry . ] .
following the traditional notation ( see e.g. ref .
@xcite ) , the labels @xmath61 and @xmath60 are used more generally , i.e. , for combinations and angular momentum projections of momentum states , and also for spin and isospin states , that are , respectively , anti - symmetric and symmetric under the exchange of quarks ( 12 ) .
the starting point for the construction of the flavor - spin - momentum - space wave function is to impose that it is symmetric under the exchange of any pair ( the color part , which is omitted , makes it anti - symmetric at the end , as required ) .
the second step is to write the non relativistic limit of the wave function in terms of @xmath60-type and @xmath61-type mixed - symmetric states , labeled @xmath64 and @xmath65 , that couple orbital states @xmath27 ( in principle in both @xmath59 and @xmath66 jacobi momenta ) with total three - quark spin @xmath29 states , and to multiply them with the adequate flavor states that make the function symmetric .
next , we assume a pointlike diquark . in this approximation , effectively , one has @xmath67 . with this suppression of the diquark internal p states ,
the orbital wave function is reduced to p - states in the momentum @xmath58 of the quark - diquark motion only .
additionally , the non relativistic wave function is calculated in the 3 body center of mass frame , where @xmath68 , and the diquark three momentum becomes @xmath69 .
then , the spin - orbital part of the non - relativistic wave function is , in our approximation , written as a function of @xmath70 only .
finally , one makes the relativistic generalization of the coupled spin - orbital states @xmath71 and @xmath72 .
the corresponding relativistic states , labeled respectively @xmath73 and @xmath74 , include a @xmath75 matrix , exhibiting the negative parity of the state explicitly .
all the details concerning the full non - relativistic wave function in the pointlike diquark limit , and its relativistic generalization , are presented in appendix [ apwf ] . to conclude this section , we write in the pointlike diquark approximation , the final expression for the covariant structure of the spin - flavor - orbital wave function of the @xmath1 .
it depends on the baryon four - momentum @xmath28 and on the diquark four momentum @xmath40 , and is given by _
s11(p , k ) = _ s11(p , k ) , [ eqpsis11 ] where @xmath76 and @xmath77 are the flavor states , and & & _ ( ) = -_5 n + & & _ ( ) = + _ 5 n .
+ & & in the last equations @xmath78 and @xmath79 . the four momentum @xmath80 can be interpreted as the diquark three momentum in the @xmath1 rest frame [ where @xmath81 and @xmath82 .
the spinors @xmath83 and @xmath84 have the same meaning as @xmath41 and @xmath85 , defined for the nucleon before @xcite , as in eq .
( [ equn ] ) , but are here associated with the @xmath1 baryon . the scalar wave function @xmath86 will be discussed later ( see sec . [ secwf ] ) .
here it suffices to say that this function carries all the information on the momentum distribution of the quark - diquark relative motion , it is purely phenomenological and normalized to one .
we make two more notes about eq .
( [ eqpsis11 ] ) : the wave function in our model does not contain the contribution of three - quark states with total spin @xmath30 , included in other works @xcite . additionally , the minus sign for the @xmath60-type spin - orbital in the wave function is needed to ensure orthogonality between the @xmath1 and the nucleon wave functions in the non relativistic limit @xcite .
we can write the transition current in relativistic impulse approximation @xcite as j^= 3 _ _ k _ s11(p_+,k ) j_i^_n(p_-,k ) , [ eqji1 ] where @xmath87 ( scalar diquark @xmath88 and vector diquark polarization @xmath89 ) and @xmath90 is the covariant integration element in the diquark on - mass - shell momentum @xmath40 ( mass @xmath91 and energy @xmath92 ) .
the factor 3 accounts for the contributions of all possible diquark pairs , since , due to the symmetry of the wave function , pairs ( 13 ) and ( 23 ) give the same contribution as pair ( 12 ) .
[ the magnitude of the electron charge @xmath93 was not included in the current for simplicity ] . in the previous equation , @xmath94 is the quark current j_i^= j_1 ( ^-)+ j_2 . to obtain the @xmath0 transition current
we take the wave functions ( [ eqpsin ] ) and ( [ eqpsis11 ] ) . to work the spin algebra one uses @xmath95 and @xmath96 , obtaining j_i & = & f_i+ + f_1- _ 3 + j_(i+2 )
& = & f_i+ - f_i- _ 3 .
the coefficients @xmath97 and @xmath98 follow the definitions in ref .
note that the result is a sum over the flavor of the anti - symmetric ( @xmath99 and @xmath100 ) and symmetric components ( @xmath101 and @xmath102 ) as done in refs .
@xcite for the su(3 ) case . for convenience one introduces also the notation ^= ^- .
using the definitions above one can write _
j_i^_n & = & \ { j_1
_ ^_s^0 + j_2 _ _ s^0 } + & - & \ { j_3
_ ^_s^1 + j_4 _ _ s^1 } , + & & [ eqsum ] where @xmath103 . for the vector diquark contributions ( terms in @xmath104 ) the sum in the diquark polarization
@xmath105 is implicit .
the isovector components include a sum in the diquark polarizations @xmath105 vectors associated with the @xmath1 , @xmath106 , and the nucleon , @xmath107 .
those polarization vectors are functions of the @xmath1 mass ( @xmath51 ) and the nucleon ( @xmath43 ) mass , respectively ( see details in ref .
@xcite where this basis of states is explained and built ) . by adding the diquark polarizations
, one has @xcite ^ & & _ _ d _ p_+^(_d ) _
p_-^ ( _ d ) + & = & -(g^ - ) + & & - a ( p_ p_+)^ ( p_+- p_-)^ , + & & [ eqdelta ] where a= .
the decomposition ( [ eqsum ] ) reduces the determination of the current ( [ eqji1 ] ) to the calculation of a few current elements .
the details are presented in appendix [ apcurrent ] .
the final result is j^&= & ( 3j_1 + j_3 ) i_0 |u_s ^_5 u + & - & ( 3j_2- j_4 ) i_0 |u_s _ 5 u , [ eqjsp ] where _ 0 ( q^2)= _ k n ( _ 0 k ) _ s11 _ n. [ eqint0 ] the integral @xmath108 is covariant and includes the dependence of the form factors on the initial and final state scalar wave functions .
we call @xmath108 the overlap integral . in this appendix
we calculate the electromagnetic transition current defined by eq .
( [ eqji1 ] ) , using the nucleon and @xmath1 wave functions given by eqs .
( [ eqpsin ] ) and ( [ eqpsis11 ] ) .
the @xmath1 wave function is given by eq .
( [ eqpsirel2 ] ) with the spin states defined by ( [ eqxr])-([eqxl ] ) . from the relations @xmath261 and @xmath262
, one can write & & _ ( ) = - n _ 5 + & & _ ( ) = n _ 5 .
+ & & in the previous equations u_s^= - |u_s ( ^- ) _ 5 . to reduce the transition current to the standard form one uses the properties of the nucleon and @xmath1 spin states @xmath84 , @xmath83
, @xmath263 and @xmath41 : & & p_- u(p_-)=m u(p_- ) + & & p_- u^(p_-)= m u^(p_- ) + & & p_+
u_s(p_+)=m_s u_s(p_+ ) + & & p_+ u_s^(p_+)= m_s u_s^(p_+ ) . also & & ( p_+)_u_s^=0 + & & ( p_-)_u^=0 . in the following we consider the symmetries in the @xmath40 integration .
the evaluation of the transition current requires the determination of the integrals _ ^= _ k n ( _ ^ k ) _ s11 _
n , where @xmath264 . it is easy to prove that _
= 0 , for any value of @xmath9 .
the demonstration is trivial in the @xmath1 rest frame , since the product of wave functions can be written as a function of @xmath265 and @xmath140 .
then @xmath266 , because the integrand function is odd in the integration variables @xmath267
. then only @xmath108 survives the @xmath40 integration for a given @xmath9 .
the case @xmath142 will be discussed in appendix [ apint ] .
the important point here , is that in the final state rest frame we have to keep in the wave function only the terms in @xmath268 .
considering the expression for the spin states and by performing the integral for the current , one obtains , for arbitrary ( initial and final ) spin projections : & & _
^_s^0= n ( _ 0 k ) |u_s ^ _ 5 u [ eqphi1a ] + & & _ _
s^0= - n ( _ 0 k ) |u_s _ 5 u [ eqphi1b ] + & & _
^_s^1= n ( _ 0 k ) _ [ eqphi2a ] + & & _ _ s^1= - n ( _ 0 k ) _ .
+ & & [ eqphi2b ] in the previous equations @xmath269 is given by eq .
( [ eqdelta ] ) . for the terms
@xmath270 one can use the generalized gordon identity : i ^ q_= p_+ ^+ ^ p_- - ( p_+ + p_-)^. the following relations holds when multiplied by @xmath269 : & & u_s^^_5 u^= - |u_s ^^ ^_5 u + & & u_s^ _ 5 u^= |u_s ^ ^_5 u. considering the results : & & _ = ^_5 + &
= -_5 , one obtains & & u_s^^_5 u^= |u_s ^_5 u + & & u_s^ _ 5 u^= - |u_s _ 5 u. using the formulas of the previous section one can write the result of the integration in @xmath40 for ( [ eqphi1a])-([eqphi2b ] ) including also @xmath271 . in the integration
the relations with @xmath272 are replaced by @xmath108 , defined by eq .
( [ eqint0 ] ) . then & &
_ k _ s11 _ n = i_0 |u_s ^ _ 5 u + & & _ k _ s11 _ n = - i_0 |u_s _ 5 u + & & _ k _ s11_n= -_0 |u_s ^_5 u + & & _ k _ s11_n= _ 0 |u_s _ 5 u. replacing the previous results in the expression for the current , we obtain j^&= & + ( 3 j_1 + j_3 ) i_0 |u_s ^_5 u + & & - ( 3 j_2- j_4 ) i_0 |u_s _ 5 u.
the previous current defines the electromagnetic transition form factors given by eqs .
( [ eqf1])-([eqf2 ] ) .
the sign of the normalization constant @xmath18 with magnitude @xmath273 has to be fixed by the experimental data .
as the data for @xmath5 is negative near @xmath142 , we choose n= - .
the transition current can be written ( suppressing the charge factor @xmath93 ) as @xcite : j^= |u_s _ 5 u , [ eqjs ] where @xmath109 defines the transition form factors .
one should note that there are alternative but equivalent conventions for the two form factors @xcite . from the eqs .
( [ eqjsp ] ) and ( [ eqjs ] ) , we conclude that & & f_1^(q^2)= ( 3j_1 + j_3 ) i_0 [ eqf1 ] + & & f_2^(q^2)= -(3j_2- j_4 ) _ 0 [ eqf2 ] the experimental data is usually presented in terms of the helicity amplitudes in the final state ( excited resonance ) rest frame .
the helicity amplitudes are defined from the projection of the current on the photon polarization states , @xmath110 and nucleon and resonance spin projections ( in the resonance frame ) . for a resonance @xmath111 with spin 1/2 , there are two independent amplitudes : & & a_1/2(q^2)= k n^,+| _ + j |n , - , [ eqa12 ] + & & s_1/2(q^2)= k n^,+| _ 0 j |n , + . [ eqs12 ] considering @xmath112 , the multiplicative constant is = , with @xmath113 is the magnitude of the electron charge with @xmath114 , and @xmath115 . the variable @xmath116 is the photon three momentum in the excitation , in the @xmath1 rest frame , |*q*|= , [ eqq2 ] where @xmath117 , with @xmath118 . the helicity amplitudes can be represented in terms of the form factors @xcite : & & a_1/2= -2b [ eqa12x ] + & & s_1/2= b(m_s+m ) , [ eqs12x ] with @xmath119 and b= e . from * equations ( [ eqf1])-([eqf2 ] )
* one can make predictions for the form factors and compare the obtained results with the experimental data .
our model is now completely defined , for the baryons and for the current , except for the scalar function @xmath120 which is part of the wave function . in the spectator quark model
the wave functions depend on @xmath121 only , as the baryon and diquark are taken on - mass - shell .
that dependence can be re - written in terms of the adimensional variable _ _
b= , where @xmath38 is the baryon mass [ nucleon or @xmath1 ] and @xmath91 the diquark mass . within the s - wave approach ,
the scalar function in the nucleon wave function is given by @xcite : _
n(p , k)= , [ eqpsins ] where @xmath123 is the normalization constant and @xmath124 are adimensional parameters which measure the momentum scale of the quark - diquark interaction . as @xmath125
, @xmath126 defines the scale for the short distance range and @xmath127 the long distance range .
as the @xmath1 corresponds to a spin 1/2 quark core with the same content of the nucleon , it is reasonable to consider a form for the scalar wave function similar to the one taken for the nucleon _
s11(p , k)= , where @xmath128 is the normalization constant and @xmath129 a new range parameter . to start with
, the same parameter @xmath126 ( @xmath130 ) can be used for the two cases , the @xmath1 and the nucleon , if one assumes that the two baryons differ only in the structure at large distances .
moreover , on the other hand , and inspired by the relativistic quark models with an harmonic oscillator confinement @xcite , we consider that the nucleon and the @xmath1 may as well have the same momentum distributions at large distances as expected for excitations of the same state and we will thus also take @xmath131 .
then , the nucleon and the @xmath1 are described by the same scalar wave functions in their rest frame .
we may say that this assumption is justified since in the chiral limit the nucleon and the @xmath1 will have the same mass and become two different parity states of the same particle .
the difference between the momentum distributions in the nucleon and the @xmath1 come from the difference in the orbital angular momentum in their total wave functions . in the non relativistic limit
, this angular dependence corresponds to @xmath132 , a constant , for the nucleon , and @xmath133 , the p - state , for the @xmath1 .
an alternative parametrization for the scalar wave functions would be to force the fit of @xmath129 to the data and to introduce a new parameter in our model .
since we will see that our parameter - free description was surprisingly successful , we did not face a good reason to assume different scalar functions for the nucleon and the @xmath1 , and our results can be considered true predictions , once the nucleon is correctly described .
the transition form factors depend on the orbital wave functions through their overlap integral @xmath108 , defined by eq .
( [ eqint0 ] ) .
terms that include integrations in @xmath134 or @xmath135 vanish because of the symmetries of the scalar wave function ( as function of @xmath136 ) , as shown in appendix [ apcurrent ] , and the integral @xmath108 carries the signature of the angular momentum dependence of the nucleon and @xmath1 wave functions .
the overlap integral is covariant and it can be evaluated in any frame .
one of the simplest calculations is the one that proceeds in the @xmath1 ( final state ) rest frame , ( see appendix [ apint ] ) , where _ 0 ( q^2)= _ k _ s11 ( p_+ k)_n .
( p_- k ) [ eqint00 ] in the @xmath1 rest frame all the angular dependence of the wave functions is contained in @xmath137 , given by eq .
( [ eqpsins ] ) .
this dependence is expressed by p_- k= e e_d + |*q*| k_z , where @xmath116 is the photon three momentum in the @xmath1 rest frame , as defined in eq .
( [ eqq2 ] ) , @xmath138 is the nucleon energy , and @xmath92 the diquark energy .
the numerical value of @xmath139 depends therefore on the existing symmetries in the variable @xmath140 .
the properties of the overlap integral @xmath141 are discussed in appendix [ apint ] . in particular for small @xmath116
, one has _ 0 ( q^2 ) |*q*| .
[ eqint2 ] this result has important consequences and allows us to define the domain of validity of our model .
in what follows we will label @xmath116 in the @xmath142 limit by @xmath143 .
as the photon energy @xmath144 equals @xmath143 at @xmath142 , one has then @xmath145 , and according to eq .
( [ eqint2 ] ) , @xmath146 , if @xmath147 .
the relation @xmath148 is then equivalent to the orthogonality condition between the @xmath1 and the nucleon wave functions . however , if @xmath149 , the integral @xmath139 will be proportional to @xmath150 .
consequently , @xmath151 , and the @xmath1 and the nucleon wave functions are not exactly orthogonal .
this result has a dramatic implication since the nucleon and the @xmath1 should in fact be orthogonal .
this is an artifact of the construction of the wave function from its non relativistic behavior , and of having imposed to it a covariant form with multiplicative scalar functions that were not derived from an ab - initio calculation .
a simple picture of what happens is that the nucleon orbital ( s - state ) wave function ( defined unambiguously only in the rest frame of the nucleon ) is distorted by the boost to the rest frame of the @xmath1 , and therefore is not orthogonal to the @xmath1 orbital ( p - state ) wave function .
this implies that the overlap integral @xmath139 does not vanish .
still , if the masses of the initial and final state are equal , @xmath142 implies @xmath152 , as mentioned , and there is no problem since there is no boost .
the fact that the integral ( [ eqint00 ] ) is not zero for @xmath142 is therefore a limitation of our model when the initial final and initial states have different masses .
however , the relation ( [ eqint2 ] ) can be used to establish the range of application of the model .
the non orthogonality between the model wave functions of the initial and final state decreases as @xmath51 approaches @xmath43 .
if the mass difference is negligible there is orthogonality to a certain extent .
then , @xmath150 is a parameter that measures the quality of our model approximations to the wave function .
as @xmath143 corresponds to the photon energy at @xmath142 ( at the photon point the energy equals the three momentum ) , it defines the natural momentum scale of the reaction . in the regime @xmath153
, one has @xmath154 , meaning that the nucleon and the @xmath1 states are _ almost _ orthogonal . as for the physical case @xmath155 gev , @xmath156 for @xmath157 gev@xmath4 , and therefore , one can say that @xmath158 gev@xmath4 establishes the threshold for the application of our model .
summarizing , the present model has limitations in its applications at low @xmath9 , in particular near @xmath142 , but can be used in the high @xmath9 regime , for @xmath159 gev@xmath4 .
with the model for the baryons and for the current depicted in the previous sections we have calculates the @xmath0 transition form factors given by eqs .
( [ eqf1])-([eqf2 ] ) and the helicity amplitudes given by eqs .
( [ eqa12x])-([eqs12x ] ) .
no parameters of our model were adjusted to these observables .
we calculated only the positive isospin case ( @xmath160 ) , corresponding to the excitation reaction from the proton , where the data at finite @xmath9 for the helicity amplitudes is available @xcite .
we did not consider the neutron case ( @xmath161 ) , since there is data only for @xmath142 , and our model is valid only for @xmath3 gev@xmath4 .
the data from desy @xcite and from jefferson lab @xcite are restricted only to the @xmath7 amplitude , assuming that the amplitude @xmath8 was negligible .
that assumption was contradicted by the recent clas @xcite and maid @xcite analysis . in the following we use ref .
@xcite where @xmath7 and @xmath8 were determined simultaneously .
we will also compare our results with the dalton @xmath162 data @xcite , for @xmath7 at high @xmath9 ( @xmath163 gev@xmath4 ) , which is determined under the assumption that @xmath164 [ for large @xmath9 the approximation @xmath164 is better justified due to the falloff of @xmath8 at high @xmath9 ] .
the results for the @xmath0 form factors are shown in fig .
[ figff ] .
the data for @xmath5 and @xmath6 was obtained by inverting the relations ( [ eqa12x])-([eqs12x ] ) . in the figure
we represent also the clas data from ref .
@xcite and the maid analysis of ref .
@xcite , as well as the results from @xcite ( where @xmath164 ) .
one can see that our model describes well the @xmath5 data for for @xmath166 gev@xmath4 , in particular that the model works in its regime of application @xmath167 gev@xmath4 . as for @xmath6 ,
our model fails completely when compared with the experimental data .
we predict positive values for @xmath6 , contrarily to the data .
also , the magnitude differs strikingly from the data : the clas data is very close to zero for @xmath168 gev@xmath4 , in the region where our model gives a strong positive contribution .
this disagreement can be interpreted in two ways .
one possibility is that our model is limited because the internal diquark p - states were neglected in our model , and we will have to confirm their effects in a future work .
other possible interpretation is that , for @xmath6 the valence quark effects , the only ones considered in our model are strongly canceled by the effect of the meson cloud polarization , not included in our model . if this last interpretation is correct , one has to conclude that meson cloud effects are very significant , even in the region @xmath168 gev@xmath4 .
this finding is at odds with what was observed till now in similar systems , like the nucleon @xcite and the roper @xcite .
nevertheless , the @xmath37 quadrupole form factors reveal a strong contribution of strong pion cloud in the region 26 gev@xmath4 @xcite .
to test the last interpretation , we compared our valence quark model predictions with the calculations from a different framework , the ebac dynamical coupled - channel model based in sato - lee model @xcite . in the ebac analysis
@xcite the effects of the meson cloud dressing are subtracted , and the pure quark core contributions calculated from the model .
the ebac data can then be directly compared with our results , as shown also in fig .
[ figff ] ( upper triangles ) . as for @xmath5 ,
the ebac results overestimates ( in absolute value ) the experimental data ( clas and maid ) but seems to approach the data for @xmath169 gev@xmath4 . as for @xmath6 , the ebac results are surprisingly consistent with our own predictions , both in sign and magnitude for @xmath170 gev@xmath4 , near the threshold where our model starts to be applicable , @xmath158 gev@xmath4 .
future ebac determination of the quark core contributions , already planed for higher @xmath9 @xcite , will be very important to test our predictions and interpretations . an independent confirmation of the large contribution of the valence quarks for @xmath6 may also come from lattice qcd at high @xmath9 .
we note that our covariant spectator quark model was already successful in the description of lattice qcd simulations for the nucleon , roper @xcite and @xmath39 systems @xcite .
to summarize , our results for the form factor @xmath5 are consistent with the data for @xmath168 gev@xmath4 , in the domain of validity of our model . as for @xmath6 ,
our model supports the idea that meson cloud contributions are comparable with the valence quark contributions , which is also validated by the ebac studies of the @xmath1 system @xcite . using our results for the form factors
we have also calculated the helicity amplitudes in the @xmath1 rest frame , corresponding to the transformations ( [ eqa12x])-([eqs12x ] ) .
some comments are necessary before showing the results .
the first note is that our quark model should be compared with the data only in the region @xmath167 gev@xmath4 .
a second important note is that in our model @xmath171 because of the violation of the orthogonality condition between the nucleon and the @xmath1 wave functions . therefore the amplitude @xmath8 in our model is singular for @xmath142 , in opposition to the finite result expected from the data .
this effect was already reported in the relativistic quark model of ref .
@xcite , where the quark current was modified to restore gauge - invariance . with those limitations in mind
, we represent in fig .
[ figamp ] the amplitudes corresponding to the form factors in fig .
[ figff ] , by the solid line .
the dramatic deviation from the data is not surprising , since our model disagrees already with the @xmath6 data .
the disagreement is evident for @xmath7 where our large @xmath6 contribution spoils a excellent result that would be obtained if the @xmath6 could be neglected .
the results obtained in that scenario ( @xmath172 ) are represented by the dashed line . in that case
the agreement of our model with the data is excellent for @xmath168 gev@xmath4 for both amplitudes .
it is moreover interesting to note that the model ( solid line ) agrees well with the ebac results for @xmath8 .
that comes from the @xmath5 suppression in the @xmath8 amplitude by the factor @xmath173 [ see eq .
( [ eqs12x ] ) ] .
we conclude that the helicity amplitudes are not the best representation to test our model , since those amplitudes amplify the limitations of our model , like @xmath171 or the large magnitude of @xmath6 . combining our results for @xmath174with the assumption that @xmath6 is negligible for @xmath168 gev@xmath4 , as a consequence of the meson cloud effect , which is substantiated by the data and the ebac results
, one can achieve a very good description of the helicity amplitudes data .
the study of the @xmath0 electromagnetic structure was in the past based almost only on the representation of the helicity amplitudes [ in the @xmath1 rest frame ] .
then , the comparison with other works has to be done in this representation . from the previous section we know that the data corresponds to positive values for @xmath7 and negative values for @xmath8
. we will start by discussing the constituent quark models .
different quark model predictions , including non - relativistic @xcite and relativistic @xcite formulations , agree qualitatively with the data for @xmath7 . in particular , in ref .
@xcite , calculations based on the light - front formalism give an excellent description of the @xmath7 data for @xmath168 gev@xmath4 @xcite . also qcd sum rules @xcite are consistent with the @xmath7 data for @xmath175 gev@xmath4 . in a non relativistic model with harmonic - oscillator confinement potential
the relative sign between @xmath176 and @xmath177 is positive , and determined by the relative sign between the @xmath178 and the @xmath179 coupling constants @xcite . for non relativistic models
we should expect then positive values for @xmath8 at low @xmath9 .
this feature is also shared by light - front and relativistic quark models @xcite although sometimes negative results are obtained for @xmath168 gev@xmath4 @xcite .
still , in general one has the same sign for @xmath176 and @xmath177 .
exceptions to this feature are obtained by the qcd sum rules @xcite and our model .
qcd sum rules predict the sign but underestimate in absolute value the result for @xmath8 .
it has also been suggested that the state @xmath1 may have a strong contribution from quark - antiquark states , or even been dynamically generated by the meson - baryon interaction .
an and zou @xcite considered a quark model with explicit quark - antiquark dressing , and concluded that those effects can be of the order of 20% for low @xmath9 . in the overall ,
the signs and magnitudes are consistent with the data . in ref .
@xcite the meson cloud dressing is calculated within the cloudy bag model . in that case
the quark core is dominant at low @xmath9 and is consistent with the data for @xmath7 ( with @xmath180 of meson cloud ) , although the @xmath8 data is overestimated . for @xmath181
the model predictions are suppressed compared with the data indicating that short range behavior is not well simulated by the bag model @xcite .
the helicity amplitudes were also determined using a chiral unitary approach @xcite .
the authors conclude that the @xmath1 seems to be largely dynamically generated from the interaction of mesons and baryons but also that a genuine quark component is necessary particularly at high @xmath9 @xcite .
qualitatively , the meson dressing explains roughly 50 - 60% of the @xmath7 amplitude .
also , the calculations of the ebac group sugest the importance of the meson dressing at low @xmath9 , although there is a dominance of the quark core @xcite
. one may conclude that in general , from quark models and hadronic models with meson dressing , the meson cloud can be important , but genuine valence quark contributions are equally necessary to explain the data .
the study of the asymptotic dependence of the @xmath0 transition form factors attracts some attention , because pqcd predicts a very slow falloff for @xmath7 @xcite and also because precise experimental data have been extracted at high @xmath9 , in particular for @xmath182 gev@xmath4 @xcite and @xmath183 , and @xmath184 gev@xmath4 @xcite . the estimate from pqcd @xcite for large @xmath9 is q^3
a_1/2(q^2 ) = e , [ eqa12assymp ] where @xmath185 gev@xmath186 , in the more optimistic estimate ( upper limit ) @xcite . as for the form factors ,
one expects @xmath187 and @xmath188 , apart @xmath189 corrections .
then , for large @xmath9 , one has @xmath190 , and according to eq .
( [ eqa12x ] ) q^4 f_1^(q^2 ) = - .
[ eqf1assymp ] the asymptotic results from eqs .
( [ eqa12assymp])-([eqf1assymp ] ) are presented in fig .
[ figassymp1 ] for both @xmath5 and @xmath7 . in the last case
we show the result obtained by making @xmath191 , as discussed earlier . in the figure
it is clear that the pqcd estimation underestimates the data and our model for high @xmath9 .
the asymptotic behavior of the form factors can be better understood scaling the functions by a convenient power of @xmath9 to check if the results converge to a constant , apart the logarithm corrections . in this case
we should take the functions @xmath192 and @xmath193 .
the results for @xmath5 are presented in fig .
[ figf1as ] .
the representation of @xmath7 would be equivalent . in the figure it is clear that pqcd estimation fails the description of the data by a factor larger than 2 .
the same was reported in ref .
@xcite for @xmath7 .
the pqcd prediction differs then from the spectator quark model . at @xmath194
gev@xmath4 the ratio is 2.3 .
also in the figure it is clear a non constant slope for both pqcd and the spectator quark model results in the region shown , indicating corrections for the @xmath195 behavior . in the pqcd case , the slope is a consequence of the @xmath9-dependent factor of the r.h.s . of eq .
( [ eqf1assymp ] ) , which became a constant only for @xmath196 gev@xmath4 [ see the slow variation of the dotted line in fig . [ figf1as ] ] . as for the spectator quark model
, the logarithm dependence at larger @xmath9 , comes from the parametrization of the nucleon wave functions by eq .
( [ eqpsins ] ) , as product of two monopole factors in the variable @xmath121 .
that choice was considered in the applications to the nucleon electromagnetic structure @xcite in order to reproduce the expected pqcd behavior for the nucleon form factors ( dirac @xmath197 and pauli @xmath198 ) , but also contains logarithm corrections .
see appendix g from ref .
@xcite for details . for @xmath199 gev@xmath4
one can represent the spectator quark model form factor @xmath5 as f_1^(q^2 ) - , where @xmath200 gev@xmath4 . in conclusion
, our model reveals a scaling with the same power as pqcd for @xmath201 gev@xmath4 , apart logarithm corrections .
the scaling due to pqcd , if it is confirmed , will be revealed only for much larger @xmath9 values than in our model .
in this work we applied the covariant spectator quark model to the @xmath1 system .
we considered the simplest case where @xmath1 is made of states with core spin 1/2 , and we neglected the effect of the core spin 3/2 state , as in ref .
we took also the diquark as a pointlike particle ( no internal p - states ) .
these approximations have the advantage of reducing the degrees of freedom of our model to the minimum , and to allow us to perform calculations with no adjustable parameters , since all parameters ( in the quark current and wave functions ) were already fixed by the study of the nucleon system @xcite .
our results in this paper are then true predictions .
the extension of this work to include spin 3/2 cores ( which are also part of the nucleon d - states ) is in progress @xcite .
once our model is calibrated for the spin 3/2 component we will also be able of making predictions for the @xmath24 form factors .
our model takes contributions for the form factors from the valence quarks alone , and neglects possible meson cloud effects ( in principle dominated by @xmath202 and @xmath203 clouds ) .
this approximation involving the meson cloud suppression simplifies the construction of the @xmath1 wave function ( as a three - quark system ) .
another approximation , intrinsic to the relativistic generalization that we make for the wave function , is that the @xmath1 state is exactly orthogonal to the nucleon state only in the case of equal masses for the two baryons @xmath204 .
however , as the orthogonal condition can be written in powers of @xmath205 , one can show that our results are accurate for @xmath158 gev@xmath4 . in that region
meson cloud effects are expected to be negligible , the reason why one can make predictions for the form factors , which otherwise would contain , apart from valence quark effects , important meson cloud contributions . for the @xmath5 form factor
our results are in excellent agreement with the data in the domain of applicability of our model .
this is remarkable since there is no parameter adjustment .
our results for @xmath5 are also close to the ebac analysis of the quark core effects , although the ebac results are restricted to the region @xmath206 gev@xmath4 . as for the @xmath6 form factor
, our predictions fail completely to describe the experimental data in their sign and magnitude , which is consistent with @xmath207 for @xmath168 gev@xmath4 .
our results are however in good agreement with the estimations of the ebac group of the quark core contribution to the @xmath6 form factor near @xmath208 gev@xmath4 .
these two last points suggest that our failure in describing @xmath6 is caused by a large negative contribution from the meson cloud which cancels almost exactly the valence quark contribution .
although meson cloud contributions are expected to decrease with increasing @xmath9 , there are some exceptions to that rule , as the observed for the @xmath37 quadrupole transition form factors @xcite , where pion cloud are in fact the dominant effect .
the other possible explanation for the failure of our model in the description of the @xmath6 , is the internal structure of the diquark which was not considered here . but
this explanation is excluded by the comparison of our results with the ebac result , which seems to indicate that the pointlike diquark approximation is apparently good , at least for @xmath5 .
a true test of the @xmath6 suppression can come from the extraction of the core contributions by ebac model for higher @xmath9 , planned for a near future @xcite , and which can confirm our results for the valence quark contributions .
that test will also be useful to assert and consolidate our @xmath5 results .
a third independent test can be the direct comparison with lattice qcd simulations , particularly for large pion masses ( say @xmath209 gev ) , a regime where quark - antiquark ( @xmath203 and @xmath202 cloud ) contributions are believed to be very small .
lattice qcd simulations are nowadays viable since they were performed previously for the @xmath37 and @xmath210 reactions @xcite . although the comparison of phenomenological model results , at the physical pion mass point , with lattice qcd can be problematic due to the necessity of extrapolating to the physical limit , that is not a problem for the spectator quark model : it is based on a vector meson dominance parametrization of the current , and therefore can be extended successfully to the lattice conditions , as was shown for the nucleon @xcite and roper @xcite reactions , for the @xmath37 transition @xcite and also for the baryon decuplet form factors @xcite .
in addition to the form factors , we calculated as well the helicity amplitudes @xmath7 and @xmath8 . as in our calculations the violation of the orthogonality condition between the initial and final states , gives @xmath211 , implying that the amplitude @xmath8 diverges for @xmath212 and the results for @xmath6 differ from the data , we conclude that helicity amplitudes are not the more convenient representation to test our model in particular , and quark models in general .
combining our results with the hypothesis that @xmath6 is negligible , because of the actual cancellation of valence quark contributions and meson cloud contributions , which is suggested by the successful comparison of the our results and the ebac quark core contribution , we obtain an excellent description of the helicity amplitudes data , @xmath7 and @xmath8 ( see dashed line in fig . [ figamp ] ) . as for @xmath7
the agreement is remarkable for @xmath175 gev@xmath4 , even before the region of validity of our model is reached . as for @xmath8 , although it is singular for @xmath142 , the model describes the data for @xmath213 gev@xmath4 . in summary ,
the @xmath0 reaction is very interesting from the constituent quark model perspective .
the possibility of the @xmath6 form factor to vanish at intermediate @xmath9 values , in contrast to what happens with all other known resonances , provides a unique challenge to theoretical models , in order to understand the role of the valence quarks , and their interplay with the meson cloud .
all effort from quarks models , dynamical coupled - channel reaction models , chiral effective models and lattice qcd , are welcome in attempts that have to be harmonized and supplemented together , in order to interpret the @xmath0 reaction data .
* acknowledgments : * the authors thank hiroyuki kamano for providing the ebac results from ref .
@xcite and viktor mokeev for helpful discussions .
the authors also thank franz gross for the invitation to visit the jefferson lab theory group .
g. r. was supported by the fundao para a cincia e a tecnologia under the grant no .
sfrh / bpd/26886/2006 .
this work is also supported partially by the european union ( hadronphysics2 project `` study of strongly interacting matter '' ) .
the work of the two authors was also financed by the fundao para a cincia e a tecnologia , under grant no .
ptdc / fis/113940/2009 , `` hadron structure with relativistic models '' .
the @xmath1 is defined as an excitation of the nucleon corresponding to the state @xmath52 . to represent the @xmath1 state in a constituent quark model framework we need to consider the momentum , spin , isospin of each quark and relate it with the @xmath1 proprieties .
we will follow the construction based on the @xmath214 as in refs .
@xcite .
we label the momentum of quark @xmath215 by @xmath53 . the center of mass momentum @xmath28
is then given by @xmath216 .
at this point we do not distinguish between non relativistic and relativistic kinematics .
the jacobi momentum are k_= ( k_1-k_2 ) , for the relative momentum of the quark in the quark - pair ( 12 ) , and k_= ( k_1+k_2 - 2k_3 ) , to measure the relative momentum between the diquark center of mass and the third quark .
note that @xmath59 is anti - symmetric in the exchange of quarks 1 and 2 , and that @xmath58 remains unchanged ( symmetric ) in the same exchange .
we note that in the non relativistic limit and in the baryon center of mass frame ( @xmath217 ) one has @xmath218 .
_ = , [ eqklambda ] where @xmath219 is the diquark three momentum .
we will use the @xmath61 and @xmath60 labels to characterize the baryon states , as it was defined in the main text , and as it is usual practice in the literature , e.g. in ref .
@xcite .
in the coupling of the spins of the 3 quarks there are different combinations for @xmath220 , where @xmath221 is the sum of the spins of quarks ( 12 ) and @xmath88 the spin of quark ( 3 ) .
the possible combinations are ^=(0 , ) , ^= ( 1 , ) , ^s= ( 1 , ) , respectively the @xmath61-type ( @xmath222 ) and the @xmath60-type ( @xmath223 ) states with mixed symmetry , and the state ( @xmath224 ) which is symmetric the change of any of the three quarks
. the spin states @xmath222 and @xmath223 are defined in terms of combinations of two spin states [ quark pair ( 12 ) ] , anti - symmetric and symmetric respectively , with the spin of the quark 3 .
this construction is similar to what was done for the nucleon @xcite .
one has for the spin projection @xmath225 : ^(+ ) & & | , + _ + & = & ( - ) + ^(+ ) & & | , + _ + & = & ( 2 - - ) .
identical expression hold for the isospin states .
for example , for the proton ( isospin projection @xmath226 ) , we write the isospin states as & & _ i^0 ( + ) ( ud - du)u + & & _ i^1 ( + ) ( 2uud - udu - duu ) , preserving the notation used in the nucleon wave function @xcite . here
the anti - symmetric state in the pair is identified by 0 and the symmetric state by 1 . for completeness , we represent also the state corresponding to isospin and spin projections @xmath227 : & & _ i^0 ( - ) ( ud - du)d + & & _ i^1 ( - ) -(2ddu - udd - dud ) , ^(- ) & & | , - _ + & = & ( - ) + ^(- ) & & | , - _ + & = & - ( 2 - - ) . later we will write the spin states in a covariant form . in the following
we suppress the isospin projection index from @xmath44 [ @xmath226 as in the proton , and @xmath227 as in the neutron ] .
with the previous notation we write the nucleon wave function for spin projection @xmath228 as _ n= \ { _ i^0 | , s_+ _
i^1 function for the quark momentum distribution .
see ref .
@xcite for details about the nucleon wave function .
the @xmath1 state has the same isospin structure of the nucleon . for the orbital angular momentum excitation of that state ,
we consider @xmath27 .
we have then the form _ s11= \ { _
i^0 x_- _ i^1 x _ } _ s11 , [ eqpsis11b ] with the states @xmath71 and @xmath72 , functions of @xmath228 , to be defined next .
the minus sign in the @xmath60-type term is included to ensure the orthogonality with the nucleon wave function ( [ eqpsin1 ] ) . by construction
, @xmath229 is anti - symmetric @xcite .
the normalization constant @xmath230 will be determined later .
here we take the @xmath1 state to be composed by states with core spin 1/2 only .
the same approximation is used in ref .
alternative models , like the classical karl - isgur model @xcite , where the baryons are confined quarks with color hyperfine interaction , describe @xmath1 as a mixture of states with core spin 1/2 and 3/2 @xcite .
the states @xmath71 and @xmath72 are combinations of the three quark system mixed - symmetric states , with total spin 1/2 ( @xmath222 or @xmath223 ) and orbital angular momentum @xmath27 .
those states are the direct product of orbital angular momentum @xmath27 with a spin @xmath231 state . considering the product for the projection @xmath88 ,
one has , for the mixed - symmetric states @xmath71 : x _ ( s ) = _ m 1 m ; , + | , s y_1,m(k _ ) | , s - m_. + the factor @xmath232 was introduced by convenience .
possible terms in @xmath233 , associated with p states in the diquark , are not considered here .
this corresponds to a pointlike approximation for the diquark ( @xmath67 ) .
note that the inclusion of structure in the diquark , which demands that a dependence of the scalar wave function in @xmath59 is included in general @xcite . here , the pointlike diquark is a first approximation .
as for the @xmath234 states , one has x _ ( s ) = _ m 1 m ; , + | , s y_1,m(k _ ) | , s - m_. + once again , we took a pointlike diquark [ no terms in @xmath233 ] .
the spherical harmonics allows us to write the angular momentum states as & & |k_| y_1,+1(k_)= k_+ + & & |k_| y_1,0(k_)= k_0 + & & |k_| y_1,-1(k_)= k_- where @xmath236 , and k_= ( k_x i k_y ) .
replacing the clebsch - gordan coefficients , and using the compact notation @xmath237 to represent @xmath45 , one obtains : & & x _ ( ) = n \ { k_0 | , _ - k _ | , _ } .
+ & & x _ ( ) = n \ { k_0 | , _ - k _ | , _ } , + & & [ eqx12 ] where @xmath238 .
these expressions reproduce the results from refs .
@xcite , in the pointlike diquark limit . in that case
only the normalization factor differs .
the normalization of @xmath229 is given by eq .
( [ eqpsis11b ] ) [ non relativistic form ] .
details associated with parity will be discussed later in the relativistic generalization .
the wave function ( [ eqpsis11b ] ) must be normalized in order to reproduce the @xmath1 charge : q_s11 & = & _ _ k _ s11^(|p , k ) ( 3 j_1 ) _
s11(|p , k ) + & = & ( 1 + _ 3 ) .
[ eqqs11 ] where @xmath239 represents the scalar component ( @xmath88 ) and the vectorial component ( polarizations @xmath89 ) of the intermediate diquark , and @xmath240 [ the momentum configuration correspondent to the rest frame ] .
the operator @xmath241 is the quark charge operator , where @xmath242 acts on the @xmath1 isospin states . in the following we use the notation introduced in the paper with calculations for the nucleon @xcite .
we project the states into isospin components , for the case @xmath142 , according to & & j_1 ( _ i^0 ) ^j_1 _ i^0 = + _ 3 + & & j_3 ( _ i^1 ) ^j_1 _ i^1 = - _ 3 . then considering ( [ eqx12 ] ) ,
one can write q_s11= ^2 _ k |_s11(|p , k)|^2 . from eqs .
( [ eqx12 ] ) , and working the spin algebra , for , one concludes that x_^(s ) x_(s ) & = & 1 + x_^(s ) x_(s ) & = & 1 . then q_s11&= & n^2 ( j_1 + j_3 ) _ k |_s11(|p , k)|^2 + & = & ( 1 + _ 3 ) n^2 _ k |_s11(|p , k)|^2 , because @xmath243 . choosing _ k |_s11(|p , k)|^2=1 , and one reproduces the @xmath1 charge ( [ eqqs11 ] ) , if we set @xmath244 . the relativistic generalization of @xmath58 is the diquark three momentum in the rest frame @xmath80 :
k _ k= k - p , where @xmath28 is the @xmath1 momentum .
the factor between @xmath58 and @xmath40 from eq .
( [ eqklambda ] ) was dropped .
that factor is included into the normalization of the states .
as @xmath245 , where @xmath246 is the quark three momentum in the rest frame , one has |k_| .
the diquark momentum components can also be defined in terms of the diquark polarization vectors : & & k_0 - k _ p ( 0 ) + & & k_+ - k _ p ( + ) + & & k_- - k _ p ( - ) . [ eqke ] in the following
we will use @xmath247 and @xmath248 for , respectively , @xmath249 and @xmath250 . to obtain the relativistic generalization of eq .
( [ eqpsis11b ] ) , one has to write the relativistic generalization of the spin states states @xmath251 .
we use the the covariant generalizations , as in the applications to the nucleon system @xcite : & & | , s _
^s u(p , s ) [ eqphir ] + & & | , s _
-(_p^)_u^(p , s ) , [ eqphil ] where u^(p , s)= _ 5 ( ^- ) u(p , s ) .
[ equa ] in the previous equations @xmath252 is the scalar diquark polarization @xmath253 and @xmath42 the spin 1 polarization vector in the fixed - axis polarization base @xcite . as @xmath252 is a scalar it can replaced by 1 in the wave functions of the nucleon and @xmath1 . the expressions for @xmath71 and @xmath72 from eqs .
( [ eqx12 ] ) can now be written in a relativistic form using eqs .
( [ eqke ] ) , ( [ eqphir ] ) and ( [ eqphil ] ) .
the states @xmath71 and @xmath72 are then functions of @xmath28 , @xmath40 ( or @xmath28 and @xmath80 ) and @xmath88 , but the momentum dependence will be suppressed in our the notation . to avoid the dependence of the spin polarization in eqs .
( [ eqx12 ] ) on the normalization factor , in the relativistic generalization we replace the factor @xmath254 by @xmath255 , obtaining a unique expression for both polarizations .
the final expression is then x _ ( ) & = & n [ eqxr ] + x _ ( ) & = & n , [ eqxl ] where we include the sub - index @xmath256 to label the @xmath1 states . in the previous equations we have replaced the non relativistic constant @xmath257 by a new constant such that @xmath258 .
the absolute value of @xmath18 will be fixed by the comparison with the experimental data and is discussed later .
the final expression for the covariant @xmath1 wave function , with respect to spin flavor , orbital angular momentum and parity , is then _
s11(p , k ) = _ 5 _ s11(p , k ) .
[ eqpsirel1 ] the operator @xmath75 was introduced to represent the parity of the state .
the scalar wave functions were discussed in the main text ( see sect . [ secwf ] ) .
equation ( [ eqpsirel1 ] ) reproduces also the @xmath1 charge . with the form ( [ eqpsirel1 ] ) , one has p _
-m_s _ s11 .
[ eqdirac ] this relation ( with the minus sign ) is a consequence of the introduction of the operator @xmath75 required by parity .
note that the @xmath1 dirac equation ( [ eqdirac ] ) differs from the equations corresponding to the previous applications of the spectator quark model @xcite ( nucleon , @xmath39 , and roper ) . in the following
we will use _ s11(p , k ) = _ s11(p , k ) , [ eqpsirel2 ] where @xmath259 and @xmath260 .
in this appendix we consider the integral of eq .
( [ eqint0 ] ) : _ 0
= _ k n(_0 k ) _ s11(p_+,k ) _ n(p_-,k ) .
[ eqint0a ] first , we derive an analytical expression for @xmath108 , next we explore the limit cases . consider the expression for the overlap integral ( [ eqint0a ] ) , in the @xmath1 rest frame _ 0 = _ k _ s11(p_+,k ) _
n(p_-,k ) , [ eqint0b ] where we used @xmath274 and eq . ( [ eqn ] ) . in the same frame one has @xmath275 , @xmath276 and , with @xmath277 and & & e=. in this case @xmath120 is independent of the azimuthal angle and we can write , using @xmath278 : _ 0= _ 0^+ _ s11(p_+ k ) i_z , [ eqint02 ] where i_z = _
-1 ^ 1 dz . in the previous equation
we use the simplified notation for the arguments of the wave functions , since they can be represented as a scalar function of @xmath279 : p_+ k= m_s e_d , p_- k= e e_d + k_z |*q*| . the separation of the function that depends on @xmath280 as @xmath281 from the ones depending only of @xmath40 , as @xmath120 is possible because in the @xmath1 rest frame @xmath282 is angle independent . as for @xmath281
, it is represented by the simple analytical form ( [ eqpsins ] ) . in these conditions
one can evaluate @xmath283 analytically using simple integration techniques .
the result is i_z & = & ( ) ^2 + & & [ eqiz ] where & & |_i= ( _ i -2 ) + 2 + & & g_i(k , |*q*|)= | | .
the expression obtained for @xmath283 from eq .
( [ eqiz ] ) does not help us to explore the limit @xmath212 . to have a clearer idea of the @xmath9 or @xmath116 dependence one considers the case @xmath284 . in that limit one case use
| |= 2 x + x^3 + o(x^5 ) , to simplify @xmath283 .
using the previous equation with x= , one can conclude that i_z= - ( ) |*q*| . for the equal mass case ( @xmath204 ) , where |*q*|= q , using @xmath285 , one can conclude that _ 0 ( q^2 ) q , implying that @xmath286 as @xmath212 .
this dependence is atypical and unexpected .
recall that the nucleon to roper form factors vanish for @xmath212 with the power @xmath9 ( in that specific case , independently of the mass difference ) .
result implies that the @xmath287 amplitude diverges for @xmath288 . as that amplitude scales with @xmath289 , for @xmath212 , if @xmath290 , the amplitude diverges for @xmath212 . for the form factors @xmath5 and @xmath6 there is no divergence for @xmath212 .
e. pace , g. salme and s. simula , few body syst .
* 10 * , 407 ( 1999 ) [ arxiv : nucl - th/9810053 ] ; e. pace , g. salme , f. cardarelli and s. simula , nucl .
a * 666 * , 33 ( 2000 ) [ arxiv : nucl - th/9909025 ] .
f. gross , g. ramalho and m. t. pea , phys .
c * 77 * , 035203 ( 2008 ) .
g. ramalho and m. t. pea , j. phys .
g * 36 * , 115011 ( 2009 ) [ arxiv:0812.0187 [ hep - ph ] ] .
f. gross , g. ramalho and k. tsushima , phys .
b * 690 * , 183 ( 2010 ) [ arxiv:0910.2171 [ hep - ph ] ] .
g. ramalho and m. t. pea , j. phys .
g * 36 * , 085004 ( 2009 ) [ arxiv:0807.2922 [ hep - ph ] ] g. ramalho , m. t. pea and f. gross , phys .
b * 678 * , 355 ( 2009 ) [ arxiv:0902.4212 [ hep - ph ] ] ; g. ramalho , m. t. pea and f. gross , phys . rev .
d * 81 * , 113011 ( 2010 ) [ arxiv:1002.4170 [ hep - ph ] ] . c. alexandrou , g. koutsou , h. neff , j. w. negele , w. schroers and a. tsapalis , phys .
d * 77 * , 085012 ( 2008 ) [ arxiv:0710.4621 [ hep - lat ] ] .
h. w. lin , s. d. cohen , r. g. edwards and d. g. richards , phys .
d * 78 * , 114508 ( 2008 ) [ arxiv:0803.3020 [ hep - lat ] ] . | a relativistic constituent quark model is applied to the @xmath0 transition .
the @xmath1 wave function is determined by extending the covariant spectator quark model , previously developed for the nucleon , to the @xmath2 resonance .
the model allows us to calculate the valence quark contributions to the @xmath0 transition form factors . because of the nucleon and @xmath1 structure the model is valid only for @xmath3 gev@xmath4 .
the results are compared with the experimental data for the electromagnetic form factors @xmath5 and @xmath6 and the helicity amplitudes @xmath7 and @xmath8 , at high @xmath9 . |
fibrous media display different degrees of meso - scale variability from ideal fibre paths .
this can be due to the manufacturing of the reinforcement , handling and preparing the moulding step .
resin flow inside porous media is influenced by fibres spatial variability and heterogeneity , and neglecting this causes errors in process analysis and uncertainty in measurement .
thus a reliable model of fluid flow in heterogeneous media must include multiscale phenomena and capture the multiscale nature of fluid transport behaviour , at microscale , mesoscale and macroscale . in the sense the dominant processes and governing equations may vary with scales . therefore , extending from microscale level to a mesoscale one needs upscaling that allows the essence of physical processes at one level to be summarized at the larger level .
however , a detailed understaning of upscaling process from microscale to mesocale has not been completed .
mesoscopic and macroscopic properties of fibrous media such as porosity , fibre size distribution and permeability can be characterized through lab - scale methods while the microscale properties are uncaptured .
the lack of ability for measurements on the microscale can lead to uncertainty in interpretations of the data captured at macroscale .
a major challenge arising from this non homogeneity is how macroscale flow is influenced by the microscale structure ( pore spaces ) , as well as by the physical properties of the resin .
permeability is an important macroscale variable representing average of microscale properties of porous media .
this average permeability is the fundamental property arising from dracy s law ( [ eq.darcy ] ) : [ eq.darcy ] = .p , which describes the relation between the volume averaged fluid velocity , , the pressure gradient , @xmath0p , the fluid s viscosity , @xmath1 , and the equivalent permeability tensor k. empirical equations , such as kozeny@xmath2carmen ( [ eq.kozeny ] ) , have been developed to relate the meso - scale permeability with the microscale properties , including porosity : [ eq.kozeny ] k=.here @xmath3 and @xmath4 , where @xmath5 is porosity , @xmath6 is fibre radius , @xmath7 is kozeny constant , @xmath8 is tortuosity , @xmath9 is the length of streamlines , l is the length of sample and c is a proportionality constant @xcite . because the nonhomogeneous nature of porous media originates in the randomness of fibre diameter distributions , porosities and pore structure , permeability is subjected to uncertainty @xcite .
causes and effects of this uncertainty have been reviewed @xcite , assuming a normal @xcite and a lognormal permeability probability density function @xcite , previous studies have modeled the effect of this uncertainty on fluid flow in porous media by employing stochastic analysis .
considered as a vast area in stochastic processes , such analyzes require to identify the sources of uncertainties and to select probabilistic methods for uncertainty propagation up to different modeling levels .
+ in simulation of mould filling process , such as resin transfer moulding , which are described by dracy s law ( [ eq.darcy ] ) , permeability directly affects filling time and flow pattern .
an accurate probability density function for permeability is therefore vital for reliable simulations . a number of studies @xcite have used a normal probability density functions for experimentally determined permeability at macroscale .
but their measurements have been mostly for small sample sizes and hence may still be the subject to experimental and statistical inaccuracy .
in other words , data obtained from the experiments may not be enough to choose between two ( normal or lognormal distribution ) or more competing distribution functions .
furthermore , these studies ignore the effect of microscale uncertainties on macroscale permeability uncertainties .
+ different approximation methods have been used to determine the impact of microscale uncertainties on macroscale permeability uncertainties , e.g , finite element based monte carlo and lattice - boltzmann methods have been used to estimate permeability and superficial velocity of representative volume elements of porous media @xcite .
the accuracy of the analytical methods has been debated because they have considered homogeneous periodic arrays of parallel fibres instead of random distributions .
in addition , all of the modeling approaches require that either the distribution of at least one property of fibrous media or the distribution of macropors of fibrous media be known .
another criticism is that estimating permeability by curve fitting with empirical constants is known to generate significant systematic errors .
+ in view of the fact that finding an appropriate distribution function describing the spatial variation of permeability in fibrous media is a challenging problem .
therefore , a method that measures the microstructural variability as input for stochastic simulations is required . in the @xcite , we analyzed the effect of tortuosity on the variability of permeability at the average local fibre volume fraction(microscale level ) .
we showed that the gaussian distribution is not necessarily the most appropriate distribution for representing permeability @xcite . in this study
, we capture the influence of a distribution of local fibre volume fraction ( @xmath5 ) on the variability of permeability . the uncertainty in this variable ( e.g,@xmath5 ) propagates to a larger level and is reflected in the variability of the geometry of flow affecting the final quality of composites .
in order to establish a probability density function for permeability , in this study we propose that ( i ) the best probability density function may be approximated for distribution of fibre volume fraction by a normal model , ( ii ) the values of fibre volume fraction were used to compute the distribution of permeability applying the kozeny - carmen equation , ( iii ) we used the kozeny - carmen equation together with change of variable technique to determine the probability density function of permeability and subsequently the analytical approach is compared with the distribution of permeability @xcite ( figure [ fig : flowchart1 ] ) .
there is no universal established relationship between fibre volume fraction and permeability . in this paper
we recall the carmen kozeny equation in order to find the pdf of permeability ( [ eq.kozeny ] ) .
observe that in ( [ eq.kozeny ] ) the random variable ( rv ) @xmath10 is an increasing function of porosity @xmath5 ( assuming @xmath11 , @xmath12 are constant ) , see figure [ change ] .
note also that @xmath10 is significantly affected by @xmath11 . however
this will not be addressed in this paper .
here we use change - of - variables technique , in this study called `` the change of @xmath5 '' , to investigate pdfs of permeability .
this technique is a common and well - known way of finding the pdf of @xmath13 if @xmath14 be a continuous random variable with a probability density function @xmath15 .
therefore to establish the pdf of random variable @xmath10 , it is required to have the pdf of @xmath5 . thus subsequently in the next section ,
we determine numerically the pdf of @xmath5 . ]
variable fibre volume fraction for a fibrous media depends on areal weight density ( @xmath16 ) and thickness(t ) and can be expressed as equation ( [ eq.poro ] ) : [ eq.poro ] v_f= where @xmath17 stands the fibre density , n is number of layers . to attain a closed form expression of the density of @xmath18 ,
both @xmath19 and @xmath20 are considered normal random variables with correlation coefficient , -1<@xmath21<1 .
when @xmath21=0 , the two variables @xmath19 and @xmath20 are independent , the distribution of @xmath18 would have a cauchy distribution .
note that the the cauchy distribution does not have finite moments of any order hence the mean and variance of @xmath5 are undefined .
therefore , assuming a cauchy distribution would not be an appropriate model for fibre volume fraction .
now set @xmath22 and @xmath23 . in this stage of work
we consider the first order taylor expansion about @xmath24 for @xmath25 ) : [ eq.taylor ] v_f:=v_f(a_w , t)=v_f(_a_w,_t)+v_f_a_w(_a_w,_t)(a_w-_a_w)+v_f_t(_a_w,_t)(t-_t)+o(n^-1),where @xmath26 and @xmath27 are the derivatives of @xmath5 with respect to @xmath19 and @xmath28 respectively . in agreement with @xcite , the approximation for @xmath29
is given by [ exp : v ] ccl _
v_f&=&(v_f(_a_w,_t)+v_f_a_w(_,_t)(a_w-_a_w)+v_f_t(_a_w,_t)(t-_t)+o(n^-1 ) ) + & & ( v_f(_a_w,_t))+v_f_a_w(_a_w,_t)(a_w-_a_w)+v_f_t(_a_w,_t)(t-_t ) + & = & ( v_f(_a_w,_t))=n/ a_w_f _ t. by virtue of the definition of variance we can write [ eq.var ] ^2(v_f)=^2()()^2.next , use the first order taylor expansion once again around @xmath24 . then owing to ( [ exp : v ] ) we approximate [ eq : apx.v ] ^2(v_f)\{(v_f(a_w , t)-v_f(_a_w,_t))^2}.substitute ( [ eq.taylor ] ) in ( [ eq : apx.v ] ) , then ( [ eq.var ] ) becomes the following : [ eq.var.1 ] ^2(v_f)&()^2 ^2(+(-_)-(h-_h ) ) + & = ( ) ^2 ^ 2(-t ) + & = ( ) ^2 ^ 2(a_w)+^2(t)-2cov(a_w , t ) + & = ( ) ^2(^2(a_w)+^2(t)-2cov(a_w , t ) ) in ( [ eq.var.1 ] ) , @xmath30 expresses the covariance of @xmath31 while as we said @xmath32 and @xmath33 are average of local areal density and thickness , respectively .
now denote @xmath34 , @xmath35 as the standard deviation of rvs @xmath19 , @xmath28 and moreover @xmath21 as their correlation coefficient and their relationship can be expressed as : [ eq.stan ] cov(a_w , t)=(a_w)(t)_corr.in addition we define the coefficient of variation ( @xmath36 ) of a rv , such as @xmath37 : cv(v_f)= [ eq.cvvv ] consequently eqn .
( [ eq.var.1 ] ) can be recast as : ^2(v_f)(_v_f)^2(cv^2(a_w)-2cv(a_w)cv(t)a_w_corr+cv^2(t))[eq.var.2]consider ( [ eq.var.2 ] ) and replace it in ( [ eq.cvvv ] ) , then the @xmath36 of fibre volume fraction yields : cv(v_f ) [ eq.var.3]equation ( [ eq.var.3 ] ) shows that the @xmath36 of rv fibre volume fraction@xmath38 is approximately independent from @xmath39 .
owing to ( [ eq.var.3 ] ) at point @xmath40 , evidently 3d scatter plots ( [ fig : hhh ] ) exhibits variation of the @xmath36 of the rv fibre volume fraction @xmath38 as a function of the coefficient of variations thickness @xmath41 and local areal density @xmath42 .
+ in figure [ fig : hhh ] , it is possible to observe that when @xmath21 increases from zero to one , a saddel is formed with @xmath43 , @xmath44 and then @xmath45 .
furthermore , figure [ fig : hhh ] shows that as @xmath42 and @xmath41 approach each other too closely , @xmath37 moves from the top to the lowest level .
+ as to the pdf of @xmath5 , section 2.1.1 establises experimentally that the random variable @xmath5 has normal distribution .
although by calling the following cases we claim that this assertion in independent case is not theoretically artificial : * @xmath19 and @xmath28 are * independent*. it is straightforward that when a random variable @xmath28 follows cauchy distribution with parameter @xmath46 then @xmath47 has also cauchy distribution with parameter @xmath48 .
further , according to our explanation above , we can easily check that the ratio of two independent normal random variables determines the cauchy pdf .
hence as result if we consider random variables @xmath19 and @xmath28 takes normal and cauchy distributions respectively then @xmath5 has normal pdf . *
@xmath19 and @xmath28 are * dependent*. this case is more complicated but practical .
first note that if one is interested in bivariate normal distribution for pair @xmath49 .
we address to @xcite which study the ratio of two correlated normal random variables .
the author indeed has established that if @xmath19 and @xmath28 be normally distributed random variables with means @xmath50 , variances @xmath51 , @xmath52 and correlated coefficient @xmath21 , then the exact pdf of @xmath5 takes the form ( 1 ) in page 636 in @xcite .
for simplicity we omit the form ( 1 ) here .
however , in @xcite we could observe that in this special case , i.e. normally distributed @xmath49 the pdf of @xmath5 is not necessarily symmetric and normal .
therefore by virtue of author s investigations , it is not clear yet that what kind of distributions should be considered for @xmath49 to prove analytically a normal pdf for @xmath5 .
consequently , the solution which may come cross the mind is experimental results represented in next subsection .
table [ input ] shows the specifications of a @xmath53 twill carbon woven fabric used for production of composite parts in this study .
the fabric was cut in the warp direction .
composite parts were produced with same @xmath38 using high injection pressure resin transfer moulding(hiprtm ) .
full production details have been presented in @xcite .
[ input ] in order to measure the h of each tow , series of samples were cut perpendicular to the fibre direction , the samples were polished manually in four steps ( sandpaper grits 320 , 400 , 800 , and 1000 ) , and subsequently photographed using an olimpus png3 optical microscope equipped with a ccd camera .
the analysis of each tow is carried out by image processing matlab 2015 . before importing the images into the matlab workspace , arears such as edges or borders which are not in the interest of tow geometry characterizations were cropped . in projective geometry
every tow section is equivallent to an ellipse .
figure [ ellipse]a illustrates how an ellipse is fitted to a tow to locate the center point based on @xcite and @xcite .
then the tow thickness ( @xmath28 ) is equivalent to twice the length of the minor radius ( @xmath54 ) ( figure [ ellipse]b ) .
a total of 200 optical images were collected and an ellipse is fitted to each tow . to determine the @xmath55 of each tow , the area of each tow ( @xmath56 ) in warp direction was approximated by the equivalent ellipse .
then , the number of filaments per tow ( @xmath57 ) were counted .
knowing the radius of the fibre cross section(r ) , length of warp tow ( l ) , and @xmath58 , the @xmath55 of each tow@xmath59 was computed from the following equation ( [ eq.ad ] ) : = [ eq.ad ] to determine distribution of @xmath18 , the `` r '' statistical software @xcite was employed .
for each pair ( @xmath20,@xmath19 ) , @xmath18 was computed , afterward a histogram was generated . a scatter plot of @xmath55 and @xmath20 for each tow is shown in figure [ fig : cv ] .
the coefficient of variation for the data was 0.72 .
an ellipse is fitted to the data .
we can see that the ellipse extends between 100 and 200 @xmath60 on the 45 degree line .
hence @xmath55 and @xmath20 follow a bivariate normal distribution with mean components 168.5 @xmath61 and 102.3 @xmath62 , respectively . therefore ,
our results suggest that @xmath55 and @xmath20 are dependent and can be well approximated by the bivariate normal distribution . for each pair of @xmath55 and @xmath20 ,
@xmath18 was computed .
figure [ fig : hist1 ] shows that the distribution of the local average fibre volume fraction follows a bell - curve distributions : the distribution of fibre volume fraction values are well approximated with a normal distribution model .
a large distribution of @xmath18 was observed , ranging from @xmath63 to @xmath64 .
the graphical analysis indicates the closeness of the local average fibre volume fraction data to the normal distribution with cumulative distribution function(cdf)(figure [ fig : cdf ] ) .
the cdf plot is following a typical s curve indicative of normal distribution .
it follows from the section of @xmath65 that the fibre volume fraction of a preform , @xmath5 , has a truncated normal pdf , @xmath66 , with mean @xmath1 and variance @xmath67 over @xmath68 $ ] such that @xmath69 : [ pdf.tn ] f^tn(v_f)= , v_f1<v_f < v_f2.in eqn .
( [ pdf.tn ] ) and ( [ pdf.tn1 ] ) below , @xmath70 introduces the distribution function of a standard normal rv @xmath71 , that is [ pdf.tn1](z)=(zz)=_-^z ( ) t . passing to the limits @xmath72 and @xmath73 eqn .
( [ pdf.tn ] ) yields the pdf of normal distribution with mean @xmath1 and variance @xmath67 .
now assume the ratio @xmath74 is constant , the rv @xmath10 is given by an increasing function of @xmath5 , @xmath75 . set @xmath76 as inverse function of @xmath76 , i.e. @xmath77 .
+ the figure [ change ] presents a graphical representation of function @xmath78 where @xmath10 is a monotonically increasing function of @xmath5 .
+ next , let @xmath79 be the distribution function of rv @xmath10 , more precisely @xmath80 . therefore , recalling @xmath75 one can do : f_k(k)=(g(v_f)k)=(v_fg^-1(k))=f_v_f(g^-1(k)).where @xmath81 is the distribution function of @xmath5 .
note that the second probability derives from the of that @xmath78 is an increasing function of @xmath5 . + consequently by taking the derivative of @xmath79 with respect to @xmath82 , the pdf of rv @xmath10 is obtained by l f_k(k)=f_k(k)= f_v_f(g^-1(k ) ) + = g^-1(k ) .
f_v_f(g^-1(k ) ) . introduce @xmath83 and @xmath84 , since @xmath78 is an increasing function then @xmath85 .
coming back to the assumption says that the rv @xmath5 has pdf @xmath86 in @xmath68 $ ] , we obtain [ eq.k1 ] l f_k(k)=g^-1(k ) .
f^tn_v_f(g^-1(k ) ) + = g^-1(k ) . , k_1<k <
k_2 . as figure [ change ] shows this function is a one to one function so it is invertible .
however , while it may be possible to find a closed form solution to this inverse , we will begin by presenting a numerical solution to this problem .
a summary of statistical properties of permeability for the different cov of @xmath87 is represented in figure [ fig : cvflow ] .
it is observed that an in - plane distribution of local fibre volume fraction results in a distribution of local permeabilities for flow perpendicular to the plane of the fibrous medium .
it is established in figure [ fig : cvflow ] that domain with larger local average areal densities possess larger local average permeability values .
this phenomenon can be explained in terms of probability of number of contact points between fibres .
the probability of number of contact points of fibres is larger for the domains with higher local average areal density .
higher contact points means stiffer arrangement of adjacent fibres , causing increased frictional resistance to fluid flow , and hence leading to less permeable area .
furthermore the normalized local average permeability ( i ) initially decreases when @xmath88 increases due to its influence on @xmath37 , ( ii ) subsequently increases due to the increasing influence of @xmath88 on the normalized local average permeability .
therefore , as expected , this implies that the heterogeneity of fibrous media has a significant impact on the magnitude of variation in permeability .
the pdf ( [ eq.k1 ] ) is computed applying monte - carlos simulations ( using the `` r '' statistical software @xcite ) .
monte - carlo simulation was carried out on the domain with a average fibre volume fraction of 0.78 .
subsequently this analytical model fit ( [ eq.k1 ] ) was compared with what would be predicted by physically based permeability equations ( e.g. kozeny - carman([eq.kozeny ] ) ) .
to do so , using the values of @xmath5 computed earlier , the distribution of permeability was obtained in ( [ eq.kozeny ] ) .
then , probability values calculated through ( [ eq.k1 ] ) were fit to permeability values . finally , in order to find the best distribution model to the analytical pdf , different statistical distributions are examined .
+ figure [ map ] shows , obtained through the method described above , ( i ) the data are not symmetric and skewed to the right , ( ii ) there is significant agreement between the analytical approach and the simulation results of permeability ( kozeny - carmen ) .
the observed skewness in permeability distribution is close to what one was observed in @xcite at macrolevel .
the same principle was applied to various kinds of corrolations between @xmath10 and @xmath5 .
table [ com ] lists the distribution behaviour of @xmath10 derived from models found in the literature .
the appearance of the distributions presented in table [ com ] is the same as one shown in figure [ map ] .
table [ com ] shows that the skewness of @xmath10 lies between 1.78 - 3.45 , implying non - normal distribution of the permeability .
in addition , the kurtosis values range between 9 and 24 , deviating extremely from normality .
as expected , there is a significant difference between the calculated skewness and kurtosis and that of the normal distribution .
[ com ] furthermore , figure [ fa ] illustrates how well the different distributions including normal , lognormal , gamma , and beta fit to the analytical probability function .
it is clear that the normal and beta distributions fail to represent the distribution of permeability values .
figure [ fa ] also shows a lognormal distribution as well as a gamma distribution provide a good fit for permeability values .
it has been derived that both lognormal and gamma distributions can be used effectively in analyzing a non - negative right - skewed data set @xcite . as shown in figure [ fa ] , a comparison between the lognormal and gamma distributions reveals that their pdfs have similar results . to find which of these distributions gives better fit to the data ,
we have considered two data transformation methods , one based on the normal approximation to the log of the data set , working on lognormal distributions and the other based on the cube root of the data , working on gamma distribution : if the data looks symmetric after log transformation , the lognormal distribution would work better to represent the variation of the permeability .
if the data looks symmetric after the cube root transformation , the gamma distribution would work better to represent the variation of the permeability @xcite .
according to figure [ gl ] , although the log transformation seems to fit well in the body of permeability values , the data shows to be left - skewed on log scale .
the significant left - skewed on log scale also is observable on the work of zhang et al.@xcite(see figure [ zh ] ) , who evaluated the permeability from local areal weight combined with the kozeny - carmen model , suggesting that the lognormal distribution can be used to describe the permeability distribution
. one final comment on which we would like to conclude this section regards entropy . in the literature
it has been proven that under constraints , known mean and variance , the normal distribution maximizes entropy .
using this principle that maximizes entropy is a selecting factor for a model we calculated the entropy values of the two fits . using straightforward computations we obtain the entropies for lognormal pdf -0.931 and gamma pfd -0.887 , which coincids with the previous result claiming gamma distribution with more entropy determines the permeability data . in order to examine the applicability of gamma distributions to other empirical permeability equations ,
these equations are subjected to the kolmogorov - smirnov statistics ( kss ) test for normality in terms of normal , lognormal , gamma , weibull and beta distributions .
if , in the kss test p < 0.05 , there is significant probability of deviation from normality .
the p value computations are listed in table [ gamma ] .
gamma distribution shows the largest p - values among the given distributions for the different empirical permeability equations .
hence , based on the information in table [ gamma ] , we conclude that the gamma distribution provides the best fit .
furthermore , it is shown in table [ gamma ] the permeability covs of all empirical equations range between 0.43 and 0.74 , which is about 8 times larger than the fibre volume fraction cov of 0.086 .
this suggests that permeability is subjected to larger uncertainty than fibre volume fraction .
an adequate representation of microstructural variability of fibre arrangement in fibre - reinforced composites is of critical importance for the analysis of the flow in the fibrous media . the distribution of fibre volume fraction was quantified by the measurement of areal weight density and areal thickness from optical images of tows in a @xmath53 twill carbon - epoxy composite .
then , the pdf of the permeability was determined by a known pdf of @xmath5 and assuming a constant @xmath12 .
to do so , we proposed a method to determine the probability density function of the permeability of porous media .
we employed the kozeny - carmen equation and combined it with the change of variable technique .
our results suggest that @xmath89 the relationship between the local areal weight density and thickness is well approximated by a bivariate normal distribution . @xmath90
the distribution of local fibre volume fraction exhibits a bell - shaped curve and fit well to a normal distribution model . @xmath91 assuming constant @xmath12 , a gamma distribution could more accurately describe the variation in permeability data . + as conclusion , the understanding of the probability distribution of permeability is still taking further clarification but that the hypothesis of normality has been refuted .
sys thanks the capes pnpd - ufscar foundation for the financial support in the year 2014 - 5 .
sys thanks the federal university of sao carlos , department of statistics , for hospitality in 2014 - 5 .
carmen p. fluid flow through a granular bed .
_ transactions of the institution of chemical engineers _ 1937 ; 15:150167 .
wong c , long a. modelling variation of textile fabric permeability at mesoscopic scale .
_ plast rubber compos _ 2006 ; 35(3):101111 .
mesogitis ts , skordos aa , long ac . uncertainty in the manufacturing fibrous thermosetting composites : a review .
_ composite part a : applied science and manufacturing _ 2014 ; 57:67 - 75 .
bodaghi m , gonalves ct , correia nc . a quantitative evaluation of the uncertainty of permeability measurements in constant thickness fibre reinforcement .
_ in : proceedings of eccm-16 conference .
june , 2014 .
arbter r , et al .
experimental determination of the permeability of textiles : a permeability benchmark exercise .
_ composite part a : applied science and manufacturing _ 2011 ; 42:1157 - 1168 .
vernet n , et al
. experimental determination of the permeability of engineering textiles : permeability benchmark ii . _
composite part a : applied science and manufacturing _ 2014 ; 61:172 - 184 .
pan r , liang z , zhang c , wang b. statistical characterization of fibre permeability for composite manufacturing .
_ polymer composites _ 2000 ; 21(6 ) : 996 - 1006 .
hoes k , dinescu d , sol h , vanheule m , parnas rs , luo y , verpoest i. new set up for measurement of permeability properties of fibrous reinforcements for rtm .
_ composite part a : applied science and manufacturing _ 2002 ; 33:959 - 969 .
li j , zhang c , liang z , wang b. stochastic simulation based approach for statistical analysis and characterization of composites manufacturing processes . _
journal of manufacturing systems _ 2006 ; 25(2 ) : 108 - 121 .
endruweit a , long ac , robitaille f , rudd cd .
influence of stochastic fibre angle variations on the permeability of bi - directional textile fabrics .
_ composite part a : applied science and manufacturing _ 2006 ; 37:122 - 132 .
wong cc . modelling the effects of textile preform architecture on permeability .
_ phd thesis , university of nottingham , nottingham , england _ 2006 .
padmanabhan sk , pitchumani r. stochastic modelling of nonisothermal flow during resin transfer moulding .
_ international journal of heat and mass transfer _ 1999 ; 42:3057 - 3070 .
zahng f , cosson b , comas - cardiba s , binetruy c. efficient stochastic simulation approach for rtm process with random fibrous permeability .
_ composites science and technology _ 2011 , 71:1478 - 1485 .
zahng f , comas - cardiba s , binetruy c. statistical modelling of in - plane permeability of non - woven random fibrous reinforcement .
_ composites science and technology _ 2012;72(12):136879 .
parseval y , roy rv , advani sg .
effect of local variations of preform permeability on the average permeability during resin transfer molding of composites .
_ antec95 _ 1995;2:3040 - 3044 .
gebart b.r.permeability of unidirectional reinforcements for rtm
. _ journal of composite materials _ 1992 , 26:1100 - 1133 .
sahraoui m , kaviany m. slip and no - slip boundary conditions at interface of porous , plain media .
_ journal of heat and mass transfer _ 1992 , 35:927 - 94 .
bruschke m. , advani s.g .
flow of generalized newtonian fluids across a periodic array of cylinders . _ journal of rheology _ 1993 , 37:479 - 498 .
westhuizen j.v.d , plessis j.p.d .
an attempt to quantify fibre bed permeability utilizing the phase average navier - stokes equation . _ composite part a : applied science and manufacturing _ 1996 , 27a:263 - 269 .
lee s.l , yang j.h .
modeling of darcy
forchheimer drag for fluid flow across a bank of circular cylinders .
_ international journal of heat and mass transfer _ 1997 40:3149 - 3155 .
catalanotti g , bodaghi m , correia n. on the statistics of transverse permeability of randomly distributed fibres .
_ submitted to journal of science and technology _ 2015 ; under review .
yu b. analysis of flow in fractal porous media . _ applied mechanics reviews _ 2008 , 61 : 050801 - 1 - 19 .
haldar a , mahadevan s. probability , reliability , and statistical methods in engineering design . _ ohn wiley , new york _ 2000 .
stuart a , ord k. _ kendall s advanced theory of statistics _ 1998 ; arnold , london , 6@xmath92 edition , vol 1 : p. 351 .
elandt - johnson r c , johnson n l. _ survival models and data analysis _ 1980 ; john wiley and sons ny , p. 69 .
hinkley d v. on the ratio of two correlated normal random variables . _ biometrica _ 1969 ; 56 ( 3 ) : 635 - 639 .
cedilnik a , komelj k , blejec a. ratio of two random variables : a note on the existence of its moments . _
metodoloki zvezki _ 2006 ; 3 ( 1 ) : 1 - 7 .
xie y , jie q. a new efficient ellipse detection method .
_ ieee _ 2002 ; 1051 : 4651 - 4652 .
basca c.a , talos m , brad r.randomized hough transform for ellipse detection with result clustering . _ ieee _ 2005 ; 1397 - 1400 .
dodson c , oba y , sampson w. on the distributions of mass , thickness and density in paper .
_ appita j _ 2001,54:385 - 389 .
team , r.d.c .
r : a language and environment for statistical computing . _
vienna , austria : r foundation for statistical computing _ 2010 .
liu q , parnas r.s , giffard h.s .
new set - up for in - plane permeability measurment _
composite part a : applied science and manufacturing _ 2007 , 38:954 - 962 .
gebart b.r.permeability of unidirectional reinforcements for rtm .
_ journal of composite materials _ 1992 , 26:1100 - 1133 .
bruschke b.r , advani s.g .
flow of generalized newtonian fluids across a periodic array of cylinders .
_ _ journal of rheology__1993 , 37:479 - 498 .
gutowski t.g , cai z , bauer s , boucher d , kingery j , wineman s. consolidation experiments for laminate composites._journal of composite materials _ 1987 , 21:650 - 669 .
happel j. viscous flow relative to arrays of cylinders .
_ alche _ 1959 , 5:174 - 177 .
lee s.l , yang j.h .
modeling of dracy - forchheimer drag for fluid flow across a bank of circular clinders .
_ international journal of heat and mass transferr _ 1997 , 40:3149 - 3155 .
sharaoui m , kaviany m. slip and no - slip boundary conditions at interface of porous , plain media .
_ international journal of heat and mass transferr _ 1992 , 35:927 - 943 .
johnson , norman l. ; kotz , samuel ; balakrishnan , narayanaswamy.continuous univariate distributions .
_ wiley _ 1994,1:173 .
e.b.wilson , m.m.hilferty.the distribution of chi - square .
_ proceedings of the nationa academy of sciences _ 1931,17:684 - 688 .
p.zhang , p.x.k.song , a.qu , t.greene .
effiecient estimation for patient - specific rates of disease progression using nonnormal linear mixed models .
_ biometrics _
2008,64:29 - 38 . | permeability of fibrous porous media at the micro / meso scale - level is subject to significant uncertainty due to the heterogeneity of the fibrous media .
the local microscopic heterogeneity and spatial variability porosity , tortuosity and fibre diameter affect the experimental measurements of permeability at macroscopic level .
this means that the selection of an appropriate probability density function ( pdf ) is of crucial importance , in the characterization of both local variations at the microscale and the equivalent permeability at the experimental level ( macroscale ) .
this study addresses the issue of whether or not a normal distribution appropriately represents permeability variations . to do so , ( i ) the distribution of local fibre volume fraction for each tow is experimentaly determined by estimation of each pair of local areal density and thickness , ( ii ) the kozeny - carmen equation together with the change of variable technique are used to compute the pdf of permeability , ( iii ) using the local values of fibre volume fraction , the distribution of local average permeability is computed and subsequently the goodness of fit of the computed pdf is compared with the distribution of the permeability at microscale level
. finally variability of local permeability at the microscale level is determined .
+ the first set of results reveals that ( 1 ) the relationship between the local areal density and local thickness in a woven carbon - epoxy composite is modelled by a bivariate normal distribution , ( 2 ) while fibre volume fraction follows a normal distribution , permeability follows a gamma distribution , ( 3 ) this work also shows that there is significant agreement between the analytical approach and the simulation results .
the second set of results shows that the coefficient of variation of permeability is one order of magnitude larger than that of fibre volume fraction .
future work will consider other variables , such as type of fabrics , the degree of fibre preform compaction to determine whether or not the bivariate normal model is applicable for a broad range of fabrics
. 1 truecm .5 truecm |
continuous - wave gravitational radiation from isolated and accreting neutron stars is a key target of long - baseline interferometers like the laser interferometer gravitational wave observatory ( ligo ) and virgo detector @xcite .
theory predicts that the signal is quasi - monochromatic .
emission occurs at simple rational multiples of the star s spin frequency @xmath8 , for example @xmath8 and @xmath9 for mass quadrupole radiation from thermoelastic and magnetic mountains @xcite , @xmath10 for r - modes @xcite , and @xmath8 for current quadrupole radiation from nonaxisymmetric flows in the neutron superfluid pinned to the stellar crust @xcite . if the source exhibits electromagnetic pulsations , so that an ephemeris can be derived from absolute pulse numbering , i.e. @xmath11 is known as a function of time @xmath12 , it is customary to search for a signal using coherent matched filters like the maximum likelihood @xmath0-statistic @xcite .
if an ephemeris is unavailable , coherent searches over multiple @xmath13 templates indexed by the taylor coefficients @xmath14 become expensive computationally @xcite , and semi - coherent methods are often preferred .
cross - correlation @xcite , stackslide @xcite , the hough transform @xcite , powerflux @xcite and twospect @xcite are all examples of semi - coherent algorithms implemented by the ligo scientific collaboration and applied to data from science runs 5 or 6 .
search methods that scan templates without guidance from a measured ephemeris are compromised if @xmath11 wanders randomly .
radio and x - ray timing of pulsating neutron stars reveal that spin wandering is a widespread phenomenon . in isolated objects
, it manifests itself as timing noise @xcite , exhibits a red fourier spectrum with an auto - correlation time - scale of days to years @xcite , and has been attributed variously to magnetospheric changes @xcite , superfluid dynamics in the stellar interior @xcite , spin microjumps @xcite , and fluctuations in the spin - down torque @xcite . in accreting objects , spin wandering results from fluctuations in the magnetized accretion torque @xcite , due to transient accretion disk formation @xcite or disk - magnetospheric instabilities and reconnection events @xcite .
again the auto - correlation time - scale is of the order of days @xcite , and fluctuations in @xmath11 are accompanied by fluctuations in the x - ray flux @xcite .
even when the ephemeris is measured electromagnetically , the gravitational - wave - emitting quadrupole may not be phase locked to the stellar crust , and wandering must still be accommodated @xcite .
hidden markov model ( hmm ) methods offer one powerful strategy for detecting and tracking a wandering frequency @xcite .
the essential idea is to model @xmath11 probabilistically as a markov chain of transitions between unobservable ( hidden " ) frequency states and relate the hidden states to the observed data via a detection statistic .
hmm frequency tracking enjoys a long track record of success in engineering applications ranging from radar and sonar analysis @xcite to mobile telephony @xcite .
it has been refined substantially since its introduction by @xcite to include information about amplitude and phase @xcite and embrace simultaneous tracking of multiple targets and frequencies @xcite .
it delivers accurate estimation , when the signal - to - noise ratio ( snr ) is low but the sample size is large @xcite , the situation normally confronting gravitational wave data analysis targeting continuous - wave sources . in this paper , we implement and test a specific hmm scheme based on the classic viterbi algorithm @xcite .
the scheme is efficient and practical : its computational demands are modest , and it co - opts existing technology for ligo continuous - wave searches based on the @xmath0-statistic .
the paper is organized as follows . in section
ii , we formulate the search problem in hmm language and describe how to solve it with the viterbi algorithm . in section iii , we apply viterbi tracking to an isolated neutron star with @xmath11 walking randomly , using the @xmath0-statistic to relate the hidden and observable states .
we quantify the performance of the tracker as a function of signal strength using synthetic data . in section
iv we apply viterbi tracking to a neutron star in a binary , again with @xmath11 wandering randomly , replacing the @xmath0-statistic with a bessel - weighted variant , and quantify its performance using synthetic data . finally , to illustrate how the tracker performs in a
realistic " scenario , we apply it to the data set prepared for stage i of the scorpius x-1 ( sco x-1 ) mock data challenge in section v @xcite .
hmm principles can be applied to gravitational wave frequency tracking in various ways , depending on the data format , the detection statistic , and any known constraints on the frequency evolution . in this paper
, we specialize to continuous - wave searches , where the raw data are packaged in 30-min short fourier transforms ( sfts ) , during which @xmath11 remains localized within one fourier bin .
the sfts are fed into a frequency - domain estimator @xmath15 , like the @xmath0-statistic ( isolated target ) or bessel - weighted @xmath0-statistic ( binary target ) .
it is safe to assume that @xmath11 remains localized within one estimator frequency bin over a short enough time interval , even when @xmath11 walks randomly .
let @xmath16 be the total observation time , and let @xmath17min be the length of each sft .
for any particular astrophysical source , there exists an intermediate time - scale @xmath18 ( @xmath19 ) , over which @xmath11 wanders by at most one estimator bin . for timing noise in isolated pulsars ,
radio timing experiments imply @xmath20 @xcite . for accretion noise in low - mass x - ray binaries , such as sco x-1 , one can estimate @xmath21days theoretically @xcite .
our general strategy is to compute @xmath15 for blocks of data of length @xmath18 then use the viterbi algorithm to track peaks in @xmath15 over the full observation interval @xmath22 , effectively summing the estimator output semi - coherently by tracking @xmath11
. can be as short as @xmath23 , the minimum time over which it makes sense to calculate @xmath15 , without harming the performance of the tracking algorithm .
as the algorithm is semi - coherent , the strain sensitivity scales approximately as @xmath24 @xcite . ]
an important practical advantage of this approach is that it leverages the existing , efficient , thoroughly tested @xmath0-statistic software infrastructure within the ligo algorithms library ( lal ) , which is used extensively by the continuous - wave data analysis community @xcite . in this section
we formulate the problem as an hmm ( section [ sec : hidden_markov][sec : setup_markov ] ) and describe the recursive viterbi algorithm for solving the problem ( section [ sec : viterbi ] ) .
we defer to future work the treatment of glitches , i.e. random , impulsive , spin - up events , where @xmath13 jumps over many estimator bins instantaneously @xcite .
an hmm is a probabilistic finite state automaton defined by a hidden ( unobservable ) state variable @xmath25 , which takes one of a finite set of values @xmath26 at time @xmath12 , and an observable state variable @xmath27 , which takes one of the values @xmath28 .
the automaton jumps between states at discrete times @xmath29 .
the jump probability from time @xmath30 to time @xmath31 depends only on the hidden state @xmath32 at time @xmath30 the markovian assumption and is described by the transition probability matrix @xmath33.\ ] ] the likelihood that the system is observed in state @xmath34 at time @xmath30 is described by the emission probability matrix @xmath35.\ ] ] the model is completed by specifying the probability that the system occupies each hidden state initially , described by the prior vector @xmath36.\ ] ] suppose that we observe the system transitioning through the sequence of observable states @xmath37 .
in general there exist @xmath38 possible paths @xmath39 through the hidden states which are consistent with the observed sequence @xmath40 . for a markov process
, the probability that the hidden path @xmath41 gives rise to the observed sequence @xmath40 equals @xmath42 the most probable path @xmath43 is the one that maximizes @xmath44 , i.e. , @xmath45 where @xmath46 returns the argument that maximizes the function @xmath47 .
the viterbi algorithm , presented in section [ sec : viterbi ] , provides a recursive , computationally efficient route to computing @xmath48 from ( [ eqn : prob_matrix])([eqn : best_path ] ) . for computational reasons ,
we actually evaluate @xmath49 , whereupon ( [ eqn : prob ] ) becomes a sum of @xmath50 likelihoods . in our first application to isolated neutron stars ( section [ sec : isolated_ns ] ) , the hidden state variable is one - dimensional and equals the star s spin frequency at time @xmath12 , i.e. @xmath51 .
its allowed , discretized values correspond one - to - one to the frequency bins in the output of the frequency - domain estimator @xmath52 computed over an interval of length @xmath18 , indexed by their central ( say ) frequencies @xmath53 , i.e. @xmath54 . and frequency @xmath11 jointly ( outside the scope of this paper ) , then the state space enlarges appropriately , with @xmath55.$ ] ] if the total search band covers the frequency range @xmath56 , and if the bin width of @xmath52 is @xmath57 , then the number of hidden states is given by @xmath58 .
radio and x - ray timing observations demonstrate that neutron star spin wandering is a continuous stochastic process in the absence of glitches @xcite . in what follows
we approximate it by an unbiased random walk or wiener process .
choosing @xmath18 to satisfy @xmath59 guarantees that the transition probability matrix takes a simple tridiagonal form , with @xmath60 and all other entries zero .
in other words , at each time step , @xmath11 jumps at most one frequency bin up or down or stays in the same bin with equal probability @xmath61 .
studies show that the performance of a hmm tracking scheme is insensitive to the exact form of the probabilities @xmath62 , as long as they capture broadly the behaviour of the underlying jump process @xcite .
equation ( [ eqn : trans_matrix ] ) can be generalized to accommodate discontinuous glitches , but doing so lies outside the scope of this paper .
be the probability of a glitch occurring in an interval @xmath18 , where @xmath63 is the mean glitch waiting time , and let @xmath64 be the maximum fractional glitch size .
then we replace @xmath61 in equation ( [ eqn : trans_matrix ] ) with @xmath65 and also have @xmath66 for @xmath67 . ] in our second application to binary neutron stars ( sections [ sec : binary ] and [ sec : mdc ] ) , the hidden state variable is two - dimensional : we search over not only @xmath11 but also the projected semimajor axis @xmath68 , where @xmath69 is the semimajor axis of the orbit , and @xmath70 is the angle of inclination . in practice
we divide the one - standard - deviation error bar on @xmath71 from electromagnetic measurements into @xmath72 bins , whereupon the total number of hidden states equals @xmath73 .
it is known astrophysically that @xmath71 does not change significantly over a typical observation ( @xmath74yr ) , so the transition matrix is still given by equation ( [ eqn : trans_matrix ] ) , with the subscript @xmath75 indexing the frequency bin @xmath53 but not @xmath71 . in other words , there is no dependence on @xmath71 in the two - dimensional version of equation ( [ eqn : trans_matrix ] ) .
therefore no separate subscript is needed to index @xmath71 .
the observable state variable @xmath27 comprises the data collected during the interval @xmath76 .
formally it is the vector @xmath77 $ ] , whose dimension @xmath78 equals the interferometer sampling frequency ( @xmath79khz ) times @xmath18 , where @xmath80 is the output of the interferometer signal channel , @xmath81 denotes the gravitational wave strain , and @xmath82 denotes the noise . the emission probability matrix is then given by @xmath83\\ \label{eqn : matrix_propto } & \propto & \exp[g({f_\star}_i)],\end{aligned}\ ] ] where ( [ eqn : matrix_propto ] ) follows from ( [ eqn : emi_prob_matrix ] ) by the definition of the frequency domain estimator . since we have no initial knowledge of @xmath11 , as prior we chose uniform distribution , i.e @xmath84 for all @xmath85 .
the maximum entropy of uniform distribution serves us to avoid any unwarranted assumptions about the signal . in binary neutron star applications
, @xmath86 depends on @xmath71 as well as @xmath8 ; an explicit formula is given in section [ sec : bessel ] .
the prior is the product of a uniform prior in @xmath8 and a uniform or gaussian prior in @xmath71 , the latter centred on the electromagnetically measured value with a standard deviation equal to the measurement uncertainty .
let @xmath87 $ ] be the first @xmath88 steps in the most probable path maximizing ( [ eqn : prob ] ) for the observation sequence @xmath89 $ ] .
for a markov process , @xmath90 satisfies the nesting property @xmath91 , i.e. @xmath92 makes up the first @xmath93 steps of @xmath90 .
this is a special case of the principle of optimality @xcite : if @xmath90 is optimal , then all subpaths within @xmath90 must be optimal too .
in other words , given the optimal path @xmath94 $ ] , one must have @xmath95 | [ o(t_i ) , \cdots , o(t_j)]\ } \leq
\pr \{[q^*(t_i ) , \cdots , q^*(t_j ) ] | [ o(t_i ) , \cdots , o(t_j)]\}$ ] for all possible choices of @xmath96 $ ] .
when applied to equation ( [ eqn : prob ] ) , the principle of optimality naturally defines a recursive algorithm , proposed by viterbi @xcite , to find @xmath43 by backtracking . at every forward step @xmath97 in the recursion
, the viterbi algorithm eliminates all but @xmath98 possible state sequences .
the retained sequences end in different states by construction ; if more than one sequence ends in a given state , the sequence with maximum @xmath99 $ ] is retained . at time @xmath100
, we save each of the @xmath98 maximum probabilities in the vector @xmath101 with components @xmath102\ ] ] and we save the state at @xmath103 leading to each retained sequence in the vector @xmath104 with components @xmath105,\ ] ] with @xmath106 = l_{o(t_k)q_i } a_{q_i q_j}\delta_{q_j}(t_{k-1})\ ] ] in both ( [ eqn : delta_t_k ] ) and ( [ eqn : phi_t_k ] ) . after stepping @xmath93 forward through @xmath107
, we backtrack through the @xmath108 vectors to find the optimal path . in summary , therefore , the viterbi algorithm comprises four stages .
@xmath109 @xmath110 for @xmath111 .
note that @xmath112 is never used .
@xmath113 @xmath114 , \\
\phi_{q_i}(t_k ) & = & \mathop{\arg \max } \limits_{1 \leq j \leq n_q } [ a_{q_i q_j}\delta_{q_j}(t_{k-1})],\end{aligned}\ ] ] for @xmath111 and @xmath107 .
@xmath115 @xmath116 for @xmath117 .
@xmath118 @xmath119 for @xmath120 . by pruning the tree of possible paths efficiently at each step
, the viterbi algorithm reduces the number of comparisons from @xmath121 to @xmath122 , which can be reduced further to @xmath123 by binary maximization @xcite .
we first consider an isolated neutron star , whose spin frequency wanders randomly by at most plus or minus one @xmath0-statistic frequency bin on the time - scale @xmath18 .
we take @xmath124d , @xmath1d , and @xmath125hz for the sake of illustration and with an eye to later comparison with the tests in section [ sec : binary ] and [ sec : mdc ] motivated by binaries like sco x-1 . over time intervals that are short
compared to @xmath18 , the optimal matched filter for a biaxial rotor with no orbital motion is the maximum - likelihood @xmath0-statistic , which accounts for the rotation of the earth and its orbit around the solar system barycentre ( ssb ) .
the time - dependent data @xmath126 collected at a single detector take the form @xmath127 where @xmath82 represents stationary , additive noise , @xmath128 are the four linearly independent signal components @xmath129 @xmath130 and @xmath131 are the antenna - pattern functions defined by equations ( 12 ) and ( 13 ) in ref .
@xcite , @xmath132 is the amplitude associated with @xmath133 , and @xmath134 is the signal phase at the detector .
the @xmath0-statistic is a frequency - domain estimator maximizing the likelihood of detecting a signal in noise with respect to the four amplitudes @xmath132 @xcite)([eqn : h4 ] ) assume emission at @xmath135 only , i.e. @xmath136 , as appropriate for a perpendicular rotor .
if the star s wobble angle is less than @xmath137 , emission also occurs at @xmath8 , and eight amplitudes are involved , i.e. @xmath138 . ] .
it is defined as @xmath139 where we write @xmath140 , and @xmath141 denotes the inverse matrix of @xmath142 .
the scalar product @xmath143 featuring in the definitions of @xmath144 and @xmath145 is defined as a sum over single - detector scalar products , @xmath146 where @xmath147 indexes the detector , @xmath148 is the one - sided noise spectral density of detector @xmath147 , the tilde denotes a fourier transform , and @xmath149 returns the real part of a complex number @xcite .
the expectation value of the @xmath0-statistic is @xmath150=4+\rho_0 ^ 2,\ ] ] where @xmath151 stands for the optimal snr given a signal in gaussian noise .
the random variable @xmath152 is distributed according to a non - central chi - squared probability density function with four degrees of freedom , @xmath153 , whose non - centrality parameter @xmath154 is related to the amplitudes by @xmath155,\ ] ] with @xmath156 , @xmath157 , and @xmath158 .
when there is no signal , the probability density function centralizes to give @xmath159 .
it can be shown that @xmath160 equals the snr , with @xmath161 where the constant @xmath162 depends on the right ascension , declination , polarization and inclination angles of the source .
averaging without bias over these angles yields @xmath163 for a perpendicular rotor and an interferometer with perpendicular arms @xcite .
we begin by illustrating the performance of the viterbi algorithm with some representative examples .
seven sets of synthetic data are created for @xmath124d at two detectors ( h1 and l1 ) with @xmath165 superposed on noise at a level typical of advanced ligo s design sensitivity , viz .
@xmath166hz@xmath167 , near the instrument s most sensitive frequency @xcite . sky position and source orientation are specified in table [ tab : sig - params ] .
the synthetic sfts are generated using _
makefakedata _ version 4 from the ligo data analysis software suite lalapps . for each test , we take @xmath1d , divide one year of data into @xmath168 segments , and create a 1-hz band of @xmath0-statistic output containing the injected @xmath11 for each segment .
we take @xmath169 as the prior and @xmath170 from the @xmath0-statistic output for each 10-day segment .
we also use the transition probability matrix given by equation ( [ eqn : trans_matrix ] ) . for each data set
, we calculate the root - mean - square deviation @xmath171 ( in hz ) between the optimal viterbi path and the injected @xmath11 .
tracking is deemed successful , if @xmath171 is smaller than one @xmath0-statistic frequency bin ( width @xmath172 ) . a systematic monte - carlo calculation of the threshold for detection lies outside the scope of this paper . 74 ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1 '' '' @noop [ 0]secondoftwo sanitize@url [ 0 ] + 12$12 & 12#1212_12%12 @startlink[1 ] @endlink[0 ] @bib@innerbibempty link:\doibase 10.1016/j.ppnp.2012.08.001 [ * * , ( ) ] link:\doibase 10.1046/j.1365 - 8711.2000.03938.x [ * * , ( ) ] link:\doibase 10.1086/428600 [ * * , ( ) ] link:\doibase 10.1103/physrevd.58.084020 [ * * , ( ) ] link:\doibase 10.1103/physrevd.79.104003 [ * * , ( ) ] link:\doibase 10.1088/0004 - 637x/807/2/132 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1093/mnras / stv1584 [ * * , ( ) ] link:\doibase 10.1103/physrevd.77.082001 [ * * , ( ) ] link:\doibase 10.1111/j.1365 - 2966.2011.18585.x [ * * , ( ) ] @noop ( ) link:\doibase 10.1103/physrevd.70.082001 [ * * , ( ) ] link:\doibase 10.1103/physrevd.87.042001 [ * * , ( ) ] link:\doibase 10.1103/physrevd.93.042007 [ * * , ( ) ] @noop ( ) @noop ( ) @noop ( ) link:\doibase 10.1103/physrevlett.102.111102 [ * * , ( ) ] link:\doibase 10.1103/physrevd.85.022001 [ * * , ( ) ] link:\doibase 10.1103/physrevd.93.042006 [ * * , ( ) ] link:\doibase 10.1088/0264 - 9381/28/21/215006 [ * * , ( ) ] link:\doibase 10.1103/physrevd.90.062010 [ * * , ( ) ] link:\doibase 10.1111/j.1365 - 2966.2009.15938.x [ * * , ( ) ] link:\doibase 10.1088/0004 - 637x/725/2/1607 [ * * , ( ) ] link:\doibase 10.1103/physrevd.91.062009 [ * * , ( ) ] link:\doibase 10.1086/158150 [ * * , ( ) ] link:\doibase 10.1111/j.1365 - 2966.2012.21863.x [ * * , ( ) ] link:\doibase 10.1126/science.1186683 [ * * , ( ) ] link:\doibase 10.1086/164765 [ * * , ( ) ] @noop * * ( ) link:\doibase 10.1093/mnras / stt1828 [ * * , ( ) ] link:\doibase 10.1086/191076 [ * * , ( ) ] \doibase 10.1051/0004 - 6361:20065267 [ * * , ( ) ] link:\doibase 10.1086/165672 [ * * , ( ) ] link:\doibase 10.1086/165673 [ * * , ( ) ] link:\doibase 10.1111/j.1745 - 3933.2006.00192.x [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1086/313060 [ * * , ( ) ] link:\doibase 10.1086/185143 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1086/426586 [ * * , ( ) ] @noop ( ) link:\doibase 10.1086/591526 [ * * , ( ) ] @noop _ _ ( , ) p. link:\doibase 10.1109/taes.2003.1207256 [ * * , ( ) ]
link:\doibase 10.1109/78.995087 [ * * , ( ) ] link:\doibase 10.1109/tnn.2002.1021891 [ * * , ( ) ] link:\doibase 10.1109/29.52700 [ * * , ( ) ] link:\doibase 10.1109/78.277803 [ * * , ( ) ] link:\doibase 10.1109/78.107416 [ * * , ( ) ] link:\doibase 10.1109/78.205738 [ * * , ( ) ] link:\doibase 10.1109/tit.1967.1054010 [ * * , ( ) ] link:\doibase 10.1103/physrevd.92.023006 [ * * , ( ) ] link:\doibase 10.1103/physrevd.89.043001 [ * * , ( ) ] link:\doibase 10.1103/physrevd.91.062008 [ * * , ( ) ] @noop ( ) link:\doibase 10.1086/523349 [ * * , ( ) ] in link:\doibase 10.1063/1.3615093 [ _ _ ] , vol .
( ) pp .
link:\doibase 10.1108/eb059970 [ _ _ ] , vol .
link:\doibase 10.1103/physrevd.75.023004 [ * * , ( ) ] link:\doibase 10.1088/0264 - 9381/32/7/074001 [ * * , ( ) ] link:\doibase 10.1109/tit.1974.1055282 [ * * , ( ) ] @noop _ _ ( , ) link:\doibase 10.1088/0004 - 637x/781/1/14 [ * * , ( ) ] link:\doibase 10.1103/physrevd.91.102005 [ * * , ( ) ] link:\doibase 10.1088/0264 - 9381/23/8/s23 [ * * , ( ) ] link:\doibase 10.1103/physrevd.76.082003 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.271102 [ * * , ( ) ] link:\doibase 10.1088/0264 - 9381/24/19/s10 [ * * , ( ) ] link:\doibase 10.1088/1742 - 6596/228/1/012005 [ * * , ( ) ] link:\doibase 10.1103/physrevd.91.062008 [ * * , ( ) ] link:\doibase 10.1111/j.1365 - 2966.2008.13594.x [ * * , ( ) ] link:\doibase 10.1088/0264 - 9381/33/10/105017 [ * * , ( ) ] | gravitational wave searches for continuous - wave signals from neutron stars are especially challenging when the star s spin frequency is unknown a priori from electromagnetic observations and wanders stochastically under the action of internal ( e.g. superfluid or magnetospheric ) or external ( e.g. accretion ) torques .
it is shown that frequency tracking by hidden markov model ( hmm ) methods can be combined with existing maximum likelihood coherent matched filters like the @xmath0-statistic to surmount some of the challenges raised by spin wandering .
specifically it is found that , for an isolated , biaxial rotor whose spin frequency walks randomly , hmm tracking of the @xmath0-statistic output from coherent segments with duration @xmath1d over a total observation time of @xmath2yr can detect signals with wave strains @xmath3 at a noise level characteristic of the advanced laser interferometer gravitational wave observatory ( advanced ligo ) . for a biaxial rotor with randomly walking spin in a binary orbit ,
whose orbital period and semi - major axis are known approximately from electromagnetic observations , hmm tracking of the bessel - weighted @xmath0-statistic output can detect signals with @xmath4 .
an efficient , recursive , hmm solver based on the viterbi algorithm is demonstrated , which requires @xmath5 cpu - hours for a typical , broadband ( 0.5-khz ) search for the low - mass x - ray binary scorpius x-1 , including generation of the relevant @xmath0-statistic input . in a
realistic " observational scenario , viterbi tracking successfully detects 41 out of 50 synthetic signals without spin wandering in stage i of the scorpius x-1 mock data challenge convened by the ligo scientific collaboration down to a wave strain of @xmath6 , recovering the frequency with a root - mean - square accuracy of @xmath7hz .
pacs numbers : : 95.85.sz , 97.60.jd |
the reciprocity relation [ 2,3 ] , @xmath6 , also known as reciprocity law or reciprocity theorem , which links the luminosity distance @xmath2 with the angular diameter distance @xmath1 at redshift z , is a fundamental result for astronomical observations in cosmology .
it is extensively used in various domains ranging from gravitational lensing studies to analyses of the cosmic microwave black body radiation observations , as well as for galaxy and galaxy cluster observations [ 4 - 22 ] . as is known to everyone ,
one fundamental hypothesis in general relativity is that light travels along null geodesics in a riemannian spacetime ( see [ 23 ] for instance ) , based on which the reciprocity law is generally established . actually , the reciprocity relation is also valid for other cosmological models based on riemannian geometry besides the flrw cosmology .
the only condition is that sources and observers are connected by null geodesics in a general riemannian spacetime . up to now
, diverse astrophysical mechanisms such as gravitational lensing and dust extinction have been proved to be capable of causing obvious deviation from the distance duality [ 24 ] and testing this equality with high accuracy can also provide a powerful probe of exotic physics [ 25 ] .
therefore , the verification of the observational validity of the reciprocity law is , perhaps , one of the major challenging open problems in modern cosmology . due to this
, many previous works have been done regarding some attempts to test the identity with diverse model parameterizations of @xmath3 [ 25 - 27 ] . in this paper , we assume that the reciprocity relation can be a constant or a function of the redshift , namely : @xmath7 where @xmath8 denotes a constant around the standard case ( @xmath9 ) and @xmath10 quantifies a possible time - dependent departing from the standard result ( @xmath11 ) .
for instance , bassett & kunz ( 2004 ) [ 25 ] used current supernovae ia data as a measurement of the luminosity distance @xmath12 , and estimated @xmath13 from friib radio galaxies [ 28 ] and ultra compact radio sources [ 9,29,30 ] to detect possible new physics .
meanwhile , observations from sunyaev - zeldovich effect ( sze ) and x - ray surface brightness from galaxy clusters may also provide us a test for the distance duality relation . in order to quantify the @xmath3 parameter , uzan et al .
( 2004 ) [ 26 ] fixed @xmath1 by using the cosmic concordance @xmath14cdm model [ 31 ] with @xmath15 from 18 galaxy clusters [ 32 ] .
they got @xmath16
( 1@xmath17 ) by assuming @xmath3 a constant .
more recently , de bernardis et al . (
2006 ) [ 33 ] have also carried out semblable works using the angular diameter distances from 38 galaxy clusters provided by the sample of ref . [
34 ] , and they obtained @xmath18 at @xmath19 in the context of @xmath14cdm model .
more recently , holanda et al . (
2010 ) [ 27 ] used the reciprocity relation to show how one may assess the cluster geometry by using the constraint provided by the reciprocity relation , with the final result that the elliptical geometry for clusters as advocated by de filippis et al .
is more favorable . in this paper , instead of testing the reciprocity relation , we take it for granted in order to show how we may assess the cluster temperature profiles by using the constraint provided by the reciprocity relation .
a robust method is that we could obtain the observational @xmath10 at redshift @xmath20 from the cluster gas mass fraction , @xmath21 , which could be inferred from x - ray observations of clusters of galaxies .
notice that the presence of temperature profiles plays an essential part in the determination of measured total mass , the fraction of mass in stars and , hence , the gas fraction .
in this context , with two groups of cluster gas mass fraction data from the chandra satellite [ 1 ] , we test the influence on the distance duality relation when different assumptions about the cluster temperature profiles are considered .
our analysis here will be based on two parametric representations defined by eq.([rec ] ) , namely : i. @xmath22 ; ii .
the first expression is on the assumption that @xmath3 does not evolve with redshifts while the second is a continuous and smooth linear one - parameter expansion .
obviously , the above parameterizations are inspired by similar expressions for @xmath24-the equation of state in the constant @xmath25 and time - varying dark energy models ( see ref.[17,35,36 ] ) .
more importantly , for our subsequent analysis , concerning a given data set , the likelihood of @xmath8 or @xmath26 must be peaked ( or @xmath27 minimizes equivalently ) at @xmath9 or @xmath28 in order to satisfy the etherington theorem . as we shall see , assuming the strict validity of the reciprocity theorem , our analysis indicates that the constant temperature assumption tends to be more compatible with no violation of the reciprocity relation . in principle , this kind of result is an interesting example of how a cosmological condition correlates to the local physics .
this paper is organized as follows . in section [ sec : sample ] we present two samples of cluster gas mass fraction data from 52 x - ray luminous galaxy clusters obtained by chandra .
we then briefly describe the analysis method and show the results of constraining parameters including @xmath8 and @xmath26 in section [ sec : analysis ] .
finally , we summarize our main conclusions in section [ sec : conclusions ] .
in order to constrain the possible values of @xmath29 and @xmath30 , we consider two samples of cluster gas mass fraction data from 52 x - ray luminous galaxy clusters obtained by chandra in the redshift range @xmath4 and temperature range @xmath31 kev , one of which is estimated by assuming a constant temperature equal to @xmath32 [ 1 ] , while the other is on the hypothesis of a different temperature profile which reproduces well the temperature gradients in nearby relaxed systems [ 37 ] : @xmath33^{0.45 } } , \label{eq : tr}\ ] ] where @xmath34 .
moreover , ettori et al .
( 2009 ) [ 1 ] found this temperature profile might increase @xmath35 by about 16 percent with respect to the estimates obtained under the assumption of isothermality . because of this , in the following analysis , we make use of the gas mass fraction data obtained with different temperature profiles to limit parameters .
we recall here that in the overall sample of 52 objects at @xmath36 , @xmath37 is assumed , which is also used in the computations throughout this paper . .
the open blue diamonds and red circles with the associated error bars stand , respectively , for the constant and vikhlinin et al . temperature profile samples .
[ t ] ] in fig .
1 we plot two samples of cluster gas fraction data from 52 x - ray luminous galaxy clusters obtained in the redshift range @xmath4 [ 1 ] , which were also used as a proxy for the cosmological parameters such as @xmath38 and @xmath25 [ 1 ] . however , since these samples are endowed with different temperature profile assumptions , our major interest here is to confront these underlying hypotheses with the validity of the reciprocity relation .
thereafter , our attention is paid to the process of extracting @xmath3 from the gas mass fraction .
as is well known , old , relaxed , rich galaxy clusters should provide us with a characteristic sample of the matter content of the universe if they are large enough .
moreover , the baryonic fraction @xmath39 , should remain the same during cosmic evolution if the baryonic - to - total mass ratio in clusters is equal to the cosmological baryonic mass fraction ratio @xmath40 .
since clusters are observed at different redshifts and the reconstructed @xmath39 depends on the assumed distance to the cluster , this data can be used to constrain @xmath3 . following allen et al .
( 2008 ) [ 38 ] , we compute the gas mass fraction given by @xmath41 where five parameters ( @xmath42 , @xmath43 , @xmath44 , @xmath45 , @xmath46 ) are related to modeling the cluster gas mass fraction .
for example , the factor @xmath43 is the angular correction factor and is close to unity for all redshifts [ 38 ] .
.[tab1 ] summary of the standard systematic allowances and priors included in the chandra @xmath39 analysis . [ cols="^,^,^,^",options="header " , ] the depletion parameter @xmath45 indicates the amount of cosmic baryons thermalized within the cluster potential .
this `` bias '' factor was modeled as @xmath47 with the uniform priors @xmath48 and @xmath49 suggested by cosmological simulations [ 38 ] .
more recently , according to various sph and eulerian simulations of a single cluster presented in the santa barbara comparison project , frenk et al .
( 1999 ) [ 39 ] gained @xmath50 .
we adopt , in our paper , the result from the simulated dataset in ref . [ 1 ] : @xmath51 , which is only related to the most massive systems , subjected only to the gravitational heating , high - temperature objects .
the parameter @xmath52 denotes the baryon gas mass fraction in stars .
allen et al . (
2008 ) [ 38 ] modeled it as @xmath53 , using the uniform prior @xmath54 as well as the gaussian prior @xmath55 .
however , recent works found that an additional intracluster light ( icl ) at very low surface brightness might also contribute to the total cold baryonic content of galaxy clusters : @xmath56 [ 40,41 ] .
more recently , lagana et al .
( 2008 ) [ 42 ] discussed in detail the baryonic content of five massive galaxy clusters , including an icl contribution .
they concluded that the stellar - to - gas mass ratio relating to the gas temperature could be expressed through the relation : @xmath57 , where @xmath58 is measured in kev and @xmath59 kev is one selection criterion .
this parametrization was extensively discussed in ref .
[ 1 ] and is also adopted in our work .
@xmath60 is the present dimensionless density parameter of the baryonic matter and the wmap observations give @xmath61 [ 31 ] .
@xmath62 is the angular diameter distance computed in a reference , spatially - flat @xmath14cdm model with @xmath63 ( @xmath64 ) , and @xmath65 is the true angular diameter distance .
the complete set of standard priors and allowances included in the gas mass fraction ( @xmath39 ) analysis are summarized in table [ tab1 ] . with @xmath66 ( see ref . [ 26 ] ) , eq.([eq : fraction ] ) could be rewritten as : @xmath67 in our statistical analysis ( see the next section ) , the observational @xmath68 and its corresponding uncertainty are also calculated through eq.([eq : eta2 ] ) .
let us now estimate the @xmath29 and @xmath30 parameters for both samples in two parameterizations for @xmath69 , e.g. , @xmath70 , @xmath71 . to begin with , we evaluate the likelihood distribution function @xmath72 , where @xmath73 } ^{2}}{\sigma^2_{\eta^{3}_{obs } } } , \ ] ] with @xmath74 gained from the @xmath39 data [ eq.([eq : eta2 ] ) ] .
as for @xmath75 , the uncertainty of @xmath74 , is calculated through the propagation of uncertainty statics with a combination of the observational @xmath39 , the best fit values of parameters such as @xmath42 , @xmath44 as well as their corresponding standard systematic allowances ( fig .
1 and table [ tab1 ] ) . plane for the parameterization @xmath76 .
the filled ( green ) circles and filled ( red ) diamonds stand for the constant and vikhlinin et al .
temperature profile samples , respectively .
[ 1 ] ] plane for the parameterization @xmath77 .
the denotation of different data is the same as fig.[1 ] .
[ 2 ] ] planes for two parameterizations @xmath76 and @xmath77 with marginalized @xmath39 parameters .
the blue stars and filled purple diamonds stand for the constant and vikhlinin et al .
temperature profile samples , respectively .
[ 2.1],title="fig : " ] planes for two parameterizations @xmath76 and @xmath77 with marginalized @xmath39 parameters .
the blue stars and filled purple diamonds stand for the constant and vikhlinin et al .
temperature profile samples , respectively .
[ 2.1],title="fig : " ] in fig . 1 and 3
, we plot the likelihood distributions on the @xmath78 and @xmath79 planes for both samples .
we obtain @xmath80 ( @xmath81 ) and @xmath82 ( @xmath83 ) at @xmath84 cl for the vikhlinin et al .
temperature profile sample and @xmath85 ( @xmath86 ) and @xmath87 ( @xmath88 ) at @xmath84 cl for the constant temperature sample .
we can see that the @xmath3 from the latter is in agreement with no violation of the reciprocity relation while the former , where a vikhlinin et al .
temperature profile is assumed to describe the clusters , is incompatible even at 99% cl .
therefore , we have not found any evidence for distance duality violation when the constant temperature sample is considered .
however , the same kind of analysis shown in fig . 2 and 3 points to be in contradiction with the vikhlinin et al .
temperature profile hypothesis assumed in the vikhlinin et al . sample .
furthermore , considering that many parameters listed in table [ tab1 ] are fixed to their best - fit values , we perform a mcmc analysis and marginalize over these nuisance parameters [ 43 ] .
the constraint results shown in fig .
4 further confirm our conclusions .
in this paper , we have explored the consequences of the distance duality relation @xmath89 based on two samples of cluster gas mass fraction data from 52 x - ray luminous galaxy clusters obtained by chandra in the redshift range @xmath4 and temperature range @xmath5 kev [ 1 ] .
we discussed the consistency between the strict validity of the distance duality relation and the assumptions about temperature profiles ( the constant and vikhlinin et al .
temperature profiles ) used to describe the galaxy clusters .
the @xmath3 parameter was parametrized in two different functional forms , @xmath70 and @xmath90 . by comparing the constant and vikhlinin et al .
temperature profile samples , we show that the constant temperature profile is more consistent with no violation of the distance duality relation in the context of two groups of cluster gas mass fraction data . in the case of constant temperature sample ( see fig .
2 and 3 ) we find @xmath91 and @xmath92 for constant and linear parametrizations , respectively . on the other hand , the vikhlinin et al
. temperature profile model ( see fig .
2 and 3 ) seems to be marginally incompatible with @xmath80 and @xmath82 for constant and linear parameterizations , respectively .
our analysis indicates that the constant temperature assumption tends to be more compatible with the etherington theorem at 68.3% cl ( between 68.3% and 95.4% cl for the first parametrization ) compared with the vikhlinin et al .
temperature profile hypothesis , while the latter is incompatible even at 99% cl . the results with marginalized @xmath39 parameters ( fig .
4 ) further confirm these conclusions . in summary , we find that , according to the statistical analysis presented here , the constant temperature profile of clusters tends to be more consistent with no violation of the distance duality relation regarding two groups of cluster gas mass fraction data .
this reinforces the interest in the observational search for such kind of data from clusters at high redshifts . with better data ,
the method proposed here based on the validity of the distance duality relation should lead to improved limits on the temperature profiles of gas in galaxy clusters .
this work was supported by the national natural science foundation of china under the distinguished young scholar grant 10825313 and grant 11073005 , and by the ministry of science and technology national basic science program ( project 973 ) under grant no . | the validity of distance duality relation , @xmath0 , an exact result required by the etherington reciprocity theorem , where @xmath1 and @xmath2 are the angular and luminosity distances , plays an essential part in cosmological observations and model constraints . in this paper , we investigate some consequences of such a relation by assuming @xmath3 a constant or a function of the redshift . in order to constrain the parameters concerning @xmath3
, we consider two groups of cluster gas mass fraction data including 52 x - ray luminous galaxy clusters observed by chandra in the redshift range @xmath4 and temperature range @xmath5 kev , under the assumptions of two different temperature profiles [ 1 ] .
we find that the constant temperature profile is in relatively good agreement with no violation of the distance duality relation for both parameterizations of @xmath3 , while the one with temperature gradient ( the vikhlinin et al .
temperature profile ) seems to be incompatible even at 99% cl . |
in recent years , there has been an increasing interest in possible use of improved actions to lattice qcd .
there exist different improvement schemes for lattice actions and observables , all which aim at the extraction of continuum physics from simulations on coarse lattices .
the simplest of these improvement schemes is the clover action , an @xmath1 improvement of the wilson fermion action suggested by sheikholeslami and wohlert @xcite , who implemented the symanzik improvement @xcite for the dirac operator .
several calculations in quenched and some in full qcd have been performed using this action especially to calculate both heavy and light hadron masses . the results obtained seem promising , so that one might want to check the improvement effects of the clover action in a simple and well - known toy model . here
we study the effects of the @xmath1 improvement of the wilson action in the full lattice schwinger model .
this model seems optimally suited for this purpose .
massless qed in @xmath4 dimension was first studied by schwinger @xcite as an example of an explicitly solvable qft . in subsequent papers
@xcite the model was generalized to allow for fermion masses and include different flavors .
the schwinger model became of considerable interest during the early 1970 s , because it shows some remarkable features like fermion confinement or a @xmath5-vacuum structure which are reminiscent of qcd .
for the purpose of this paper , it is sufficient to note , that the massless @xmath6-flavor schwinger model has a spectrum which consists of one massive isosinglet boson and a isomultiplet of @xmath7 massless fermion - antifermion bound states ( for further discussion cf .
e.g. @xcite ) .
the mass of the isosinglet state , the schwinger mass ( in lattice units ) is given by @xmath8 where @xmath9 denotes the dimensionless gauge coupling .
in the next section some technical details about the definition and the numerical simulation will be given and in sec .
3 we discuss the effect of the improvement on the rotation invariance of the propagator , the phase diagram and the dispersion relations of the schwinger boson .
the fermionic part of the wilson action for the massive , n - flavor schwinger model is @xmath10 the fields have @xmath11 components , @xmath12 are the pauli matrices . this action has a classical @xmath1 error introduced by the wilson term , which can be removed @xcite by rotating the fermion fields @xmath13 and adding a term @xmath14 to the action .
@xmath15 denotes the clover - leaf discretization of the field strength tensor .
the resulting lattice dirac operator now has discretization errors of @xmath16 .
we note in passing that for this model one may introduce a topological charge in its ` geometric ' definition @xmath17 @xcite .
thus @xmath18 introduces a local coupling of @xmath19 to @xmath20 .
the coefficient @xmath21 , which generally has to be determined non - perturbatively at finite @xmath22 , is just @xmath23 in 2d , since the gauge coupling @xmath24 is dimensionful and therefore @xmath25 , which only gives corrections to higher order in @xmath26 .
since the gauge part of the wilson action @xmath27 has @xmath16 discretization errors , the total action @xmath28 together with ( [ frot ] ) gives a lattice version of the schwinger model with discretization errors of @xmath16 .
we have done a mc simulation of the schwinger model , using both the wilson and the sw action , with two degenerate flavors of dynamical fermions .
we used a hybrid monte carlo algorithm with 10 integration steps and a trajectory length tuned for a @xmath29 acceptance rate of the mc step . for fermion matrix inversion we used a standard conjugate gradient algorithm , and there was no problem with its convergence or stability .
in that algorithm the doubling of the fermionic species is required in order to guarantee positivity of the fermionic determinant one of our reasons to study the schwinger model with two flavors . in our implementation
, the speed loss when including the clover term into the action is consistently @xmath30 .
depending on the lattice size ( from @xmath31 up to @xmath32 ) and couplings , we skipped @xmath33 to @xmath34 configurations between measurements and finally had between @xmath35 and @xmath36 independent configurations . to check for ergodicity
, we measured the topological charge ( using the geometric definition ) and observed , that in each run the system tunneled several times between different topological sectors .
the additional term in the action affects the critical @xmath37 value where the theory has effectively massless fermions .
since the clover action is supposedly closer to the continuum ( has smaller corrections to leading scaling behavior ) , we might expect less renormalization of @xmath38 with a value closer to its continuum value of @xmath39 even for small @xmath22 . in order to determine @xmath40 we follow suggestions @xcite to utilize the pcac relation as discussed in @xcite for the schwinger model . at each @xmath22
we measured an observable proportional to the effective fermion mass for @xmath41 different @xmath37 values and determined @xmath38 .
our result for a @xmath31 lattice is compared with the result for the original wilson fermion action in fig .
[ phd ] .
the values @xmath38 for the clover action are in fact closer to @xmath39 .
however , both , wilson and clover action have corrections @xmath42 the dependence on the finite lattice volume is weak and for the wilson action is discussed elsewhere @xcite .
the spectral distribution of the dirac operator is of some interest with regard to topological and chiral properties of the system .
recent studies for the schwinger model @xcite and for the sw - action in 4d @xcite have emphasized these aspects .
[ swspectrum ] should be compared with typical spectra for the pure wilson dirac operator ( cf .
e.g. @xcite ) . as compared to the wilson action with its @xmath43 and @xmath44 symmetries we observe for the sw operator an agglomeration of eigenvalues at larger distances from 0 .
this feature may improve somewhat the separation of the low - lying states from the doubling states already at moderate @xmath22 .
for the continuum limit the distribution density of small eigenvalues is of relevance .
however , the distribution is not much closer to the circular shape one obtains for fixed point actions @xcite or neuberger - projected actions @xcite .
we can not identify a clear signal of improvement in this respect for the sw dirac operator .
we measured the propagators of the pseudoscalar isotriplet ( @xmath45 ) mesonic bound states . in the 2-flavor model these modes are expected to be massless at @xmath38 . for short distances on the lattice ,
the rotation invariance of these propagators is clearly broken when using wilson action .
generally , improved actions are expected to show better rotation invariance ( cf .
the schwinger model fp action study in @xcite ) .
one must be careful not to compare these quantities simply at the same values of @xmath22 and @xmath37 , but at values corresponding to comparable discretization of the same continuum theory .
we choose to compare the propagators at @xmath46 and the respective @xmath38 , so that we have a discretization of the massless ( fermion ) theory in both cases .
the results are shown in fig.[rot ] .
we find no noticeable improvement of rotation invariance when using the clover action .
the relative error of the diagonal @xmath47 propagator compared to the @xmath48 and @xmath49 propagators is @xmath50 for both actions .
this is the first indication , that short distance ( high momentum ) observables are not improved .
[ disp ] shows the dispersion relations for the massive ( isosinglet vector ) and the massless ( isotriplet vector ) meson at @xmath46 and @xmath51 on a @xmath52 lattice .
no improvement for the high momentum behavior is observed . concerning the mass of the massive boson
, however , the improvement is significant .
[ mass ] shows the ratio of the observed schwinger mass over its continuum value at @xmath53 on a @xmath52 and @xmath54 , @xmath51 on a @xmath32 lattice . assuming the continuum scaling behavior @xmath55 these are different discretizations of the same physical system size @xmath56 .
the improvement when including the clover term is quite obvious .
the scaling corrections are remarkably smaller for the clover results .
this behavior , the considerable improvement of low momentum states and the lack thereof for high momentum states , is no surprise when one considers , that the construction of symanzik improved actions aims at improving the fermion dispersion relations for low momenta ( cf .
the overall improvement of the fixed point action in @xcite ( which , however has more than hundred terms per site in the action ) is definitely more pronounced and extends to high momenta as well .
we performed a monte carlo simulation on the 2-flavor lattice schwinger model , comparing standard wilson to an @xmath1 improved sheikholeslami - wohlert action .
we found the following results : * the critical value of the hopping parameter @xmath38 moves closer to its continuum value @xmath39 .
* the eigenvalue spectrum of the lattice sw dirac operator changes but does not appear to improve significantly towards , e.g. a circular shape .
* there was no improvement in the rotation invariance of meson propagators . for the meson dispersion relations
, there is no improvement in the high momentum states .
* there is considerable improvement in the scaling behavior of the schwinger mass of the massive boson .
these observations are consistent with the fact that the sw action is constructed as an improved action for low momentum fermionic states @xmath57 . while there is considerable improvement in low momentum observables , high momentum observables are largely unaffected by the addition of the clover term .
for the full fermion qcd simulations this suggests that it is preferable to use the clover action as long as one is only interested in the mass spectrum of the theory . on the other hand
, some observables will likely show no improvement at all , especially those connected with high momentum states .
we would like to thank prof .
claudio rebbi and ivan hip for many helpful discussions .
c.b.l . also wants to thank claudio for the kind hospitality at boston university and wants to acknowledge support by heinrich - jrg - stiftung , univ .
the work was supported by fonds zur frderung der wissenschaftlichen forschung in sterreich , project p11502-phy . | we perform a model study on the @xmath0-flavor lattice schwinger model using standard wilson and @xmath1 improved sheikholeslami - wohlert ( sw ) action .
we find , that the phase diagram is altered , the critical line shifted closer towards its continuum value @xmath2 .
we find no improvement in the rotation invariance of meson propagators ; the scaling of the schwinger mass is considerably improved , high momentum states are not .
the additional cost of @xmath3 cpu - time is highly justified when calculating masses , but not for high momentum observables .
pacs : 11.15.ha , 11.10.kk + key words : lattice field theory , schwinger model , clover ( sw ) improved dirac operator , eigenvalue spectrum , scaling behaviour |
type iin supernovae ( sne iin ) have raised important questions about the latest phases of evolution in massive stars , because they require enhanced or episodic mass loss shortly before core collapse that far exceeds known examples of steady winds ( see @xcite for a general review of mass loss and its connection to interacting sne ) .
sne iin have relatively narrow lines of hydrogen in their spectra ( see @xcite for a review of sn spectral classification ) caused by slow - moving circumstellar material ( csm ) that is hit by the blast wave or illuminated by ultraviolet ( uv ) radiation from the sn . in some extreme cases , as much as 1025m@xmath3 seems to have been ejected in just the decade before core collapse @xcite . in other cases ,
the luminosity enhancement is more modest , but the csm is still dense enough to correspond to the strongest known winds from extreme and unstable red supergiants ( rsgs ; e.g. , @xcite ) .
sne iin are about 89% of all observed core - collapse sne @xcite , so the physical trigger of the most violent precursors only operates in a subset of core - collapse events .
a key issue is the timing of this mass loss .
are these eruptions synchronized to go off just before the sn owing to instabilities in the final ne , o , and si burning stages of their progenitors @xcite ?
we suspect that pre - sn instability may be more widespread than indicated by the observed fraction of traditional sne iin noted above , since these may be only the most extreme manifestation of a more generic phenomenon @xcite . at a less extreme level , some sne show type iin signatures for just a day or so after explosion @xcite , which can only be seen if they are discovered early . at the other extreme , major mass - loss episodes also seem to precede core collapse by centuries or millennia . in the latter case , it could take months or years for the blast wave to reach the csm .
this may often go undetected as follow - up sn observations rarely extend past the first few months because of sensitivity limits or observers compulsion to turn their attention to newer sne .
there are some observed cases of significantly delayed onset of csm interaction .
most famously , sn 1987a began to collide with its csm ring nebula after roughly a decade @xcite , and in this case velocities of the ring nebula indicate that it was ejected 1020 thousand years prior @xcite .
this interaction was relatively modest compared to that of sne iin , however , and did not lead to a huge increase in luminosity that would have been observable in a distant galaxy .
an interesting case is sn 2008iy , which did show a large increase in luminosity after a delay .
initially it had an absolute magnitude of about @xmath4 , similar to a normal sn ii - p with no csm interaction , but then brightened to about @xmath5mag after 400 days as the sn caught up to a massive shell ejected a century earlier @xcite .
sn 2001em and more recently sn 2014c represent cases in which a stripped - envelope explosion crashed into a h - rich shell after a year or more and developed strong , relatively narrow h@xmath0 emission @xcite .
infrared ( ir ) echoes from csm dust illuminated by the sn itself @xcite or by ongoing interaction ( see , e.g. , @xcite ) also provide evidence for massive distant shells .
an extreme case is sn 2006gy , which in addition to its @xmath6m@xmath3 shell ejected only 8yr before core collapse @xcite also had another more distant shell of 1020m@xmath3 ejected @xmath2yr earlier @xcite .
yet another case of a very massive dust shell seen as an echo is sn 2002hh @xcite , although interestingly , this aging sn does not yet show signs of shock interaction with this shell even though it probably should have reached it by now @xcite . for sn 2005ip , the sn considered in this paper , @xcite interpreted the ir echo properties as indicating a past major mass - loss outburst that was followed by a less intense wind .
we therefore have mounting evidence of major mass - loss episodes that precede core collapse by centuries or millennia , and not just in the few years beforehand .
if the precursor outbursts of sne iin in the decade prior to core collapse can be linked to final nuclear burning phases @xcite , then what can give rise to older mass ejections ?
wave - driven mass loss during ne and o burning can only drive significant mass loss for a few years before core collapse , according to current models @xcite .
additional observations are needed to constrain the physical parameters and time dependence , but there is mounting evidence of enhanced mass loss on much longer timescales than ne and o burning .
@xcite notes that the pulsational pair mechanism can produce mass - loss eruptions that precede core collapse on a variety of timescales , but these come from very massive stars and should be relatively rare ( and the eruptions might not be punctuated by an energetic sn , as material falls back to a black hole ) .
this poses an interesting observational question : what fraction of sne ( including otherwise normal sne ) have much more extended dense shells ?
for this reason , we have been monitoring some old , nearby sne to look for changes in their late - time interaction . in this paper
we revisit the aging event sn
2005ip , an unusual sn iin located in the host galaxy ngc 2906 at a distance of @xmath730 mpc . @xcite presented the optical photometric and spectroscopic evolution over the first @xmath8d , while @xcite discussed the ir photometric evolution over the same time period .
sn 2005ip was notable among sne iin for its sustained strong csm interaction at late times , its rather unusual high - ionization spectrum with strong coronal emission lines and a strong blue pseudocontinuum , and its strong ir emission from dust .
sn 2005ip showed some evidence for the formation of new dust at these early times @xcite , although @xcite concluded that the ir emission was likely produced by a combination of new dust formation and an ir echo from pre - existing csm dust .
@xcite considered distant dust shells of various radii to explain the ir echo , and @xcite favoured a dust shell located @xmath9pc from the star to explain the ir echo evolution .
this is similar to the radius where we infer late - time csm interaction in sn 2005ip , so it is possible that the variability of h@xmath0 emission that we report here is caused by the ejecta hitting shells that were seen as an ir echo at earlier times .
this is discussed below .
overall , the early - time data suggested a scenario wherein the sn blast wave was interacting with a very dense and clumpy rsg wind having a mass - loss rate of more than @xmath10m@xmath3yr@xmath11 and likely around @xmath12m@xmath3yr@xmath11 @xcite .
this points to an extreme rsg akin to vy cma , rather than a normal rsg like @xmath0 ori @xcite .
subsequent papers followed the continuing evolution of sn 2005ip as it faded during the decade after explosion . by 18002400d after discovery , the h@xmath0 flux had dropped by about a factor of 10 @xcite , although this was still much stronger than typical interacting sne at a comparable epoch .
@xcite also drew similar conclusions as @xcite concerning the evidence for dust formation and the nature of the progenitor s mass loss , and @xcite mentioned that the x - ray and h@xmath0 fluxes seem to provide consistent results .
@xcite also presented a late - time ( day 3024 ) optical spectrum of sn 2005ip , showing evidence for continued interaction .
@xcite studied the evolution of x - ray emission from sn 2005ip during the @xmath13d after discovery .
they found that the x - ray luminosity was roughly constant until 2009 and dropped by a factor of two by 2012 .
they also showed that the absorbing column density ( @xmath14 ) appeared to be steadily declining during the same time period .
this prompted @xcite to suggest that the progenitor had ejected a massive csm shell in the centuries before explosion , and that after @xmath15yr , the blast wave was finally emerging from this shell , which they estimated to have a total mass of @xmath16m@xmath3 swept up by that time .
this implied that the epoch of csm interaction was coming to an end in sn 2005ip . here
we show that csm interaction clearly has not yet finished .
the csm interaction intensity , traced by h@xmath0 luminosity , continued to decline very slowly until about mid 2015 .
after that time , however , the h@xmath0 showed abrupt variations , including a brightening by late 2015 and early 2016 .
the radio and x - ray emission was also bright in early 2016 , as described below .
this indicates that the blast wave has continued to run into a distant csm shell or torus .
we explore possible explanations for the origin of this csm and implications for the pre - sn evolution of massive stars .
for comparison , we also discuss similar continued interaction via late - time h@xmath0 emission seen in newly obtained spectra of sne 1988z , 1993j , and 1998s .
.late - time optical spectroscopy of sn 2005ip [ cols="<,>,^,^,^ " , ] [ tab : radio ]
we obtained late - time optical spectroscopy of sn 2005ip over several epochs since our previous study @xcite .
these observations included the bluechannel ( bc ) spectrograph on the 6.5 m multiple mirror telescope ( mmt ) , the multi - object double spectrograph ( mods ; @xcite ) on the large binocular telescope ( lbt ) , the low - resolution imaging spectrometer ( lris ; @xcite ) mounted on the 10 m keck i telescope , and the deep imaging multi - object spectrograph ( deimos ; @xcite ) on keck ii .
details of the spectral observations including measurements of the total h@xmath0 line flux are summarized in table [ tab : spec ] , and the spectra are shown in figure [ fig : spectra ] .
standard reductions were carried out using iraf including bias subtraction , flat - fielding , and optimal spectral extraction ( see , e.g. , @xcite ) .
flux calibration was achieved using spectrophotometric standards observed at similar airmass to that of each science frame , and the resulting spectra were median combined into a single one - dimensional spectrum . for one observation noted in table [ tab : spec ] , the flux calibration was suspect owing to clouds that came in during subsequent exposures . for this date ( day 3425 ) we increase the uncertainty in figure [ fig : lha ] to @xmath17% . _
sn 1988z : _ in our study , we find it useful to compare the evolution of sn 2005ip to that of a few other well - studied , geriatric sne with late - time interaction .
sn 1988z is a prototypical , long - lasting , and radio - luminous sn iin that has been studied in detail by several authors @xcite . in particular , @xcite have compiled multiwavelength data , including h@xmath0 fluxes , up to day @xmath73000 . we have obtained additional unpublished late - time spectra of this aging sn iin using the keck observatory , with both lris and deimos .
these observations , including estimated h@xmath0 line fluxes , are summarized in table [ tab:88z ] and almost triple the time baseline for this object .
this series of new late - time spectra of sn 1988z is shown in figure [ fig : spectra88z ] .
these spectra have been corrected for a redshift of @xmath18 , but no reddening correction has been applied as in previous studies @xcite . when we plot the h@xmath0 line luminosity in figure [ fig : lha ] , we first subtract a constant flux of @xmath19ergs@xmath11@xmath20 , which is the mean value of the narrow h@xmath0 + [ n ii ] line fluxes in late - time spectra ( this is only significant for the last epochs ) .
we attribute this constant narrow emission to contamination by coincident or nearby h ii region emission , or perhaps unshocked distant csm .
since we compare the h@xmath0 luminosity of sn 2005ip to sn 1988z and also to other well - studied late interactors , we obtained late - time spectra to extend the temporal coverage of two additional prototypical sne . _
sn 1993j : _ we procured several late - time spectra of sn 1993j , most recently on 2016 feb .
10 using lris on keck i. sn 1993j was a type ib event rather than a type iin , but it was very nearby in m81 , was discovered within about 1 day of explosion , and has shown strong , long - lived csm interaction that has given rise to a high radio luminosity and prominent optical / uv emission lines for more than a decade after explosion @xcite .
table [ tab:93j ] summarizes our late - time spectra reported here , including a measurement of the total h@xmath0 line flux in the calibrated spectra . here
the uncertainty is dominated by the fact that the blue side of h@xmath0 is blended with a strong oxygen line , as well as uncertainty from possible slit losses .
qualitatively , the spectral appearance of sn 1993j changes very little at these late epochs after day 4000 , and even resembles the day 2454 spectrum presented by @xcite .
the most recent epoch is illustrated in the bottom panel of figure [ fig : spectra88z ] in dark green ; we also show two earlier epochs on days 4249 and 6159 in a lighter shade of green for comparison .
although some of the emission lines have weakened with time , the line profiles and overall appearance of the spectrum have not changed much .
_ sn 1998s : _ this has long been considered a prototypical sn iin , which was discovered within about a day of explosion , but its early csm interaction was on the weak side compared to most sne iin @xcite .
sn 1998s did , however , exhibit strong continued csm interaction at late times @xcite .
we obtained a late epoch of sn 1998s on 2016 mar .
1 using deimos on keck ii , extending the spectral evolution by @xmath21d compared to published data . the last previously published spectrum of sn 1998s @xcite was obtained on day 5079 , and our new spectrum corresponds to day 6574 .
qualitatively , the appearance of this spectrum shows little change compared to the last one published by @xcite . for this spectrum , however , we measure a strong h@xmath0 line flux of @xmath22ergs@xmath11@xmath20 , which is several times higher than our previous flux measurement @xcite .
the uncertainty is likely to be dominated by absolute flux calibration and slit placement as noted above for sn 2005ip and sn 1988z , but this increase appears to be significant .
the new spectrum of sn 1998s is shown in the bottom panel of figure [ fig : spectra88z ] .
we observed the position of sn 2005ip with the karl g. jansky very large array ( vla ; program 16a-101 , pi kilpatrick ) starting on 2016 feb . 4 ( day 3743 after discovery ) at 09:47:29 ( ut dates are used throughout this paper ) , at mean frequencies of 6.2 ghz and @xmath23 ghz ( with side - bands centred at 5.0 , 7.4 , 8.5 , and 11.0 ghz ) .
we obtained 14min of on - source time for each frequency , and used 3c286 and j0925 + 0019 for bandpass / flux and gain calibration , respectively .
following standard procedures in the astronomical image processing system ( aips ; @xcite ) for data calibration and analysis , we detect a bright source located at @xmath0 = , @xmath24 = @xmath25 ( j2000 ; @xmath26 , @xmath27 ) , consistent with the optical position .
we measure flux densities and @xmath28 uncertainties for the upper and lower side - bands using aips/ jmfit .
the radio detections are listed in table [ tab : radio ] .
the radio spectral index of sn 2005ip is @xmath29 , which is consistent with optically thin emission observed from sne ii @xcite .
we postpone a more detailed analysis of the radio data to a later paper that will examine its temporal variability . the x - ray telescope ( xrt ; @xcite ) onboard the _ swift _ satellite @xcite started observing sn 2005ip on 2007 feb .
14 ( 475 days since discovery ) .
late - time observations of sn 2005ip have been carried out until 3630.5 days , as part of a _ swift _ fill - in program ( pi margutti ) .
xrt data have been analyzed using heasoft ( v6.18 ) and corresponding calibration files , following standard procedures ( see @xcite for details ) .
in particular , our rebinning scheme requires a minimum of 20 photons in the source region and an inferred source count rate at least @xmath30 above the background level to yield a detection .
compared to the previous compilation by @xcite , our campaign extends the x - ray monitoring of sn 2005ip from @xmath13 days to @xmath31 days since explosion , and reveals a continuation of the fading x - ray emission with time .
the resulting luminosity light curve is shown in fig .
[ fig:05iplx ] .
-0.0 true cm for the flux calibration of _ swift_-xrt data acquired at @xmath32 days we employ the spectral parameters of the one - component model derived by @xcite ( their table 2 ) with a shock temperature evolution @xmath33 .
swift_-xrt data acquired at @xmath34 days we adopt the spectral parameters constrained by our latest _ chandra _ spectrum ( see below ) .
figure [ fig:05iplx ] shows the x - ray evolution of sn 2005ip in the 0.310kev energy range as captured by the _ swift_-xrt and _ chandra _ in the first @xmath35 days since explosion .
the steep decline ( @xmath36 ) at late times suggests that the shock is sampling a steeply decaying environment density profile @xmath37 with @xmath38 @xcite .
-0.0 true cm sn 2005ip was observed with the _ chandra x - ray observatory _ ( cxo ) and advanced ccd imaging spectrometer ( acis ) on 2016 apr .
3 ( 3812 days ) under a single - epoch ddt program ( pi mauerhan , i d 18802 ) .
the acis - s array was used and the total on - source exposure time was 35.59 ks .
photometry and energy spectra were extracted using the ` specextract ` package within the heasoft _ ciao _ software suite .
the point source associated with sn 2005ip yielded 820 counts within a circular aperture 2 in radius .
a 5 radius background annulus surrounding the source aperture yielded 121 counts .
the energy spectrum of the source is shown in figure [ fig : cxo_spec ] .
note the apparent detection of fe k@xmath0 emission near 6.8 kev .
the source and background spectra were modeled simultaneously using the _ sherpa _ package .
an absorbed single - temperature thermal plasma model ( _ apec _ ) having solar abundances as defined by @xcite was fit to the source for photon energies in the range 0.58.0 kev , and a simple power law was used to fit to the background .
the he , c , n , o , and fe elemental fractions , by number relative to h , are @xmath39 , @xmath40 , @xmath41 , @xmath42 , and @xmath43 , respectively .
c_-statistic was used in the fitting .
we fixed an equivalent interstellar neutral hydrogen column density of @xmath44@xmath20 , adopting this value from @xcite , and allowed for an additional intrinsic source of absorption for the sn .
our best fit yielded @xmath45@xmath20 and a plasma temperature of @xmath46kev .
the associated uncertainties are 90% confidence envelopes ( @xmath47 ) .
the absorbed energy flux is @xmath48ergs@xmath11@xmath20 ( 0.58.0 kev ) and the unabsorbed model flux is @xmath49ergs@xmath11@xmath20 . for our adopted distance
the flux implies an x - ray luminosity of @xmath50ergs@xmath11 for the sn at epoch 3812 days .
we also experimented with fitting the data using a @xmath51 statistic with a gehrels variance function , and obtained a reduced @xmath51 value of 43.6 for 65 degrees of freedom .
the resulting physical parameters are consistent with those from the @xmath52 statistic , quoted above .
the model fit is also shown in figure [ fig : cxo_spec ] .
line profile , with velocity relative to the centroid of the narrow component .
this shows two of the higher - resolution spectra obtained with mmt / bluechannel ( day 2249 ) and keck / deimos ( day 3770 ) , before and after the dip in 2015 , respectively.,width=288 ]
figure [ fig : spectra ] shows our newly obtained series of late - time optical spectra of sn 2005ip , and figure [ fig : halpha ] displays the h@xmath0 line profile in a few epochs with the best signal - to - noise ratio .
in addition to h@xmath0 , the spectrum shows consistent emission from other lines commonly seen in late - time spectra of strongly interacting sne iin .
here we focus mainly on the h@xmath0 line , since it is the dominant line in the spectrum of sne iin observed at late times , and is a good proxy for the strength of interaction . our higher - resolution spectra ( i.e. , all but the 2014 may and 2015 nov .
spectra ) are able to cleanly resolve the intermediate - width emission of the post - shock gas from the narrow pre - shock csm or underlying h ii region emission ( fig .
[ fig : halpha ] ) . in these spectra
, the intermediate - width component of h@xmath0 ( within about @xmath53 km s@xmath11 ) has a consistent profile shape , which is asymmetric and shifted to the blue .
the peak of this emission is located roughly at @xmath54 to @xmath55 km s@xmath11 , shifting toward more positive velocities with time .
this asymmetric blueshifted profile was seen at early times in the broad and intermediate - width lines @xcite , and was attributed to new dust formation in the sn ejecta or in the post - shock gas .
blueshifted lines with this type of shape can arise simply from occultation of the far side by the sn photosphere at early times @xcite , but the fact that the blueshifted asymmetry persists for a decade in sn 2005ip , long after the continuum opacity of the photosphere has vanished , means that the blocking of the red side of the lines must be caused by dust mixed within the sn ( or to intrinsically asymmetric csm ) .
much of the dust is likely to be mixed in the post - shock region where the intermediate - width h@xmath0 originates , since even early - time spectra showed a blueshift in lines formed within the post - shock shell , as well as in the broad sn ejecta lines @xcite .
the narrowest components from the pre - shock csm are unresolved or only marginally resolved in our spectra , indicating expansion speeds of 80 km s@xmath11 or less ( fig .
[ fig : halpha ] ) .
the temporal evolution of the h@xmath0 line luminosity is shown in figure [ fig : lha ] .
this plot uses h@xmath0 line fluxes in the first 1000 d from our earlier paper @xcite , and supplements these with our new spectra of sn 2005ip reported here .
the line fluxes were converted to h@xmath0 luminosity with the same assumptions as in our earlier study .
the h@xmath0 luminosity remains roughly constant for the first 1000 d , as we noted earlier , and then declines slowly over the next several years as found previously by @xcite .
our new spectra show that this slow decline continued to around day @xmath31 ( mid 2015 ) , when the h@xmath0 luminosity was in decline , and then suddenly reversed to be on a quick rise that has continued to the time of writing .
additional monitoring will be needed to determine how high it will rise , if it levels off , or fades again .
possible causes of this are discussed further in section 4 .
figure [ fig : lha ] also compares the h@xmath0 luminosity of sn 2005ip to that of several other sne with strong late - time interaction .
we see here that sn 2005ip really is quite unusual , with a sustained h@xmath0 luminosity that was essentially the same as the current record holder for such interaction : sn 1988z ( see below ) . at days 10002000 ,
the h@xmath0 luminosity was 2 orders of magnitude stronger than in some other well - studied objects like sn 1998s @xcite , sn 1980k @xcite , and sn 1993j @xcite , and an order of magnitude stronger than in the sn iin ptf11iqb @xcite .
the late - time csm interaction in these other objects can be explained by a very strong rsg wind with @xmath56 m@xmath3 yr@xmath11 , but sn 2005ip and sn 1988z are clearly more extreme situations .
thus , the h@xmath0 luminosity implies that around 4000 days after explosion , the shock is interacting with a wind that had a mass - loss rate of at least @xmath12 m@xmath3 yr@xmath11 . with a csm speed resolution limit of our data , and similar to csm speeds for sn 1998s seen in echelle spectra @xcite .
] around 40 km s@xmath11 , this implies that the high mass - loss rate was occurring within a few thousand years before core collapse .
this , in turn , requires several m@xmath3 being lost from the star in the last few millennia of its life .
our estimate is conservative compared with some others , which favour higher rates of @xmath57 m@xmath3 yr@xmath11 and total csm masses greater than 10 m@xmath3 @xcite . in our earlier paper @xcite
, we drew comparisons between sn 2005ip and sn 1988z , including a direct comparison of the h@xmath0 luminosity . during the first several hundred days ,
sn 2005ip was significantly less luminous in both h@xmath0 and continuum emission .
however , during this time sn 2005ip remained roughly constant in h@xmath0 luminosity , while sn 1988z steadily declined .
figure [ fig : lha ] extends this same comparison , and shows that by day 1000 the two objects had about the same luminosity , and for the rest of its first decade , sn 2005ip was an almost exact twin of sn 1988z in terms of its csm interaction strength traced by h@xmath0 .
as in our previous paper , figure [ fig : lha ] incorporates h@xmath0 luminosities for sn 1988z published by @xcite , extended to almost day 3000 by @xcite .
we also supplement these published h@xmath0 luminosities with measurements from our own spectra of sn 1988z ( see table [ tab:88z ] ) . with its recent resurgence in h@xmath0 luminosity ,
sn 2005ip equals sn 1988z again at a comparable epoch , making it tied as the record holder for the highest late - time h@xmath0 luminosity at + 10yr . integrating through time in figure [ fig : lha ] , the total energy radiated by sn 2005ip in the h@xmath0 line alone is @xmath58erg . for comparison ,
the total integrated 0.310kev x - ray energy is about a factor of 10 larger at @xmath59erg , and the integrated optical luminosity yields a similar total radiated energy of @xmath59erg .
this is certainly an underestimate of the total radiated energy , since we appled no correction for flux at other wavelengths .
the asymmetric blueshifted profiles of h@xmath0 , in particular , suggest some dust extinction in sn 2005ip ( with a lack of such evidence for dust extinction in sn 1988z ) ; correcting for this could raise the h@xmath0 and continuum luminosities further , and thus make sn 2005ip significantly more luminous than sn 1988z .
the radiated energy so far is therefore likely to be at least @xmath60erg for sn 2005ip .
the kinetic energy imparted to the swept - up csm shell ( about 10m@xmath3 accelerated to @xmath61kms@xmath11 ) is roughly @xmath62erg .
thus , csm interaction may have already tapped a substantial fraction of the explosion energy in sn 2005ip ; remaining kinetic energy of the freely expanding ejecta that are still inside the reverse shock may power sn 2005ip for the next decade
. it will be interesting to see where it goes from here .
a significant detail is that after it rebrightened , the spectrum showed stronger emission from the intermediate - width components of [ o i ] @xmath636300 , 6364 , which had only shown narrow pre - shock csm emission before that time .
this may hint that we are starting to see an increased contribution from the reverse - shock luminosity ( see section 3.3 below ) . despite its redshift of 0.022 , sn 1988z has remained detectable for nearly three decades .
it was a very luminous radio sn that has been studied extensively as it faded slowly after discovery , showing remarkable longevity in the radio , x - rays , and optical emission lines like h@xmath0 @xcite .
it has been compared to sn 1986j , and most studies of its first decade and a half favour the interpretation that it is powered by an energetic explosion ( as much as @xmath64erg ; @xcite ) from a massive progenitor ( an initial mass of something like 2030m@xmath3 or more ) , that was interacting with roughly 10m@xmath3 produced by a star that had a wind with a mass - loss rate of @xmath65m@xmath3yr@xmath11 for @xmath66yr before core collapse .
the rate of decline in h@xmath0 and radio suggested that the mass - loss rate ramped up in the final millennium before core collapse @xcite .
even the somewhat lower mass - loss rate traced by later csm interaction was extreme , however , allowing sn 1988z to remain as one of the most radio luminous sne even after a decade , matched only by sn 2005ip .
here we present a series of optical spectra , listed in table [ tab:88z ] and shown in figure [ fig : spectra88z ] , which extend this late - time evolution from the mid-1990s to the present epoch .
we find that sn 1988z s strong csm interaction has continued to fade very slowly , indicating a remarkably extended and dense wind into which the blast wave continues to crash .
this traces mass loss over @xmath67yr preceeding core collapse .
based on the luminosity of h@xmath0 and the character of the optical spectra , we group sn 1988z s late evolution into a few different epochs : \(1 ) _ the first 5003000 days_. this is the phase that has already been studied extensively as noted above . during this time , sn 1988z showed a steady and very slow decline in radio and h@xmath0 luminosity , although remaining far more luminous than most sne having signs of csm interaction .
@xcite noted a possible break in the decline rate of the h@xmath0 luminosity around day 1000 , although this break is not clear in figure [ fig : lha ] where the decline looks rather continuous within the uncertainties .
of interest in the present paper is that throughout the time period of roughly 10003000 days , sn 2005ip was a nearly identical twin of sn 1988z in terms of its decline rate , its luminosity in the radio ( admittedly only a single epoch so far for sn 2005ip ) , h@xmath0 , and x - rays , and in the appearance of its spectrum . in our previous paper
@xcite we noted similarities in the spectra of sn 2005ip and sn 1988z at early times , although sn 2005ip exhibited stronger coronal lines . from day 1000 to 3000 , however , their spectra are almost indistinguishable .
figure [ fig : spectra88z ] includes a recent spectrum of sn 2005ip on day 3098 for comparison , which appears very similar to the days 2006 and 2542 spectra of sn 1988z in the same plot .
this phase is characterized by very strong intermediate - width h@xmath0 , a steep h@xmath0/h@xmath68 decrement of @xmath69 , several intermediate - width and narrow coronal lines such as [ fe vii ] , and intermediate - width lines that indicate very high electron densities in the post - shock gas , such as [ n ii ] @xmath705755 ( which is , however , blended with an [ fe vii ] line ) .
\(2 ) _ a `` plateau '' over days 30007000_. during this time period , the h@xmath0 luminosity appears to level - off for a decade .
while h@xmath0 remains constant , the rest of the spectrum shows some interesting changes .
the coronal lines and other indicators of high density fade away , leaving mostly narrow h ii region lines and broad , weak oxygen lines ( possibly from the reverse shock ; see below ) that are also seen in young sn remnants .
the intermediate - width h@xmath68 line also fades , yielding an h@xmath0/h@xmath68 ratio that is higher than before ( @xmath71 ) .
\(3 ) _ after day 7000 .
_ eventually , other lines fade away too , leaving only intermediate - width h@xmath0 , which also resumed a decline in flux at these very late times .
even the broad / intermediate - width oxygen lines mostly disappear .
note that in figure [ fig : spectra88z ] , we have subtracted a constant value for the contribution of narrow h@xmath0+[n ii ] emission from underlying h ii regions in the spectrum , which typically contributes about @xmath19ergs@xmath11@xmath20 in our keck slit aperture after subtraction of adjacent background .
this suggests that sn 1988z resides in or near a complex of h ii regions , providing another possible indication of a high initial mass for the progenitor . at sn
1988z s distance of @xmath72mpc , however , @xmath73 corresponds to almost 500pc , which could include a large complex of h ii regions like the tarantula nebula , and is not a very reliable or precise indicator of initial mass .
it is , however , a good indicator of a relatively strong ambient uv radiation field .
this last point is interesting in the context of something we discuss later .
namely , @xcite have suggested that external ionization of an otherwise normal rsg wind can produce a stalled dense shell , which might offer an alternative to very strong or eruptive mass loss as an origin for sne iin . despite its probable location in an h ii region , this idea does not seem to apply well to sn 1988z , which from its earliest phases shows an uninterrupted decline in csm interaction strength , consistent with a strong shock moving through a freely expanding wind .
the total mass of 1520m@xmath3 or more of csm that has been swept up so far by sn 1988z rules out a normal rsg wind that has been confined to a thin shell , because such shells only contain a fraction ( about a third ) of the total rsg mass loss , and are therefore expected to have much lower total mass @xcite .
with such a strong wind , the expected location of a stalled shell would probably be different from the models presented by @xcite , however .
we return to this issue later , because despite its similarity to sn 1988z , sn 2005ip does show a rapid interruption of the smooth decline in its csm interaction strength , which might be indicative of external influence of this sort .
this may hint that two otherwise very similar events may have their csm modified in different ways .
then again , the relatively coarse sampling of sn 1988z in its past decade does not rule out the possibility of some brief drops or spikes in luminosity indicative of density fluctuations in its csm .
indeed , one may speculate that the late plateau phase in sn 1988z s h@xmath0 luminosity might represent the shock traversing a relatively constant density , ionized portion of the wind outside such a neutral dense shell .
for this reason as well , it will be interesting to see how sn 2005ip behaves in the coming decade as compared to sn 1988z .
all the sne discussed in this paper show signs of long - lived , strong csm interaction , but there are interesting differences . examining figures [ fig : spectra ] and [ fig : spectra88z ] , it is clear that both sn 2005ip and sn 1988z have late - time spectra dominated by their intermediate - width components of h@xmath0 and other lines , with line widths of about 2000kms@xmath11 .
both sn 1993j and sn 1998s have significantly broader line profiles of 500010,000kms@xmath11 .
this difference is probably closely linked to their total h@xmath0 luminosity . both sn 2005ip and sn 1988z are @xmath74 orders of magnitude more luminous than the other two at a comparable late epoch .
the combination of higher h@xmath0 luminosity ( as well as higher radio and x - ray luminosities ) , high - density tracers , and narrower lines likely stems from the same cause : _ denser csm and a radiative forward shock_. if a blast wave is expanding into extremely dense and extended csm , the forward shock can remain radiative , giving higher csm interaction luminosity and causing the intermediate - width components to continue to dominate the spectrum as the blast wave continues to expand into the csm and to decelerate slowly .
( strong global asymmetry or a high degree of clumping can provide very high csm densities without necessarily having a huge total csm mass , so an intermediate - width component might also be present at lower luminosity in some cases . ) at lower average csm densities and luminosities , as in sn 1993j and sn 1998s , radiation from the reverse shock dominates the late - time emission , and so the spectrum exhibits much broader lines from oxygen - rich sn ejecta crossing the reverse shock . in our late spectra of sn 1993j , the flat - topped and double - peaked profiles described by @xcite still persist
. these may be the result of interaction with a disk - like csm , which pinches the waist of the csm interaction region and makes the reverse shock brighter at the equator .
one may envision a reverse - shock geometry similar to that of sn 1987a @xcite .
interestingly , we may see a transitional phase in sn 1988z , where weak and broad oxygen lines remain after the intermediate - width coronal lines have faded ( days 5138 and 6640 in figure [ fig : spectra88z ] ) .
in general , though , the lack of strong oxygen lines in most sne iin at a few hundred days after explosion is most likely a consequence of the fact that the forward shock still dominates the emission spectrum . since this is tracing shock - heated csm material , it reflects the composition of the csm and not the composition of the inner ejecta .
thus , a lack of strong oxygen features should not be taken as any indication of the abundances in the sn ejecta as long as intermediate - width lines are still seen .
another relevant detail concerns the line profiles . while sn 2005ip and sn 1988z are otherwise very similar , they differ in that sn 2005ip shows asymmetric blueshifted line profiles , especially in its intermediate - width h@xmath0 line at late times ( figure [ fig : spectra ] ) .
the intermediate - width h@xmath0 components of sn 1988z remain much more symmetric at all epochs ( figure [ fig : spectra88z ] ) .
this could signify a difference in dust - formation efficiency in the two sne , or perhaps viewing - angle effects if the csm interaction is not spherically symmetric ( for example , sn 1988z could be a ring or disk - like interaction seen pole - on , in which case dust can not block emission from receding material ) .
interestingly , the late - time line profiles of sn 1998s , and to a somewhat lesser extent sn 1993j , do show blueshifted asymmetry that may be indicative of dust formation in these sne .
this was already discussed in detail for the case of sn 1998s by @xcite .
there have been no published studies of early - time spectropolarimetry of sn 1988z or sn 2005ip , however . ,
100 , 24 , 33 mpc , respectively ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
we have assumed that the radio emission from each event is emitted isotropically .
the dotted curve shows the best - fit model to late - time ( @xmath75 day ) emission from sn 1988z.,width=336 ] very few radio sne have been observed beyond 10 yr .
previous examples such as sne 1986j , 1988z , and 1995n ( see , e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) were noted for having unusually luminous radio emission that peaked around 900 to 1100 days after the inferred explosion date . in , we compare the 5 ghz radio luminosity observed toward sn 2005ip to these radio sne observed at late times ( 15006000 days ) .
sn 2005ip appears very similar in its radio luminosity to sn 1988z at a comparable epoch , which is perhaps not surprising in light of the h@xmath0 results above .
the best fit to the observed radio light curve of sn 1988z is based on a five - parameter model derived from @xcite : @xmath76 with @xmath77 , and @xmath78 as free parameters .
the explosion date @xmath79 can be fixed from early - time optical photometry . in @xcite ,
the fit parameters for the late - time ( @xmath75 day ) light curve of sn 1988z were @xmath80 , @xmath81 , @xmath82 , @xmath83 , and @xmath84 .
these parameters imply that the sn progenitor underwent mass loss of at least @xmath85m@xmath3yr@xmath11 for the last 10,000 yr before exploding .
sn 2005ip is about 20% less luminous than sn 1988z at 5 ghz and @xmath86 days , but it is a factor of a few more luminous in x - rays at the same epoch .
analysis of x - rays suggests significantly larger mass - loss rates of @xmath12 to @xmath87m@xmath3yr@xmath11 @xcite .
it is difficult to infer the physical parameters of the csm surrounding sn 2005ip based on a single epoch of radio observations .
however , given the similarity between sne 2005ip and 1988z at both optical and radio wavelengths , we can hypothesize that sn 2005ip has an extremely dense and clumpy csm profile out to large radii .
we plan to continue to monitor the radio and x - ray evolution of sn 2005ip in light of its recent h@xmath0 brightening , and the results will be presented in a future paper .
-0.0 true cm , but allows for a more direct comparison with sn 1988z .
the two sne overlap in x - ray luminosity in this energy band , but have somewhat different deecay rates.,title="fig:",width=336 ] as noted in section 2.3 , the rate of fading of the x - ray luminosity of sn 2005ip implies a csm density that falls off more steeply than for a steady wind with @xmath88 and @xmath89 , which should yield a @xmath90 profile in the case of free - free emission from a wind .
this would imply that the star s mass - loss rate was ramping up in the last few thousand years of its life as it approached core collapse .
this is qualitatively consistent with implications from the h@xmath0 evolution noted above .
the wind density being steeper than a steady wind is also consistent with the earlier analysis of x - rays by @xcite .
looking at x - rays alone , the rate of fading appears to be consistent with a continuation of the trend inferred by @xcite .
this is interesting , since the h@xmath0 luminosity shows a more sudden drop followed by a resurgence in recent data .
it is possible that the coarse time sampling of the x - ray data has missed this peculiar variability , so future additional x - ray observations are required to resolve this .
although the x - rays from sn 2005ip imply a csm that falls off more steeply than a steady wind , it is interesting to note that the x - ray luminosity of sn 1988z falls off even faster with time at a similar epoch after explosion than sn 2005ip ( e.g. , figure [ fig:05ip88z ] and @xcite ) . note that a direct comparison between x - rays from sn 2005ip and sn 1988z is a bit tricky , because @xcite made somewhat different assumptions in their analysis .
therefore , for the purpose of comparing the two objects in fig [ fig:05ip88z ] , we have recalibrated our sn 2005ip x - ray data with the same assumptions ( integrated over 0.52kev , uncorrected for intrinsic absorption ) . here
we find that the x - ray luminosities are about the same at @xmath91d , although the somewhat faster decline of sn 2005ip is still apparent . yet , despite this faster decline in x - rays around a decade after explosion , the h@xmath0 luminosity of sn 1988z then hit a plateau and did not decline for another decade after this epoch .
this constant h@xmath0 luminosity seems inconsistent with the steep drop in wind density implied by the x - rays .
overall , it seems that a clear lesson from comparing various diagnostics from different late - time interactors is that any single diagnostic by itself provides an incomplete picture , so readers should be aware that some of the corresponding mass - loss rate estimates are likely to be lower limits .
note that the value of @xmath92 that we derive from radio emission is an order of magnitude lower than from x - ray and h@xmath0 observations of sn 2005ip at the same epoch .
consider the comparison of sn 1988z and sn 2005ip a decade after explosion .
the two have roughly equal x - ray and h@xmath0 luminosities , but sn 1988z is more luminous in radio emission .
a caveat is that the quoted x - ray luminosities are not bolometric x - ray luminosities , although we have tried to recalibrate our sn 2005ip x - ray data in the same way as had been done for sn 1988z @xcite .
one might imagine that these inconsistencies could be attributed to some combination of clumping or aspherical geometry , which can influence the emissivity of h@xmath0 or the escape of x - rays and radio differently for a given csm mass .
dense regions that emit strong h@xmath0 emission might be self absorbed in the radio , for example , and a range of densities may exist simultaneously if the emitting region is clumpy .
@xcite discussed how asymmetry might influence x - ray emission in sne iin .
we noted earlier that there are some indications of asymmetry and different viewing angles between the two sne ( sn 2005ip shows asymmetric blueshifted line profiles in h@xmath0 that imply dust formation blocks the far side , whereas sn 1988z does not ; this could perhaps be ascribed to a pole - on view of sn 1988z and a nearly edge - on view of sn 2005ip ) .
while such aspects are difficult to constrain with high confidence , the large binary fraction among massive stars @xcite and the high incidence of aspherical geometry in resolved csm around nearby massive stars suggests that effects such as asymmetry and clumping may be the norm rather than an exception .
in this paper we have emphasized the remarkable similarity between the late - time evolution of sn 2005ip and that of sn 1988z , which is the prototypical long - lived sn iin .
so far there are three main differences in their evolution . first , the csm interaction in sn 1988z was stronger at early times and declined for the first 1000 days , whereas sn 2005ip had a more delayed onset of its strongest csm interaction .
sn 2005ip s interaction remained roughly constant until day 1000 , after which it tracked that of sn 1988z .
this relatively delayed onset in sn 2005ip was accompanied by stronger narrow coronal emission lines at early times , probably signifying a more highly clumped inner - wind region ( compared to sn 1988z ) , allowing the densest pre - shock regions to be photoionized by x - rays from a very fast blast wave @xcite .
the second major difference is that sn 2005ip has just recently shown a drop and then a quick resurgence in its h@xmath0 luminosity over the course of a few months , indicative of strong density fluctuations in its distant csm .
the recent sudden increase in h@xmath0 would reflect a density increase by a factor of 25 .
sn 1988z has shown no such behaviour , although brief blips can not be ruled out by the coarse time sampling of our late - time sn 1988z spectra . for a blast wave expanding at 20005000kms@xmath11 , fluctuations seen @xmath93 days after sn 2005ip s explosion would correspond to radii of ( 0.61.5 ) @xmath94 cm , or roughly 0.020.05pc from the star .
interestingly , this is consistent with the radius of a warm dust shell responsible for the near / mid - ir echo seen at early times @xcite , which was inferred to reside at @xmath9pc from the star @xcite .
this dust was ascribed to a shell of enhanced density at that location .
it therefere seems likely that the sudden increase in the h@xmath0 luminosity that we observe around day 3500 could in fact be caused by the forward shock reaching and overtaking this same dusty csm shell .
if so , continued study of this interaction may provide a novel way to investigate the destruction or survival of csm dust that is hit by a sn blast wave .
third , despite their similar h@xmath0 luminosity evolution , sn 2005ip declines more slowly in x - rays than sn 1988z , but it is also somewhat less luminous at radio wavelengths than sn 1988z at a comparable epoch ( see sections 3.4 and 3.5 for details and caveats ) .
this suggests that any one diagnostic taken alone gives an incomplete picture of the progenitor mass loss .
the recent h@xmath0 variability suggests that sn 2005ip has strong deviations from a monotonically decreasing wind density at large radii , but what caused this ?
a simple interpretation would be that changes in csm density reflect previous changes in wind mass - loss rate . at a radius of 0.03pc , and with a constant wind speed of @xmath95kms@xmath11 (
somewhat faster wind speeds are appropriate for more luminous rsgs ; see @xcite ) , this would point to substantial changes in mass loss about 1000yr preceding core collapse . as noted in section 1 , sudden bursts of mass loss on these timescales before core collapse do not fall within the realm of currently suggested ideas that involve ne and o burning , which are too brief for objects like sn 2005ip and sn 1988z @xcite . instead , it appears as if the mass - loss rate was ramping up in the last few thousand years before core collapse in sn 2005ip and sn 1988z .
some evidence for such behaviour is observed among nearby populations of rsgs @xcite , and it may be expected theoretically @xcite .
although rsg winds may ramp up near the end of a star s life , strong fluctuations in csm density may point to episodic mass loss akin to eruptions of luminous blue variables ( lbvs ) or unsteady binary mass - transfer episodes .
indeed , lbv - like eruptive mass loss has been suggested as the possible origin of the csm shells around sn 2005ip @xcite . @xcite
have discussed how binary interaction triggered by pre - sn evolution might instigate lbv - like mass loss through a rapid onset of common - envelope evolution . recalling that lbvs are
a phenomenological class and binary interaction is a physical mechanism , these might be two names for the same phenomenon . some current ideas for lbvs do favour binary interaction as a necessary ingredient in their evolution @xcite .
a drop in pre - shock density followed immediately by an increase in density , as indicated by the h@xmath0 fluctuation of sn 2005ip , might also result from a temporary increase in wind speed that swept material into a thinner shell .
thus , variations in wind speed , as opposed to just wind mass - loss rate , might also give rise to such structures .
wind speeds could vary significantly if the progenitor experienced a blue loop on the hertzsprung - russell diagram , for example .
perhaps the common assumption of ballistic speeds is naive , but nevertheless , this material must be far from the star and quite old compared to the csm overtaken by normal sne iin during their early bright phases .
another puzzling aspect of sn 2005ip s late - time behaviour concerns its x - ray emission . in 20132014 , analysis of x - rays from sn 2005ip @xcite indicated a drop in both the x - ray flux and the absorbing neutral h column density ( @xmath14 ) along our line of sight ( reaching the expected galactic line - of - sight value by the last epoch ) .
this prompted @xcite to suggest that sn 2005ip s blast wave had reached the outer boundary of the dense shell .
however , the similarity to sn 1988z ( which continued unabated for another decade after this epoch ) , the recent resurgence in h@xmath0 emission from sn 2005ip , and strong radio and x - ray emission , indicate that its csm interaction is not yet going away . in that case , how should we reconcile the observed drop in @xmath14 ?
@xcite considered the possibility that the outer rsg wind is ionized by the luminosity from csm interaction , but found the ionization to be insufficient for expected shock parameters .
however , they did not discuss the possibility of external influences on the rsg wind , and interpreted the drop in density as a shell resulting from a past lbv - like eruption , as noted above .
( they also noted that a change in the degree of clumping might explain the drop in @xmath14 . )
an alternative explanation for the origin of a sudden jump in density at a large radius may be an otherwise normal steady wind that becomes confined by external pressure at large radii @xcite . if the progenitor was indeed a more massive star ( above 1820m@xmath3 ) so that it was an o - type star when on the main sequence , then its hot shocked main - sequence wind will provide an external pressure that might decelerate the rsg wind and confine it to a thin shell .
( if the progenitor has similarly massive o - type neighbours in a young stellar cluster , the hot interior of the h ii region might create the required external pressure and ionization . ) @xcite have discussed the interesting possibility that some sne iin may arise from the interaction between the sn shock and this type of externally confined , thin neutral shell . in practice
, this may be quite rare : most normal rsg progenitors are at relatively low initial masses ( i.e. , 1015m@xmath3 ) , which means that they were not o - type stars on the main - sequence , and they reach the rsg phase long after associated o - type stars have died .
on the other hand , more - massive rsgs that die sooner have much higher wind momentum in the rsg phase than assumed in the models , so it remains unclear if normal rsg winds are likely to yield sne iin through this process .
although this scenario was proposed as a way to explain sne iin without appealing to eruptive pre - sn mass loss , it may also provide a way for an interacting sn to be `` rejuvenated '' in its old age , since the stalled shell is denser than a freely expanding wind .
@xcite estimate that such a shell can have its density enhanced by a factor of 80 , and may contain a significant fraction ( perhaps a third ) of the total mass lost during the rsg phase .
their models demonstrate that a sn can show a bolometric luminosity spike at very late times from such a collision . in this case
it is difficult to estimate the age of the shell , since its expansion may have stalled , and its radius depends on the external pressure as much as on the momentum of the rsg wind .
future monitoring of sn 2005ip may help to test this hypothesis , as the shock continues to make its way through this dense neutral shell and out into the ionized portion of the rsg wind , or if sn 2005ip resumes a very slow decline as sn 1988z did .
also , higher - resolution spectra may be able to place tighter constraints on the pre - shock csm expansion speed , since the narrow components are unresolved in our moderate - resolution spectra presented here .
regardless of the true explanation for the density fluctuation in sn 2005ip s outer csm , it is useful to realize that mass - loss rates this high ( several times @xmath56 to @xmath12m@xmath3yr@xmath11 or more ) are quite extreme , and do not correspond to normal rsgs like betelgeuse , where the wind is two orders of magnitude less dense @xcite .
the cool evolved stars with winds this dense are rare ( see , e.g. , @xcite for an overview of mass - loss rates ) , corresponding to the most luminous and extreme rsgs that are enshrouded by their own dusty winds , like vy cma , nml cyg , s per , etc . , or to self - enshrouded super - asymptotic giant branch ( agb ) stars . in the cases of sn 1988z and sn 2005ip , however ,
the very large csm mass needed for their sustained high - luminosity csm interaction has been estimated as 10m@xmath3 or more .
this confidently rules out super - agb stars ( total mass including a neutron star of only @xmath96m@xmath3 ) as their possible progenitors .
this would be unlikely anyway , as these super - agb stars that may undergo o - ne - mg core collapse in an electron - capture event are expected to produce a low sn kinetic energy of @xmath97 ergs @xcite .
thus , electron - capture sne that produce sne iin through interaction may be expected to fade quickly ( see , e.g. , @xcite ) .
sn 1988z was hyperenergetic ; up to day 3000 , @xcite estimate a total radiated energy for sn 1988z of @xmath98erg ( model - dependent @xmath99 ) , and at least @xmath100erg ( observed @xmath101 ) .
sn 2005ip appears very similar in many respects , and is even more luminous and slower to decline in x - rays .
the extreme rsgs mentioned have initial masses of order 2535m@xmath3 , higher than normal unobscured rsgs , with more tightly bound cores that may require more energetic explosions than normal sne ii - p .
this reinforces earlier suggestions that these extreme rsgs ( as opposed to normal rsgs ) are the likely progenitors of some sne iin , including long - lived sn 1988z - like events @xcite .
more massive lbv stars may also be able to supply the required amount of csm mass @xcite .
however , the steady late - time decline over decades in the case of sn 1988z argues in favour of a freely expanding , slow , and sustained dense rsg wind , rather than the sporadic eruptive mass loss characteristic of lbvs . for sn 2005ip , future observations of its continuing evolution through the extended csm may help us choose more confidently between an extreme rsg perhaps exposed to external ionizing radiation from its environment or an eruptive progenitor . in any case
, the strong variability in h@xmath0 around day 3500 in sn 2005ip has not been seen so clearly before at such late times in sne iin , and suggests that similar late - time observations of other objects can provide important constraints on the variability in their pre - sn mass loss .
strong and variable mass loss of the type seen here occurring 1000yr before death presents a challenge for models of pre - sn evolution .
once we have a larger number of known sne iin with such enduring strong csm interaction , it would be interesting to investigate their statistical proximity to h ii regions as compared to other populations of sne .
we thank an anonymous referee for helpful suggestions .
support was provided by the national science foundation ( nsf ) through grants ast-1210599 and ast-1312221 to the university of arizona .
c.d.k.s research receives support from nasa through contract number 1255094 issued by jpl / caltech . w.f .
was supported by nasa through an einstein postdoctoral fellowship .
the supernova research of a.v.f.s group at u.c .
berkeley is supported by gary & cynthia bengier , the richard & rhoda goldman fund , the christopher r. redlich fund , the tabasgo foundation , and nsf grant ast-1211916 .
we thank the staffs at the mmt and keck observatories for their assistance with the observations .
observations using steward observatory facilities were obtained as part of the large observing program aztec : arizona transient exploration and characterization .
some of the observations reported in this paper were obtained at the mmt observatory , a joint facility of the university of arizona and the smithsonian institution .
this research was also based in part on observations made with the lbt .
the lbt is an international collaboration among institutions in the united states , italy , and germany .
the lbt corporation partners are the university of arizona on behalf of the arizona university system ; the istituto nazionale di astrofisica , italy ; the lbt beteiligungsgesellschaft , germany , representing the max - planck society , the astrophysical institute potsdam , and heidelberg university ; the ohio state university and the research corporation , on behalf of the university of notre dame , university of minnesota , and university of virginia . some 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 nasa ; 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 .
smith n , chornock r , li w , ganeshalingam m , silverman jm , foley rj , filippenko av , barth aj , 2008a , apj , 686 , 467 smith n , et al . , 2008b , apj , 686 , 485 smith n , et al . , 2009a , apj , 695 , 1334 smith n , hinkle kh , ryde n , 2009b , aj , 137 , 3558 | supernova ( sn ) 2005ip was a type iin event notable for its sustained strong interaction with circumstellar material ( csm ) , coronal emission lines , and infrared ( ir ) excess , interpreted as shock interaction with the very dense and clumpy wind of an extreme red supergiant .
we present a series of late - time spectra of sn 2005ip and a first radio detection of this sn , plus late - time x - rays , all of which indicate that its csm interaction is still strong a decade post - explosion
. we also present and discuss new spectra of geriatric sne with continued csm interaction : sn 1988z , sn 1993j , and sn 1998s . from 310 yr post - explosion ,
sn 2005ip s h@xmath0 luminosity and other observed characteristics were nearly identical to those of the radio - luminous sn 1988z , and much more luminous than sne 1993j and 1998s . at 10 yr after explosion , sn 2005ip showed a drop in h@xmath0 luminosity , followed by a quick resurgence over several months .
we interpret this h@xmath0 variability as ejecta crashing into a dense shell located @xmath1pc from the star , which may be the same shell that caused the ir echo at earlier epochs .
the extreme h@xmath0 luminosities in sn 2005ip and sn 1988z are still dominated by the forward shock at 10 yr post - explosion , whereas sn 1993j and sn 1998s are dominated by the reverse shock at a similar age .
continuous strong csm interaction in sne 2005ip and 1988z is indicative of enhanced mass loss for @xmath2yr before core collapse , longer than ne , o , or si burning phases .
instead , the episodic mass loss must extend back through c burning and perhaps even part of he burning .
[ firstpage ] circumstellar matter stars : evolution stars : winds , outflows supernovae : general supernovae : individual ( sn 2005ip ) |
the evolution of high mass clustered star forming regions is complex and poorly understood . only the nearby ( @xmath2400 pc ) , optically revealed , orion nebula cluster ( onc ) is well studied ( see muench et al . and odell et al.in this handbook ) . yet , a wide variety of high mass embedded clusters is found within 2 kpc of the sun . within this limit , the young cluster rcw 38 ( 08@[email protected]@xmath5 -47@xmath631@xmath757@xmath8 ( j2000 ) , _ l , b _
= @xmath9 ) is one of the few regions other than the onc to contain over 1000 members ( lada & lada 2003 ; wolk et al .
rcw 38 has an embedded and dense stellar population , comparable to other regions that have been studied recently with spitzer ( e.g. , m 20 and rcw 49 ; rho et al . 2004 ,
whitney et al .
2004 , churchwell et al .
rcw 38 provides a unique opportunity to study the evolution of a rich cluster during the phase where its most massive members , a pair of o5.5 stars ( derose et al.2008 ) have just completed their ultracompact hii region ( uchii ) phase and are now greatly influencing its natal environment and the evolution of its low mass members .
rcw 38 was catalogued in both the h@xmath10 survey by rodgers , campbell & whiteoak ( 1960 ) and the earlier survey of southern hii regions ( gum 1955 ) as a moderately bright region of emission about 40 on a side ( figure [ dss ] ) .
radio surveys during the 1960 s indicated that it was one of the brightest hii regions at radio wavelengths ( e.g. , wilson et al .
the radio brightness made it an early candidate supernova remnant until its spectrum was shown to be thermal .
a complicating factor in the study of rcw 38 is that it is adjacent to the vela molecular ridge ( see the chapter by pettersson ) .
in fact , there is a faint ring shape structure perhaps a supernova remnant ( snr ) in apparent contact with the star forming cloud ( rx j0852.0 - 4622 ) .
however , this seems to be a chance superposition as the snr is about a factor of two closer than rcw 38 based on absorption arguments ( aschenbach et al .
the 5 ghz survey by shaver & goss ( 1969 ) shows rcw 38 as both a bright continuum source at multiple wavelengths ( shaver & goss 1970 ) and broad in its extent of greater than 30 in diameter .
it was not until the early 1970 s that this region was clearly associated with massive young stars ( johnson 1973 ) .
the next year , with improving resolution and detectors , h90@xmath10 observations indicated that the radio peak of the region had a ring like shape about 2 across ( huchtmeier 1974 ) . figure [ dss ] shows an optical view of the rcw 38 region .
note the central region is opaque due to high extinction .
this has confined the study of the central region mostly to infrared and radio wavelengths .
several infrared studies indicated a `` ring like '' or `` horseshoe '' structure around irs 2 ( figures [ 2mass],[vlt ] ) , which is also seen in high resolution radio and millimeter wavelength studies ( vigil 2004 ; wolk et al .
these structures are shown in detail in figure [ vlt ] .
there are two defining infrared sources in rcw 38 ( frogel & persson 1974 ) the brightest at 2 is labeled irs 2 and the brightest at 10 is referred to as irs 1 .
frogel & persson ( 1974 ) studied rcw 38 at 10 @xmath11 m with moderate spatial resolution ( 14.5 ) and found that the warm dust emission follows a horse - shoe shape across a @xmath12 arcmin@xmath13 area ; they labeled the brightest unresolved peak in its western side irs 1 .
epchtein & turon ( 1979 ) mapped the central region of rcw 38 , at 10 , with about twice the resolution of frogel & persson and resolved irs 1 into several discrete sources ; they suggested that this may be a cluster of embedded stars , younger than those associated with irs 2 . in msx observations
at 8.3 @xmath11 m the central region of rcw 38 exhibits a circular emission structure , perhaps slightly elongated about 2 across , in agreement with the earlier mid - infrared studies .
this is resolved by @xmath14/irac observations at 3.6 - 8.0 @xmath11 m ( wolk pi ) to show a roughly heart shaped region of polycyclic aromatic hydrocarbon ( pah ) emission which includes numerous bubbles filled with 4.5 diffuse emission ( see figure [ 2mass ] ) .
the central region appears to lie within a large bubble about 0.4 pc across .
the o5.5 binary irs 2 appears to lie in the center of the bubble as seen in the mid- and near - infrared data ( sect . 4 and figures 2 and 3 ) .
the inner face of the eastern part of the central bubble appears to be illuminated , giving the impression that we are not looking straight down the opening but rather offset by between 30 and 45 degrees .
smith et al .
( 1999 ) made high resolution mid - infrared maps of the core region of rcw 38 .
they concentrated on the irs 1 region that has the peak radio and 10 emission .
they found that irs 1 is a dust ridge extending 0.1 - 0.2 parsec predominantly in the north south direction less than 0.1 pc west of irs 2 .
the dust ridge has a color temperature of about 175 k and has several condensations within it .
they identify irs 1 as the brightest peak among them , lying interior to all the other peaks about 12 west of irs 2 .
from its near - infrared ( nir ) colors and assuming a distance of 1.5 kpc , frogel & persson ( 1974 ) suggested that the brightest 2 source , irs 2 , is an early o - star ( specifically o4 ) having about 12.8 magnitudes of visual extinction , or a group of later spectral type o stars .
this is sufficient to account for the observed continuum emission at 5 ghz ( shaver & goss 1970 ) .
furniss , jennings , & moorwood ( 1975 ) combined broad - band 40 - 350 @xmath11 m measurements with the previous work of frogel & persson , deriving a total luminosity of 7@xmath15 l@xmath16 and arguing that an o4 star would supply more than enough luminosity to account for all the infrared luminosity and emit enough ly photons to yield twice the observed radio flux , unless significant absorption by dust is occurring .
they suggest however that irs 2 is an o5 star with little or no dust absorption of the continuum photons in the ionized region .
the higher resolution imaging of ligori et al .
( 1994 ) confirmed the presence of a cluster of at least 5 optically invisible young stellar objects embedded in nebular emission in the central 1.5 near irs 2 .
frogel & persson ( 1974 ) produced detailed 1.25 20 mapping of the region . at 10
they confirm a ring - like structure with the bright 2 source
irs 2 very near the center and a dominating 10 source ( irs 1 ) about 10 to the west
. a third 10 source about 40 to the east of irs 2 was also noted .
persson et al . ( 1976 ) revisited this region using narrow band photometry and found no evidence of silicates .
other narrow line infrared imaging such as br-@xmath17 at 2.16 also indicated an open center morphology centered on irs 2 with enhanced emission toward irs 1 ( mizutani et al .
1987 ) . mid - infrared imaging and spectroscopy by smith et al . ( 1999 ) provide further evidence that the ionizing source of the region is at least o5 or earlier . through modeling of the ratio of emission lines at irs1 ,
the ionizing radiation field constrains the ionizing source to be between o4.5 and o5.5 .
smith et al .
find that the region surrounding irs2 is depleted of dust , suggesting the region can be explained in terms of a wind - blown cavity , where winds from irs2 have blown out the surrounding material .
derose et al .
( 2008 ) used adaptive optics imaging to directly show that irs2 is a binary with equal mass members and a projected separation of @xmath2 500 au .
the presence of two o stars of equal mass combined with the earlier data on luminosity and ionization imply that they have spectral types of o5.5 .
muzzio ( 1979 ) and muzzio & celotti de frecha ( 1979 ) identified ob stars throughout the southern sky using photometry and objective prism photography .
they identified about 20 ob star candidates in the vicinity of rcw 38 .
all have absolute v magnitudes less than zero but are concentrated about 1020 southwest of irs 2 .
wolk et al .
( 2006 ) use the absolute k - band magnitudes of x ray sources to determine the existence of 31 candidate ob stars over a 5 pc@xmath18 region centered on irs 2 .
more than 20 of these are found to be in the central @xmath21 pc .
no unusual clustering to the southeast was noted . to make mass estimates they estimated the age of the cluster at 0.5 myr and noted that if they assumed an age of 1 myr , the number of ob stars would have more than doubled to a somewhat less reasonable value .
however , the wolk et al .
sample was x - ray based and hence biased against finding soft ob stars behind more than 20 a@xmath19 of extinction .
the ob candidates remain to be confirmed .
nir imaging and photometry of the region by storey & bailey ( 1982 ) showed a number of highly reddened point sources and they deduced an a@xmath19 of order 60 to the cluster .
the 2mass data show that the nebulosity associated with rcw 38 is extensive across a large area , with dust lanes and patches running throughout .
wolk et al .
( 2006 ) measure the extinction and hydrogen column to over a score of sources with good x - ray and nir data and find a@xmath19 ranging from 3 to 20 with strong evidence of nir sources embedded in more than 30 magnitudes of visual extinction .
smith et al .
( 1999 ) find that the region around irs 2 has a gas to dust ratio much lower than 100 to 1 .
thus , this area is a true cavity and not just an extinction effect .
millimeter radio observations indicate that several hundred magnitudes of extinguishing material may lie behind the observed stars ( see section 11 ) . combining these results with the near
ir extinction observations suggests that rcw 38 is an embedded blister hii region lying just inside the front edge of a giant molecular cloud .
further , it appears that the hii region is compacted by the overlying material and beginning to break out in some locations .
even in the remarkable @xmath14/irac image ( figure [ 2mass ] ) the scope and extent of the cluster is hard to ascertain .
only a few dozen sources are identifiable within the bright nebular emission . outside the ir nebula cluster members are difficult to distinguish from background stars .
we have found that 116 sources in the 2mass catalog , corresponding to the region shown in figure 3 , have ir - excesses consistent with an optically thick disk at k band . a number of the cluster members are directly exposed to irs 2 in the cavity , and could be undergoing the same fate as the proplyds in the onc ( derose et al . 2008 ) .
further , about 250 sources in the regions are class i or class ii objects based on their irac colors .
this group includes the 2mass selected candidates .
most of these are clearly associated with rcw 38 . however , at least 2 are associated with a background star forming region
bran 231a .
coincidences among the x - ray detected sources and nir excess sources are found over 10 from irs 2 indicating that the cluster extends at least 5 pc .
the first hint of the extent of the low mass star population is provided by the spectacular nir image of the heart of the region ( figure [ vlt ] ) . despite the small ( 2.5 on a side ) field of view
, one can immediately ascertain several important features of the system from this image .
first , the blue to red gradient indicates a steep dust extinction gradient increasing from northwest to southeast .
second , two regions are cleared of dust , one about 0.1 pc in diameter centered on irs 2 , and another of similar size , just west of irs 1 . both of these cleared regions are visible in the optical plates indicating that extinction is not particularly high for these regions .
the bright ridge separating the two cleared regions is the irs 1 ridge which was mapped in detail by smith et al .
third , there is a plethora of stars in this image .
over 480 stars are visible in this image and the vast majority of these are likely to be cluster members ( wolk et al . 2006 ,
derose et al . 2008 ) .
the data are complete to the brown dwarf limit for extinctions less than an a@xmath19 of 20 ( wolk et al .
about 130 of the nir sources have x ray counterparts there are fewer than a dozen x - ray sources without nir counterparts in the central 2.5 .
this leaves about 350 nir detected cluster members not detected in x - rays an incompleteness of over 70% .
the x - ray counterparts demonstrate both membership in the cluster and masses of more than about 0.5 m@xmath16 .
overall about 345 x - ray sources detected by @xmath20 in a 17x 17 field centered here are associated with rcw 38 ( wolk et al .
completeness arguments based on the cluster distance , obscuration , and x - ray sensitivity lead to a total estimated cluster size of between 1500 and 2400 stars .
this makes rcw 38 the largest embedded cluster by membership within 2 kpc of the sun , after the onc .
as shown in figure [ diffuse ] , rcw 38 hosts very bright diffuse x - ray emission ( wolk et al . 2002 ) .
this x - ray emission differs from that of most other massive star forming regions because of its power - law spectrum indicative of synchrotron , not thermal , emission . since the spectral signature is not that of stellar coronae this emission can not be the sum of unresolved stars .
further , it implies the existence of a strong magnetic field .
synchrotron emission requires an electron population to be driven along a magnetic field .
indeed a significant magnetic field has been measured in the gas associated with rcw 38 ( 38@xmath213 @xmath11 g ; bourke et al.2001 ) .
astrophysically , this is most common along the expanding shock front of a supernova remnant or during violent accretion onto a compact object with a large magnetic field .
we noted earlier the existence of a possible supernova remnant in the field , it is offset from rcw 38 by about 6 and there is known x - ray emission associated with this location .
further , the snr is foreground to rcw 38 and the diffuse emission becomes more absorbed towards the southeast .
this is the same pattern seen among the stars and nebulosity associated with rcw 38 indicating that the diffuse x - ray emission is co located with the stars and dust of rcw 38 .
rcw 38 is the first and currently only known source of diffuse synchrotron x - ray emission from a very young star forming region ( feigelson et al .
the morphology of the diffuse x - ray emission is striking .
it is strongest in the central region near irs 2 where radio and ir - nebular emission appears cleared .
further it is confined on the southeast along a ridge that also confines the mm continuum ( cool dust ) emission .
there is also a distinctive kink where the diffuse x - ray emission is deflected around the dust ridge near irs 1 .
it also appears that the plasma is breaking out of some form of confinement to the northwest where it traces around the diffuse nir emission
akin to the way a stream moves around a stone .
this break out to the northwest is visible as reflection nebulosity in figure 1 and corresponding bubble - like structures in figure 2 .
data ( red ) .
the lack of x - ray emission to the southeast is likely due to high extinction . ]
the first reasonable estimate for the distance was from radhakrishnan et al .
they estimate a kinematic distance of 1.5 kpc . muzzio ( 1979 )
identified about 10 ob star candidates in the vicinity of rcw 38 and derived a distance of about 1.7 kpc by placing the stars on the main sequence of an hr diagram .
murphy ( 1985 ) examined the co structures in the vela ridge in addition to the photometric data and derived a distance of 1.6 @xmath22 kpc .
avedisova & palous ( 1989 ) performed photometric measurements of 218 star forming regions and also achieve a result of 1.7 kpc .
the x - ray luminosity function is consistent with the 1.7 kpc distance ( wolk et al .
2006 ) . at this distance , 1
arcmin corresponds to 0.495 pc .
early low resolution radio observations of continuum emission and recombination lines indicate typical hii region conditions , i.e. , an electron temperature of @xmath28000 k and an emission measure of @xmath23 pc @xmath24 ( e.g. , mcgee & newton 1981 , caswell & haynes 1987 ) .
a relatively high electron density of @xmath25 @xmath26 supports the view that the hii region is confined by the surrounding molecular cloud . resolved radio studies
reveal a clumpy ring - like structure around irs 2 showing the general morphology seen in the infrared ( huchtmeier 1974 ; vigil 2004 ) . as in the mid - infrared , the brightest radio peak coincides with irs 1 ( figure 5 ) .
high resolution observations indicate the ring is mostly optically thick at centimeter wavelengths , with no evidence for synchrotron emission ( vigil 2004 ) .
line widths observed in recombination lines are typically 30 km s@xmath27 or greater .
sensitive high resolution 18 cm ( 1.6 ghz ) observations show weak emission extending to the nw - se outside of the ring , suggesting that , like the x - ray emission , some diffuse radio emission has escaped the compact hii region .
figure [ 2mass ] shows a bubble - like feature associated with this emission in the nw .
the radio ring is not well centered on irs 2 , likely indicating that the surrounding gas is not homogeneous .
detailed molecular line observations of rcw 38 are rare .
gillespie et al.(1979 ) mapped a large region ( 10 @xmath28 5 pc ) in co 1 - 0 with a 3.2 beam .
they found two clouds separated by a bridge of emission .
yamaguchi et al . ( 1999 ) also mapped co 1 - 0 with similar resolution over a larger region , confirming the earlier result , and estimated a mass of @xmath29 m@xmath16 for the two clouds as a whole .
the rcw 38 cluster is directly associated with the eastern cloud .
zinchenko et al .
( 1995 ) mapped the eastern cloud over a 5 arcmin@xmath18 region in cs 2 - 1 with 1 resolution .
their maps show that cs , like the radio , follows a ring - like structure , but close comparison indicates that the cs is located exterior to the radio , tracing the dense molecular gas in which the compact hii region is embedded .
the mass traced by cs is @xmath30 m@xmath16 ( correcting for their incorrectly assumed distance of only 0.7 kpc ) with a mean density of @xmath31 @xmath26 .
a number of other high density tracers have been observed with 1 or better resolution but maps are lacking . deep absorption is observed in the main lines of the lowest rotational levels of oh at 18 cm , and these have been used to measure the magnetic field strength of 38@xmath213 @xmath11 g ( bourke et al .
2001 ) . high resolution ( @xmath32 ) observations of oh show that the absorption is seen against the radio ring , and is directly associated with the gas surrounding the hii region , sharing the same velocity as the recombination lines ( t.bourke , private communication ) .
no maser emission is observed in oh . water maser emission at 22 ghz is observed from a position near to irs 1 , and it is time variable ( kaufmann et al . 1977 ; batchelor et al . 1980 ; caswell et al .
water maser emission is a sure sign of active star formation , and high resolution observations are needed to precisely located this emission .
the dust continuum has been mapped at wavelengths near to 1 mm ( cheung et al .
1980 ; vigil 2004 ) .
low resolution observations revealed only one clump centered near irs 1 ( cheung et al .
1980 ) , with a peak density of @xmath33 @xmath26 , a total mass of @xmath34 m@xmath16 , and a visual extinction of 800 mag .
higher resolution observations reveal the ring - structure seen at other wavelengths ( vigil 2004 ) , with a central hole around irs 2 , but mostly peaking exterior to the radio ring . in particular , the position of peak intensity is west of the irs 1 ridge and associated with the extincted region seen in figure [ vlt ] .
the mass calculated by vigil ( 2004 ) is an order of magnitude less than that of cheung et al .
( 1980 ) , from similar fluxes . as the morphology of the dust emission observed by vigil suggests that it is associated with the radio ring , the extinction determined by cheung et al .
( 1980 ) may be an overestimate .
the ring - like shape of the two dimensional radio and infrared continuum image suggests the possibility of a shell like structure surrounding irs2 .
the bulk of the emission seen in the cm images is concentrated in a ring which would be the thick walls of a three dimensional shell .
the center cavity would contain only the emission from the cover of the shell perpendicular to the direction of observation ( figure [ vigil3 ] ) .
we see two possible interpretations of the relationship among the observational data
one is wind driven ( smith et al .
1999 ) , the other supernova driven ( wolk et al .
the supernova scenario has , as its key piece of evidence , a symmetrical ring of emission in the radio continuum .
a second line of evidence for a supernova in this region is the hot x - ray plasma with a power law spectrum , which appears to be getting harder as it moves away from the cluster center .
this is as one would expect with synchrotron emission being accelerated from a compact object near the cluster center .
the total x - ray luminosity is about @xmath35 ergs sec@xmath27 . in these attributes
it evokes the image of a shell supernova remnant such as sn 1006 .
further , if one assumes equipartition , we derive a very reasonable magnetic field of about 4.9 @xmath11 g and a reservoir of @xmath36 ergs of accelerated electrons , giving a lifetime of about 20,000 years .
but , the radio continuum of rcw 38 reported here does not show the brightening at the edges seen in sn 1006 ( reynolds & gilmore 1986 ) .
the positive spectral index of the radio continuum also favors thermal emission .
further the x - ray emission seen in shell supernovae tends to follow the form of the radio emission ( c.f .
allen et al . 2001 ) . here
, the plasma seems confined by the continuum radio emission with the exception of the breakout to the northwest .
this confinement naturally lends itself to a wind driven scenario .
smith et al .
( 1999 ) note that the central region around irs 2 is dominated by ly@xmath10 emission and is dust free .
this suggests that rcw 38 is a compact hii region expanding into and confined by a non homogeneous molecular cloud .
the wind from the massive o - star ( irs 2 ) can give rise to collisional shocks near to the star as described by cant et al .
such shocks centered near irs 2 would fill the cavity with a limb darkened x - ray emission as seen .
but the derived thermal spectrum for the plasma is much warmer than expected by these models ( @xmath210 kev ( @xmath2 100 mk ) vs. @xmath2 1 kev ( @xmath2 10 mk ) ) .
this high temperature may have to do with confinement due to the surrounding molecular cloud .
heated plasma will tend to move outward and cool , but in the case of rcw 38 the molecular cocoon prevents meaningful cooling and causes a greenhouse effect " on a parsec scale .
smith et al .
( 1999 ) derived a similar result using ly@xmath10 data .
analysis of the smoothed x - ray data supports the hypothesis that the molecular material is trapping the plasma that appears to be partially heated by the trapped ly@xmath10 photons .
there appears to be a break out of the plasma from the ring of the molecular gas in the northwest corner while elsewhere the gas is confined much closer to the o star .
further evidence for this interpretation is found in other regions of massive star formation .
several massive star forming regions have been seen to have much cooler diffuse x - ray emission .
townsley et al .
( 2003 ) report thermal spectra of below 1 kev for the rosette nebula and m17 . other massive star forming regions , including the arches cluster ( yusef - zadeh et al .
2002 ) and ngc 3603 ( moffat et al .
2002 ) have thermal spectra above 3 kev . while a commonality among these clusters is a massive o star ( o6 or earlier ) the key difference between the former group and the latter group seems to be the presence of absorbing ( confining ) material .
all the clusters with high temperature plasma have @xmath37 @xmath38 while the clusters with cooler plasma have less absorption .
the wind driven interpretation involves winds blowing from irs 2 that are excavating the material in its near vicinity creating this expanding shell , seen as a radio continuum ring of emission .
irs 2 is near the center of the cavity and is likely the driving source .
however , irs 2 is displaced from the geometric center of the ring that is offset by about 18 to the northeast . as noted earlier
, this appears due to our viewing angle at the system not being face on and the non - homogeneous surroundings . from inspection of the k band image ( figure 5 ) , an arc - like feature , reminiscent of a bow shock , coincident with the bright radio ridge , can be seen in the southwest of the ring .
it suggests either a massive star forming region @xmath39 to the west , which is creating a bow shock by deflecting the winds from irs 2 around its edge , a wind source from the southwest which could be counteracting the wind from irs 2 causing a compression ridge to build up along the edge of the ridge irs 1 , or a dense globule being ablated ( bertoldi & mckee 1997 ) . indeed , there are about a half dozen x - ray sources within this arc , and they all are found to have column densities in the range of 1.9 - 2.9@xmath40 @xmath38 , equivalent to about 10 - 15 a@xmath19 .
the southwest source , be it a massive star driving a wind or simply a dense core , is preventing the expansion in the southwest direction , causing the shell to expand non - uniformly and mostly toward the northeast , thus explaining the offset of the center to the northeast . from close inspection of the vlt k - band images
, an arc - like feature outlines the bright radio ridge ( figure 5 ) .
this specific feature implies dense gas that is likely photo - evaporating and could suggest that this is a site of triggered star formation .
additional evidence of active star formation occurring just west of irs 1 is a possible protostar found embedded in the hydrogen gas and molecular jets eminating from this region ( derose et al .
2008 ) , as well as the previously mentioned water masers .
the winds from irs 2 appear to be confined by the surrounding molecular cloud . assuming a typical o - star wind of 1000 km / s ( cant et al .
2000 ) and an age for irs 2 of about 500,000 years suggests that a freely expanding wind would have traveled about 500 pc since the star was formed .
the inner width of the shell is close to a tenth of a parsec , much smaller than expected from a freely expanding wind .
also , in the smooth x - ray data , plasma is observed as far as 3.5 pc from irs 2 , far short of the 500 pc predicted for a freely expanding wind .
rcw 38 appears to be a blister compact hii region with winds originating from the central o - star irs 2 .
these winds are excavating the mass in the immediate vicinity of irs 2 creating a shell like structure , detected as a radio continuum ring and also evident at infrared and millimeter wavelengths .
a few hundred young low mass stars are found in the immediate vicinty of irs 2 , and may be directly exposed to its winds and ionizing radiation .
the region as a whole is estimated to contain around 2000 young stars , with 30 ob star candidates .
the region of the ring to the west of irs 2 , containing irs 1 , appears to be particularly active and is likely the site of ongoing ( triggered ) star formation .
the irs 1 ridge appears to be the interface between the irs 2 wind and either a similar wind from a high mass star further to the west , or a dense clump of gas that is being ablated by irs 2 .
further observations are needed to test these scenarios . in many ways ,
rcw 38 appears to be a younger more embedded version of the onc and deserves further study at higher angular resolution and across the spectrum .
we acknowledge many useful comments from the referee leisa townsley .
joo alves openly shared his unpublished data and ideas as part of our ongoing collaboration on this region .
we thank tom megeath for fruitful discussions .
we gratefully acknowledge the financial support from nasa contract nas8 - 39073 ( cxc ) . this work is based in part on observations made with the spitzer space telescope , operated by the jet propulsion laboratory , california institute of technology under a contract with nasa .
support for this work was provided by nasa . | rcw 38 is a uniquely young ( @xmath01 myr ) , embedded ( @xmath1 ) stellar cluster surrounding a pair of early o stars ( @xmath2o5.5 ) and is one of the few regions within 2 kpc other than orion to contain over 1000 members .
x - ray and deep near - infrared observations reveal a dense cluster with over 200 x - ray sources and 400 infrared sources embedded in a diffuse hot plasma within a 1 pc diameter .
the central o star has evacuated its immediate surroundings of dust , creating a wind bubble @xmath20.1 pc in radius that is confined by the surrounding molecular cloud , as traced by millimeter continuum and molecular line emission .
the interface between the bubble and cloud is a region of warm dust and ionized gas , which shows evidence for ongoing star formation .
extended warm dust is found throughout a 23 pc region and coincides with extended x - ray plasma .
this is evidence that the influence of the massive stars reaches beyond the confines of the o star bubble .
rcw 38 appears similar in structure to rcw 49 and m 20 but is at an earlier evolutionary phase .
rcw 38 appears to be a blister compact hii region lying just inside the edge of a giant molecular cloud .
10 pc ) on a side , based on images from the digitized sky survey ( dss ) .
blue plates are printed as blue , red plates are printed as yellow , near - infrared data are printed in red .
the region in the green box is shown at mid - infrared wavelengths in figure 2 . ]
/irac observations .
irac band 1 ( 3.6 ) is assigned to blue , band 2 ( 4.5 ) to green , and band 3 ( 5.6 ) to red . [
3.6 ] band = blue , [ 4.5 ] band = green and [ 5.6 ] band = red .
voids in the 5.6 image tend to be filled by 4.5 emission .
the yellow box is 2.5 on a side and is shown in more detail in figures 3 and 5 . ] |
in this short note we show how to endow with a canonical hopf algebra structure the tensor product @xmath0 of two braided hopf algebras living in a monoidal category , provided that there is an isomorphism @xmath1 satisfying suitable conditions ( see proposition [ prop1.1 ] ) . in particular
, the square tensor @xmath2 is always a braided hopf algebra with this structure . also , the tensor product of two hopf algebras living in a braided category is again a braided hopf algebra ( although in a new braided category ) . when seen in this context ,
the key point is to replace the braid @xmath3 with @xmath4 .
we work in a monoidal category @xmath5 , for instance , the category of vector spaces over a field @xmath6 .
we write @xmath7 and @xmath8 for the tensor product and the unit of @xmath5 , respectively .
the associativity and unit constraints are assumed without referring to them .
we assume the reader is familiar with the notions of algebras and coalgebras in monoidal categories .
all the algebras are associative unitary and the coalgebras are coassociative counitary . given an algebra @xmath9 and a coalgebra @xmath10 , we let @xmath11 , @xmath12 , @xmath13 and @xmath14 denote the multiplication , the unit , the comultiplication and the counit , respectively , specified with a subscript if necessary .
we are going to use the nowadays well known graphic calculus for monoidal and braided categories .
as usual , morphisms will be composed downwards and tensor products will be represented by horizontal concatenation in the corresponding order .
the identity map of an object will be represented by a vertical line . given an algebra @xmath9 , the diagrams @xmath15 and stand for the multiplication map and the unit of @xmath9 respectively .
given a coalgebra @xmath10 , the comultiplication and the counit of @xmath10 will be represented by the diagrams @xmath16 and respectively .
let @xmath17 , @xmath18 be objects in @xmath19 and let @xmath20 be an arrow .
* if @xmath17 is an algebra , then we say that @xmath21 is _ compatible _ with the algebra structure of @xmath17 if @xmath22 and @xmath23 . *
if @xmath17 is a coalgebra , then we say that @xmath21 is compatible with the coalgebra structure of @xmath17 if @xmath24 and @xmath25 .
of course , there are similar compatibilities when @xmath18 is an algebra or a coalgebra .
recall that a braided bialgebra in @xmath5 is an object @xmath26 of @xmath5 endowed with an algebra structure , a coalgebra structure and an isomorphism @xmath27 ( called the braid of @xmath26 ) satisfying the braid equation @xmath28 such that : @xmath29 is compatible with the algebra and coalgebra structures of @xmath26 , @xmath30 is a coalgebra morphism , @xmath31 is an algebra morphism and @xmath32 .
if moreover there exists a map @xmath33 which is the inverse of the identity in the monoid @xmath34 with the convolution product , then we say that @xmath26 is a braided hopf algebra and we call @xmath35 the antipode of @xmath26 .
let @xmath26 be a braided bialgebra in @xmath19 .
it is well known that if the braid of @xmath26 is involutive ( i.e. , @xmath36 ) , then @xmath37 is a braided bialgebra in a natural way .
the following proposition is the main result in this note .
it shows in particular that the involutivity hypothesis can be removed .
[ prop1.1 ] let @xmath26 and @xmath38 braided bialgebras in @xmath19 and let @xmath39 be a invertible arrow . if @xmath40 and @xmath41 is compatible with the bialgebras structures of @xmath26 and @xmath38 , then @xmath42 is a braided bialgebra , via * @xmath43 * @xmath44 , * @xmath45 , * @xmath46 , * @xmath47
. moreover , if @xmath26 and @xmath38 are braided hopf algebras , then so is @xmath42 , with antipode @xmath48 .
the fact that @xmath0 is an algebra and a coalgebra is well - known and standard .
we check now the compatibility between multiplication and comultiplication : @xmath49\object{h } & \step[1]\object{l } \\ \hh\operatorname{\mathsf{id } } & \xx & \step[1]\operatorname{\mathsf{id}}\\ \hh \cu & & \cu \\
\hh \cd & & \cd \\ \hh\operatorname{\mathsf{id } } & \x & \step[1]\operatorname{\mathsf{id}}\\ \hh\object{h } & \object{l } \step[1]\object{h } & \step[1]\object{l } \\
\end{tangles } \;=\ ; \begin{tangles}{lccr } \hh\step[.5]\object{h } \step[2 ] \object{l } \step[2]\object{h } \step[2]\object{l }
\\ \step[.5]\operatorname{\mathsf{id}}\step[2]\xx\step[2]\operatorname{\mathsf{id}}\\ \hh\cd \step[1]\cd \step[1 ] \cd \step[1 ] \cd \\ \hh \operatorname{\mathsf{id}}\step[1 ] \xx \step[1 ] \operatorname{\mathsf{id}}\step[1 ] \operatorname{\mathsf{id}}\step[1 ] \xx \step[1 ] \operatorname{\mathsf{id}}\\ \hh\cu \step[1]\cu \step[1 ] \cu \step[1 ] \cu \\ \step[.5]\operatorname{\mathsf{id}}\step[2]\x\step[2]\operatorname{\mathsf{id}}\\ \hh\step[.5]\object{h } \step[2 ] \object{l } \step[2]\object{h } \step[2]\object{l } \\ \end{tangles } \;=\ ; \begin{tangles}{lccr } \hh\step[.5]\object{h } \step[2 ] \object{l } \step[2]\object{h } \step[2]\object{l } \\ \hh\cd \step[1]\cd \step[1 ] \cd \step[1 ] \cd \\ \hh\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\step[1]\xx\step[1]\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\\ \hh \operatorname{\mathsf{id}}\step[1 ] \operatorname{\mathsf{id}}\step[1 ] \xx \step[1 ] \xx \step[1 ] \operatorname{\mathsf{id}}\step[1 ] \operatorname{\mathsf{id}}\\ \hh \operatorname{\mathsf{id}}\step[1 ] \xx \step[1 ] \operatorname{\mathsf{id}}\step[1 ] \operatorname{\mathsf{id}}\step[1 ] \xx \step[1 ] \operatorname{\mathsf{id}}\\ \hh \operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\step[1]\x\step[1]\x\step[1]\operatorname{\mathsf{id}}\step[1 ] \operatorname{\mathsf{id}}\\ \hh\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\step[1]\x\step[1]\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\\ \hh\cu \step[1]\cu \step[1 ] \cu \step[1 ] \cu \\ \hh\step[.5]\object{h } \step[2 ] \object{l } \step[2]\object{h } \step[2]\object{l } \\ \end{tangles } \;=\ ; \begin{tangles}{lccr } \hh\step[.5]\object{h } \step[2 ] \object{l } \step[2]\object{h } \step[2]\object{l } \\ \hh\cd \step[1]\cd \step[1 ] \cd \step[1 ] \cd \\ \hh\operatorname{\mathsf{id}}\step[1]\x\step[1]\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\step[1]\x\step[1]\operatorname{\mathsf{id}}\\ \hh\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\step[1]\xx\step[1]\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\\ \hh\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\step[1]\xx\step[1]\xx\step[1]\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\\ \hh\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\step[1]\x\step[1]\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\step[1]\operatorname{\mathsf{id}}\\ \hh\operatorname{\mathsf{id}}\step[1]\xx \step[1]\operatorname{\mathsf{id}}\step[1 ] \operatorname{\mathsf{id}}\step[1]\xx \step[1]\operatorname{\mathsf{id}}\\ \hh\cu \step[1]\cu \step[1 ] \cu \step[1 ] \cu \\ \hh\step[.5]\object{h } \step[2 ] \object{l } \step[2]\object{h } \step[2]\object{l } \\ \end{tangles } \end{aligned}\ ] ] we leave to the reader to prove that @xmath50 satisfies the braid equation , and that it is compatible with @xmath51 , @xmath52 , @xmath53 and @xmath54 . the proof of the last assertion in the statement is also straightforward .
the aim of this section is to give a more categorical proof of proposition [ prop1.1 ]
. the methods presented here could be useful in generalizing this result to braided versions of bicrossproducts , matched pairs , etc .
( see @xcite ) .
let @xmath55 be a family of objects in @xmath5 .
a family @xmath56 of isomorphisms @xmath57 is said to be _ braided _ if @xmath58 the braid equations @xmath59 are satisfied .
given such a family and a string @xmath60 of elements in @xmath61 , we call @xmath62 the _ length _ of @xmath63 , we let @xmath64 denote the string @xmath65 and we put @xmath66 . for each pair @xmath67 of such strings
we define the map @xmath68 , recursively by let @xmath81 , where @xmath82 is a family of maps compatible with @xmath56 .
we want to embed this datum into a braided category in a natural way .
to this end , let @xmath83 be the category whose objects are pairs @xmath84 , where @xmath18 is an object in @xmath5 and @xmath85 is a family of isomorphisms @xmath86 , subject to we consider @xmath103 , the center of @xmath83 , which , we recall from ( * ? ? ?
9.1.6 ) , has as objects the pairs @xmath104 , where @xmath105 is an object in @xmath83 and @xmath106 is a natural isomorphism @xmath107 such that @xmath108 an arrow @xmath109 is an arrow @xmath110 in @xmath83 such that @xmath111 we take @xmath121 .
@xmath56 consists of the arrows @xmath122 , @xmath123 , @xmath124 , @xmath125 and the canonical maps @xmath126 , @xmath127 .
we take also @xmath128 . by proposition [ pr:32 ] , these data are included , via a map @xmath129 , in a braided category .
moreover , the braiding between @xmath130 and @xmath131 is involutive , i.e. , @xmath132 .
hence , the first item of ( * ? ? ?
* corollary 2.17 ) applies to give the desired result . | we show how to define the tensor product of two braided hopf algebras . |
the _ r_-modes are non - radial pulsations modes of rotating stars that have the coriolis force as their restoring force and a characteristic frequency comparable to the rotation speed of the star @xcite .
these modes are driven unstable by gravitational radiation in all rotating stars @xcite
. a deeper understanding of @xmath1-modes and their astrophysical relevance requires taking into account nonlinear effects in the evolution of the @xmath1-mode instability .
one such a nonlinear effect , differential rotation induced by @xmath1-modes , has been studied by several authors .
rezzolla , lamb and shapiro @xcite were the first to suggest that differential rotation drifts of kinematical nature could be induced by @xmath1-modes , deriving an approximate analytical expression for this differential rotation .
afterwards , the existence of such drifts was confirmed numerically both in general relativistic @xcite and newtonian hydrodynamics @xcite .
recently , an exact @xmath1-mode solution , describing differential rotation of pure kinematic nature that produces large scale drifts along stellar latitudes , was found within the nonlinear theory up to second order in the mode s amplitude @xcite . in this paper
we discuss the role of differential rotation in the evolution of the @xmath0 _ r_-mode instability of a newly born , hot , rapidly - rotating neutron star .
a few years ago , owen _ et al . _
@xcite proposed a simple model to study the evolution of the @xmath1-mode instability in newly born , hot , rapidly - rotating neutron stars . within this model
, it is assumed that the time evolution of the system ( star and @xmath1-mode perturbation ) is characterized by two parameters : the angular velocity of the star , @xmath2 , and the amplitude of the mode , @xmath3 .
the total angular momentum of the star is then given by @xmath4 where the momentum of inertia @xmath5 of the equilibrium configuration is @xmath6 , with @xmath7 , and the canonical angular momentum @xmath8 of the _
r_-mode perturbation is @xmath9 , with @xmath10 . here ,
only the @xmath0 @xmath1-mode is being considered and it is assumed that the mass density @xmath11 and the pressure @xmath12 of the fluid are related by a polytropic equation of state @xmath13 , with @xmath14 such that the mass and radius of the star take the values @xmath15 and @xmath16 km , respectively . within this model
it is also assumed that the total angular momentum of the star decreases due to the emission of gravitational radiation and that the physical energy of the @xmath1-mode perturbation ( in the co - rotating frame ) increases due to the emission of gravitational radiation and decreases due to the dissipative effect of viscosity .
these assumptions lead then to a system of differential equations for the angular momentum of the star , @xmath17 , and the amplitude of the @xmath1-mode , @xmath18 : @xmath19 where @xmath20 , the gravitational - radiation and viscous timescales , @xmath21 and @xmath22 , are given by @xcite @xmath23 in the above expressions , the fiducial timescales are @xmath24 , @xmath25 and @xmath26 . for a newly born , hot
, rapidly - rotating neutron star there is an interval of relevant temperatures and angular velocities of the star for which the gravitational timescale is much smaller than the viscous timescale , @xmath27 .
therefore , for this interval of temperatures and angular velocities , we can neglect in the right - hand side of eqs .
( [ omega ] ) and ( [ alfa ] ) the terms proportional to @xmath28 and obtain the solution @xmath29 and @xmath30 . if the initial angular velocity is chosen to be @xmath31 , where @xmath32 is the keplerian angular velocity at which the star starts shedding mass at the equator , then @xmath33 , implying that the perturbation grows exponentially from an initial amplitude @xmath34 to values of order unity in just about @xmath35 @xcite .
after this first stage of evolution of the @xmath1-mode instability , the amplitude @xmath18 has to be forced , by hand , to take a certain saturation value @xmath36 ; in the second stage of the evolution it is then assumed that @xmath37 and @xmath17 is determined from the equation @xmath38 which yields the solution @xmath39}^{-1/6 } , \label{omega - k54}\end{aligned}\ ] ] where @xmath40 is the time at which occurs the transition from the first to the second stage of evolution . during the second stage of evolution , the star loses its angular momentum , spinning down to its final angular velocity . as can be seen from the above equations , the final angular velocity of the star depends critically on the saturation value of the mode s amplitude @xmath41 ; for instance ,
after one year of evolution , for @xmath31 and @xmath42 one obtains @xmath43 , while for @xmath44 the angular velocity is @xmath45 .
the fact that the growth of the mode s amplitude has to be stopped by hand at a certain saturation value introduces an element of arbitrariness into the solution , permitting , for instance , that agreement between the predicted final value of the angular velocity of the star and the value inferred from astronomical observations can always be achieved by fine - tuning the value of the saturation amplitude .
as will be seen in the next section , this arbitrariness disappears when differential rotation is taken into account .
differential rotation induced by @xmath1-modes _ does _ contribute to the physical angular momentum of the _
r_-mode perturbation @xcite .
therefore , a model for the evolution of the _ r_-mode instability should include the effect of differential rotation . here , we adopt the model of owen _ et al . _
@xcite , described in the previous section , with the important difference that the total angular momentum of the star is given by @xmath46 where the physical angular momentum of the _
r_-mode perturbation @xmath47 is @xcite @xmath48 in the previous expression , @xmath49 is a constant fixed by the choice of initial data and gives the initial amount of differential rotation associated with the _ r_-mode .
the specific case @xmath50 , for which the physical angular momentum of the _
r_-mode perturbation coincides with the canonical angular momentum , corresponds to the model of owen _ et al .
@xcite described in great detail in the previous section .
there is not , to our knowledge , any physical condition that forces @xmath49 to take such a particular value @xmath50 .
therefore , we will consider in this paper arbitrary values of @xmath49 in the interval @xmath51 .
results from the fact that one wishes to impose the condition that @xmath52 ( the initial amplitude @xmath53 is considered throughout the paper to be @xmath54 ) .
the lower limit for @xmath49 results from the fact that we do not want to saturate the amplitude of the mode by hand , a procedure needed for the case @xmath55 in order to avoid that the total angular momentum of the star becomes negative . ]
following the procedure described in the previous section , we arrive at a system of two first - order coupled differential equations determining the time evolution of the amplitude of the _ r_-mode @xmath3 and of the angular velocity of the star @xmath2 : @xmath56 \frac{\alpha}{\tau_{gr } } , \label{alfa-1}\end{aligned}\ ] ] valid when the damping effect of viscosity can be neglected relatively to the driving effect of gravitational radiation .
this system of equations was solved analytically and discussed in great detail in ref . [ 9 ] . in the initial stages of the evolution of the _ r_-mode instability @xmath18 increases exponentially @xcite , @xmath57 while for later times
, @xmath18 increases very slowly as @xmath58 the amplitude of the mode saturates in a natural way and a smooth transition between the regimes ( [ l - fase ] ) and ( [ m - fase ] ) occurs for @xmath59 seconds ( see fig .
[ fig : alpha ] ) .
this contrasts with the model of owen _ et al . _
@xcite , in which , after the short initial period of exponential growth , the amplitude @xmath18 has to be forced by hand to take a certain saturation value @xmath60 . .
the initial values of the amplitude of the mode and of the angular velocity of the star are , respectively , @xmath34 and @xmath31.,scaledwidth=50.0% ] as can be seen from eq .
( [ m - fase ] ) , the saturation value of the amplitude of the mode depends crucially on the parameter @xmath49 , namely , @xmath61 .
if @xmath62 , corresponding to a situation in which the initial amount of differential rotation is small , then the @xmath1-mode saturates at values of order unity . if , on the other hand , the initial differential rotation associated to _
r_-modes is significant , then the saturation amplitude @xmath41 can be as small as @xmath63 .
let us now turn our attention to the time evolution of the angular velocity of the star , @xmath17 . in the initial stages of the evolution of the _ r_-mode instability @xmath17 decreases as @xcite @xmath64 while for later times @xmath17 decreases slowly as @xmath65 the smooth transition between the regimes ( [ i - fase ] ) and ( [ f - fase ] ) occurs for @xmath59 seconds ( see fig .
[ fig : omega ] ) . .
the initial values of the amplitude of the mode and of the angular velocity of the star are , respectively , @xmath34 and @xmath31.,scaledwidth=50.0% ] remarkably , in the later phase of the evolution , the angular velocity @xmath17 does not depend on the value of @xmath49 and , consequently , does not depend on the saturation value of @xmath18 .
this contrasts with the results obtained in ref . [ 7 ] , where the value of @xmath17 depends critically on the choice of @xmath41 .
after about one year of evolution , when the dissipative effect of viscosity becomes dominant and starts damping the mode , the angular velocity of the star becomes @xmath66 ( for @xmath31 ) , in good agreement with the inferred initial angular velocity of the fastest pulsars associated with supernova remnants .
in this paper we have discussed the role of differential rotation in the evolution of the @xmath0 _ r_-mode instability of a newly born , hot , rapidly - rotating neutron star .
we have shown that , a few hundred seconds after the mode instability sets in , the amplitude of the _ r_-mode saturates in a natural way at values that depend on the initial amount of differential rotation associated to the @xmath1-mode .
if the initial differential rotation of _ r_-modes is small , then the _ r_-mode saturates at values of order unity .
on the other hand , if the initial differential rotation is significant , then the saturation amplitude can be as small as @xmath63 .
these low values for the saturation amplitude of _
r_-modes are of the same order of magnitude as the ones obtained in recent investigations on wind - up of magnetic fields @xcite and on nonlinear mode - mode interaction @xcite .
we have also shown that the value of the angular velocity of the star becomes , after a short period of evolution of the _ r_-mode instability , very insensitive to the saturation value of the mode s amplitude .
after about one year of evolution the angular velocity is @xmath67 ( for any @xmath41 ) , in good agreement with the inferred initial angular velocity of the fastest pulsars associated with supernova remnants .
we thank scar dias and luciano rezzolla for helpful discussions .
this work was supported in part by the _ fundao para a cincia e a tecnologia _ ( fct ) , portugal .
bt acknowledges financial support from fct through grant praxis xxi / bd/21256/99 . | we discuss the role of differential rotation in the evolution of the @xmath0 _ r_-mode instability of a newly born , hot , rapidly - rotating neutron star .
it is shown that the amplitude of the _ r_-mode saturates in a natural way at a value that depends on the amount of differential rotation at the time the instability becomes active .
it is also shown that , independently of the saturation amplitude of the mode , the star spins down to a rotation rate that is comparable to the inferred initial rotation rates of the fastest pulsars associated with supernova remnants . |
a number of `` problems '' have been discussed in relation to the cosmic mean metallicity and/or mean metal density .
pagel ( 1999 ) discussed the low redshift ( @xmath0@xmath10 ) `` excess metals problem '' showing that a comparison of the average metal to stellar densities in the low-@xmath0 universe indicates that the average cosmic stellar initial mass function ( imf ) has a higher yield , than what is obtained for the `` standard '' salpeter ( 1955 ) imf .
this is a well known result for ellipticals in clusters ( e.g. , romeo 2006 and references therein ) , but pagel s analysis indicates that it could be more universal .
conversely , pettini ( 1999 ) formulated the high-@xmath0 `` missing metals problem '' as follows : studies of the co - moving rest - frame uv luminosity density of high-@xmath0 lyman break galaxies ( lbgs ) allow us to trace the cosmic star formation density ( or history , sfh ) , @xmath13 , up to redshifts @xmath14 . assuming an imf of such stars
, one can compute the specific fraction of heavy elements ( `` metals '' ) they produce , @xmath15 , and derive the metal production rate @xmath16 .
the integral from @xmath17 gives , at any given @xmath0 , the density of cosmic metals @xmath18 , or , expressed in units of the critical density , @xmath19 .
moreover , if one restricts the analysis to elements , such as the @xmath9-elements , produced almost exclusively in massive stars undergoing core collapse , the apparent dependence of @xmath18 on the imf is essentially removed , since both the uv light and the oxygen production originate from massive stars . this will be discussed further in section [ sect : uv ] , and was already noticed by , e.g. , songaila ( 1990 ) , madau ( 1996 ) , pagel ( 1999 ) and pettini ( 1999 ) .
early searches in cosmic structures for which the metal / baryon mass ratio ( metallicity , @xmath20 ) can be derived either via intergalactic gas quasar absorption line experiments ( dlas or the ly@xmath9 `` forest '' ) or through direct spectroscopic studies of lyman break galaxies have found that only @xmath21 is stored in these components , the large majority of the metals are `` missing '' .
similar missing metal problems have been formulated by wolfe ( 2003 ) and prochaska ( 2003 , 2006 ) on the basis of star formation rates and metallicities of dlas alone .
ferrara ( 2005 , fsb05 ) attempted to quantify more precisely the extent of the missing metals deficit , and also to suggest where the `` missing metals '' might be found .
they found from considering typical stellar masses and metallicities of lyman break galaxies , and an estimate of the co - moving density of lbgs a contribution of @xmath22 from lyman break galaxies .
they also estimated the contribution from metals in dlas to about @xmath23 . using the sfh estimate of bouwens ( 2004b ) , assuming a salpeter imf , and applying a dust correction factor of 4.5 ( reddy & steidel 2004 ) they estimated that the universe should be characterised by a total metallicity density of @xmath24 at @xmath0=2.3 . from the above numbers they concluded that about 80% of the metals in the universe are missing , in the sense that they are not directly associated with the above two galactic components .
fsb05 suggested that the `` missing metals '' are associated with galaxies , mainly residing in their `` hot '' gas halos , having been deposited there by star - burst driven super - winds ( see also pettini 2004 ) .
this hypothesis seems reasonable , given that outflow velocities of 300 - 400 km / s are routinely inferred from the spectra of lbgs ( e.g. , pettini 2001 , shapley 2003 ) .
moreover , semi - analytical as well as fully hydrodynamical models of galaxy formation , based on cold dark matter ( cdm ) , require such `` feedback '' in order to enable the formation of realistic galaxies , solving the `` over - cooling '' , `` angular momentum '' , `` missing satellites '' and other problems ( e.g. , sommer - larsen 1999 , cole 2001 , thacker & couchman 2001 , sommer - larsen 2003 , sgp03 ) . considering the widely used probes of ly@xmath9 forest metal absorption in qso spectra , and ( e.g. , songaila 2001 , bergeron 2002 , carswell 2002 , boksenberg , sargent & rauch 2003 , schaye 2003 , simcoe 2004 , aracil 2004 , scannapieco 2005 ; and bergeron & herbert - fort 2005 , tripp 2006 ) , fsb05 inferred observed integrated and column densities over the redshift range 1.7@xmath25@[email protected] , and the corresponding cosmic average densities of these . assuming a two - phase intergalactic medium ( igm ) , consisting of a `` cold '' part ( @xmath2@xmath110@xmath26k ) , responsible for the hydrogen ly@xmath9 forest absorption , and a `` hot '' component ( @xmath2@xmath110@xmath27k )
, they showed , using a number of simplifying assumptions , that more than 90% of the metals can reside in the hot component , without violating the and ly@xmath9 forest metallicity constraints ( and with the cold phase as the major contributor to the and column densities ) .
this result follows provided that @xmath2810@xmath29k , and the density of the hot metal - enriched gas is only factors of a few times the mean cosmic baryonic density , which , e.g. , at @xmath0=3 , corresponds to a hydrogen number density of @[email protected]@xmath3010@xmath31@xmath8 .
fsb05 suggest this metal containing gas to be identified as wind - blown galaxy halo gas .
the low luminosity galaxies at high redshift are traced by the dlas ( e.g. fynbo 1999 ; haehnelt 2000 , schaye 2001a ; mller 2002 ) . from very early on it
was found that dlas contain different regions with very different temperatures and ionisation states ( turnshek et al .
the dla metal content discussed by fsb05 only accounts for the cold , mainly neutral component .
the and lines in dlas corresponds to a warmer phase with more turbulent kinematics .
the cross - section for strong and absorption is much larger than for dlas consistent with the picture that this gas is located in an extended wind - blown halo around the stellar components ( petitjean & bergeron 1994 , see also adelberger 2003 , 2005 , porciani & madau 2005 , scannapieco 2005 )
. an even hotter component , traced by and ( temperatures of order @xmath32 k ) , has recently been identified in dlas ( fox 2007 ) .
fox find that if the temperature of the bearing gas is 10@xmath27 k or higher , then this hot phase can contribute significantly to the metal mass budget qualitatively consistent with the proposal of fsb05 .
bouch ( 2006a , b ) discussed the missing metals problem at @[email protected] - 2.5 and found that the problem is not quite as severe as at @xmath0@xmath13 . counting metal contributions from stars in bx galaxies ( the equivalent of lbgs at @xmath0=2.2 ; e.g. , adelberger 2004 ) and drgs ( `` distant red galaxies '' ; e.g. , franx 2003 , van dokkum 2003 ) , as well as from gas in smgs ( sub - millimeter galaxies ; e.g. , blain 2004 , greve 2005 ) and dlas ( e.g. , pettini 2003 ) they could account for about 1/3 of the metals expected .
the bx galaxies contain the largest fraction of the identified metals , about 55% .
moreover , bouch ( 2007 ) found that another about 1/3 of the metals expected can be identified in the igm at such redshifts .
recently , dav & oppenheimer ( 2006 , do06 ) modeled the enrichment history of the universe and address the missing metals problem using a fully numerical approach .
they perform a moderate resolution cosmological hydro / gravity simulation of a cubic region of the universe of 32 @xmath33mpc box size .
their simulation invokes metallicity and uv background dependent radiative cooling , star - formation and chemical evolution , in the instantaneous recycling approximation , and based on the salpeter imf .
the simulation resolves galaxies of stellar masses down to @xmath110@xmath34 @xmath35 , corresponding to a v - band absolute magnitude @xmath36@xmath37 at @xmath0@xmath13 . due to the limited resolution
, the authors adopt a parameterized description of star - burst driven outflows in the form of a `` momentum - driven '' wind , found by oppenheimer & dav ( 2006 , od06 ) to yield the best match of the results of their simulations to various observational data .
do06 quantify their results in terms of cosmic metal fractions in five phases : a ) stars , b ) star - forming gas , c ) halo gas ( gas inside of the virial radius of galaxy halos , which is not star - forming ) , d ) shocked igm ( gas outside of galaxy halos of @xmath2@xmath123@xmath3010@xmath26 k ) , and e ) diffuse igm ( gas outside of galaxy halos of @xmath2@xmath253@xmath3010@xmath26 k ) . at @xmath0@xmath13 , they find that the dominant metal phase is the diffuse igm , which contains about 40% of the cosmic metals , with the other four phases containing approximately equal fractions of about 15% .
hence , do06 find that indeed only about 20% of the metals in the @xmath0@xmath13 universe reside in stars , while the `` missing '' about 80% are located in the gas phase .
moreover , they find that the hot halo gas is not the dominant , metal containing phase . instead , a large fraction of the missing metals are `` hidden '' in the diffuse igm .
do06 suggest that this is possible , in relation to observations of , e.g. , low density igm abundances , because the diffuse gas is somewhat hotter than assumed in previous estimates , which at gas over - densities of @xmath38@xmath39 implies larger to c ionisation corrections , and hence that larger amounts of c can be `` hidden '' in this phase . in this paper
we take another approach from that of do06 ( see also calura & matteucci 2004 , 2006 ) .
as we will show in the paper , due to resolution limitations do06 likely fail to account for about half of the metals in the universe produced by @xmath0@xmath13 , on top of which adds cosmic variance effects , given the relatively small computational box of do06 . to undertake simulations of larger cosmological volumes , at the same time probing 8 - 10 magnitudes deeper ( sec .
[ sect : cosmic ] ) is at present , as well as in any foreseeable future , computationally prohibitive .
using k - band observations of a sample of lbgs , shapley ( 2001 ) determined the @xmath0@xmath13 rest - frame v - band galaxy luminosity function . in sommer - larsen & fynbo ( 2007 , paper i ) we correct the shapley luminosity function for extinction and other effects , to obtain a `` true '' @xmath0@xmath13 lyman break galaxy lf
subsequently , cosmological high - resolution hydro / gravity simulations of the formation and evolution of individual galaxies are combined with the corrected galaxy luminosity function to obtain estimates of the average cosmic density of metals ( in particular oxygen ) residing in stars at @xmath0=3 .
we consider models based on three different stellar imfs , spanning almost a factor of five in chemical yield as well as thermal / kinetic energy feedback from supernova type ii ( snii ) explosions .
these are the kroupa ( 1998 , k98 ) imf , a typical imf suited for chemical evolution models of the solar neighbourhoohd , e.g. , boissier & prantzos ( 1999 ) , the `` standard '' salpeter ( 1955 , s ) imf and the arimoto - yoshii ( 1987 , ay ) imf , which is well suited for describing the chemical evolution of elliptical galaxies .
the simulations have sufficiently high resolution to allow a two - phase modeling of the star - forming interstellar medium ( ism ) , consisting of a `` cold '' @xmath2@xmath110@xmath26 k star - forming phase and a `` hot '' @xmath2@xmath110@xmath40- 10@xmath27 k phase , intermixed with the cold gas .
star - bursts drive galactic winds by depositing thermal energy from multiple snii explosions in the gas , part of which is subsequently converted into kinetic energy self - consistently by the hydro - code . in paper
i we conclude that for none of the imfs is it possible to reconcile the amount of oxygen locked in stars with the amount predicted from the ( observed ) cosmic uv luminosity density history , hence confirming the `` missing metals problem '' as stated by pettini ( 1999 ) and fsb05 .
in this paper we combine the high - resolution galaxy formation models with the v - band luminosity function to determine the cosmologically averaged amount and properties of @xmath0=3 gas - phase metals , mostly focusing on oxygen , but also on carbon , in particular in relation to qso absorption line determinations of the cosmic density , as well as iron , in relation to dla abundances . combining the results obtained with those of paper
i the total cosmic metal distribution is obtained for the three imfs considered .
comparing in turn these results to inferences from the observed cosmic uv luminosity density history allows us to significantly constrain the `` true '' cosmic stellar imf .
finally , relations to the `` missing metals problem '' , as well as results obtained by other authors , are discussed .
the paper is organised as follows : in section [ sect : uv ] we derive constraints on cosmic metal production from the cosmic uv luminosity density history , in section [ sect : simulations ] we briefly describe the hydro / gravity galaxy formation simulations , and in sections [ sect : approach ] and [ sect : cosmic ] we present the approach used in this paper to determine the temperature distribution of the cosmic gas - phase metals .
section [ sect : civ ] presents results on gas - phase abundances , and , relating these to observations of absorption lines in qso spectra , we derive a constraint on the cosmic imf . in section [
sect : comb ] we combine the results of paper i and those obtained here to derive the total cosmic metal distribution . comparing this
to what is obtained from the cosmic uv luminosity density history we obtain an additional constraint on the cosmic stellar imf . in section [
sect : mm ] we relate our results to the `` missing metals problem '' and compare them to those of other workers in the field , in section [ sect : numres ] we demonstrate that our results are robust to changes of the numerical resolution and , finally , section [ sect : conclusions ] summarises our conclusions . in the paper
we assume the flat @xmath41 cosmology , with @xmath42 = 0.3 and @xmath43=0.7 , and @xmath44 km / s / mpc , with @xmath45=0.7 , unless it is explicitly stated otherwise .
constraints on metal production from the cosmic uv luminosity density history are discussed in paper i , but for coherence we present the main discussion in the following as well .
the luminosity of young galaxies at rest - frame @xmath461500 is a measure of the rate of formation of massive stars ( mainly o and b type ) as shown by , e.g. , madau ( 1998 ) .
moreover , @xmath9-elements like oxygen are almost exclusively produced in such massive stars , and hence there is a direct link between the cosmic average oxygen production rate density and the average uv luminosity density , which is essentially independent of the stellar initial mass function see , e.g. , pettini ( 1999 ) . to estimate the cosmic oxygen production rate density as a function of redshift we use various recent estimates of the average cosmic star formation rate .
the estimates have been obtained from steidel ( 1999 ) , bouwens ( 2004a , b ) , bunker ( 2004 ) , giavalisco ( 2004 ) , schimonovich ( 2005 ) , bouwens ( 2006 ) and sawicki & thompson ( 2006 ) .
the estimates used are all based on assuming a `` standard '' salpeter imf in converting from uv luminosity to star formation rate . the oxygen ( mass ) yield of a salpeter imf is 0.01 ( e.g. , lia 2002a , b ) hence in fig .
[ fig : ofr ] we have multiplied the published sfrs by 0.01 to obtain the ( essentially ) imf independent oxygen production rate densities .
the estimates have moreover been corrected to correspond to ( at any redshift ) the uv luminosity density of galaxies brighter than 0.1@xmath47 following an approach similar to that of bouwens ( 2006 ; note though that the @[email protected] data point of bouwens 2004b only represents @xmath480.3@xmath47 ) .
finally , the values shown in the plot result from multiplying the observed values by a dust - correction factor of 5.5 .
this value is intermediate between the @xmath493 values of 4.5 and 6.5 suggested by reddy & steidel ( 2004 ) and dahlen ( 2007 ) , respectively ( but see below ) .
the cosmic average oxygen density at @xmath0=3 can now be obtained by integrating the oxygen production rate density , @xmath50 , from @xmath0=3 and back in time .
we shall consider three models for @xmath50 : the `` median '' model , the `` maximum '' model , and the `` minimum '' model .
the median model is obtained by calculating the median values of @xmath51 in the bins @xmath0=3 to 4 and @xmath0= 4 to 6 ( assigning equal weight to each data point , but excluding the @xmath0=5.9 value of bunker 2004 ; see below ) , and connecting these values with the @xmath527.5 value of bouwens ( 2004b ) . this model
is shown in fig .
[ fig : ofr ] by the long - dashed curve .
the maximum model is obtained by connecting the @xmath0=3 to 4 median value with the giavalisco ( 2004 ) @xmath525.7 value , and the bouwens ( 2004b ) @xmath526.0 value , and multiplying the resulting `` upper envelope '' by a factor 6.5/5.5 to maximize also the dust correction .
this model is shown by the upper short - dashed curve in fig .
[ fig : ofr ] .
finally , the minimum model is obtained by connecting the @xmath0=3 to 4 median value with the giavalisco ( 2004 ) @xmath524.8 value , the bouwens ( 2004a ) value and the bouwens ( 2006 ) @xmath527.5 value , and multiplying this `` lower envelope '' by a factor 4.5/5.5 to minimize the dust correction ( but see below ) .
this model is shown by the lower short - dashed curve in fig .
[ fig : ofr ] . assuming a flat space world model , the @xmath0=3 cosmic average oxygen density can now be evaluated as @xmath53 @xmath54 where values of @xmath44 km / s / mpc , with @xmath45=0.7 , @xmath42 = 0.3 and @xmath43=0.7 are assumed in this paper . using eq.1
we obtain @xmath55=3 ) = 0.32 , 0.42 and 0.56 @xmath56 @xmath35/(@xmath33mpc)@xmath57 for the `` minimum '' , `` medium '' and `` maximum '' models , respectively .
these numbers represent an integral constraint that any successful model of galaxy formation must meet .
as will be shown in section [ sect : cosmic ] this enables us to strongly constrain the `` cosmic '' stellar imf at @xmath583 . in determining the `` median '' and `` minimum '' models we excluded the @xmath0=5.9 value of bunker ( 2004 ) .
this is simply done on the basis of the large discrepancy between this value , and all other @xmath0=4 to 6 values . to indicate the effect of this
, we determined an alternative median model in which the bunker ( 2004 ) value is included in the @xmath0=4 to 6 bin .
this changed the median model estimate of @xmath55=3 ) from 0.42 to 0.40 @xmath56 @xmath35/(@xmath33mpc)@xmath57 , i.e. a @xmath12 % change , and hence quite small effect , which , given all other uncertainties , we shall ignore in the following .
bouwens ( 2006 ) propose that the @xmath461350 - 1500 extinction correction decreases with increasing @xmath0 from @xmath493 .
converting their proposed , extinction - corrected sfr(@xmath0 ) to a corresponding @xmath50 ( including changing the uv luminosity density from their limit of 0.04@xmath47 to the limit of 0.1@xmath47 used here ) results in a value of @xmath55=3 ) = 0.33 @xmath56 @xmath35/(@xmath33mpc)@xmath57 , hence within the bounds derived above .
the integral constraint obtained in this section can also be expressed in units of the critical density : we obtain @xmath59(@xmath0=3 ) = @xmath55=3)/@xmath60 = 0.81 , 1.05 and 1.41@xmath61 for the minimum , median and maximum models , respectively .
fsb05 obtained for the total density of heavy elements in units of the critical @xmath62 = ( [email protected])@xmath61 , by integrating the observed cosmic star formation history ( sfh , based on a salpeter imf ) , as presented by bouwens ( 2004b ) , to @xmath0=2.3 and applying a dust correction factor of 4.5 .
the above value translates into @xmath59(@xmath0=2.3 ) = ( [email protected])@xmath61 .
if the oxygen density production rate models shown in fig . [ fig : ofr ] are continued @xmath0=2.3 , we would obtain 1.22 , 1.55 and 2.01 @xmath61 .
the reason for the apparent discrepancy is two - fold : a ) we use a dust correction of a factor of 5.5 , and b ) more importantly , the bouwens ( 2004b ) sfh was based on a uv luminosity to a limit of 0.3@xmath47 , rather than the limit of 0.1@xmath47 adopted in this work . using the @xmath64 uv luminosity function of adelberger & steidel ( 2000 )
we estimate the latter correction to be a factor of 1.5 .
multiplying this by a factor of 5.5/4.5 the findings of fsb05 would translate into about ( [email protected])@xmath61 , in good agreement with our values above .
extremely dust - obscured galaxies such as high-@xmath0 scuba sources ( e.g. , smail 1997 ; barger 1998 ; eales 1999 ; chapman 2005 ) are not included in the dust - corrected star formation rate densities shown in fig . [
fig : ofr ] .
most such galaxies have uv luminosities that heavily under - represent their star formation rates and thus such galaxies are not properly included in the estimate of the oxygen production rate density .
their numbers may be sufficiently large that their contribution is significant , however we neglect this contribution for two reasons : a ) the contribution from such galaxies would be largest at @xmath0@xmath12 , where the redshift distribution of scuba sources appears to peak ( chapman 2005 , reddy 2005 ) , whereas we are concerned with the range @xmath0@xmath73 , and b ) more importantly , we are concerned with the oxygen production rate density history of lyman break galaxies _ only _ , as we wish to compare the integral constraint thus obtained to results of combining the observed lbg @xmath0@xmath13 optical rest - frame luminosity function with detailed , high - resolution models of galaxy formation .
the code used for the simulations is a significantly improved version of the treesph code we used for our previous work on galaxy formation ( sgp03 ) .
a similar version of the code has been used recently to simulate clusters of galaxies , and a detailed description can be found in romeo ( 2006 ) .
here we briefly mention its main features and the upgrades over the previous version of sgp03 see also sommer - larsen ( 2006 ) . 1
. the basic equations are integrated by incorporating the `` conservative '' entropy equation solving scheme of springel & hernquist ( 2002 ) , which improves the numerical accuracy in lower resolution regions .
cold high - density gas is turned into stars in a probabilistic way as described in sgp03 . in a star - formation event
an sph particle is converted fully into a star particle .
non - instantaneous recycling of gas and heavy elements is described through probabilistic `` decay '' of star particles back to sph particles as discussed by lia ( 2002a ) . in a decay event
a star particle is converted fully into a sph particle , so that the number of baryonic particles in the simulation is conserved .
non - instantaneous chemical evolution tracing 10 elements ( h , he , c , n , o , mg , si , s , ca and fe ) has been incorporated in the code following lia ( 2002a , b ) ; the algorithm includes supernov of type ii and type ia , and mass loss and chemical enrichment from stars of all masses .
most of the simulations presented in this paper have been undertaken using three different initial mass functions : the kroupa ( 1998 ) imf ( denoted k98 in the following ) , derived for field stars in the solar neighborhood , the standard salpeter imf ( s ) , and the more top - heavy arimoto & yoshii ( 1987 ) imf ( ay ) , which is well suited for the modeling of elliptical galaxies , as well as galaxy clusters .
more detail is given in lia ( 2002a , b ) .
4 . atomic radiative cooling
is implemented , depending both on the metallicity of the gas ( sutherland & dopita 1993 ) and on the meta galactic uv field , modeled after haardt & madau ( 1996 ) . moreover ,
a simplified treatment of radiative transfer , switching off the uv field where the gas becomes optically thick to lyman limit photons on scales of @xmath1 1 kpc , is invoked .
star - burst driven winds are incorporated in the simulations at early epochs ( @xmath0@xmath75 - 6 ) , as strong early feedback is crucial to largely overcome the angular momentum problem ( sgp03 ) .
a burst of star formation is modeled in the same way as in sgp03 : when a star particle is formed , further self - propagating star formation is triggered in the surroundings ; the energy from the resulting , correlated snii explosions is released initially into the interstellar medium as thermal energy , and gas cooling is locally halted to reproduce the adiabatic super
shell expansion phase ; a fraction of the supplied energy is subsequently converted ( by the hydro code itself ) into kinetic energy of the resulting expanding super winds and/or shells .
the super shell expansion also drives the dispersion of the metals produced by type ii supernov ( while metals produced on longer timescales are restituted to the gaseous phase by the `` decay '' of the corresponding star particles , see point 2 above ) .
+ at later epochs , only a fraction ( typically , 20% ) of the stars induce efficient feedback , and star formation is no longer self propagating so that no strong star - bursts are triggered by correlated sn explosions .
this allows the smooth settling of the disc ( see sgp03 for all details ) .
star formation efficiencies have been chosen such that realistic disc galaxy gas fractions result at @xmath0=0 ( sgp03 ) , and can also be shown to match the kennicutt ( 1998 ) star formation relation quite well .
+ agn driven feedback has not been invoked in the simulations , as it is unlikely to play a major role in the formation of , at least , disc galaxies ( e.g. , sommer - larsen 2006 ) the most common galaxy type in the universe .
the higher stellar uv escape fractions as well as lyc luminosities found by razoumov & sommer - larsen ( 2007 ) at @xmath65 , compared to @xmath66 , also support the notion that massive stars in galaxies become progressively more important sources of ionizing photons as one goes back in time , as the comoving number density of quasars declines rapidly at @xmath67 ( e.g. , richards 2006 ) .
this provides additional circumstantial evidence that agn feedback is of minor importance , at least at @xmath68 , the relevant redshift range in this paper .
the galaxies were drawn and re - simulated from a @xmath69 mpc box - length dark matter ( dm)-only cosmological simulation , based on the `` standard '' flat @xmath41 cold dark matter cosmological model ( @xmath70 , @xmath71 , @xmath72 ) ; our choice of @xmath45 and @xmath73 is slightly different from presently more popular values ( 0.7 and 0.9 respectively ) , but this has little impact on the resulting galaxy properties ( e.g. , portinari & sommer - larsen 2007 ) . when re - simulating with the hydro - code , baryonic particles were added " to the original dm ones , which were split according to an adopted baryon fraction @xmath74 .
the gravity softening lengths were fixed in physical coordinates from @xmath0=6 to @xmath0=0 and in co - moving coordinates at earlier times .
the simulations are run with resolutions of @xmath75 - 4.5@xmath56 @xmath35 , @xmath76 - 3.3@xmath56 @xmath35 and @xmath77 kpc going from the smallest individual galaxy simulations to the largest proto - cluster simulation .
each region containing an individual proto - galaxy was simulated using 0.2 - 2.2 million particles in total ; in addition two `` low '' and `` medium '' galaxy density regions were simulated using 1.4 and 1.6 million particles , a proto - elliptical region using 1.4 million particles , and a proto - cluster region using 2.3 million particles ( the high resolution `` virgo '' simulation of romeo 2006 ) .
images of some of the simulated galaxies are available at http://www.tac.dk/@xmath1jslarsen .
the approach adopted is , based on the simulations , to characterize a given galaxy at @xmath0=3 and of absolute magnitude @xmath36 by a distribution of ism / igm oxygen mass @xmath78 , where @xmath79=3.5 , 4 , 4.5 , .. , 7 , and a given value of @xmath79 corresponds to gas temperatures in the range @xmath80 $ ] k. almost no oxygen is found at temperatures lower than 10@xmath81 k , mainly because the radiative cooling function is effectively truncated below about 8000 k. if molecule formation and molecular cooling had been invoked in the hydro / gravity simulations , part of this gas would be in a much colder , high - density , predominantly h@xmath82 phase ( see also sec . [
sect : dla ] ) .
moreover , it is found that the @xmath79=3.5 as well as 4 ism / igm oxygen is spatially associated with the stellar galaxies ( similar spatial extents ) , so the @xmath79=3.5 and 4 bins have been merged , and are denoted by @xmath79=4 .
hence , all gas colder than 10@xmath26 k is included in this phase .
moreover , for field galaxies at @xmath0@xmath13 almost no oxygen is found in the @xmath79@xmath837 phases , which will therefore be ignored in the following . once the functions @xmath78 have been determined for a given imf , the cosmologically averaged ( gaseous ) oxygen temperature distribution can be determined by folding these functions with the ( dust corrected ) rest - frame v - band luminosity function at @xmath64 .
we stress that @xmath78 pertains only to the galaxy itself , not including any of its satellite galaxies see further below . to determine @xmath78 we proceed as follows : in fig .
[ fig : omassr ] ( left ) is shown the cumulative mass of ism / igm oxygen , @xmath84 , around three ( mostly ) isolated galaxies of @xmath36 = @xmath85 , @xmath86 and @xmath87 , respectively , simulated using the k98 imf . fig .
[ fig : omassr ] ( right ) is similar , but for the _ same _ three galaxies , of @xmath36 = @xmath88 , @xmath89 and @xmath90 , simulated with the ay imf for all three galaxies the region from @xmath91 to 8 @xmath91 is shown , is the radius corresponding to , at @xmath0=3 , an over - density 181 times the mean matter density of the universe at this redshift , as appropriate for a top hat collapse in the adopted @xmath41cdm cosmology ( e.g. , bryan & norman 1998 ) . ] , and the different curves correspond to @xmath79=4 , 4.5 , .. , 7 .
as can be seen from the figure , for a given galaxy and @xmath79 , the cumulative gas oxygen mass gently increases with @xmath92 by factors @xmath11.5 - 4 , over the radial range shown .
the one exception is the @xmath79=4 case , where the gas metals are mainly spatially correlated with the central galaxies , and the curves are quite flat ( with the exception of the very bright galaxy , which has been selected from the proto - cluster simulations ) .
we stress that the mass of gas - phase oxygen around a given ( isolated ) galaxy has been contributed not only by the stars in the galaxy itself , but also by stars in all its satellite galaxies ( and in principle also by `` inter - galactic '' stars , as will be discussed below ) .
we select a sample of `` base '' galaxies consisting of ( mainly ) isolated galaxies ranging in absolute v - band magnitude from about @xmath93 to @xmath94 .
it is assumed that in order to account for the @xmath78 of a given base galaxy one should include all gas oxygen mass inside of a radius @xmath95 , to be determined in the following , such that @xmath78=@xmath96 .
`` isolated '' galaxies are in this paper defined as galaxies not containing companions of stellar mass larger than 1/3 of the mass of the galaxy itself within 8 @xmath91 .
the very brightest galaxies , however , of @xmath36@xmath97 , were selected from the higher density proto - cluster simulations , and the criterion was relaxed to not containing companions of stellar mass larger than 1/2 of the mass of the galaxy itself within 4 @xmath91 .
the relation between @xmath91 and absolute v - band ( rest - frame ) magnitude @xmath36 is shown in fig . [
fig : rvir ] for the three imfs under consideration . for a given galaxy
, @xmath98 will be finite for two reasons : a ) no galaxy is perfectly isolated , such that , even for isolated galaxies , with increasing @xmath98 increasing amounts of gas - phase oxygen associated with other ( smaller ) companion galaxies are included , and b ) for certain values of @xmath79 ( depending on the imf ) galaxies are found to `` share '' a common pool of gas - phase oxygen . in practice
it is found that for none of the imfs considered , and any @xmath79 , does @xmath98 exceed 8 .
for the brightest galaxies , @xmath36@xmath97 , @xmath98 does not exceed 4 .
[ fig : t ] shows , for @xmath79=4.5 , 5 , 5.5 and 6 , the spatial distribution of the gas - phase oxygen for a region containing 52 resolved galaxies with @xmath36 in the range @xmath99 to @xmath100 , simulated using the k98 imf .
each dot represents , for a given @xmath79 , an equal amount of gas - phase oxygen mass , so the figure displays the actual distribution of gas - phase oxygen mass , _ not _ gas mass .
as can be seen , the spatial oxygen mass distributions depend strongly on the gas temperature . in fig .
[ fig : imfs ] is shown , for the three imfs considered , the distribution of @xmath79=6 oxygen mass . as can be seen
, the oxygen mass distribution also depends strongly on the imf adopted . for comparison
is also shown the corresponding plot for a k98 simulation of a proto - elliptical , high - density region , which is biased towards many large galaxies ( bottom left ) it is seen how the @xmath79=6 oxygen mass distribution in this case is concentrated near the large galaxies , more so than for the more representative `` field '' galaxy case .
we now approximate @xmath101 by @xmath102 , independent of @xmath36 .
we show further below that this is a good approximation , except for the very brightest galaxies , for which a correction has to be made . for each imf , @xmath102 is determined as follows : i ) for a number of regions containing up to about 200 resolved galaxies we a ) determine the absolute v - band magnitudes of the @xmath103 galaxies in the region , @xmath104 , and b ) the total gas - phase oxygen mass temperature distribution of the region @xmath105 ii ) for a given set of base galaxies , and given values of the @xmath102 s , one can predict the total mass of oxygen as a function of @xmath79 as @xmath106 where @xmath107 is determined by linear interpolation in @xmath36 as well as @xmath102 ( @xmath98=0,1, ... 8 ) using the corresponding values for the base galaxies .
the values of @xmath102 are then obtained such as to provide the best simultaneous match of @xmath108 to @xmath109 for a number of simulated regions , using standard least squares fitting .
for all three imfs this is done for the _ same _ two regions , representing `` low - density '' and `` medium - density '' `` field '' environments , with galaxies of @xmath110 .
the two regions are simulated using a total of 1.4 and 1.6 million particles , respectively . for the ay imf ,
two additional regions centered on smaller galaxies are also used ( contrary to the k98 and s imf cases , for the ay simulations such small galaxies have substantial amounts of not only @xmath79=4 - 6 , but also @xmath79=6.5 and 7 gas - phase oxygen associated ) these regions are each simulated using about 300,000 particles in total .
finally , to test the approach also in a proto - cluster environment , for all three imfs the same such region was simulated using a total of 2.3 million particles ( the high - resolution `` virgo '' region of romeo 2006 ) .
as will be shown in the next section , in particular for @xmath79=4.5 , 5 and 5.5 , a significant contribution to the cosmic gas - phase oxygen budget comes from small galaxies with @xmath111 to @xmath112 . to check
that the @xmath102 fits also work in this range , comparison was made to combined data - sets of regions with @xmath36 of the central galaxy ranging from @xmath113 to @xmath112 , all simulated at high resolution , with typically 200 - 300000 particles per region .
the results for the k98 imf are shown in fig .
[ fig : gmpk ] , including also the result for a proto - elliptical region . in general , the predicted distributions match the actual ones very well , except for the very smallest galaxies , where the predicted distribution falls somewhat short of the actual one in the @xmath79=4 , 4.5 and 5 temperature bins .
this implies that the results obtained in the following section for the faintest galaxies are actually lower limits .
given the small magnitude of the effect we shall , however , neglect this in the following . in fig .
[ fig : gmpay ] a similar comparison is shown for ay imf simulations .
again , the match between predicted and actual distributions is , in general , good . for the smallest galaxies the distributions are somewhat different
( see above ) . for the proto - cluster region ,
the predicted distribution exceeds the actual one , with @xmath108 being up to about twice as large as @xmath109 .
this is due to the very large galaxy density in such environments .
using this region to solve for an alternative set of @xmath102 results in @xmath102 values of about half the ones obtained as described above . as the main contribution to @xmath109
is associated with the larger galaxies in the region , @xmath114 to @xmath115 , @xmath109 in all regions can be matched by a model , where @xmath102 changes linearly from the `` field '' values to the `` proto - cluster '' values between @xmath36 = @xmath116 to @xmath117 , as shown in fig .
[ fig : gmpk ] .
however , in any case this correction for the very large galaxies is of no consequence to the main results presented in the following section .
this is because the contribution from such galaxies to the overall cosmic metal budget is very minor , as will be detailed in the following section .
finally we note , that even in typical `` field '' environments , some fraction of the stellar systems originally formed have subsequently been disrupted through tidal stripping and other dynamical processes .
the gas - phase metals produced by these `` inter - galactic '' stars are `` automatically '' accounted for by the above approach , but in any case the fraction of such stars is quite small , @xmath1185 - 10% ( paper i ) .
the average cosmic density of gas - phase oxygen at @xmath0=3 , as a function of gas temperature , can now be obtained as @xmath119 where @xmath120 is the rest - frame v - band , extinction corrected galaxy luminosity function at @xmath0=3 . in paper
i the `` true '' @xmath493 galaxy luminosity function was determined by extinction correcting a modified version of the observationally determined luminosity function ( lf ) of shapley ( 2001 ) of faint end slope @xmath121 .
the original shapley ( 2001 ) lf , as well as the extinction corrected , modified lf , are shown in fig .
[ fig : lf ] . also shown
are the corresponding lfs obtained by fitting a lf of faint end slope @xmath122 ( see below ) to the shapley ( 2001 ) data .
the faint end slopes of these @xmath64 lfs are steeper than what is found for @xmath123 lfs . results of semi - analytical galaxy formation models ( e.g. , lacey 2005 ) indicate that this is a natural outcome of hierarchical strucure formation scenarios . using the extinction corrected ,
modified shapley lf in eq.4 , and the @xmath124 functions determined for the three imfs as described in the previous section , results in the distributions shown in fig .
[ fig : cosmic ] . for the k98 imf ,
the amount of oxygen in the @xmath1254.5 phase is about 1.1 times the amount in the @xmath79=4 `` cold '' phase . for the s and ay imfs ,
the corresponding numbers are about 1.7 and 3.2 , respectively .
hence , in particular , for the latter two imfs , the majority of the gas - phase oxygen is in the `` warm''/hot rather than cold phase .
moreover , for the ay imf , the hot to cold / warm gas - phase oxygen mass ratio is 1.1 , so the bulk of the oxygen is actually in the hot @xmath79@xmath75 phase , rather than in the cold / warm phase .
these are some of the main results of this paper . in fig .
[ fig : cosmic ] is also shown the results of adopting a galaxy luminosity function with less steep faint end slope than the shapley one .
following paper i , results are presented for a luminosity function of faint end slope @xmath122 , as found for the rest - frame uv @xmath493 luminosity function by , e.g. , adelberger & steidel 2000 ) . although the results for this lf are qualitatively similar to the ones described above , the amounts of warm and hot phase oxygen decrease somewhat compared to the cold phase ones .
the warm / hot to cold gas - phase oxygen mass ratios are for this lf 0.8 , 1.1 and 2.1 for the k98 , salpeter and ay imfs , respectively . moreover , for the ay imf
, the hot to cold / warm gas - phase oxygen mass ratio is unity . an obvious question is now : for the different gas temperatures , which galaxies contribute mainly to the cosmic gas - phase oxygen budget ? in fig .
[ fig : cosmic_abunt_m_v ] this is shown using the modified shapley luminosity function , for the k98 and ay imfs , respectively .
it is seen that for the k98 imf the main contribution to the @xmath79=4.5 phase comes from small galaxies , @xmath126 , and for the ay imf this is the case for the @xmath79=5 and 5.5 phases as well ( the reason for the large difference between the @xmath36-distributions of @xmath79=4 and @xmath79=4.5 , say , gas - phase metals is that , whereas the @xmath79=4 gas is of fairly high density and associated with all galaxies , @xmath79=4.5 gas is typically diffuse and associated with halos of comparable virial temperature , i.e. fairly small galaxies ) .
given this , and in relation to observational studies , it is clearly of interest to determine how large a fraction of the gas - phase oxygen is associated with galaxies to a certain limiting magnitude . in fig .
[ fig : ocum ] this is shown by the long - dashed curves ( for the three imfs considered ) for the @xmath121 and @xmath127 lfs .
typical lyman break galaxy metallicity determinations at @xmath493 only probe to about one magnitude below @xmath128 , @xmath129 . for the @xmath121
lf , it is found that less than about 20% of the cosmic gas - phase oxygen will be associated with galaxies brighter than this , irrespective of the choice of imf . for the @xmath122 lf
the corresponding fraction is about 30% .
hence , most of the cosmic gas - phase oxygen is associated with galaxies considerably fainter than @xmath128 , and is also in this sense `` missing '' . in calculating the results shown in fig .
[ fig : ocum ] it has been assumed that the observational lfs maintain a constant faint end slope down to @xmath130 . as can be seen from the figure ,
in particular for the @xmath121 case this is somewhat critical to the shape of the curves shown . at redshifts @xmath0@xmath13 the lbgs luminosity function
is only probed down to a few magnitudes below @xmath128 ( but see below ) . selecting galaxies using the ly@xmath9 line it is possible to probe 2 - 3 magnitudes fainter still ( fynbo 2001 , 2003 ; gawiser 2006 ; nilsson 2007 ) .
this , however , only demonstrates the existence of star - forming galaxies at these redshifts with m@xmath131 to @xmath132 , but the data are not good enough to put strong constraints on the shape of the luminosity function .
jakobsson ( 2005 ) find that the magnitudes of @xmath133 grb selected galaxies are consistent with being drawn from the steep luminosity function , but the statistics are still poor .
we note , however , that bouwens ( 2007 ) were actually able to probe the uv luminosity function at @xmath134 down to 4 - 5 magnitudes below @xmath135 and find steep faint end slopes , @xmath136 .
our calculations show , however , that if the slope of the faint end lf flattens _ significantly _ at some brighter @xmath36 , then the curves still fairly well represent the result , after appropriate renormalisation at this limiting @xmath36 ( if the faint end slope is assumed to be constant to even fainter than @xmath130 , the results are not much changed compared to the @xmath130 results ) .
moreover , we note that neither the two - degree field galaxy redshift survey ( 2dfgrs ) @xmath137 local luminosity function ( norberg 2002 ) , nor the sloan digital sky survey ( sdss ) u - band local luminosity function ( baldry 2005 ) show any flattening of the faint end slope to 5 - 6 magnitudes fainter than @xmath138 . moreover , some galaxy cluster luminosity functions tend to show a _ steepening _ faint end slope : milne ( 2007 ) find a steepening of the slope in the coma cluster at a @xmath139 range of about @xmath113 to @xmath140 , and yamanoi ( 2007 ) find a similar result for the hydra i cluster . given the significance of the results obtained in this section on the distribution of cosmic gas - phase metals , it is important to consider the following two sources of potential model dependence .
first , it is possible that the results depend on the specific choice of `` base '' galaxies .
we verified explicitly , however , that choices of alternative sets of isolated sets of `` base '' galaxies , and subsequently going through the procedure described in the previous section , lead to very similar results as the ones described here .
hence , the results presented are very robust to this .
second , it would be expected that the amount of ( in particular ) @xmath79=4 cold gas and metal mass available in a simulation at any given time would depend on the star - formation efficiency adopted ( sec .
[ sect : simulations ] ) . to this end , three k98 galaxy simulations , of galaxies of @xmath0=3 @xmath36=@xmath141 , @xmath142 , and @xmath143 and surrounding regions ,
were run with twice the standard star - formation efficiency . comparing the results of these simulations to the corresponding standard ones
, it was found that the total mass of gas - phase oxygen at @xmath0=3 was virtually unchanged , and that the mass of @xmath79=4 oxygen was reduced by @xmath1185 - 10% compared to the standard case .
hence , none of the main conclusions about @xmath79=4 metals , obtained in this and the following sections , are affected by such a change . moreover
, a doubling of the star - formation efficiency would lead to too small @xmath0=0 disc galaxy gas fractions , and likely also too small disc galaxy angular momenta ( sgp03 ) , likely too small @xmath0@xmath13 dla cross - sections ( ellison 2007 , sommer - larsen 2007 ) , and star - formation rates above the kennicutt ( 1998 ) relation .
interesting tests of the abundance and thermal properties of the low to intermediate density igm are provided by oso absorption line studies of the lyman metal forest , as probed by .
songaila ( 2001 ) first determined @xmath144 ( ) by integrating the total column density of systems between @xmath145()@xmath146 @xmath147 see also schaye ( 2003 ) .
as shown by , e.g. , od06 , this range of column densities trace moderate igm gas over - densities , @xmath148 , or @xmath149 @xmath8 , at @xmath0@xmath13 . to obtain a corresponding estimate of @xmath144 ( ) based on our simulations we adopted the following approach : to convert from total c ( as given in our simulations ) to abundances ,
ionisation corrections are required .
we base our conversion on ionisation corrections determined by od06 using cloudy ( ferland 1998 ) .
od06 assumed the haardt & madau ( 2001 ) @xmath0@xmath13 uv background field divided by a factor 1.6 at all redshifts , in order to match the observed mean ly@xmath9 flux decrement , @xmath150 .
this may seem inconsistent with our use of the haardt & madau ( 1996 ) uvb in the cosmological simulations , but as shown by croft ( 1998 ) , because photo - ionisation is sub - dominant in gas dynamics such a correction yields virtually identical results as having done the simulations with the above ( reduced ) background .
the resulting ionisation corrections were kindly supplied to us by ben oppenheimer and romeel dav , in the form of look - up tables . in fig .
[ fig : civ ] ( left ) is shown the cumulative @xmath144 ( ) as a function of @xmath151 .
for the two simulated `` field '' galaxy regions the average ratio of the cumulative mass to total ( gas - phase ) oxygen mass has been determined .
this has then been multiplied by @xmath152 , determined on the basis of the results obtained in the previous section .
as is clear from this section , the value of @xmath152 depends on the actual form of galaxy luminosity function . to obtain lower and upper limits on @xmath152
, we evaluate it at @xmath36=@xmath153 and @xmath154 , respectively , assuming the @xmath121 luminosity function .
@xmath36=@xmath153 corresponds approximately to the detection limit of shapley ( 2001 ) , whereas @xmath36=@xmath154 corresponds to assuming that the faint end slope remains constant to this very faint magnitude . as can be seen from fig .
[ fig : ocum ] , using the @xmath122 lf provides less conservative limits . in the figure
is shown the result of applying the lower and upper limits for the k98 , salpeter and ay imfs , respectively .
using other simulated regions , centered on smaller galaxies , yields similar results . using the proto - cluster region to determine the average ratio of the cumulative mass to total ( gas - phase ) oxygen mass results in lower values of @xmath144 ( ) , due to the temperature dependence of the c/ ionisation correction combined with the higher average igm temperature of the proto - cluster region .
however , proto - cluster regions are very special sites , and , as can be seen from the results of the previous section , the contribution to the cosmic metal production associated with the very luminous galaxies typical of such regions , @xmath155 , is very small ( in the proto - cluster regions containing 100 - 200 galaxies , almost half of the gas - phase metals are associated with five galaxies of @xmath156 ) . also shown in the figure
is the median value of 10 observational estimates in the redshift range @xmath0=2 - 4 , where the run of observed @xmath144 ( ) with redshift is essentially flat ( e.g. , songaila 2001 , schaye 2003 , songaila 2005 ) , as well as the observational variance .
the observational estimates were taken from songaila ( 2001 ) , boksenberg , sargent & rauch ( 2003 ) and songaila ( 2005 ) .
as can be seen , the chemical yield of the k98 imf appears too low to match the observational estimates , whereas the salpeter and ay imfs can match the observations , and an imf with a yield in between these two appears optimal .
we caution , however , that our estimates of @xmath144 ( ) depend on the adopted gas - phase oxygen to mass conversions . the optimal would be to carry out simulations of @xmath7100 @xmath33mpc box size cosmological volumes at the ( high ) resolution of our galaxy formation runs , but as mentioned previously , this is currently computationally prohibitive .
hence , our result above on the cosmic imf should be seen as indicative only .
the main point is that both the salpeter and ay imf simulations can match the observational constraints on @xmath144 ( ) .
observations of line widths for the above column density range can be used to probe the thermal properties of the low - density igm .
three components contribute to the line widths : a ) thermal broadening , b ) spatial broadening due to hubble expansion across the physical extent of the absorber , and c ) turbulent broadening .
the last component can be ignored , since the igm at these over - densities is very quiescent ( rauch 2005 ) .
observational estimates of line widths by boksenberg , sargent & rauch ( 2003 ) indicate @xmath15710 km / s in the range @xmath491.5 - 4.5 for @xmath158()@xmath159 @xmath147 absorbers .
this can be used to derive upper limits to the abundance weighted igm temperature .
as the thermal component of the line width is given by @xmath160 , abundance weighted igm temperatures of less than @xmath15@xmath3010@xmath26 k are indicated . in fig .
[ fig : civ ] ( right ) we show abundance weighted igm temperatures a function of @xmath151 in the `` field '' galaxy regions for the three imfs considered ( results for other regions are quite similar ) .
it is seen that models in general satisfy the above igm temperature criterion .
moreover , our predictions of the abundance weighted igm temperatures agree well with the predictions of od06 for their best fitting `` momentum - driven '' wind model .
finally we note , that in the future it will be possible to probe the metal enrichment of the igm to even lower densities using see , e.g. , schaye ( 2000 ) .
with the results on the cumulative cosmic gas - phase oxygen distributions presented in the previous section , and the results on the similar distributions for the stellar oxygen from paper i , we can now assess the relative importance of the two components . in fig .
[ fig : ocum ] results for the normalised gas - phase oxygen distributions are shown by long - dashed curves , the stellar distributions are shown by the short - dashed curves , and the combined distributions by the solid curves ( for the three imfs considered ) for the @xmath121 and -1.57 lfs ( we note that stellar oxygen refers to oxygen in galactic stars only , but since the amount of oxygen in intergalactic stars is , in comparison , very small ( paper i ) , we neglect this component in the following ) . as can be seen ,
the combined distributions are for all three imfs dominated by the gas - phase oxygen , and increasingly so going from the k98 to the ay imf . for the @xmath121
lf , it is found that less than about 25% of the total cosmic oxygen is associated with galaxies brighter than @xmath161 , irrespective of the choice of imf . for the @xmath122 lf
the corresponding fraction is about 35% for the ay imf , and about 40% for the k98 and salpeter imfs .
[ fig : gasstar ] shows , for the @xmath121 luminosity function , and the three imfs , the cumulative partition between gas - phase oxygen and stellar oxygen as a function of @xmath36 .
it is seen , that if the faint end slope of -1.85 holds down to @xmath162 , then the cumulative gas - phase oxygen fractions are 72 , 79 and 92% for the k98 , salpeter and ay imfs , respectively , so , as stated above , for all three imfs , the amount gas - phase oxygen dominates over the amount stellar oxygen . for the @xmath122 luminosity function these fractions drop slightly to 64 , 71 and 88% . as discussed in sec .
[ sect : cosmic ] the calculations assume a constant faint end slope down to @xmath36=@xmath163 . if the luminosity function is assumed to display a significant flattening at a brighter magnitude than this , then the figure can still be used to determine the gas - phase / stellar oxygen partition to such a limiting @xmath36 .
assuming , for example , that the faint end slope is constant to at least 6 magnitudes below @xmath164 ( @xmath36@xmath165 ) , which is the case for local luminosity functions ( sec .
[ sect : cosmic ] ) , results in lower limits on the cumulative gas - phase oxygen fractions of 65 , 72 and 90% for the @xmath121 luminosity function , not much different from the results quoted above .
we can now finally compare the integral constraints on the average cosmic oxygen density at @xmath0=3 , obtained in sec .
[ sect : uv ] , to what is obtained by combining detailed high - resolution galaxy formation simulations with the observed ( corrected ) @xmath0@xmath13 , v - band luminosity function . in fig .
[ fig : cumodens ] we show the cumulative oxygen density versus @xmath36 for the three imfs and the two faint end luminosity function slopes considered . also shown
are the constraints from the `` minimum '' , `` median '' and `` maximum '' oxygen production rate density models discussed in sec .
[ sect : uv ] . in order to enable a meaningful comparison ,
the latter results should be compared to the former evaluated at certain limiting magnitudes .
these magnitudes should be chosen to be brighter than or equal to about @xmath166 , @xmath167 and @xmath168 for the k98 , salpeter and ay imfs , respectively , for the following reasons : kennicutt ( 1998 ) finds a relation between uv luminosity and star formation rate ( sfr ) , viz .
@xmath169 with @xmath170=1.0 for the salpeter imf .
given that the yield of the kroupa ( 1998 ) imf is smaller than for the salpeter imf , and vice versa for the arimoto - yoshii imf , @xmath170 will be larger and smaller than unity for these two imfs , respectively .
quantitatively we find that @xmath1711.7 and 0.4 for the k98 and ay imfs , respectively .
the absolute uv magnitude ( @xmath461500 ) characterising the observed galaxy luminosity function at @xmath0@xmath13 ( corresponding to @xmath172 ) is @xmath173@xmath174 ( e.g. , steidel 1999 , after correction to the adopted cosmology , sawicki & thomson 2006 ) .
following sawicki & thomson ( 2006 ) , and using eq.5 above , one can show that this absolute magnitude corresponds to a ( un - extincted ) star formation rate of about 15 @xmath170 @xmath35/yr . with a dust correction of about a factor 5.5 at @xmath0@xmath13 ( sec .
[ sect : uv ] ) the above observed @xmath173 hence corresponds to a true ( un - obscured ) sfr of about 82 @xmath170 @xmath35/yr .
we shall now consider two models for the @xmath461500 extinction as a function of redshift for @xmath583 .
model a assumes a constant extinction factor of 5.5 at all @xmath583 ( cf .
[ sect : uv ] ) , whereas model b is a low - extinction model assuming factors of 4.2 , 3 , 2 and 1.5 at @xmath0=3 , 4 , 5 and 6 , respectively ( cf . bouwens 2006 ) . for model a , a galaxy of _ observed _ @xmath175@xmath172 will be characterised by a sfr@xmath18 @xmath170 @xmath35/yr . from our large sample of galaxy models we find that galaxies with such star formation rates at _ any _ @xmath583 will have an ( un - extincted ) @xmath36 at @xmath0=3 brighter than about @xmath176 , @xmath142 and @xmath167 for the k98 , salpeter and ay imfs , respectively . for model
b , the above @xmath173 corresponds to a true sfr of about 63 @xmath170 @xmath35/yr , and galaxies of observed @xmath175@xmath172 will have sfrs of about 6.3 , 4.5 , 3.0 and 2.3 @xmath35/yr at @xmath0=3 , 4 , 5 and 6 , respectively .
we find that only galaxies of ( un - extincted ) @xmath36 at @xmath0=3 brighter than about @xmath142 , @xmath177 and @xmath167 for the above three imfs , respectively , will satisfy this .
moreover , even assuming ( very conservatively ) zero extinction at @xmath1786 , corresponding to a `` limiting '' ( un - extincted ) sfr of about 1.5 @xmath35/yr , all galaxies of such @xmath1786 sfrs will be be brighter than about @xmath36=@xmath142 , @xmath177 and @xmath167 at @xmath0=3 for the three imfs , respectively . assuming lower luminosity limits of @xmath36=@xmath177 , @xmath167 and @xmath168 for the three imfs
is hence very conservative these limits are indicated in fig .
[ fig : cumodens ] by vertical dashed lines . from fig .
[ fig : cumodens ] it follows that galaxy models based on the k98 imf can not meet the constraint set by the observed uv luminosity density history the oxygen ( and general metal ) yield of this imf is simply not sufficiently large .
the same is the case for the salpeter imf , though the @xmath122 model comes close to matching the lower bound of 0.32@xmath56 @xmath35/(@xmath33mpc)@xmath57 ( see also below ) . on the other hand , models based on the arimoto - yoshii imf ,
match the integral constraint well .
summarising , one of the main results obtained in this paper is that galaxy formation models based on the kroupa ( 1998 ) imf ( and any other imf of similar chemical yield ) are strongly excluded by the cosmic enrichment history constraint .
other arguments , why the average cosmic imf must have a larger chemical yield than what is typical for a solar neighborhood one , have been given by , e.g. , portinari ( 2004 ) , dantona & caloi ( 2004 ) , serjeant & harrison ( 2005 ) , lucatello ( 2005 ) , loewenstein ( 2006 ) , prantzos & charbonnel ( 2006 ) and weidner & kroupa ( 2006 ) , but see also elmegreen ( 2006 ) .
perhaps even more interestingly , models based on the salpeter imf , the arguably most extensively used model of the stellar imf , are also excluded , though the @xmath122 models only marginally so . given that the salpeter imf is routinely used in translating uv luminosities into star formation rates for high redshift galaxies , this is also an important result .
although it would be inappropriate to assign a high statistical significance to this result , given all the uncertainties involved , it is obvious from fig . [
fig : cumodens ] that the models based on the more top - heavy arimoto - yoshii imf provide a much better match to the median value for the integral constraint .
this is in particular the case , since the above lower absolute luminosity limits of @xmath36=@xmath177 , @xmath167 and @xmath168 for the three imfs , respectively , are likely to be very conservative .
in sec . [ sect : cosmic ] it was shown that for the @xmath121 @xmath0@xmath13 luminosity function , less than about 1/4 of the cosmic oxygen is associated with lyman break galaxies sufficiently bright , @xmath36@xmath179 , for direct abundance determination , using oxygen lines emitted from hii regions around young stars in the galaxies ( e.g. , pettini 2001 ) . for the @xmath122 luminosity function
this fraction increases slightly , to about 35% .
furthermore , in sec . [
sect : comb ] it was shown that for both luminosity functions the major part of the cosmic oxygen is in the gas - phase , rather than in stars .
in particular , for the salpeter and arimoto - yoshii imfs , which emerge from the previous section as the more plausible , the gas - phase oxygen fraction exceeds 70% .
fsb05 found a factor of about five discrepancy between the amount of metal in the stars of lyman break galaxies and what is predicted from the integrated uv luminosity density history .
they assumed typical lyman break galaxy stellar masses of 2 @xmath180 @xmath35 , corresponding to @xmath36@xmath181 , close to @xmath182 . in fig .
[ fig : gasstar ] is shown , for the @xmath121 luminosity function , the cumulative stellar cosmic oxygen density . for the k98 imf ,
the ratio between the stellar oxygen density to @xmath183 and what is obtained from the median model ( sec .
[ sect : comb ] ) is 0.10 . for the salpeter and ay imfs , evaluated at @xmath184 and @xmath185 , respectively ,
the corresponding ratios are 0.17 and 0.20 . for the @xmath122 luminosity function the corresponding ratios are 0.14 , 0.24 and 0.29 , for the k98 , salpeter and ay imfs , respectively . given that the ay models over - predict the metallicities of lyman break galaxies of @xmath129 to @xmath186 by about 0.2 dex relative to observations ( paper i ) , the above ratios for the ay models should be reduced to 0.13 and 0.18 for the @xmath121 and -1.57 luminosity functions , respectively .
the above results are obtained , however , by including stellar oxygen mass all the way down to @xmath36@xmath187 .
if one only includes stellar oxygen mass to @xmath36@xmath181 , all the above fractions are reduced by about a factor of two .
hence the discrepancy discussed above , denoted by fsb05 and others as `` the missing metals problem '' , is actually about _
twice _ larger than originally found by fsb05 .
do06 predicted that at @xmath0@xmath13 , about 50% of the gas - phase metals should reside in the diffuse igm ( gas outside the virial radii of galaxy halos , and of @xmath18810@xmath26 k ) .
we can not compare the results found in the previous sections directly to theirs , since different gas - phase criteria have been used .
however , our @xmath79=4.5 gas criterion is similar to theirs for diffuse igm , though we also include @xmath79=4.5 metals inside of galaxy virial radii in our estimate ( see also below ; note also that we find almost all the @xmath79=4 phase metals to reside in gas of fairly high density , @[email protected] @xmath8 , so there is essentially no overlap between this phase and do06 s diffuse phase ) . for the @xmath121 lf
we find @xmath79=4.5 to total gas phase metal fractions of 34 , 31 and 26% for the k98 , salpeter and ay imfs , respectively . for the @xmath122 lf , the corresponding fractions are 21 , 20 and 18% .
if we try to mimick the criterion of do06 better , by selecting all metals in gas of @xmath2@xmath253@xmath3010@xmath26 k and @xmath189@xmath190100 , then for the field galaxy regions we find `` diffuse igm '' oxygen fractions of 4 , 4 and 6% for the k98 , salpeter and ay imfs , respectively .
if we use the proto - cluster regions , which are the only simulations at our disposal with a resolution comparable to the ( fairly modest ) numerical resolution of do06 s 32 @xmath33mpc box size simulation , the corresponding fractions drop to @xmath11% for all imfs .
however , we stress that the temperature of the proto - cluster igm is larger than that of the average igm , but in any case significantly lower `` diffuse igm '' metal fractions , than predicted by do06 , are indicated .
do06 used the springel & hernquist ( 2003 ) `` sub - grid '' approach to model star - formation .
they predict that about 20% of the gas - phase metals reside in star - forming gas . in our models , which invoke explicit two - phase modeling of the ism ,
gas of @xmath2@xmath1@xmath191 k and @[email protected] @xmath8 is potentially star - forming .
hence we compare our results for the @xmath79=4 gas - phase metal content to the above of do06 ( as noted above , almost all @xmath79=4 metals reside in gas of @[email protected] @xmath8 ) . for the @xmath121 lf
we find @xmath79=4 to total gas phase metal fractions of 47 , 37 and 26% for the k98 , salpeter and ay imfs , respectively . for the @xmath122 lf
, the corresponding fractions are 57 , 47 and 32% .
hence we find somewhat larger metal fractions in star - forming gas , than do do06 .
a more important difference is , however , that do06 at @xmath0@xmath13 find no metal containing gas at densities @[email protected] @xmath8 .
this result is strongly at variance with our results , and it would seem difficult for such models to match the large amount of dla cross section observationally detected at such redshifts , although further analysis obviously is required to clarify this . finally , we can compare the @xmath0=3 gas phase metal temperature distribution predicted by do06 to our results .
qualitatively , our predictions for the k98 and salpeter imfs agree with that of do06 , in yielding temperature distributions , which peak at @xmath2@xmath1@xmath191 k and decrease towards larger temperatures .
however , for the ay imf the results are very different . moreover , quantitatively our results for the k98 and salpeter imfs disagree significantly with those of do06 at the `` high-@xmath2 '' end : do06 find that @xmath1185% of the gas - phase metals reside in gas of @xmath2@xmath83@xmath193 k. in comparison , we find for the @xmath121 lf that the corresponding fractions are 20 , 32 and 51% for the k98 , salpeter and ay imfs , respectively , and for the @xmath122 lf , 21 , 32 and 50% .
hence for any of the imfs considered , we predict significantly larger fractions of `` high-@xmath2 '' gas - phase metals , than do do06 . the @xmath79=4 phase is found to be the most prominent metal containing gas - phase for the k98 and salpeter imfs , and also to be quite prominent for the ay imf . as the metals in this phase typically reside in gas of densities @[email protected] @xmath8 , one would expect a significant fraction of these metals to be situated at column densities typical of dlas , viz .
n(hi)@xmath1210@xmath194 @xmath147 .
however , in general dlas have very low abundances - significantly less than galactic stars with similar ages ( pettini 1990 , 1994 ; lu 1996 ; kulkarni & fall 2002 ; prochaska 2003 ; akerman 2005 ; zwaan 2005 ; erni 2006 ) . expressed in terms of the global metal content of the lbg stars
, we predict similar to three times as large amounts of metals in the @xmath79=4 phase ( depending on the imf ) , whereas observations of dlas only indicate a fraction of about 10 - 20% ( e.g. , fsb05 , prochaska 2006 ) .
it is obviously of importance to understand the reason for this apparent discrepancy .
although we will defer a thorough discussion of dlas to a forthcoming paper ( sommer - larsen 2007 ) , we briefly address the above issue in the following .
ellison ( 2007 ) analysed two of the very high resolution simulations , described in the following section , in relation to dla properties .
the two simulations represent the formation and evolution of two disc galaxies , of @xmath0=0 characteristic circular velocities @xmath195=245 and 180 km / s .
the galaxies were selected from a larger sample to represent two different disc formation evolutionary paths : for the @xmath195=180 km / s galaxy the disc starts growing steadily already by @[email protected] , whereas for the other galaxy disc growth is merger induced , and the disc grows strongly from @xmath0@xmath11 to 0 ( see also robertson 2004 ) .
ellison determined the dla / sub - dla characteristics of these two proto - galaxies at @xmath0=3.6 , 3.0 and 2.3 , with focus on determining neutral column density distributions , and the probability of detecting coincident 100 kpc - scale dla / sub - dla absorption in individual galaxies at such redshifts we refer the reader to ellison for more detail , as well as images of the objects . here
we build on the analysis by ellison of these two ( proto-)galaxies . in fig .
[ fig : dla ] ( left ) we show the normalised cumulative oxygen mass in the absorber above log(n(hi))=20.3 versus the cumulative absorber area , starting at the highest hi surface densities , log(n(hi))@xmath124 , and going down to log(n(hi))=20.3 .
for each galaxy results for `` face - on '' and one `` edge - on '' projections are shown .
as can be seen , about 90% of the total oxygen mass of the absorber resides in 1% of the total absorber area for both projections . from fig .
[ fig : dla ] ( center ) it is moreover seen that this oxygen mass is associated with column densities log(n(hi))@xmath722.5 .
dlas of such high column densities have never been detected in qso spectra , which of course is not surprising given the above findings .
moreover , at such high densities , formation of molecular hydrogen is likely to take place ( schaye 2001b ; zwaan & prochaska 2006 ) .
although the process of h@xmath82 formation is not included in the hydro / gravity simulations , it is clear that the properties of the high density gas can be significantly affected by h@xmath82 formation ( e.g. , pelupessy 2006 ) .
greve & sommer - larsen ( 2007 ) showed that the effects of h@xmath82 formation can be approximately determined post - process on the basis of the cosmological galaxy formation simulations this will be one of the main topics of a forthcoming paper on dlas ( sommer - larsen 2007 ) .
molecules have been detected in dlas , but only at relatively low fractions of less than a tenth relative to atomic hydrogen ( ledoux 2003 ) .
moreover , h@xmath82 is preferentially detected in the highest metallicity dlas ( petitjean 2006 ) .
at this point it is sufficient to note that a ) only a very small fraction of qso sight - lines will probe the high - metallicity regions of dlas , and b ) due to h@xmath82 formation these regions may possibly be difficult to probe in neutral hydrogen ( zwaan & prochaska 2006 ) .
we also note from fig .
[ fig : dla ] ( right ) that the metal abundances of the high column density regions are significantly larger than what is typically observed in dlas , whereas at log(n(hi))@xmath11822 they are not ( e.g. , pettini 2003 ) .
finally , we note that the high column density regions will likely be characterised by significant dust contents , which could further bias against selecting qso sight - lines passing through such regions .
evidence for dust in dlas has been reported ( pei , fall & bechtold 1991 ; vladilo 2006 ) .
this may naturally explain why dla column densities of log(n(hi))@xmath722 are very rare ( vladilo & proux 2005 , vladilo 2006 ) .
however , studies of radio selected dlas ( free from dust - bias ) have found similar column density and metallicity distributions as for optically selected dla samples ( ellison 2001 , 2004 ; ellison , hall & lira 2005 ; akerman 2005 ; jorgenson 2006 ) .
also , murphy & liske ( 2004 ) find no excess absorption towards qsos with dlas ( but this study is based on an optically selected dla sample ) .
it remains to be clarified to which extent dust bias is important for dla column density and metallicity studies , especially at log(n(hi))@xmath722 . in particular , a larger sample of radio selected dlas is needed ( jorgenson 2006 , ellison 2007 , private communication ) .
we note that dlas at @xmath196 with near solar metallicity have been found ( ledoux 2002 ) .
also , the sloan digital sky survey has been used to strongly increase the sample of dlas ( prochaska 2005 ) and among these 5% have very strong metal - line absorption ( herbert - fort 2006 ) . some of these 5% are metal rich dlas , although some could be relatively metal poor dlas with very large hydrogen column densities .
about 60% of the metal - strong dlas have detected molecular hydrogen ( milutinovic in preparation ) .
interestingly , dlas with log(n(hi))@xmath722 have been detected in the spectra of the optical afterglows of gamma - ray bursts ( watson 2006 ; jakobsson 2006 ; prochaska 2007 ) .
the selection function for grb sight - lines is completely different than the cross - section selection of qso dlas .
grb sight - lines are strongly correlated with the blue light of their host galaxies ( bloom , kulkarni & djorgovski 2002 ; fruchter 2006 ) most likely tracing the location of the most massive stars ( m@xmath720m@xmath197 , larsson 2007 ) .
significant dust obscuration is observed for at least some of the log(n(hi))@xmath722 grb dlas ( watson 2006 ) .
furthermore , it has been found that the metallicities inferred for dlas in grb hosts are systematically higher than for qso dlas and that grbs hence may offer a new probe of early cosmic chemical evolution complementary to the qso dlas and lbgs ( fynbo 2006 ; savaglio 2006 ; prochaska 2007 ) . in conclusion , the low metallicities inferred from qso dlas do not exclude that significant amounts of metals can reside in the @xmath198 phase .
the reason for this is most likely that the total cross - section for this component is only of order a few percent of the total dla cross - section ( see also johansson & efstathiou 2006 ) and that the current dla samlpes are too small to uncover them ( see also zwaan & prochaska 2006 ) , but dust bias , as well as h@xmath82 formation , will probably also be of importance .
in sgp03 and sommer - larsen ( 2006 ) it has been shown that the results of cosmological galaxy formation and evolution using our code are , in general , robust to changes of the numerical resolution . however , in this paper we present gas - phase metal temperature distributions and other results , which have not been presented before .
it is hence clearly important to demonstrate that these results are resolution robust as well . to this end
we carried out three very high resolution k98 imf simulations , of eight times higher mass and two times higher force resolution compared to `` standard '' .
we simulated three ( proto-)disc galaxies , which at @xmath0=0 have characteristic circular velocities of 245 , 180 and 66 km / s .
the galaxy forming regions were represented by a total of about 2.2 , 1.2 and 1.3 million particles , compared to the about 290000 , 150000 and 210000 particles used at `` normal '' resolution ( for computational reasons the region simulated at very high resolution for the last galaxy was slightly smaller than the corresponding `` normal '' resolution one ) .
gas and star particle masses were @xmath199=@xmath200=9.1@xmath3010@xmath26 and 2.5@xmath3010@xmath57 @xmath33m@xmath197 for the first two galaxies and the last one , respectively . the corresponding dark matter particle masses were @xmath201=5.2@xmath3010@xmath40 and 1.4@xmath3010@xmath26 @xmath33m@xmath197 .
moreover , gravitational ( spline ) softening lengths of @xmath202=@xmath203=190 and @xmath204=340 @xmath33pc , respectively , were adopted for the first two galaxies , and 57 and 100 @xmath33pc for the latter .
other results for the first two runs have already been presented in sommer - larsen ( 2006 ) , razoumov & sommer - larsen ( 2006 ) , greve & sommer - larsen ( 2007 ) , laursen & sommer - larsen ( 2007 ) and ellison ( 2007 ) . in fig .
[ fig : numres ] is shown the gas - phase oxygen temperature distributions for the three very high resolution runs , together with the results of the normal resolution ones note that the oxygen masses for the smaller galaxy have been multiplied by a factor 50 for clarity .
it is seen that the general agreement is very good ; the slight disagreement for the small galaxy is due to the fact that the region simulated at very high resolution was slightly smaller than the corresponding normal resolution one . due to computational limitations
we did not perform similar resolution tests for the two other imfs considered , but we have no reason to believe that simulations based on these imfs would perform less well in resolution tests .
we conclude that the results obtained in this paper are robust to resolution changes .
the global temperature distribution of the cosmic gas - phase oxygen at @xmath0@xmath13 has been determined by combining high resolution cosmological simulations of individual proto - galactic , as well as larger , regions with extinction - corrected , observationally based , v - band ( rest - frame ) galaxy luminosity functions ( lfs ) of faint end slopes @xmath121 and @xmath127 .
the simulations have been performed with three different stellar initial mass functions ( imfs ) , a kroupa ( k98 ) , a salpeter ( s ) and an arimoto - yoshii ( ay ) , spanning a range of a factor of five in chemical yield and specific snii energy feedback .
gas - phase oxygen is binned according to @xmath2 as log@xmath3 ( `` cold '' ) , log@xmath4 ( `` warm '' ) , and [email protected] , 5.5 , 6.0 , 6.5 , 7.0 ( `` hot '' phases ) .
below we summarize results for the @xmath121 lf , but results for the @xmath122 lf are similar .
oxygen is found to be distributed over all @xmath2 phases , in particular for the ( `` top - heavy '' ) ay imf .
but , at variance with previous works , it is found that , for the k98 and salpeter imfs , the most important phase is the cold one , which contains 47 and 37% of all gas - phase oxygen , mainly in gas at fairly high density , @[email protected] @xmath8 , and potentially star - forming .
moreover , the cold phase alone contains 1.3 , 1.5 and 3.2 times the mass of oxygen in galactic stars for the three imfs .
the implications of this in relation to observational damped lyman-@xmath9 absorber ( dla ) studies are discussed on the basis of very - high resolution simulations of two ( proto-)disc galaxies , with emphasis on oxygen and iron abundances .
it is concluded , that the reason why current dla surveys only detect a cold ism metal fraction of about 20% relative to the metal mass in galactic stars is that the total cross - section for the high - metallicity component is only of order a few percent of the total dla cross - section , and that the current dla samlpes are too small to uncover it .
moreover , in addition , dust bias , as well as h@xmath82 formation , will likely also be of importance . in relation to `` missing metals ''
it is found that the ratio of gas - phase to stellar oxygen mass is 2.7 , 3.9 and 13 , and the ratio of warm+hot to cold gas - phase oxygen mass is 1.1 , 1.7 and 3.2 for the three imfs . for the ay imf ,
the hot phases actually contain more oxygen than the cold+warm . in conclusion
, a significant fraction of the cosmic oxygen may be difficult or impossible to detect .
in addition , it is found that less than about 20 - 30% of the cosmic oxygen will be associated with galaxies brighter than @xmath129 , the faintest galaxy luminosities probed by current lbg metallicity determinations ( about one mag . below @xmath128 ) .
hence , 70 - 80% of the cosmic oxygen is also in this sense `` missing '' . from the lbg based , @xmath11@xmath11500
uv luminosity density history at @xmath0@xmath123 , we obtain an essentially imf independent constraint on the mean cosmic oxygen density at @xmath0=3 .
we compare this to what is obtained from our models , for the three different imfs .
we find that the ( solar neighbourhood type ) k98 imf is strongly excluded , as the chemical yield is simply too small , the salpeter is marginally excluded , and the ay matches the constraint well .
the optimal imf would have a yield intermediate between the s and ay .
the k98 imf can only match the data if the @xmath11@xmath11500 extinction corrections have been overestimated by factor of @xmath14 , which seems highly unlikely , cf .
reddy & steidel ( 2004 ) . using carbon abundances , and c to ionisation corrections
, we estimate @xmath144 ( ) at moderate igm gas over - densities , @xmath205 for the three imfs , and compare to observational results . as
above , we find that the yield of the k98 imf is too small to match the data , whereas models based on the salpeter and ay imfs can match the data , with the optimal imf in between the two .
moreover , we show that for all imfs , abundance weighted igm temperatures are moderate , @xmath2@xmath1184@xmath3010@xmath26 k , consistent with observational constraints on line widths .
we are very grateful to ben oppenheimer and romeel dav for supplying us with the c/ ionisation correction look - up tables used in this work .
we have benefited from discussions with romeel dav , sara ellison , peter johansson , cedric lacey , peter laursen , ben oppenheimer , laura portinari and alex razoumov .
the comments by the anonymous referee considerably helped in improving the presentation of the paper .
the treesph simulations were performed on the sgi itanium ii facility provided by dcsc .
the dark cosmology centre is funded by the dnrf .
this research was supported by the dfg cluster of excellence `` origin and structure of the universe '' .
haardt , f. & madau , p. 2001 , in clusters of galaxies and the high redshift universe observed in x - rays , recent results of xmm - newton and chandra " , xxxvith rencontres de mor iond , eds .
neumann & j.t.t .
van lacey , c. g. , baugh , c. m. , frenk , c. s. , granato , g. l. , silva , l. , bressan , a. , cole , s. , 2005 , in the proceedings of the conference `` the fabulous destiny of galaxies : bridging past and present '' , 20 - 24 june 2005 , ( astro - ph/0509715 ) ) | the global temperature distribution of the cosmic gas - phase oxygen at @xmath0@xmath13 is determined by combining high resolution cosmological simulations of individual proto - galactic , as well as larger , regions with the observed , extinction - corrected , rest - frame v - band galaxy luminosity function .
the simulations have been performed with three different stellar initial mass functions ( imfs ) , a kroupa ( k98 ) , a salpeter ( s ) and an arimoto- yoshii ( ay ) , spanning a range of a factor of five in chemical yield and specific snii energy feedback .
gas - phase oxygen is binned according to @xmath2 as log@xmath3 ( `` cold '' ) , log@xmath4 ( `` warm '' ) , and [email protected] , 5.5 , 6.0 , 6.5 , 7.0 ( `` hot '' phases ) .
oxygen is found to be distributed over all @xmath2 phases , in particular for the top - heavy ay imf .
but , at variance with previous works , it is found that for the k98 and s imfs the cold phase is the most important . for these imfs
it contains 47 and 37% , respectively , of all gas - phase oxygen , mainly at fairly high density , @[email protected] @xmath8 .
the implications of this in relation to observational damped lyman-@xmath9 absorber ( dla ) studies are discussed . in relation to `` missing metals ''
it is found that a significant fraction of the oxygen is located in a warm / hot phase that may be very difficult to detect .
moreover , it is found that less than about 20 - 25% of the cosmic oxygen is associated with galaxies brighter than @xmath10 , the faintest galaxy luminosities probed by current metallicity determinations for lyman break galaxies ( lbgs ) .
hence , 75 - 80% of the oxygen is also in this sense `` missing '' . from the lbg based , @xmath11@xmath11500
uv luminosity density history at @xmath0@xmath123 , we obtain an essentially imf independent constraint on the mean oxygen density at @xmath0=3 .
we compare this to what is obtained from our models , for the three different imfs .
we find that the k98 imf is strongly excluded , as the chemical yield is simply too small , the salpeter is marginally excluded , and the ay matches the constraint well .
the k98 imf can only match the data if the @xmath11@xmath11500 extinction corrections have been overestimated by factor of @xmath14 , which seems highly unlikely .
the yields for k98 are also far too small to match the observational data for .
the optimal imf should have a yield intermediate between the s and ay .
-1 cm galaxies : galaxies : formation and evolution ; cosmology : simulations |
the concern of this paper is the effect of a rough boundary on a viscous fluid . in most situations of physical relevance
, such effect can not be described in detail : either the precise shape of the roughness is unknown , or its spatial variations are too small for computational grids .
therefore , one may only hope to account for the averaged effect of the irregularities .
this is the purpose of _ wall laws _ : the irregular boundary is replaced by an artificial smoothed one , and an artificial boundary condition ( a wall law ) is prescribed there , that should reflect the mean impact of the roughness .
this paper is a mathematical study of wall laws , in the following simple setting : we consider a two - dimensional rough channel @xmath5 where @xmath6 is the _ smooth part _ , @xmath7 is the rough part , and @xmath8 their interface .
we assume that the rough part has typical size @xmath9 , that is @xmath10 for a @xmath11lipschitz function @xmath12 , @xmath13 .
more will be assumed on the boundary function @xmath14 hereafter ( see figure for an example of such a rough domain ) .
we assume that in this channel domain , the viscous fluid obeys to the stationary incompressible navier - stokes equations : @xmath15 where @xmath16 denotes any vertical cross - section of @xmath17 and @xmath18 .
the third equation in expresses that a flux @xmath19 is imposed across the channel .
note that this flux does not depend on the cross - section , due to the incompressibility and no - slip condition at the boundary .
we also stress that , up to minor changes , we could apply our analysis to many variants of this problem , notably to elliptic type systems or to unstationary navier - stokes . in this simple
setting , the search for wall laws resumes to the following problem : to find a boundary operator @xmath20 , regular in @xmath9 , acting at the _ artificial _ boundary @xmath21 , such that the solution of @xmath22 approximates well the solution @xmath23 of in @xmath24 .
this type of homogenization problems has been considered in many mathematical works .
on wall laws for scalar elliptic equations , we refer to @xcite . on wall laws for fluid flows , see @xcite .
see also @xcite on porous boundaries .
these works go along with more formal computations , grounded by empirical arguments ( _ cf _ for instance @xcite ) .
we finally mention @xcite for the study of roughness - induced effects on geophysical systems .
all these studies have been carried under two assumptions : * compact domains , for instance bounded channels or periodic in the variable @xmath25 . * _ periodic irregularities _ , meaning that the boundary function @xmath14 is periodic .
the first restriction is just a small mathematical convenience , that gives direct compactness properties through rellich type theorems .
the second assumption is of course a big simplification , both from the point of view of mathematics and physics .
these assumptions were considerably relaxed in the recent article @xcite by a. basson and the author . as the present note extends this article , we now describe shortly its main results and underlying difficulties . in all papers on wall laws ,
the starting point is a formal expansion of @xmath23 : @xmath26 formally , the leading term @xmath27 satisfies with the simple no - slip condition @xmath28 the solution of this approximate system is the famous poiseuille flow : @xmath29 note that @xmath27 is defined in all @xmath30 .
this zeroth order asymptotics can be mathematically justified , at least for small fluxes @xmath19 : we prove in article @xcite [ direstimates ] there exists @xmath31 , such that for all @xmath32 , system has a unique solution @xmath23 in @xmath33 .
moreover , @xmath34 we stress that these estimates hold without further assumption on the boundary : we only assume that @xmath14 has values in @xmath35 and is @xmath11lip .
a look at the proof shows that the constants @xmath36 and @xmath37 are only increasing functions of @xmath38 . theorem [ direstimates ] expresses that the wall law provides a @xmath39 approximation of @xmath23 in @xmath40 .
see also @xcite for a similar result in a bounded channel .
however , this wall law does not account for the behaviour of @xmath23 near the boundary , and can therefore be refined . indeed ,
as the poiseuille flow @xmath27 does not vanish at the lower part of @xmath41 , a boundary layer corrector @xmath42 must be added to the expansion .
the ( normalized ) boundary layer @xmath43 is defined on the rescaled infinite domain @xmath44 and formally satisfies the following stokes problem @xmath45 note the inhomogeneous dirichlet condition , that cancels the trace of @xmath27 . .
first , the well - posedness is not clear .
as the boundary function @xmath14 is not decreasing at infinity , one can expect only local integrability of the solution @xmath46 in variable @xmath47 .
the derivation of local bounds is not obvious : the stokes operator being vectorial , one can not use scalar tools such as the maximum principle or harnack inequality .
moreover , as @xmath48 is unbounded in all directions , the poincar inequality ( which allows to get @xmath49 estimates in the channel ) is not available .
besides the well - posedness issue , the qualitative properties of @xmath46 seem also out of reach without further hypothesis . under an assumption of periodic irregularities
, the analysis of becomes straightforward . if @xmath14 is say @xmath50 periodic in @xmath47 , it is easy to show well - posedness in the space @xmath51 moreover , a simple fourier transform in @xmath47 shows that @xmath52 that is exponential convergence to a constant field @xmath53 at infinity .
the constant @xmath54 at infinity is then responsible for a @xmath39 tangential slip .
namely , chosing as a wall law the navier - slip condition @xmath55 it can be shown ( in this periodic framework ) that the solution of provides a @xmath4 approximation of @xmath23 in @xmath3 .
we refer to @xcite for all necessary details .
the error estimate @xmath56 comes from the fact that the boundary layer term satisfies @xmath57 .
the periodicity hypothesis is a stringent one , and one may wonder if the use of navier slip condition can be justified in more general configurations .
this issue has been adressed rigorously in the recent article @xcite .
inspired by the probabilistic modeling of heterogeneous media ( see for instance @xcite ) , we considered irregularities that are not distributed periodically , but randomly , following a stationary stochastic process .
namely , the rough boundary is seen as a realization of a stationary spatial process . following the well - known construction of kolmogorov ,
this amounts to consider the space @xmath58 of all possible rough boundaries , together with the cylindrical @xmath59 field @xmath60 ( that is generated by the coordinates @xmath61 ) and with a stationary measure @xmath62 .
stationary means that @xmath62 is invariant by the group of translation @xmath63 as a consequence of this modeling , the domains @xmath17 , @xmath48 , as well as the velocity fields @xmath23 or @xmath46 depend on the parameter @xmath14 . as discussed earlier , the existence result and estimates of theorem [ direstimates ] are uniform on @xmath64 . moreover ,
it was shown in article @xcite that the function @xmath65 ( extended by @xmath66 outside @xmath67 ) is measurable as a function from @xmath64 to @xmath68 . using this probabilistic structure ,
we have been able to extend partially the results of the periodic case .
key elements of our analysis are : * the well - posedness of the boundary layer system , obtained in functional spaces encoding the relation @xmath69 * the convergence of @xmath70 to @xmath71 as @xmath72 , both in @xmath73 and almost surely , locally uniformly in @xmath47 .
such convergence is deduced from the ergodic theorem .
more on the boundary layer system will be provided in the next sections . as regards the navier wall law , the main result of @xcite resumes to [ navestimates ] there exists @xmath74 such that the solution @xmath75 of , satisfies @xmath76 we remind that @xmath77 theorem [ navestimates ] shows that a slip condition of navier type improves the approximation of @xmath23 . as in the periodic case ,
the random variable @xmath54 in comes from the convergence of the boundary layer @xmath46 .
if the measure @xmath62 is ergodic , @xmath54 does not depend on @xmath14 , as pointed out in @xcite . a natural concern about this result
is the @xmath78 bound , which is only a slight improvement of the @xmath39 in theorem [ direstimates ] .
a look at article @xcite shows that this poor bound is due to the lack of information on the way @xmath46 converges at infinity .
contrary to the periodic case , where convergence at exponential rate is established , the simple use of the ergodic theorem does not yield any speed rate .
the present paper aims at clarifying this point .
losely , _ we will show that for a large class of boundaries , the navier wall law provides a latexmath:[$o({{\varepsilon}}^{3/2 } namely , we will make the two following assumptions on our random roughness : + _ ( h1 ) the measure @xmath62 is supported by @xmath80 for some @xmath81 and some @xmath82 . _
+ _ ( h2 ) the randon boundary has no correlation at large distances , that is the @xmath83-fields @xmath84 are independent for @xmath85 , for some @xmath86 . _
+ under these assumptions , the main theorem of the paper reads : [ refinedestimate ] for small enough @xmath19 and under ( h1)-(h2 ) , the following refined estimate holds : @xmath87 before entering the proof of this theorem , let us give a few hints .
theorem [ refinedestimate ] is deduced from a central limit theorem for the quantity @xmath88 .
broadly , this theorem comes from good properties of the random variables @xmath89 due to the elliptic nature of the stokes operator , such random variables are not independent .
however , under assumption ( h2 ) , we are able to prove that the correlation terms @xmath90 decay fast enough as @xmath91 . as a result , one can prove a central limit theorem on @xmath92 , and then a similar one on @xmath93 .
we point out that such type of results for dependent variables with strong decay of correlations is quite classical and has been used in various fields .
we refer to @xcite for a review paper related to dynamical systems , and to recent articles @xcite for applications in a pde context . as a consequence of this central limit theorem
, we show that the boundary layer converges to a constant as @xmath94 .
note that this is in sharp contrast with the periodic case , where exponential convergence holds ( we stress that periodic boundaries are highly correlated , thus far from satisfying ( h2)).this speed of convergence is resposible for the @xmath95 in the navier wall law .
the main difficulty is to obtain the decay of correlations of variables like @xmath92 .
the proof relies on precise estimates of the green function for the stokes operator above a non flat boundary .
such estimates follow from sharp elliptic regularity results , where one must pay attention to the oscillation of the boundary .
this is achieved under the regularity assumption ( h1 ) , using ideas of avellaneda and lin for homogenization of elliptic systems @xcite .
in this section , we explain how theorem [ navestimates ] follows from estimates on the solution @xmath46 of .
such estimates will be established in the following sections . at first , we remind the main features of @xmath46 , as stated in article @xcite . as emphasized in the introduction , to solve in a deterministic way , that is for each possible boundary @xmath14 , is still unclear .
hence , one must take advantage of the probabilistic setting . first ,
notice that a reasonable solution @xmath46 should satisfy : @xmath96 together with the stationarity assumption , this relation sort of substitutes to the identity @xmath97 used in the treatment of @xmath98periodic roughness .
it allows to extend the well - posedness result , through an appropriate variational formulation .
this formulation has been described in article @xcite .
first , one introduces the new unknown @xmath99 and replace system by @xmath100\vert_{y_2 = 0 } = 0 , \quad [ { { \partial}}_2 w - ( 0,q)]\vert_{y_2 = 0 } = ( -1,0 ) , \end{aligned } \right.\ ] ] where @xmath101\vert_{y_2 = 0}$ ] denotes the jump at @xmath102 .
then , one multiplies formally the stokes equation by a test function @xmath103 that satisfies @xmath104 , @xmath105 . integrating by parts over @xmath106 yields @xmath107 finally , if @xmath108 , @xmath109 satisfy relation , one can integrate with respect to @xmath14 , and thanks to the stationarity of @xmath62 , get rid of the annoying boundary term at the r.h.s : @xmath110 afterwards , this formal variational formulation can be rigorously defined and solved : in short , one can apply the riesz theorem in a functional space of sobolev type , made of functions @xmath108 such that @xmath111 and satisfying almost surely , together with @xmath112 .
we refer to @xcite for all details .
note that stationarity implies : @xmath113 back to the original system , this variational solution @xmath108 provides almost surely a solution @xmath114 in the sense of distributions .
moreover , the ergodic theorem yields ( see @xcite ) @xmath115 in order to understand the origin of the navier approximation , the next step is to describe the behavior of @xmath46 as @xmath116 . for periodic roughness , one can show exponential convergence of @xmath46 to a constant vector field @xmath117 .
however , the rate of convergence goes to zero with the period @xmath50 .
when dealing with stationary random boundaries , that broadly speaking contain all periods , the exponential decay does not hold _ a priori_. in other words , there is a problem associated to the fourier spectrum , that is discrete in the periodic case , and may accumulate to zero in the random case .
again , this problem has been ( partially ) overcome in @xcite .
the first step is to obtain a representation of @xmath46 in terms of a stokes double layer potential , _ cf _ proposition ? ? . almost surely , for any @xmath118 where @xmath119 is the poisson type kernel for the stokes operator over a half space .
then , the ergodic theorem and a few calculations yield : @xmath120 where the convergence holds almost surely ( locally uniformly in @xmath47 ) , as well as in @xmath73 ( uniformly in @xmath47 ) . in the case where the stationary measure @xmath62 is ergodic ,
the constant @xmath54 does not depend on @xmath14 . finally , back to the integral representation , and with similar treatment for derivatives of @xmath46 : @xmath121 { } \ : 0.\ ] ] we refer to ( * ? ? ?
* proposition 13 ) for all details .
most of the analysis of the present paper will be devoted to a refined asymptotic estimate of the boundary layer : [ theosharp ] under assumptions ( h1 ) , ( h2 ) , for all @xmath122 , @xmath123 { } \sigma_\beta \ge 0.\ ] ] it is of course a much sharper convergence result than . before tackling its proof ,
we explain how it implies theorem [ navestimates ] .
arguments are direct adaptation from section 5 in @xcite . on the basis of the boundary layer analysis
, one can build an approximation of @xmath23 of boundary layer type .
namely , we introduce @xmath124 in this approximation , @xmath27 is the poiseuille flow and @xmath125 is the boundary layer solution of . as @xmath46 does not converge to zero at infinity , we add a large scale corrector @xmath126 satisfying : @xmath127 it is just a combination of a couette and a poiseuille flow : @xmath128 .
still , this approximation does not vanish at the boundary , which explains the addition of another term @xmath129 .
it must satisfy @xmath130 this remainder can be taken small in the sense of this problem possesses a ( non unique ) solution @xmath131 such that @xmath132 _ proof : _ the proof of this result mimics the one of proposition 14 in @xcite .
the corrector @xmath131 can be chosen in the form @xmath133 the streamfunction @xmath134 is determined up to a constant and polynomial in @xmath135 .
its coefficients have explicit dependence on @xmath136 .
hence , the @xmath137 estimate on @xmath131 follows from the control of various terms involving @xmath136 . for instance
, one must bound the @xmath138 norm of @xmath139 where the last equality comes from the stokes formula .
using stationarity of @xmath62 , we get @xmath140 thanks to the refined estimate , we finally obtain @xmath141 all other terms involve similar computations .
these are straightforwardly adapted from the proof of proposition 14 in @xcite , using instead of .
once the approximate solution @xmath142 is built , one can obtain by energy estimates the following bounds , for @xmath19 small enough : @xmath143 which of course imply theorem [ navestimates ] . as the proof is very similar to what was done in paper @xcite
, we do not expand more and refer to it for all details .
up to the end of the paper , we will assume ( h1)-(h2 ) , and focus on theorem [ theosharp ] .
it is classical that ( h2 ) implies ergodicity of @xmath62 , so that the constant @xmath54 does not depend on @xmath14 .
we start again from an integral representation @xmath144 where the matrix kernel @xmath119 is given by @xmath145 we introduce @xmath146 a simple change of variable leads to @xmath147 our first goal is to show that the l.h.s
. converges in law to a gaussian distribution , for all @xmath148 .
we will focus on the case @xmath149 , the other cases being handled in the exact same way .
we state [ tcl ] the function @xmath150 satisfies the following properties : i ) : : @xmath151 ii ) : : the random process @xmath152 converges weakly to a gaussian process @xmath153 as @xmath154 goes to infinity .
iii ) : : the covariance matrices also converge , that is for all indices @xmath155 and for all @xmath156 , @xmath157 { } { { \mathbb e}}b_i(\cdot , s ) b_j(\cdot , t)\ ] ] we remind that the process @xmath158 with values in @xmath30 _ converges weakly _ to @xmath159 if , for all @xmath160 and all continuous bounded function @xmath161 , { { \mathbb r}}^2 \right ) \mapsto { { \mathbb r}}$ ] , @xmath162 { } \ : { { \mathbb e}}{\cal f}(x).\ ] ] theorem [ theosharp ] is then a direct consequence of the random process @xmath163 converges in law to a gaussian vector with zero average .
moreover , for all @xmath155 , @xmath164 { } \sigma_{ij},\ ] ] where @xmath83 is the covariance matrix of this gaussian vector .
_ proof of the corollary : _ to prove convergence in law to a gaussian vector @xmath165 of covariance matrix @xmath83 , we need to show that for any @xmath166 , @xmath167 { } { { \mathbb e}}\ , f\left({\cal n}_{\sigma}\right).\ ] ] unsurprisingly , we take @xmath168 we decompose , for any @xmath160 , @xmath169 we now show that expressions @xmath170 , @xmath171 ,
converge to zero ( @xmath172 is similar to @xmath171 and simpler ) .
let @xmath173 .
we have @xmath174 where the last line comes from point ii ) of proposition [ tcl ] .
thus , for @xmath175 large enough , independently of @xmath154 , @xmath176 . such @xmath175 being fixed , for @xmath154 large enough , we get @xmath177 by point i ) of proposition [ tcl ] . this yields convergence in law . the convergence of the covariance matrix @xmath178 follows from the dominated convergence theorem , using i ) and iii ) of proposition [ tcl ] .
we get @xmath179 this concludes the proof of the corollary . it remains to prove theorem [ tcl ] .
note that point ii ) is essentially a central limit theorem for the sequence of random variables @xmath180 the problem is that these random variables are not independent , due to `` propagation of information at infinite speed '' in the stokes system . to establish a central limit theorem for such type of sequences is a classical question .
the basic idea is that one can extend the central limit theorem to non independent sequences that feature a good decay of correlations as @xmath181 goes to infinity .
we now illustrate this general principle on our problem , using assumption ( h2 ) .
we follow the presentation of article @xcite , in which a similar question arises for a semilinear heat equation with random source .
let @xmath182 the @xmath83-algebra generated by the applications @xmath183 , @xmath184.we state the following lemma : [ lemmeun ] suppose that @xmath185 satisfies @xmath186 for some @xmath187
. then , proposition [ tcl ] holds .
_ proof of the lemma : _ we write the decomposition @xmath188 with the sum converging in @xmath73 .
the corresponding sum for @xmath150 is @xmath189 where @xmath190 .
then , we have : @xmath191 . by the assumption of independence at large distances , the correlations @xmath192 and @xmath193 vanish for @xmath194 .
we introduce @xmath195.\ ] ] if @xmath196 , we just write @xmath197 if @xmath198 , we decompose @xmath199 using the bound on the conditional expectations , we end up with @xmath200 for some constant @xmath201 .
we thus get i ) . to prove ii ) , we just write the decomposition @xmath202 it is well - known that each finite sum satisfies a central limit theorem , that is @xmath203 { } b^k(\omega , t)\ ] ] in the sense of weak convergence , to some gaussian process @xmath204
. moreover , the covariance matrix also converges , that is @xmath205 { } { { \mathbb e}}b^k_l(\omega , s ) \
, b^k_m(\omega , t).\ ] ] in short , it is due to the fact that the random variables @xmath206 have zero correlations at large distances : see ( * ? ? ?
* theorem ( 7.11 ) p424 ) for a similar result and detailed proof .
moreover , thanks to estimate , the remainder @xmath207 converges to zero as @xmath208 , locally uniformly in @xmath209 , uniformly in @xmath154 .
hence , points ( ii ) and ( iii ) of proposition [ tcl ] hold , which ends the proof of the lemma .
we still have to estimate the variance of @xmath210 .
following @xcite , we can turn this question into a question of domain of dependence for solutions of .
precisely , starting from the measure @xmath62 on @xmath64 , we define a new measure @xmath211 on the product space @xmath212 endowed with its cylindrical @xmath59field .
namely , @xmath211 is defined in the following way : 1 .
@xmath213 , which determines @xmath211 over the sub @xmath59field @xmath214 generated by the applications @xmath215 .
2 . for all @xmath216 , for all @xmath217 with @xmath218 , for all borelian subsets @xmath219 , @xmath220 of @xmath221 , and @xmath222 @xmath223 , which determines @xmath211 conditionaly to @xmath214 .
it is then easy to derive the following identity , see @xcite : @xmath224 thus , if @xmath225 and @xmath226 are two boundary layer domains with boundaries that coincide over @xmath227 $ ] , we need to estimate the difference of the corresponding boundary layer solutions @xmath228 and @xmath229 .
this is the purpose of the next section .
_ throughout the rest of the paper , we will assume ( h1 ) . _
the main difficuly is to prove the following [ propdecay ] under assumption ( h1 ) , for all @xmath230 , for almost every @xmath231 , @xmath232 where @xmath36 does not depend on @xmath233 .
together with the results of the preceding section , this proposition concludes the proof of theorem [ theosharp ] ( take @xmath234 ) , and therefore the proof of the main theorem [ navestimates ] .
in fact , the sharper bound @xmath235 that is with @xmath236 would still be true .
we will discuss this briefly in the last section of the paper .
for the sake of brevity , we only prove here the weaker form .
the main difficulty is that the boundary layer solutions @xmath237 and @xmath238 of are not defined on the same domain , so that estimates on the difference are not directly available .
if the poisson equation rather than the stokes system was considered , representation of the solution in terms of brownian motion would allow to conclude quite easily . again , this will be explained in the last section of the paper . in the case of system , we are not aware of such representation , and the bound will come from an accurate description of the ( matrix ) green function of the stokes operator above a humped boundary .
we consider for all @xmath239 , and for all @xmath240 , the system : @xmath241 where @xmath242 is the dirac mass at @xmath243 , and @xmath244 is the @xmath245 identity matrix .
let us remind how to build the matrix green function @xmath246 .
up to a vertical translation of the domain , we can first assume that @xmath247 .
we then introduce the green function @xmath248 for the stokes operator in the upper - half plane , see @xcite . extending @xmath249 by @xmath66 for @xmath250 ,
the functions @xmath251 satisfy formally @xmath252 = 0 , & \quad \bigl [ { { \partial}}_2 h(z,\cdot ) - q(z , \cdot ) \otimes \,e_2 \bigr ] = - \left [ { { \partial}}_2 g_0(z,\cdot ) - p_0(z , \cdot)\ , \otimes e_2 \right ] , \end{aligned } \right.\ ] ] where @xmath253 $ ] is the jump along @xmath254 .
the jump on the derivative is explicit , as @xmath255 standard variational formulation yields existence and uniqueness of a solution @xmath256 with @xmath257 in @xmath3 . in turn
, this provides a unique solution @xmath258 to . the corresponding pressure field @xmath259 is determined up to the addition of a constant matrix .
note that uniqueness yields the relation @xmath260 our key estimate is provided by [ green2 ] for all @xmath230 , for all @xmath261 satisfying @xmath262 , we have @xmath263 where @xmath264 denotes the distance to the boundary @xmath265 , and @xmath36 is a constant depending only on @xmath266 and on @xmath267 .
note that by symmetry of @xmath119 , we also have @xmath268 moreover , in the course of the proof of lemma [ green2 ] , we will show that for all @xmath269 , @xmath270 .
we first need to connect the solution @xmath125 of to the green function @xmath271 . for this purpose
, we rather consider @xmath272 which satisfies .
note that @xmath46 and @xmath108 coincide at @xmath273 .
formally , @xmath108 should be equal to @xmath274 using estimates , , it is standard to show that @xmath275 is a solution of in @xmath276 . using bound , one has even @xmath277 for all @xmath278 . extending @xmath279 by @xmath66 outside @xmath280
, one can show that @xmath281 is measurable from @xmath282 to @xmath68 ( see the appendix for details ) .
moreover , thanks to , @xmath275 satisfies the stationarity relation @xmath283 . finally , using that @xmath108 and @xmath275 both satisfy , a simple energy estimate on the difference leads to @xmath284 which shows that @xmath285 almost surely .
it remains to estimate the difference @xmath286 for every @xmath287 , @xmath288 in @xmath289 which coincide over @xmath227 $ ] .
this integral is bounded by @xmath290 the use of gives @xmath291 where @xmath36 , which depends _ a priori _ on @xmath292 , can be chosen uniformly over @xmath289 , as all @xmath293 norms are bounded by @xmath294 . to bound the second term ,
we first assume that @xmath295 for @xmath296 , which is always possible up to introduce an intermediate third boundary .
hence , @xmath297 . to lighten notations ,
we introduce @xmath298 as well as @xmath299 which defines a continuous extension of @xmath300 outside @xmath226 .
finally , we define the vector fields @xmath301 they satisfy @xmath302\vert_{\gamma_{1,2 } } & = 0 , \quad \left [ { { \partial}}_n u - q \otimes
n \right]\vert_{\gamma_{1,2 } } = -{{\partial}}_n g_{\omega_2}((0,0),y)\vert_{\gamma_{1,2 } } . \end{aligned } \right.\ ] ] a direct energy estimate yields @xmath303 note that all @xmath304 in both @xmath305 and @xmath306 satisfy @xmath307 .
using , we end up with @xmath308 back to @xmath244 , we obtain @xmath309 this ends the proof of proposition .
_ this section is devoted to the proof of lemma [ green2 ] _ , that is sharp estimates on the green function @xmath310 for the stokes operator above the humped boundary @xmath311 , where @xmath14 belongs to @xmath293 .
a fundamental remark is that the green function satisfies the scaling @xmath312 we want estimates to hold for @xmath313 large , that is for @xmath314 small . by relation , to establish such estimates amounts to get local estimates for the green function @xmath315 .
thus , this is again a homogenization problem : more precisely , we must show that the oscillations of the boundary at frequency @xmath316 do not affect too much the estimates on @xmath315 , so that it behaves as the green function for a half - plane .
a very close problem has been considered in the papers @xcite by avellaneda and lin , namely the derivation of local estimates for elliptic systems @xmath317 , in which @xmath318 is a positive definite matrix with periodic coefficients .
our reasoning follows these papers . for all @xmath319 , @xmath320
, we will denote @xmath321 the disk of center @xmath322 and radius @xmath323 , and @xmath324 an important property is that for all @xmath325 , @xmath326 independent of @xmath9 .
more precisely , @xmath327 only involves the lipschitz norm of @xmath328 , wich is bounded uniformly in @xmath9 .
this will be used implicitly throughout the sequel .
the core of the proof is to derive elliptic estimates uniform with respect to @xmath9 on the following stokes problem : @xmath329 where @xmath330 . more precisely , there are two steps in the proof : 1 .
we show a @xmath9-uniform hlder estimate on @xmath331 : for all @xmath332 , @xmath333 and for @xmath334 , @xmath335 where @xmath36 depends only on @xmath336 .
thanks to this hlder estimate , we prove .
the two next paragraphs correspond to these steps . to obtain a hlder regularity result ,
a classical approach is to use a characterization of hlder spaces due to campanato ( see @xcite ) : for @xmath24 an open connected bounded set , @xmath337 iff @xmath338 and @xmath339 one then tries to control such local integrals through energy estimates .
this approach has been successful in the study of elliptic systems , see the work of giaquinta and coauthors @xcite .
it extends to the stokes type equations , _ cf _ article @xcite . for us
, it amounts to controlling @xmath340 where @xmath331 is solution of .
note that , thanks to ( see @xcite ) , @xmath341 with @xmath342 independent of @xmath9 . in our case
, the main problem is to keep track of the dependence of @xmath343 on @xmath9 .
it involves a discussion of the way @xmath9 relates to @xmath323 . broadly speaking ,
the idea is the following : if @xmath323 is large compared to @xmath9 , then the oscillations have small enough amplitude to apply the regularity results of the flat case . on the contrary ,
when @xmath323 gets as small or even smaller than @xmath9 , one can rescale everything by a factor @xmath9 , so that oscillations of the boundary have frequency @xmath344 , and are no longer annoying .
implementation of this idea is a bit technical , and follows closely the work of avellaneda and lin .
we first remind a few elements of regularity theory for stokes type systems .
let @xmath24 an open connected bounded set , with _
lipschitz boundary_. then , for any @xmath345 satisfying @xmath346 , the problem @xmath347 has one solution @xmath108 satisfying @xmath348 , where _ @xmath36 can be taken as an increasing function of @xmath349 and of the lipschitz constant @xmath38 of the boundary _ , see @xcite . thanks to this result , one has quite easily , @xmath350 where @xmath351 satisfies ( in the distributional sense ) @xmath352 again , the constant @xmath36 in depends only on @xmath349 and the lipschitz constant of the boundary .
we now state the famous cacciopoli inequality : for all @xmath325 , any solution @xmath353 of satisfies @xmath354 _ sketch of proof : _ we remind the main elements of proof .
let @xmath327 a smooth function with compact support in @xmath355 , with @xmath356 on @xmath321 .
note that @xmath357 .
multiplying by the test function @xmath358 , and integrating by parts , one has easily @xmath359 using , we get @xmath360 where @xmath361 is the solution of @xmath362 note that the previous bound is uniform in @xmath9 , as it only involves the lipschitz constant of @xmath328 which is uniformly bounded .
a simple energy estimate on @xmath46 gives @xmath363 as @xmath364 , and using hlder inequality on @xmath365 , we end up with @xmath366 where @xmath173 is arbitrary small .
we conclude as in ( * ? ? ?
* theorem 1.1 , page 180 ) .
inequality of type has been used by giaquinta and modica in the study of elliptic regularity . in the context of stokes type system
, they obtain a local estimate , see @xcite : [ modica ] let @xmath24 of class @xmath367 , and @xmath368 , @xmath369 , satisfying @xmath370 then , @xmath371 for @xmath372 , and @xmath373 unfortunately , we can not use this theorem assuch .
indeed , the constant @xmath36 in the last regularity estimate involves the modulus of continuity of @xmath374 , where @xmath375 describes the boundary . in our case @xmath376 ,
such modulus of continuity is not uniformly bounded in @xmath9 .
we must proceed in several steps to control the local integrals @xmath343 .
note that theorem [ modica ] implies estimate when @xmath377 is far from the boundary .
thus , we can restrict ourselves to a case in which @xmath378 is close to the oscillating boundary , for instance belongs to the axis @xmath379 .
[ lemme1 ] for all @xmath380 small enough , there exists @xmath381 such that for all @xmath382 , and for all solutions of satisfying @xmath383 , @xmath384 , one has @xmath385 _ proof of the lemma : _ suppose that the result does not hold
. then one can find @xmath380 arbitrary small , and sequences @xmath386 , @xmath387 satisfying @xmath388 { } 0 , \quad \| u^{{{\varepsilon}}_j } \|_{l^2(d^{{{\varepsilon}}_j}(x_0,\theta ) ) } > \theta^{\mu+1}.\ ] ] one can extend all the @xmath389 , @xmath387 by @xmath66 outside @xmath390 so that all these functions are defined on the fixed domain @xmath391 . from the @xmath3 bound on @xmath386 , up to extract a subsequence , we get @xmath392 and by cacciopoli inequality , @xmath393 one can then take the limit in , which yields @xmath394 as the upper half plane is a regular domain , we can apply theorem [ modica ] , so that for all @xmath395 , for all @xmath380 , @xmath396 for @xmath380 small enough , it contradicts the lower bound on @xmath397 .
we fix @xmath380 , @xmath398 as in lemma [ lemme1 ] .
we then state [ lemme2 ] for all @xmath399 satisfying @xmath400 ( @xmath216 ) , @xmath401 _ proof of the lemma : _ the lemma is deduced from an induction argument on @xmath402 . for @xmath403 , the bound is given by lemma [ lemme1 ] .
assume now that this bound holds for @xmath216 .
up to a horizontal translation , we can assume that @xmath404 .
then , we set @xmath405 and introduce the rescaled functions @xmath406 they satisfy @xmath407 moreover , one has @xmath408 , and by the induction assumption @xmath409 applying lemma [ lemme1 ] to @xmath46 and @xmath410 yields the result .
we can now finish the proof of estimate .
let @xmath411 .
we need to bound @xmath412 , for @xmath320 .
there are two cases to handle : * _ the distance between @xmath322 and the boundary @xmath413 satisfies @xmath414 . _ up to take a smaller @xmath398
, we can suppose that @xmath415 , which implies that there exists @xmath416 on the axis @xmath417 with @xmath418 .
by lemma [ lemme2 ] , for all @xmath419 , @xmath420 if @xmath421 , @xmath422 , and the previous line implies @xmath423 on the contrary , if @xmath424 , then @xmath425 ( it does not intersect the boundary ) . a simple rescaling of yields @xmath426 thus , @xmath427 now , by lemma [ lemme2 ] , as @xmath428 , @xmath429 using the two last inequalities , we end up again with @xmath430 which in turn clearly implies @xmath431 as @xmath432 , this gives the required estimate .
* _ the distance between @xmath322 and the boundary @xmath413 satisfies @xmath433 . _ this time , there exists @xmath416 on the axis @xmath434 such that @xmath435 .
again , for all @xmath436 , @xmath437 and implies @xmath438 it remains to handle the case @xmath439 .
up to a horizontal translation , we can assume that @xmath440 .
we introduce the rescaled functions @xmath441 they satisfy in particular @xmath442 it is a stokes type system set in a domain independent of the small parameter @xmath9 .
hence , we can apply theorem [ modica ] , which yields : for all @xmath443 , @xmath444 back to the original unknowns @xmath23 , @xmath365 , we obtain the control of @xmath343 for @xmath445 . this ends the proof . from the above hlder estimate
, we can deduce the estimate .
the arguments are again adapted from article @xcite . for the sake of completeness
, we describe the ideas at play .
the first step is to establish the following bound : for all @xmath446 in @xmath447 , @xmath448 where @xmath36 only involves @xmath336 .
note that it implies . to lighten the notations
, we drop the suffix @xmath14 , denoting @xmath449 , @xmath119 instead of @xmath450 .
let us introduce the green function @xmath451 for the stokes operator over @xmath452 .
namely , it satisfies for all @xmath453 @xmath454 one has easily that @xmath455 the point is to show that @xmath456 the estimate is then obtained by integration with respect to @xmath457 , at @xmath458 . such bound on @xmath459 will be deduced from a repeated use of the hlder estimate .
note that this estimate extends without difficulty to similar stokes problems in dimension @xmath460 , with @xmath461 and @xmath462 .
in particular , it holds when the domain is @xmath463 .
namely , let @xmath464 and @xmath465 .
let @xmath466 , and @xmath467 .
we consider the quantity @xmath468 the field @xmath23 satisfies a stokes equation with source term @xmath365 over @xmath469 , with a dirichlet boundary condition .
we therefore apply the estimate to @xmath23 .
properly rescaled , it yields @xmath470 where we used the fact that @xmath365 vanishes over @xmath471 .
thus , we get @xmath472 where the last two inequalities come respectively from the sobolev imbedding theorem ( note that @xmath23 is zero at the boundary so that the imbedding does not involve lower order terms ) , and from the standard energy estimate on the stokes system . by duality
, we infer that @xmath473 using that @xmath474 satisfies a homogenenous stokes system over @xmath475 , one more application of leads to @xmath476 inequality at hand , we can derive the final estimate on @xmath119 .
let @xmath477 . set this time @xmath478
. for all @xmath479 such that @xmath480 , implies @xmath481 applying to the function @xmath482 , we get for any @xmath483 @xmath484 that leads to @xmath485 now reversing the roles of @xmath322 and @xmath486 , we obtain @xmath487 using the scaling relation , we get for all @xmath488 for @xmath489 , @xmath490 using classical local regularity results for the stokes equation in a @xmath293 domain ( see ( * ? ? ?
* theorem 1.3 , page 198 ) , which extends theorem [ modica ] ) : for @xmath262 , @xmath491 that is exactly estimate .
as mentioned several times in this paper , the well - posedness of system is not known without a structural assumption on @xmath14 , like periodicity or stationarity .
we stress however that thanks to our estimates on the green function @xmath271 , the representation formula defines a solution of for any @xmath293 boundary , _ cf _ the fourth section .
hence , the open issue is rather to find the appropriate functional space for uniqueness .
such difficulty does not arise when the stokes operator is replaced by the laplacian , or more generally by a scalar elliptic operator .
hence , one can show well - posedness in @xmath492 of @xmath493 if the function @xmath14 is bounded and lipschitz .
for the existence part , one may consider , for all @xmath198 , the solution @xmath494 of @xmath495 by the maximum principle , @xmath496 , so that up to a subsequence it converges to some @xmath46 in @xmath492 weak*. straightforwardly , @xmath46 satisfies .
for the uniqueness part , let @xmath497 satisfying @xmath498 let us show that @xmath499 . as we do not know the behavior of @xmath46 at infinity
, it does not follow directly from the maximum principle . in the case of a lipschitz boundary @xmath14
, we can conclude to the uniqueness in the following way . by the change of variables @xmath500 ,
the previous equation becomes @xmath501 for some elliptic matrix @xmath502 with bounded coefficients .
we extend @xmath318 and @xmath46 to @xmath503 by the formulas , @xmath504 @xmath505 in this way , we get @xmath506 on all @xmath30 .
harnack s inequality for elliptic equations ( see @xcite ) leads to @xmath507 for any @xmath508 and @xmath509 such that @xmath510 . with @xmath511 and @xmath512 going to infinity
, we obtain that @xmath513 .
a key element in the paper is the estimate on the green function for the stokes operator above an oscillating boundary .
this estimate relies itself on the hlder regularity result .
in fact , with some more calculations in the same spirit , one could show a refined bound : for a @xmath514 boundary @xmath14 and a source term @xmath515 ( @xmath516 , the solution @xmath517 of satisfies @xmath518 with @xmath36 independent of @xmath9 .
this yields an optimal bound on the green function , that is @xmath519 estimate can in turn be improved as @xmath520 such bound is far easier to prove when is replaced by the scalar system . in this case , one may use a representation in terms of the standard two - dimensional brownian motion @xmath521 .
if we denote @xmath522 the probability space on which this brownian motion is defined , @xmath523 where @xmath266 is the exit time from @xmath280 ( see @xcite ) .
we now want to bound @xmath524 where @xmath525 is the exit time from @xmath526 .
we remind that @xmath527 over @xmath227 $ ] , so the exit times are the same for brownian particles leaving in the region @xmath528 $ ] .
hence , @xmath529 where we denote by @xmath530 the first time for which @xmath531 , and @xmath532 the first time for which @xmath533 .
it is well - known that the distributions of these hitting times are @xmath534 a straightforward calculation provides @xmath535 which gives the result .
theorem [ theosharp ] shows that the boundary layer solution @xmath46 converges _ at least as _ @xmath536 .
one may wonder if this result is optimal , that is if we can find roughness distributions for which the speed of convergence is exactly given by @xmath536 . in other words , is the constant @xmath537 of the theorem positive for some random distribution of roughness ? we have not so far been able to show optimality in this setting , but it can be established for the easier dirichlet problem @xmath538 where @xmath14 is a given boundary data .
although simpler , this system shares many features with the original system : * if @xmath539 is say @xmath50-periodic , the solution @xmath540 converges exponentially fast to the constant @xmath541 , as @xmath154 goes to infinity .
* if @xmath14 belongs as before to the probability space @xmath542 , one can show under assumption ( h2 ) that @xmath543 { } \ : \sigma^2 \ : \ge 0 , \quad \alpha : = { { \mathbb e}}(\omega \mapsto \omega(0)).\ ] ] along the lines of @xcite , we will exhibit a stationary measure @xmath62 for which @xmath544 .
of course , @xmath62 is the law of the random process @xmath545 , so that we just need to characterize the random initial data .
let @xmath546 a gaussian random process , of zero mean and covariance @xmath547 , where @xmath548 is a smooth even function with compact support .
note that such process exists : take @xmath549 , with @xmath550 an even smooth function with compact support .
then , its fourier transform satisfies @xmath551 , which ensures the required positivity property @xmath552 for any family @xmath553 with compact support .
note moreover that this process defines almost surely smooth functions of @xmath47 : indeed , a simple calculation yields @xmath554 } |{{\partial}}_{y_1}^k x(\cdot , y_1)|^2 \ , dy_1 \right ) \ : = \ : 2r \ , ( -1)^k \ , \rho^{(2k)}(0 ) \ : < \ : \infty\ ] ] so that @xmath555 is almost surely in the space @xmath556 and therefore smooth .
finally , we introduce @xmath557 for a smooth increasing function @xmath558 with values in @xmath559 .
we stress that @xmath560 satisfies ( h2 ) as @xmath561 has compact support .
we will show that the corresponding @xmath83 is positive .
suppose _ a contrario _ that @xmath562 .
for @xmath563 , we introduce the measure @xmath564 associated to the gaussian process with variance @xmath547 but mean @xmath565 where @xmath410 will be given later .
note that @xmath564 is associated to the random initial data @xmath566 standard computation yields @xmath567 and @xmath568 if @xmath569 , then a simple cauchy - schwartz inequality @xmath570 and _ goes to zero as @xmath116 if @xmath571 is bounded from above_. let @xmath331 and @xmath46 solutions associated to the initial data @xmath572 and @xmath573 . as @xmath574 , by monotonicity , @xmath575 .
we can express @xmath331 and @xmath46 in terms of the poisson kernel , so that @xmath576 now , we define for @xmath563 @xmath577 where @xmath578 has compact support , @xmath579 over @xmath580 .
on one hand , with this definition of @xmath410 , one can check that @xmath581 on the other hand , one has @xmath582 this implies that the quantity @xmath583 does not go to zero , leading to a contradiction .
10 constructing wall laws with domain decomposition or asymptotic expansion techniques .
, 1 - 2 ( 1998 ) , 215232 .
symposium on advances in computational mechanics , vol . 3 ( austin , tx , 1997 ) . domain decomposition & wall laws . in _ recent developments in domain decomposition methods and flow problems ( kyoto , 1996 ; anacapri , 1996 ) _ , vol .
11 of _ gakuto internat .
_ gakktosho , tokyo , 1998 , pp .
effective boundary conditions for laminar flows over periodic rough boundaries .
, 1 ( 1998 ) , 187218 . effect of rugosity on a flow governed by stationary navier - stokes equations . , 4 ( 2001 ) , 769785 .
compactness methods in the theory of homogenization . ,
6 ( 1987 ) , 803847 .
bounds on singular integrals in homogenization .
, 8 - 9 ( 1991 ) , 897910 .
decay of correlations . in _
smooth ergodic theory and its applications ( seattle , wa , 1999 ) _ , vol .
69 of _ proc .
pure math .
soc . , providence , ri , 2001 , pp .
297325 .
wall laws for fluid flows at a boundary with random roughness .
available at ` http://www.dma.ens.fr/~dgerardv/publications.html ` . the viscous flow on surfaces with longitudinal ribs . , 1 ( 1989 ) , 105129 .
roughness - induced effects on the quasi - geostrophic model . , 1 ( 2005 ) , 81119 .
higher order boundary layer correctors and wall laws derivation : a unified approach . .
long wave expansions for water waves over random topography .
( 2007 ) . , second ed . duxbury press , belmont , ca , 1996 . ,
38 of _ springer tracts in natural philosophy_. springer - verlag , new york , 1994 .
linearized steady problems . highly rotating fluids in rough domains . ,
11 ( 2003 ) , 14531498 .
105 of _ annals of mathematics studies_. princeton university press , princeton , nj , 1983 .
nonlinear systems of the type of the stationary navier - stokes system .
( 1982 ) , 173214 . .
classics in mathematics .
springer - verlag , berlin , 2001 .
reprint of the 1998 edition . on the roughness - induced effective boundary conditions for an incompressible viscous flow .
, 1 ( 2001 ) , 96122 .
couette flows over a rough boundary and drag reduction .
, 3 ( 2003 ) , 429455 .
asymptotic analysis of the laminar viscous flow over a porous bed . ,
6 ( 2000 ) , 20062028 ( electronic ) . .
springer - verlag , berlin , 1994 . translated from the russian by g. a. yosifian [ g. a. iosif yan ] .
asymptotic analysis of laminar boundary - layer flow over finely grooved surfaces .
, 2 ( 1995 ) , 169195 .
behavior of the solution of a random semilinear heat equation .
available at ` http://www-rcf.usc.edu/~zygouras/research.html ` . .
notes based on a course given at new york university during the year 1967/68 .
courant institute of mathematical sciences new york university , new york , 1968 .
the author warmly thanks s.r.s varadhan for pointing to reference @xcite , as well as luis silvestre and thierry levy for fruitful discussions .
we want to show here that @xmath584 defines a measurable function from @xmath64 to @xmath68 .
let @xmath585 a sequence of smooth functions with compact support , @xmath586 .
we define @xmath587 note that @xmath588 is the ( unique ) solution of @xmath589\vert_{y_2 = 0 } = 0 , \quad [ { { \partial}}_2 w_n - ( 0,q_n)]\vert_{y_2 = 0 } = ( -\varphi_n,0 ) , \end{aligned } \right.\ ] ] satisfying @xmath590 . by the dominated convergence theorem applied to the integral formula , we get that @xmath591 in @xmath592 . by the cacciopoli inequality ,
the convergence is also true in @xmath593 .
thus , we just have to show measurability of @xmath588 .
let us define @xmath594 following the lines of @xcite it can be shown that the application @xmath595 , where @xmath596 is the orthogonal projection from @xmath150 to @xmath597 , is measurable .
now , @xmath588 is the unique fixed point of the contraction @xmath598 where @xmath599 is the unique function of @xmath600 satisfying @xmath601 the measurability of @xmath588 follows . | we consider the navier - stokes equation in a domain with irregular boundaries .
the irregularity is modeled by a spatially homogeneous random process , with typical size @xmath0 . in the parent paper @xcite
, we derived a homogenized boundary condition of navier type as @xmath1 .
we show here that for a large class of boundaries , this navier condition provides a @xmath2 approximation in @xmath3 , instead of @xmath4 for periodic irregularities .
our result relies on the study of an auxiliary boundary layer system .
decay properties of this boundary layer are deduced from a central limit theorem for dependent variables . _
keywords : _ wall laws , rough boundaries , stochastic homogenization , decay of correlations |
data management systems , like database , information retrieval ( ir ) , information extraction ( ie ) or learning systems , store , organize , index , retrieve and rank information units , like tuples , objects , documents , items .
a wide range of applications of these systems have emerged that require the management of uncertain or imprecise data .
important examples of data are sensor data , webpages , newswires , imprecise attribute values . what is common to all these applications is uncertainty and then that they have to deal with decision and statistical inference .
ranking is perhaps the most crucial task performed by the data management systems which have to deal with uncertainty . in many applications , ranking aims at deciding or inferring , for example ,
the class assigned to a unit or the order by relevance , usefulness , or utility of the units delivered to another application or to an end user .
in addition , ranking is performed to decide whether a unit is placed at a given rank .
the management of imprecise data require means for ranking information units by probability .
ranking places information units in a list ordered by a measure of utility , cost , relevance , etc .. a probability theory measures the uncertainty of the decision . to this end , the definition of an event space and the estimation of probabilities are necessary steps for representing imprecise data and making predictions within many contexts of data management like machine learning , information retrieval or probabilistic databases .
the measurement of the imprecision and the uncertainty in the data leads to the definition of regions of acceptance of a predefined set of hypotheses , thus bringing many decision problems to the calculation of a probability of detection and of a probability of false alarm .
although the data management systems reach good results thanks to classical probability theory and parameter tuning , ranking is far from being perfect because useless units are often ranked at the top or useful units are missed .
classical probability theory describes events and probability distributions using sets and set measures , respectively , according to kolmogorov s axioms @xcite .
in contrast , quantum probability theory describes events and probability distributions using hermitian operators in the complex hilbert vector space .
whereas parameter tuning is performed within a fixed probability theory , the adoption of quantum probability entails a radical change . furthermore , whereas classical probability is based on sets such that the regions of acceptance or rejection are set - based detectors ( i.e. , indicator functions ) , quantum probability is based on subspace - based detectors and the detectors are projector - based .
note that the use of quantum probability does not imply that quantum phenomena are investigated in the paper ; we are interested in the formalism based the hilbert vector spaces instead .
the main question asked in the paper is whether further improvement may be obtained if the classical probability theory is replaced by the quantum probability theory .
the paper shows that ranking information units by quantum probability yields different outcomes which are in principle more effective than ranking them by classical probability given the same data available for parameter estimation .
the effectiveness is measured in terms of probability of detection ( also known as recall or power ) and probability of false alarm ( also known as fallout or size ) .
we structure the paper as follows .
section [ sec : class - prob - quant ] illustrates the basics of the probability theory through a view that encompasses both theories .
section [ sec : quant - prob - detect ] compares quantum detection with classical detection .
section [ sec : optim - rank - quant ] shows that the ranking by quantum probability more effective than the ranking by classical probability .
section [ sec : interpr - quant - proj ] provides an interpretation of the projectors which define the regions of acceptance and rejection .
section [ sec : impl - rank ] describes the algorithm for ranking information units by quantum probability .
section [ sec : related - work ] provides an overview of the related work . onto @xmath0 . ]
in this section , we introduce a special view of probability distributions for the classical theory of probability .
the same view is also introduced for quantum probability , which is a non - classical theory and does not admit the distributive law , to provide a general framework for quantum and classical probabilities ; the view is depicted in figure [ fig : correspondence ] . before introducing the view of probability theory ,
some basic definitions are provided .
a probability space is a set of mutually exclusive events such that each event is assigned a probability between @xmath1 and @xmath2 and the sum of the probabilities over the set of events is @xmath2 .
for the sake of clarity , we introduce the case of binary event spaces because it is the simplest and most common in data management
keyword occurrence in webpages , binary features in sample records or binary attribute values in relational tables are some examples .
the case of binary event spaces are usually represented by mutually exclusive scalars like @xmath1 and @xmath2 .
if binary scalars are used , the mutual exclusiveness is given by the scalar product , for example , @xmath3 ( see @xcite ) . whereas the scalars @xmath4 is a possible representation of events , the vectors of a complex finite - dimensional
are another option .
when using vectors , an event is @xmath5 and its complement is @xmath6 .
the representation of the events must encode the mutual exclusiveness .
if binary vectors are used , the mutual exclusiveness is given by the inner product , i.e. , @xmath7 .
when the event space is not binary ( e.g. , when the events are represented by @xmath8 natural numbers @xmath9 ) , a binary representation can again be used .
the vector @xmath10 is assigned to symbol @xmath1 , the vector @xmath11 is assigned to symbol @xmath2 , and so on until @xmath12 is assigned to symbol @xmath13 .
whatever the representation is used , the inner product between two vectors must be @xmath1 and their norm must be @xmath2 .
the mapping between the probabilities and the events is called `` probability distribution '' which is a function mapping a mathematical object which represents an event to a real number ranging between @xmath1 and @xmath2 . the difference between classical probability and quantum probability ; the difference is due to the way the event space and the probability distribution are represented .
the starting point of the view of probability used in the paper is the algebraic form of the probability space . to this end , hermitian ( or self - adjoint ) linear operators are used . in quantum mechanics , `` operator '' is preferred to `` matrix '' yet in the paper , for the sake of clarity , `` matrix '' is preferred because , for a fixed basis , the matrices are isomorphic to the operators .
a matrix is hermitian when it is equal to its conjugate transpose .
hermitian matrices are important because their eigenvalues are always real .
in particular , hermitian matrices with trace @xmath2 is the key notion in quantum probability because the sum of the eigenvalues is @xmath2 and , thus , the eigenvalues can be viewed as a probability distribution .
the projector is an idempotent hermitian matrix .
every subspace has one projector and then the projectors are 1:1 correspondence with the subspaces .
each vector corresponds to one projector with rank one defined as the outer product of the vector by its conjugate transpose .
there are two main instructions for representing events using projectors : * the projectors must be mutually orthogonal for representing the mutual exclusiveness of the events , and * the projectors must have trace @xmath2 for making probability calculation consistent with the probability axioms . an event space and a probability function defined over it are represented using hermitian matrices with trace @xmath2 . in particular , a projector represents an event and an event space is modeled by a collection of projectors . as the union of the events results in the whole event space , the sum of the projectors of a collection corresponding to an event space results in the unity . more specifically ,
if @xmath14 is a collection of mutually orthogonal projectors , @xmath15 the latter being termed `` resolution to the unity '' .
for example , using the dirac notation introduced in appendix [ sec : dirac - notation ] , the projector of two events are represented by @xmath16 0 \end{array } \right ) \qquad { |0\rangle } = \left ( \begin{array}{c } 0 \\ [ 5pt ] 1 \end{array } \right)\ ] ] and @xmath17 0 & 0 \end{array } \right ) \qquad { |0\rangle\langle0| } = \left ( \begin{array}{cc } 0 & 0 \\ [ 5pt ] 0 & 1 \end{array } \right)\ ] ] however , there is not a unique representation of an event space .
for example , the following vectors are also representing mutually exclusive events : @xmath18 \frac{1}{\sqrt{2 } } \end{array } \right ) \qquad \left ( \begin{array}{r } \frac{1}{\sqrt{2 } } \\ [ 5pt ] -\frac{1}{\sqrt{2 } } \end{array } \right)\ ] ] thus leading to a different resolution to the unity given by the following projectors @xmath19 \frac{1}{2 } & \frac{1}{2 } \end{array } \right ) \qquad \left ( \begin{array}{rr } \frac{1}{2 } & -\frac{1}{2 } \\ [ 5pt ] -\frac{1}{2 } & \frac{1}{2 } \end{array } \right)\ ] ] the second kind of hermitian matrix of a probability space is the density matrix ; the density matrix encapsulates the probability values assigned to the events . in physics
, a density matrix represents the _ state _ of a microscopic system , such as a particle , a photon , etc .. the structure of a microscopic system is unknown . yet
a device can measure the system to obtain some information .
a microscopic system is similar to a urn of colored balls .
the internal composition of the urn is always unknown even if opened and observed because the device disturbs the state ( i.e. , the distribution of the colors ) of the urn . in data management and in other domains different from particle physics , a system is macroscopic instead .
examples of macroscopic systems in data management are webpages , customers , queries , clicks , tuples , attributes , and so on .
the states of these systems correspond to the probability densities according to which keywords , reviews , attribute values are observables to be measured from such systems .
density matrices are a powerful formalism in the macroscopic worlds too because they allow us to introduce the algebraic approach adopted for incorporating the more powerful probability space and decision rule suggested in the paper . to the end of introducing the way density matrices
are defined , consider two equiprobable events , e.g. , the occurrence of a feature or a positive / negative customer review .
the probability distribution is @xmath20 where each value refers to an event . as an alternative to a list
, the probability distribution can be arranged along the diagonal of a two - dimensional matrix and the other matrix elements are zeros .
for example , the matrix corresponding to the probability distribution of two equally probable events is @xmath21 0 & \frac{1}{2 } \end{array } \right)\ ] ] in general , the probability distribution @xmath22 of a @xmath8-event space can be written as @xmath23 0 & p_2 & \cdots & 0 \\ [ 5pt ] \vdots & \vdots & \ddots & \vdots\\ [ 5pt ] 0 & 0 & \cdots & p_k \\ [ 5pt ] \end{array } \right)\ ] ] a probability distribution is _ pure _ when the density matrix is a projector , otherwise , the distribution is _
mixed_. a distribution is mixed when the density matrix is a mixture of density matrices ; a pure distribution is an instance of mixture with one matrix .
the density matrix representing a pure distribution is 1:1 correspondence with a density vector such that the projector is the outer product between the vector and its conjugate transpose .
a classical probability distribution is pure when the probability is concentrated on a single elementary event which is the certain event and then has probability @xmath2 . given a density matrix , the spectral theorem helps find the underlying events and the related probabilities .
because of the importance of the spectral theorem , we provide its definition below : [ sec : class - prob - quant-1 ] to every hermitian matrix @xmath24 on a finite - dimensional complex inner product space there correspond real numbers @xmath25 and rank - one projectors @xmath26 so that the @xmath27 s are pairwise distinct , the @xmath28 are mutually orthogonal , @xmath29 , @xmath30 and @xmath31 . see @xcite .
the eigenvalues are the spectrum and @xmath26 are the projectors of the spectrum of @xmath24 . from theorem [ sec : class - prob - quant-1 ] , thus , a pure distribution is always a rank - one projector .
the spectral theorem says that any hermitian matrix corresponding to a distribution can be decomposed as a linear combination of projectors ( i.e. pure distributions ) where the eigenvalues are the probability values associated to the events represented by the projectors .
the eigenvalues are real because the decomposed matrix is hermitian , are non - negative and sum to @xmath2 . for example
, when the matrix corresponding to the distribution of two equally probable events is considered , the spectral theorem says that @xmath21 0 & \frac{1}{2 } \end{array } \right ) = \frac{1}{2 } \left ( \begin{array}{cc } 1 & 0 \\ [ 5pt ] 0 & 0 \end{array } \right ) + \frac{1}{2 } \left ( \begin{array}{cc } 0 & 0 \\ [ 5pt ] 0 & 1 \end{array } \right)\ ] ] a mixed distribution have more non - zero eigenvalues , a pure distribution has a single eigenvalue @xmath2 . in classical probability , every pure distribution represented by a diagonal density matrix corresponds to a projector .
however , in general , a density matrix is not necessarily diagonal , yet the matrix is necessarily hermitian .
for example , are trace - one projectors and correspond to pure distributions , thus there is a certain event ( with probability @xmath2 ) and an impossible event ( with probability @xmath1 , of course ) . yet , they are not diagonal .
when , for example , keyword occurrence in webpages is represented , the first projector may be assigned eigenvalue @xmath2 and the other is assigned eigenvalue @xmath1 .
thus the former represents the certain event and the latter represents the impossible event in the probability space .
when using the algebraic form to represent probability spaces , the function for computing a probability is the trace of the matrix obtained by multiplying the density matrix by the projector corresponding to the event .
the usual notation for the probability of the event represented by projector @xmath32 when the distribution is represented by density matrix @xmath33 is @xmath34 also known as born s rule @xcite .
for example , when @xmath35 @xmath36 0 & \frac{1}{2 } \end{array } \right ) \qquad \mathbf{e } = \left ( \begin{array}{cc } 1 & 0 \\ [ 5pt ] 0 & 0 \end{array } \right)\ ] ] the probability is @xmath37 0 & \frac{1}{2 } \end{array } \right ) \left ( \begin{array}{cc } 1 & 0 \\ [ 5pt ] 0 & 0 \end{array } \right ) \right ) = { \mathrm{tr}}\left ( \left ( \begin{array}{cc } \frac{1}{2 } & 0 \\ [ 5pt ] 0 & 0 \end{array } \right ) \right ) = \frac{1}{2}\ ] ] when @xmath38 is a rank - one projector , the trace - based probability function can be written as @xmath39 when @xmath33 is a rank - one projector @xmath40 , then @xmath41 from the example , the definition of a function that computes the probability of an event when the probability is already allocated in the diagonal of the density matrix may be odd .
however , we have shown that not all the density matrices corresponding to a distribution need to be diagonal matrices and the diagonal elements do not necessarily correspond to probability values , although they do have to sum to @xmath2 .
a density matrix encapsulate the values assigned to the events by a probability function because of gleason s theorem stated below and proved in @xcite . to every probability distribution on the set of all projectors in a complex vector space with dimension greater than @xmath42
there corresponds a unique density matrix @xmath33 on the same vector space for which the probability of the event represented by a projector @xmath43 is @xmath44 for every unit vector @xmath45 in the vector space .
basically , the theorem tells us that corresponding to a probability distribution is one density matrix such that the probability of any event represented as a projector is calculated by the trace function . the probability of an event when computed using a mixture differs from the probability computed using a pure state , yet they share the classical probability term whereas the difference is called _ interference term_. using a mixture , @xmath46 0 & |a_0|^2 \end{array } \right)\ ] ] using superposition , @xmath47 where @xmath48 and @xmath49 is the angle of the polar representation of the complex number @xmath50 .
suppose , as an example , that @xmath51 represents the event `` the keyword occurs '' and the density matrix represents the probability distribution of keyword occurrence in useful webpages .
the common factor ( i.e. ) is the sum of two probabilities ; the probability that the webpage is not useful ( @xmath52 ) multiplied by the probability that the keyword occurs in a useless webpage ( @xmath53 ) , and the probability that the webpage is useful ( @xmath54 ) multiplied by the probability that the keyword occurs in a useful webpage ( @xmath55 ) .
the sum is nothing but an application of the law of total probability .
the quantity @xmath56 is the interference term .
as the interference term ranges between @xmath57 and @xmath58 , the probability of keyword occurrence computed when usefulness is superposed with uselessness becomes different from the common factor in which usefulness and uselessness are mutually exclusive and their probability distribution is described by a mixture .
the interference term can be so large that the law of total probability is violated and any probability space obeying kolmogorov s axioms can not admit the probability values @xmath54 and @xmath55 , thus requiring the adoption of a quantum probability space .
in general , the information stored in the data is acquired and delivered through information unit representation and ranking , these processes are described in terms of decision and estimation , and they are therefore affected by error . the error could be eliminated only if precise and exhaustive methodological tools and computer systems were developed .
nevertheless , there is a trade - off between precision , exhaustivity and the computation cost because high level of the former can be achieved only if a high computation cost is devoted .
thus , a certain amount of error is unavoidable yet can be controlled and limited below a given threshold . either a set of statements , or hypotheses , must be decided to best describe the information unit insofar as data permit to judge ( e.g. , the best topic(s ) to which a webpage is assigned ) , or the values of certain quantities ( also known as parameters ) characterizing the information unit must be estimated , the probability of detection and the probability of false alarm related to a decision must be calculated . in the paper , a great deal of attention is paid to decision whereas estimation is set apart not because estimation is little important , but because estimation would require another research stream had it to be addressed to the appropriate level of exhaustivity .
many tasks in data management are decision problems , examples are the classification of images with respect to predefined patterns , the categorization of webpages to topics , contextual advertising ( i.e. , the decision whether an ad has to displayed in a search engine result page ) , the retrieval and ranking of webpages ( i.e. , the decision as to whether a webpage has to put at rank @xmath59 of a search engine result page ) , probabilistic databases ( i.e. , the decision about the correct value of an attribute and then the computation of the associated probability ) .
our illustration of decision theory is necessarily brief and confined to its simplest aspects and examples .
the illustration is also organized in such a way as to bring out most clearly the parallels between classical probability - based decision and quantum probability - based decision .
the examples are chosen from elementary information retrieval or machine learning theory and perhaps provide a basis for comparison with the quantum case .
a certain information unit ( e.g. , a webpage or an store item ) is observed in such a way as to obtain numbers ( e.g. , the pagerank or the number of positive reviews ) on the basis of which a decision has to be made about its state
. the numbers observed are , for example , the frequency of a feature in the information unit , the simplest example being the frequency of a keyword in a webpage used for calculating search engine statistical ranking functions . for the sake of clarity
, we use the binary frequency and the feature presence / absence case in the paper .
the state might be , for example , the relevance of the webpage to the search engine user s interests or the customer s willingness to buy the store item .
the use of the term `` state '' is not coincidental because the numbers are observed depending upon the density matrix , which is indeed the mathematical notion implementing the state of a system .
thus , quantum probability ascribes the decision about the state of an information unit to testing the hypothesis that the density matrix has generated the observed numbers . consider the hypothesis that the state of the system is the density matrix @xmath60 and the alternative hypothesis that the state of the system is the density matrix @xmath61 .
the two hypotheses can be labeled @xmath62 and @xmath63 , respectively . in data management ,
hypothesis @xmath63 asserts , for example , that a customer does not buy an item or that a webpage shall be irrelevant to the search engine user whereas hypothesis @xmath62 asserts that an item shall be bought by a customer or that a webpage shall be relevant to the user .
therefore , the probability that , say , a feature occurs in an item which shall not be bought by a customer or a keyword occurs in a webpage which shall be irrelevant to the search engine user depends on the state ( i.e. , the density matrix ) .
statistical decision theory is a old topic and neyman - pearson s lemma is by now one out of the most important results which provides a criterion for deciding upon hypotheses instead of the bayesian approach .
the lemma provides the rule to govern the decider s behaviour and decide the true hypothesis without hoping to know whether it is true .
given an information unit and an hypothesis about the unit , such a rule calculates a specified number ( e.g. , a feature ) and , if the number is greater than a threshold reject the hypothesis , otherwise , accept it .
such a rule tells nothing whether , say , the item shall be bought by the customer , but the lemma proves that , if the rule is followed , then , in the long run , the hypothesis shall be accepted at the highest probability of detection ( or power ) possible when the probability of false alarm ( or size ) is not higher than a threshold .
the set of the pairs given by size and power is the power curve which is also known as the receiver operating characteristic ( roc ) curve .
neyman - pearson s lemma implies that the set of the observable numbers ( e.g. , features ) can be partitioned into two distinct regions ; one region includes all the numbers for which the hypothesis shall be accepted and is termed acceptance region , the other region includes all the numbers for which the hypothesis shall be rejected and is termed rejection region . for example , if a keyword is observed from webpages and only presence / absence is observed , the set of the observable numbers is @xmath64 and each region is one out of possible subsets , i.e. , @xmath65 .
the paper reformulates neyman - pearson s lemma in terms of subspaces instead of subsets to utilize quantum probability .
therefore , the region of acceptance and the region of rejection must be defined in terms subspaces . in the following ,
we illustrate the algorithm for calculating the most efficient test in hilbert spaces .
the following result holds : [ the : helstrom ] let @xmath66 be the density matrices under @xmath67 , respectively .
the region of acceptance at the highest power at every size is given by the projectors of the spectrum of @xmath68 whose eigenvalues are positive .
see @xcite .
an optimal projector is a projector which identifies the region of acceptance and the region of rejection according to theorem [ the : helstrom ] .
we define the _ discriminant function _ as @xmath69 where @xmath32 is a projector . if the discriminant function is positive , the observed event represented by @xmath32 is placed in the region of acceptance .
suppose that the density matrix that corresponds to @xmath62 is a mixed , classical probability distribution .
the mixed case is the usual method for dealing with uncertainty in data management , even though more than one feature may exist or the feature may not be binary ; however , the number of features or the number of values of a feature is not essential in the paper .
let @xmath70 be such a mixed distribution and @xmath71 0 & 1-p_1 \end{array } \right)\ ] ] where @xmath72 0 & 0 \end{array } \right ) \qquad \mathbf{p}_0 = \left ( \begin{array}{cc } 0 & 0 \\ [ 5pt ] 0 & 1 \end{array } \right)\ ] ] similarly , @xmath73 0 & 1-p_0 \end{array } \right)\ ] ] when @xmath74 is observed , the power and the size are , respectively , @xmath75 in the classical case , @xmath76 represents the absence and the presence , respectively , of a feature .
hence , the possible acceptance or rejection regions are @xmath77 which correspond respectively to `` never accept '' , `` accept when the feature does not occur '' , `` accept when the feature occurs '' and `` always accept '' .
thus , the decision on , say , webpage classification , topic categorization , item suggestion , can be made upon the occurrence of one or more features because @xmath76 represent `` physical '' events .
furthermore , the discriminant function in the mixed case is @xmath78 the power curve can be built as follows .
suppose , as an example , that a keyword describes webpage content and that that webpage either includes ( @xmath79 ) or does not include ( @xmath80 ) the keyword .
@xmath81 only if @xmath82 and this point corresponds to the event represented by @xmath83 .
let @xmath84 be the probability that the keyword occurs in a relevant webpage or in a non - relevant webpages , respectively .
when the keyword is the unique observed feature , the webpage is presented to the user if @xmath85 and the keyword occurs , or @xmath86 and the keyword does not occur .
the power curve includes the points @xmath87 and @xmath88 .
the key point is that a mixture is not the unique way to implement the probability distributions .
as we illustrate in section [ sec : class - prob - quant ] , the superposed vectors@xmath89 \sqrt{1-p_1 } \end{array } \right ) \qquad { |\varphi_0\rangle } = \left ( \begin{array}{c } \sqrt{p_0}\\ [ 5pt ] \sqrt{1-p_0 } \end{array } \right)\ ] ] yield the pure densities @xmath90
\sqrt{p_1(1-p_1 ) } & 1-p_1 \end{array } \right ) = { |\varphi_1\rangle\langle\varphi_1|}\ ] ] @xmath91 \sqrt{p_0(1-p_0 ) } & 1-p_0 \end{array } \right ) = { |\varphi_0\rangle\langle\varphi_0|}\ ] ] which replace the mixed densities .
theorem [ the : helstrom ] instructs us to define the optimal projectors as those of the spectrum of whose eigenvalues are positive , the spectrum being @xmath92 where the @xmath93 s are eigenvalues , @xmath94 and @xmath95 ( see @xcite ) .
@xmath96 is the distance between densities defined in @xcite ; @xmath96 is the squared cosine of the angle between the subspaces corresponding to the density vectors .
the justification of viewing @xmath96 as a distance comes from the fact that `` the angle in a hilbert space is the only measure between subspaces , up to a costant factor , which is invariant under all unitary transformations , that is , under all possible time evolutions . ''
@xcite @xmath97 are the optimal projectors in the pure case and @xmath76 are the optimal projectors in the mixed case .
the probability of detection ( i.e. , the power ) @xmath98 and the probability of false alarm ( i.e. , the size ) @xmath99 in the pure case are defined as follows : @xmath100 finally , @xmath98 can be defined as function of @xmath99 : @xmath101 1 & { |x|^2 } < q_0 \leq 1 \end{array } \right.\ ] ] so that the power curve is obtained ( see @xcite ) .
expressions and have no counterpart in classical probability and are among the essential points of the paper because they allow us to improve ranking yet using the same amount of evidence as the evidence used in the classical probability distribution and .
at this point , there are three main issues : * the numerical difference between the classical and the quantum probabilities of detection at every given probability of false alarm , * the interpretations of @xmath76 and @xmath97 and whether the interpretations can be tied together , * how the optimal projectors @xmath97 in the pure case can be used for ranking information units in a data management system .
the issues are addressed in section [ sec : optim - rank - quant ] , [ sec : interpr - quant - proj ] and [ sec : impl - rank ] , respectively .
the following lemma shows that the power of the decision rule in quantum probability is greater than , or equal to , the power of the decision rule in classical probability with the same amount of information available from the training set to estimate @xmath102 .
[ sec : optim - rank - quant-2 ] @xmath103 at every given false alarm probability .
the equality holds only if @xmath104 : @xmath105 0 \end{array } \right ) = p_1 = { \mathrm{tr}}(\mu_1\mathbf{p}_1)\ ] ] + @xmath106 0 \end{array } \right ) = p_0 = { \mathrm{tr}}(\mu_0\mathbf{p}_1)\ ] ] let @xmath107 be a certain false alarm probability and let @xmath108 be the real , continuous functions yielding the detection probabilities at @xmath45 .
@xmath98 admits the first and the second derivatives in the range @xmath109 $ ] .
in particular , @xmath110 in @xmath109 $ ] .
@xmath111 is a continuous function .
consider the polynomial @xmath112 of order @xmath2 passing through the points @xmath113 and @xmath87 at which @xmath114 intersects @xmath98 .
then , the lagrange interpolation theorem can be used so that @xmath115 the latter being non negative because @xmath110 and @xmath116 .
the number @xmath117 $ ] exists due to the rolle theorem . as @xmath118 $ ] , hence , @xmath119 $ ] .
similarly , consider the polynomials @xmath120 and @xmath121 of order @xmath2 passing through the points @xmath87 , @xmath88 and @xmath122 , @xmath123 at which @xmath124 and @xmath125 intersect @xmath98 , respectively .
then , the lagrange interpolation theorem can again be used so that @xmath126 @xmath127 then , @xmath128 for all @xmath129 $ ] . the power @xmath98 can plotted against the size @xmath99 , thus producing the power curve of the classical decision rule and the power curve of the quantum decision .
a graphical representation is provided in figure [ fig : roc ] .
is the curve above the polygonal curve depicting @xmath111 .
the classical probability roc curve intercepts the quantum probability roc curve at @xmath87 , @xmath88 and @xmath88 , @xmath123 where @xmath130 are observed .
@xmath131 for @xmath132 and @xmath103 for all @xmath99 s . ]
[ sec : optim - proj - quant ] suppose that ten information units have been used for training a data management system .
each unit has been indexed using one binary feature and has been marked as useful ( 1 ) or useless ( 0 ) .
the training set is summarized by table [ tab : example ] : @xmath133 + therefore , @xmath134 the computation of @xmath135 follows from and the computation of @xmath136 follows from . when @xmath137 , we have that @xmath138 .
in this section , some interpretations of the optimal projectors representing the region of acceptance are provided .
the optimal projectors @xmath97 in the pure case have a more difficult interpretation than @xmath76 because the latter represent `` physical '' observations ( e.g. , a customer review does exist or does not ) whereas @xmath97 can not be expressed in terms of @xmath139 and we can not explain the @xmath140 s by saying that , for example , they represent the presence and/or the absence of a feature . in quantum theory , the impossibility of expressing a projectors as functions of other projectors is termed incompatibility which is expressed mathematically as @xmath141 .
the interpretation of the optimal projectors reflects on the interpretation of what means that they are `` observed '' in an information unit ; for example , if the information unit is a commercial item suggested to a customer , what does the `` observation '' of @xmath142 mean ?
what should we observe from an information unit so that the observation outcome corresponds to the projector ?
the question is not a futile because the answer(s ) would effect the algorithms ( e.g. , automatic indexing ) used for representing the informative content of the unit .
specifically , the interpretation of @xmath143 provides what the retrieval algorithm must do when a feature is observed .
either the interpretation of an optimal projector is implemented at indexing time or at query time , an internal memory representation in terms of data structures is necessary for automatic processing and the representation needs the observation of physical properties which are then converted into numbers . in the mixed case ,
the answer is quite straightforward because the optimal projectors correspond to the feature occurrence and separate the units indexed by the feature from those not indexed . in the pure case , the answer is not straightforward at all . if @xmath76 represent feature occurrence , the @xmath144 s can not be a feature occurrence , but they are something new which can not be described in convential way .
indicate the angles between the vectors . ]
geometrically , each vector is a superposition of other two independent vectors .
figure [ fig : geometry ] depicts the way the vectors and the spanned subspaces ( i.e. , projectors ) interact and shows that @xmath145 are placed symmetrically `` around '' the density vectors and the probability of error is minimized @xcite .
the observation of a binary feature places the observer upon either @xmath146 or @xmath147 and there is no way to move upon @xmath148 or @xmath149 .
probabilistically , the optimal projectors and the density vectors are related as follows : @xmath150 where @xmath151 logically , the projectors are assertions , thus a parallel can be established with assertions and subsets an assertion defines the elements of the universe ( e.g. , an event space ) which belong to a subset .
the basic difference between subspaces and subsets is that the vectors belong to a subspace if and only if they are spanned by a basis of the subspace .
a containment relationship can be established between subspaces such that if a subspace ( e.g. , a line ) includes a point , then every subspace ( e.g. , a plane ) containing the line includes the point too .
the subspace spanned by the projector @xmath24 is termed as @xmath152 and the containment relationship between @xmath152 and @xmath153 can be encoded as @xmath154 such that for every vector @xmath51 , @xmath155 ( * ? ? ?
* ch . 5 ) .
the paper considers the information units relevant , useful or interesting when they are included by the subspace @xmath156 spanned by @xmath157 .
suppose that a subspace @xmath152 is given and that @xmath158 is the metric defined on the whole space to which @xmath152 belongs .
@xmath159 see @xcite . as @xmath160
, @xmath161 maximizes the probability @xmath162 note that when @xmath51 is in @xmath163 , is the distance between the @xmath163 s defined in @xcite .
the result establishes a connection between the geometric , probabilistic and logical interpretation of the projectors even though they seems different . in the next section
, these interpretations are tied together , thus allowing us to look for a criterion for assigning an information unit to the best region as explained .
the problem is to decide whether an information unit represented by a binary feature is considered relevant , useful , interesting , etc .. an algorithm implementing such a decision rule should perform as follows .
it reads the feature occurrence symbol ( i.e. , either @xmath1 or @xmath2 ) ; check whether the feature is included by the region of acceptance . if the feature is not included , the hypothesis of relevance , usefulness , interest , etc . is rejected .
another view of the preceding decision rule is the ranking of the information units .
when ranking information units , the system returns the units whose features lead to the highest probability of detection , then those whose features lead to the second highest probability of detection , and so on . when a binary feature is considered , the ranking ends up to placing the units whose features lead to the highest probability of detection whereas the other units are not retrieved . as we point out in section [ sec : interpr - quant - proj ] , the observation of features corresponding to @xmath130 can not give any information about the observation of the events corresponding to @xmath164 due to the incompatibility between these pairs of events .
thus , we can not design an algorithm implementing the decision rule so that the observation of a feature can be translated into the observation of the events corresponding to @xmath143 .
a possible approach can be based on the probabilistic interpretation of the optimal projectors .
according to such an approach , the probability that the event corresponding to an optimal projector occurs provided that a feature occurs can provide a measure of the degree to which the event had occurred if it could have been observed .
when the subspaces represent these events , the probability that the event corresponding to an optimal projector is observed if the state is described by @xmath60 is @xmath111 .
such an approach is partially satisfactory because @xmath165 . as an alternative approach ,
consider the geometric interpretation depicted in figure [ fig : geometry ] .
note that the asymmetry of @xmath166 with respect to the densities causes the suboptimality of @xmath167 .
indeed , if @xmath166 coincided with @xmath168 , the power and the size would be the same in both cases .
we propose a method which reaches the optimality without neither resorting to probabilistic approximations nor undergoing high computational costs .
the density vectors @xmath169 are a superposition of both @xmath166 and @xmath168 which are two different bases and induce different coordinates .
when @xmath166 is the basis , the coordinates of @xmath170 are @xmath171 . when @xmath168 is the basis , the coordinates of @xmath170 are @xmath172 such that @xmath173 and @xmath174 as @xmath175 is often assumed when ranking information units , the coordinates have a quite simple and intuitive meaning provided by the following expressions : @xmath176 where @xmath177 . as the asymmetry of @xmath166 with respect
the density vectors is due to @xmath178 which summarize the statistics observed from the training set , we leverage them for improving the ranking . in particular , in the paper
we show that changing the estimation of @xmath84 is sufficient to reach the optimality .
then , we wonder how we should define the density vectors or matrices so that @xmath179 were obtained instead of @xmath135 .
the basis vectors ( i.e. , @xmath166 or @xmath168 ) are rotated , thus changing the coordinates .
therefore , we define the density vectors @xmath180 in @xmath181 according to in such a way that if a feature is observed under @xmath62 , the probability of detection is @xmath98 and , if a feature is observed under @xmath63 , the probability of false alarm is @xmath99 .
the simple solution is defining the new density vectors as follows : @xmath182 thus obtaining @xmath183 at first sight , the increase of the probability of detection is due to the higher probability values assigned to the region of acceptance in the pure case than those assigned to the region of acceptance in the mixed case , and not to a different ranking . in the following , we show that the superiority of the discriminant function in the pure case is due to the different ranking induced by a different partition of the event space into region of acceptance and region of rejection .
we state the problem as follows .
are there @xmath184
such that the region of acceptance in the pure case differs from that in the mixed case ?
consider theorem [ the : helstrom ] to answer the question .
the region of acceptance in the mixed case is defined through table [ tab : region - of - acceptance - mixed ] whereas the region of acceptance in the pure case is defined through table [ tab : region - of - acceptance - pure ] .
furthermore , the discriminant function derived from is @xmath185 where @xmath186 .the regions of acceptance corresponding to the sign of the eigenvalues of the spectrum of the discriminant function in the mixed case .
the equality case is addressed in @xcite . [ cols="<,^,^ " , ] suppose that @xmath187 .
thus , @xmath188 and the region of acceptance in the mixed case is represented by @xmath189 , whereas the region of acceptance in the pure case is represented by @xmath74 .
the counter - example just mentioned proves the following [ sec : impl - optim - rank ] the discriminant function ranks information units in a different way from the discriminant function because an alternative ranking is computed . in the densities that are considered
are those associated to a mixed state , while in the densities are the one associated to a pure state .
so the equations look like the same , what differs is the type of densities that are used in the two cases .
we have shown that the improvement of ranking measured in terms of probability of detection given a probability of false alarm is due to the ranking induced by @xmath190 . hence ,
we state the problem of finding @xmath97 into the problem of defining the coordinates of the representation of the density vectors in @xmath181 .
the problem of defining the new coordinates for the density vectors might be viewed as a problem of feature weighting .
in such a context , the traditional estimations of the probability of feature occurrence ( i.e. , @xmath84 ) under two different hypothesis are replaced by the @xmath191 s .
feature re - weighting is explored in ir whose state - of - the - art is given by the bm25 weighting scheme surveyed , for example , in .
the main drawback of the weighting schemes like bm25 is the parameter tuning necessary for optimizing the effectiveness with a given database or query , thus making the understanding of how and why a scheme is more effective than others rather problematic .
in contrast , the paper illustrate the decision rule in such a way that if the decision rule is followed , then @xmath62 shall be accepted when it is true at a higher probability of detection than @xmath111 when the probability of false alarm is not more than a given threshold .
the formulation of the decision rule provided in this section allows us to design an efficient algorithm for indexing and retrieving ( or classifying ) information units .
the algorithm is just an instance of those employed currently in ir ( see for example ) where the maximum likelihood or bayesian estimations of @xmath102 are replaced by and .
the foundations of quantum mechanics and theory has been illustrated in plenty of books such as @xcite and @xcite .
quantum probability , for example , has been introduced in @xcite . in particular ,
the interference term is addressed in @xcite .
the view of probability illustrated in section [ sec : class - prob - quant ] is based on @xcite .
the utilization of quantum theory in computation , information processing and communication is described in .
recently , investigations have started in other research areas , for example , in ir .
the paper is inspired by helstrom s book @xcite which provides the foundations and the main results in quantum detection ; an example of the exploitation of the results in quantum detection is reported within communication theory .
this paper links to @xcite as far it concerns density matrices and projectors ; however , the paper develops quantum detection for data management .
this paper departs from the probability ranking principle ( pr ) proposed in the context of classical probability ; we propose quantum probability to improve ranking in a principled way . in information retrieval ,
the probability ranking principle ( prp ) states that `` if a reference retrieval system s response to each request is a ranking of the documents in the collection in order of decreasing probability of relevance to the user who submitted the request , where the probabilities are estimated as accurately as possible on the basis of whatever data have been made available to the system for this purpose , the overall effectiveness of the system to its user will be the best that is obtainable on the basis of those data . ''
@xcite . however , some assumptions undermine the general applicability of the prp .
we state a similar principle yet replace classical probability , which is implied in @xcite , with quantum probability
parameter estimation data are kept the same , bu we instead use subspaces to define alternative regions of acceptance and rejection . to our knowledge
, the use of quantum probability for ranking information units has not yet been addressed in the same way of this paper although a few papers that are somehow comparable can be found .
perhaps , the closest paper is . that paper proposes to rank documents by quantum probability and suggests that interference ( which must be estimated ) might model dependencies in relevance judgements such that documents ranked until position @xmath192 interfere with the degree of relevance of the document ranked at position @xmath193
this means , the optimal order of documents under the prp differs from that of the quantum prp .
note that they empirically show that quantum probability is more effective that classical one in specific rankings tasks .
in contrast , in this paper , we do not need to address interference because quantum probability can be estimated using the same data used to estimate classical probability .
we rather show that not only ranking by quantum probability provides a different optimal ranking , it is also more effective than classical probability . with this regard , the effectiveness of quantum probability measured in stems from the estimation of classical probability and that of interference .
but , the regions of acceptance and rejection are still based on sets .
it follows that the optimality of the quantum prp strongly depends on the optimality of the prp and on the interference estimantion effectiveness . in this paper , on the contrary , ranking optimality only depends on the region of acceptance defined upon subspaces .
another paper somewhat related to ours is .
the authors discuss how to emply quantum formalisms for encompassing various information retrieval tasks within a single framework . from an experimental point of view
, what that paper demonstrates is that ranking functions based on quantum formalism are computationally feasible .
the best experimental results of rankings driven by quantum formalism are comparable to bm25 , that is , to prp , thus limiting the contribution within a classical probability framework .
probabilistic databases systems manage imprecise data and provide tools for structured complex queries .
a survey is provided in . beside scalability and query plan execution ,
these systems do probabilistic inference which may be defined upon classical or quantum probability and they concentrate on top-_k _ query answering where the tuples are assigned a probability distribution . the results of this paper may be applied to probabilistic databases systems too .
the main result of the paper is the demonstration that quantum probability can be incorporated into a data management system for ranking information units .
as ranking by quantum probability is more effective than ranking by classical probability when it has been used in other domains , it is our belief that an analogous improvement can be achieved within data management .
the future developments are threefold .
first , we will work on the intepretation of the optimal projectors in the pure case because the detection of them in an information unit may open further insights .
second , feature classical correlation and quantum entanglement will be investigated .
third , evaluation is crucial to understand whether the results of the paper can be confirmed by the experiments . 10 l. accardi . on the probabilistic roots of the quantum mechanical paradoxes . in s.
diner and l. de broglie , editors , _ the wave - particle dualism _ , pages 297330 .
d. reidel pub .
co. , 1984 .
l. accardi . .
il saggiatore , 1997 . in italian .
l. accardi and a. fedullo . on the statistical meaning of complex numbers in quantum mechanics .
, 34(7):161172 , june 1982 .
g. boole . .
walton and maberly , 1854 .
p. bruza , d. sofge , w.f .
lawless , c.j .
van rijsbergen , and m. klusch , editors . ,
volume 5494 of _ lecture notes in computer science _ ,
saarbrcken , germany , 2009 .
springer .
g. cariolaro and g. pierobon .
performance of quantum data transmission systems in the presence of thermal noise .
, 58:623630 , february 2010 .
croft , d. metzler , and t. strohman . .
addison wesley , 2009 .
n. dalvi , c. r , and d. suciu .
probabilistic databases : diamonds in the dirt . ,
52:8694 , july 2009 .
r. b. griffiths . .
cambridge university press , 2002 .
undergraduate texts in mathematics .
springer , 1987 .
helstrom . .
academic press , 1976 .
harvard university press , 1989 .
kolmogorov . .
chelsea publishing company , new york , second edition , 1956 .
m. melucci . a basis for information retrieval in context .
, 26(3 ) , 2008 .
j. neyman and e.s .
pearson . on the problem of the most efficient tests of statistical hypotheses .
, 231:289337 , 1933 .
m. nielsen and i.l .
cambridge university press , 2000 .
parthasarathy . .
birkhuser , 1992 .
b. piwowarski , i. frommholz , m. lalmas , and k. van rijsbergen .
what can quantum theory bring to information retrieval ? in _ proc .
19th international conference on information and knowledge management _
, page 5968 , 2010 .
e. rieffel .
certainty and uncertainty in quantum information processing . in _ proceedings of the quantum interaction symposium _ , 2007 .
the probability ranking principle in information retrieval .
, 33(4):294304 , 1977 .
robertson and h. zaragoza .
the probabilistic relevance framework : bm25 and beyond .
, 3(4):333389 , 2009 .
keith van rijsbergen . .
cambridge university press , uk , 2004 .
j. von neumann . .
princeton university press , 1955 .
w. k. wootters . statistical distance and hilbert space . , 23(2):357362 , jan 1981 .
g. zuccon , l. azzopardi , and c.j .
van rijsbergen .
the quantum probability ranking principle for information retrieval . in _ proceedings of the international conference on the theory of information retrieval ( ictir ) _ , pages 232240 , 2009 .
a complex vector @xmath45 is represented as @xmath51 and is called `` ket '' .
the conjugate transpose of @xmath45 is represented as @xmath194 and is called `` bra '' ( therefore , the dirac notation is called the bra(c)ket notation ) .
moreover , @xmath202 and the properties of trace allow us to write @xmath203 . in general ,
if @xmath204 are trace-1 hermitian operators , and @xmath205 is a projector , @xmath206 is the probability that the event represented by @xmath205 occurs given a density operator @xmath24 . | data management systems , like database , information extraction , information retrieval or learning systems , store , organize , index , retrieve and rank information units , such as tuples , objects , documents , items to match a pattern ( e.g. classes and profiles ) or meet a requirement ( e.g. , relevance , usefulness and utility ) . to this end , these systems rank information units by probability to decide whether an information unit matches a pattern or meets a requirement .
classical probability theory represents events as sets and probability as set measures .
thus , distributive and total probability laws are admitted .
quantum probability is a non - classical theory nor does admit distributive and total probability laws .
although ranking by probability is far from being perfect , it is optimal thanks to statistical decision theory and parameter tuning .
the main question asked in the paper is whether further improvement over the optimality provided by probability may be obtained if the classical probability theory is replaced by quantum probability theory . whereas classical probability ( and detection theory ) is based on sets such that the regions of acceptance / rejection are set - based detectors , quantum probability is based on subspace - based detectors .
the paper shows that ranking information units by quantum probability differs from ranking them by classical probability provided the same data used for parameter estimation . as probability of detection (
also known as recall or power ) and probability of false alarm ( also known as fallout or size ) measure the quality of ranking , we point out and show that ranking by quantum probability yields higher probability of detection than ranking by classical probability provided a given probability of false alarm and the same parameter estimation data . as quantum probability
provided more effective detectors than classical probability within other domains that data management , we conjencture that , the system that can implement subspace - based detectors shall be more effective than a system which implements a set - based detectors , the effectiveness being calculated as expected recall estimated over the probability of detection and expected fallout estimated over the probability of false alarm . [ 1]*proof of # 1 * |
there is a sense of youthfulness to saturn s system , as if the system were now being caught in the process of evolving to something quieter .
signs of youth abound , each of which has something to tell us about what happened at saturn : ( i ) saturn s rings appear both by color and dynamics to be tens or hundreds of millions of years old , rather than the billions of years that describe the solar system .
( ii ) enceladus eruptions are too energetic to have been going on as they are now for more than a small fraction of the age of the solar system .
( iii ) the methane in titan s atmosphere will be chemically destroyed on a time scale of order 10@xmath4 yrs , with the hydrogen escaping to space ; there is no known source of new methane , nor is there any sign on the surface of the hundreds of meters of photochemical product which would have accumulated had methane been supplied to the atmosphere over the age of the solar system .
( iv ) saturn s satellite system is crowded and has at least three examples of coorbital moons , a phenomenon not seen elsewhere in the solar system .
one set of coorbital satellites consists of the two small moons janus and epimetheus ( @xmath5 km and @xmath6 km in mean radius , respectively ) , traveling on mutual `` horseshoe '' orbits .
the other two sets each consist of a pair of small trojan companions of larger moons : tethys is accompanied by telesto ( smith et al . 1980 ) and calypso ( pascu et al .
1980 ; seidelmann et al . 1981 ) , while dione is accompanied by helene ( lecacheux et al . 1980 ) and polydeuces ( porco et al . 2007 ) .
telesto , calypso , and helene are all 11 - 17 km in radius , and were discovered by ground - based observations , while tiny polydeuces is only 1.3 km in radius and was found with the cassini orbiter ( porco et al .
from other spectacular cassini images , telesto and calypso appear to be covered smoothly with fine material .
all of these trojan moons are intriguing because tethys and dione are adjacent ; furthermore , tethys is in a 4:2 mixed - inclination - type mean - motion resonance with mimas , while dione , helene , and polydeuces are all in a 2:1 inner - eccentricity - type mean - motion resonance with enceladus .
the overall impression is paradoxical .
small number statistics precludes surety , but it seems unlikely that chance alone can explain why dione and tethys each have two trojans , while all other major satellites have no such attendants . rather , there is a sense of underlying order , as if all of this came about in a perfectly ordinary way .
the better to characterize this system , in this study we simulate the orbital evolution of test particles launched from tethys coorbital satellites telesto and calypso .
first we describe the methodology of the numerical integrations and how we prepare the initial conditions .
next we discuss evolution and fates of ejecta from telesto and calypso .
the results turn out to be rather interesting : both the speed and direction of ejecta launch are important , with well - defined regimes which correspond to tadpoles , horseshoes , and passing orbits .
the key points are best illustrated by strongly reductionist numerical experiments in which ejecta are launched vertically into space at various places along a satellite s equator .
we then broaden our compass to consider more realistic distributions of ejecta launched from impact craters .
our results imply that telesto and calypso readily exchange ejecta .
we use the swift integrator package ( levison and duncan , 1994 ) to evolve the saturn system and the orbits of the ejecta forward in time .
swift , which is based on work by wisdom and holman ( 1991 ) , is able to propagate the orbits of the massless test particles as well as the massive bodies using a choice of integrators . for this study
we use its regularized , mixed variable symplectic integrator ( rmvs3 ) ; one of its advantages is the ability to use relatively large time - steps .
another advantage of the rmvs3 method is that it can handle very close approaches between a test particle and a massive body , which are of great interest for this study .
we use a customized version of swift , which we have modified to detect collisions with saturn s rings and to include saturn s optical flattening , along with its dynamical oblateness up to order @xmath7 ( dobrovolskis and lissauer , 2004 ) .
all other massive bodies are treated as spherical ; although the trojan moons and hyperion are some tens of percent out of round , so that their surface gravity and escape speed should vary by a comparable amount over their surfaces , this variation is minor compared with the speeds of their impactors and ejecta . in our ejecta integrations
we include the gravitational attractions of saturn , mimas , methone , enceladus , tethys and its coorbitals calypso and telesto , dione and its coorbitals helene and polydeuces , rhea , titan , hyperion , iapetus and phoebe .
the initial state vectors and gravitational constants are from r. jacobson ( sat263 model , pers .
communication ) in the form of barycentric initial locations and velocities of the saturn system for epoch 26-jul-2006 , 00:00:00.0 , expressed in the icrf system ( arias et al .
in addition , we modified swift to include the gravitational perturbations of the sun by treating it as a special massive , distant `` satellite '' of saturn ; we also added the mass of jupiter to that of the sun ( danby , 1962 ) .
initial conditions for the sun and jupiter were obtained from the de200 jpl analytical ephemeris .
finally , a rotation and a translation were performed to align the fundamental plane with saturn s equator and to set the origin at the center of saturn for our swift integrations .
the initial state vector for polydeuces is from jacobson et al .
( 2008 ) ; however , its j2000 barycentric initial conditions are for the epoch 2-jan-2004 , 00:00:00.0 .
hence , the previous initial conditions for the saturn system ( minus polydeuces ) were integrated backwards in time using swift until 2-jan-2004 , at which point we `` drop '' the initial conditions for polydeuces ( rotated and translated as well , of course ) into our final , `` augmented '' saturn system to be used in the integrations .
table 1 gives an overview of the saturn satellite system used in our integrations .
radii and density ( @xmath8 = 0.6 g/@xmath9 ) of the coorbitals are from porco et al .
( 2007 ) .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ table 1 .
overview of the saturn system used in this paper .
the values of @xmath10 and @xmath11 are from r. jacobson ( sat263 model ) , except for methone and the coorbitals , for which the values of @xmath10 were computed assuming a density of 0.6 g/@xmath9 ( porco et al .
orbital elements are mean values from murray and dermott ( 1999 ) , except for methone and polydeuces , which are from jacobson et al .
inclinations are given relative to the local laplace plane .
the semi - major axis values are shown in kilometers as well as in terms of saturn s radius @xmath12 . for saturn
, we use @xmath13 km@xmath14/s@xmath15 ; @xmath12 = 60,330 km ; @xmath16 ; @xmath17 ; see dobrovolskis and lissauer ( 2004 ) for @xmath18 , @xmath19 , @xmath20 and @xmath7 . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ [ cols="<,^,^,^,^,^,^,^ " , ] @xmath21 antapex initial conditions generated assuming the spallation model ( melosh 1989 ) as modified by zahnle et al .
in this paper we have investigated the fates of ejecta from the trojan companions of tethys , namely telesto and calypso . in the first numerical experiment
, we systematically investigated the fates of ejecta launched from their equatorial regions .
the initial conditions are translated and rotated from a topocentric coordinate system ( with origin at the trojan moon s surface whence the ejecta are launched ) to a saturnicentric system and fed to the swift numerical integrator ( levison and duncan , 1994 ) to propagate the particle trajectories forward in time for 100,000 years . as expected , the lowest velocity ejecta can not fully escape the source moon s gravity and fall back onto its surface in less than one day , following suborbital trajectories .
ejecta do not need to achieve full escape speed to go into orbit about saturn , though ; the launch speed @xmath22 only needs to be greater than @xmath23 ( where @xmath24 is the classical escape speed , and @xmath25 ; see eq .
1 ) . particles with @xmath26 or slightly higher barely escape the trojan moon s hill sphere and are confined to tadpole orbits librating about tethys @xmath0 point ( for ejecta from telesto , leading ) or @xmath1 point ( for ejecta from calypso , trailing ) .
these particles can not reach tethys , and instead are eventually re - accreted by telesto or calypso respectively .
an interesting finding is that the removal times of ejecta which re - impact telesto or calypso within the first dozen years or so are discretized to integer multiples of half the tadpole period .
somewhat faster ejecta with azimuthal velocities @xmath2 10 m / s go into horseshoe orbits , which enclose not only the @xmath0 and @xmath1 points but also @xmath3 ; these particles still can not reach tethys or its hill sphere and are again accreted by either of the trojan moons .
ejecta from telesto or calypso reaching tangential velocities @xmath2 40 m / s go into `` passing '' orbits ; beyond this point tethys accretes most particles , but some are strongly scattered inward or outwards towards other saturnian moons in adjacent orbits such as enceladus or dione ( on timescales from a few to several hundred years ) .
it is interesting to note that the timescale for removal of an ejectum in a horseshoe orbit is usually orders of magnitude longer than for ejecta in either tadpole or passing orbits . in order to investigate a more realistic distribution of ejecta
, we assumed an impact producing a 1-km - radius crater at five `` cardinal '' points of each trojan moon ( apex , subsaturn , antisaturn , north pole , and south pole ) .
the ejection model adopted here is that of housen et al .
( 1983 ) , consistent with soft regolith : the range of ejecta speeds is roughly 5 m / s to 30 m / s .
as before , initial conditions are integrated for 100,000 years . as per our earlier findings
, none of these particles can go into passing orbits ; indeed , all of these go into tadpole or horseshoe orbits . if the former , all particles come back to the source trojan ; if the latter , particles can hit either coorbital .
however , we noted that most ( between 89% and 97% ) of particles come back to the source trojan moon .
most of the rest go into the other trojan moon , with a tiny minority ( @xmath27 1% ) surviving the 100,000 year integrations .
ejecta from the antapex of motion was launched according to a `` spallation '' model , with higher speeds of 5 m / s to 56 m / s , so a few percent reached tethys .
the ease with which telesto and calypso exchange ejecta suggests that this recycling may help to maintain their similar sizes in the face of impact erosion .
we intend to test this hypothesis in future work on the lopsided helene - dione - polydeuces system . a recent preprint ( izidoro _ et al .
_ 2010 ) finds that a swarm of planetesimals tadpoling with a proto - satellite generally results in the accretion of 14 coorbital companions ; this appears to be consistent with our conclusions as well .
we would like to acknowledge patrick hamill for his contributions , and for previewing the manuscript , and to thank r. a. jacobson for providing the initial conditions for the saturn system .
j. l. a. also would like to acknowledge the patience and encouragement of alejandra , joselito , isabella and danielito .
this work has made use of nasa s astrophysics data system ( ads located at http://adswww.harvard.edu ) , and was supported by nasa s planetary geology & geophysics program through wbs 811073.02.01.03.89 .
alvarellos , j. l. , k. j. zahnle , a. r. dobrovolskis , and p. hamill 2002 .
orbital evolution of impact ejecta from ganymede .
icarus * 160 * , 108 - 123 . seidelmann p. k. , harrington , r. s. , pascu d. , baum w. a. , curie , d. g. , westphal , j. a. and g. e. danielson 1981 .
saturn satellite observations and orbits from the 1980 ring plane crossings .
icarus * 47 * 282 - 287 . | we have numerically integrated the orbits of ejecta from telesto and calypso , the two small trojan companions of saturn s major satellite tethys .
ejecta were launched with speeds comparable to or exceeding their parent s escape velocity , consistent with impacts into regolith surfaces .
we find that the fates of ejecta fall into several distinct categories , depending on both the speed and direction of launch . the slowest ejecta follow sub - orbital trajectories and re - impact their source moon in less than one day .
slightly faster debris barely escape their parent s hill sphere and are confined to tadpole orbits , librating about tethys triangular lagrange points @xmath0 ( leading , near telesto ) or @xmath1 ( trailing , near calypso ) with nearly the same orbital semi - major axis as tethys , telesto , and calypso .
these ejecta too eventually re - impact their source moon , but with a median lifetime of a few dozen years .
those which re - impact within the first ten years or so have lifetimes near integer multiples of 348.6 days ( half the tadpole period ) .
still faster debris with azimuthal velocity components @xmath2 10 m / s enter horseshoe orbits which enclose both @xmath0 and @xmath1 as well as @xmath3 , but which avoid tethys and its hill sphere . these ejecta impact either telesto or calypso at comparable rates , with median lifetimes of several thousand years .
however , they can not reach tethys itself ; only the fastest ejecta , with azimuthal velocites @xmath2 40 m / s , achieve `` passing orbits '' which are able to encounter tethys .
tethys accretes
most of these ejecta within several years , but some 1% of them are scattered either inward to hit enceladus or outward to strike dione , over timescales on the order of a few hundred years . *
28 pages * 5 figures * 5 tables _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ * prepared for icarus + * submitted 5 jan 2010 + * revised 2010 april 27 + * * * * proposed running head : exchange of ejecta between telesto and calypso * * editorial correspondance , proofs and reprints : + anthony r. dobrovolskis + 245 - 3 nasa ames research center + moffett field , ca 94035 - 1000 + ( 650 ) 604 - 4194 + [email protected] * _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ * keywords : * cratering , satellite dynamics , saturn satellites |
determining when neutral hydrogen in the intergalactic medium ( igm ) was reionized is an important question in observational cosmology and a precursor to understanding whether star forming galaxies provided the necessary uv photons @xcite .
one of the most practical probes of reionization with current facilities utilizes the frequency of occurrence of ly@xmath0 emission in star forming galaxies . as ly@xmath0 photons
are resonantly scattered by neutral hydrogen , the abundance of ly@xmath0 emitters should decrease as observations probe into the era where there is neutral gas ( e.g. , @xcite , @xcite , @xcite ) .
the recent discovery of large numbers of candidate galaxies beyond @xmath6 through multi - color imaging undertaken with the infrared wide field camera ( wfc3/ir ) onboard hubble space telescope ( hst ) ( e.g. , @xcite ) now makes it feasible to track the occurrence of ly@xmath0 line emission to interesting redshifts where neutral hydrogen may be present . of course ,
astrophysical factors other than a neutral igm can also affect the presence of ly@xmath0 emission . because of this , an alternative approach for gauging when reionization occurred , introduced by @xcite ( hereafter paper i ) , is to spectroscopically measure _ the fraction of ly@xmath0 emission within color - selected lyman break galaxy ( lbg ) populations_. tracking the redshift - dependent fraction in a well - defined population avoids consideration of absolute changes in the host galaxy number density such as has been the case in studies based on the luminosity function of narrow - band selected lyman @xmath0 emitters ( laes ) .
furthermore , evolution in dust obscuration can be independently tracked via correlations seen with the colors of the rest - frame uv continuum .
paper i presented a comprehensive survey of over 600 lbgs with deep spectra , mostly undertaken with the deimos instrument on keck , but including published samples from the very large telescope ( vlt , @xcite and references therein ) . in that paper
we demonstrated the utility of the method and discussed the paucity of line emission in gravitationally - lensed @xmath7 candidates from the sample of @xcite . in @xcite ( hereafter paper ii )
, through ultra - deep exposures with deimos we significantly improved the line emission statistics at @xmath8 , providing a robust measure of the rest - frame ew distribution of ly@xmath0 emission at the highest redshift when the universe is believed to be fully ionized ( @xcite , c.f .
this provides a sound basis for predicting the likelihood of emission at higher redshift and thereby enabling a test of whether there is absorption by neutral gas .
significantly , we found that over 50% of moderately - faint ( @xmath9 ) @xmath8 lbgs exhibit strong emission with rest frame ews @xmath10
. as this fraction increases over @xmath2 , we argued on continuity grounds that we should expect a high success rate in recovering line emission from the newly - found wfc3/ir samples of @xmath7 lbgs _ unless _ we encounter a more neutral igm in the short time interval prior to @xmath11 .
some evidence for this is seen in the recent studies of @xcite and @xcite .
the present paper is concerned with an initial application of this test to the newly - available sample of wfc3/ir candidates with photometric redshifts in the redshift range @xmath3 .
two important factors have motivated and shaped our program .
firstly , it is important to note that ly@xmath0 emission is the only spectroscopic redshift indicator for galaxies beyond @xmath8 . since it is the absence of strong ly@xmath0 emission that provides the basis for considering an increased neutral fraction , it is important to be sure that the targets are truly at the expected redshifts .
many early candidate lbgs believed to lie beyond @xmath12 remained controversial because of their limited or marginal photometry .
the improved filter set and superior performance of wfc3/ir has given us confidence that the current list of @xmath13 candidates is more robust than those based on earlier nicmos data @xcite .
secondly , to match the lower redshift data , sampling from a similarly wide range of lbg luminosities , as we do here , will be advantageous .
as shown in papers i and ii , the fraction of line emission increases in intrinsically fainter sources and so by comparing fractions with respect to their lbg luminosities , we may gain additional evidence for the onset of the neutral era . throughout the paper , we adopt a @xmath14-dominated , flat universe with @xmath15 , @xmath16 and @xmath17 . all magnitudes in this paper are quoted in the ab system @xcite .
in compiling a target list for this program , we are guided by the need for a robust photometric redshift for each galaxy based on improved photometry from wfc3/ir and a range of rest - frame uv luminosities ( @xmath18 ( paper i ) ] ) .
our primary source of targets for the wide - field multi - slit capabilities of the low resolution imaging spectrometer ( lris ) on the keck i telescope @xcite equipped with a new red - sensitive ccd was @xmath19 and @xmath20-drop candidates whose photometric redshifts @xmath21 from the hst early release science ( ers ) field @xcite .
the grating for these observations was blazed at 600 lines mm@xmath22 . on january 7 and february 4 2011
we secured 7 hours of on - source integration for 8 suitable targets on a single mask using slit widths of 1@xmath23 , observed through a median seeing of 0.98@xmath23 .
in a more ambitious campaign probing to higher redshift we also targeted 3 @xmath20-drop sources from the hubble ultradeep field ( hudf ) p34 field ( go 11563 , pi : illingworth ) and an additional gravitationally - lensed source in the cluster ms0451 - 03 ( go 11591 , pi : kneib ) using the near - infrared spectrograph nirspec @xcite during november 14 - 17 2010 and jan 14 - 15 2011 .
this extends our search for ly@xmath0 emission up to a redshift @xmath24 . although we undertook extended integrations on all 4 sources with a 0.76@xmath23 slit , tracking difficulties affected some exposures . to determine the effective on - source integration time , we secured our astrometric position for each exposure by locating objects visible in the slit viewing camera to a precision of @xmath25 0.2@xmath23 .
we continued this campaign over may 15 - 18 2011 . during these 4 nights
, we did not encounter any tracking difficulties and , in excellent conditions , successfully used nirspec to study an additional 7 wfc3-ir dropouts drawn from numerous surveys : the borg pure parallel survey @xcite ( also independently discovered by @xcite ) , the egs region in candels ( @xcite ; @xcite ) , and the lensing clusters abell 1703 ( go 10325 , pi : ford , @xcite ) and abell 2261 ( clash survey , @xcite ) .
we reduce the lris data following standard procedures , with bias subtraction and flat - fielding using dome exposures .
we used the @xcite code to remove spatial and spectral distortion and to model and subtract the sky emission .
wavelength calibration was determined directly from sky lines .
a final two - dimensional spectrum was extracted for each object with pixels binned logarithmically by @xmath26log@xmath27 . as in papers i and ii , we search through the two - dimensional spectrum visually to identify emission lines , and confirm these with a boxcar extracted one - dimensional spectrum . our exposures with nirspec were conducted with typical spatial dithering of 5@xmath23 . in the case of some lensed sources
, we dithered by longer amounts to ensure the arc was oriented along the slit , and was not spatially coincident with any background objects .
we flat - fielded and sky - subtracted the spectra using idl routines written by g. becker ( 2010 , private communication ) . to compute the camera distortion and spectral curvature
, we fit traces of standard stars along the slit , and skylines perpendicular to the slit . from this , we then derive a wavelength and sky position for all pixels in each two - dimensional spectrum , which were used to align the individual exposures .
a final spectrum was created for each object by median stacking all exposures to eliminate signals from cosmic ray strikes . [ cols="<,^,^,^,^,^,^,^,^,^,^",options="header " , ] [ params ] in total
, this paper therefore presents the results of keck spectroscopy for 19 wfc3-ir selected sources whose photometric redshifts lie in the range @xmath3 .
a summary of the new observations is given in table 1 .
to this sample , we add a further 7 @xmath28 sources discussed by @xcite .
figure 1 compares the uv absolute magnitude distribution of the combined sample with that presented for the redshift range @xmath29 in paper ii ; clearly the samples span a similar luminosity range .
this luminosity range is broader than the recent work of @xcite and @xcite . in similar spectroscopic follow - up campaigns , they target brighter dropouts , primarily with m@xmath30 , and @xmath31m@xmath32 , respectively . , for @xmath19-drop sources discussed in paper ii with z @xmath25 5.5 - 6.3 ( top panel ) compared with those for the present survey of @xmath19 and @xmath20-drops at z @xmath33 6.3 ( lower panel ) .
there is an additional dropout , a2261_1 , not shown on this histogram at m@xmath34 .
dark shading in the lower panel refers to sources selected on the basis of wfc3/ir imaging in the keck campaign , ( table 1 ) ; light shading refers to additional data drawn from the vlt campaign of @xcite.,scaledwidth=50.0% ] remarkably , from the new keck exposures , we find very few convincing detections of line emission .
figure 2 ( bottom panel ) shows a 2-d spectroscopic montage of 4 sources for which line emission may be present , of which one case ( hudf09_1596 ) is marginal ( 2@xmath4 ) and the other ( ers 8290 ) lies outside the expected redshift range if it is ly@xmath0 .
the emission line apparently seen in ers 8290 ( a @xmath20-band dropout ) is detected at @xmath355@xmath4 with a flux of 6.7 @xmath36 0.8 @xmath37 @xmath38 erg @xmath39 s@xmath22 at @xmath407644 @xmath36 2 .
it also exhibits the asymmetric profile characteristic of ly@xmath0 in the 1d extraction , but this would place the object at @xmath41 , quite discrepant from our photometric estimate of @xmath42 = 6.52 .
however , upon examining the positioning of our lris slits more carefully ( bottom panel of figure 2 ) , we find there is a faint @xmath43-drop candidate with @xmath19 = 27.5 and @xmath20 = 27.9 only @xmath44 away which would have been at least partially visible through our slit during the exposure . after subtracting the line flux from the @xmath19-band photometry
, we determine a photometric redshift of @xmath42 = 4.91 , in reasonable agreement with the keck spectroscopy , particularly given a greater line flux ( as is likely given the object s poor positioning in the slit ) would increase the photometric redshift estimate .
the resolution of the confusion arising from these two proximate sources emphasizes again the prominence of ly@xmath0 emission in low luminosity @xmath29 sources ( papers i and ii ) .
the two satisfactory detections refer to emission at @xmath409045 @xmath45 for ers 8496 in the lris mask and emission at @xmath409780 @xmath46 in the nirspec exposure of the lensed source abell 1703_zd6 .
both objects are detected at @xmath33 5@xmath4 in our final exposures . in our 1d extraction of ers 8496
, the emission line has a fwhm of 9 @xmath36 1 , and displays an asymmetric profile with a steeply rising blue edge and slowly decaying red tail , characteristic features of ly@xmath0 at high redshift . because our spectral resolution is significantly lower ( 6.5 for nirspec versus 4.1 for lris ) in our spectrum of a1703_zd6 , so we are unable to determine any line profile information .
the emission feature is seen independently in coadditions on two separate nights , indicating its reality . if both lines are ly@xmath0 , the implied spectroscopic redshifts for ers 8496 and abell 1703_zd6 are @xmath47 and @xmath48 , respectively , in excellent agreement with our photometric predictions of 6.52 and that of 7.0 derived by @xcite .
the measured line fluxes for the two objects are 9.1 @xmath36 1.4 and 28.4 @xmath36 5.3 @xmath37 @xmath38 erg @xmath39 s@xmath22 .
we then assume a spectral slope of @xmath49 = -2 , which is characteristic of galaxies at this redshift @xcite , though there may be evidence for steeper slopes at sub-@xmath50 luminosities @xcite . under this assumption , taking the magnitude from the first filter in which the object is detected ( @xmath51 for ers 8496 , and @xmath52 for 1703_zd6 ) , we calculate ews of 69 @xmath36 10 and 65 @xmath36 12 , respectively . because our objects have additional coverage longward of the detection filter
, we can also compute a value for @xmath49 , and extrapolate to find the continuum flux at @xmath53 . using the formulae of @xcite
, we find @xmath49 = -2.39 @xmath36 0.55 , and -2.44 @xmath36 0.64 . when computing ews using this method , we obtain 67 @xmath36 11 and 59 @xmath36 12 , respectively .
emission detected from 4 sources in our keck survey , along with boxcar - extracted 1d spectrum .
wavelength ranges contaminated by strong skylines are shaded in grey in the 1d extraction .
the top row shows the two robust detections of ers 8496 and a1703_zd6 at @xmath54 and @xmath42=7.045 respectively .
the bottom row shows a marginal detection at @xmath42=6.905 for hudf09_1596 and a likely @xmath55 line at @xmath41 arising from a serendipitous @xmath43-drop close to ers 8290 as illustrated in the bottom left slit image .
the bottom right panel shows the photometric redshift distribution for this serendipitous @xmath43-drop , with a vertical line indicating the observed spectroscopic redshift ( see text for further details).,scaledwidth=50.0% ]
our approach in this paper is to compare the rate of occurrence of ly@xmath0 in our new @xmath3 sample with that expected from our reference sample of @xmath19-drops with @xmath56 drawn from paper ii ( figure 1 ) . in both papers i and ii we showed that the rest - frame ew distribution is a function of rest - frame uv absolute magnitude , @xmath18 , and
thus we additionally take this luminosity dependence into account .
we estimate the luminosities of all our sources in table 1 from their photometric redshift and incorporate the lensing magnification @xmath57 for our lensed sources from @xcite for abell 1703 and from the mass model of @xcite for ms0451 - 03 and abell 2261 . for our baseline ly@xmath0 ew distribution , we use the data from of paper ii at @xmath58 , separated into high and low luminosity regimes . the fraction of emitters within each bin of ew @xmath35 25 a is taken directly from fig . 2 of paper ii .
we set the slope of the distribution within an ew bin equal to the slope between the two lowest bins in paper ii , 25 @xmath59 ew @xmath60 55 , and 55 @xmath59 ew @xmath61 85 .
this slope is equal to @xmath62ew@xmath63ew = @xmath64 for the lower luminosity sample ( @xmath65m@xmath66 ) , and @xmath67 for the higher luminosity sample ( @xmath31m@xmath68 ) . to create the probability distribution for galaxies with ew less than 25
, we extrapolate to ew = 0 using this slope , and assign the remaining fraction of galaxies as non - emitters . in papers
i and ii we also showed the fraction of emitters is a function of redshift , rising significantly for lower luminosity sources over @xmath2 , most likely as a result of reduced dust extinction in the early lyman break population . therefore ,
as discussed in paper ii , we have also used a projected rest - frame ew distribution at @xmath11 assuming this evolutionary trend continues beyond @xmath69 6 .
two key factors enter into the calculation of the visibility of line emission in a ground - based survey .
firstly , for any target with a particular photometric redshift likelihood function @xmath70 , it may be that the spectral region surveyed by lris or nirspec does not completely cover the expected wavelength range where ly@xmath0 might be present .
secondly , the ew limit for ly@xmath0 emission will be a highly non - uniform function of wavelength due to the mitigating effect of night sky emission .
provided the photometric redshift solution we derive is robust , we can estimate both factors and hence derive the likelihood of seeing ly@xmath0 for each of our 26 sources assuming the relevant wavelength range studied and the exposure time secured , if the particular source of a given @xmath71 has a ew distribution drawn from the sample with @xmath56 . in the case of those sources for which the wavelength range searched does not fully sample the extended @xmath70 , we reduce the detection likelihood by the fraction of the integrated @xmath70 that lies outside our search range . for each target , we determine its redshift probability function @xmath70 using the photometric redshift code eazy @xcite . to determine the varying visibility function within our search range
, we first estimate the the noise within an aperture encompassing the expected profile of the line , assuming an emission line width of 10 fwhm which is typical of those detected in paper ii .
figure 3 illustrates the 5@xmath4 ew limit as a function of wavelength ( and ly@xmath0 redshift ) for most sources in our 12 hour lris exposure ( @xmath72 ) and a typical source studied with nirspec ( @xmath73 ) during a 5 hour exposure .
we note that although our survey spans a large range in redshift ( @xmath3 ) , the @xmath70 distributions for individual galaxies typically span a much smaller range .
the average 1@xmath4 redshift confidence interval for sources in our sample is only @xmath74 = 0.43 . since the nirspec exposures were usually single - object exposures ,
the limits vary from source to source depending on the conditions and exposure times .
we then apply a completeness correction to account for the fact that an emission line may fall in a noise trough and lie undetected , despite having an intrinsic flux above the 5@xmath4 limit . to estimate this completeness correction
, we follow the methods discussed in detail in paper ii , where we simulate the addition and recovery of fake line emission in our actual spectra , again assuming a fwhm of 10 .
as the absolute limits vary from source to source our keck survey is not complete to a fixed ew limit but , provided the limits are well - understood for each source , we can readily estimate the probability of seeing ly@xmath0 in our exposures . in the case of the @xcite fors2 survey we estimated the night sky emission from our own lris exposures normalizing the limits from numerical data supplied in that paper .
emission in our new keck spectroscopic campaign : the panels show the 5@xmath4 limiting ew calculated for a typical source studied in our 12 hour multi - slit lris exposure ( top ) and an example 5 hour long slit nirspec observation ( bottom ) .
the limits vary from source to source depending on the continuum brightness and the exposure times .
an additional completeness correction is taken into account by adding and attempting to recover fake emission lines with fluxes equal to the 5@xmath4 flux limit at the wavelength of insertion .
see text and paper ii for more detailed discussion .
, scaledwidth=50.0% ] the above simulations can be used to verify that our keck survey is well - placed to search for ly@xmath0 emission .
out of the combined 26 targets from our survey and that of @xcite , 24 are covered spectroscopically over more than half the integrated probability of their photometric redshift distribution , and 17 are covered over 95% of the range . additionally , we are able to determine the fraction of our spectra occulted by oh sky emission . for example , for a j=27 galaxy in one of our lris exposures , we are sensitive to lines with ew @xmath33 30 over @xmath75 of our usable spectral range ( see figure 3 ) . similarly , for a 5 hr nirspec exposure of a j@xmath76 galaxy , we are sensitive to lines with ew @xmath35 55 over @xmath77 our spectral range .
the results of the simulations are shown in figure 4 where , depending on whether we adopt the ew distribution observed at @xmath56 or that extrapolated to @xmath6 in paper ii assuming continuity in redshift - dependent increase in line emission seen over @xmath2 , we would expect to recover 7 - 8 emission lines in the combined keck and vlt surveys .
in contrast , we have only 2 robust detections ( both in the keck sample , figure 2 ) and at most 4 including the marginal candidate discussed by @xcite and hudf09_1596 at @xmath78 shown in figure 2 . assuming all the targeted sources are at @xmath28 , given our previously mentioned assumptions , our results reject the input ew distributions at the 99.3@xmath79 level of significance ( 91.4@xmath79 if the two marginal detections are included ) . emission lines with greater than or equal to 5@xmath4 significance in the combined keck and vlt survey of 26 sources .
the blue histogram shows the likelihood function for 10,000 monte carlo realizations assuming the intrinsic line emission properties follow the luminosity dependence seen in our @xmath56 @xmath19-drop sample ( figure 1 ( top ) .
the open histogram shows the expectation if the fraction of line emitters continues to increase with redshift at the rate described in paper ii .
vertical lines show the recovered number of emitters ( robust and maximal including marginal detections in both the keck and vlt surveys).,scaledwidth=50.0% ] we can display the significance of this downturn with increasing redshift in the terms of the fraction of ly@xmath0 emission seen in lyman break galaxies , @xmath80(ly@xmath0 ) as in paper i. the difficulty we face in creating such a figure is the non - uniform ew limit across the various targets in the keck and vlt campaigns , in contrast to the more straightforward uniform search we undertook with deimos at @xmath2 . to account for this ,
we assume a simple model in which ly@xmath0 emission is transmitted without igm absorption for a fraction @xmath81 of galaxies , while it is fully extinguished by the igm for a fraction of galaxies ( 1-@xmath81 ) .
we assume that @xmath82 at redshifts below 6 , where the universe is believed to be highly ionized @xcite , and that @xmath81 is independent of the intrinsic ew of a ly@xmath0 emission line .
we caution that an interpretation in terms of absolute values of @xmath81 is premature , as there is still some debate on whether the igm is fully ionized at @xmath83 @xcite , but emphasize that our value of @xmath81 at @xmath84 is computed relative to the value assumed at @xmath83 .
additionally , with the increased fraction of emitters in our @xmath83 sample from paper ii , we do not see any evidence for a decrease in @xmath81 prior to @xmath83 , though we can not rule it out .
it is important to note that our @xmath81 is differs from @xmath85 , commonly defined in the literature as the total escape fraction of ly@xmath0 photons ( e.g. , @xcite ) .
@xmath85 represents the total transmission of ly@xmath0 , accounting for both attenuation of photons within the galaxy by mechanisms such as dust , as well as any attenuation by the igm .
our definition of @xmath81 is only intended to account for any downturn in the fraction of lbgs which show observable ly@xmath0 emission from the @xmath86 ( or @xmath87 ) extrapolated ew distributions from paper ii , and represents an igm extinction averaged over the entire population .
to compute the most likely value of @xmath81 , we undertake monte carlo simulations using the previously described ew distributions , but with @xmath81 now added as a free parameter .
we vary @xmath81 from 0 to 1 in steps of 0.01 , and compute n=1000 simulations for each step .
we can then calculate the probability distribution for @xmath81 given our n@xmath88=2 confirmed sources using bayes theorem : @xmath89 here , @xmath90 is the prior probability for @xmath81 , which we take to be uniform for @xmath91 , and @xmath92 is the probability , drawn from our monte carlo simulations , that we would find n@xmath93 sources for a given value of @xmath81 . assuming that the intrinsic ew distribution for our observed sources is that of paper ii at @xmath86 , we find @xmath94 , while using the @xmath87 extrapolated distribution yields @xmath95 . in the figure 5 , we plot the value of x(ly@xmath0 ) in the same luminosity bins of paper ii , as predicted by our best fit values of @xmath81 .
we stress that this figure is intended to serve as a continuation of the visualization provided in papers i and ii , rather than a statistiscal result of our study . due to our strongly varying limiting ew sensitivity ( as a function of both wavelength and object magnitude ) , choosing
a fixed ew limit will exclude a non - negligible fraction of useful data from our analysis .
our monte carlo simulations are able to utilize the full data set , simulating whether we would have likely seen a line even when our ew limits are above the fixed thresholds used in fig .
5 , and thus represent the major statistical result of this study . using the models of @xcite to predict what global neutral hydrogen fraction , @xmath96
would be required to account for this decline , we find @xmath97 , and @xmath98 , respectively .
the models of @xcite , which provide a more comprehensive treatment of ly@xmath0 radiative transfer through outflows , result in an increased value for @xmath96 in both cases . in emission , @xmath80(ly@xmath0 ) , adjusted as discussed in the text to approximate one within a similar luminosity range with a rest - frame ew in excess of 25 .
data points for the galaxies with @xmath99 are displaced by @xmath100 in redshift for clarity .
data over @xmath2 is from paper i and paper ii , and new estimates beyond @xmath28 are derived from the present paper , including sources discussed by @xcite .
the curves shown represent the aggregate redshift probability distributions for our sources in the z @xmath101 6 bin ( black ) , and the z @xmath101 7 bin ( blue ) ; probability distributions for individual sources are typically much sharper.,scaledwidth=50.0% ]
although we consider the most likely explanation for our observed decrease in the number of lbgs which show observable ly@xmath0 emission to be an increase with redshift in the neutral fraction of the igm , it is important to remember our assumptions .
foremost we have assumed that _ all _ of our 26 targets have true redshifts beyond @xmath102 .
should there be low redshift interlopers or galactic stars in our new sample , we will overestimate the decline in the ly@xmath0 fraction .
secondly , we have assumed the deimos spectra from paper ii constitute a representative sample for calculating the expected ew distribution for 6.3 @xmath103 8.2 .
although the uncertainties here are not as great , we plan further studies with deimos to increase the statistical sample of 5.5 @xmath103 6.3 lbgs .
of course our observed decrease in the ly@xmath0 fraction could also be attributed to an increased opacity arising from dust within the lbgs .
however , given the blue uv continuum slopes observed for galaxies with @xmath28 @xcite , we consider this explanation unlikely .
our diagnosis of a possible increase in the neutral hydrogen fraction beyond @xmath69 6.3 is supported by the earlier study of @xcite .
they found 1 marginal candidate out of 7 targets whereas we find 2 robust and 1 marginal cases out of our 19 targets spanning a larger luminosity and redshift range .
our conclusion is also supported by laes studies at @xmath104 and @xmath105 by @xcite and @xcite .
compared to @xmath104 , their lae sample at @xmath106 displays systematically lower ews for ly@xmath0 .
they also derive little evolution in the rest uv luminosity function for laes , but a decrease in the ly@xmath0 lf , which could be explained by an increase in x@xmath107 .
our derived values of x@xmath107 are slightly higher than that of @xcite , perhaps consistent with our survey probing to higher redshifts than their @xmath106 laes .
@xcite have recently compiled results from numerous ly@xmath0 and uv luminosity function studies to derive a volumetrically averaged ly@xmath0 escape fraction , and find very similar results .
their derived ly@xmath0 escape fraction steadily increases with redshift below @xmath108 , then tentatively drops off at higher redshifts . very recently , @xcite report the convincing detection of ly@xmath0 emission in a small fraction ( 3/11 ) of lbgs that , by virtue of their selection using subaru imaging , are more luminous ( @xmath109 ) than most of the objects considered here .
such a complementary campaign targeting luminous lbgs selected from larger volumes will provide further insight into whether reionization is responsible for the declining fraction of line emission .
we note that our measured decrease in the fraction of lbgs with strong ly@xmath0 potentially agrees with the result of @xcite .
although they argue against any evidence for reionization at @xmath110 , they find that @xmath111 of galaxies at this redshift show strong ly@xmath0 emission , comparable to the fraction we detect in this work , spread across a larger redshift range . with the new generation of multi - object , near infrared spectrographs , such as mosfire , set to come online soon ,
the prospects for this field are bright .
in addition to the significant multiplexing advantage , the increased sensitivity of these detectors will allow us to probe the lower luminosity ranges at @xmath112 to ew limits comparable to those in paper ii between sky lines .
having such a statistical sample is key for allowing the quantification of any change in the hydrogen neutral fraction .
we thank the referee for valuable comments which improved the manuscript , george becker for allowing us to use his nirspec reduction pipeline , and mark dijkstra for his helpful discussion .
rse and mas thank the institute of astronomy , cambridge , where this work was completed , for its support .
we also wish to recognize and acknowledge the very signicant 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 .
dps acknowledges financial support from a postdoctoral fellowship from the science technology and research council .
ber is supported by a hubble fellowship grant , program number hst - hf-51262.01-a provided by nasa from the space telescope science institute , which is operated by the association of universities for research in astronomy , incorporated , under nasa contract nas5 - 26555 .
jsd acknowledges the support of the royal society via a wolfson research merit award , and the support of the european research council via an advanced grant .
rjm acknowledges the support of the royal society via a university research fellowship . ,
m. , et al .
2003 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol . 4834 , society of photo - optical instrumentation engineers ( spie ) conference series , ed .
p. guhathakurta , 161 , s. m. , et al .
2003 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol . 4841 , society of photo - optical instrumentation engineers ( spie ) conference series , ed . m. iye & a. f. m. moorwood , 1657 | using deep keck spectroscopy of lyman break galaxies selected from infrared imaging data taken with the wide field camera 3 onboard the hubble space telescope , we present new evidence for a reversal in the redshift - dependent fraction of star forming galaxies with detectable lyman alpha ( ly@xmath0 ) emission in the redshift range @xmath1 .
our earlier surveys with the deimos spectrograph demonstrated a significant increase with redshift in the fraction of line emitting galaxies over the interval @xmath2 , particularly for intrinsically faint systems which dominate the luminosity density . using the longer wavelength sensitivities of lris and nirspec , we have targeted 19 lyman break galaxies selected using recent wfc3/ir data whose photometric redshifts are in the range @xmath3 and which span a wide range of intrinsic luminosities .
our spectroscopic exposures typically reach a 5@xmath4 sensitivity of @xmath5 for the rest - frame equivalent width ( ew ) of ly@xmath0 emission . despite the high fraction of emitters seen only a few hundred million years later
, we find only 2 convincing and 1 possible line emitter in our more distant sample .
combining with published data on a further 7 sources obtained using fors2 on the eso vlt , and assuming continuity in the trends found at lower redshift , we discuss the significance of this apparent reversal in the redshift - dependent ly@xmath0 fraction in the context of our range in continuum luminosity .
assuming all the targeted sources are at their photometric redshift and our assumptions about the ly@xmath0 ew distribution are correct , we would expect to find so few emitters in less than 1% of the realizations drawn from our lower redshift samples .
our new results provide further support for the suggestion that , at the redshifts now being probed spectroscopically , we are entering the era where the intergalactic medium is partially neutral . with the arrival of more sensitive multi - slit infrared spectrographs
, the prospects for improving the statistical validity of this result are promising . |
in the interaction picture the effective interaction hamiltonian for the four - wave mixing ( fwm ) parametric process takes the form@xcite @xmath93\\ & \cdot e^{-i(\omega_{p}+\omega_{c}-\omega_{as}-\omega_{s})t}+h.c.,\\ & = \frac{i\hbar l}{2\pi}\int d\omega_{as}d\omega_{s}\kappa \text{sinc}(\frac{\delta kl}{2})\frac{1}{\sqrt{2}}[\hat{a}^{+}_{\sigma^{1}_{p}}(\omega_{s } ) \hat{a}^{+}_{\sigma^{1}_{c}}(\omega_{as } ) + \hat{a}^{+}_{\sigma^{2}_{p}}(\omega_{s } ) \hat{a}^{+}_{\sigma^{2}_{c}}(\omega_{as})e^{i\phi_{1}}e^{i\phi_{2}}]e^{-i(\omega_{p}+\omega_{c}-\omega_{as}-\omega_{s})t}+h.c.\\ \end{split } \tag{s.1}\end{aligned}\ ] ] where @xmath94 @xmath95 is the creation operator of the stokes photons with pump polarization @xmath96 , @xmath97 @xmath95 is the creation operator of the anti - stokes photons with coupling polarization @xmath98 , @xmath99 ( @xmath100 ) is the length of the coupling ( pump ) laser arm j in the mach - zehnder interferometer , @xmath101 , @xmath102 . based on perturbation theory @xcite , the two - photon ( biphoton ) state @xmath103 can be expressed as @xmath104|0\rangle\\ & = l\int d\omega_{as } \kappa(\omega_{as } ) \text{sinc}(\frac{\delta k(\omega_{as})l}{2})\frac{1}{\sqrt{2}}[|\omega_s=\omega_{p}+\omega_{c}-\omega_{as } , \sigma^{1}_{p}\rangle |\omega_{as } , \sigma^{1}_{c}\rangle + |\omega_s=\omega_{p}+\omega_{c}-\omega_{as } , \sigma^{2}_{p}\rangle |\omega_{as } , \sigma^{2}_{c}\rangle e^{i\phi}]\\ & = l\int d\omega_{as}\kappa(\omega_{as})\text{sinc}[\frac{\delta k(\omega_{as})l}{2}]|\omega_s=\omega_{p}+\omega_{c}-\omega_{as}\rangle|\omega_{as}\rangle \otimes \frac{1}{\sqrt{2 } } [ \end{split } \tag{s.2}\end{aligned}\ ] ] where @xmath105 . equation ( s.2 ) shows generation of both time - frequency entanglement ( energy conservation due to the time translation symmetry of the system @xmath106 ) and polarization entanglement . for simplify , equation ( s.2 ) can also be further described in the time and polarization domain as @xmath107 in eq .
( s.3 ) , the time - domain wave function @xmath33 , which indicates the time - frequency entanglement , is the fourier transform of the photon spectrum @xmath108 $ ] and can be measured through coincidence counts in the time domain directly .
equations ( s.2 ) and ( s.3 ) correspond to eqs .
( 1 ) and ( 2 ) in the main text , respectively .
in order to produce the polarization entangled bell state , the phase difference @xmath21 in eq.(2 ) in the main text should be locked . for the experimental setup ,
the phase difference is given by @xmath109 where @xmath110 ( @xmath111 ) is the wave length of the coupling ( pump ) laser beam .
let s define @xmath112 then the phase difference can be rewritten as @xmath113+\delta [ ( l_{c2}-l_{p2})-(l_{c1}-l_{p1 } ) ] .
\tag{s.6}\end{aligned}\ ] ] under the condition @xmath114 , equation ( s.6 ) can be further reduced to @xmath115=\frac{k_{0}}{k_{l}}\varphi+\varphi_{0 } , \tag{s.7}\end{aligned}\ ] ] here @xmath116 $ ] is the phase difference of the mach - zehnder interferometer for the locking laser ( @xmath117 , 795 nm ) , @xmath118 is the phase shift between the two mach - zehnder interferometers ( the locking laser does nt overlap with the pump - coupling laser fields ) .
we use the two - photon polarization correlation measurement to determine the relation between the locking point @xmath119 and the phase difference @xmath21 . as shown in fig .
( s.1 ) , with four groups of correlation measurements ( @xmath120 and @xmath46 ) , the locking point @xmath119 are found to be linearly relative to the phase difference @xmath21 that fitted very well with the theoretical results from eq .
( s.7 ) .
( color online ) .
( a ) polarization correlation of the paired stokes and anti - stokes photons .
the polarization angle for stokes photons are set at @xmath23 , the phase difference @xmath21 is @xmath45 ( black line and cycle ) , @xmath121 ( red line and triangle ) , @xmath122 ( blue line and star ) and @xmath46 ( green line and square ) , respectively .
( b)the locking point versus the phase difference of the polarization entangled state .
all of the solid lines are theoretically fitted.,width=321 ]
the properties of the four bell states generated in our experiments are described in the main text .
the detailed density matrix results are showing in the following : | we demonstrate an efficient experimental scheme for producing polarization - entangled photon pairs from spontaneous four - wave mixing ( sfwm ) in a laser - cooled @xmath0rb atomic ensemble , with a bandwidth ( as low as 0.8 mhz ) much narrower than the rubidium atomic natural linewidth . by stabilizing the relative phase between the two sfwm paths in a mach - zehnder interferometer configuration ,
we are able to produce all four bell states .
these subnatural - linewidth photon pairs with polarization entanglement are ideal quantum information carriers for connecting remote atomic quantum nodes via efficient light - matter interaction in a photon - atom quantum network .
the connectivity of a long - distance photon - atom quantum network strongly depends on efficient interactions between flying photonic quantum bits and local long - lived atomic matter nodes @xcite .
such efficient quantum interfaces , which convert quantum states ( such as time - frequency waveform and polarizations ) between photons and atoms , require the photons to have a bandwidth sufficiently narrower than the natural linewidth of related atomic transitions ( such as 6 mhz for rubidium d1 and d2 lines ) . as a standard method for producing entangled photons
, spontaneous parametric down conversion ( spdc ) in a nonlinear crystal usually has a wide bandwidth ( larger than terahertz ) and very short coherence time ( less than picosecond ) .
many efforts have been investigated in the past more than one decade to narrow down the spdc photon bandwidth using optical cavities @xcite . however , the bandwidth of spdc polarization - entangled photon pairs is still wider than most atomic transitions and leads to a very low efficiency of storing these polarization states in a quantum memory @xcite .
our motivation was stimulated by the recent progress in generating subnatural - linewidth biphotons by using continuous - wave spontaneous four - wave mixing ( sfwm ) in a laser - cooled atomic ensemble with electromagnetically induced transparency ( eit ) @xcite .
photons produced from this method not only have narrow bandwidth but also automatically match the atomic transitions .
the applications of these narrow - band photons include the demonstration of a single - photon memory with a storage efficiency of about 50% @xcite , a single - photon precursor @xcite , and coherent control of single - photon absorption and reemission @xcite .
however , while this method provides a natural entanglement mechanism in the time - frequency domain , it is extremely difficult to produce polarization entanglement because of the polarization selectivity of eit in a nonpolarized atomic medium @xcite .
it is possible to generate the polarization entanglement by scarifying the eit effect , but the photon generation efficiency is low and the bandwidth is not narrower than the atomic natural linewidth @xcite . the `` writing - reading '' technique with optical pumping provides a solution to polarization entanglement but results in reducing time - frequency entanglement @xcite . in this letter ,
we report our work on producing subnatural - linewidth polarization - entangled photon pairs by using the continuous - wave sfwm cold - atom eit configuration .
we demonstrate that the polarization entanglement can be efficiently produced by making use of a mach - zehnder interferometer in a two - path sfwm setup while maintaining the eit effect in controlling the photon bandwidth . by tuning the phase difference between the two sfwm paths and properly setting the driving laser polarizations
, we can generate all four bell states .
these photons have a coherence time of up to 900 ns and an estimated bandwidth of about 1 mhz that is much narrower than the rb atomic natural linewidth ( 6 mhz ) .
these subnatural - linewidth polarization - entangled photon pairs are ideal flying qubits for connecting remote atomic quantum nodes in a quantum network .
our experimental setup is illustrated in fig .
1 . we work with a two - dimensional @xmath0rb magneto - optical trap ( mot ) with a longitudinal length of @xmath1 cm @xcite .
the experiment is run periodically . in each cycle , after 4.5 ms mot time , the atoms are prepared in the ground level @xmath2 and followed by a 0.5 ms sfwm biphoton generation window . along the longitudinal direction ,
the atoms have an optical depth of 32 in the @xmath3 transition .
the pump laser ( 780 nm , @xmath4 ) is 80 mhz blue detuned from the transition @xmath5 , and the coupling laser ( 795 nm , @xmath6 ) is on resonance with the transition @xmath7 . the linear polarized pump laser beam , with a @xmath8 diameter of 1.8 mm , is equally split into two beams after a half - wave plate and the first polarization beam splitter ( pbs1 ) .
these two beams , with opposite circular polarizations ( @xmath9 and @xmath10 ) after two quarter - wave plates , then intersect at the mot with an angle of @xmath11 to the longitudinal axis .
similarly , the two coupling laser beams after pbs2 with opposite circular polarizations overlap with the two pump beams from opposite directions . in the presence of these two pairs of counterpropagating pump - coupling beams , phase - matched stokes ( @xmath12 ) and anti - stokes ( @xmath13 ) paired photons
are produced along the longitudinal axis and coupled into two opposing single - mode fibers ( smfs ) . in each sfwm path , the polarizations of the stokes and anti - stokes photons follow those of the corresponding pump and coupling field , respectively .
the two smf spatial modes are focused at the mot center with a @xmath8 diameter of 0.4 mm .
after two narrow - band filters ( @xmath14 and @xmath15 , 0.5 ghz bandwidth ) , the photons are detected by two single - photon counter modules ( spcm , perkin elmer spcm-4q4c ) and analyzed by a time - to - digital converter ( fast comtec p7888 ) with a time bin width of 1 ns . two sets of quarter - wave plates , half - wave plates , and pbss are inserted for measuring polarization correlation and quantum state tomography . to obtain the polarization entanglement
, we must stabilize the phase difference between the two sfwm spatial paths .
this is achieved by injecting a reference laser beam ( 795 nm , 110 mhz blue detuned from the transition @xmath16 ) from the second input of pbs2 .
the two reference beams split after pbs2 are then recombined after pbs1 and detected by a photodetector ( a half - wave plate and a pbs are used to obtain the interference ) , as shown in fig .
1 .
this is a standard mach - zehnder interferometer to the reference laser . locking the phase difference of the two arms of the mach - zehnder interferometer with
a feedback electronics stabilizes the phase of the two sfwm paths ( see supplementary material @xcite ) . to avoid interaction with the cold atoms , the reference beams
are slightly shifted away from the pump - coupling beams but pass through the same optical components .
( color online ) .
( a)-(c ) two - photon coincidence counts as a function of the relative time delay between the stokes and anti - stokes photons .
( a ) path 1 : @xmath17 and @xmath18 ; path 2 : blocked .
( b ) path 1 : blocked ; path 2 : @xmath19 and @xmath20 .
( c ) both paths are present .
( d ) the temporal length of the paired photons as a function of the coupling laser power .
all the black lines are theoretical plots.,width=321 ] ( color online ) .
polarization correlation of the paired stokes and anti - stokes photons at @xmath21=0 .
the polarization angle for stokes photons are set at @xmath22 ( red solid line and circle ) and @xmath23 ( blue dashed line and triangle ) .
the black dotted line and squares are the data taken without stabilizing @xmath21.,width=283 ] following the perturbation theory @xcite , the produced two - photon state can be described as ( see supplemental material @xcite for the derivation ) @xmath24|\omega_s=\omega_{p}+\omega_{c}-\omega_{as}\rangle\\ & & |\omega_{as}\rangle d\omega_{as}\otimes \frac{1}{\sqrt{2}}(|\sigma_{p}^{1}\rangle_{s}|\sigma_{c}^{1}\rangle_{as } + e^{i\phi}|\sigma_{p}^{2}\rangle_{s}|\sigma_{c}^{2}\rangle_{as } ) , % \end{split}\end{aligned}\ ] ] which shows hyperentanglement in frequency ( continuous ) and polarization ( discrete ) . the upper indices ( 1 and 2 ) represent the two sfwm spatial paths .
@xmath25 is the phase mismatching , and @xmath26 is the nonlinear parametric coupling coefficient .
$ ] determines the photon spectrum .
@xmath21 is the phase difference between the two sfwm paths .
equation ( 1 ) can also be rewritten in the time and polarization domains @xmath28 where @xmath29 and @xmath30 are the detection time of the stokes and anti - stokes photons , respectively .
@xmath31 and @xmath32 are their central frequencies , respectively .
the time - domain wave function @xmath33 results from the frequency entanglement ( @xmath34 ) and is the fourier transform of the two - photon joint spectrum .
meanwhile , as shown in eqs .
( 1 ) and ( 2 ) , the polarization entanglement can be manipulated by controlling the pump - coupling polarizations and phase difference .
we first characterize the two - photon nonclassical correlation in the time domain .
figure 2 ( a)-2(c ) show the two - photon coincidence counts as functions of the relative time delay ( @xmath35 ) with the polarization configurations @xmath36 , @xmath37 , and @xmath38 , respectively .
we carefully balance the pump and coupling laser powers on the two sfwm paths to make their correlation indistinguishable in the time domain for achieving the maximally polarization - entangled states . at each path , the pump beam has a power of 8 @xmath39w and the coupling beam of 2 mw . as shown in fig .
2 , these paired photons have a temporal correlation length of 300ns . normalizing the coincidence counts to the accidental uncorrelated background counts , we obtain the normalized cross correlation @xmath40 with a peak value of 35 at @xmath41=25 ns . with measured autocorrelations
@xmath42 , we confirm that the cauchy - schwartz inequality @xmath43^{2}\leq g_{s , s}^{(2)}(0)g_{as , as}^{(2)}(0)$ ] is violated by a factor of @xmath44 , which clearly indicates the quantum nature of the paired photons . with an integration time bin of 300 ns , the normalized cross correlation has a reduced peak value of 10 , which still violates the cauchy - schwartz inequality by a factor of 25 . the solid curves in fig
. 2(a)-2(c ) are calculated from eq .
( 2 ) .
the nearly perfect agreement between the theory and experiment indirectly verifies the time - frequency entanglement .
the shorter correlation time in fig .
2(c ) compared to that in fig.2(a ) and ( b ) is caused by the addition of the powers from the two coupling beams that widen the eit and biphoton bandwidth .
the coherence time of about 300 ns corresponds to an estimated bandwidth of 2.9 mhz ( also confirmed from the theory ) .
the photon bandwidth can be further reduced by lowering the coupling laser power to narrow the eit window .
figure 2(d ) shows the measured correlation time versus the coupling laser power . with 0.13
mw coupling laser power , we obtain a coherence time of up to 900 ns , which corresponds to a bandwidth of about 0.8 mhz . the solid curve in fig .
2(d ) is also obtained with eq .
( 2 ) .
we next demonstrate that all four polarization - entangled bell states can be realized by locking the phase @xmath21 as well as properly choosing the polarizations of the coupling and pump laser beams .
as shown in eq .
( 2 ) , locking the phase difference @xmath21 as @xmath45 or @xmath46 and setting @xmath47 , @xmath48 , we can produce two polarization - entangled bell states : @xmath49 similarly , by setting @xmath50 , @xmath51 , we obtain other two bell states : @xmath52 ( color online ) .
tomography measurement results of the photon pairs at the target polarization - entangled bell states ( a ) @xmath53 , ( b ) @xmath54 , ( c ) @xmath55 , and ( d ) @xmath56.,width=302 ] figure 3 displays the measured two - photon polarization correlations for @xmath57 by locking @xmath58 .
the coincidence counts are integrated from @xmath59 to 300 ns for a total measurement time of 150 s. the circle data ( @xmath60 , red line ) are collected by fixing the stokes photon polarization angle at @xmath22 and the triangle data ( @xmath61 , blue line ) at @xmath23 .
other parameters during the measurement remain the same as those for fig .
2(c ) .
the solid cosine- and sine - wave curves are the theoretical fits with adjustable background and amplitude parameters .
we obtain the visibility @xmath62 , which is beyond the requirement of @xmath63 for violating the bell - clauser - horne - shimony - holt ( bell - chsh ) inequality@xcite . for comparison
, we plot the data without locking the phase as the square points ( @xmath64 , black line ) which shows no quantum interference . to obtain a complete characterization of the polarization entanglement
, we also make a quantum state tomography to determine the density matrix following the maximum likelihood estimation method @xcite . with two additional quarter - wave plates ,
the circular polarization basis @xmath65,@xmath66,@xmath67 , and @xmath68 can be converted into linear polarization basis @xmath69,@xmath70,@xmath71 , and @xmath72 .
then we use a half - wave plate followed by a pbs as the polarization selector .
the density matrix is constructed from the coincidence counts at 16 independent projection states ( see supplementary material @xcite ) .
the graphical representations of the obtained density matrix for @xmath73 and @xmath74 are shown in fig .
4 , from which we obtain the fidelities of 93.6% , 91.8% , 92.9% , and 95.2% , respectively .
we also use the density matrix to test the violation of the bell - chsh inequality ( @xmath75 ) and get the values @xmath76 , @xmath77 , 2.3 @xmath78 0.02 , and [email protected] for the obtained four bell states .
( color online ) .
visibility versus the biphoton generation rate .
the solid line is linear fit .
the dashed line marks the boundary for violating the bell - chsh inequality .
, width=283 ] now , we turn to the brightness of our photon source . for the data shown in fig . 2 ( c ) , by taking into account the fiber coupling efficiency ( @xmath80 ) , filter transmission ( @xmath80 ) , detector quantum efficiency ( @xmath81 ) , and duty cycle ( @xmath82 ) , our photon source spontaneously generates about @xmath83 photon pairs per second . with the pump power of @xmath84 and a linewidth of @xmath85 mhz , we estimate a spectrum brightness of @xmath86 and the normalized spectrum brightness of @xmath87 . we can further increase the photon pair generation rate by increasing the pump laser power .
figure 5 shows the dependence of the visibility of the polarization correlation to the photon pair generation rate .
the visibility is estimated from @xmath88 @xcite .
as long as the generation rate is less than @xmath89 ( here the averaged normalized cross correlation is under consideration ) , the visibility is larger than @xmath63 , which is the boundary to violate the bell - chsh inequality . in summary
, we have demonstrated an efficient experimental scheme for producing subnatural - linewidth photon pairs with polarization entanglement .
the polarization entanglement results from the interference between the two sfwm spatial paths . by stabilizing the phase difference between these two paths and setting properly the driving laser polarizations
, we produce all four bell states , confirmed by the quantum state tomography measurements .
their long coherence time ( up to 900 ns ) and narrow bandwidth ( about 1 mhz ) make them a promising entangled photon source for interacting with rubidium atomic quantum nodes .
we thank j. f. chen and c. liu for their helpful discussion .
this work was supported by the nsfc ( grants no .
11104085 , no . 11125417 , and no .
61378012 ) , the major research plan of the nsfc ( grant no .
91121023 ) , the skpbr of china ( grants no .
2011cb922104 and no .
2011cba00200 ) , the foytheg ( grant no .
yq2013050 ) , the prnpgz ( grant no .
2014010 ) , and the pcsirt ( grant no .
irt1243 ) .
s.d . was support by hong kong research grants council ( project no .
hku8/crf/11 g ) . 99 l. m. duan , m. d. lukin , j. i. cirac , and p. zoller , nature ( london ) * 414 * , 413 ( 2001 ) .
h. j. kimble , nature ( london ) * 453 * , 1023 ( 2008 ) . z. y. ou and y. j. lu , phys .
rev .
lett . * 83 * , 2556 ( 1999 ) .
h. wang , t. horikiri , and t. kobayashi , phys .
rev .
a * 70 * , 043804 ( 2004 ) .
x. h. bao , y. qian , j. yang , h. zhang , z .- b .
chen , t. yang , and j .- w .
pan , phys .
rev .
lett .
* 101 * , 190501 ( 2008 ) .
m. scholz , l. koch , and o. benson , phys .
rev .
lett . *
102 * , 063603 ( 2009 ) .
j. fekete , d. riel@xmath90nder , m. cristiani , and h. de riedmatten , phys .
rev .
lett . * 110 * , 220502 ( 2013 ) .
m. f@xmath91rtsch , j. u. f@xmath92rst , c. wittmann , d. strekalov , a. aiello , m. v. chekhova , c. silberhorn , g. leuchs , and c. marquardt , nat
. commun . * 4 * , 1818 ( 2013 ) .
h. zhang , x. m. jin , j. yang , h. n. dai , s. j. yang , t. m. zhao , j. rui , y. he , x. jiang , f. yang , g. s. pan , z. s. yuan , y. deng , z. b. chen , x. h. bao , s. chen , b. zhao , and j. w. pan , nat .
photonics * 5 * , 628 ( 2011 ) .
s. w. du , p. kolchin , c. belthangady , g. y. yin , and s. e. harris , phys .
rev .
lett . * 100 * , 183603 ( 2008 ) .
s. w. du , j. m. wen , and m. h. rubin , j. opt .
soc . am .
b * 25 * , c98 ( 2008 ) .
s. y. zhou , s. c. zhang , c. liu , j. f. chen , j. m. wen , m. m. t. loy , g. k. l. wong , and s. w. du , opt .
express * 22 * , 24124 ( 2012 ) .
s. zhang , j. f. chen , c. liu , m. m. t. loy , g. k. l. wong , and s. du , phys .
rev .
lett . * 106 * , 243602 ( 2011 ) .
s. zhang , c. liu , s. zhou , c .- s .
chuu , m. m. t. loy , and s. du , phys .
rev .
lett . * 109 * , 263601 ( 2012 ) .
y. c. chen , c. w. lin , and i. a. yu , phys .
rev .
a * 61 * , 053805 ( 2000 ) .
h. yan , s. c. zhang , j. f. chen , m. m. t. loy , g. k. l. wong , and s. w. du , phys .
rev .
lett .
* 106 * , 033601 ( 2011 ) .
d. matsukevich and a. kuzmich , science * 306 * , 663 ( 2004 ) .
s. zhang , j. f. chen , c. liu , s. zhou , m. m. t. loy , g. k. l. wong , and s. du , rev .
sci .
instrum . * 83 * , 073102 ( 2012 ) .
see supplemental material for the derivation of eqs .
( 1 ) and ( 2 ) , the method to lock the phase @xmath21 , and the density matrix results of the four bell states .
j. f. clauser , phys .
rev .
d. * 9 * , 853 ( 1974 ) .
a. g. white , d. f. v. james , p. h. eberhard , and p. g. kwiat , phys .
rev .
lett . * 83 * , 3103 ( 1999 ) .
d. f. v. james , p. g. kwiat , w. j. munro , and a. g. white , phys .
rev .
a * 64 * , 052312 ( 2001 ) .
h. de riedmatten , j. laurat , c. w. chou , e. w. schomburg , d. felinto , and h. j. kimble , phys .
rev .
lett . * 97 * , 113603 ( 2006 ) .
j. k. thompson , j. simon , h. q. loh , and v. vuleti , science * 313 * , 74 ( 2006 ) .
* supplemental material : * |
for those who study the interstellar medium ( ism ) , the manifold states that atoms can assume in space are both a benefit and a hindrance . on the favorable side ,
the apportionments of atoms in various ionization stages and the fractions that become bound in molecular compounds ( including solids ) reveal the outcomes of fundamental processes that arise from various physical and chemical influences , and these , in turn , disclose the nature of nearby objects or conditions that created these circumstances .
conversely , if it turns out that we have a poor understanding of the fractions of atoms that exist in unseen states , we are hampered in our ability to gauge the relative total abundances of different elements .
these two considerations underscore the importance of our mastering the principles that govern how the atoms are subdivided into different states within environments that range from the local part of our galaxy to the most distant gas systems that we can detect in the universe
. a major theme in this article will be to cover the underlying principles of this topic .
viewing atomic species by their absorption features in the spectra of background continuum sources , such as stars or quasars , offers the simplicity of not having to understand the details of how the atoms are excited , as we must do if we are observing emission lines arising from collisional excitation or recombination .
moreover , the radiative transfer of absorption features in the optical and ultraviolet regions is straightforward , since the attenuation of light follows a simple exponential absorption law .
studies of absorption lines have revealed a broad assortment of findings on the state and composition of interstellar gases , and we review here many of these conclusions . at this point , it is appropriate to digress and offer a brief historical perspective .
the papers in this volume arose from one of four symposia that celebrated the founding of the carnegie observatories exactly 100 years ago .
we may reflect on the fact that some two years after that event , hartmann ( 1904 ) reported that stationary ca ii lines in the spectrum of the spectroscopic binary @xmath2 ori arose from the intervening medium , rather than from the atmospheres of the stars . as with many important discoveries , doubts about the interpretation remained ( young 1922 ) , but
eventually the interstellar nature of these features was firmly established ( plaskett & pearce 1933 ) . there
soon followed large surveys of interstellar lines , with carnegie astronomers playing an important role in the process ( merrill et al . 1937 ; adams 1949 ) . nowadays , modern echelle spectrographs with very high resolving powers show exquisite recordings of the complex superpositions of narrow velocity features , and these data allow one to study either weak features or those arising from very small parcels of gas along the line of sight to the target stars ( e. g. , blades , wynne - jones , & wayte 1980 ; welsh , vedder & vallerga 1990 ; welty , hobbs , & kulkarni 1994 ; barlow et al .
1995 ; crane , lambert & sheffer , 1995 ; welty , morton , & hobbs 1996 ; crawford 2001 ; welty & hobbs 2001 ; knauth , federman , & lambert 2003 ; welty , hobbs , & morton 2003 ) . against the backdrop of almost 100 years of studies of the interstellar medium using absorption features in the visible part of the spectrum ,
we celebrate yet another anniversary .
some 30 years ago the first papers describing results from the _ copernicus _ satellite ( rogerson et al .
1973a ) , launched in 1972 , demonstrated the extraordinary utility of using the ultraviolet part of the spectrum to study a vast array of features arising from important interstellar atomic constituents ( morton et al .
1973 ; rogerson et al .
1973b ) , along with our first insights on the excitation and widespread presence of molecular hydrogen ( spitzer et al . 1973 ; spitzer & cochran 1973 ) .
spitzer & jenkins ( 1975 ) reviewed many of the principal conclusions that came from this initial venture into the ultraviolet .
additional inferences on the nature of the ism arose from the _ international ultraviolet explorer ( iue ) _
( kondo et al . 1987 ) , which bridged the _ copernicus _ era and that of the _ hubble space telescope ( hst ) _ , launched in 1990 . _
hst _ still functions as a premier facility for gathering useful observations of interstellar absorption lines through the use of a new instrument , the space telescope imaging spectrograph ( stis ) that was installed in 1997 ( kimble et al .
1998 ; woodgate et al . 1998 ) .
a comprehensive review of the findings on interstellar abundances gathered from _
hst _ up to 1996 [ using a first - generation spectrograph , the goddard high - resolution spectrograph ( ghrs ) ] , has been written by savage & sembach ( 1996 ) . finally , the _ far ultraviolet spectroscopic explorer ( fuse ) _ ( moos et al . 2000 ) offers an opportunity to revisit the spectral region @xmath3
with high sensitivity , a band that has been mostly inaccessible since the days of _ copernicus_. in parallel with investigations of the galactic ism using observatories in earth orbit , large - aperture telescopes on the ground have allowed us to study absorptions by the same uv transitions arising from gaseous material at very large redshifts .
much of the progress in understanding the abundances in these distant systems has drawn upon lessons learned from studies of galactic material , where we have better insights on the fundamental abundance patterns ( from the sun and stars ) and the physical conditions .
recent reviews by pettini ( 2003 ) and calura et al ( 2003 ) summarize the current state of many of the conclusions on highly redshifted systems that have arisen from such studies . in the sections that follow
, we will explore how the apparent abundances of elements are altered by the effects of ionization and depletion onto dust grains .
we bypass discussions of the techniques employed to study the absorption features .
the reader is referred to a number of papers that cover some key points on the methodology of this type of research ( cowie & songaila 1986 ; jenkins 1986 , 1996 ; savage & sembach 1991 ; sembach & savage 1992 ) , plus a list of possible pitfalls in its conduct ( jenkins 1987 ) .
of all the species that have absorption features in the visible part of the spectrum , ti ii is the only one that represents an element in its favored stage of ionization within h i regions .
the remaining species , li i , ca i , ca ii , na i , and k i all have an ionization potential less than that of h i , with the consequence that they are mostly ionized to higher stages by uv starlight photons that can easily penetrate the regions .
one may attempt to derive the gas - phase abundance of an element @xmath4 from solutions of the ionization equilibrium equation involving the element s ionization rate @xmath5 from starlight photons , balanced against its rate of recombination governed by the rate coefficients @xmath6 with free electrons and @xmath7 with negatively charged or neutral , very small dust grains [ at a rate that is normalized to the hydrogen density @xmath8(h ) also depends on many incidental factors , such as the character and concentration of dust grains , the local density of starlight photons , and the local electron density . ] ] ( weingartner & draine 2001 ) .
that is , in principle the balance @xmath9\ ] ] may be solved to yield a total abundance of @xmath4 if the parameters are well defined .
if the local values of @xmath5 and the electron and hydrogen densities @xmath10 and @xmath8(h ) are poorly known , it is still possible to take equation [ ioniz_equlib ] for element @xmath4 and divide it by that for another element ( @xmath11 ) to derive a relative abundance of the two ( @xmath12 ) .
while this may seem to be an attractive solution for studying abundance trends , recent comparisons made by welty , hobbs & morton ( 2003 ) indicate that different elements showing features in the ultraviolet show mutually inconsistent values for the ratios of adjacent ionization stages when equation [ ioniz_equlib ] is invoked in the respective cases .
these disparities may arise from either errors in the atomic constants or , perhaps in some cases , from rapid changes in conditions that prevent the equilibria from being fully established .
until these effects are better understood , comparing abundances from minor ionization stages is probably a risky undertaking . for the major stages of ionization of different elements in h
i regions , i.e. , the lowest stages with ionization potentials greater than 13.6 ev , it is generally assumed that nearly all of the atoms are in that stage .
this is probably a trustworthy notion for individual regions with hydrogen shielding depths @xmath13 , but below this column density photons with energies @xmath14ev can penetrate the region , as illustrated in the bottom section of figure [ qplot_4 ] .
once this happens , some fractions of both the element in question and the accompanying hydrogen can be elevated to higher stages of ionization , but by different amounts .
it follows that a comparison of an element s most abundant stage to that of h i could lead to a misleading abundance ratio , which differs from the true one by a logarithmic difference @xmath15~,\ ] ] where @xmath16~.\ ] ] these equations are discussed in more detail by sofia & jenkins ( 1998 ) as @xmath2 , but @xmath2 is not adopted here because @xmath2 has often been used in much of the literature to denote a depletion factor , as expressed here in eq .
[ depl_def ] .
] , who also present their more complex modified forms that account for the additional effects of charge exchange with hydrogen .
alterations arising from charge exchange can be very important for certain elements , such as n and o , which have large reaction rate coefficients ( field & steigman 1971 ; butler & dalgarno 1979 ) .
the curves for @xmath17 for several elements are shown in figure [ qplot_4 ] .
this figure illustrates that under conditions of partial ionization , @xmath0(ar i)/@xmath0(h i ) underestimates the true value of ar / h because ar is more easily ionized than h , while the converse is true for @xmath0(mg ii)/@xmath0(h i ) . at first glance
, s might appear to be more immune to ionization corrections because its @xmath17 is not far from zero .
however , the second ionization potential of s is 23.4 ev , i.e. , far above the 13.6 ev needed to ionize hydrogen .
thus , there is some danger that significant amounts of s ii could arise from a fully ionized h ii region in front of a target star , and this might increase the apparent abundance of s ii over that which should have been identified with an h i region under study . to illustrate how important this effect can be
, we note that for the extreme case of @xmath18 cma the ratio of foreground s ii to h i is 10 times solar ( jenkins , gry , & dupin 2000 ) .
the picture just presented is an oversimplification , but one that is instructive . in real life , the incident radiation has a distribution of fluxes over different energies , and it is scattered and attenuated by different amounts as deeper layers of a cloud are penetrated .
what we view is the composite absorption produced by these layers .
one popular approach for calculating these effects is to use the program cloudy ( ferland et al .
1998 ) , which is designed to evaluate the various ion fractions at different locations within a cloud that is irradiated from the outside .
this program has been made generally available to the public ( http://www.nublado.org ) and is continuously updated .
another complication is that in some circumstances anomalously large temperatures created by shock heating may lead to collisional ionization , which can elevate even further the populations of high - ionization states ( trapero et al . 1996 ) .
useful general discussions of the effects that can arise from a mixture of ionized and neutral regions are given by sembach et al .
it is clear that they are especially important in low - density gases within about 100 pc of the sun ( jenkins et al . 2000 ; jenkins , gry , & dupin 2000 ; gry & jenkins 2001 ) .
howk & sembach ( 1999 ) point out the dangers inherent in deriving abundances in damped ly@xmath19 systems , and they back up their cautions by noting that the velocity profiles of al iii are very similar to those from species that are expected to arise from h i regions .
the notion that such systems may have a considerable amount of partially ionized gas is reinforced by the findings of vladilo et al .
( 2003 ) who noted that ar i seems to be deficient relative to other @xmath19-capture elements that are not expected to be appreciably depleted onto dust grains ( see [ depl ] below ) . by definition ,
damped ly@xmath19 systems have column densities @xmath20 , well in excess of the limit @xmath21 for shielding out photons with @xmath22ev .
perhaps these systems are either ( 1 ) subjected to appreciable fluxes with @xmath23ev , ( 2 ) are made up of many thinner regions that are each separately exposed to the ionizing radiation , or ( 3 ) have strong sources of internal ionization arising from rapid star formation . the fact that appreciable fractions of some elements in the ism have condensed into solid form
is supported by a number of independent observations .
first , starlight is absorbed and scattered by dust particles or very large molecules
the opacity takes the form of continuous attenuation in the visible and uv , together with discrete absorption bands ( 2200 feature , diffuse interstellar bands in the visible , and infrared absorption features ) .
when compared with the amount of hydrogen present , the magnitude of the attenuation indicates that the proportions of some atoms in solid form are large ( see the review by draine 2003 ) .
second , it is clear from circumstellar emissions in the infrared that many types of evolved stars cast off their dusty envelopes into the ism .
while such dust probably does not account for most of the material that is in solid form within the general ism , it probably provides nucleation sites for further growth in dense clouds .
finally , we may compare interstellar atomic abundances with solar ( or stellar ) abundances in the galactic disk and surmise that the inferred depletions measure the fractional amounts of material hidden in some kind of solid form ( or within large molecules that are difficult to identify spectroscopically ) .
specifically , we may define a depletion of free atoms using the usual bracket notation , @xmath24=\log(n(x)/n({\rm h}))-\log(x/{\rm h})_\odot\ ] ] and the amount of an element locked up in solids , relative to h , is @xmath25})~,\ ] ] assuming that solar abundances are good reference values that truly reflect the total abundances ( see the second footnote in [ reinvestigation ] ) .
much of the discussion that follows in this paper concentrates on the findings on depletions , ones that are strikingly evident from the uv absorption data , but which have also been evident even from the visible absorption lines of ti ii ( wallerstein & goldsmith 1974 ; stokes & hobbs 1976 ; stokes 1978 ) and the features of ca ii ( snow , timothy , & seab 1983 ; crinklaw , federman , & joseph 1994 ) ( notwithstanding the inherent difficulties in dealing with the uncertainties of the ionization equilibrium of the latter ; see eq .
[ ioniz_equlib ] ) . to obtain a better understanding of the underlying causes of the depletions of free atoms in space and the various factors that are influential ,
it is important to identify systematic effects that occur from one element to the next and from one environmental factor to the next .
three important effects have been observed : 1 .
elements that reside within refractory compounds are more severely depleted than ones that mostly condense into volatile ones . field ( 1974 )
was the first to identify this effect by comparing the depletions of different elements toward @xmath26 oph with their condensation temperatures , defined according to the temperature at which some fixed proportion of an element in a gaseous cosmic mixture freezes out into some solid compound(s ) .
the interpretation offered by field was that in cooling and expanding mass loss outflows from stars , the dense , inner zones of the flows gave rise to nearly complete depletions of elements making up the refractory compounds , whereas the outer , less dense zones that were cool enough to freeze out more volatile substances could not fully complete the process .
while this is an attractive concept , we must also acknowledge that condensation temperatures are an indication of general strengths of chemical bonds and may thus also apply to the balance of creation and destruction of dust grains in the general ism ( cardelli 1994 ) . a more modern presentation of the correlations of depletions with condensation temperature ( again for @xmath26 oph )
is shown in figure 4 of the review article by savage & sembach ( 1996 ) .
2 . the strengths of depletions decrease for gas moving at large velocities relative to the undisturbed material in a given region .
this effect was first pointed out by routly & spitzer ( 1952 ) who compared lines of ca ii to those of na i , and it has been substantiated in greater detail by siluk & silk ( 1974 ) , vallerga et al .
( 1993 ) , and sembach & danks ( 1994 ) .
misgivings that this effect could arise from changes in ionization have been overcome by the ultraviolet data ( jenkins , silk , & wallerstein 1976 ; shull , york , & hobbs 1977 ) , where several ionization stages for certain elements could be observed .
the interpretation of this effect is that gas at high velocities was recently subjected to a shock that was strong enough to partly ( or completely ) destroy the grains ( barlow 1978a , b ; draine & salpeter 1979 ; jones et al .
1994 ; tielens et al . 1994 ) .
3 . the strengths of depletions scale in proportion to the average density of gas ( i.e. , @xmath0(h ) divided by distance ) along a line of sight ( savage & bohlin 1979 ; harris , gry , & bromage 1984 ; spitzer 1985 ; jenkins , savage , & spitzer 1986 ; jenkins 1987 ; crinklaw et al . 1994 ) .
such a measure is a very crude indication of conditions , since there is no way to discriminate between very compact clouds with high densities and a small filling factor from far more extended regions having only moderately high densities but large filling factors .
nevertheless , the trend supports the propositions that the growth of dust grains is markedly enhanced in dense regions and/or that within these regions the destructive effects of shocks are muted .
successful alternative ways to gauge the predisposition for elements to deplete include the fraction of hydrogen in molecular form ( cardelli 1994 ) and , toward very distant sources at high galactic latitudes , simply @xmath0(h i ) ( wakker & mathis 2000 ) .
seven years ago , savage & sembach ( 1996 ) summarized the interstellar abundances derived from observations using the ghrs instrument aboard _
hst_. since that time , the continued success of _ hst _ ( with the stis instrument that replaced ghrs ) and emergence of _ fuse _ have enabled a significant number of important new surveys of interstellar abundances to be carried out . for this reason , it is useful to revisit the general trends , now that a larger suite of results , both new and old , can be compiled and analyzed . unlike previous studies that directly compared the strengths of depletions of specific elements with tangible properties of the lines of sight ( e.g. , jenkins 1987 or cardelli 1994 ) , we will initially ignore these external factors and instead characterize depletions simply to arise as the result of some measure of each element s general propensity to deplete , which works in combination with an index @xmath27 for the overall strength of depletions ( applied to all elements ) along the line of sight to a star .
this method has the advantage that links in the depletions from element to element are evaluated directly . as a result ,
they are not weakened by the inevitable inaccuracies in the relationships between individual depletions and some independently measured quantity ( either average density or the fraction of h@xmath1 ) .
after values of @xmath27 have been defined for many different stars , we can then probe how they correlate with the average density ( or any other interesting attribute for a line of sight ) .
when we investigate how depletions change under different conditions , we enjoy the benefit of being able to draw upon a wide selection of examples and not just those for a single element . for the study presented here , abundances toward 144 stars
were identified from over 70 investigations reported in the literature ( far too numerous to cite here ) .
the targets were limited to those for which we can extract reliable measures of hydrogen in the form of neutral atoms and h@xmath1 , derived from measurements of the ly@xmath19 damping profiles and the h@xmath1 lyman and werner bands.(h@xmath1 ) were not available . however , when upper limits for @xmath28 were known to be smaller than @xmath29(h i )
, we could disregard the effect of h@xmath1 in increasing the value of @xmath30 .
also , based on general findings for h@xmath1 ( savage et al .
1977 ; rachford et al . 2002 ) , it is safe to ignore h@xmath1 if the star s @xmath31 color excess is less than 0.1 , or ( equivalently ) @xmath32 . ] since it is usually impossible ( or very difficult ) to identify the amounts of hydrogen associated with individual velocity components along a line of sight ( without applying simplifying assumptions arising from `` undepleted '' species ) , only total column densities over all components were considered .
exceptions to this principle could be made for a few cases where one component strongly dominated the total .
we begin with the simplest assumption that from one line of sight to the next , the depletions of different elements scale with each other in some fixed proportion , as is illustrated for zn and p shown in the left - hand panel of figure [ abund0 ] .
however , this simple prediction does not hold for all elements , as one can see for the case of fe and p in the right - hand panel of this figure .
this behavior seems likely to be explained by the concept that typical dust grains have a mantle that is easily created and destroyed in the ism , and that this mantle surrounds a relatively indestructible core that survives all but the most energetic ( and very rare ) shock events ( spitzer 1985 ; jenkins , savage , & spitzer 1986 ; joseph 1988 ; spitzer & fitzpatrick 1993 ) .
we are thus forced to consider at least three parameters for a general law that characterizes the depletion of an element @xmath4 , @xmath33=a_xf_*+a_{0,x},\ ] ] where @xmath34 represents the relative ease with which an element s individual depletion changes as the general line - of - sight depletion multiplier @xmath27 varies , and @xmath35 represents an offset attributable to special compounds within the most rugged portions of the grains . for the study presented here ,
the normalization of @xmath27 was defined to make a value of 1.0 equal to the depletions on line of sight toward @xmath26 oph , a star that shows for one of its velocity components a significant amount of depletion .
it is also a case that has probably received the most attention of any star for depletions of a broad range of different elements [ see table 5 of savage & sembach ( 1996 ) for a summary ] .
c&8.39&@xmath40&@xmath41 & 2.2 & 6 & 0.899 + n&7.93&@xmath42&@xmath43 & 27.9 & 15 & 0.022 + o&8.69&@xmath44&@xmath45 & 38.7 & 28 & 0.086 + mg&7.54&@xmath46&@xmath47 & 42.9 & 62 & 0.969 + si&7.54&@xmath48&@xmath49 & 5.0 & 7 & 0.660 + p&5.57&@xmath50&@xmath51 & 31.9 & 52 & 0.987 + cl&5.27&@xmath52&@xmath53 & 41.3 & 35 & 0.215 + ti&4.93&@xmath54&@xmath55 & 30.5 & 32 & 0.543 + cr&5.68&@xmath56&@xmath57 & 16.8 & 12 & 0.158 + mn&5.53&@xmath58&@xmath59 & 52.7 & 59 & 0.704 + fe&7.45&@xmath60&@xmath61 & 59.7 & 54 & 0.275 + ni&6.25&@xmath62&@xmath63 & 31.0 & 9 & 0.000 + cu&4.27&@xmath64&@xmath65 & 7.7 & 9 & 0.561 + zn&4.65&@xmath66&@xmath67 & 12.6 & 12 & 0.399 + kr&3.23&@xmath68&@xmath69 & 3.1 & 9 & 0.961 + weighted least - squares determinations for the constants @xmath34 and @xmath35 associated with all of the elements covered in this study are given in table [ parameters ] .
figure [ abundf2 ] shows plots of the observed depletions as a function of @xmath27 ; in each case a dashed line is drawn to show the relationship expressed by equation [ law ] for the appropriate values of @xmath34 and @xmath35 shown in the table . within the uncertainties of @xmath34 ,
the elements c , n , o , and kr all have slopes that are consistent with zero , in accord with the findings of cardelli et al .
( 1996 ) , meyer , cardelli , & sofia ( 1997 ) , meyer , jura , & cardelli ( 1998 ) and cartledge , meyer & lauroesch ( 2003 ) .
table [ parameters ] also indicates the values of @xmath39 for the fits , the associated degrees of freedom ( d.f .
equal to the number of cases minus 2 ) , and the probability of obtaining a worse value of @xmath39 assuming that the errors are accurate and a linear fit of the depletions @xmath70 $ ] to @xmath27 is a correct model .
values of these probabilities that are unreasonably close to 0 or 1 probably reflect unrealistically large or small errors assigned by the investigators .
an exception is cl , which seems to exhibit a genuinely poor linear relationship .
figure [ abundf1 ] shows how well the values of @xmath27 correlate with the average hydrogen densities along the lines of sight .
there is a clear trend between the two , which substantiates the statements made earlier in [ trends ] ( point nr . 3 ) .
the lack of any apparent separation between the darker and lighter shaded points indicates that for this sample , there is no clear effect that distinguishes paths toward stars in the lower halo from those well within the galactic plane ( but see [ halo ] ) .
savage & sembach ( 1996 ) summarized depletions of the elements mg , si , s , mn , cr , fe , and ni in terms of broad categories that were labeled `` warm disk '' and `` cool disk , '' in addition to identifying circumstances that included some gas arising from the galactic halo ( see their table 6 ) . as a point of reference , if one takes their listed depletions and evaluates the corresponding values of @xmath27 for all of the elements - values of the 1240 doublet determined by fitzpatrick ( 1997 ) .
likewise , abundances of ni should be increased by 0.27 dex ( fedchak & lawler 1999 ) ] , the warm disk gas is equivalent to @xmath71 while the cool disk material is equivalent to @xmath72 .
the rms dispersion of these two numbers from one element to the next is 0.05 . from strong deficiencies of al iii compared to s iii along certain lines of sight , howk & savage ( 1999 ) concluded that dust grains reside within ionized regions as well . as an aid to help us visualize the composition of dust grains ,
figure [ abundf5 ] indicates what fractions of the more abundant elements exist in the simplest characterizations of ( 1 ) resilient grain cores depicted by the lower , black portions of the bars , ( 2 ) the outer , somewhat interchangeable forms ( sometimes in mantles and other times as free atoms ) shown by the gray extensions , and ( 3 ) atoms perpetually in free gaseous form shown by the clear portions that extend to the top of each bar ( with a length equal to the total abundance ) .
the divisions between the three zones are based on evaluations of equation [ law ] for @xmath73 ( tops of the gray zones ) and @xmath74 ( the boundary between the gray and black zones ) .
the presumption here is that these two values of @xmath27 offer good representations of the quantities of material found in grains that are , respectively , either stripped or well nourished .
note from a previous statement in [ reinvestigation ] that it is not clear from the observations that c , n , and o show significant changes in depletion as @xmath27 progresses from one extreme to the other , so in truth the differences between the gray and black zones are not well established .
two elements shown in the figure were not studied here , primarily because there were problems with the ionization effects discussed in [ ionization ] : na is seen only as na i ( an unfavored stage of ionization ) in the visible , and significant contributions of s ii may arise from h ii regions .
sofia et al .
( 1994 ) , spitzer & fitzpatrick ( 1993 ) , and fitzpatrick & spitzer ( 1994 ) have considered the possible mix of compounds in grains that were consistent with the observed depletions ; the current state of this topic is reviewed by draine ( 2003 ) . in order to study the composition of gases in the lower halo ,
it is necessary to eliminate the inevitable foreground contributions arising from material in the galactic disk .
this can be accomplished by selecting velocity components that are displaced from that of the local gas by virtue of either differential galactic rotation or special kinematics ( usually infall , as with material that is often referred to as `` intermediate velocity clouds '' . ] ) . however , in this circumstance it is difficult to obtain useful information on the amount of hydrogen that accompanies this gas .
that is , @xmath0(h i ) derived from the ly@xmath19 damping wings applies to all of the hydrogen present , regardless of velocity .
measures of 21-cm emission show hydrogen as function of velocity , but the results are often not very satisfactory because the radio beam covers a solid angle in the sky that is much larger than a typical cloud size and thus may detect material that is not on top of the source viewed in the uv . also , some of the signal may arise from hydrogen beyond the target if it is a halo star and not a distant qso or agn .
one method of overcoming the lack of good information on hydrogen is to use some other element as a proxy , preferably one that is known to be very lightly depleted .
for instance , sembach & savage ( 1996 ) summarized the abundances of various elements relative to the corresponding amounts of zn in halo gas components along various sightlines .
those findings are reproduced in table [ halo_abund ] in the form of a value followed by the observed dispersion of results from one halo star to the next . for comparison
, the table also shows similar ratios for disk gas at two extremes for the depletion index , @xmath74 and 1.0 .
c c c c c c c c c c c c c c & & & & halo&&&rms&&&disk with&&&disk with + ratio&&&&value&&&dispersion&&&@xmath74&&&@xmath73 + & & & & @xmath75&&&@xmath76 & & & @xmath77&&&@xmath78 + [ si / zn ] & & & & @xmath79 & & & 0.14&&&@xmath80&&&@xmath81 + [ s / zn ] & & & & @xmath82 & & & 0.14 & & & & & & + [ cr / zn ] & & & & @xmath83 & & & 0.13&&&@xmath84&&&@xmath85 + [ mn / zn ] & & & & @xmath86 & & & 0.09&&&@xmath81&&&@xmath87 + [ fe / zn ] & & & & @xmath88 & & & 0.04&&&@xmath89&&&@xmath90 + [ ni / zn]&&&&@xmath91 & & & 0.07&&&@xmath92&&&@xmath93 + we must be cautious when interpreting the data shown in table [ halo_abund ] . the column densities of zn ii typically indicate that @xmath94 , which means that some of the gas could be partially photoionized , and this in turn would indicate that the column density comparisons may be subject to the ionization corrections discussed in [ ionization ] . nevertheless , even with the possibility that the apparent abundances may be perturbed in this manner , we can still formulate some useful conclusions .
for instance , we note from figure [ qplot_4 ] that fe ii is more easily ionized to higher stages than mg ii , regardless of the characteristic photon energy . however , according to the entries in table [ halo_abund ] , @xmath95}_{\rm halo}>{\rm [ fe / zn]}_{\rm disk}$ ] , while @xmath96}_{\rm halo } \approx { \rm [ mg / zn]}_{\rm disk}$ ] for @xmath74 .
it is therefore safe to say that fe is liberated from dust grains more easily in the halo than in the disk , regardless of the strength of the ionization ( or the value of @xmath27 adopted for the disk gas ) .
evidently , in the halo the amount of residual iron left in grain cores is not as large as in the most lightly depleted gas in the disk , i.e. , @xmath97}_{\rm halo } > a_{0,fe}=-0.95 $ ] .
this effect might arise from the condition that much of the gas in the halo was heavily shocked as it was ejected from the disk and thus the grains were more completely destroyed , or , alternatively , that perhaps ejecta from type ia supernovae in the halo have had a chance to enrich the fe in a time that is shorter than the circulation time between the disk and halo ( jenkins & wallerstein 1996 ) .
we now move on to examine the abundances in distinct cloud complexes at high galactic latitudes that are moving at extraordinarily large velocities ( oort 1966 ; mathewson , cleary , & murray 1974 ) , i.e. , with kinematics clearly not associated with differential galactic rotation .
a current and very comprehensive review of this material has been presented by wakker ( 2001 ) . here ,
our objective is not to assume some reference abundances and then study the condensation into dust grains , but instead , we seek to use abundances as a tool to gain a better understanding about the origin and history of the material .
however , we must still apply the concepts that we have considered earlier and account for the alterations caused by ionization effects and the removal of atoms as they condense into dust grains .
c c c c c c c c c & [ n i/&[o i/&[s ii/&[si ii/&[fe ii/&[ar i/&@xmath98 + object&h i]&h i]&h i]&h i]&h i]&h i]&@xmath99&refs .
+ fairall 9 ( ms ) & & & @xmath100&@xmath101& & &19.97&[1,2 ] + ngc 3783 ( ms ) & @xmath102 & & @xmath100 & @xmath103&@xmath104&@xmath105&19.90&[3,4 ] + mrk 279 ( c ) & @xmath106&@xmath107 & @xmath108 & @xmath109&@xmath103&@xmath110&19.49&[5,6 ] + mrk 290 ( c ) & & & @xmath111&@xmath112& & &19.96&[5,7 ] + mrk 817 ( c ) & @xmath113&@xmath107 & @xmath109 & @xmath107&@xmath100&@xmath114&19.48&[5,6 ] + mrk 876 ( c ) & @xmath115 & & & @xmath116&@xmath109&@xmath110&19.37&[5,6,8 ] + pg 1259 + 593 ( c ) & @xmath117&@xmath118 & @xmath103 & @xmath103&@xmath119 & @xmath120&19.92&[6,9 ] + 3c 351 ( c ) & @xmath114&@xmath121&@xmath122 & @xmath109&@xmath108&
&18.62&[10 ] + ioniz .
corr.&*[n / h]*&*[o / h]*&*[s / h]*&*[si / h]*&*[fe / h]*&*[ar / h ] * + 3c 351 ( c)&@xmath123&@xmath124&@xmath125&@xmath126&@xmath127& &&[10 ] + [ hivel_abund ] key to references : [ 1 ] gibson et al . 2000 , [ 2 ] lu , savage , & sembach 1994 , [ 3 ]
1998 , [ 4 ] sembach et al .
2001 , [ 5 ] gibson et al .
2001 , [ 6 ] collins , shull , & giroux 2003 , [ 7 ] wakker et al .
1999 , [ 8 ] murphy et al .
2000 , [ 9 ] richter et al .
2001b , [ 10 ] tripp et al .
the values of [ o i / h i ] listed by collins et al .
( 2003 ) have been adjusted to a reference solar abundance of 8.69 given by allende prieto et al .
( 2001 ) . for sightlines that penetrate through well - known high - velocity cloud complexes called complex c and the magellanic stream , table [ hivel_abund ] lists the observed deficiencies of the dominant ion @xmath128 of element @xmath4 in h i regions with respect to neutral hydrogen , @xmath129=\log\{n(x_i)/n({\rm h~i})\}-\log(x/{\rm h})_\odot,\ ] ] from which one may define a true depletion of the element @xmath130=[x_i/{\rm h~i}]+\log\{f({\rm h~i})/f(x_i)\}\ ] ] if the ion ratio @xmath131 can be calculated .
tripp et al . ( 2003 ) argue that for the complex c gas in front of 3c 351 the corrections of equation [ corr ] are substantial and are most likely to arise from collisional ionization .
the last row of table [ hivel_abund ] shows the element abundances after these corrections were made .
one can argue that the other four directions through complex c are probably less susceptible to ionization effects , particularly from photoionization , because their high - velocity gas components have much larger values of @xmath0(h i ) than that toward 3c 351 . while this may be true
, the fact that [ ar i / h i ] @xmath132 [ o i / h i ] for mrk 279 , mrk 817 , and pg 1259 + 593 warns us that ionization corrections for other elements , while probably small , are not completely inconsequential .
( of all the elements , ar is the most responsive one to the effects arising from partial photoionization . )
the complex c determinations shown in table [ hivel_abund ] indicate that the generally very mildly depleted species n and o ( see [ reinvestigation ] ) are substantially below their corresponding solar abundances .
the findings for o are particularly secure since their corrections arising from equation [ corr ] are very small ( e.g. , 0.03 dex the correction appears to be 0.1 dex , but this is due to roundoff error . ] for 3c 351 ) , principally because the ionization of o is strongly coupled to that of h through charge exchange reactions that have large rate constants ( see [ ionization ] ) .
the fact that n seems even more deficient than o [ even after applying the corrections , which amount to @xmath133 dex according to tripp et al .
( 2003 ) ] indicates that the gas in complex c has not only a low metallicity , but it has not undergone multiple generations of secondary element processing that are characteristic of the disk of our galaxy ( henry , edmunds , & kppen 2000 ) .
these two considerations disfavor the proposition that complex c arose from a recent ejection of gas from the galactic disk , even after some heavy dilution from more pristine material in the galactic halo or local group .
finally , the fact that [ fe / h ] is not substantially less than the values for other elements indicates that the fraction of elements locked into grains is much smaller than for gas within the galactic disk .
this finding is consistent with an apparent lack of h@xmath1 for material arising from complex c in front of pg 1259 + 593 ( richter et al .
2001a ) , since h@xmath1 is produced primarily on the surfaces of grains . a question may arise about a possible enhancement of fe over @xmath19-capture elements , but tripp et al .
caution that currently the errors in both the measurements and correction factors are too large to make a sound evaluation of this proposition .
the picture for the magellanic stream is not as complete as that for complex c. nevertheless , the pattern appears to be very similar , except for evidence that fe is deficient toward ngc 3783 , which in turn suggests that some dust is present [ further support for the presence of dust in the magellanic stream is suggested by the presence of h@xmath1 ( richter et al . 2001a ; sembach et al .
the evidence presented in table [ hivel_abund ] is consistent with the picture that the magellanic stream is material that has been tidally stripped from the magellanic clouds .
the topics discussed in the preceding sections all lead to interesting insights on the nature of processes that govern the apparent abundances of elements in the gas phase within and near our own galaxy .
a new and exciting frontier is the exploration of abundances in very distant systems ( prochaska , howk , & wolfe 2003 ; prochaska et al .
2003a , b ) , once again through the observations of absorption lines against continuum sources ( qsos ) , but ones at very large redshifts .
this is a productive way to understand the chemical state of systems that are otherwise invisible , owing to their great distances .
the lessons we have learned from studying the interstellar material in an environment that we know well ( our galaxy ) can be applied elsewhere . for instance , we may start with the assumption that the patterns of dust depletion are the same everywhere , regardless of the overall level of metallicity . at the very least ,
this assumption has been shown to hold true for galaxy and the lmc by vladilo ( 2002 ) .
provided one has measurements of two or more elements with very different indices @xmath34 and @xmath35 , but with somewhat similar production origins and responses to ionization , it should be possible to solve simultaneously the different forms of equation [ law ] to arrive at a representative @xmath27 .
once this has been done , the @xmath27 and observations of all of the elements may once again be substituted into equation [ law ] to derive the absolute abundances .
these , in turn , lead to our understanding of the chemical histories of such systems , along with key properties of their internal environments .
the preparation of this paper was supported by nasa contract nas5 - 30110 .
the author thanks b. t. draine , j. x. prochaska , t. m. tripp , b. d. savage , and d. welty for helpful comments on early drafts of this paper .
calura , f. , matteucci , f. , dessauges - zavadsky , m. , dodorico , s. , prochaska , j. x. , & vladilo , g. 2003 , in carnegie observatories astrophysics series , vol .
4 : origin and evolution of the elements , ed . a. mcwilliam & m. rauch ( cambridge : cambridge univ .
press ) , in press | a broad array of interstellar absorption features that appear in the ultraviolet spectra of bright sources allows us to measure the abundances and ionization states of many important heavy elements that exist as free atoms in the interstellar medium . by comparing these abundances with reference values in the sun , we find that some elements have abundances relative to hydrogen that are approximately consistent with their respective solar values , while others are depleted by factors that range from a few up to around 1000 .
these depletions are caused by the atoms condensing into solid form onto dust grains .
their strengths are governed by the volatility of compounds that are produced , together with the densities and velocities of the gas clouds .
we may characterize the depletion trends in terms of a limited set of parameters ; ones derived here are based on measurements of 15 elements toward 144 stars with known values of @xmath0(h i ) and @xmath0(h@xmath1 ) . in turn , these parameters may be applied to studies of the production , destruction , and composition of the dust grains .
the interpretations must be done with care , however , since in some cases deviations from the classical assumptions about missing atoms in unseen ionization stages can create significant errors .
our experience with the disk of our galaxy offers important lessons for properly unravelling results for more distant systems , such as high- and intermediate - velocity clouds in the galactic halo , material in the magellanic stream and magellanic clouds , and otherwise invisible gas systems at large redshifts ( detected by absorption features in quasar spectra ) .
inferences about the total ( gas plus dust ) abundances of such systems offer meaningful information on their origins and/or chemical evolution .
[ 1996/06/01 ] |
the benefits of cooperative work were explored by nature well before the advent of the human species as attested by the collective structures built by slime molds and social insects @xcite . in the human context
, the socio - cognitive niche hypothesis purports that hominin evolution relies so heavily on social elements to the point of viewing a band of hunter
gatherers as a ` group - level predator ' @xcite . whether physical or social , those structures and organizations may be thought of as the organisms solutions to the problems that endanger their existence ( see , e.g. , @xcite ) and have motivated the introduction of the concept of social intelligence in the scientific milieu @xcite . in the computer science circle ,
that concept prompted the proposal of optimization heuristics based on social interaction , such as the popular particle swarm optimization algorithm @xcite and the perhaps lesser - known adaptive culture heuristic @xcite . despite the prevalence of the notion that cooperation can aid a group of agents to solve problems more efficiently than if those agents worked in isolation , and of the success of the social interaction heuristics in producing near optimal solutions to a variety of combinatorial optimization problems , we know little about the conditions that make cooperative work more efficient than independent work .
in particular , we note that since cooperation ( or communication , in general ) reduces the diversity or heterogeneity of the group of agents , it may actually end up reducing the efficiency of group work @xcite .
efficiency here means that the time to solve a problem scales superlinearly with the number of individuals or resources employed in the task . in this contribution we study the performance of a group of cooperative agents following a first - principle research strategy to study cooperative problem solving set forth by huberman in the 1990s that consists of tackling easy tasks , endowing the agents with simple random trial - and - test search tools , and using plain protocols of collaboration @xcite .
here the task is to find the global maxima of three realizations of the nk - fitness landscape characterized by different degrees of ruggedness ( i.e. , values of the parameter @xmath0 ) .
we use a group of m agents which , in addition to the ability to perform individual trial - and - test searches , can imitate a model agent the best performing agent at the trial with a probability @xmath1 .
hence our model exhibits the two critical ingredients of a collective brain according to bloom : imitative learning and a dynamic hierarchy among the agents @xcite .
the model exhibits also the key feature of distributed cooperative problem solving systems , namely , the exchange of messages between agents informing each other on their partial success ( i.e. , their fitness at the current trial ) towards the completion of the goal @xcite .
we find that the presence of local maxima in the fitness landscape introduces a complex trade - off between the computational cost to solve the task and the group size @xmath2 .
in particular , for a fixed frequency of imitation @xmath1 , there is an optimal value of @xmath2 at which the computational cost is minimized .
this finding leads to the conjecture that the efficacy of imitative learning could be a factor determinant of the group size of social animals @xcite .
our study departs from the vast literature on cooperation that followed robert axelrod s 1984 seminal book the evolution of cooperation @xcite since in that game theoretical approach it is usually assumed a priori that mutual cooperation is the most rewarding strategy to the individuals . on the other hand , here we consider a problem solving scenario and a specific cooperation mechanism ( imitation ) aiming at determining in which conditions cooperation is more efficient than independent work .
the rest of the paper is organized as follows . in section [ sec : nk ] we offer a brief overview of the nk model of rugged fitness landscapes . in section [ sec : imi ] we present a detailed account of the imitative learning search strategy and in section [ sec : res ] we study its performance on the task of finding the global maxima of three realizations of the nk landscape with @xmath3 and ruggedness parameters @xmath4 , @xmath5 and @xmath6 .
the rather small size of the solution space ( @xmath7 binary strings of length @xmath8 ) allows the full exploration of the space of parameters and , in particular , the study of the regime where the group size is much greater than the solution space size .
finally , section [ sec : conc ] is reserved to our concluding remarks .
the nk model of rugged fitness landscapes introduced by kauffman @xcite offers the ideal framework to test the potential of imitative learning in solving optimization problems of varied degrees of difficulty , since the ruggedness of the landscape can be tuned by changing the two integer parameters @xmath9 and @xmath0 of the model . roughly speaking
, the parameter @xmath9 determines the size of the solution space whereas the value of @xmath10 influences the number of local maxima and minima on the landscape .
the solution space consists of the @xmath11 distinct binary strings of length @xmath9 , @xmath12 with @xmath13 . to each string
we associate a fitness value @xmath14 which is an average of the contributions from each component @xmath15 in the string , i.e. , @xmath16 where @xmath17 is the contribution of component @xmath15 to the fitness of string @xmath18 .
it is assumed that @xmath17 depends on the state @xmath19 as well as on the states of the @xmath0 right neighbors of @xmath15 , i.e. , @xmath20 with the arithmetic in the subscripts done modulo @xmath9 .
in addition , the functions @xmath21 are @xmath9 distinct real - valued functions on @xmath22 . as usual , we assign to each @xmath17 a uniformly distributed random number in the unit interval @xcite .
hence @xmath23 has a unique global maximum . for @xmath4
there are no local maxima and the sole maximum of @xmath24 is easily located by picking for each component @xmath15 the state @xmath25 if @xmath26 or the state @xmath27 , otherwise . for @xmath28 , finding the global maximum of the nk model is a np - complete problem @xcite , which essentially means that the time required to solve the problem using any currently known deterministic algorithm increases exponentially fast as the size @xmath9 of the problem grows @xcite .
the increase of the parameter @xmath0 from @xmath29 to @xmath30 decreases the correlation between the fitness of neighboring configurations ( i.e. , configurations that differ at a single component ) in the solution space and for @xmath31 those fitness values are uncorrelated so the nk model reduces to the random energy model @xcite . and @xmath4 ( _ circles _ ) , @xmath32 ( _ triangles _ ) and @xmath33 ( _ inverted triangles _ )
. the lines are guides to the eye.,scaledwidth=48.0% ] to illustrate the effect of varying @xmath0 on the ruggedness of the fitness landscape , in fig .
[ fig:1 ] we show the relative fitness of a string @xmath34 as function of its hamming distance @xmath35 to the global maximum for @xmath36 and different values of @xmath0 . here
@xmath37 stands for the fitness of the global maximum . for each @xmath0 , the figure shows the results for a single realization of the fitness landscape and for a single trajectory in the solution space that begins at the maximum ( @xmath38 ) and changes the state components sequentially until all @xmath9 states are reversed ( @xmath39 ) .
we note that the ruggedness of the landscapes ( essentially , the number of local maxima ) can vary wildly between landscapes characterized by the same values of @xmath9 and @xmath28 @xcite and so can the performance of any search heuristic based on the local correlations of the fitness landscape . hence in order to appreciate the role of the parameters that are relevant to imitative learning , namely , the group size @xmath2 and the imitation probability @xmath1 , for fixed @xmath9 and @xmath0 we consider a single realization of the nk fitness landscape .
we assume a group or system composed of @xmath2 agents .
each agent operates in an initial binary string drawn at random with equal probability for the digits @xmath29 and @xmath40 . in the typical situation
that the size of the solution space is much greater than the group size @xmath41 we can consider those initial strings as distinct binary strings , but here we will consider the case that @xmath42 as well , so that many copies of a same string are likely to appear in this initial stage of the simulation .
in addition , we assume that the agents operate in parallel . at any trial @xmath43 , each agent can choose with a certain probability between two distinct processes to operate on the strings .
the first process , which happens with probability @xmath44 , is the elementary or minimal move in the solution space , which consists of picking a string bit @xmath45 at random with equal probability and then flipping it .
the repeated application of this operation is capable of generating all the @xmath11 binary strings starting from any arbitrary string . the second process , which happens with probability @xmath1 , is the imitation of a model string .
we choose the model string as the highest fitness string in the group at trial @xmath43 . the string to be updated ( target string ) is compared with the model string and the different digits are singled out
. then the agent selects at random one of the distinct bits and flips it so that the target string becomes now more similar to the model string .
the parameter @xmath46 $ ] is the imitation probability and the case @xmath47 corresponds to the baseline situation in which the @xmath2 agents explore the solution space independently .
the specific imitation procedure proposed here was inspired by the mechanism used to model the influence of an external media @xcite in the celebrated agent - based model proposed by axelrod to study the process of culture dissemination @xcite .
it is important to note that in the case the target string is identical to the model string , and this situation is not uncommon since the imitation process reduces the diversity of the group , the agent executes the elementary move with probability one .
this procedure is different from that used in @xcite , in which strings identical to the model string are not updated within the imitation process .
however , both procedures yield qualitatively similar results , except in the regime where imitation is extremely frequent , i.e. , for @xmath48 . in particular , in the present implementation , a small group can find the global maximum in the case @xmath49 since the model string can explore the solution space through the elementary move whereas the other strings are simply followers .
the collective search ends when one of the agents finds the global maximum and we denote by @xmath50 the number of trials made by the agent that found the solution .
of course , @xmath50 stands also for the number of trials made by any one of the @xmath2 agents , since they operate in parallel and the search halts simultaneously for all agents .
in other words , the trial number @xmath43 is updated , or more specifically , incremented by one unit , when the @xmath2 agents have executed one of the two processes on its associated string .
we note that except for the case @xmath47 , the update of the @xmath2 agents is not strictly a parallel process since the model strings may change several times within a given trial .
nonetheless , since in a single trial all agents are updated , the total number of agent updates at trial @xmath43 is given by the product @xmath51 .
the efficiency of the search strategy is measured by the total number of agent updates necessary to find the global maximum ( i.e. , @xmath52 ) which can then be interpreted as the computational cost of the search .
in addition , since we expect that the typical number of trials to success @xmath50 scales with the size of the solution space @xmath11 , we will present the results in terms of the rescaled computational cost , defined as @xmath53 .
an interesting variant of our imitation rule is obtained by relaxing the condition that only the string with the highest fitness value can be imitated and allowing any string to pose as a model according to an imitation probability function proportional to the string relative fitness ( e.g. , a fermi function ) .
such fitness dependent imitation probability functions are frequently used in the study of cooperation dilemmas @xcite and may be useful to prevent the string population being trapped in the local maxima . in this paper , however , we will consider only the noiseless or zero - temperature limit of those imitation functions , where only the string with the maximum relative fitness can be imitated .
here we report the results of extensive monte carlo simulations of groups of agents that use imitative learning to search for the global maxima of three representative realizations of the nk fitness landscape . for fixed @xmath9 and @xmath0
we generate a single realization of the fitness landscape and carry out @xmath54 searches in order to determine the dependence of the mean rescaled computational cost @xmath55 on the imitation probability @xmath1 and on the group size @xmath2 .
in addition , for the baseline case in which the @xmath2 agents explore the landscape independently ( @xmath47 ) we derive an analytical expression for the computational cost . as function of the binary string length @xmath9 that determines the size of the solution space @xmath11 .
these results , which were obtained by the numerical calculation of the eigenvalues of the tridiagonal stochastic matrix @xmath56 , do not depend on the ruggedness of the landscape .
, scaledwidth=48.0% ] in this case there is no imitation ( @xmath47 ) and the ruggedness of the landscape has no effect on the performance of the search , which depends only on the size of the solution space , @xmath11 .
because of this independence on the landscape we can derive exact results for the time needed for @xmath2 independent agents to find the global maximum . for this analysis
we can assume that the global maximum is the string with the @xmath9 digit values equal to 1 , i.e. , @xmath57 , without loss of generality .
let us consider first the case of a single agent that operates on a string with @xmath58 digits 1 . according to the elementary move
, the probability that the resulting string has @xmath15 digits 1 is @xmath59 for @xmath60 , @xmath61 , and @xmath62 , where @xmath63 is the kronecker delta . since @xmath64 and @xmath65 for @xmath66 the matrix @xmath56 is a tridiagonal stochastic matrix .
the associated stochastic process has a single absorbing state @xmath67 and @xmath68 yields the probability that the agent finds the solution before or at trial @xmath43 . for large @xmath43
this quantity is given @xmath69 where @xmath70 is the second largest eigenvalue of @xmath56 .
( the largest eigenvalue is 1 because @xmath56 is a stochastic matrix . )
the probability that the agent finds the solution exactly at trial @xmath71 is @xmath72 , i.e. , @xmath73 which is a geometric distribution with success probability @xmath74 .
the value of @xmath70 can be easily obtained numerically ( see , e.g. , @xcite ) provided @xmath9 is not too large since @xmath75 is on the order of the unity and so it becomes practically impossible to distinguish @xmath70 from @xmath40 for @xmath76 .
the mean rescaled computational cost @xmath77 $ ] is shown in fig .
[ new:2 ] for different string lengths @xmath9 .
( recall that @xmath78 at this stage . ) in particular for @xmath3 we find @xmath79 that implies @xmath80 . for very large @xmath9 ,
the probability of success becomes @xmath81 .
we note that although eq .
( [ geom1 ] ) is valid strictly for large @xmath50 only , the fact that @xmath50 is typically on the order of @xmath11 makes this geometric distribution an exceedingly good approximation to the correct distribution of absorbing times .
now let us consider the case of @xmath82 agents searching independently for the global maximum .
since the process halts when one of the agents finds the global maximum , the halting time is @xmath83 where @xmath84 and @xmath85 are independent random variables distributed by the geometric distribution ( [ geom1 ] ) .
it is easy to show that the distribution of @xmath86 is also geometric with probability of success @xmath87 @xcite . in the case of @xmath88 agents the halting time @xmath89 and since
both @xmath86 and @xmath84 are geometrically distributed , though with distinct success probabilities , we find that @xmath90 is also geometrically distributed with probability of success @xmath91 .
the generalization of this reasoning to @xmath2 agents yields that the mean scaled cost is @xmath92}.\ ] ] since @xmath93 is close to @xmath40 we can write @xmath94 so that @xmath95 $ ] for @xmath96 and @xmath97 for @xmath98 .
we recall that for @xmath3 we have @xmath99 . as expected , eq .
( [ cind ] ) matches the simulation data very well ( see fig .
[ fig:2 ] ) . as function of the group size @xmath2 for the imitation probability @xmath47 ( _ circles _ ) , @xmath100 ( _ triangles _ ) , @xmath101 ( _ inverted triangles _ ) , and @xmath49 ( _ squares _ ) .
the solid curve is eq .
( [ cind ] ) and the dashed line is the linear function @xmath102 .
the parameters of the nk landscape are @xmath3 and @xmath4 .
the landscape exhibits a single maximum .
, scaledwidth=48.0% ] the nk fitness landscape with @xmath4 exhibits a single global maximum and no local maxima .
the results of the performance of the imitative search for a landscape with @xmath3 and @xmath4 are summarized in fig .
[ fig:2 ] .
as shown in the previous subsection , the mean computational cost for non - interacting agents ( @xmath47 ) is a constant @xmath103 provided the group size @xmath2 is not too large compared to @xmath99 , which is close to the size of the solution space , @xmath7 . in the case
the group size @xmath2 is very large , the agents begin to duplicate each other s work leading to a linear increase of @xmath104 with increasing @xmath2 .
more pointedly , in this regime we find @xmath102 ( see the straight line in fig .
[ fig:2 ] ) .
we stress that a constant computational cost means that the time @xmath50 the group requires to find the global maximum decreases with the inverse of the group size ( i.e. , the time to solve the problem decreases linearly with the number of agents ) , whereas a computational cost that grows linearly with the group size means that @xmath50 does not change as more agents are added to the group . allowing the agents to imitate the best performer ( model ) at each trial
leads to a rapid reduction of the computational cost provided the group size remains small .
the best performance is achieved for @xmath105 and @xmath49 and corresponds to a fiftyfold decrease of the computational cost with respect to the independent search ( @xmath47 ) .
the fact that the best performance is obtained when the imitation probability is maximum is due to the absence of local maxima in the landscape for @xmath4 .
we recall that for @xmath49 only the model string , which is likely to be represented by several copies in the group , can perform the elementary move ; all other strings must imitate the model . as a result , the effective size of the search space is greatly reduced , i.e. , the strings are concentrated in the vicinity of the model string which can not accommodate many more than @xmath106 strings without duplication of work .
this is the reason we observe the degradation of the performance when the group size increases beyond its optimal value .
note that for @xmath4 the imitative learning search always performs better than the independent search .
now we consider a somewhat more complex nk fitness landscape by setting @xmath3 and @xmath5 . in the particular realization of the landscape
studied here there are 5 maxima in total , among which 4 are local maxima .
the relative fitness of those maxima ( @xmath107 ) , as well as their hamming distances to the global maximum ( @xmath35 ) , is presented in table [ table : k2 ] .
.local maxima for the studied realization of the nk fitness landscape with @xmath3 and @xmath5 [ cols="^,^,^ " , ] [ table : k2 ] the results of the imitative learning search are summarized in fig .
[ fig:3 ] where the mean rescaled computational cost is plotted against the group size for different values of the imitation probability .
the performance of the independent search ( @xmath47 ) is identical to that shown in fig .
[ fig:2 ] for @xmath4 , since that search strategy is not affected by the complexity of the landscape .
the results for the cooperative search ( @xmath108 ) , however , reveal a trade - off between the group size @xmath2 and the imitation probability @xmath1 .
in particular , for @xmath109 we observe a steep increase of the computational cost for intermediate values of @xmath2 .
this happens because the group can be trapped near one of the local maxima . for large groups ( @xmath110 in this case )
, chances are that some of the initial strings are close to the global maximum and end up attracting the rest of the group to its neighborhood , thus attenuating the effect of the local maxima .
the robust finding is that for any @xmath108 there is an optimal group size , which depends on the value of @xmath1 , that minimizes the computational cost of the search . as function of the group size @xmath2 for the imitation probability @xmath100 ( _ triangles _ ) , @xmath101 ( _ inverted triangles _ ) , @xmath111 ( _ squares _ ) and @xmath112 ( _ circles _ ) .
the solid line is the linear function @xmath102 .
the parameters of the nk landscape are @xmath3 and @xmath5 .
the landscape exhibits 4 local maxima and a single global maximum.,scaledwidth=48.0% ] agents and probability of imitation @xmath112 .
the thin horizontal lines indicate the relative fitness of the 4 local maxima given in table [ table : k2 ] .
the parameters of the nk landscape are @xmath3 and @xmath5.,scaledwidth=48.0% ] a better understanding of the dynamics of the search is offered in fig .
[ fig:4 ] which shows the relative fitness of the model string as function of the number of trials for @xmath88 and @xmath112 .
the role of the local maxima as transitory attractors of the search is evident in this figure . a typical run with @xmath113 agents , which is approximately the location of the maximum of @xmath114 in fig . [ fig:3 ] , yields essentially a flat line at the highest fitness local maximum ( maximum 4 in table [ table : k2 ] ) and a sudden jump to the global maximum .
because there are many copies of the model string ( the mean hamming distance between the @xmath113 strings is less than 1 ) , the variants produced by the elementary move are attracted back to the local maximum .
this is the reason why a small group can search the solution space much more efficiently than a large one in the case imitation is very frequent .
the particular realization of the nk fitness landscape with @xmath3 and @xmath6 that we consider now has 53 maxima , including the global maximum , which poses a substantial challenge to any hill - climbing type of search strategy . as in the previous case , we observe in fig .
[ fig:5 ] a trade - off between @xmath2 and @xmath1 , but now the results reveal how imitative learning may produce disastrous performances on rugged landscapes for certain ranges of those parameters . the strategy of following or imitating a model string can backfire if the fitness landscape exhibits high fitness local maxima that are distant from the global maximum .
a large group may never escape from those maxima due to the attractive effect of the clones of the model string .
this is what we observed in the case of @xmath101 for which we were unable to find the global maximum with groups of size @xmath115 .
as function of the group size @xmath2 for the imitation probability @xmath116 ( _ circles _ ) , @xmath117 ( _ squares _ ) , @xmath100 ( _ triangles _ ) and @xmath101 ( _ inverted triangles _ ) .
for @xmath101 we find @xmath118 for @xmath115 .
the solid line is the linear function @xmath102 .
the parameters of the nk landscape are @xmath3 and @xmath6 .
the landscape exhibits 52 local maxima and a single global maximum.,scaledwidth=48.0% ] as function of the probability of imitation @xmath1 for the group size @xmath105 ( solid line ) , @xmath113 ( dashed line ) and @xmath119 ( dotted line )
. the parameters of the nk landscape are @xmath3 and @xmath6 .
the landscape exhibits 52 local maxima and a single global maximum.,scaledwidth=48.0% ] we note , however , that for a fixed group size @xmath2 it is always possible to tune the imitation probability @xmath1 so that the imitative learning strategy performs better than ( or , in a worst - case scenario , equal to ) the independent search .
this point is illustrated in fig .
[ fig:6 ] that shows the computational cost as function of @xmath1 . for any fixed value of @xmath2 , the computational cost exhibits a well - defined minimum that determines the value of the optimal imitation frequency .
as hinted in the previous figures , this optimal value decreases with increasing group size . in order to verify the robustness of our findings ,
which were obtained for specific realizations of the nk fitness landscape , we have considered four random realizations of the landscape with @xmath3 and @xmath6 in addition to the realization studied above .
the comparison between the mean computational costs to find the global maxima of the five realizations of the landscape is shown in fig .
[ fig:8 ] for the imitation probability @xmath100 .
the results are qualitatively the same , despite the wild fluctuations of @xmath114 in the regime where the search is trapped in the local maxima .
it is reassuring to note that the initial decrease of the mean cost with increasing group size and the existence of an optimal group size that minimizes that cost , which are exhibited by all five realizations of the nk landscapes shown in fig .
[ fig:8 ] , are robust properties of the imitative learning search . for the probability of imitation @xmath100 as function of the group size @xmath2 for five realizations ( different symbols ) of the nk fitness landscape with @xmath3 and @xmath6 .
the solid line is the linear function @xmath102 .
, scaledwidth=48.0% ]
in this paper we study quantitatively the problem solving performance of a group of agents capable to learn by imitation .
the performance measure we consider is the computational cost to find the global maximum of three specific realizations of the nk fitness landscape .
the computational cost is defined as the product of the number of agents in the group and the number of attempted trials till some agent finds the global maximum .
our main conclusion , namely , that for a fixed probability of imitation @xmath1 there is a value of group size that minimizes the computational cost corroborates the findings of a similar study in which the task was to solve a particular cryptarithmetic problem @xcite . hence our conjecture that the efficacy of imitative learning could be a factor determinant of the group size of social animals ( see @xcite for a discussion on the standard selective pressures on group size in nature ) .
we note that in the case the connectivity between agents is variable , i.e. , each agent interacts with @xmath120 distinct randomly picked agent ( here we have focused on the fully connected network @xmath121 only ) then there is an optimal connectivity value that minimizes the computational cost @xcite .
it would be most interesting to understand how the network topology influences the performance of the group of imitative agents and how the optimal network topology correlates with the known animal social networks @xcite .
the main aim of our contribution is to show that the existence of an optimal group size that maximizes performance for imitative learning @xcite is insensitive to the choice of the fitness landscape , so it is likely a robust property of populations that use imitation as a cooperative strategy .
although we have focused on the effect of the parameter @xmath0 , which roughly determines the ruggedness of the nk landscape , the parameter @xmath9 ( the length of the strings ) also plays a relevant role on the search for the global maximum , besides the obvious role of fixing the size @xmath11 of the search space .
( of course , since the typical time to find the global maximum scales with @xmath11 , even moderate values of @xmath9 , say @xmath122 , would render the simulations unfeasible . )
the nontrivial role is that the value of @xmath9 seems to pose an upper bound to the optimal size of the group @xmath2 in the regime that imitation is very frequent ( see figs .
[ fig:2 ] , [ fig:3 ] and [ fig:5 ] ) .
this is so because in that regime the strings are concentrated in the close vicinity of the model string , which can not accommodate more than @xmath123 different strings .
we do not purport to offer here any novel method to explore efficiently rugged landscapes , but the finding that for small group sizes imitative learning decreases considerably the computational cost of the search , even in a very rugged landscape ( see data for @xmath101 in fig . [ fig:5 ] ) motivates us to address the question whether in such landscapes that strategy could achieve a better - than - random performance for all group sizes .
this is achieved automatically for smooth landscapes ( see figs .
[ fig:2 ] and [ fig:3 ] ) but not for rugged ones ( see fig . [ fig:5 ] and @xcite ) .
clearly , the way to accomplish this aim is to decrease the probability of imitation @xmath1 as the group size @xmath2 increases , following the location of the minima shown in fig .
[ fig:6 ] .
it is interesting to note that the finding that too frequent interactions between agents may harm the performance of the group ( see fig .
[ fig:6 ] ) may offer a theoretical justification for henry ford s factory design in which the communication between workers was minimized in order to maintain the established efficiency and maximal productivity @xcite . | the idea that a group of cooperating agents can solve problems more efficiently than when those agents work independently is hardly controversial , despite our obliviousness of the conditions that make cooperation a successful problem solving strategy .
here we investigate the performance of a group of agents in locating the global maxima of nk fitness landscapes with varying degrees of ruggedness .
cooperation is taken into account through imitative learning and the broadcasting of messages informing on the fitness of each agent .
we find a trade - off between the group size and the frequency of imitation : for rugged landscapes , too much imitation or too large a group yield a performance poorer than that of independent agents . by decreasing the diversity of the group , imitative learning may lead to duplication of work and hence to a decrease of its effective size .
however , when the parameters are set to optimal values the cooperative group substantially outperforms the independent agents . |
the integer quantized hall effect ( iqhe ) , observed at two dimensional charge systems ( 2dcs ) subject to strong perpendicular magnetic fields @xmath0 , is usually discussed within the single particle picture , which relies on the fact that the system is highly disordered @xcite .
these quantized ( spinnless ) single particle energy levels are called the landau levels ( lls ) and the discrete energy values are given by @xmath1 , where @xmath2 is the landau index and @xmath3 is the cyclotron frequency of an electron with an effective mass @xmath4 ( @xmath5 , @xmath6 being the bare electron mass at rest ) and @xmath7 is the speed of light in vacuum . in single particle models
the disorder plays several roles , such as landau level broadening @xcite , leading to a finite longitudinal conductivity @xcite , spatial localization @xcite _ etc_. disorder can be created by inhomogeneous distribution of dopant ions which essentially generates the confinement potential @xcite for the electrons . in the absence of disorder ,
the density of states are dirac delta - functions @xmath8 , where @xmath9 is the magnetic length , and the longitudinal conductivity ( @xmath10 ) vanishes . for a homogeneous two dimensional electron system ( 2des ) , by the inclusion of disorder and due to collisions ,
lls become broadened .
therefore the longitudinal conductance becomes non - zero in a finite energy ( in fact magnetic field ) interval .
long range potential fluctuations generated by the disorder result in the so called _ classical localization _
@xcite , _ i.e. _ the guiding center of the cyclotron orbit moves along closed equi
potentials @xcite .
in contrast to the above mentioned bulk theories , the edge theories usually disregard the effect of disorder to explain the ( quantized ) hall resistance @xmath11 and accompanying ( zero ) longitudinal resistance @xmath12 . however , the non - interacting edge theories still require disorder to provide a reasonable description of the transition between the plateaus . the landauer - bttiker approach ( known as the edge channel picture )
@xcite and its direct coulomb interaction generalized version , _
i.e. _ the non - self - consistent chklovskii picture @xcite , also needs localization assumptions in order to obtain quantized hall ( qh ) plateaus of finite width ( see for a review _
e.g. _ datta s book @xcite and ref .
@xcite for the estimates of plateau widths at the high disorder limit ) .
in contrast to above discussions very recent experimental @xcite and theoretical @xcite results point the incomplete treatment of the disorder potential and scattering mechanisms .
fairly recent theoretical approaches @xcite , the qh plateaus are obtained by the inclusion of direct coulomb interaction self - consistently @xcite and the effect of the disorder was handled via conductivity tensor elements @xcite , however , the source of the disorder and its properties was left unresolved @xcite . whereas , the influence of potential fluctuations on the qh plateaus were discussed briefly @xcite .
this work provides a systematic investigation of the disorder potential and its influence on the quantized hall effect including direct coulomb interaction .
the investigation is extended to realistic experimental conditions in determining the widths of the quantized hall plateaus .
we , essentially study the effect of disorder in two distinct regimes , namely the short range and the long range .
the short range part is included to the density of states ( dos ) , thereby influences the widths of the current carrying edge - states and the entries of the conductivity tensor .
whereas , the long range part is incorporated to the self - consistent calculations , which determines the extend of the plateaus in turn . in sec [ impurity ]
we introduce two types of single impurity potentials , namely the coulomb and the gaussian , and compare their range dependencies considering damping of the dielectric material . in the next step
we discuss the _ screened _ disorder potential within a pure electrostatic approach , by considering an homogeneous two dimensional electron system ( 2des ) without an external magnetic field and show that the long range part is well screened , whereas the short range part is almost unaffected .
section [ treed ] is devoted to investigate impurities numerically , where we solve the poisson equation self - consistently in three dimensions .
the numerical and analytical calculations are compared , considering the estimations of the disorder potential range and its variation amplitude .
we finalize our discussion with sec .
[ quanhp ] , where we calculate the plateau widths under experimental conditions for different sample widths and mobilities .
the disorder potential experienced by the 2des , resulting from the impurities has quite complicated range dependencies . since , the potential generated by an impurity is ( i ) _ damped _ by the dielectric material in between the impurity and the plane where the 2des resides ( ii ) is screened by the homogeneous 2des depending on the density of states , which changes drastically with and without magnetic field .
it is common to theoreticians to calculate the conductivities from single impurity potentials , such as lorentzian @xcite , gaussian @xcite or any other analytical functions @xcite .
however , the landscape of potential fluctuations is also important to define the actual mobility of the sample at hand , in particular in the presence of an external magnetic field .
we first discuss the different range dependencies of the coulomb and gaussian donors , assuming open boundary conditions .
next , the effect of the spacer thickness on the disorder potential is discussed , namely the damping of the external ( coulomb ) potential , and is compared with the thomas - fermi screening .
the different damping / screening dependencies of the resulting potentials are discussed in terms of range .
the coulomb potential presents long range part , which leads to long range fluctuations due to overlapping if several donors are considered . whereas , the gaussian potential decays exponentially on the length scale comparable with the separation thickness .
since the gaussian potential is relatively short ranged , no overlapping of the single donor potentials occur .
hence , the external potential experienced by the electrons can be approximated to a homogeneous potential fairly good .
thus one can conclude that approximating the total disorder potential by gaussians is not sufficient to recover the long range part .
similar arguments are also found in the literature @xcite . in order to overcome the difference observed at the long range potential fluctuations between the coulomb and the gaussian impurities , the following procedure is applied :
first we calculate the total disorder potential considering many impurities then we perform a two - dimensional fourier transformation of the coulomb potential and make a back transformation keeping the first few momentum @xmath13 components in each direction , hence only the long range part of the potential is left @xcite . then we add the long range part of the coulomb potential to the potential created by donors , _
i.e. _ to the confinement potential .
we take this as a motivation to simulate the short range part of the impurity potential by gaussian impurities , and calculate the landau level broadening and the conductivities , described within the self - consistent born approximation ( scba ) @xcite . here
we point to the effect of the spacer thickness on the impurity potential experienced in the plane of 2des .
it is well known from experimental and theoretical investigations that , if the distance between the electrons and donors is large , the mobility is relatively high and it is usually related with suppression of the short range fluctuations of the disorder potential .
these results agree with the experimental observations of high mobility samples and are easy to understand from the @xmath14 dependence of the fourier expansion of the coulomb potential , @xmath15where @xmath16 contains all the information about the in - plane donor distribution and @xmath17 is the total number of the ionized donors .
we observe that if the spacer thickness is increased , the amplitude of the potential decreases rapidly .
we also see that the short range potential fluctuations , which correspond to higher order fourier components , are suppressed more efficiently .
next , we discuss electronic screening of the external potential created by the donors discussed above . for a dielectric material
the relation between the external and the screened potentials are given by , @xmath18where @xmath19 is the dielectric _ function _ and is given by @xmath20 with the constant 2d density of states @xmath21 in the absence of an external @xmath0 field , and is known as the thomas - fermi ( tf ) function .
the simple linear relation above , together with the tf dielectric function essentially describes the electronic screening of the coulomb potential given in eq .
[ fourier1 ] , if there are sufficient number of electrons @xcite ( @xmath22 m@xmath23 ) .
consider a case where the @xmath13 component approaches to zero , then the external ( damped ) potential is well screened , hence the long range part of the disorder potential . whereas , the short range part remain unaffected , _ i.e. _ high @xmath13 fourier components .
now we turn our attention to the second type of impurities considered , the gaussian ones . as well known ,
the fourier transform of a gaussian is also of the form of a gaussian , therefore , similar arguments also hold for this kind of impurity
. we should emphasize once more the clear distinction between the effect of the spacer on the external potential and the screening by the 2des , _ i.e. _ via @xmath19 .
the former depends on the fourier transform of the coulomb potential and the important effect is the different decays of the different fourier components ( see eq . [ fourier1 ] ) , so that the short range part of the disorder potential is well dampened , whereas the latter depends on the relevant dos of the 2des and the screening is more effective for the long range part .
we continue our investigation by solving the 3d poisson equation iteratively for randomly distributed single impurities , where three descriptive parameters ( _ i.e. _ the number of impurities , the amplitude of the impurity potential and the separation thickness ) are analyzed separately .
next , we discuss the long range parts of the potential fluctuations investigating the coulomb interaction of the 2des , numerically . the range is estimated from these investigations by performing fourier analysis and is related to the samples used in experiments @xcite ( sec .
[ exper ] ) . in the previous section we took a rather simple way to study the effect of interactions by assuming an homogeneous 2des and screening
is handled by the tf dielectric function . here , we present our results obtained from a rather complicated numerical method .
we solve the poisson equation in 3d starting from the material properties of the wafer at hand , the typical material we consider is sketched in fig .
[ fig:5 ] .
namely , using the growth parameters , we construct a 3d lattice where the potential and the charge distributions are obtained iteratively assuming open boundary conditions , _
i.e. _ @xmath24 . for such boundary conditions , we chose a lattice size which is considerably larger than the region that we are interested in .
we preserve the above conditions within a good numerical accuracy ( absolute error of @xmath25 ) .
a forth order grid approach @xcite is used to reduce the computational time , which is successfully used to describe similar structures @xcite .
figure [ fig:5 ] presents the schematic drawing of the hetero - structure which we are interested in .
the donor layer is @xmath26 doped by a density of @xmath27 m@xmath23 ( ionized ) silicon atoms , @xmath28 nm above the 2des , which provide electrons both for the potential well at the interface and the surface .
it is worthwhile to note that most of the electrons ( @xmath29 ) escape to the surface to pin the fermi energy to the mid - gap of the gaas . in any case , for such wafer parameters there are sufficient number of electrons ( @xmath30 m@xmath23 ) at the quantum well to form a 2des . to investigate the effect of impurities we place
positively charged ions at the layer where donors reside . from eq .
[ fourier1 ] we estimate the amplitude of the potential of a single impurity to be @xmath31 ev and assume that _ some _ percent of the ionized donors are generating the disorder potential , that defines the long range fluctuations . in our simulations
we perform calculations for a unit cell with areal size of @xmath32m@xmath33 m which contains @xmath34 donors per square meters , thus with 10 percent disorder we should have @xmath35 3300 impurities . , title="fig : " ] , title="fig : " ] figure [ fig:7]a depicts the actual density distribution , when considering 3300 impurities , whereas fig .
[ fig:7]b presents only the long range part of the density fluctuation .
the arrows show the average distance between two maxima , which is calculated approximately to be @xmath36 nm . to estimate an average range of the disorder potential
, we repeated calculations for such randomly distributed impurities , where number of repetitions scales with @xmath37 .
such a statistical investigation , sufficiently ensembles the system to provide a reasonable estimation of the long range fluctuations .
we also tested for larger number of random distributions , however , the estimation deviated less than tens of nanometers .
we show our main result of this section in fig .
[ fig:8 ] , where we plot the estimated long range part of the disorder potential considering various number of impurities @xmath38 and impurity potential amplitude @xmath39 .
our first observation is that the long range part of the total potential becomes less when @xmath38 becomes large , not surprisingly .
however , the range increases nonlinearly while decreasing @xmath38 , obeying almost an inverse square law and tend to saturate at highly disordered system . when fixing the distributions and @xmath38 , and changing the amplitude of the impurity potential we observe that for large amplitudes the range can differ as large as 200 nm at all impurity densities .
we found that for impurity concentration less than @xmath40 , the range of the potential is larger than the unit cell we consider , _
i.e _ @xmath41 m . in contrast to the long range part ,
the short range part is almost unaffected by the impurity concentration , however , is affected by the amplitude . therefore , while defining the conductivities we will focus our investigation on @xmath39 .
another important result is that the estimates of long range fluctuations does not depend strongly on the spacer thickness , if one keeps the amplitude of single impurity potential amplitude fixed , fig .
[ fig:8]b .
all of the above numerical observations coincide fairly good with our analytical investigations in the previous section .
however , the range dependency on the impurity concentration can not be estimated with the analytical formulas given .
we should also note that , similar or even complicated numerical calculations are present in the literature @xcite .
a indirect measure of the screening effects on the potential can also be inferred by capacitance measurements , supported by the above calculation scheme in the presence of external field @xcite
. 5 disorder .
[ fig:8],title="fig : " ] 5 disorder .
[ fig:8],title="fig : " ] next section is devoted to investigate the widths of the quantized hall plateaus utilizing our findings .
we consider mainly two `` mobility '' regimes , where the long range fluctuations is at the order of microns ( high mobility ) and is at the order of few hundred nanometers , low mobility .
however , the amplitude of the total potential fluctuations will be estimated not only depending on the number of impurities but also depending on the spacer thickness , range and amplitude of single impurity potential .
the main aim of this section is to provide a systematic investigation of the quantized hall plateau ( qhp ) widths within the screening theory of the iqhe @xcite , therefore here we summarize the essential findings of the mentioned theory . in calculating the qhps one needs to know local conductivities , namely the longitudinal @xmath42 and the transverse @xmath43 . to determine these quantities it is required to relate the electron density distribution @xmath44 to the local conductivities explicitly .
here we utilize the scba @xcite .
however , the calculation of the electron density and the potential distribution including direct coulomb interaction is not straightforward , one has to solve the schrdinger and the poisson equations simultaneously .
this is done within the thomas - fermi approximation which provides the following prescription to calculate the electron density @xmath45where @xmath46 is the appropriate density of states calculated within the scba , where @xmath47 is the boltzmann constant and @xmath48 temperature .
the total potential is obtained from @xmath49and the kernel @xmath50 is the solution of the poisson equation satisfying the boundary conditions to be discussed next . in the following
we assume a translation in variance in @xmath51-direction and implement the boundary conditions @xmath52 ( @xmath53 being the sample width ) , proposed by chklovskii _ et.al . _
@xcite , such a geometry allows us to calculate the kernel in a closed form .
hence , eqs .
( [ density ] ) and ( [ potential ] ) forms the self - consistency . for a given initial potential distribution
, the electron concentration can be calculated at finite temperature and magnetic field , where the density of states @xmath46 contains the information about the quantizing magnetic field and the effect of short range impurities . here
we implicitly assume that the electrons reside in the interval @xmath54 ( where , @xmath55 is called the depletion length ) , and is fixed by the fermi energy , _
i.e. _ the number of electrons , hence donors . as a direct consequence of landau quantization and the locally varying electrostatic potential ,
the electronic system is separated into two distinct regions , when solving the above self - consistent equations iteratively : i ) the fermi energy equals to ( spin degenerate ) landau energy and due to dos the system illustrates a metallic behavior , the compressible region , ii ) the insulator like incompressible region , where @xmath56 falls in between two consequent eigen - energies and no states are available @xcite .
it is usual to define the filling factor @xmath57 , to express the electron density in terms of the applied @xmath0 field as , @xmath58 .
since all the states below the fermi energy are occupied the filling factor of the incompressible regions correspond to integer values ( _ e.g. _ @xmath59 ) , whereas the compressible regions have non - integer values , due to partially occupied higher most landau level . the spatial distribution and widths of these regions
are determined by the confinement potential @xcite , magnetic field @xcite , temperature @xcite and level broadening @xcite . for the purpose of the present work we fix the confinement potential profile by confining ourselves to the chklovskii geometry and keeping the donor concentration ( and distribution ) constant .
moreover we perform our calculations at a default temperature given by @xmath60 , where @xmath61 is the fermi energy calculated for the electron concentration at the center of the sample and is typically similar to 10 mev .
the next step is to calculate the global resistances , _
i.e. _ the longitudinal @xmath12 and hall @xmath11 resistances , starting from the local conductivity tensor elements .
such a calculation is done within a relaxed local model that relates the current densities @xmath62 to the electric fields @xmath63 , namely the local ohm s law : @xmath64the strict locality of the conductivity model is lifted by an spatial averaging process @xcite over the quantum mechanical length scales and an averaged conductivity tensor @xmath65 is used to obtain the global resistances .
it should be emphasized that , such an averaging process also simulates the quantum mechanical effects on the electrostatic quantities . to be explicit :
if the widths of the current carrying incompressible strips become narrower than the extend of the wave functions , these strips become `` leaky '' which can not decouple the two sides of the hall bar and back - scattering takes place . therefore , to simulate the leakiness " of the incompressible strips we perform coarse - graining over quantum mechanical length scales .
now let us relate the local conductivities with the local filling factors . since the compressible regions behave like a metal within these regions there is finite scattering leading to finite conductivity .
in contrast , within the incompressible regions the back - scattering is absent , hence , the longitudinal conductivity ( and simultaneously resistivity ) vanishes .
therefore , all the imposed current is confined to these regions .
the hall conductivity , meanwhile is just proportional to the local electron density .
the explicit forms of the conductivity tensor elements are presented elsewhere @xcite .
having the electron density and local magneto - transport coefficients at hand , we perform calculations to obtain the widths of the quantized hall plateaus utilizing the above described , microscopic model assisted by the local ohm s law at a fixed external current @xmath66 .
further details of the calculation scheme is reviewed in ref .
@xcite .
since the very early days of the charge transport theory , collisions played an important role .
such a scattering based definition of conduction also applies for the system at hand , _
i.e. _ a two - dimensional electron gas subject to perpendicular magnetic field . among many other approaches @xcite the scba emerged as a reasonable model to describe the dos assuming gaussian impurities , considering short range scattering .
a single impurity has two distinct parameters that represents the properties of the resulting potential , the range @xmath67 ( at the order of separation thickness ) and the amplitude of the potential ( in relevant units ) , @xmath68 .
however , these two parameters are not enough to define the widths of the landau levels ( @xmath69 ) , another important parameter is the number of the impurities , @xmath38 . in the previous section
we have already investigated these three parameters in scope of potential landscape , now we utilize our findings to define the level widths and the conductivities .
it is more convenient to write the single impurity potential of the form , @xmath70 together with the impurity concentration , the relaxation time is defined as @xmath71 and in the limit of delta impurities ( _ i.e. _ @xmath72 ) the landau level width @xmath69 takes the form @xmath73 .
it is useful to define the impurity strength parameter to investigate the effect of disorder by @xmath74 given in units of magnetic energy @xmath75 and as a normalization parameter we fix the magnetic energy at 10 t. at this point we would like to make a remark on the concepts short / long range impurities and short / long range potential fluctuations , which is commonly mixed . by short range impurity potential we mean that @xmath76 , however , by short range potential fluctuation a length scale of the order of @xmath77 nm is meant .
the long range impurity potential corresponds to @xmath78 and long range potential fluctuation is of the order of micrometers .
thus , when considering short range impurities the potential fluctuations may be long range , if @xmath38 is not large ( @xmath79 of the donor concentration ) . we have also observed that , the long - range potential fluctuations are more efficiently screened by the 2des and their range can be at the order of 500 nm at most , when assuming large impurity concentration ,
_ i.e. _ @xmath80 . in light of the above findings and formulation we now investigate the widths of the quantized hall plateaus .
figure [ fig:9 ] presents the calculated hall resistances at a fixed temperature for typical single impurity ranges .
we observe that , when increasing @xmath67 the plateau widths remain approximately the same , with a small variation , which is in contrast to the experimental findings , _
i.e. _ if the system is low mobility ( small @xmath81 highly broadened dos ) the plateau are larger .
in fact changing @xmath67 from 10 nm to 20 nm should increase the zero @xmath0 field mobility almost an order of magnitude , when fixing the other parameters ( see _ e.g _ table i of ref .
the contradicting behavior is due to the fact that the levels become broader when increasing the single impurity range , therefore the incompressible strips become narrower , which results in a narrower plateau .
however , the long range potential fluctuations are completely neglected , therefore the effect(s ) of disorder on the quantized hall plateaus can not be described in a complete manner . to investigate the effect of the single impurity range we systematically calculated the plateau widths ; table [ table:1 ]
depicts the calculated widths of the hall plateaus considering different sample widths , depletion lengths , filling factors and @xmath67 .
one sees that the plateau widths are affected by the increase of impurity range , however , in a completely wrong direction , _
i.e. _ plateaus become narrower when decreasing the mobility . as we show in the next section , it is not sufficient to describe mobility only considering the range of a single impurity .
moreover , we also show that the other two parameters defining @xmath82 mobility are either not important or behaves in the opposite direction when calculating the resistances .
next we investigate the effect of the remaining two parameters , @xmath68 and @xmath38 .
however , these two parameters both effect the level width simultaneously , thereby the widths of the incompressible strips .
hence , one can not to distinguish their influence on the qhps separately .
typical hall resistances are shown in fig .
[ fig:10 ] calculated at default temperature considering different impurity parameters .
similar to the range parameter , we observe that the plateau widths become narrower when the mobility is low , which also points that our single particle based level broadening calculations are not in the correct direction .
such a behavior is easy to understand , when we decrease the mobility either by increasing the impurity concentration or by the amplitude of the impurity potential , the landau levels become broader due to collisions .
this means that , both the energetic and spatial gap between two consequent levels is reduced , hence the resulting incompressible strips are also narrower and fragile even at low temperatures .
a detailed investigation on the incompressible widths depending on impurity parameters are reported in ref .
it is known that if there exists an incompressible strip wider than the fermi wavelength the system is in the quantized hall regime @xcite , therefore , if the gap is reduced the incompressible strips are smeared , thus the quantized hall plateau vanish . as a general remark on the single particle theories , we should note that such a reduced gap is also a gross problem for the non - interacting models @xcite , however , one can overcome this discrepancy by making localization assumptions @xcite .
namely , one assumes that even within the broadened landau levels there are states , which are localized , therefore electrons can not contribute to transport .
hence , although the gap is small ( levels are broad ) these localized states serves as a reasonable candidate to explain the low mobility behavior . in the early days of iqhe it was a great challenge to describe and observe these localized states @xcite
recent experiments @xcite show clearly that , the localization assumptions are not relevant in all the cases , _ i.e. _ narrow and high mobility samples .
moreover , the universal behavior of the localization length dictated by these theories fail @xcite .
an explicit treatment of the activation energy @xcite and critical exponents are left to an other publication .
interval at default temperature , setting @xmath83 m , @xmath84 nm and @xmath85 , while changing the depletion length .
it is clearly seen that depletion length is much more important than the single impurity parameters in determining the plateau widths .
( b ) the direct comparison of the plateau widths considering different sample sizes .
the impurity parameters and depletion lengths are kept constant .
calculations are done at @xmath60 , whereas the donor density is @xmath86 m@xmath23 for all sample sizes .
[ fig:11],title="fig : " ] interval at default temperature , setting @xmath83 m , @xmath84 nm and @xmath85 , while changing the depletion length .
it is clearly seen that depletion length is much more important than the single impurity parameters in determining the plateau widths .
( b ) the direct comparison of the plateau widths considering different sample sizes . the impurity parameters and depletion lengths are kept constant .
calculations are done at @xmath60 , whereas the donor density is @xmath86 m@xmath23 for all sample sizes . [ fig:11],title="fig : " ]
another important parameter in defining the plateau widths is the depletion length @xmath87 .
the slope of the confinement potential close to the edges essentially determines the widths of the incompressible strips @xcite , which in turn determines the plateau widths . in fig .
[ fig:11 ] we show the @xmath57=2 plateau calculated for two different depletion lengths , we see that for the larger depletion the plateau is more extended .
since , the larger the depletion is , the smoother the electron density is . therefore ,
resulting incompressible strips are wider , hence the plateau .
such an argument will fail if one considers a highly disordered large sample , which we discuss in sec .
[ manyimp ] .
next , we compare the plateau widths of different sample sizes while keeping constant the disorder parameters and depletion length .
figure [ fig:11 ] depicts the sample size dependency of @xmath88 plateau width .
it is seen that the larger samples present wider plateaus , if the magnetic field is normalized with the center fermi energy , @xmath61 .
one can understand this by similar arguments given above , _
i.e. _ if the sample is narrow the variation of the confinement potential is stronger , therefore the incompressible strips become narrower , hence , the plateaus .
the discrepancy between the experimental results and the screening theory of the iqhe is solved if one considers not only the single impurity potentials but also the overall _ disorder _ potential landscape generated by the impurities . in the next part of this section
, we investigate the effect of the long range potential fluctuations on the quantized hall plateaus and find that , when the mobility is reduced the plateaus become wider and stabile , as it is observed in many experiments , ( see _ e.g. _ refs .
@xcite ) .
m. the depletion lengths and other single impurity parameters are kept fixed , whereas the parity of the modulation period is set 5.[fig:13 ] ] so far we have investigated the effect of single impurity potentials on the overall potential landscape in sec .
[ treed ] and on the widths of the plateaus in sec .
[ singleimp ] .
we have seen that , at high impurity concentration the overall potential fluctuates over a length scale of couple of hundred nanometers , whereas for low @xmath38 concentration such length scale can be as large as micrometers .
now we include the effect of this long range potential fluctuations _ into _ our screening calculations via modulation potential defined as @xmath89 where , the modulation period @xmath90 , is chosen such that the boundary conditions are preserved . at the moment , we consider two modulation periods regardless of the sample width and vary the amplitude of the modulation potential . in the next section , however , we select these two parameters from our estimations obtained in sec . [ impurity ] and sec .
[ treed ] .
figure [ fig:13 ] depicts the self - consistently calculated hall resistances , considering different modulation amplitudes @xmath91 for a fixed sample width ( @xmath92 @xmath93 m ) and @xmath94 .
we observe that , the plateaus become wider from the high @xmath0 field side , when @xmath91 is increased , _
i.e _ mobility is reduced .
such a behavior is now consistent with the experimental findings .
since the qhps occur whenever an incompressible strip is formed ( somewhere ) in the sample and the modulation forces the 2des to form an incompressible strip at a higher magnetic field , therefore the plateau is also extended up to higher field compared with the ( approximately ) non - modulated calculation , @xmath95 .
our investigation of the impurities lead us to conclude that , one has to define mobility at high magnetic fields also taking into account screening effects in general and furthermore also the geometric properties of the sample such as the width and depletion length . as an example if we consider an impurity concentration of @xmath96 the long range part of the potential fluctuation can be approximated to 900 nm .
however , note that the amplitude of this fluctuation varies between @xmath97 of the fermi energy , considering different separation thicknesses , therefore the wafer changes from low mobility to intermediate one .
another important parameter is the number of modulations within the system : a sample with an extend of @xmath98 m and @xmath99 is a high mobility sample with the same @xmath100 ( only 2 maximum ) , however , sample with a width of @xmath101 m is low mobility ( 10 maximum ) . in the next section
we study the plateau widths of different mobility samples , while keeping constant the extend and the amplitude of long range potential fluctuations ( _ i.e. _ @xmath91 and @xmath100 ) and short range impurity parameters ( @xmath68 , @xmath38 and @xmath67 ) under experimental conditions .
in this final section , we harvest our findings of the previous sections to make quantitative estimations of the plateau widths , considering narrow gate defined samples .
our aim is to show the qualitative and quantitative differences between `` high '' and `` low '' mobility samples , by taking into account properties of the single impurity potentials and the resulting disorder potential .
the experimental realizations of these samples are reported in the literature @xcite .
we estimated in sec .
[ treed ] that , the range of the potential fluctuations is @xmath102 nm for low mobility ( @xmath103 ) and is @xmath104 m at high mobility .
therefore , the modulation period is chosen such that many oscillations correspond to low mobility , and few oscillations correspond high mobility . as an specific example
let us consider a @xmath101 m sample , for the low mobility we choose @xmath105 and for the high mobility @xmath100 is taken as 9 or 10 .
the amplitude of the disorder potential is damped to @xmath106 of the fermi energy when considering the effect of spacer thickness , however , including screening this amplitude is further reduced to few percents . in light of this estimations the low mobility will be presented by a modulation amplitude of @xmath107 , whereas high mobility corresponds to @xmath108 . therefore , we have 4 different combinations of the disorder potential parameters yielding four different mobilities considering two sample widths , as tabulated in table [ table:2 ] . the second important aspect of the disorder is the single impurity parameters , for low mobility set we choose @xmath84 nm and @xmath109 , whereas for high mobility @xmath110 nm and @xmath85 is set
. remember that , the range of the single impurity is much less important than @xmath111 in determining the plateau width ( see sec .
[ singleimp ] ) .
.a qualitative comparison of the mobility in the presence of magnetic field also taking into account self - consistent screening .
mobility also depends on the size of the sample when screening is also considered .
[ cols="^,^,^,^",options="header " , ] figure [ fig:15 ] summarizes our results considering above discussed mobility regimes for two different sample widths . in fig .
[ fig:15]a , we show the calculated hall resistances for a sample of 10 microns with the highest mobility ( solid ( black ) line ) and intermediate 1 mobility ( broken ( red ) line ) .
the solid line is the highest mobility since the range of the fluctuations are at the order of 1 @xmath93 m and the amplitude of the modulation potential is five percent of the fermi energy .
the broken line presents the intermediate mobility considering a modulation amplitude of fifty percent .
we observe that the lower mobility wafer presents a larger quantized hall plateau , which is now in complete agreement with the experimental results .
moreover , our calculation scheme is free of localization assumptions in contrast to the known literature and we only considered a very limited level broadening , _
i.e. _ @xmath85 . in fact
our results also hold for dirac - delta landau levels , however , for the sake of consistency we choose the broadening parameters according to the selected disorder parameters . in fig .
[ fig:15]c , we show two curves for even lower mobilities , the solid line corresponds to the intermediate 2 case , whereas the broken line is the lowest mobility considered here .
the potential fluctuation range ( _ i.e. _ the modulation period ) is chosen to present the low mobility wafer .
we again see that for the lowest mobility the quantized hall plateau is enlarged considerably from both edges of the plateau .
these results explicitly show that the quantized hall plateaus become broader if one strongly modulates the electronic system by long range potential fluctuations , either by changing the range _ or _ the amplitude of the modulation .
similar results are also obtained for a relatively narrower sample @xmath112 m , fig .
[ fig:15]b and [ fig:15]d , however , we see that decreasing the range of the potential fluctuation is more efficient in enlarging the quantized hall plateaus when compared to the effect of the amplitude of the modulation .
the last interesting investigation is on the parity of the modulation period , _
i.e. _ whether @xmath100 is odd or even . figure [ fig:16 ] presents the different behavior when considering even ( a ) or odd ( b ) periods . here ,
all the disorder parameters are kept fixed , other than the parity .
we see that for the even parity the plateau is shifted towards the high field edge , both for @xmath88 and 4 , whereas for the odd parity the plateau is enlarged from both sides .
this tendency is also observed for the larger sample ( not shown here ) .
we attribute this behavior again to the formation of the incompressible strips , however , this time only to the one residing at the center of the sample , _
i.e. _ the bulk incompressible strip .
the picture is as follows : if the maxima of the modulation potential is at the center of the sample , the incompressible strip is formed at a higher magnetic field value , whereas , the edge incompressible strips become narrower at the lower field side . hence , due to the larger incompressible strip at the bulk of the sample the plateau is shifted to the higher field , in contrast , due to the narrower ( compared to the unmodulated system ) edge strips the plateau is cut off at higher fields . since , the edge incompressible strip becomes narrower than the extend of the wave function . for the odd parity , the edge
incompressible strips become wider , therefore , the plateau extends to the lower @xmath0 fields . the enhancement at the high field edge results from the two maximum in the proximity of the center . for a better visualization of the incompressible strip distribution we suggest reader to look at fig.2 of ref .
@xcite and fig.1 of ref .
such a shift of the quantized hall plateaus is also reported in the literature @xcite and is attributed to the asymmetrical density of states due to the acceptors in the system @xcite .
we claim that , the shift due to the modulation parity change observed in our calculations overlap with their findings . note that in our calculations we only consider symmetric dos , however , replacing a maxima with a minima at the confinement potential corresponds to the acceptor behavior of the dopants . a systematic experimental investigation is suggested to understand the underlying physical mechanism , where the system is doped with small number of acceptors .
in this work we tackled with the long standing and widely discussed question of the effect of disorder on the quantized hall plateaus .
the distinguishing aspect of our approach relies on the separate treatment of the long and short range of the disorder potential .
we show that assuming gaussian impurities is not sufficient to describe long range potential fluctuations , however , is adequate to give a prescription in defining the density of states broadening and conductivities .
the discrepancy in handling the long range potential fluctuations is cured by the inclusion of a modulation potential to the self - consistent calculations .
we estimated the range of these fluctuations from our analytical and numerical calculations considering the effect of dielectric spacer and the screening of the 2des .
it is observed that spacer damps the short range fluctuations effectively , whereas the direct coulomb interaction is dominant in screening the long range fluctuations . utilizing the estimations of the range and the amplitude of potential fluctuations , we classified mobility in four groups and calculated the hall resistances within the screening theory of the quantized hall effect .
we found that the hall plateaus are wider when decreasing the mobility , not surprisingly .
however , the most important point of our theory is that , we do not consider any localization assumptions , still obtain correct behavior of the plateau widths .
we show that @xmath82 and/or short range impurity defined mobility is not adequate to describe the actual mobility at high magnetic fields , moreover , one has to include geometrical properties of the sample at hand .
a natural persecutor theoretical investigation of the present work should deal with the activated behavior of the longitudinal resistance within the screening theory . as it is well known , the properties of the localized states , _
e.g_. the localization length , is usually obtained from the activation experiments @xcite . moreover ,
spin generalization of the screening theory @xcite is necessary to describe and investigate odd integer quantized plateaus also considering level broadening , namely disorder .
one of the authors ( a.s . ) would like to thank e. ahlswede , s. c. lok and j. weiss for the enlightening discussions on the disorder from `` an experimentalist '' point of view .
the authors acknowledges , the feza - grsey institute for supporting the iii .
nano - electronic symposium , where this work has been conducted partially and would like to acknowledge the scientific and technical research council of turkey ( tubitak ) for supporting under grant no 109t083 .
10 jiri j mares , afif siddiki , dobroslav kindl , pavel hubik , and jozef kristofik . electrostatic screening and experimental evidence of a topological phase transition in a bulk quantum hall liquid .
, 11(8):083028 , 2009 .
a. siddiki and r. r. gerhardts .
nonlinear thomas - fermi - poisson theory of screening for a hall bar under strong magnetic fields . in a.
r. long and j. h. davies , editors , _ proc .
15th intern .
conf . on high magnetic fields in semicond
_ , bristol , 2002 .
institute of physics publishing . | we study effects of disorder on the integer quantized hall effect within the screening theory , systematically .
the disorder potential is analyzed considering the range of the potential fluctuations .
short range part of the single impurity potential is used to define the conductivity tensor elements within the self - consistent born approximation , whereas the long range part is treated self - consistently at the hartree level . using the simple , however , fundamental thomas - fermi screening , we find that the long range disorder potential is well screened
. while , the short range part is approximately unaffected by screening and is suitable to define the mobility at vanishing magnetic fields . in light of these range dependencies
we discuss the extend of the quantized hall plateaus considering _ the _ mobility " of the wafer and the width of the sample , by re - formulating the ohm s law at low temperatures and high magnetic fields .
we find that , the plateau widths mainly depend on the long range fluctuations of the disorder , whereas the importance of density of states broadening is less pronounced and even is predominantly suppressed .
these results are in strong contrast with the conventional single particle pictures .
we show that the widths of the quantized hall plateaus increase with increasing disorder , whereas the level broadening is negligible .
this work focuses on the disorder effects on the integer quantized hall effect within the screening theory . since the early days of qhe
, disorder played a very important role , however , interactions were completely neglected . here
we present our results which also includes interactions in a self - consistent manner and show that even without localization one can obtain the quantized hall plateaus .
we investigated different aspects of the impurity potential and suggested a criterion on mobility at high magnetic fields .
we think that our work will shed light on the understanding of the qhe and is interest to condensed matter community .
_ keywords _ : article preparation , iop journals |
the expressions ( [ rhopert ] ) , ( [ rhog2 ] ) and ( [ rhog3 ] ) are common to the three resonances .
next , we list the other terms that are not common . for a review and references to original works , see e.g. , s.narison , _ qcd as a theory of hadrons _ , cambridge monogr .
* 17 * , 1 ( 2002 ) [ hep - h/0205006 ] ; _ qcd spectral sum rules _ , world sci .
notes phys .
* 26 * , 1 ( 1989 ) ; acta phys . pol . *
b26 * , 687 ( 1995 ) ; riv .
* 10n2 * , 1 ( 1987 ) ; phys .
rept . * 84 * , 263 ( 1982 ) .
s. narison , nucl .
* 86 * ( 2000 ) 242 ( hep - ph / 9911454 ) ; s. narison , hep - ph / 0202200 ; s. narison , hep - ph / 0510108 ; s. narison , phys .
lett . b341 ( 1994 ) 73 ; h.g .
dosch and s. narison , phys . lett . *
b417 * ( 1998 ) 173 ; s. narison , phys . lett .
* b216 * ( 1989 ) 191 . | using the qcd sum rule approach we investigate the possible four - quark structure for the new observed @xmath0 narrow structure ( d0 ) .
we use a diquak - antidiquark scalar current and work to the order of @xmath1 in full qcd , without relying on @xmath2 expansion .
our study indicates that although it is possible to obtain a stable mass in agreement with the state found by the d0 collaboration , a more constraint analysis ( simultaneous requirement of the ope convergence and the dominance of the pole on the phenomenological side ) leads to a higher mass .
we also predict the masses of the bottom scalar tetraquark resonances with zero and two strange quarks .
recently the d0 collaboration reported the observation of a narrow structure , called @xmath3 , in the decay @xmath4 @xcite .
this is the first observation of a hadronic state with two quarks and two antiquarks of four different flavors and , therefore , can only be explained as a tetraquark or molecular state .
the mass and width of the observed state were reported to be : @xmath5 and @xmath6 . as pointed out in ref .
@xcite , considering the large mass difference between the mass of the @xmath3 and the sum of the @xmath7 and @xmath8 masses , it can be difficult to explain the @xmath3 as a molecular state .
therefore , the @xmath3 is an excellent candidate for a tetraquark state .
if the @xmath0 pair in the @xmath3 decay is produced in @xmath9-wave , its quantum numbers are @xmath10 as the very narrow @xmath11 state , first discovered in the @xmath12 decay channel by the babar collaboration @xcite .
due to its low mass , the structure of the @xmath13 meson has been extensively debated .
it has been interpreted as a @xmath14 state @xcite , two - meson molecular state @xcite , @xmath15- mixing @xcite , four - quark states @xcite or a mixture between two - meson and four - quark states @xcite . in this paper
we use the qcd sum rule ( qcdsr ) approach @xcite to investigate the possible four - quark structure for the @xmath3 and , therefore , to test if the @xmath3 could be the isovector bottom partner of the @xmath16 .
the qcdsr for scalar mesons are constructed from the two - point correlation function written in terms of a scalar current @xmath17 : ( q)=id^4x e^iq.x0 |t[j_s(x)j^_s(0)]|0 .
the key idea of the qcdsr method is to consider that this correlation function is of dual nature and it depends on the value of the momentum @xmath18 . for large momentum , i.e. , short distances
, the correlation function can be calculated using perturbative qcd . in this case , the current @xmath17 is written in terms of the quark content of the studied mesons . however , since we are interested in studying the properties of hadrons , the relevant energies are lower and contributions from quark condensates , gluon condensates , etc . , need to be included in the evaluation of eq .
( [ 2po ] )
. this can be done by using the wilson operator product expansion ( ope ) of the correlation function . in this case , eq .
( [ 2po ] ) is expanded in terms of local condensates and a series of coefficients .
the local operators incorporate nonperturbative long - distance effects , while the coefficients , by construction , include only the short - distance domain and can be determined perturbatively .
this way of evaluating the correlation function is customarily named as the calculation on the ope side " . at large distances , or , equivalently , small momentum , the currents @xmath19 and @xmath17 of eq .
( [ 2po ] ) can be interpreted as operators of creation and annihilation of the scalar mesons . in this case , the correlation function is obtained by inserting a complete set of scalar states .
this interpretation of the correlation function is called as the phenomenological side " .
the assumption made in the qcdsr approach is that there must be a range of @xmath20 values in which both descriptions must be equivalent . calculating the correlation function of eq .
( [ 2po ] ) using these two approaches and equating them , it is possible to obtain information about the properties of the hadronic states generated in the system . in ref .
@xcite the @xmath16 state was considered as a diquark - antidiquark tetraquark state and was studied by using the qcdsr approach .
a very good agreement with the experimental mass was obtained . here
we follow ref .
@xcite to write an analogous but isovector scalar - diquark scalar - antidiquark tetraquark current for @xmath3 : j_s&=&_abc_dec(u_a^tc _ 5s_b)(|d_d_5c|b_e^t ) , [ int ] where @xmath21 are colour indices , @xmath22 is the charge conjugation matrix .
of course a scalar - scalar diquark - antidiquark form is not the only possible choice for a scalar tetraquark current , and one could use pseudoscalar - pseudoscalar , vector - vector or axial - axial diquark - antidiquark form .
however , it was shown in ref .
@xcite that the scalar - scalar type of current gives the more stable results .
therefore , we use the current given by eq .
( [ int ] ) for the @xmath23 .
the coupling of the state , @xmath24 , to the scalar current , @xmath17 , can be parametrized in terms of the constant @xmath25 as : @xmath26 , therefore , the phenomenological side of eq .
( [ 2po ] ) can be written as ^phen(q^2)=f_x^2m_x^2-q^2 + , where the dots denote the contribution from higher resonances , which is usually parametrized through the introduction of a continuum threshold parameter @xmath27 @xcite .
on the ope side we work at leading order and consider condensates up to dimension six .
we deal with the strange quark as a light one and consider the diagrams up to order @xmath1 .
to keep the bottom quark mass finite , we use the momentum - space expression for the bottom quark propagator .
we follow ref .
@xcite and calculate the light quark part of the correlation function in the coordinate - space , which is then fourier transformed to the momentum space in @xmath28 dimensions .
the resulting light - quark part is combined with the charm - quark part before it is dimensionally regularized at @xmath29 .
we can write the correlation function on the ope side in terms of a dispersion relation : ^ope(q^2)=_m_b^2^ds ( s)s - q^2 , where the spectral density is given by the imaginary part of the correlation function : @xmath30 $ ] . after making a borel transform on both sides , and transferring the continuum contribution to the ope side , the sum rule for the scalar meson @xmath24 can be written as f_x^2e^-m_x^2/m^2=_m_b^2^s_0ds e^-s / m^2 ( s ) , where @xmath31 is the borel mass and @xmath32 , with [ rhopert ] ^pert(s)=12 ^ 10 3 ^ 6_^1 d(1- ) ^3(m_b^2-s)^4 , ^m_s(s)=0,^(s)=12 ^ 6 ^ 4_^1 d 1- ( m_b^2-s)^2 , ^g^2(s)&=&2 ^ 10 ^ 6_^1 d ( m_b^2-s ) , [ rhog2 ] ^mix(s)&=&12 ^ 6 ^ 4_^1 d ( m_b^2-s ) , ^^2(s)&=&-124 ^ 2_^1 d ( 2|ss(m_b^2-s+m_bm_s ) + & & -^2m_bm_s ) , ^g^3(s)=32 ^ 12 9 ^ 6_^1 d(1-)^3(3m_b^2-s),[rhog3 ] where the lower limit of the integrations is given by @xmath33 . in order to compute the mass of the state , @xmath34 , we first take the derivative of eq .
( [ sr ] ) with respect to @xmath35 and then we divide the result by eq .
( [ sr ] ) , obtaining m_x^2=_m_b^2^s_0ds e^-s / m^2 s ( s)_m_b^2^s_0ds e^-s / m^2 ( s ) .
this expression will be used to evaluate the mass of the state .
the numerical values for the quark masses and condensates are listed in table [ tab1 ] @xcite .
.qcd input parameters . [
cols="<,<",options="header " , ] the remaining input to the calculation is the continuum threshold parameter which , in general , is related to the mass of the state to be studied ( @xmath3 , in the present case ) as @xmath36 . therefore , to start our analysis we choose @xmath37 gev . in order to determine the values of the borel mass parameter
, we analyze the pole contribution , the ope convergence and the borel stability . in the qcdsr approach
we extract information only from the ground state , therefore we must ensure that the pole contribution is greater that the continuum contribution . here
we fix the borel mass in such a way that the pole contribution is always between 80% and 50% of the total contribution . from fig .
[ fig1 ] we can see that this condition is satisfied for values of the borel mass in the range @xmath38 . in fig .
[ fig2 ] we plot the ground state mass as a function of @xmath39 , considering three different values of the threshold parameter .
we can see that there is a good @xmath39-stability for the borel window considered .
using the central values of the parameters in table [ tab1 ] and @xmath40 we get m_x~5.58 .
to evaluate the uncertainties inherent to the qcd sum rule approach , we consider the variation of the mass in borel window , as a function of the continuum threshold , changed within a small range : @xmath41 , and the quark masses and condensates errors indicated in table [ tab1 ] . considering these uncertainties
we get : m_x=(5.580.17 ) , [ result1 ] which is in excellent agreement with the experimental mass of the x(5568 ) determined by the d0 collaboration @xcite .
the result in eq .
( [ result1 ] ) was obtained considering only the pole dominance and the stability with the borel mass .
there is , however , a stronger constraint to the lower bound of the @xmath39 , that comes from imposing the ope convergence .
we analyze the convergence of the ope by comparing the relative contribution of each term given by eqs .
( [ rhopert ] ) to ( [ rhog3 ] ) , to the right hand side of eq .
( [ sr ] ) .
the requirement of a good convergence sets a lower limit to @xmath39 .
this analysis in shown in fig .
[ fig3 ] . as can be seen from fig .
[ fig3 ] , there is no ope convergence in any region allowed by the upper bound given by pole / continuum analysis : @xmath42 .
this means that the lower bound given by ope convergence will be higher than the upper bound , and there is no valid `` sum rule window '' where we can completely trust the results for this current . to overcome this problem
we can consider higher dimension condensates in the ope side and test if the series starts converging with such contributions .
we include the condensates of dimension seven and eight whose expressions are given below : & & ^|ssg^2(s)= , ^g^2(s)&=&- , & & ^|qgq(s)= , & & ^|ss|qgq(s)=- , ^|sgs(s)=-(6+(s - m_b^2 ) ) , ^g^4g^4(s)&=&- .
on continuing with our analysis we find that , even after considering condensates up to dimension eight , a valid `` sum rule window '' exists only for values of @xmath43 . in figs .
[ fig4 ] and [ fig5 ] we show the ope convergence and the pole versus continuum contribution for @xmath43 . from fig .
[ fig4 ] we can see that there is an ope convergence , the dimension eight condensate contribution is smaller than 20% of the total contribution , only for values of @xmath44 . on the other hand , from fig .
[ fig5 ] we can see that the pole contribution is bigger than the continuum contribution for values of @xmath45 .
although very small , there exists a valid borel window in the region @xmath46 , which provides a good `` sum rule '' to extract a reliable value for the mass of the state . in fig .
[ fig6 ] , we show the resulting value for the mass of the state , as a function of the borel mass , for three different values of the continuum threshold .
the crosses in the figure indicate the `` sum rule window '' . finally , considering all the restrictions described above we get : m_x=(6.390.10 ) , [ result2 ] which is not in agreement with the experimental mass of the @xmath3 determined by the d0 collaboration @xcite . as a matter of fact , recently the lhcb collaboration has not confirmed the observation of the @xmath3 . in their preliminary analysis @xcite
no structure is found in the @xmath0 mass spectrum from the @xmath47 threshold up to @xmath48 .
more analyses are required to clarify this situation .
our work predicts a tetraquark state decaying in this channel with a mass around 6.39 gev .
the uncertainty given in eq .
( [ result2 ] ) is only related with the range of values of the borel mass window , a small variation in the continuum threshold , @xmath49 , and the quark masses and condensates errors indicated in table [ tab1 ] .
the difference between the values in eqs .
( [ result1 ] ) and ( [ result2 ] ) can be associated mainly with the change in the value of the continuum threshold . however , as discussed above , there is no allowed borel window for values of @xmath50 .
therefore , the result obtained with @xmath51 given in eq .
( [ result1 ] ) , although being obtained in a borel region where one has pole dominance , can not be trusted .
this result illustrates very well how we can reproduce the mass of a given state and then after a more careful analysis conclude that the state is not the particle associated with the chosen current .
we also would like to point out that the difference between @xmath52 and the result in eq .
( [ result2 ] ) is about 0.5gev , as the general supposition of the qcdsr approach for the start of the continuum threshold .
it is also important to notice that the difference : @xmath53 increases with the value of the continuum threshold . as an example , for @xmath54 we get @xmath55 which implies @xmath56 , much larger than 0.5 gev .
this could be an indication that there is a contribution from higher resonances below the continuum threshold and , therefore , once again , the estimated mass can not be trusted . therefore , to fix a `` good range '' of the values of @xmath27 we test if it provides an allowed borel window ( where both constraints of the pole dominance and the ope convergence are satisfied ) , and that the value of the obtained mass falls within the range 0.4 gev to 0.6 gev smaller than @xmath57 . using these criteria ,
we have obtained @xmath27 in the range @xmath58 .
we can extend the formalism to study bottom scalar mesons states that contain zero and two strange quarks . in order to calculate the correlation function for these states we use the following interpolating fields for these states ( zero and two strange quarks , respectively ) : j_0&=&_abc_dec(u_a^tc_5q_b ) ( |d_d_5c|b_e^t ) , + j_ss&=&_abc_dec(u_a^tc _
5s_b)(|s_d_5c|b_e^t ) , [ int2 ] where @xmath18 represents the quark @xmath59 or @xmath60 according to the charge of the meson .
the expression for the resulting spectral densities are given in appendix a. we call @xmath61 and @xmath62 the scalar bottom tetraquark mesons represented by @xmath63 and @xmath64 respectively ( eq . [ int2 ] ) . in figs .
[ fig7 ] and [ fig8 ] we show the masses of the states as a function of the borel mass for different values of the continuum threshold , and as the previous case there is good @xmath39-stability in the allowed borel window , represented by the crosses in these figures .
the values obtained for the masses of the states are : m_b_0 ^ 0s=6.100.16 , + m_b_0 ^ 2s=6.390.17.[result3 ] the sources of errors are the same that were used for the strange - bottom scalar meson . comparing the results of the masses in eqs .
( [ result2 ] ) and ( [ result3 ] ) we can see that the @xmath24 ( the state with one strange quark ) and @xmath65 resonance masses are basically degenerated , while the mass of @xmath66 is around @xmath67 smaller than the others .
the same behavior is observed for the scalar mesons in the charm sector , as it was pointed out in ref .
@xcite .
the increase in the mass is expected with the inclusion of one strange quark ( from zero to one strange quarks ) .
the fact this is not observed when one goes from one to two strange quarks can be traced when we compare the contribution of the quark condensate term , that is smaller for the @xmath65 , but this is compensated by the inclusion of a term @xmath68 ( eq . [ rhoms ] ) . in summary , we have presented a qcd sum rule study for a bottom scalar meson considered as diquark - antidiquark state .
the motivation of the study is to look for a possible state associated with the recently claimed @xmath3 by the d0 collaboration .
we find that it is possible to obtain a stable mass in agreement with the state found by the d0 collaboration while satisfying the condition of the pole dominance on the phenomenological side , but by sacrificing the simultaneous constraint of the ope convergence .
this last missing ingredient casts a doubt on the reliability of the result , leading us to conclude that the @xmath3 state can not be represented by the scalar tetraquark current .
we find that a rigorous application of qcd sum rule constraints leads to a higher mass .
thus , we predict the existence of a scalar bottom - strange tetraquark state with a mass around 6.4 gev .
we also have obtained the masses for the resonances with zero and two strange quarks , which we call @xmath69 and @xmath70 .
these resonances have also not been observed yet , but our calculations show that they should exist . : this work has been supported by cnpq and fapesp .
[ apa ] |
among the recently studied superconducting compounds there are several groups of materials ( copper oxides , fullerides , chevrel phases , barium bismuthates ... ) which share some nonconventional features like high @xmath2 with relatively low carrier concentration , short coherence length together with extremely small @xmath15 ratio and universal dependence of @xmath16 vs. @xmath11 ^ 2 $ ] @xcite . above features suggest that a short range almost nonretarded attraction is responsible for pairing in these systems @xcite .
one of the simplest effective models to describe this situation is the negative@xmath0 hubbard hamiltonian .
this model constitutes a common basis for the description of the superconductors with weak local electron pairing , being in many ways similar to the classical bcs systems , and superconductors with strong attraction where the local pairs conform to weakly charged hard core bosons @xcite .
it has been considered as an effective model of superconductivity in the family of cuprates @xcite , the barium bismuthates @xcite , the fullerides @xcite and the chevrel phases @xcite .
in our paper we present some new results of an analysis of the several basic thermodynamic and electromagnetic properties of this model for various types of lattice structure .
we have concentrated on the case of ( i ) square ( sq , @xmath17@xmath18@xmath19 ) , ( ii ) simple cubic ( sc , @xmath17@xmath18@xmath20 ) and ( iii ) face centered cubic ( fcc , @xmath17@xmath18@xmath20 ) lattices , i.e. structures representing ( i ) the family of cuprates , ( ii ) the chevrel phases and barium bismuthates and ( iii ) the fullerides , respectively . for comparison
we also present the results obtained for a model rectangular density of states @xmath21 .
to calculate quantities of interest we use the ( broken symmetry ) hartree fock approximation ( hfa ) which is known to give credible results at @xmath22 k as concerns energy of the ground state , the energy gap , the chemical potential as well as the collective excitations in the whole interaction range , interpolating smoothly between the weak and strong@xmath0 limit , where it matches the results of the perturbation theory developed from the zero bandwidth limit @xcite . on the other hand the hfa leads to qualitatively erroneous results for the the thermodynamic critical field @xmath4 and ginzburg landau correlation length @xmath5 in the strong attraction limit .
this failure is due to the fact that the hfa greatly overestimates the energy of the _ normal _ state which is used in a standard calculation of @xmath4 and next @xmath5 . in order to discuss the behaviour of @xmath4 , @xmath5 and ginzburg ratio @xmath23 in the whole interaction range we have completed the results of the hfa calculation with the ones of the perturbation theory .
we will also point out a possible relevance of our results to the recent experimental data for the short coherence length superconductors .
we consider the negative@xmath0 hubbard hamiltonian in the hypercubic lattice @xmath24 where @xmath25 is the hopping integral , @xmath26 is a unit lattice vector and the exponential ( peierls ) factors in ( [ hubbard ] ) account for the coupling of electrons to the magnetic field via its vector potential @xmath27 .
@xmath28 , and @xmath29 is the electron charge . for weak vector potential the expectation value of the fourier transform of the total current operator in @xmath30 direction @xmath31 can be obtained from the linear response theory @xcite as @xmath32 a_{\beta}(\vec{q}\,,\omega)\,\,\,\,\ , .\ ] ] here we have separated the total response kernel @xmath33 into a diamagnetic part and a paramagnetic part .
the diamagnetic contribution evaluated within hfa is @xmath34\cos k_\alpha \,\,\,\,\ , , \end{aligned}\ ] ] where @xmath35 , @xmath36 and @xmath37 is the quasiparticle energy with the superconducting gap parameter + @xmath38 .
the chemical potential is @xmath39 , @xmath40 and @xmath41 denote the number of lattice sites and the number of electrons per site , respectively and @xmath42 .
the values of @xmath1 and @xmath43 are obtained as a solution of the system of the selfconsistent equations @xcite : @xmath44 \,\,\,\ , , \ ] ] whereas the free energy per site of the superconducting ( ss ) phase and the normal ( no ) phase is given by : @xmath45 where @xmath46 is determined from : @xmath47 the expression for the paramagnetic part of the kernel reads @xmath48\times\nonumber \\ & & \times\left(\frac{1}{\omega-(e_{\vec{k}-\vec{q } } -e_{\vec{k}})+i\varepsilon}- \frac{1}{\omega+e_{\vec{k}-\vec{q } } -e_{\vec{k}}+i\varepsilon } \right)+\nonumber \\ & & + \left(1-\frac{\lambda_{\vec{k } } \lambda_{\vec{k}-\vec{q}}+\delta^{2 } } { e_{\vec{k}}e_{\vec{k}-\vec{q}}}\right ) \left[1-n_{f}(e_{\vec{k } } ) -n_{f}(e_{\vec{k}-\vec{q}})\right]\times \nonumber \\ & & \left.\times \left(\frac{1}{\omega- ( e_{\vec{k}-\vec{q}}+e_{\vec{k}})+i\varepsilon}- \frac{1}{\omega+e_{\vec{k } } + e_{\vec{k}-\vec{q}}+i\varepsilon } \right ) \right]\,\,\,\,\ , . \ ] ] in the london superconductors the magnetic field penetration depth @xmath49 is determined by the sum of diamagnetic and paramagnetic part of the total kernel in the static limit @xmath50^{-1/2}$ ] @xcite . at @xmath22 k the paramagnetic part of the kernel becomes important in determining @xmath3 when the correlation length becomes greater than the penetration depth and we deal in this case with nonlocal ( pippard ) superconductor .
this situation is common in many low@xmath2 systems .
the short coherence length materials , including the high @xmath2 superconductors represent the opposite , i.e. the london limit . in the latter case the ground state penetration depth is determined entirely by the @xmath51 limit of the kernel where the paramagnetic part of the kernel vanishes and @xmath3 is given by : @xmath52 in our calculation of @xmath3 we restrict ourselves to the london limit having in mind the properties of the systems of interest and finaly determine the area in the @xmath0@xmath6 parameters space where the local approximation may be valid .
the value of the penetration depth calculated in this way is qualitatively good both in the weak and strong@xmath0 limits , in the latter case approaching the results of the perturbation theory , as will be shown below .
using the value of the penetration depth and the difference of the free energy between normal and superconducting phase one is able to determine the thermodynamic critical field @xmath4 and the ginzburg
landau correlation length @xmath5 as @xmath53 where @xmath54 , and to obtain the estimations for the critical fields @xmath55 and @xmath56 , where @xmath57 . at @xmath58 . within hfa
the expressions determining @xmath59 and @xmath60 are given by : @xmath61 where @xmath62 is the average value of the kinetic energy term in the no(ss ) phase .
the hfa calculation of the energy of the ordered state at @xmath22 k are reliable for any @xmath0 @xcite and @xmath63 reduces correctly to the exact value : @xmath64 in the zero bandwidth limit where the electrons form a system of on site pairs and singly occupancy of sites is prohibited . on the other hand
the hfa energy of the normal phase in this limit is equal to @xmath65 instead of @xmath64 thereby producing the incorrect energy difference of @xmath65 . in order to remedy this inconsistency , at least in part , we resort to the results of the perturbation theory in the calculation of @xmath66 for @xmath67 .
the hamiltonian ( [ hubbard ] ) reduces in this case to the pseudospin@xmath68 heisenberg model working in the subspace of states with no singly occupied sites @xcite : @xmath69 where @xmath70 , @xmath71 , @xmath72 and @xmath73 denotes a number of nearest neighbours . the hard core boson operators : @xmath74 satisfy the commutation rules of the @xmath75 operators .
the electron number condition is : @xmath76 for @xmath77 the hamiltonian ( [ model - eff ] ) describes the superconducting state and the mfa result for the @xmath60 is equal to the corresponding hfa one up to terms of order @xmath78 inclusive ( see sec.3 ) .
the mfa calculation of the ground state energy for the normal state , @xmath79 , gives : @xmath80 using this expresion in the analysis one should keep in mind that the mfa neglects the intersite correlation of fluctuation of the pseudospin operators what can lead to the error of order of @xmath81 .
the system of the selfconsistent equations ( [ delta],[mu ] ) can be approximately solved in the limiting cases of the weak and strong attraction limits and here we present these results : * the weak@xmath0 case + provided that the chemical potential is not located too close to van hove singularity of the density of states ( dos ) function @xmath21 , the gap parameter at @xmath22 can be obtained from equation ( [ delta ] ) as : @xmath82 } { 2{\cal d}(\bar{\mu})}\right ) \exp\left(-\frac{1}{|u|{\cal d}(\bar{\mu})}\right ) \,\,\,\,\
, , \ ] ] where : @xmath83 , @xmath84 with @xmath85,@xmath86 denoting lower and upper boundary of the band , respectively ; @xmath87 $ ] is a functional defined by + @xmath88\!=\!\int_{-d_1}^{d_2 } d\epsilon \mbox{\rm sign}(\epsilon-\bar{\mu})f^{\bar{\mu}}(\epsilon ) \,\,\,\,\
, , \ ] ] + @xmath89 is a function of electron density , which depends on details of dos and is defined by @xmath90 in eqs.([delta - weaku]),([cmu ] ) the chemical potential may be approximated by its value in the normal state , @xmath46 and it is determined by : @xmath91 the critical temperature in this limit is given by : @xmath92 which is the well known weak coupling bcs ratio .
+ analogously , performing the weak coupling expansions of eqs.([k_dia ] , [ mu ] , [ hc ] ) at @xmath22 one can find the analytical expressions determining @xmath3 , @xmath93 , @xmath4 , @xmath5 and @xmath94 .
the london penetration depth is calculated as @xmath95 where + @xmath96 + \frac{\bar{\mu}}{2(d_2-\bar{\mu})^2}\nonumber \\ + & & \left.{\cal d}(\bar{\mu})+\frac{1}{2}i\left[{\cal c}^{\bar{\mu}}\right ] -i\left[{\cal f}_3^{\bar{\mu}}\right]+\bar{\mu } i\left[{\cal g}_3^{\bar{\mu}}\right]\right\ } \,\,\,\,\ , , \end{aligned}\ ] ] + where @xmath97 and @xmath98 . obviously , for @xmath99 approaches the value of the band energy in the normal state .
+ in a case when we can neglect variation of @xmath100 and @xmath101 eq.([ekss_uweak ] ) reduces to + @xmath102\,\,\,\,\ , , \end{aligned}\ ] ] + whereas @xmath93 and @xmath103 are given by + @xmath104 + where @xmath46 is given by eq.([mun ] ) , and + @xmath105 + the weak coupling formulas for @xmath5 and @xmath106 follow directly from eqs.([xi ] ) , ( [ lambda_uweak ] ) , ( [ hc_uweak ] ) . in the case of @xmath107 and
@xmath101 their explicit forms are : + @xmath108 + @xmath109 * the strong@xmath0 case + in this case the integrands in eqs.([delta],[mu ] ) can be expanded in series with respect to @xmath110 and we have : + @xmath111 + @xmath112 + where @xmath113 , @xmath46 is given by ( [ mun ] ) and the lattice sums @xmath114 represent the moments of dos and are given by : @xmath115 one can note that odd moments vanish for the alternating ( i.e. sq and sc ) lattices but not for the fcc ones . for alternating lattices @xmath116 , @xmath117 and for fcc lattice @xmath118 and @xmath119 . the corresponding formula for @xmath3 obtained by expanding of eq.([k_dia ] ) reads @xmath120 \,\,\,\,\ , . \ ] ] in the structures with asymmetric dos ( like fcc ) the odd terms in @xmath110 from eq.([lambda - strongu ] ) are again nonzero . from ( [ eno ] ) and ( [ ess ] )
we obtain : @xmath121 @xmath122 @xmath123 @xmath124^{-1 } \,\,\,\,\
, , \ ] ] @xmath125 it is worth while to compare the strong coupling results of hfa with the ones obtained for the effective pseudospin model hamiltonian ( [ model - eff ] ) .
recently , the electromagnetic properties of that model have been analysed in @xcite and below we only quote the expressions derived within mfa ( the rpa treatment , taking into account quantum corrections , yields qualitatively similar results , except the low density limit @xcite ) : @xmath126 @xmath127 @xmath128 @xmath129 where @xmath130 , @xmath131 and @xmath132 , @xmath133 , @xmath134 @xmath135 and @xmath66 is given by eq.([en - strongu ] ) . up to terms of order
@xmath78 the hfa expressions for @xmath136 ( eq.([x0_ustrong ] ) ) , @xmath137 ( eq.([ess_ustrong ] ) ) and @xmath138 ( eq.([lambda - strongu ] ) ) are equal to those given by eqs.([mfax0 ] ) , ( [ mfaess ] ) and ( [ mfalambda ] ) . on the contrary
the hfa results for @xmath4 , @xmath139 and @xmath94 are qualitatively erroneous due to the great overestimation of @xmath66 by hfa in this limit ( comp .
eq.([eno_uweak ] ) and eq.([en - strongu ] ) ) .
thus , to get realistic values of @xmath4 , @xmath5 and @xmath94 for arbitrary @xmath140 one should combine the results obtained using @xmath66 given by eq.([eno ] ) ( reliable for @xmath141 ) with those obtained using @xmath66 given by eq.([en - strongu ] ) ( reliable for @xmath142 ) .
this will be done in the next section , where we will also compare some of the predictions given above with the results of numerical solutions of eqs.([k_dia])-([mu ] ) , ( [ hc])-([ess ] ) .
figs.1a and 1b show the numerical plots of @xmath143 , @xmath144 and @xmath145 for sc and fcc lattices , respectively ( the corresponding plots for sq lattice are given in fig.1 of ref.@xcite ) , whereas fig.1c presents the concentration dependence of @xmath1 and @xmath145 for the model rectangular @xmath21 : where @xmath147 is the effective bandwidth .
all these plots were made for the values of @xmath0 from the weak coupling regime ( @xmath148 ) where the effects of the density of states , @xmath21 , are most clearly seen . in this regime the superconducting critical temperature @xmath2 and the @xmath149 gap parameter @xmath1 increase rapidly with @xmath150 approaching van hove singularities in @xmath21 , in agreement with a modified bcs expression eq.([delta - weaku ] ) .
this is the reason why @xmath2 for sq and sc lattices is symmetric with respect to @xmath6@xmath18@xmath151 and so strongly peaked for sq lattice at @xmath6@xmath18@xmath151 while the position of @xmath2 peak in the fcc structure moves with increase of @xmath152 from the vicinity of @xmath6@xmath153@xmath19 toward @xmath6@xmath18@xmath151 . for the fcc lattice
the effect of density of states on the penetration depth consist mainly in @xmath0-dependent shift of the maximum of @xmath138 from @xmath154 towards @xmath155 ( fig.1b ) .
the inverse squared penetration depth , being proportional to the bandwidth for weak@xmath0 decreases like @xmath156 in the strong coupling limit .
the plot of this quantity as a function of @xmath6 for a few values of @xmath140 is shown in fig.2 ( sq lattice ) .
examples of the evolution of the thermodynamic critical field @xmath4 and of the correlation length @xmath5 , with @xmath152 and @xmath6 are shown in fig.3,4 .
note the steep decrease of @xmath5 with @xmath6 in a relatively narrow range of @xmath6 values for the fcc lattice and substantial reduction of @xmath5 near @xmath154 for the sq lattice ( fig .
these features can be understood on the basis of the @xmath144 behaviour presented in figs 1,2 .
the substantial variation of @xmath5 with @xmath6 for @xmath141 is largely due to the strong @xmath6 dependence of @xmath157 which in the weak coupling limit is proportional to @xmath158 ( cf .
eqs.([hc_uweak ] ) , ( [ ksi_uweak ] ) ) . with increasing @xmath152
the @xmath6dependence of the correlation length is less pronounced and in the strong coupling regime , @xmath159@xmath160@xmath151 , @xmath5 calculated with in hfa goes like @xmath161 being almost independent on @xmath6 .
however , for large @xmath152 the energy of the _ normal _ phase , entering ( [ hc ] ) , ( [ xi ] ) is drastically overestimated in the hfa ( comp .
3 ) . to get more realistic values of @xmath4 and @xmath5 for @xmath140@xmath162@xmath151
we have made an estimation of @xmath5 and @xmath4 exploiting perturbation approach in the calculation of the energy of the normal phase ( eq.([en - strongu ] ) ) and the numerical results for sq lattice and @xmath154 are shown in figs 3,4 . in the @xmath163
limit the analytical results for @xmath4 , @xmath5 and @xmath94 can be obtained in this way by using eqs.([en - strongu ] , [ ess_ustrong ] , [ lambda - strongu ] and [ ux0_ustrong ] ) : @xmath164 @xmath165 @xmath166
they coincide ( to the second order in @xmath167 ) with the mfa results for the pseudospin model ( [ model - eff ] ) ( cf . eqs.([mfahc ] ) , ( [ 15 ] ) , ( [ 151kappa ] ) ) .
finally let us conclude the evolution of the ginzburg ratio @xmath171 with @xmath152 and @xmath6 .
concentration dependence of this quantity is plotted in fig 6 .
notice universal behaviours ( i ) @xmath172 for @xmath173 ( arbitrary @xmath140 ) and ( ii ) @xmath174^{-1/2}$ ] for @xmath168 ( arbitrary @xmath6 ) . in the weak - to - intermediate coupling regime ( @xmath141 )
@xmath175 is not universal and strongly depends on the details of @xmath100 . from the equation
@xmath176 one can estimate the boundaries between local and nonlocal electromagnetic behaviour in the considered model @xcite . in fig .
7 we plot this boundary in the @xmath177 parameter space for fixed value of @xmath25 and @xmath178 . in the low concentration limits ( @xmath179 ) the local behaviour extends up to very small values of @xmath140 .
the high-@xmath2 systems ( not only the cuprates but also the barium bismuthates and the fullerides ) are extreme type ii superconductors with local electromagnetic behaviour and with very large ginzburg ratio estimated to be of order @xmath180 @xcite .
thus , from the equality @xmath181 one can estimate the values of @xmath140 and @xmath6 which could be reliable for these materials .
examples of such estimations are given in fig .
8 and they indicate that best agreement can be obtained for intermediate values of the local attraction . in order to make some more detailed comparisons with experiment we have evaluated the plots @xmath182 vs. @xmath183 , where @xmath184 denotes the maximum critical temperature in the @xmath185@xmath186@xmath6@xmath186@xmath19 range and @xmath187 corresponds to @xmath188 . for a given @xmath140 the data corresponding to sq and sc lattices represent the one valued function ( see figs.9ab ) , whereas the ones for the fcc lattice ( see fig.9c ) exhibit substantial hysteresis
the behaviour for the fcc structure being due to the difference in positions of @xmath143 and @xmath189 maxima seen in fig.1b . at present the available experimental data @xcite correspond mainly to sq lattices ( cuprates ) and alternating cubic lattices ( chevrel phases , doped babi0@xmath190 ) and except of the overdoped cases they compare well with the present plots for these lattices in the intermediate coupling range ( @xmath191 ) .
the experimentally found deviations from the @xmath182 vs. @xmath183 universal dependence in the strongly overdoped regime can be explained either by the asymmetry in @xmath21 ( cf fig 9c ) due to further neighbour hopping in sc and sq structures or by the effects of intersite coulomb interactions @xcite , which can be different for various families of materials .
very recently locquet @xcite have reported the penetration depth measurements for la@xmath192 sr@xmath193cuo@xmath194 films as a function of doping , extending for the first time in the range from heavily underdoped to heavily overdoped . as we see from fig.10 the theoretical plots @xmath195 for our simple effective model for sq lattice fit surprisingly well these experimental data .
this work has financial supports from k.b.n .
poland , projects 2 p03b 104 11 , 2p03b05709 .
+ we are grateful to prof .
r. micnas for very helpful discussion .
one of us ( w.r.c . ) wants to thank to the center de recherches sur les trs basses tempratures ( crtbt ) of the centre national de la recherche scientifique in grenoble for hospitality and to the uesco
/ icsu / twas for a grant nr .
sc / rp 206.770.5 which made possible his stay in the crtbt during the course of this work .
thanks are also due to the french polish scientific and technological cooperation joint project for 1996 , proj .
6448 . for rectangular dos defined by ( [ 2d ] ) .
one can derive from eqs.([lambda]-[xi ] ) analytical expressions determining @xmath3 , @xmath4 , @xmath5 and @xmath94 as a function of @xmath6 and @xmath196 at @xmath22 .
+ in particular , from eqs.([delta]-[freeno ] ) taken at @xmath22 one obtains @xcite for @xmath218 the results for @xmath4 ( and consequently also for @xmath5 and @xmath94 ) are unrealistic as @xmath66 in this limit largely overestimated by mfa . as we have quoted in sec . 2
, one can correct these results by using eq.([en - strongu ] ) instead of ( [ a3 ] ) in the calculation of @xmath4 .
1 y. j. uemura et al .
nature * 352 * , 605 ( 1991 ) ; phys .
* 66 * , 2665 ( 1991 ) .
t. schneider , k. keller , phys .
69 * , 3374 ( 1992 ) ; int .
* b 8 * , 487 ( 1994 ) .
r. micnas , r. ranninger , s. robaszkiewicz , rev .
phys . * 62 * , 113 ( 1990 ) .
t. schneider , h. beck , d. bormann , t. meintrup , s. schafroth , a. schmidt , physica * c 216 * , 432 ( 1993 ) .
j. a. wilson , j. phys .
* c 20 * l 911(1987 ) ; ibid.*21 * , 2067 ( 1988 ) ; idem , physica * c 233 * , 322 ( 1994 ) .
n. f. mott , physica * a 200 * , 127 ( 1993 ) ; + idem , j. phys . condens .
matt . * 5 * , 3487 ( 1993 ) .
a. taraphder et al .
phys . rev . *
b 52 * , 1368 ( 1995 ) ; europhys .
* 21 * , 79 ( 1993 ) . c. m. varma , phys .
* 61 * , 2713 ( 1989 ) .
s. k. sarker , phys .
* b 49 * , 12047 ( 1994 ) .
f. c. zhang , m. ogata and t. m. rice , phys .
* 67 * , 3452 ( 1991 ) .
s. chakravarty , s. kivelson , europhys . lett .
* 16 * , 751 ( 1991 ) .
s. chakravarty , m. p. gelfand , s. kivelson , science * 254 * , 970 ( 1991 ) . j. a. wilson , physica * c 182 * , 1 ( 1991 ) .
t. kostyrko and r. micnas phys . rev . * b 46 * , 11025 ( 1992 ) . s. robaszkiewicz , r. micnas , k. a. chao , phys . rev . * b 24 * , 4018 ( 1981 ) .
a. l. fetter , j. d. walecka , _ quantum theory of many particle systems _ , mcgraw hill , new york , ( 1971 ) .
d. j. scalapino , s. r. white , s. zhang , phys . rev . * b 47 * , 7995 ( 1993 ) ; t. kostyrko , r. micnas , k. a. chao , phys . rev . * b 49 * , 6158 ( 1994 ) .
s. robaszkiewicz , t. kostyrko , r. micnas , physica * c 234 - 240 * , 1827 ( 1994 ) .
r. micnas , s. robaszkiewicz , t. kostyrko phys .
* b 52 * , 6863 ( 1995 ) .
w. czart , t. kostyrko and s. robaszkiewicz , j. mag . mag . mat .
* 140 - 144 * , 2059 ( 1995 )
. k. holczer , int .
* b 6 * , 3967 ( 1992 ) .
niedermayer , c. bernhard , u. binninger , h. glcker , j. l. tallon , e. j. ansaldo , j. i. budnick , phys .
lett . * 71 * , 1764 ( 1993 ) .
locquet et al .
( 1995 ) preprint ; t. schneider in `` polarons and bipolarons in high @xmath2 superconductors and related materials '' , ( eds .
e. k. h. salje , a. s. alexandrov , w. y. liang ) , + cambridge univ . press 1995 ) , p. 258 .
* figure captions + * * fig .
1 * concentration dependence of @xmath2 , the gap parameter @xmath1 and @xmath138 at @xmath219@xmath18@xmath185 k : @xmath140@xmath18@xmath220 ( @xmath147@xmath18 bandwidth ) ; * a ) * sc lattice , * b ) * fcc lattice , * c ) * model rectangular density of states .
( here : @xmath221 , @xmath178lattice constant , @xmath147bandwidth ) . + * fig .
2 * concentration dependence of @xmath138 at @xmath219@xmath18@xmath185 k for different values of @xmath222 for sq lattice . + * fig .
3a * reduced square value of thermodynamic critical field @xmath4 at @xmath22 ( @xmath223 ) as a function of @xmath140 for @xmath154 for sq lattice .
hfa ( @xmath224 ) , perturbation theory : @xmath225 order ( obtained using eqs.([hc],[en - strongu],[mfaess ] ) for @xmath226 ) ( @xmath227 ) , @xmath228 order ( obtained using eqs.([hc],[en - strongu],[mfaess ] ) for @xmath229 ( + ) . +
* fig . 3b * @xmath230 vs @xmath6 in @xmath22 plotted for increasing values of @xmath140 in hfa for sq lattice .
( numbers to the curves ) . + * fig .
4 * the g
l correlation length @xmath5 as a function of @xmath140 for @xmath231 for sq lattice .
hfa ( solid line ) , perturbation theory : @xmath225 order ( obtained by putting @xmath226 in eqs.([xi],[en - strongu],[mfaess ] ) ) ( dotted line ) , @xmath228 order ( obtained using eqs.([xi],[en - strongu],[mfaess ] ) for @xmath229 ( dashed line ) , @xmath22 ( @xmath232 ) . + * fig . 5 * @xmath233 vs @xmath6 for sq ( full circles ) and fcc ( triangles ) lattices : @xmath234 . + * fig .
6 * concentration dependence of @xmath235 ( @xmath106 at @xmath22 , @xmath236 ) calculated for sq lattice for several fixed values of @xmath140 ( numbers to the curves ) .
the corresponding plot obtained for a model rectangular dos for arbitrary @xmath140 is shown by a dashed line . + * fig .
7 * the boundary between local and nonlocal electromagnetic behaviour in the @xmath140 - @xmath6 parameter space calculated from the equation @xmath237 for @xmath238 ev , @xmath239 and @xmath240 for rectangular dos .
8 * the values of @xmath140 for the attractive hubbard model which provide the ginzburg ratio @xmath181 .
the plots as a function of @xmath6 for sq lattice , @xmath238 ev , @xmath239 , @xmath240 ( solid line ) and for rectangular dos @xmath238ev , @xmath241 ( dashed line ) . + * fig .
9 * @xmath242 vs. @xmath243 at @xmath219@xmath18@xmath185 k * a ) * for the sq lattice , * b ) * sc lattice @xmath244@xmath18@xmath19 ( ) , @xmath244@xmath18@xmath245 ( @xmath246 ) , @xmath244@xmath18@xmath247 ( - - ) and * c ) * for fcc lattice : @xmath248 ( ) , @xmath249 ( - - ) , @xmath250 ( ) .
@xmath187 corresponds to the maximum transition temperature @xmath188 .
( @xmath224)-tl@xmath251ba@xmath252 ca@xmath251cu@xmath253 o@xmath254 , tl@xmath255 pb@xmath256sr@xmath252 ca@xmath251cu@xmath253 o@xmath257 , bi@xmath258 pb@xmath193sr@xmath252 ca@xmath251cu@xmath253 o@xmath259 ; ( @xmath260)-y@xmath261 pr@xmath193ba@xmath252 cu@xmath262o@xmath263 ; ( @xmath264)-yba@xmath265 cu@xmath262o@xmath193 ; ( @xmath266)-la@xmath258 sr@xmath193cuo@xmath194 ; ( @xmath267)-bi@xmath251sr@xmath252 ca@xmath261 y@xmath193cu@xmath252 o@xmath268 ; ( @xmath269)-lamo@xmath270 se@xmath271 , pbmo@xmath272 s@xmath273 , snmo@xmath272 s@xmath194se@xmath194 , snmo@xmath270s@xmath274se , snmo@xmath270s@xmath275 se@xmath274 , lamo@xmath276 s@xmath273 , pbmo@xmath272 s@xmath194se@xmath194 ; ( @xmath277)-tl@xmath252 ba@xmath251 cuo@xmath278 .
experimental data taken from @xcite . + * fig .
10 * @xmath279 at @xmath219@xmath18@xmath185 k for sq lattice ( @xmath244@xmath18@xmath220 ( @xmath227 ) , @xmath244@xmath18@xmath245 ( @xmath224 ) ) , compared with experimental results ( full squares ) @xcite la@xmath192 sr@xmath193cuo@xmath194 films ( -@xmath280 nm ) ( squares ) , @xmath281$] . the results for a model square density of states almost coincide with the curve ( @xmath224 ) for arbitrary @xmath282 . | we study the basic thermodynamic and electromagnetic properties of the superconductor described by the negative@xmath0 hubbard model ( gap parameter @xmath1 , critical temperature @xmath2 , london penetration depth @xmath3 , thermodynamic critical field @xmath4 and ginzburg landau correlation length @xmath5 ) .
the calculations are performed for square ( sq , d=2 ) , simple cubic ( sc , d=3 ) and face centered cubic ( fcc , d=3 ) lattices .
we analyze the results as a function of electron density @xmath6 and interaction @xmath0 using the hartree
fock approximation in the weak to intermediate @xmath0 case combined with the conclusions from the perturbation theory valid in the strong coupling limit . from the equality : @xmath7 we find the boundaries between local and nonlocal electronic behaviour of the @xmath8 hubbard model in the @xmath9 parameter space . using the calculated values of @xmath2 and @xmath3
we compare the ` universal plot ' of @xmath10 vs. @xmath11 ^ 2 $ ] ( where : @xmath12 denotes maximum @xmath13 as a function of @xmath6 and @xmath14 corresponds to @xmath12 ) with the recent experimental data for various families of nonconventional superconductors and conclude that the best agreement can be obtained for intermediate values of the local attraction .
+ keywords : negative@xmath0 hubbard model , london penetration depth , high@xmath2 superconductivity superfluid characteristics of the attractive hubbard model for various lattice structures institute of physics , a.mickiewicz university , umultowska 85 , 61 - 614 pozna , poland .
+ pacs numbers : 74.30 , 74.20.-z , 71.28.+d , 74.25.ha |
the values of the exchange interactions between the magnetic moments carried by metallic atoms constitute a crucial piece of information for the theoretical investigation of the magnetic properties ( magnetization , susceptibility , specific heat , etc . ) of low - dimensional quantum s=@xmath3 systems , for instance vanadates.@xcite there are several ways to obtain this information .
one of them is based on the estimation of the exchange integrals from the hopping parameters obtained by fitting the band - structure with an extended tight - binding model @xcite or through a downfolding @xcite and projection @xcite procedure .
the first - principle method for the calculation of the exchange interaction parameters was proposed by lichtenstein _
@xcite within this scheme the exchange interaction parameters are determined by calculating the second variation of the total energy @xmath4 with respect to small deviations of the magnetic moments from the collinear magnetic configuration .
the first theoretical investigations @xcite of the magnetic couplings of na@xmath0v@xmath1o@xmath2 , a mott insulator with the rather exotic topology of a nanotube , were based on the former method .
they relied on different estimates of the hopping integrals and lead to conflicting results .
whangbo and koo @xcite have employed a spin dimer analysis in order to estimate the relative magnitudes of the spin exchange interaction parameters expected for na@xmath0v@xmath1o@xmath2 .
they have concluded that na@xmath0v@xmath1o@xmath2 could be described by six helical spin chains and should show a gap in the spin excitations .
dasgupta _ et al .
_ @xcite have performed the first _ ab - initio _ microscopic analysis of the electronic and magnetic properties of na@xmath0v@xmath1o@xmath2 using linear augmented plane wave ( lapw ) method.@xcite since , due to the geometry , the 3d _ xy _ bands of the vanadium atoms are well separated from other states , these authors have derived a v 3d _ xy _ orbital model hamiltonian using the downfolding procedure within the framework of the n - th order muffin - tin orbital method.@xcite based on the calculated hopping parameters between _ xy _ orbitals , they have estimated the exchange interactions between vanadium atoms in na@xmath0v@xmath1o@xmath2 using standard superexchange theory and proposed an effective heisenberg model . the calculated magnetic susceptibility as a function of temperature determined by exact diagonalization of finite clusters is in good agreement with experimental data . in this paper
we report a detailed investigation of the isotropic exchange interactions in na@xmath0v@xmath1o@xmath2 based on a first - principle lda+u calculation of its electronic and magnetic structure .
the exchange interactions between vanadium atoms were calculated using the green - function method developed by lichtenstein _
@xcite in order to provide a microscopic explanation of the resulting exchange interactions , which are _ qualitatively _ different from previous estimates , we have used an extended kugel - khomskii model that includes the hybridization between filled and vacant orbitals using the hopping parameters calculated by the projection procedure .
we conclude that estimating the exchange interactions of a system with a very exotic geometry such as na@xmath0v@xmath1o@xmath2 within simple superexchange between _ xy _ orbitals with on - site coulomb interaction is not possible .
the paper is organized as follows . in section
ii we shortly describe the crystal structure of na@xmath0v@xmath1o@xmath2 and present the results of the lda calculation . in section iii , we present the results of the lda+u calculation , and we discuss the origin of the exchange interaction parameters . in section iv , we briefly summarize our results .
a simplified view of the crystal structure of na@xmath0v@xmath1o@xmath2 is presented in fig.1 .
the main building blocks of the crystal structure are distorted vanadium - oxygen pyramids .
vanadium atoms are roughly in the center of a pyramid which is built by five oxygen atoms .
there are three types of vanadium atoms v1-v2-v3 which form a basic structure unit .
three of these units connected by edges are needed to form a ring of the tube .
the lowest energy orbital is the v 3d orbital of @xmath5 symmetry @xcite ( using a convention where the axes of the coordinate system are directed towards the oxygen ions in the plane ) , which is the orbital whose lobes point away from the oxygen .
more detailed information about the crystal structure of na@xmath0v@xmath1o@xmath2 can be found in ref .. v@xmath1o@xmath2 .
equivalent interaction paths have the same color .
the tube structure is made by putting m and n into contact.,scaledwidth=50.0% ] the electronic structure calculation of na@xmath0v@xmath1o@xmath2 was performed using the tight binding linear - muffin - tin - orbital atomic sphere approximation ( tb - lmto - asa ) method in terms of the conventional local - density approximation .
@xcite the band structure of na@xmath0v@xmath1o@xmath2 obtained from lda calculations is presented in fig.2 .
there are 18 bands near the fermi level which have @xmath5 symmetry ( since there are 18 vanadium atoms in the unit cell ) and are separated from the other v 3d bands as well as from the 2p oxygen bands .
the corresponding partial densities of states are shown in fig.3 .
v@xmath1o@xmath2 near the fermi level ( 0 ev).,scaledwidth=45.0% ] six bands have an energy significantly lower than the others , which are located near the fermi level .
two of them have an energy near -0.7 ev , while the other four are located between -0.3 and -0.4 ev .
such a band location of active orbitals ( occupied by electrons ) is unusual in comparison with the band structure of other s=@xmath3 transition metal compounds .
@xcite these bands are in good agreement with those presented in ref . . with one subset of energy bands well separated from the rest
, one can expect that a tight - binding model with a single _ xy_-orbital per v site ( i.e. a 18-band model ) will provide a good approximation to the full band structure close to the fermi level . and the total density of 3d states , respectively.,scaledwidth=45.0% ] for this we express the lda band structure results in terms of the low - energy model hamiltonian @xmath6 where _ i _ and _ j _ are site indexes , @xmath7 is the effective hopping integral between sites @xmath8 and @xmath9 , and @xmath10 is the operator which creates an electron with a spin @xmath11 in a state with a wannier function centered at the _ i_th atom . in order to obtain the hopping parameters @xmath7 , we have employed the projection procedure .
@xcite according to this procedure , the hopping integrals in the wannier function basis @xmath12 ( _ m _ is orbital index , _
i _ is site index ) , which is defined as the fourier transform of a certain linear combination of bloch functions @xmath13
( @xmath14 is the band index and k is the wave - vector in reciprocal space ) can be written in the general form : @xmath15 where @xmath16 is the eigenvalue in the basis of linearized muffin - tin orbitals .
the calculated inter - ring and intra - ring hopping parameters between 3d _ xy _ orbitals of vanadium atoms are presented in table i. .calculated hopping integrals between _ xy _ orbitals of vanadium atoms for the one - orbital model and estimated exchange interaction parameters @xmath17 ( @xmath18 , @xmath19 ) of na@xmath0v@xmath1o@xmath2 ( in mev ) . [ cols="<,^,^,^,^,^",options="header " , ] the values of the crystal - field splittings of different types of vanadium atoms obtained by using the projection procedure are presented in table v. if we take for the the splitting @xmath20 the average value of the crystal field splittings , namely 2 ev , the ferromagnetic contributions to the total exchange interactions estimated within the extended kugel - khomskii model are about 2 times smaller than those obtained using the green function method ( presented in table iv ) .
a physically more appealing approximation is to use in the model calculation the smallest value of the crystal field splitting , which corresponds to the first excitation energy for an electron between occupied and vacant orbitals . in our case
this splitting is about 1.1 ev .
the calculated contributions to the total exchange interactions between vanadium atoms for @xmath20 = 1.1 ev are presented in table vi .
one can see that we can qualitatively reproduce the anti- and ferromagnetic contributions to the total exchange interactions between vanadium atoms using this simple model except for bond ( v2 - 9)-(v3 - 18 ) . in order to obtain full qualitative agreement with the green function method
results one should use @xmath20 = 0.8 ev , which is slightly less than the crystal - field splitting calculated by the projection procedure .
we have investigated the electronic structure and the microscopic origin of the exchange interactions of na@xmath0v@xmath1o@xmath2 . based on our lda+u results
we predict that the value of the energy gap in electronic excitation spectra in this system is about 1 ev . the different contributions to the exchange interaction parameters have been calculated using the first - principle green - function method proposed by a.i . lichtenstein .
on the basis of the obtained results we argue that it is _ not _ correct to assume that the exchange interaction which is due to hybridization between active orbitals is a good approximation to the total exchange interaction between vanadium atoms for na@xmath0v@xmath1o@xmath2 .
according to the present results , the main interactions are antiferromagnetic along the ring , ferromagnetic between the rings , and of the same order of magnitude .
the resulting heisenberg model is strongly frustrated , with no obvious starting point for perturbation theory , hence no simple way to predict its properties .
such an analysis is left for future investigation .
we would like to thank a.i .
lichtenstein , i.v .
solovyev , and m.a .
korotin for helpful discussions .
the hospitality of the institute of theoretical physics of epfl is gratefully acknowledged .
this work is supported by intas young scientist fellowship program ref .
nr 04 - 83 - 3230 , netherlands organization for scientific research through nwo 047.016.005 , the scientific program `` development of scientific potential of universities '' section 3.3 project code 315 , russian foundation for basic research grant rffi 04 - 02 - 16096 and rffi 03 - 02 - 39024 .
the calculations were performed on the computer cluster of `` university center of parallel computing '' of ustu - upi .
we also acknowledge the financial support of the swiss national fund and of manep .
korotin , i.s .
elfimov , v.i .
anisimov , m. troyer , and d.i .
khomskii , phys .
* 83 , * 1387 ( 1999 ) ; m.a .
korotin , v.i .
anisimov , t. saha - dasgupta and i. dasgupta , j. phys . : condens . matter .
* 12 , * 113 ( 2000 ) . | we have performed first - principle calculations of the electronic structure and exchange couplings for the nanotube compound na@xmath0v@xmath1o@xmath2 using the lda+u approach .
our results show that while the intra - ring exchange interactions are mainly antiferromagnetic , the inter - ring couplings are _
ferromagnetic_. we argue that this is a consequence of the strong hybridization between filled and vacant 3d vanadium orbitals due to the low symmetry of na@xmath0v@xmath1o@xmath2 , which results into strong - and often dominant - ferromagnetic contributions to the total exchange interaction between vanadium atoms . a comparison with results of previous works is included . |
since the box covering method is central to the understanding of the scale - invariant properties of networks , we describe it in more detail here .
figure [ demo]a shows the same network as in fig . [ web]a for the case @xmath60 .
we tile the system by first assigning nodes 1 and 2 to the box colored in blue .
notice that the maximum distance between the nodes of a given box is @xmath66 .
thus , node 8 would not be in the blue box since its distance from node 2 is @xmath67 ( even though its distance from 1 is @xmath68 ) .
then we cover the nodes 6 and 7 with the orange box , and the nodes 3 , 4 , and 5 with the red box .
finally , the last node 8 is assigned to the green box .
the number of boxes to cover the network is then @xmath69 .
the renormalization is then applied by replacing each box by a single node .
thus , nodes 1 and 2 will be combined into a single node as indicated by the arrow from the first panel to the second panel in fig .
this renormalized node is connected with the orange and green boxes because there is a link between nodes 2 and 7 , and 1 and 8 , respectively .
the same rule applies to the other boxes .
the renormalized network is shown in the second panel .
the system is then tiled again with boxes ; in this case two boxes ( blue and red ) are needed to cover the entire network .
the two boxes are then replaced by nodes and a second renormalized network is obtained as shown in the third panel . finally , the last two nodes belong to the same ( red ) box and are replaced by a single node .
this procedure is applied to the www in fig .
the main panel corresponds to the first stage in the renormalization of the web for @xmath36 .
the procedure is applied again obtaining the remaining panels in fig .
[ web]b until the web is reduced to a single box in the last panel .
the colors of the nodes corresponds to the boxes to which they belong . in fig .
[ fractal]d we show the invariance of the degree distribution @xmath6 under the renormalization performed as a function of the box size in the www .
the other networks analyzed in this study present the same invariant property .
it is important to mention that the networks are also invariant under multiple renormalizations applied for a fixed box size @xmath11 .
this corresponds , for instance , to the stages depicted in fig .
[ web]a in the second row for @xmath36 for the network demo .
figure [ pk1 ] shows the invariance of @xmath6 for the www after several stages of the renormalization for a fixed @xmath36 , and it is the analogous of fig .
[ fractal]d for different box size .
the stages 1 , 2 , and 3 correspond to the networks depicted in the first 3 stages in fig .
[ web]b . from the above explanation
it should be clear that there are many ways to tile the network .
for instance in fig .
[ demo]b we show another tiling . in this case
we assign nodes 4 and 7 together in a single box instead of nodes 6 and 7 as in fig .
this tiling results in an extra box needed to cover node 6 and therefore in a larger number of nodes to tile the system , @xmath70 .
while there are many ways to assign nodes to the boxes , we notice that the rigorous mathematical definition of eq .
( [ dh ] ) corresponds to the _ minimum _ number of boxes needed to cover the network @xcite .
this minimization does not have any consequence for the determination of the fractal dimension in homogeneous clusters .
however , it may become relevant when calculating the self - similar exponent of a complex network with a _ widely _ distributed number of links . finding the minimum number of boxes to cover
the network is a hard optimization problem to solve , analogous to the graph coloring problem in the np - complete complexity class .
this minimization problem has to be solved by an exhaustive numerical search since there is no numerical algorithm to solve this kind of problems .
we have performed the search over a limited part of the phase - space for the www to obtain an estimation of the average and the minimum number of boxes needed to tile the network for every value of @xmath11 .
we find that the average value of the boxes is very close to the estimated minimum number of boxes .
moreover , we find that the minimization is not relevant and any covering gives rise to the same exponent .
the underlying meaning of the existence of scale - free networks which are self - similar is yet to be deciphered , but some insight can be gained by examining the simplest structure of a known network of that kind : a _ tree _ network which has been characterized using field theoretical arguments and fractal dimensions in @xcite . the sequence of renormalization steps depicted in fig . [ web ] suggests the following scheme : one begins with a single node and then constructs the network by applying the renormalization procedure in a reversed fashion .
this can be achieved by following the procedure in fig .
[ web ] for a specific value of @xmath11 .
more specifically , a single node with a large number of links is first connected to the next generation of nodes . for every node
we assign a number of links from a power - law distribution with a given @xmath7 .
the next layer of the tree is generated in the same way .
a tree structure with a power - law degree distribution and self - similar topology emerges which is depicted in fig .
[ tree]a .
this is corroborated numerically in fig .
[ tree]b where we study a scale - free tree structure with 192,827 nodes and @xmath71 , and we find @xmath72 and @xmath73 .
the parallels between the features of such a simple structured network and those discussed in this paper suggest that this simplified view may lie at the core of more complex self - similar networks .
moreover , we also calculate the average mass of the boxes and the mass of the clusters in the box covering method and the cluster covering method , respectively , and we find the power law of eq .
( [ mass_b ] ) and the exponential behaviour of eq .
( [ mass_c ] ) ( see fig .
[ mass_model]a ) in agreement with the results of the real networks analyzed in the main manuscript , fig .
figure [ mass_model]b shows the probability distribution of @xmath65 ( power - law ) and @xmath35 ( log - normal ) in agreement with previous results as well , fig . [ mass]b .
it is interesting to note that not all complex networks show the clear self - similarity of the networks presented so far .
we analyze the internet composed of computers and routers linked by physical lines such as the database collected by the scan project ( the `` mbone '' , www.isi.edu/scan/scan.html , we also analyze the database of the internet mapping project @xcite and found similar results ) .
figure [ internet ] shows the result of @xmath24 .
we fit the curve with a modified power - law @xmath74 with @xmath75 representing a cut - off and @xmath76 , suggesting a large self - similar exponent .
the decay of @xmath10 with @xmath11 is faster than a power - law and slower than exponential as shown in the inset of fig .
[ internet ] .
thus these networks lack the clear self - similar structure found for the www , actors and the biological networks .
however , we find that the distribution of @xmath77 remains a power law and the degree distribution @xmath6 is invariant under the renormalization suggesting that some self - similar properties might still be valid for the internet .
we notice that internet maps are made by programs that use the ip protocol to trace the connections between each registered node in the internet .
these maps are incomplete since they map a few routers from each domain and also due to the existence of firewalls .
thus , the apparent lack of self - similarity might be due to incomplete information of the network .
we also analyze the protein interaction networks of the fruit fly _ d. melanogaster _ as given in @xcite , the bacterium _ h. pylori _
@xcite , the baker s yeast _ s. cerevisiae _ @xcite , and the nematode worm _ c. elegans _
@xcite , which are all available via the dip database @xcite .
figure [ proteins ] shows the results of @xmath10 versus @xmath11 indicating that their behaviour is in between a pure power - law decay and a pure exponential . as with the internet data , we are able to fit the results with eq .
( [ modified ] ) with @xmath78 and @xmath79 for _ c. elegans_. for _ h. pylori _ and _ d. melanogaster _ the fit is a pure exponential @xmath80 with @xmath81 , while for _ s. cerevisiae _ the data could be fitted either by an exponential or by large values of @xmath82 and @xmath12 ( note that the exponential is the limit of eq .
( [ modified ] ) for @xmath83 , @xmath84 and @xmath85 constant ) . on the other hand
, we observe that for small scales , @xmath10 seems to display the same power law found for _ e. coli _ and _ h. sapiens_. the lack of clear self - similarity in these networks might be due to the incompleteness of these databases which are continuously being updated with newly discovered physical interactions @xcite .
next we introduce an example of a model lacking self - similarity : the random scale - free model .
this model consists of nodes to which a number of links are assigned with a power - law degree distribution and then connected randomly .
such a network shows a small world effect and a scale - free property but is not self - similar .
we numerically find that the number of boxes decays exponentially with the box size ( see fig .
[ tree]b ) .
moreover , while eq .
( [ s ] ) is still valid in this case , the power law relation in eq .
( [ sl ] ) is replaced by an exponential law .
we conjecture that the reason for this is a clustering of hubs ; by assigning randomly the connections between the nodes , two nodes with a large number of links will have a large probability to be connected .
this induces spatial correlations in the values of @xmath40 which may explain the breakdown of self - similarity .
in contrast , the simple tree - structure proposed above does not cluster the hubs by construction .
a summary of our results is presented in table [ table ] .
we also analyzed the barabsi - albert model of complex networks @xcite ( which introduces the concepts of preferential attachment to describe the dynamics of scale - free networks ) .
the results of @xmath24 are shown in fig .
[ abmodel ] for different parameters in the model ( see @xcite for details ) reveling that the structure is not self - similar ; @xmath10 seems to decrease faster than exponential with @xmath11 .
it is interesting to compare our results with the random erds - rnyi graph @xcite at the critical percolation threshold . in this case
the largest cluster has self - similar properties and eq .
( [ mass_b ] ) , @xmath86 , is valid with @xmath87 @xcite .
we corroborate this result in fig .
[ erc ] showing the scaling of the number of boxes @xmath10 with the box size @xmath11 .
however , for this case the network is not small - world since eq .
( [ mass_c ] ) is not valid as well as eq .
( [ smallworld]) but rather the mean distance @xmath0 scales as @xmath88 , i.e. , a power - law relation rather than the logarithmic relation characteristic of small world networks .
.[table ] summary of the exponents obtained for the scale - invariant networks studied in the manuscript . [ cols="<,^,^,^,^,^ " , ]
the wit database @xcite ( http://igweb.integratedgenomics.com/igwit ) of cellular networks considers the cellular functions divided according to bioengineering principles containing datasets for intermediate metabolism and bioenergetics ( core metabolism ) , information pathways , electron transport , and transmembrane transport .
the metabolic network is a subset of all reactions that take place in the cell .
since this is the largest part of the network we analyze it separately and compare it with the full biochemical reaction network .
the data presented in fig .
[ fractal]c represents the full biochemical reaction networks of only three substrates .
here we present results of the 43 different substrates represented in the database for the metabolic and full networks .
the following figures show the results of @xmath10 vs @xmath11 .
both the metabolic and full networks display the power law relationship of self - similar networks with the same exponent ( within error bars ) for all the organisms considered ( the metabolic networks show a finite size effect due to their smaller size ) .
we find an average @xmath28 .
the solid line in the figures represent the average fit .
the values are reported in table [ table ] . | complex networks have been studied extensively due to their relevance to many real systems as diverse as the world - wide - web ( www ) , the internet , energy landscapes , biological and social networks @xcite .
a large number of real networks are called `` scale - free '' because they show a power - law distribution of the number of links per node @xcite .
however , it is widely believed that complex networks are not _ length - scale _ invariant or self - similar .
this conclusion originates from the `` small - world '' property of these networks , which implies that the number of nodes increases exponentially with the `` diameter '' of the network @xcite , rather than the power - law relation expected for a self - similar structure .
nevertheless , here we present a novel approach to the analysis of such networks , revealing that their structure is indeed self - similar .
this result is achieved by the application of a renormalization procedure which coarse - grains the system into boxes containing nodes within a given `` size '' .
concurrently , we identify a power - law relation between the number of boxes needed to cover the network and the size of the box defining a finite self - similar exponent .
these fundamental properties , which are shown for the www , social , cellular and protein - protein interaction networks , help to understand the emergence of the scale - free property in complex networks .
they suggest a common self - organization dynamics of diverse networks at different scales into a critical state and in turn bring together previously unrelated fields : the statistical physics of complex networks with renormalization group , fractals and critical phenomena
. two fundamental properties of real complex networks have attracted much attention recently : the small - world and the scale - free properties .
many naturally occurring networks are small world since one can reach a given node from another one , following the path with the smallest number of links between the nodes , in a very small number of steps .
this corresponds to the so - called `` six degrees of separation '' in social networks @xcite .
it is mathematically expressed by the slow ( logarithmic ) increase of the average diameter of the network , @xmath0 , with the total number of nodes @xmath1 , @xmath2 , where @xmath3 is the _ shortest _ distance between two nodes and defines the distance metric in complex networks @xcite .
equivalently , we obtain : @xmath4 where @xmath5 is a characteristic length .
a second fundamental property in the study of complex networks arises with the discovery that the probability distribution of the number of links per node , @xmath6 ( also known as the degree distribution ) , can be represented by a power - law ( scale - free ) with a degree exponent @xmath7 usually in the range @xmath8 @xcite , @xmath9 these discoveries have been confirmed in many empirical studies of diverse networks @xcite . with the aim of providing a deeper understanding of the underlying mechanism which leads to these common features one needs to probe the patterns within the network structure in more detail .
the question of connectivity between groups of interconnected nodes on different length - scales has received less attention .
yet , a plethora of examples in nature exhibits the importance of collective behavior , from interactions between communities within social networks , links between clusters of web - sites of similar subjects , all the way to the highly modular manner in which molecules interact to keep a cell alive .
here we show that real complex networks , such as www , social , protein - protein interaction networks ( pin ) and cellular networks are indeed constructed of self - repeating patterns on all length - scales , and are therefore invariant or self - similar under a length - scale transformation .
this result comes as a surprise since the exponential increase in eq .
( [ smallworld ] ) has led to the general understanding that complex networks are not self - similar , since self - similarity requires a power - law relation between @xmath1 and @xmath3 .
how can one reconcile the exponential increase in eq .
( [ smallworld ] ) with self - similarity , or in other words an underlying _ length_-scale - invariant topology ? at the root of the self - similar properties that we unravel in this study is a scale - invariant renormalization procedure which we show to be valid for dissimilar complex networks . in order to demonstrate this concept we first consider a self - similar network embedded in euclidean space ,
of which a classical example would be a fractal percolation cluster at criticality @xcite . in order to unfold the self - similar properties of such clusters we calculate the fractal dimension using a `` box counting '' method and a `` cluster growing '' method @xcite . in the first method we cover the percolation cluster with @xmath10 boxes of linear size @xmath11 .
the fractal dimension or box dimension @xmath12 is then given by @xcite : @xmath13 in the second method , the network is not covered with boxes , instead one seed node is chosen at random and a cluster of nodes centered at the seed and separated by a minimum distance @xmath3 is calculated .
the procedure is then repeated by choosing many seed nodes at random and the average `` mass '' of the resulting clusters ( @xmath14 , defined as the number of nodes in the cluster ) is calculated as a function of @xmath3 to obtain the following scaling : @xmath15 defining the fractal cluster dimension @xmath16 @xcite . comparing eq .
( [ df ] ) and ( [ smallworld ] ) implies that @xmath17 for complex small - world networks . for a _ homogeneous _ network characterized by a _
narrow _ degree distribution ( such as a fractal percolation cluster ) the box covering method of eq .
( [ dh ] ) and the cluster growing method of eq .
( [ df ] ) are equivalent since every node typically has the same number of links or neighbors . equation ( [ df ] ) can then be derived from ( [ dh ] ) and @xmath18 , and this relation has been regularly used . the crux of the matter is to understand how one can calculate a self - similar exponent ( analogous to the fractal dimension in euclidean space ) in complex _ inhomogeneous _ networks with a _ broad _ degree distribution such as eq .
( [ scale - free ] ) . under these conditions eqs .
( [ dh ] ) and ( [ df ] ) are not equivalent as will be shown below .
the application of the proper covering procedure in the box counting method , eq . ( [ dh ] ) , for complex networks unveils a set of self - similar properties such as a finite self - similar exponent and a new set of critical exponents for the scale - invariant topology .
figure [ web]a illustrates the box covering method using a schematic network composed of 8 nodes . for each value of the box size @xmath11
, we search for the number of boxes needed to tile the entire network such that each box contains nodes separated by a distance @xmath19 .
this procedure is applied to several different real networks : _ ( i ) _ a part of the www composed of 325,729 web - pages which are connected if there is a url link from one page to another @xcite ( http://www.nd.edu/@xmath20networks ) , _ ( ii ) _ a social network where the nodes are 392,340 actors linked if they were cast together in at least one movie @xcite , _ ( iii ) _ the biological networks of protein - protein interactions found in _
e. coli _ ( 429 proteins ) and _ h. sapiens _
( human ) ( 946 proteins ) linked if there is a physical binding between them ( database available via the database of interacting proteins @xcite , other pins are discussed in the supplementary materials ) , and _
( iv ) _ the cellular networks compiled by @xcite using a graph - theoretical representation of the whole biochemical pathways based on the wit integrated - pathway genome database @xcite ( http://igweb.integratedgenomics.com/igwit ) of 43 species from archaea , bacteria , and eukarya . here
we show the results for _ a. fulgidus , e. coli _ and _ c. elegans _
@xcite , while the full database is analyzed in the supplementary materials .
it has been previously determined that the www and actors networks are small - world and scale - free , characterized by eq .
( [ scale - free ] ) with @xmath21 and 2.2 , respectively @xcite .
for the pins of _ e. coli _ and _ h. sapiens _ we find @xmath22 and @xmath23 , respectively .
all cellular networks are scale - free with average exponent @xmath22 @xcite .
we confirm these values and show the results for the www in fig .
[ fractal ] .
figures [ fractal]a and [ fractal]b show the results of @xmath24 according to eq . ( [ dh ] ) .
they reveal the existence of self - similarity in the www , actors , and _ e. coli _ and _ h. sapiens _ protein - protein interaction networks with self - similar exponents @xmath25 , @xmath26 and @xmath27 and @xmath27 , respectively .
the cellular networks are shown in fig .
[ fractal]c and have @xmath28 .
we now elaborate on the apparent contradiction between the two definitions of self - similar exponents in complex networks . after performing a renormalization at a given @xmath11
, we calculate the mean mass of the boxes covering the network , @xmath29 , to obtain @xmath30 which is corroborated by direct measurements for all the networks and shown in fig .
[ mass]a for the www .
on the other hand , the average performed in the cluster growing method ( for this calculation we average over single boxes without tiling the system ) gives rise to an exponential growth of the mass @xmath31 with @xmath32 in accordance with the small - world effect eq .
( [ smallworld ] ) , as seen in fig .
[ mass]a .
the topology of scale - free networks is dominated by several highly connected hubs the nodes with the largest degree implying that most of the nodes are connected to the hubs via one or very few steps .
therefore the average performed in the cluster growing method is biased ; the hubs are overrepresented in eq .
( [ mass_c ] ) since almost every node is a neighbor of a hub . by choosing the seed of the clusters at random , there is a very large probability of including the hubs in the clusters . on the other hand
the box covering method is a global tiling of the system providing a flat average over all the nodes , i.e. each part of the network is covered with an equal probability .
once a hub ( or any node ) is covered , it can not be covered again .
we conclude that eqs .
( [ dh ] ) and ( [ df ] ) are not equivalent for inhomogeneous networks with topologies dominated by hubs with a large degree . the biased sampling of the randomly chosen nodes is clearly demonstrated in fig . [ mass]b .
we find that the probability distribution of the mass of the boxes for a given @xmath11 is very broad and can be approximated by a power - law : @xmath33 in the case of www and @xmath34 . on the other hand
, the probability distribution of @xmath35 is very narrow and can be fitted by a log - normal distribution ( see fig . [
mass]b ) . in the box covering method
there are many boxes with very large and very small masses in contrast to the peaked distribution in the cluster growing method , thus showing the biased nature of the latter method in inhomogeneous networks .
this biased average leads to the exponential growth of the mass in eq .
( [ mass_c ] ) and it also explains why the average distance is logarithmic with @xmath1 as in eq .
( [ smallworld ] ) .
the box counting method provides a powerful tool for further investigations of the network properties as it enables a renormalization procedure , revealing that the self - similar properties and the scale - free degree distribution persists irrespective of the amount of coarse - graining of the network .
subsequent to the first step of assigning the nodes to the boxes we create a new renormalized network by replacing each box by a single node .
two boxes are then connected , provided that there was at least one link between their constituent nodes .
the second column of the panels in fig .
[ web]a shows this step in the renormalization procedure for the schematic network , while fig .
[ web]b shows the results for the same procedure applied to the entire www for @xmath36 .
the renormalized network gives rise to a new probability distribution of links , @xmath37 , which is invariant under the renormalization : @xmath38 figure [ fractal]d supports the validity of this scale transformation by showing a data collapse of all distributions with the same @xmath7 according to ( [ pk ] ) for the www
. further insight arises from relating the scale - invariant properties ( [ dh ] ) to the scale - free degree distribution ( [ scale - free ] ) . plotting ( see inset in fig .
[ fractal]d for the www ) the number of links @xmath39 of each node in the renormalized network versus the maximum number of links @xmath40 in each box of the unrenormalized network exhibits a scaling law @xmath41 this equation defines the scaling transformation in the connectivity distribution .
empirically we find that the scaling factor @xmath42 scales with @xmath11 with a new exponent @xmath43 as @xmath44 shown in fig .
[ fractal]a for the www and actor networks ( with @xmath45 and @xmath46 , respectively ) , in fig .
[ fractal]b for the protein networks ( @xmath47 for _ e. coli _ and @xmath48 for _ h. sapiens _ ) and in fig .
[ fractal]c for the cellular networks with @xmath49 .
equations ( [ s ] ) and ( [ sl ] ) shed light on how families of hierarchical sizes are linked together .
the larger the families , the fewer links exist .
surprisingly the same power - law relation exists for large and small families represented by eq .
( [ scale - free ] ) .
from eq .
( [ pk ] ) we obtain @xmath50 , where @xmath51 is the number of nodes with links @xmath40 and @xmath52 is the number of nodes with links @xmath39 after the renormalization ( @xmath53 is the total number of nodes in the renormalized network ) . using eq .
( [ s ] ) we obtain @xmath54 .
then , upon renormalizing a network with @xmath1 total nodes we obtain a smaller number of nodes @xmath53 according to @xmath55 . since the total number of nodes in the renormalized network is the number of boxes needed to cover the unrenormalized network at any given @xmath11
, we have @xmath56 .
then , from eqs .
( [ dh ] ) and ( [ sl ] ) we obtain the relation between the three indexes @xmath57 equation ( [ relation ] ) is confirmed for all the networks analyzed here ( see supplementary materials ) . in all cases
the calculation of @xmath12 and @xmath43 and eq .
( [ relation ] ) gives rise to the same @xmath7 exponent as that obtained in the direct calculation of the degree distribution .
the significance of this result is that the scale - free properties characterized by @xmath7 can be related to a more fundamental length - scale invariant property , characterized by the two new indexes @xmath58 and @xmath59 .
summarizing , we elucidate a fundamental property of a wide variety of complex networks : that of a scale - invariant topology . concepts first introduced for the study of critical phenomena in statistical physics are shown to be valid here in the characterization of a different class of phenomena : the topology of complex networks .
one could envisage a great deal of fundamental information being understood by the application of renormalization techniques to this kind of complex system .
for instance , networks with similar degree distributions are characterized by different self - similar exponents , thus indicating that they may belong to different `` universality classes '' .
it is as though each node ( ranging from web - pages in the www , to people in social networks , to proteins and substrates in cellular networks ) were connected to other nodes under a single self - organizing principle according to which groups of nodes of all sizes self - organize too ; such that everything links with everything else , governed by one universal dynamics in nature @xcite .
acknowledgments : this work was supported by the national science foundation .
albert r. & barabsi , a .-
l .
statistical mechanics of complex networks . _ rev .
mod .
phys _ * 74 * , 47 - 97 ( 2002 ) .
dorogovtsev , s. n. & mendes , j. f. f. _ evolution of networks : from biological nets to the internet and the www _
( oxford university press , oxford , 2003 ) .
pastor - satorras , r. & vespignani , a. _ evolution and structure of the internet : a statistical physics approach _ ( cambridge university press , cambridge , 2004 ) .
newman , m. e. j. the structure and function of complex networks .
_ siam review _ * 45 * , 167 - 256 ( 2003 ) .
amaral , l. a. n. & ottino , j. m. complex networks - augmenting the framework for the study of complex systems . _
eur .
phys .
j. b _ * 38 * , 147 - 162 ( 2004 ) .
albert , r. jeong , h. & barabsi , a .-
l .
diameter of the world wide web
. _ nature _ * 401 * , 130 - 131 ( 1999 ) .
faloutsos , m. , faloutsos , p. & faloutsos , c. on power - law relationships of the internet topology .
_ computer communications review _ * 29 * , 251 - 262 ( 1999 ) .
erds , p. & rnyi , a. on the evolution of random graphs .
_ publ .
math .
inst . hung .
acad .
sci . _ * 5 * , 17 - 61 ( 1960 ) .
bollobs , b. _ random graphs _ ( academic press , london , 1985 ) .
milgram , s. _ psychol . today
_ * 2 * , 60 ( 1967 ) .
watts , d. j. & strogatz , s. h. collective dynamics of `` small - world '' networks .
_ nature _ * 393 * , 440 - 442 ( 1998 ) .
bunde , a. & havlin , s. _ fractals and disordered systems _ , edited by a. bunde and s. havlin , 2nd edition ( springer - verlag , heidelberg , 1996 ) .
vicsek , t. _ fractal growth phenomena _ , 2nd ed . , part iv ( world scientific , singapore , 1992 ) . feder , j. _ fractals _ ( plenum press , new york , 1988 ) .
barabsi , a .-
l . &
albert , r. emergence of scaling in random networks .
_ science _ * 286 * , 509 - 512 ( 1999 ) .
xenarios , i. _ et al .
_ dip : the database of interacting proteins . _
nucleic acids res .
_ * 28 * , 289 - 291 ( 2000 ) . (
dip ) .
http://dip.doe-mbi.ucla.edu
jeong , h , tombor , b. , albert , r. , oltvai z. n. & barabsi , a .-
l .
the large - scale organization of metabolic networks .
_ nature _ * 407 * , 651 - 654 ( 2000 ) .
overbeek , r. _ et al .
_ wit : integrated system for high - throughput genome sequence analysis and metabolic reconstruction .
_ nucleic acid res .
_ * 28 * , 123 - 125 ( 2000 ) .
schopenhauer , a. transcendent speculation on the apparent deliberateness in the fate of the individual , in _ parerga and paralipomena : short philosophical essays _
, vol .
i , pp .
199 - 225 ( oxford university press , oxford , 1974 ) .
burda , z. , correia , j. d. & krzywicki , a. statistical ensemble of scale - free random graphs _ phys .
rev .
e _ * 64 * , 046118 ( 2001 ) .
burch , h. & cheswick , w. mapping and visualizing the internet .
_ ieee computer _ * 32 * , 4 ( 1999 ) .
http://research.lumeta.com/ches/map giot , l. _ et al . _ a protein interaction map of _ drosophila melanogaster_. _ science _ * 302 * , 1727 - 1736 ( 2003 ) .
rain , j .- c .
_ et al . _ the protein - protein interaction map of _ helicobacter pylori_. _ nature _ * 409 * , 211 - 215 ( 2001 ) .
uetz , p. _ et al . _ a comprehensive analysis of protein - protein interactions in _
saccharomyces cerevisiae_. _ nature _ * 403 * , 623 - 627 ( 2000 ) .
li , s. _ et al .
_ a map of the interactome network of the metazoan _ c. elegans_. _ science _ * 303 * , 540 - 543 ( 2004 ) .
braunstein , l. , buldyrev , s. v. , cohen , r. , havlin , s. & stanley , h. e. optimal paths in disordered complex networks .
_ phys .
rev .
lett . _ * 91 * , 168701 ( 2003 ) .
* acknowledgements*. sh wishes to thank the israel science foundation for support .
this work is supported by the national science foundation , dmr-0239504 .
fig [ web ] .
the renormalization procedure to complex networks . * a , * demonstration of the method for different @xmath11 and different stages in a network demo .
the first column depicts the original network .
we tile the system with boxes of size @xmath11 ( different colors correspond to different boxes ) .
all nodes in a box are connected by a minimum distance smaller than the given @xmath11 .
for instance , in the case of @xmath60 , we identify four boxes which contain the nodes depicted with colors red , orange , white , and blue , each containing 3 , 2 , 1 , and 2 nodes , respectively .
then we replace each box by a single node ; two renormalized nodes are connected if there is at least one link between the unrenormalized boxes .
thus we obtain the network shown in the second column . the resulting number of boxes needed to tile the network , @xmath24 , is plotted in fig . [ fractal ] versus @xmath11 to obtain @xmath12 as in eq .
( [ dh ] ) . the renormalization procedure is applied again and repeated until the network is reduced to a single node ( third and fourth columns for different @xmath11 ) .
* b , * three stages in the renormalization scheme applied to the entire www .
we fix the box size to @xmath61 and apply the renormalization for four stages .
this corresponds , for instance , to the sequence for the network demo depicted in the second row in part * a * of this figure .
we color the nodes in the web according to the boxes to which they belong .
the network is invariant under this renormalization as explained in the legend of fig .
[ fractal]d and the supplementary materials .
fig [ fractal ] .
self - similar scaling in complex networks . *
a , * upper panel : log - log plot of the @xmath10 vs @xmath11 revealing the self - similarity of the www and actor network according to eq .
( [ dh ] ) .
lower panel : the scaling of @xmath62 vs. @xmath11 according to eq .
( [ sl ] ) .
the errors bars are of the order of the symbol size .
* b , * same as ( a ) but for two protein interaction networks : _ h. sapiens _ and _
e. coli_. results are analogous to ( b ) but with different scaling exponents .
* c , * same as ( a ) for the cellular networks of _ a. fulgidus _ , _
e. coli _ and _ c. elegans_. * d , * invariance of the degree distribution of the www under the renormalization for different box sizes , @xmath11 .
we show the data collapse of the degree distributions demonstrating the self - similarity at different scales .
the inset shows the scaling of @xmath63 for different @xmath11 , from where we obtain the scaling factor @xmath62 .
moreover , we also apply the renormalization for a fixed box size , for instance @xmath36 as shown in fig .
[ web]b for the www , until the network is reduced to a few nodes and find that @xmath6 is invariant under these multiple renormalizations as well , for several iterations ( see supplementary materials ) .
fig .
[ mass ] .
different averaging techniques lead to qualitatively different results . * a , * mean value of the box mass in the box counting method , @xmath64 , and the cluster mass in the cluster growing method , @xmath14 , for the www .
the solid lines represent the power - law fit for @xmath64 and the exponential fit for @xmath14 according to eqs .
( [ mass_b ] ) and ( [ mass_c ] ) , respectively .
* b , * probability distribution of @xmath65 and @xmath35 for @xmath34 for the www .
the curves are fitted by a power - law and a log - normal distribution , respectively . * a * * b * * b * * supplementary materials * |
in a recent paper , onatski , moreira and hallin ( 2011 ) ( hereafter omh ) analyze the asymptotic power of statistical tests in the detection of a signal in spherical real - valued gaussian data as the dimensionality of the data and the number of observations diverge to infinity at the same rate .
this paper generalizes omh s alternative of a single symmetry - breaking direction ( _ single - spiked _ alternative ) to the alternative of multiple symmetry - breaking directions ( _ multispiked _ alternative ) , which is more relevant for applied work .
contemporary tests of sphericity in a high - dimensional environment ( see ledoit and wolf ( 2002 ) , srivastava ( 2005 ) , schott ( 2006 ) , bai et al . (
2009 ) , chen et al .
( 2010 ) , and cai and ma ( 2012 ) ) consider general alternatives to the null of sphericity .
our interest in alternatives with only a few contaminating signals stems from the fact that in many applications , such as speech recognition , macroeconomics , finance , wireless communication , genetics , physics of mixture , and statistical learning , a few latent variables typically explain a large portion of the variation in high - dimensional data ( see baik and silverstein ( 2006 ) for references ) . as a possible explanation of this fact , johnstone ( 2001 ) introduces the spiked covariance model , where all eigenvalues of the population covariance matrix of high - dimensional data are equal except for a small fixed number of distinct spike eigenvalues .
the alternative to the null of sphericity considered in this paper coincides with johnstone s model .
the extension from the single - spiked alternatives of omh to the multi - spiked alternatives considered here , however , is all but straightforward .
the difficulty arises because the extension of the main technical tool in omh ( lemma 2 ) , which analyzes high - dimensional spherical integrals , to integrals over high - dimensional real stiefel manifolds obtained in onatski ( 2012 ) is not easily amenable to the laplace approximation method used in omh .
therefore , in this paper , we develop a completely different technique , inspired from the large deviation analysis of spherical integrals by guionnet and maida ( 2005 ) .
let us describe the setting and main results in more detail .
suppose that the data consist of @xmath0 independent observations @xmath1 @xmath2 of a @xmath3-dimensional gaussian vector with mean zero and positive definite covariance matrix @xmath4 .
let @xmath5 where @xmath6 is the @xmath7-dimensional identity matrix , @xmath8 is a scalar , @xmath9 an @xmath10 diagonal matrix with elements @xmath11 @xmath12 along the diagonal , and @xmath13 a @xmath14-dimensional parameter normalized so that @xmath15 .
we are interested in the asymptotic power of tests of the null hypothesis @xmath16 against the alternative @xmath17 @xmath18 for some @xmath19 based on the eigenvalues of the sample covariance matrix of the data when @xmath20 so that @xmath21 with @xmath22 , an asymptotic regime which we abbreviate into @xmath23 .
the matrix @xmath13 is an unspecified nuisance parameter , the columns of which indicate the directions of the perturbations of sphericity .
we consider the cases of specified and unspecified @xmath24 for the sake of simplicity , in the rest of this introduction , we only discuss the case of specified @xmath25 although the case of unspecified @xmath26 is more realistic . denoting by @xmath27 the @xmath28-th largest sample covariance eigenvalue ,
let @xmath29 where @xmath30 .
we begin our analysis with a study of the asymptotic properties of the likelihood ratio process @xmath31 ^{r}\right\ } , $ ] where @xmath32 @xmath33 and @xmath34 is defined as the ratio of the density of @xmath35 under @xmath36 to that under @xmath37 , considered as a @xmath38-measurable random variable .
note that @xmath34 depends on @xmath0 and @xmath39 while @xmath35 is @xmath40 -dimensional .
an exact formula for @xmath34 involves the integral @xmath41 over the orthogonal group @xmath42 , where the @xmath43 matrix @xmath44 has a deficient rank @xmath45 . in the single - spiked case ( @xmath46 ) , omh link this integral to the confluent form of the lauricella function , and
use this link to establish a representation of the integral in the form of a contour integral ( see wang ( 2010 ) and mo ( 2011 ) for independent different derivations of this contour integral representation for this particular @xmath46 case ) . then
, the laplace approximation to the contour integral is used to derive the asymptotic behavior of @xmath34 .
onatski ( 2012 ) generalizes the contour integral representation to the multispiked case ( @xmath47 ) . for complex - valued data ,
such a generalization allows him to extend omh s results to the multi - spiked context .
unfortunately , for real - valued data , which we are concerned with in this paper , this generalization is not straightforwardly amenable to the laplace approximation method . therefore , in this paper , we consider a totally different approach . for the @xmath46 case , guionnet and maida ( 2005 ) ( hereafter gm ) use large deviation methods to derive a second - order asymptotic expansion of @xmath41 as the non - zero eigenvalues of @xmath44 diverge to infinity ( see their theorem 3 ) .
we extend gm s second - order expansion to the @xmath47 case , and use that extension to derive the asymptotics of @xmath34 .
more precisely , we show that , for any @xmath48 such that @xmath49 the sequence of log - likelihood processes @xmath50^{r}\}$ ] converges weakly to a gaussian process @xmath51 ^{r}\right\ } $ ] under the null hypothesis @xmath37 as @xmath23 . the index @xmath35 in the notation @xmath52 is used to distinguish the limiting @xmath35-log - likelihood process in the case of specified @xmath25 from that of the @xmath53-log - likelihood process considered in the case of unspecified @xmath54 which we denote by @xmath55 ( see section 2 ) .
the limiting process has mean @xmath56 = % \frac{1}{4}\sum\limits_{i , j=1}^{r}\ln \left ( 1-h_{i}h_{j}/c\right ) $ ] and autocovariance function @xmath57 that convergence entails the weak convergence , in the le cam sense , of the @xmath58-indexed statistical experiments @xmath59 under which the eigenvalues @xmath60 are observed , i.e. the statistical experiments with log - likelihood process @xmath61 ^{r}\right\ } $ ] ( see van der vaart ( 1998 ) , page 126 ) .
although this limiting process is gaussian , it is not a log - likelihood process of the gaussian shift type , so that the statistical experiments @xmath59 under study are not locally asymptotically normal ( lan ) ones .
the weak convergence of @xmath62 implies , however , via le cam s first lemma ( see van der vaart 1998 , p.88 ) , that the joint distributions of the normalized sample covariance eigenvalues under the null and under alternatives associated with @xmath63 are mutually contiguous .
an asymptotic power envelope for @xmath35-based tests of @xmath37 against @xmath64 can be constructed using the neyman - pearson lemma and le cam s third lemma .
we show that , for tests of size @xmath65 the maximum achievable asymptotic power against a point alternative @xmath66 equals @xmath67 , $ ] where @xmath68 is the standard normal distribution function and @xmath69 . as we explain in the paper ,
this asymptotic power envelope is valid not only for the @xmath35-based tests , but also for all tests that are invariant under left orthogonal transformations of the data @xmath1 @xmath2 .
next , we consider previously proposed tests of sphericity and of the equality of the population covariance matrix to a given matrix .
we focus on the tests studied in ledoit and wolf ( 2002 ) , bai et al ( 2009 ) , and cai and ma ( 2012 ) .
we find that , in general , the asymptotic powers of those tests are substantially lower than the corresponding asymptotic power envelope value .
in contrast , our computations for the case @xmath70 show that the asymptotic powers of the @xmath35- and @xmath53-based likelihood ratio tests are close to the power envelope .
the rest of the paper is organized as follows .
section 2 establishes the weak convergence of the log - likelihood ratio process to a gaussian process .
section 3 provides an analysis of the asymptotic powers of various sphericity tests , derives the asymptotic power envelope , and proves its validity for general invariant tests .
section 4 concludes .
all proofs are given in the appendix .
let @xmath71 be a @xmath72 matrix with independent gaussian @xmath73 columns .
let @xmath74 be the ordered eigenvalues of @xmath75 and write @xmath76 where @xmath77 .
similarly , let @xmath78 @xmath79 and @xmath80 . as explained in the introduction ,
our goal is to study the asymptotic power , as @xmath81 of the eigenvalue - based tests of @xmath82 against @xmath83 for some @xmath84 , where @xmath85 are the diagonal elements of the diagonal matrix @xmath9 .
if @xmath86 is specified , the model is invariant with respect to left and right orthogonal transformations ; sufficiency and invariance arguments ( see appendix 5.4 for details ) lead to considering tests based on @xmath87 only .
if @xmath88 is unspecified , the model is invariant with respect to left and right orthogonal transformations and multiplications by non - zero scalars ; sufficiency and invariance arguments ( see appendix 5.4 ) lead to considering tests based on @xmath89 only .
note that the distribution of @xmath90 does not depend on @xmath54 whereas , if @xmath88 is specified , we can always normalize @xmath87 dividing it by @xmath91 therefore , we henceforth assume without loss of generality that @xmath92 .
let us denote the joint density of @xmath93 at @xmath94 as @xmath95 , and that of @xmath96 at @xmath97 as @xmath98 .
we have@xmath99where @xmath100 depends only on @xmath101 and @xmath7 ; @xmath102 ; @xmath103 is a @xmath104 diagonal matrix ; @xmath105 is the set of all @xmath43 orthogonal matrices ; and @xmath106 is the invariant measure on the orthogonal group @xmath105 , normalized to make the total measure unity .
formula ( [ common complex1 ] ) is a special case of the density given in james ( 1964 , p.483 ) for @xmath107 and follows from theorems 2 and 6 in uhlig ( 1994 ) for @xmath108 .
let @xmath109 and let @xmath110 note that the jacobian of the coordinate change from @xmath111 to @xmath112 is @xmath113 changing variables in ( common complex1 ) and integrating @xmath114 out , we obtain@xmath115where @xmath116 is a @xmath104 diagonal matrix . consider the likelihood ratios @xmath117 and @xmath118
. formulae ( [ common complex1 ] ) and ( [ common complex2 ] ) imply the following proposition .
[ proposition1]_let _ @xmath105 _ be the set of all _ @xmath43 _ _ orthogonal matrices .
denote by _ _
@xmath119 _ _ the invariant measure on the orthogonal group _ _ @xmath105 _ normalized to make the total measure unity . put _ @xmath120 @xmath121 and let @xmath122 be the @xmath43 diagonal matrix @xmath123 where @xmath124 . _ _
then,__@xmath125 in the special case where @xmath126 the rank of the matrix @xmath122 equals one , and the integrals over the orthogonal group in ( [ lr1 ] ) and ( [ lr2 ] ) can be rewritten as integrals over a @xmath7-dimensional sphere .
omh show how such spherical integrals can be represented in the form of contour integrals , and apply laplace approximation to these contour integrals to establish the asymptotic properties of @xmath127 and @xmath128 in the @xmath47 case , the integrals in ( [ lr1 ] ) and ( [ lr2 ] ) can be rewritten as integrals over a stiefel manifold , the set of all orthonormal @xmath45__-frames _ _ in @xmath129 .
onatski ( 2012 ) obtains a generalization of the contour integral representation from spherical integrals to integrals over stiefel manifolds .
unfortunately , the laplace approximation method does not straightforwardly extend to that generalization , and we therefore propose an alternative method of analysis . the second - order asymptotic behavior , as @xmath7 goes to infinity , of integrals of the form @xmath130 was analyzed in guionnet and maida ( 2005 ) ( theorem 3 ) for the particular case where @xmath131 is a fixed matrix of rank one , @xmath132 a deterministic matrix , and under the condition that the empirical distribution of @xmath132 s eigenvalues converges to a distribution function with bounded support .
below , we extend guionnet and maida s approach to cases where @xmath133 has rank larger than one , and to the stochastic setting of this paper .
we then use such an extension to derive the asymptotic properties of @xmath134 and @xmath135 .
let @xmath136 be the empirical distribution of @xmath137 , and denote by @xmath138 the marchenko - pastur distribution function , with density@xmath139where @xmath140 and @xmath141 , and a mass of @xmath142 at zero .
as is well known , the difference between @xmath136 and @xmath143 weakly converges to zero a.s . as @xmath144 .
moreover , @xmath145 and @xmath146 if @xmath147 , and @xmath148 if @xmath149 consider the hilbert transform of @xmath150 @xmath151 that transform is well defined for real @xmath114 outside the support of @xmath138 , that is , on the set @xmath152 using ( [ mp density ] ) , we get@xmath153where the sign of the square root is chosen to be the sign of @xmath154 .
it is not hard to see that @xmath155 is strictly decreasing on @xmath156 .
thus , on @xmath157 , we can define an inverse function @xmath158 , with values@xmath159the so - called @xmath160-transform @xmath161 of @xmath138 takes the form @xmath162 for @xmath163 and @xmath164 sufficiently small , consider the subset of @xmath165 @xmath166 & \text{for } c\geq 1 , \\
\left [ -\frac{1}{\sqrt{c}\left ( 1-\sqrt{c}\right ) } + \varepsilon , 0\right ) \cup \left ( 0,\frac{1}{\sqrt{c}\left ( 1+\sqrt{c}\right ) } -\varepsilon \right ] & \text{for } c<1.% \end{array}% \right.\]]from ( [ stijeltjes analytic ] ) , @xmath167 when @xmath168 , @xmath169 when @xmath170 , and @xmath171 when @xmath172 .
therefore , @xmath173 with probability approaching one as @xmath174 [ proposition2]let @xmath175 be a sequence of random @xmath43 diagonal matrices @xmath176 where @xmath177 @xmath178 further , let @xmath179 , where @xmath180 is the @xmath160-transform of the marchenko - pastur distribution @xmath138 .
assume that , for some @xmath163 and @xmath181 @xmath182 with probability approaching one as @xmath183 .
then,@xmath184 } \\ & & \times
\prod_{j=1}^{r}\prod_{s=1}^{j}\sqrt{1 - 4\left ( \theta _ { pj}v_{pj}\right ) \left ( \theta _ { ps}v_{ps}\right ) c_{p}}\left ( 1+o(1)\right ) \text { a.s.},\end{aligned}\]]where @xmath185 is uniform over all sequences @xmath175 satisfying the assumption .
this proposition extends theorem 3 of guionnet and maida ( 2005 ) to cases when @xmath186 @xmath187 depends on @xmath39 and @xmath188 is random .
when @xmath126 @xmath189 and @xmath190 are fixed , it is straightforward to verify that @xmath191 where @xmath192 in guionnet and maida s ( 2005 ) theorem 3 , the expression @xmath193 should have been used instead of @xmath194 which is a typo .
setting @xmath46 and @xmath195 in proposition [ proposition2 ] and using formula ( [ lr1 ] ) from proposition [ proposition1 ] gives us an expression for @xmath196 which is an equivalent of formula ( 4.1 ) in theorem 7 of omh .
theorem [ theorem1 ] below uses proposition [ proposition2 ] to generalize theorem 7 of omh to the multispiked case @xmath47 .
let @xmath197 and@xmath198 & \text{for } c>1 , \\
\left [ -\sqrt{c}+\delta , 0\right ) \cup \left ( 0,\sqrt{c}-\delta \right ] & \text{for } c\leq 1.% \end{array}% \right .
\label{h domain}\]]the condition @xmath199 for some @xmath200 implies that @xmath201 for some @xmath202 @xmath203 and @xmath7 sufficiently large .
below , we are only interested in non - negative values of @xmath85 , and assume that @xmath204 $ ] under the alternative hypothesis .
the corresponding @xmath205 , thus , is positive . with the above setting for @xmath206
we have @xmath207 and @xmath208 say , as in theorem 7 in omh .
[ theorem1]__suppose that the null hypothesis is true ( _ _ @xmath210)_. let _
@xmath211 _ _ be any fixed number such that _ _ @xmath212 , _ and let _ @xmath213 ^{r}$ ] _ _ be the space of real - valued continuous functions on _ _ @xmath214 ^{r}$ ] _ _ equipped with the supremum norm__. _ then , _ _ as _ @xmath215 _ _ _ _ @xmath216__almost surely , _ _ where the @xmath217 terms are uniform in @xmath218 ^{r}$ ] . _
furthermore , _
@xmath219 _ and _ @xmath220 _ _ viewed as random elements of _ _ @xmath221 ^{r}$ ] , _ converge weakly to _ @xmath222 _ and _ @xmath223 _ with gaussian finite - dimensional distributions such that _ @xmath224 _ _ , _ _ @xmath225 and , _ for any _
@xmath226 ^{r},$]@xmath227 theorem [ theorem1 ] and le cam s first lemma ( van der vaart ( 1998 ) , p.88 ) imply that the joint distributions of @xmath60 ( as well as those of @xmath228 ) under the null and under the alternative are mutually contiguous for any @xmath229 . by applying le cam s third lemma ( van der vaart ( 1998 ) , p.90 )
, we can study the local powers of tests detecting signals in noise .
the requirement that @xmath85 be positive under alternatives corresponds to situations where the signals contained in the data are independent from the noise .
if dependence between the signals and the noise is allowed , one might consider two - sided alternatives of the form @xmath230 for some @xmath28 .
values of @xmath85 between @xmath231 and @xmath232 correspond to alternatives under which the noise variance is reduced along certain directions . in view of proposition [ proposition2 ]
, it should not be difficult to generalize theorem [ theorem1 ] to the case of fully ( @xmath233 all @xmath28 s ) or partially ( @xmath234 some @xmath28 s ) two - sided alternatives .
this problem will not be discussed here , and is left for future research .
denote by @xmath235 and @xmath236 respectively , the asymptotic powers of the asymptotically most powerful @xmath35- and @xmath53-based tests of size @xmath237 of the null @xmath238 against a point alternative @xmath239 with @xmath240 @xmath12 . as functions of @xmath241 @xmath242 and @xmath243
are called the _ asymptotic power
envelopes_. [ proposition3]_let _ @xmath68 _ _ denote the standard normal distribution function .
then,__@xmath244 \text { \textit{and } } \label{local power } \\ \beta _ { \mu } \left ( h\right ) \ ! & = & \!1\!-\!\phi \left [ \!\phi ^{-1}\left ( 1\!-\!\alpha \right ) \!-\!\sqrt{-\frac{1}{2}\sum_{i , j=1}^{r}\left ( \ln \left ( 1\!-\!\frac{h_{i}h_{j}}{c}\right ) \!+\!\frac{h_{i}h_{j}}{c}\right ) } \,% \right ] . \label{local power mu}\end{aligned}\ ] ] figure [ envelope_may2012 ] shows the asymptotic power envelopes @xmath245 and @xmath246 as functions of @xmath247 and @xmath248 when @xmath249 is two - dimensional .
( upper panel ) and @xmath250 ( lower panel ) for @xmath251 as functions of @xmath252.,width=384 ] it is important to realize that the asymptotic power envelopes derived in proposition [ proposition3 ] are valid not only for @xmath35- and @xmath53-based tests but also for any test invariant under left orthogonal transformations of the observations ( @xmath253 where @xmath254 is a @xmath255 orthogonal matrix ) , and for any test invariant under multiplication by any non - zero constant and left orthogonal transformations of the observations ( @xmath256 where @xmath257 and @xmath254 is a @xmath43 orthogonal matrix ) , respectively .
let @xmath258 and @xmath259 denote the frobenius norm and the spectral norm , respectively , of a matrix @xmath44 .
let @xmath37 be the null hypothesis @xmath260 and let @xmath36 be any of the following alternatives : @xmath17 @xmath261 for some @xmath19 or @xmath262 or @xmath263 or @xmath264 , where @xmath265 is a positive constant that may depend on @xmath0 and @xmath7 .
[ proposition4 ] for specified @xmath88 , consider tests of @xmath37 against @xmath36 that are invariant with respect to the left orthogonal transformations of the data @xmath266 .$ ] for any such test , there exists a test based on @xmath35 with the same power function .
similarly , for unspecified @xmath54 consider tests that , in addition , are invariant with respect to multiplication of the data @xmath71 by non - zero constants .
for any such test , there exists a test based on @xmath53 with the same power function .
examples of the former tests include the tests of @xmath267 studied in chen et al ( 2010 ) and cai and ma ( 2012 ) .
an example of the latter test is the test of sphericity studied in chen et al ( 2010 ) .
the tests studied in chen et al ( 2010 ) and cai and ma ( 2012 ) are invariant , although they are not @xmath35- or @xmath53-based . for @xmath126
omh show that the asymptotic power envelopes are closely approached by the asymptotic powers of the @xmath35- and @xmath53-based likelihood ratio tests .
our goal here is to explore the asymptotic power of those likelihood ratio tests for @xmath47 .
unfortunately , as @xmath45 grows , it becomes increasingly difficult to compute the asymptotic critical values for the likelihood ratio tests by simulation .
for example , @xmath70 requires simulating a 2-dimensional gaussian random field with the covariance function and the mean function described in theorem [ theorem1 ] . for @xmath268 figure
[ powerlam_section ] shows sections of the power envelope ( dotted lines ) and the power of the likelihood ratio test based on @xmath38 for various fixed values of @xmath247 under the alternative .
figure [ powermu_section ] shows the same plots for the tests based on @xmath269 to enhance readability , we use a different parametrization : @xmath270 @xmath84 .
as @xmath271 varies in the region of contiguity @xmath272 @xmath273 spans the entire half - line @xmath274 note that the asymptotic mean and autocovariance functions of the log likelihood ratios derived in theorem [ theorem1 ] depend on @xmath85 only through @xmath275 therefore , under the new parametrization , they depend only on @xmath276 .
the parameter @xmath277 plays the classical role of a local parameter in our setting .
-based lr test ( solid lines ) relative to the asymptotic power envelope ( dotted lines ) for different values of @xmath278 under the alternative ; @xmath279 , width=384 ] -based lr test ( solid lines ) relative to the asymptotic power envelope ( dotted lines ) for different values of @xmath280 under the alternative ; @xmath279 , width=384 ] figure [ isoplot ] further explores the relationship between the asymptotic powers of the @xmath35- and @xmath53-based lr test and the corresponding asymptotic power envelopes when @xmath70 .
we pick all values of @xmath281 satisfying inequality @xmath282 and such that the asymptotic power envelope for @xmath35-based tests is exactly 25 , 50 , 75 , and 90% .
then , we compute and plot the corresponding power of the @xmath35-based lr test ( solid lines ) against @xmath283 the dashed lines show similar graphs for the @xmath53-based lr test .
the value @xmath284 corresponds to single - spiked alternatives @xmath285 @xmath286 the value @xmath287 corresponds to equi - spiked alternatives @xmath288 the intermediate values of @xmath289 link the two extreme cases .
we do not consider values @xmath290 , as the power function is symmetric about the 45-degree line in the @xmath291 space .
-based ( solid lines ) and @xmath292-based ( dashed lines ) lr tests plotted against @xmath293 where @xmath294 are such that the respective asymptotic power envelopes @xmath295 and @xmath296 equal 25 , 50 , 75 and 90%.,width=384 ] somewhat surprisingly , the power of the lr test along the set of alternatives @xmath294 corresponding to the same values of the asymptotic power envelope is not a monotone function of @xmath283 equi - spiked alternatives typically seem to be particularly difficult to detect by the lr tests .
however , for the set of alternatives corresponding to an asymptotic power envelope value of 90% , the single - spiked alternatives are even harder to detect .
a natural question is : how does the asymptotic power of the @xmath35- and @xmath297-based lr tests depend on the choice of @xmath45 , that is , how do those tests perform when the actual @xmath45 does not coincide with the value the test statistic is based on ?
for example , to detect a single signal , one can , in principle , use lr tests of the null hypothesis against alternatives with @xmath298 etc .
how does the asymptotic powers of such tests compare ?
figure [ wrongr ] reports the asymptotic powers of the @xmath35- and @xmath53-based lr tests designed to detect alternatives with @xmath46 ( solid line ) and @xmath299 ( dashed line ) , under single - spiked alternatives . as in figures 2 and 3
, we use the parametrization @xmath300 for the single - spiked alternative .
it appears that the two asymptotic powers are very close to each other ; interestingly , neither of them dominates the other .
using lr tests designed against alternatives with @xmath47 seems to be beneficial for detecting a single - spiked alternative with relatively small @xmath301 ( and @xmath58 ) . -based ( left panel ) and @xmath292-based ( right panel ) lr tests .
solid line : power when @xmath46 is correctly assumed .
dashed line : power when @xmath70 is incorrectly assumed.,width=384 ] in the remaining part of this section , we consider examples of some of the tests that have been proposed previously in the literature , and , in proposition [ proposition5 ] , derive their asymptotic power functions .
john ( 1971 ) proposes testing the sphericity hypothesis @xmath302 against general alternatives based on the test statistic @xmath303 , \label{john 's stat}\]]where @xmath304 is the sample covariance matrix .
he shows that , when @xmath305 such a test is locally most powerful invariant .
ledoit and wolf ( 2002 ) study john s test when @xmath23 .
they prove that , under the null , @xmath306 hence , the test with asymptotic size @xmath237 rejects the null of sphericity whenever @xmath307 .
ledoit and wolf ( 2002 ) propose @xmath308 -% \frac{p}{n}\left [ \frac{1}{p}tr\hat{\sigma}\right ] ^{2}+\frac{p}{n } \label{lw stat}\]]as a test statistic for testing the hypothesis that the population covariance matrix is the unit matrix .
they show that , under the null , @xmath309 as in the previous example , the null is rejected at asymptotic size @xmath237 whenever @xmath310 when @xmath311 bai et al .
( 2009 ) propose to use a corrected version@xmath312of the likelihood ratio statistic to test the equality of the population covariance matrix to the identity matrix against general alternatives . under the null , @xmath313 ( still , as @xmath314 ) .
the null hypothesis is rejected at asymptotic level @xmath237 whenever @xmath315 is larger than @xmath316 .
let @xmath317 be any function of the @xmath45 largest eigenvalues increasing in all its arguments .
the asymptotic distribution of @xmath318 under the null is determined by the functional form of @xmath319 and the fact that@xmath320where tw@xmath321 denotes the @xmath45-dimensional tracy - widom law of the first kind , @xmath322 and @xmath323 . call tracy - widom - type tests all tests that reject the null whenever @xmath324 is larger than the corresponding asymptotic critical value obtained from ( [ tw ] ) .
cai and ma ( 2012 ) propose to use a u - statistic@xmath325where @xmath326 to test the hypothesis that the population covariance matrix is the unit matrix . under the null , as @xmath23 , @xmath327 the null hypothesis is rejected at asymptotic level @xmath237 whenever @xmath328 is larger than @xmath329 .
cai and ma ( 2012 ) show that this test is rate - optimal against general alternatives from a minimax point of view .
consider the tests described in examples 1 , 2 , 3 , 4 and 5 , and denote by @xmath330 @xmath331 @xmath332 @xmath333 and @xmath334 their respective asymptotic powers at asymptotic level @xmath335 [ proposition5]__the asymptotic power functions of the tests described in examples 1 - 5 are__@xmath336for any @xmath239 such that @xmath337 for @xmath338 formula ( [ beta_j_lw ] ) for @xmath333 directly follows from proposition 2 of cai and ma ( 2012 ) .
the proof of the other formulae follows along the same lines as in the proof of proposition 10 in omh , and is omitted . except for the tracy - widom tests of example 4 ,
all those asymptotic power functions are non - trivial .
figures [ johnmu_section ] and [ johnlam_section ] compare these power functions to the corresponding power envelopes for @xmath70 .
since john s test is invariant with respect to orthogonal transformations and scalings of the data , figure johnmu_section compares @xmath339 ( solid line ) to the power envelope @xmath246 ( dotted line ) .
the ledoit - wolf test , the correctedlikelihood ratio test , and the cai - ma test are invariant with respect to orthogonal transformations of the data only , and figure [ johnlam_section ] thus compares the asymptotic power functions @xmath340 and @xmath341 ( solid and dashed lines , respectively ) to the power envelope @xmath342 ( dotted line ) .
note that @xmath341 depends on @xmath343 . as @xmath343 converges to one ,
@xmath341 converges to @xmath65 which corresponds to the case of trivial power . as @xmath343 converges to zero
, @xmath341 converges to @xmath344 . in figure
[ johnlam_section ] , we provide plots of @xmath341 that correspond to @xmath345 these comparisons show that , contrary to our lr tests ( see figures powerlam_section and [ powermu_section ] ) , _ all _ those tests either have trivial power @xmath237 ( the tracy - widom ones ) , or power functions that increase very slowly with @xmath346 and @xmath347 and lie very far below the corresponding power envelope .
( dotted lines ) for different values of @xmath278 under the alternative ; @xmath279 , width=384 ] ) relative to the asymptotic power envelope @xmath348 ( dotted lines ) for different values of @xmath349 under the alternative ; @xmath279 , width=384 ]
this paper extends onatski , moreira and hallin s ( 2011 ) ( omh ) study of the power of high - dimensional sphericity tests to the case of multi - spiked alternatives .
we derive the asymptotic distribution of the log - likelihood ratio process and use it to obtain simple analytical expressions for the maximal asymptotic power envelope and for the asymptotic powers of several tests proposed in the literature .
these asymptotic powers turn out to be very substantially below the envelope .
we propose the likelihood ratio test based on the data reduced to the eigenvalues of the sample covariance matrix .
our computations show that the asymptotic power of this test is close to the envelope .
all convergence statements made below refer to the situation when @xmath183 .
we start with two auxiliary results .
* proof : * let us denote the cumulative distribution function corresponding to a measure @xmath53 as @xmath357 further , let us denote @xmath358 \right\ } $ ] as @xmath359 consider the following three distances between measures @xmath53 and @xmath360 the kolmogorov distance @xmath361 the wasserstein distance @xmath362 and the kantorovich distance @xmath363 as is well known ( see , for example , exercise 1 on p.425 of dudley ( 2002 ) ) , @xmath364 therefore , we have@xmath365@xmath366as follows from theorem 1.1 of gtze and tikhomirov ( 2011 ) , there exists a constant @xmath355 such that @xmath367 for all @xmath163 .
thus , @xmath368 a.s .. since @xmath369 is @xmath370 and @xmath371 a.s . , the result follows.@xmath372 let us denote the integral @xmath376 as @xmath377 . as explained in guionnet and maida ( 2005 , p.454 ) , we can write@xmath365@xmath378where @xmath379 denotes the expectation conditional on @xmath380 and the @xmath7-dimensional vectors @xmath381 are obtained from standard gaussian @xmath7-dimensional vectors @xmath382 , independent from @xmath383 , by a schmidt orthogonalization procedure .
more precisely , we have @xmath384 , where @xmath385 and@xmath386@xmath387 in the spirit of the proof of guionnet and maida s ( 2005 ) theorem 3 , define@xmath386@xmath388where @xmath389 stands for the classical kronecker symbol . as will be shown below , after an appropriate change of measure , @xmath390 and @xmath391 are asymptotically centered gaussian .
expressing the exponent in ( [ first ip ] ) as a function of @xmath390 and @xmath392 changing the measure of integration , and using the asymptotic gaussianity will establish the proposition .
let @xmath393 , where @xmath394 . using this notation , ( [ first ip ] ) , ( [ a definition ] ) , and ( gamma definition ) , we get , after some algebra,@xmath365@xmath395where @xmath396 is the standard gaussian probability measure , and@xmath397@xmath398where @xmath399 is a @xmath400 matrix with @xmath401-th element @xmath402@xmath403 next , define the event@xmath365@xmath404where @xmath405 and @xmath406 are positive parameters to be specified later . somewhat abusing notation
, we will also refer to @xmath407 as a rectangular region in @xmath408 that consists of vectors with odd coordinates in @xmath409 and even coordinates in @xmath410 .
let@xmath365@xmath411where @xmath412 denotes the indicator function .
below , we establish the asymptotic behavior of @xmath413 as first @xmath39 and then @xmath405 and @xmath414 , diverge to infinity .
we then show that the asymptotics of @xmath415 and @xmath416 coincide .
consider infinite arrays @xmath417 @xmath19 of random centered gaussian measures@xmath418@xmath419since @xmath420 and @xmath421 , there exists @xmath422 such that , for sufficiently large @xmath423@xmath424recall that @xmath425 and @xmath426 a.s .. therefore , still a.s .
, for sufficiently large @xmath427 @xmath428 when @xmath429 and @xmath430 when @xmath431 hence , the measures @xmath432 are a.s . well defined for sufficiently large @xmath7 .
whenever @xmath432 is not well defined , we re - define it arbitrarily .
we have@xmath365@xmath433 } j_{p}^{m , m^{\prime } } , \vspace{-0.2 in } \label{immnow}\]]where@xmath365@xmath434we now show that , under @xmath435 , @xmath436 a.s .
converges in distribution to a centered @xmath437-dimensional gaussian vector , so that @xmath438 is asymptotically equivalent to an integral with respect to a gaussian measure on @xmath439 first , let us find the mean @xmath440 , and the variance @xmath441 of @xmath436 under measure @xmath442 .
note that @xmath443 and @xmath444 .
with probability one , for sufficiently large @xmath39 we have@xmath365@xmath445which , by corollary 1 , is @xmath217 uniformly in @xmath446 , a.s .. that corollary 1 can be applied here follows from the form of expression ( [ k transform ] ) for @xmath447 .
similarly,@xmath365@xmath448uniformly in @xmath449 , a.s .. thus,@xmath365@xmath450 next , with probability one , for sufficiently large @xmath7 we have@xmath451@xmath452let @xmath453 and @xmath454 . then , using corollary 1 , we get@xmath386@xmath455uniformly in @xmath449 .
similarly , we have@xmath365@xmath456and@xmath365@xmath457uniformly in @xmath449 , a.s .. a straightforward calculation , using formula ( [ k transform ] ) , shows that@xmath386@xmath458uniformly in @xmath449 , a.s . , where the matrix @xmath459 has elements@xmath365@xmath460
\text{.}\vspace{% -0.25 in } \label{v22}\end{aligned}\]]this implies that@xmath365@xmath461which is bounded away from zero and infinity for sufficiently large @xmath7 , uniformly over @xmath462 , a.s .. by construction , @xmath436 is a sum of @xmath7 independent random vectors having uniformly bounded third and fourth absolute moments under measure @xmath463 therefore , a central limit theorem applies .
moreover , since the function @xmath464 is lipshitz over @xmath465 uniformly in @xmath466 theorem 13.3 of bhattacharya and rao ( 1976 ) , which describes the accuracy of the gaussian approximations to integrals of the form ( [ jmm_first ] ) in terms of the oscillation measures of the integrand , implies that@xmath365@xmath467where @xmath468 denotes the gaussian distribution function with mean @xmath469 and variance @xmath470 and @xmath471 converges to zero uniformly in @xmath472 as @xmath473 a.s .. the rate of such a convergence may depend on the values of @xmath405 and @xmath474 note that , in @xmath465 as @xmath475 the difference @xmath476 converges to zero uniformly over @xmath477 where@xmath418@xmath478such a convergence , together with ( [ egammatil ] ) , ( [ vargammatil ] ) , and ( [ jmm_second ] ) implies that@xmath365@xmath479where @xmath480 .
note that the difference @xmath481 converges to zero as @xmath482 , uniformly in @xmath7 for @xmath7 sufficiently large . on the other hand,@xmath365@xmath483 } { 2\pi \!\sqrt{\det \left ( v_{p}^{(j , s)}\right ) } } \mathrm{d}y,% \vspace{-0.2 in } \label{mmintegral}\]]where@xmath365@xmath484using ( [ v11]-[v22 ] ) , we verify that , for sufficiently large @xmath39 @xmath485 is a.s .
positive definite , and@xmath365@xmath486therefore,@xmath365@xmath487and , uniformly in @xmath7 for @xmath7 sufficiently large@xmath488@xmath489equations ( [ immnow ] ) , ( [ jmm_third ] ) , and ( [ mmintegralfinal ] ) describe the behavior of @xmath490 for large @xmath39 @xmath491 and @xmath406 .
let us now turn to the analysis of @xmath492 let @xmath493 be the event @xmath494 , and let@xmath365@xmath495as explained in guionnet and maida ( 2005 , p.455 ) , @xmath496 @xmath497 are independent of @xmath498 @xmath338 therefore,@xmath386@xmath499 denoting again by @xmath396 the centered standard gaussian measure on @xmath500 , we have@xmath365@xmath501for @xmath502 and @xmath503@xmath504therefore , using chebyshev s inequality , for @xmath502 and @xmath505@xmath506setting @xmath507 ( here we assume that @xmath508 ) , we get@xmath397@xmath509similarly , we show that the same inequality holds when @xmath390 is replaced by @xmath510 and thus@xmath365@xmath511for @xmath512 the same line of arguments yields@xmath365@xmath513 inequalities ( [ subgaussgms ] ) and ( [ subgaussgmm ] ) imply that @xmath514 and therefore , for sufficiently large @xmath515@xmath516note that@xmath365@xmath517 } \left ( j_{p}^{m , m^{\prime } } + j_{p}^{m , m^{\prime } , \infty } \right ) , \vspace{-0.2 in } \label{ipm}\]]where@xmath365@xmath518we will now derive an upper bound for @xmath519 from the definition of @xmath520 we see that there exist positive constants @xmath521 and @xmath522 which may depend on @xmath523 and @xmath524 , such that for any @xmath187 satisfying @xmath525 @xmath526 and for sufficiently large @xmath39 when @xmath493 holds,@xmath365@xmath527let @xmath528 . clearly , @xmath529 .
therefore,@xmath397@xmath530 first assume @xmath531 denote @xmath532 as @xmath533 and @xmath534 as @xmath535 note that , under @xmath536 @xmath537 is a standard normal random variable .
further , as long as @xmath421 for @xmath538 @xmath533 considered as a function of @xmath539 is continuous on @xmath540 for sufficiently large @xmath39 a.s .. hence , the empirical distribution of @xmath541 converges .
moreover , @xmath542 and @xmath543 a.s . converge to finite real numbers .
now , for @xmath544 such that @xmath545 we have@xmath365@xmath546for sufficiently large @xmath39 a.s .. using this inequality , we get , for sufficiently large @xmath7 and any positive @xmath547 such that @xmath548@xmath549setting @xmath550 ( here we assume that @xmath405 and @xmath406 are such that @xmath547 satisfies the above requirements ) , we get@xmath365@xmath551replacing @xmath552 by @xmath553 in the above derivations and combining the result with the above inequality , we get@xmath386@xmath554when @xmath555 following a similar line of arguments , we obtain@xmath365@xmath556and thus , for sufficiently large @xmath557@xmath558 finally , combining ( [ imversusi ] ) , ( [ ipm ] ) , and ( [ jkk ] ) , we obtain for@xmath365@xmath559 } \vspace{-0.2 in } \label{jp}\ ] ] the following upper and lower bounds:@xmath560let @xmath355 be an arbitrarily small number . equations ( [ jmm_third ] ) and ( [ mmintegralfinal ] ) imply that there exist @xmath561 and @xmath562 such that , for any @xmath563 and @xmath564@xmath565for all sufficiently large @xmath566 let us choose @xmath563 and @xmath567 so that@xmath365@xmath568and@xmath365@xmath569 \sup_{\left\ { 2\theta _ { pj}\in \omega _ { \varepsilon \eta } , j\leq r\right\ } } \prod_{j=1}^{r}\prod_{s=1}^{j}\sqrt{1\!-\!4\left ( \!\theta _ { pj}v_{pj}\!\right ) \left ( \!\theta _ { ps}v_{ps}\!\right ) c_{p}}<\frac{\tau } { 4}\vspace{-0.2in}\]]for all sufficiently large @xmath39 a.s .. then , ( [ lastineq ] ) implies that@xmath386@xmath570for all sufficiently large @xmath7 , a.s .. since @xmath544 can be chosen arbitrarily , we have , from ( [ jp ] ) and ( [ jpbound]),@xmath365@xmath571 } \\ & & \times \left ( \prod_{j=1}^{r}\prod_{s=1}^{j}\sqrt{1\!-\!4\left ( \!\theta _ { pj}v_{pj}\!\right ) \left ( \!\theta _ { ps}v_{ps}\!\right ) c_{p}}+o(1)\right ) , \vspace{-0.2in}\end{aligned}\]]where @xmath572 as @xmath573 uniformly in @xmath574 a.s .. @xmath575 setting @xmath576 , we have @xmath577 , @xmath578 , and@xmath579 further , by lemma 11 and formula ( 3.3 ) of omh , @xmath580 for sufficiently large @xmath39 a.s .. with these auxiliary results , formula ( [ equivalence 1 ] ) is a straightforward consequence of ( [ lr1 ] ) and proposition 2 . in what follows , we omit the subscript @xmath7 in @xmath101 to simplify notation .
note that @xmath582 is the integral appearing in expression ( [ lr2 ] ) for @xmath583 .
let us now prove that , for some constant @xmath584@xmath585where @xmath185 is uniform in @xmath586 ^{r}.$ ] since , by corollary 1 , @xmath587 a.s . , the set @xmath588 is bounded from below , and @xmath589 a.s . , there exists a constant @xmath590 that depends only on @xmath211 and @xmath591 such that @xmath592 ^{r}}px\limfunc{tr% } \left ( d_{p}q^{\prime } \lambda _ { p}q\right ) /s_{p}\geq
-a_{1}x/2 $ ] for all @xmath593 and all sufficiently large @xmath39 a.s .. therefore , for all @xmath594 ^{r},\vspace{-0.2in}$]@xmath595and , using stirling s approximation , we get@xmath365@xmath596 next , there exists a constant @xmath597 such that , for all @xmath598 and all sufficiently large @xmath39 @xmath599 ^{r}}px% \limfunc{tr}\left ( d_{p}q^{\prime } \lambda _ { p}q\right ) /s_{p}\leq a_{2}x/2 $ ] , a.s .. therefore , a.s .
, for all sufficiently large @xmath7,@xmath365@xmath600where @xmath601 is the complementary incomplete gamma function ( see olver 1997 , p.45 ) with @xmath602 hence , for sufficiently large @xmath39 @xmath603 and we can continue@xmath365@xmath604now , @xmath605 whenever @xmath606 and @xmath607 ( olver 1997 , p.70 ) . therefore , we have , for sufficiently large @xmath7,@xmath365@xmath608comparing this to ( [ i1bound ] ) , we see that @xmath237 can be chosen so that@xmath365@xmath609 further , for sufficiently large @xmath7,@xmath365@xmath610where @xmath611 . therefore ,
for any positive @xmath612 and sufficiently large @xmath557@xmath613setting @xmath614 and using stirling s approximation , we have , a.s.,@xmath365@xmath615so that@xmath365@xmath616comparing this to ( [ i1bound ] ) , we see that @xmath237 can be chosen so [email protected] .. combining ( [ i2bound ] ) and ( [ i3bound ] ) , we get ( [ ip ] ) . now , letting @xmath618 note that there exist @xmath619 and @xmath164 such that @xmath620 \text { and } x\in \left [ p\!-\!\alpha \sqrt{p}% , p\!+\!\alpha \sqrt{p}\right ] \right\ } \subseteq \theta _ { \varepsilon \eta } $ ] for all sufficiently large @xmath39 a.s .. hence , by ( [ ip ] ) , and proposition 2 , a.s.,@xmath365@xmath621 } \label{i0inf } \\ & & \times \left ( \prod_{j=1}^{r}\prod_{s=1}^{j}\sqrt{1\!-\!4\left ( \!\tilde{% \theta}_{pj}\tilde{v}_{pj}\!\right ) \left ( \!\tilde{\theta}_{ps}\tilde{v}% _ { ps}\!\right ) c_{p}}+o(1)\right ) \!\mathrm{d}x\text{,}\vspace{-0.2 in } \notag\end{aligned}\]]where @xmath185 is uniform in @xmath586 ^{r}$ ] and @xmath622 .$ ] expanding @xmath623 and @xmath624 into power series of @xmath625 we get@xmath365@xmath626where @xmath627 and @xmath628 are @xmath370 uniformly in @xmath629 ^{r}.$ ] further , consider the integral@xmath365@xmath630splitting the domain of integration into segments @xmath631 , \left [ p\!-\!\alpha p^{\gamma } , p\!+\!\alpha p^{\gamma } \right ] $ ] and @xmath632 , $ ] where @xmath633 and denoting the corresponding integrals by @xmath634 and @xmath635 respectively , we have@xmath365@xmath636using the laplace approximation , we have@xmath365@xmath637so that @xmath638 dominates @xmath639 and @xmath640 and@xmath365@xmath641this implies that@xmath365@xmath642and hence , only constant and linear terms in the expansion of @xmath643 into power series of @xmath644 matter for the evaluation of @xmath645 let us find these terms . by corollary 1 , @xmath646 a.s .. using this fact , after some algebra , we get@xmath365@xmath647@xmath648and@xmath365@xmath649it follows that@xmath365@xmath650 } \label{last eq } \\ & & \times e^{\sum_{j=1}^{r}\theta _ { pj}v_{pj}\left ( x - s_{p}\right ) } \left ( \prod_{j=1}^{r}\prod_{s=1}^{j}\sqrt{1\!-\!4\left ( \!\theta _ { pj}v_{pj}\!\right ) \left ( \!\theta _ { ps}v_{ps}\!\right ) c_{p}}+o(1)\right ) \!\mathrm{d}x \notag \\ & = & \left ( 1+o(1)\right ) \prod_{j=1}^{r}\left ( 1+h_{j}\right ) ^{\frac{n_{p}}{2% } } l_{p}\!\left ( h;\lambda _ { p}\right ) \int_{p\!-\!\alpha \sqrt{p}% } ^{p\!+\!\alpha \sqrt{p}}\!\!\!x^{\frac{np}{2}-1}e^{-\frac{n}{2}% x}e^{\sum_{j=1}^{r}\theta _ { pj}v_{pj}\left ( x - s_{p}\right ) } \mathrm{d}x,% \vspace{-0.2 in } \notag\end{aligned}\]]where the last equality in ( [ last eq ] ) follows from ( [ lr1 ] ) and proposition 2 . the last equality in ( [ last eq ] ) , ( [ lr2 ] ) and the fact that@xmath397@xmath651imply that@xmath365@xmath652which establishes ( [ equivalence 2 ] ) . the rest of the statements of theorem 1 follow from ( [ equivalence 1 ] ) , ( [ equivalence 2 ] ) , and lemmas 12 and a2 of omh.@xmath372 to save space , we only derive the asymptotic power envelope for the relatively more difficult case of real - valued data and @xmath53-based tests . according to the neyman - pearson lemma , the most powerful test of @xmath210 against the simple alternative @xmath653 is the test which rejects the null when @xmath654 is larger than a critical value @xmath655 it follows from theorem 1 that , for such a test to have asymptotic size @xmath237 , @xmath656 must be@xmath365@xmath657where@xmath365@xmath658now , according to le cam s third lemma and theorem 1 , under @xmath659 @xmath660 the asymptotic power ( [ local power mu ] ) follows.@xmath372 before turning to the proof of proposition [ proposition4 ] , let us clarify the invariance issues in the problem under study . for basic definitions ( invariant , maximal invariant , etc . )
, we refer to chapter 6 of lehmann and romano ( 2005 ) .
suppose that @xmath71 is a @xmath661 random matrix with @xmath662 .
this model is clearly invariant under the group @xmath663 acting on @xmath664 of left - hand multiplications by a @xmath43 orthogonal matrix @xmath665 @xmath666 @xmath667 so are the null hypothesis @xmath37 and the alternative @xmath668 letting @xmath669 the @xmath670-tuple @xmath671 of non - zero eigenvalues of @xmath672 is clearly invariant under that group , since @xmath673 and @xmath674 share the same eigenvalues @xmath675 for any orthogonal matrix @xmath254 and any matrix @xmath676 however , @xmath677 is not maximal invariant for @xmath663 as @xmath678 and @xmath679 where @xmath680 is an arbitrary @xmath681 orthogonal matrix , share the same @xmath682 although , in general , there is no @xmath43 orthogonal matrix @xmath254 such that @xmath683 now , the joint density of the elements of @xmath71 is@xmath684by the factorization theorem , @xmath685 is a sufficient statistic , and it is legitimate to restrict attention to @xmath685-measurable inference procedures .
left - hand orthogonal multiplications @xmath686 of @xmath114 yields , for @xmath687 a transformation of the form @xmath688 when @xmath254 range over the family @xmath689 of @xmath255 orthogonal matrices , those transformations also form a group , @xmath690 say , now acting on the space of @xmath43 symmetric positive semidefinite real matrices of rank @xmath670 .
clearly , @xmath691 is maximal invariant for @xmath692 as @xmath693 and @xmath694 share the same eigenvalues if and only if @xmath695 for some @xmath43 orthogonal matrix @xmath696 a similar reasoning applies in the case of unspecified @xmath54 with a larger group combining multiplication by an arbitrary non - zero constant with the @xmath43 left orthogonal transformations . sufficiency and invariance then lead to restricting attention to @xmath53-measurable tests . with the same notation as above ,
write @xmath697 for the sufficient statistic .
consider an arbitrary invariant ( under the group @xmath698 of left orthogonal transformations of @xmath699 ) test @xmath700 , and define @xmath701 . then @xmath702 is a @xmath703-measurable test with the same size and power function as @xmath704 it follows from the proof of theorem 6.5.3 ( i ) in lehmann and romano ( 2005 ) that @xmath705 is _ almost invariant_. moreover , since the conditions of lemma 6.5.1 ( same reference ) hold , this test is invariant under the group @xmath706 ( acting on @xmath703 ) . since
the ordered @xmath670-tuple @xmath60 of the eigenvalues of @xmath707 is maximal invariant for @xmath692 and since any invariant statistic is a measurable function of a maximal invariant one , @xmath702 must be @xmath35-measurable .
hence , @xmath705 is a @xmath35-measurable test and has the same power function as @xmath708 as was to be shown .
the existence of a @xmath53-measurable test with the same power function as that of a test @xmath700 invariant under left orthogonal transformations and multiplication by non - zero constants is established similarly.@xmath372 | this paper deals with the local asymptotic structure , in the sense of le cam s asymptotic theory of statistical experiments , of the signal detection problem in high dimension .
more precisely , we consider the problem of testing the null hypothesis of sphericity of a high - dimensional covariance matrix against an alternative of ( unspecified ) multiple symmetry - breaking directions ( _ multispiked _ alternatives ) .
simple analytical expressions for the asymptotic power envelope and the asymptotic powers of previously proposed tests are derived .
these asymptotic powers are shown to lie very substantially below the envelope , at least for relatively small values of the number of symmetry - breaking directions under the alternative .
in contrast , the asymptotic power of the likelihood ratio test based on the eigenvalues of the sample covariance matrix is shown to be close to that envelope .
these results extend to the case of multispiked alternatives the findings of an earlier study ( onatski , moreira and hallin , 2011 ) of the single - spiked case .
the methods we are using here , however , are entirely new , as the laplace approximations considered in the single - spiked context do not extend to the multispiked case . key words : sphericity tests , large dimensionality , asymptotic power , spiked covariance , contiguity , power envelope . |
this paper concentrates on relationships of formal systems with biology . in particular , this is a study of different forms and formalisms for replication . in living systems
there is an essential circularity that is the living structure .
living systems produce themselves from themselves and the materials and energy of the environment .
there is a strong contrast in how we avoid circularity in mathematics and how nature revels in biological circularity .
one meeting point of biology and mathematics is knot theory and topology .
this is no accident , since topology is indeed a controlled study of cycles and circularities in primarily geometrical systems . in this paper
we will discuss dna replication , logic and biology , the relationship of symbol and object , the emergence of form .
it is in the replication of dna that the polarity ( yes / no , on / off , true / false ) of logic and the continuity of topology meet . here
polarities are literally fleshed out into the forms of life .
we shall pay attention to the different contexts for the logical , from the mathematical to the biological to the quantum logical . in each case
there is a shift in the role of certain key concepts . in particular , we follow the notion of copying through these contexts and with it gain new insight into the role of replication in biology , in formal systems and in the quantum level ( where it does not exist ! ) . in the end
we arrive at a summary formalism , a chapter in _ boundary mathematics _
( mathematics using directly the concept and notation of containers and delimiters of forms - compare @xcite and @xcite ) where there are not only containers @xmath0 , but also extainers @xmath1 entities open to interaction and distinguishing the space that they are not . in this formalism
we find a key for the articulation of diverse relationships .
boundary algebra of containers and extainers _ is to biologic what boolean algebra is to classical logic .
let @xmath2 and @xmath3 then @xmath4 and @xmath5 thus an extainer produces a container when it interacts with itself , and a container produces an extainer when it interacts with itself . the formalism of containers and extainers is a chapter in the foundations of a symbolic language for shape and interaction . with it
, we can express the _ form _ of dna replication succinctly as follows : let the dna itself be represented as a container @xmath6 we regard the two brackets of the container as representatives for the two matched dna strands .
we let the extainer @xmath3 represent the cellular environment with its supply of available base pairs ( here symbolized by the individual left and right brackets ) .
then when the dna strands separate , they encounter the matching bases from the environment and become two dna s .
@xmath7 life itself is about systems that search and learn and become .
perhaps a little symbol like @xmath8 with the property that @xmath9 produces containers @xmath0 and retains its own integrity in conjunction with the autonomy of @xmath0 ( the dna ) could be a step toward bringing formalism to life .
* acknowledgment . *
the author thanks sofia lambropoulou for many useful conversations in the course of preparing this paper .
the author also thanks sam lomonaco , john hearst , yuri magarshak , james flagg and william bricken for conversations related to the content of the present paper .
most of this effort was sponsored by the defense advanced research projects agency ( darpa ) and air force research laboratory , air force materiel command , usaf , under agreement f30602 - 01 - 2 - 05022 .
government is authorized to reproduce and distribute reprints for government purposes notwithstanding any copyright annotations thereon .
the views and conclusions contained herein are those of the author and should not be interpreted as necessarily representing the official policies or endorsements , either expressed or implied , of the defense advanced research projects agency , the air force research laboratory , or the u.s .
( copyright 2002 . )
we start this essay with the question : during the replication of dna , how do the daughter dna duplexes avoid entanglement ? in the words of john hearst @xcite , we are in search of the mechanism for the emaculate segregation "
. this question is inevitably involved with the topology of the dna , for the strands of the dna are interwound with one full turn for every ten base pairs . with the strands
so interlinked it would seem impossible for the daughter strands to separate from their parents .
a key to this problem certainly lies in the existence of the topoisomerase enzymes that can change the linking number between the dna strands and also can change the linking number between two dna duplexes .
it is however , a difficult matter at best to find in a tangled skein of rope the just right crossing changes that will unknot or unlink it .
the topoisomerase enzymes do just this , changing crossings by grabbing a strand , breaking it and then rejoining it after the other strand has slipped through the break .
random strand switching is an unlikely mechanism , and one is led to posit some intrinsic geometry that can promote the process . in @xcite
there is made a specific suggestion about this intrinsic geometry .
it is suggested that in vivo the dna polymerase enzyme that promotes replication ( by creating loops of single stranded dna by opening the double stranded dna ) has sufficient rigidity not to allow the new loops to swivel and become entangled . in other words , it is posited that the replication loops remain simple in their topology so that the topoisomerase can act to promote the formation of the replication loops , and these loops once formed do not hinder the separation of the newly born duplexes .
the model has been to some degree confirmed @xcite .
the situation would now appear to be that in the first stages of the formation of the replication loops topoi acts favorably to allow their formation and amalgamation .
then topoii has a smaller job of finishing the separation of the newly formed duplexes . in figure 1
we illustrate the schema of this process . in this figure
we indicate the action of the topoi by showing a strand being switched in between two replication loops .
the action of topo ii is only stated but not shown . in that action , newly created but entangled dna strands would be disentangled .
our hypothesis is that this second action is essentially minimized by the rigidity of the ends of the replication loops in vivo .
@xmath10 * figure 1 - dna replication * in the course of this research , we started thinking about the diagrammatic logic of dna replication and more generally about the relationship between dna replication , logic and basic issues in the foundations of mathematics and modeling .
the purpose of this paper is to explain some of these issues , raise questions and place these questions in the most general context that we can muster at this time .
the purpose of this paper is therefore foundational .
it will not in its present form affect issues in practical biology , but we hope that it will enable us and the reader to ask fruitful questions and perhaps bring the art of modeling in mathematics and biology forward .
to this end we have called the subject matter of this paper biologic " with the intent that this might suggest a quest for the logic of biological systems or a quest for a biological logic " or even the question of the relationship between what we call logic " and our own biology .
we have been trained to think of physics as the foundation of biology , but it is possible to realize that indeed biology can also be regarded as a foundation for thought , language , mathematics and even physics . in order to bring this statement over to physics one has to learn to admit that physical measurements are performed by biological organisms either directly or indirectly and that it is through our biological structure that we come to know the world .
this foundational view will be elaborated as we proceed in this paper .
in logic it is implicit at the syntactical level that copies of signs are freely available . in abstract logic there is no issue about materials available for the production of copies of a sign , nor is there necessarily a formalization of how a sign is to be copied . in the practical realm
there are limitations to resources .
a mathematician may need to replenish his supply of paper .
a computer has a limitation on its memory store . in biology , there are no signs , but there are entities that we take as signs in our description of the workings of the biological information process . in this category the bases that line the backbone of the dna are signs whose significance lies in their relative placement in the dna . the dna itself could be viewed as a text that one would like to copy . if this were a simple formal system it would be taken for granted that copies of any given text can be made .
therefore it is worthwhile making a comparison of the methods of copying or reproduction that occur in logic and in biology . in logic
there is a level beyond the simple copying of symbols that contains a non - trivial description of self - replication .
the schema is as follows : there is a universal building machine @xmath11 that can accept a text or description @xmath12 ( the program ) and build what the text describes .
we let lowercase @xmath12 denote the description and uppercase @xmath13 denote that which is described . thus @xmath11 with @xmath12 will build @xmath14 in fact , for bookkeeping purposes we also produce an extra copy of the text @xmath15 this is appended to the production @xmath13 as @xmath16 thus @xmath11 , when supplied with a description @xmath17 produces that which @xmath12 describes , with a copy of its description attached .
schematically we have the process shown below .
@xmath18 self - replication is an immediate consequence of this concept of a universal building machine .
let @xmath19 denote the text or program for the universal building machine .
apply @xmath11 to its own description .
@xmath20 the universal building machine reproduces itself .
each copy is a universal building machine with its own description appended .
each copy will proceed to reproduce itself in an unending tree of duplications . in practice
this duplication will continue until all available resources are used up , or until someone removes the programs or energy sources from the proliferating machines .
it is not necessary to go all the way to a universal building machine to establish replication in a formal system or a cellular automaton ( see the epilogue to this paper for examples . ) . on the other hand ,
all these logical devices for replication are based on the hardware / software or object / symbol distinction .
it is worth looking at the abstract form of dna replication .
dna consists in two strands of base - pairs wound helically around a phosphate backbone .
it is customary to call one of these strands the watson " strand and the other the crick " strand .
abstractly we can write @xmath21 to symbolize the binding of the two strands into the single dna duplex .
replication occurs via the separation of the two strands via polymerase enzyme .
this separation occurs locally and propagates .
local sectors of separation can amalgamate into larger pieces of separation as well .
once the strands are separated , the environment of the cell can provide each with complementary bases to form the base pairs of new duplex dna s .
each strand , separated in vivo , finds its complement being built naturally in the environment .
this picture ignores the well - known topological difficulties present to the actual separation of the daughter strands .
the base pairs are @xmath22 ( adenine and thymine ) and @xmath23 ( guanine and cytosine ) . thus if @xmath24 then @xmath25 symbolically we can oversimplify the whole process as @xmath26 @xmath27 @xmath28 either half of the dna can , with the help of the environment , become a full dna .
we can let @xmath29 be a symbol for the process by which the environment supplies the complementary base pairs @xmath30 , @xmath31 to the watson and crick strands .
in this oversimplification we have cartooned the environment as though it contained an already - waiting strand @xmath32 to pair with @xmath33 and an already - waiting strand @xmath33 to pair with @xmath34 _ in fact it is the opened strands themselves that command the appearance of their mates .
they conjure up their mates from the chemical soup of the environment .
_ the environment @xmath35 is an identity element in this algebra of cellular interaction .
that is , @xmath35 is always in the background and can be allowed to appear spontaneously in the cleft between watson and crick : @xmath36 @xmath37 this is the formalism of dna replication .
compare this method of replication with the movements of the universal building machine supplied with its own blueprint . here
watson and crick ( @xmath33 and @xmath32 ) are each both the machine _ and _ the blueprint for the dna .
they are complementary blueprints , each containing the information to reconstitute the whole molecule .
they are each machines in the context of the cellular environment , enabling the production of the dna .
this coincidence of machine and blueprint , hardware and software is an important difference between classical logical systems and the logical forms that arise in biology .
one can look at formal systems involving self - replication that do not make a distinction between symbol and object . in the case of formal systems
this means that one is working entirely on the symbolic side , quite a different matter from the biology where there is no intrinsic symbolism , only our external descriptions of processes in such terms .
an example at the symbolic level is provided by the lambda calculus of church and curry @xcite where functions are allowed to take themselves as arguments .
this is accomplished by the following axiom .
* axiom for a lambda algebra * : let @xmath38 be an algebraic system with one binary operation denoted @xmath39 for elements @xmath40 and @xmath19 of @xmath41 let @xmath42 be an algebraic expression over @xmath38 with one variable @xmath15 then there exists an element @xmath40 of @xmath38 such that @xmath43 for all @xmath12 in @xmath41 an algebra ( not associative ) that satisfies this axiom is a representation of the lambda calculus of church and curry .
let @xmath19 be an element of @xmath38 and define @xmath44 then by the axiom we have @xmath40 in @xmath38 such that @xmath45 for any @xmath12 in @xmath41 in particular ( and this is where the function " becomes its own argument ) @xmath46 thus we have shown that for any @xmath19 in @xmath38 , there exists an element @xmath12 in @xmath38 such that @xmath47 every element of @xmath38 has a fixed point . "
this conclusion has two effects .
it provides a fixed point for the function @xmath48 and it creates the beginning of a recursion in the form @xmath49 the way we arrived at the fixed point @xmath50 was formally the same as the mechanism of the universal building machine .
consider that machine : @xmath51 we have left out the repetition of the machine itself .
you could look at this as a machine that uses itself up in the process of building @xmath14 applying @xmath11 to its own description @xmath19 we have the self - replication @xmath52 the repetition of @xmath12 in the form @xmath53 on the right hand side of this definition of the builder property is comparable with @xmath54 with its crucial repetition as well . in the fixed point theorem ,
the arrow is replaced by an equals sign !
repetition is the core of self - replication in classical logic .
_ this use of repetition assumes the possibility of a copy at the syntactic level , in order to produce a copy at the symbolic level .
_ there is , in this pivot on syntax , a deep relationship with other fundamental issues in logic . in particular this same form of repetition
is in back of the cantor diagonal argument showing that the set of subsets of a set has greater cardinality than the original set , and it is in back of the g@xmath55del theorem on the incompleteness of sufficiently rich formal systems .
the pattern is also in back of the production of paradoxes such as the russell paradox of the set of all sets that are not members of themselves .
there is not space here to go into all these relationships , but the russell paradox will give a hint of the structure . let
ab " be interpreted as
b is a member of a "
. then @xmath56 can be taken as the definition of a set @xmath57 such that @xmath13 is a member of @xmath57 exactly when it is _ not _ the case that @xmath13 is a member of @xmath14 note the repetition of @xmath13 in the definition @xmath58 substituting @xmath57 for @xmath13 we obtain @xmath59 , which says that @xmath57 is a member of @xmath57 exactly when it is not the case that @xmath57 is a member of @xmath60 this is the russell paradox . from the point of view of the lambda calculus ,
we have found a fixed point for negation . where is the repetition in the dna self - replication ?
the repetition and the replication are no longer separated .
the repetition occurs not syntactically , but directly at the point of replication .
note the device of pairing or mirror imaging .
@xmath38 calls up the appearance of @xmath61 and @xmath62 calls up the appearance of @xmath63 @xmath33 calls up the appearance of @xmath32 and @xmath32 calls up the appearance of @xmath64 each object @xmath65 calls up the appearance of its _ dual or paired object _ @xmath66 .
@xmath65 calls up @xmath66 and @xmath66 calls up @xmath67 the object that replicates is implicitly a repetition in the form of a pairing of object and dual object .
@xmath68 replicates via @xmath69 @xmath70 whence @xmath71 the repetition is inherent in the replicand in the sense that the dual of a form is a repetition of that form .
we now consider the quantum level . here
copying is not possible .
we shall detail this in a subsection . for a quantum process to copy a state
, one needs a unitary transformation to perform the job .
one can show , as we explain in the last subsection of this section , that this can not be done .
there are indirect ways that seem to make a copy , involving a classical communication channel coupled with quantum operators ( so called quantum teleportation @xcite ) .
the production of such a quantum state constitutes a reproduction of the original state , but in these cases the original state is lost , so teleportation looks more like transportation than copying . with this in mind
it is fascinating to contemplate that dna and other molecular configurations are actually modeled in principle as certain complex quantum states . at this stage
we meet the boundary between classical and quantum mechanics where conventional wisdom finds it is most useful to regard the main level of molecular biology as classical .
we shall quickly indicate the basic principles of quantum mechanics .
the quantum information context encapsulates a concise model of quantum theory : _ the initial state of a quantum process is a vector @xmath72 in a complex vector space @xmath73 observation returns basis elements @xmath74 of @xmath75 with probability _ @xmath76 where @xmath77 with @xmath78 the conjugate transpose of @xmath79 a physical process occurs in steps @xmath80 where @xmath81 is a unitary linear transformation .
note that since @xmath82 when @xmath81 is unitary , it follows that probability is preserved in the course of a quantum process .
one of the details for any specific quantum problem is the nature of the unitary evolution .
this is specified by knowing appropriate information about the classical physics that supports the phenomena .
this information is used to choose an appropriate hamiltonian through which the unitary operator is constructed via a correspondence principle that replaces classical variables with appropriate quantum operators .
( in the path integral approach one needs a langrangian to construct the action on which the path integral is based . ) one needs to know certain aspects of classical physics to solve any given quantum problem .
the classical world is known through our biology . in this sense biology
is the foundation for physics .
a key concept in the quantum information viewpoint is the notion of the superposition of states . if a quantum system has two distinct states @xmath72 and @xmath83 then it has infinitely many states of the form @xmath84 where @xmath40 and @xmath19 are complex numbers taken up to a common multiple .
states are really " in the projective space associated with @xmath73 there is only one superposition of a single state @xmath72 with itself .
dirac @xcite introduced the bra -(c)-ket " notation @xmath85 for the inner product of complex vectors @xmath86 .
he also separated the parts of the bracket into the _ bra _ @xmath87 and the _ ket _ @xmath88 thus @xmath89 in this interpretation , the ket @xmath90 is identified with the vector @xmath91 , while the bra @xmath87 is regarded as the element dual to @xmath38 in the dual space @xmath92 .
the dual element to @xmath38 corresponds to the conjugate transpose @xmath93 of the vector @xmath38 , and the inner product is expressed in conventional language by the matrix product @xmath94 ( which is a scalar since @xmath11 is a column vector ) .
having separated the bra and the ket , dirac can write the ket - bra " @xmath95 in conventional notation , the ket - bra is a matrix , not a scalar , and we have the following formula for the square of @xmath96 @xmath97 written entirely in dirac notation we have @xmath98 @xmath99 the standard example is a ket - bra @xmath100 where @xmath101 so that @xmath102 then @xmath103 is a projection matrix , projecting to the subspace of @xmath75 that is spanned by the vector @xmath104 .
in fact , for any vector @xmath90 we have @xmath105 if @xmath106 is an orthonormal basis for @xmath75 , and @xmath107 then for any vector @xmath104 we have @xmath108 hence @xmath109 @xmath110 @xmath111\ , \,\,|a>\ ] ] @xmath112 we have written this sequence of equalities from @xmath113 to @xmath114 to emphasize the role of the identity @xmath115 so that one can write @xmath116 in the quantum context one may wish to consider the probability of starting in state @xmath104 and ending in state @xmath88 the square of the probability for this event is equal to @xmath117 .
this can be refined if we have more knowledge .
if it is known that one can go from @xmath38 to @xmath118 ( @xmath119 ) and from @xmath118 to @xmath11 and that the intermediate states @xmath120 are a complete set of orthonormal alternatives then we can assume that @xmath121 for each @xmath122 and that @xmath123 this identity now corresponds to the fact that @xmath124 is the sum of the probabilities of an arbitrary state being projected into one of these intermediate states . if there are intermediate states between the intermediate states this formulation can be continued until one is summing over all possible paths from @xmath38 to @xmath125 this becomes the path integral expression for the amplitude @xmath126 we wish to draw attention to the remarkable fact that this formulation of the expansion of intermediate quantum states has exactly the same pattern as our formal summary of dna replication .
compare them .
the form of dna replication is shown below . here
the environment of possible base pairs is represented by the ket - bra @xmath127 @xmath128 @xmath129 here is the form of intermediate state expansion .
@xmath130 @xmath131 we compare @xmath132 and @xmath133 that the unit @xmath124 can be written as a sum over the intermediate states is an expression of how the environment ( in the sense of the space of possibilities ) impinges on the quantum amplitude , just as the expression of the environment as a soup of bases ready to be paired ( a classical space of possibilities ) serves as a description of the biological environment .
the symbol @xmath134 indicated the availability of the bases from the environment to form the complementary pairs .
the projection operators @xmath135 are the possibilities for interlock of initial and final state through an intermediate possibility . in the quantum mechanics
the special pairing is not of bases but of a state and a possible intermediate from a basis of states .
it is through this common theme of pairing that the conceptual notation of the bras and kets lets us see a correspondence between such separate domains .
finally , we note that in quantum mechanics it is not possible to copy a quantum state !
this is called the no - cloning theorem of elementary quantum mechanics @xcite .
here is the proof : * proof of the no cloning theorem . * in order to have a quantum process make a copy of a quantum state we need a unitary mapping @xmath136 where @xmath75 is a complex vector space such that there is a fixed state @xmath137 with the property that @xmath138 for any state @xmath139
( @xmath140 denotes the tensor product @xmath141 ) let @xmath142 note that @xmath61 is a linear function of @xmath143 thus we have @xmath144 @xmath145 @xmath146 but @xmath147 hence @xmath148 @xmath149 from this it follows that @xmath150 since @xmath151 and @xmath74 are arbitrary complex numbers , this is a contradiction . @xmath152 the proof of the no - cloning theorem depends crucially on the linear superposition of quantum states and the linearity of quantum process . by the time we reach the molecular level and attain the possibility of copying dna molecules we are copying in a quite different sense than the ideal quantum copy that does not exist .
the dna and its copy are each quantum states , but they are different quantum states ! that we see the two dna molecules as identical is a function of how we filter our observations of complex and entangled quantum states .
nevertheless , the identity of two dna copies is certainly at a deeper level than the identity of the two letters i " in the word identity .
the latter is conventional and symbolic .
the former is a matter of physics and biochemistry .
we now comment on the conceptual underpinning for the notations and logical constructions that we use in this paper
. this line of thought will lead to topology and to the formalism for replication discussed in the last section .
mathematics is built through distinctions , definitions , acts of language that bring forth logical worlds , arenas in which actions and patterns can take place .
as far as we can determine at the present time , mathematics while capable of describing the quantum world , is in its very nature quite classical . or perhaps we make it so .
as far as mathematics is concerned , there is no ambiguity in the @xmath153 hidden in @xmath154 the mathematical box shows exactly what is potential to it when it is opened .
there is nothing in the box except what is a consequence of its construction . with this in mind , let us look at some mathematical beginnings .
take the beginning of set theory .
we start with the empty set @xmath155 and we build new sets by the operation of set formation that takes any collection and puts brackets around it : @xmath156 making a single entity @xmath157 from the multiplicity of the parts " that are so collected .
the empty set herself is the result of collecting nothing " .
the empty set is identical to the act of collecting . at this point of emergence
the empty set is an action not a thing .
each subsequent set can be seen as an action of collection , a bringing forth of unity from multiplicity .
one declares two sets to be the same if they have the same members . with this prestidigitation of language
, the empty set becomes unique and a hierarchy of distinct sets arises as if from nothing .
@xmath158 all representatives of the different mathematical cardinalities arise out of the void in the presence of these conventions for collection and identification .
we would like to get underneath the formal surface .
we would like to see what makes this formal hierarchy tick .
will there be an analogy to biology below this play of symbols ? on the one hand it is clear to us that there is actually no way to go below a given mathematical construction .
anything that we call more fundamental will be another mathematical construct . nevertheless , the exercise is useful , for it asks us to look closely at how this given formality is made .
it asks us to take seriously the parts that are usually taken for granted .
we take for granted that the particular form of container used to represent the empty set is irrelevant to the empty set itself . but how can this be ? in order to have a concept of emptiness , one needs to hold the contrast of that which is empty with everything else " .
one may object that these images are not part of the formal content of set theory .
but they are part of the _ formalism _ of set theory .
consider the representation of the empty set : @xmath159 that representation consists in a bracketing that we take to indicate an empty space within the brackets , and an injunction to ignore the complex typographical domains outside the brackets .
focus on the brackets themselves .
they come in two varieties : the left bracket , @xmath160 and the right bracket , @xmath161 the left bracket indicates a distinction of left and right with the emphasis on the right .
the right bracket indicates a distinction between left and right with an emphasis on the left .
a left and right bracket taken together become a _ container _ when each is in the domain indicated by the other .
thus in the bracket symbol @xmath162 for the empty set , the left bracket , being to the left of the right bracket , is in the left domain that is marked by the right bracket , and the right bracket , being to the right of the left bracket is in the right domain that is marked by the left bracket .
the doubly marked domain between them is their content space , the arena of the empty set .
the delimiters of the container are each themselves iconic for the process of making a distinction . in the notation of curly brackets,@xmath163 , this is particularly evident .
the geometrical form of the curly bracket is a cusp singularity , the simplest form of bifurcation .
the relationship of the left and right brackets is that of a form and its mirror image . if there is a given distinction such as left versus right , then the mirror image of that distinction is the one with the opposite emphasis
this is precisely the relationship between the left and right brackets .
a form and its mirror image conjoin to make a container .
the delimiters of the empty set could be written in the opposite order : @xmath164 this is an _
extainer_. the extainer indicates regions external to itself . in this case of symbols on a line , the extainer @xmath165 indicates the entire line to the left and to the right of itself .
the extainer is as natural as the container , but does not appear formally in set theory .
to our knowledge , its first appearance is in the dirac notation of bras " and kets " where dirac takes an inner product written in the form @xmath166 and breaks it up into @xmath167 and @xmath104 and then makes projection operators by recombining in the opposite order as @xmath168 see the earlier discussion of quantum mechanics in this paper .
each left or right bracket in itself makes a distinction .
the two brackets are distinct from one another by mirror imaging , which we take to be a notational reflection of a fundamental process ( of distinction ) whereby two forms are identical ( indistinguishable ) except by comparison in the space of an observer .
the observer _ is _ the distinction between the mirror images . mirrored pairs of individual brackets interact to form either a _ container _
@xmath169 or an _ extainer _ @xmath170 these new forms combine to make : @xmath171 and @xmath172 two containers interact to form an extainer within container brackets .
two extainers interact to form a container between extainer brackets .
the pattern of extainer interactions can be regarded as a formal generalization of the bra and ket patterns of the dirac notation that we have used in this paper both for dna replication and for a discussion of quantum mechanics . in the quantum mechanics application
@xmath173 corresponds to the inner product @xmath174 , a commuting scalar , while @xmath165 corresponds to @xmath175 , a matrix that does not necessarily commute with vectors or other matrices . with this application in mind ,
it is natural to decide to make the container an analog of a scalar quantity and let it commute with individual brackets .
we then have the equation @xmath176 by definition there will be no corresponding equation for @xmath177 .
we adopt the axiom that containers commute with other elements in this combinatorial algebra .
containers and extainers are distinguished by this property .
containers appear as autonomous entities and can be moved about .
extainers are open to interaction from the outside and are sensitive to their surroundings . at this point , we have described the basis for the formalism used in the earlier parts of this paper .
if we interpret e as the environment " then the equation @xmath178 expresses the availability of complementary forms so that @xmath179 becomes the form of dna reproduction .
we can also regard @xmath180 as symbolic of the emergence of dna from the chemical substrate .
just as the formalism for reproduction ignores the topology , this formalism for emergence ignores the formation of the dna backbone along which are strung the complementary base pairs . in the biological domain
we are aware of levels of ignored structure . in mathematics
it is customary to stop the examination of certain issues in order to create domains with requisite degrees of clarity .
we are all aware that the operation of collection is proscribed beyond a certain point .
for example , in set theory the russell class @xmath57 of all sets that are not members of themselves is not itself a set .
it then follows that @xmath181 the collection whose member is the russell class , is not a class ( since a member of a class is a set ) .
this means that the construct @xmath182 is outside of the discourse of standard set theory .
this is the limitation of expression at the high end " of the formalism .
that the set theory has no language for discussing the structure of its own notation is the limitation of the language at the low end " .
mathematical users , in speaking and analyzing the mathematical structure , and as its designers , can speak beyond both the high and low ends . in biology
we perceive the pattern of a formal system , a system that is embedded in a structure whose complexity demands the elucidation of just those aspects of symbols and signs that are commonly ignored in the mathematical context .
rightly these issues should be followed to their limits .
the curious thing is what peeks through when we just allow a bit of it , then return to normal mathematical discourse . with this in mind
, lets look more closely at the algebra of containers and extainers .
taking two basic forms of bracketing , an intricate algebra appears from their elementary interactions : @xmath183 @xmath184[\ ] ] @xmath185 @xmath186<\ ] ] are the extainers , with corresponding containers : @xmath187 , \,\,\,\,\ , [ > , \,\,\,\,\ , < ] .\ ] ] these form a closed algebraic system with the following multiplications : @xmath188 @xmath189[\ , ] [ = \ , [ ] f\ ] ] @xmath190 @xmath191<\ , ] < = \ , < ] h\ ] ] and @xmath192 [ = \ , < ] g\ ] ] @xmath193 @xmath194 < = \
, < ] e\ ] ] @xmath195[\ , >
< = \ , [ > h\ ] ] @xmath196[\ , > [ = \ , [ > f\ ] ] @xmath197[\ , ]
< = \ , [ ] h\ ] ] @xmath198 @xmath199 [ = \ , [ ] g\ ] ] @xmath200 < = \ , [ ] e\ ] ] @xmath201<\ , > < = \ , < >
h\ ] ] @xmath202<\ , ] [ = \ , < ] f\ ] ] @xmath203<\ , > [ = \ ,
< > f\ ] ] other identities follow from these .
for example , @xmath204 [ > < = \ , < ] [ > e.\ ] ] this algebra of extainers and containers is a precursor to the temperley lieb algebra , an algebraic structure that first appeared ( in quite a different way ) in the study of the potts model in statistical mechanics @xcite .
we shall forgo here details about the temperley lieb algebra itself , and refer the reader to @xcite where this point of view is used to create unitary representations of that algebra for the context of quantum computation . here
we see the elemental nature of this algebra , and how it comes about quite naturally once one adopts a formalism that keeps track of the structure of boundaries that underlie the mathematics of set theory .
the _ temperley lieb algebra _
@xmath205 is an algebra over a commutative ring @xmath206 with generators @xmath207 and relations @xmath208 @xmath209 @xmath210 where @xmath211 is a chosen element of the ring @xmath206 .
these equations give the multiplicative structure of the algebra .
the algebra is a free module over the ring @xmath206 with basis the equivalence classes of these products modulo the given relations . to match this pattern with our combinatorial algebra
let @xmath212 and let @xmath213 , @xmath214[$ ] and assume that @xmath215 = [ > $ ] while @xmath216.$ ] the above equations for our combinatorial algebra then match the multiplicative equations of the temperley lieb algebra .
the next stage for representing the temperley lieb algebra is a diagrammatic representation that uses two different forms of extainer .
the two forms are obtained not by changing the shape of the given extainer , but rather by shifting it relative to a baseline .
thus we define diagrammatically @xmath217 and @xmath218 as shown below : @xmath219 @xmath220 @xmath221 @xmath222 in this last equation @xmath223 we have used the topological deformation of the connecting line from top to top to obtain the identity . in its typographical form
the identity requires one to connect corresponding endpoints of the brackets . in figure 2
we indicate a smooth picture of the connection situation and the corresponding topological deformation of the lines . we have deliberately shown the derivation in a typographical mode to emphasize its essential difference from the matching pattern that produced @xmath204 [
> < = \ , < ] [ > e.\ ] ] by taking the containers and extainers shifted this way , we enter a new and basically topological realm . this elemental relationship with topology
is part of a deeper connection where the temperley lieb algebra is used to construct representations of the artin braid group .
this in turn leads to the construction of the well - known jones polynomial invariant of knots and links via the bracket state model @xcite .
it is not the purpose of this paper to go into the details of those connections , but rather to point to that place in the mathematics where basic structures apply to biology , topology , and logical foundations .
@xmath224 * figure 2 - a topological identity * it is worthwhile to point out that the formula for expanding the bracket polynomial can be indicated symbolically in the same fashion that we used to create the temperley lieb algebra via containers and extainers .
we will denote a crossing in the link diagram by the letter chi , .
the letter itself denotes a crossing where _ the curved line in the letter chi is crossing over the straight segment in the letter_. the barred letter denotes the switch of this crossing where _ the curved line in the letter chi is undercrossing the straight segment in the letter_. in the bracket state model a crossing in a diagram for the knot or link is expanded into two possible states by either smoothing ( reconnecting ) the crossing horizontally , , or vertically @xmath1 . the vertical smoothing can be regarded as the extainer and the horizontal smoothing as an identity operator . in a larger sense , we can regard both smoothings as extainers with different relationships to their environments . in this sense
the crossing is regarded as the superposition of horizontal and vertical extainers .
the crossings expand according to the formulas @xmath225 @xmath226 the verification that the bracket is invariant under the second reidemeister move is then seen by verifying that @xmath227 for this one needs that the container @xmath0 has value @xmath228 ( the loop value in the model ) . the significant mathematical move in producing this model
is the notion of the crossing as a superposition of its smoothings .
it is useful to use the iconic symbol @xmath1 for the extainer , and to choose another iconic symbol for the identity operator in the algebra . with these choices we have @xmath229 @xmath230
thus @xmath231 @xmath232 @xmath233 @xmath234 @xmath235 @xmath236 note the use of the extainer identity @xmath237 at this stage the combinatorial algebra of containers and extainers emerges as the background to the topological characteristics of the jones polynomial .
the approach in this section derives from ideas in @xcite .
here is another use for the formalism of bras and kets . consider a molecule that is obtained by folding " a long chain molecule .
there is a set of sites on the long chain that are paired to one another to form the folded molecule .
the difficult problem in protein folding is the determination of the exact form of the folding given a multiplicity of possible paired sites .
here we assume that the pairings are given beforehand , and consider the abstract structure of the folding _ and _ its possible embeddings in three dimensional space .
_ let the paired sites on the long chain be designated by labeled bras and kets with the bra appearing before the ket in the chain order .
_ thus @xmath238 and @xmath104 would denote such a pair and the sequence @xmath239 could denote the paired sites on the long chain .
see figure 3 for a depiction of this chain and its folding . in this formalism
we do not assume any identities about moving containers or extainers , since the exact order of the sites along the chain is of great importance .
we say that two chains are _ isomorphic _ if they differ only in their choice of letters . thus @xmath240 and @xmath241 are isomorphic chains .
note that each bra ket pair in a chain is decorated with a distinct letter .
written in bras and kets a chain has an underlying parenthesis structure that is obtained by removing all vertical bars and all letters . call this @xmath242 for a given chain @xmath243 .
thus we have @xmath244 note that in this case we have @xmath245 is a legal parenthesis structure in the usual sense of containment and paired brackets .
legality of parentheses is defined inductively : 1 .
@xmath0 is legal .
if @xmath13 and @xmath246 are legal , then @xmath247 is legal .
if @xmath13 is legal , then @xmath248 is legal .
these rules define legality of finite parenthetic expresssions . in any legal parenthesis structure
, one can deduce directly from that structure which brackets are paired with one another .
simple algorithms suffice for this , but we omit the details . in any case
a legal parenthesis structure has an intrinsic pairing associated with it , and hence there is an inverse to the mapping @xmath103 .
we define @xmath249 for @xmath13 a legal parenthesis structure , to be the result of replacing each pair @xmath250 in x by @xmath251 where @xmath38 denotes a specific letter chosen for that pair , with different pairs receiving different letters .
thus @xmath252 note that in the case above , we have that @xmath253 is isomorphic to @xmath63 @xmath254 * figure 3 - secondary structure @xmath255 * a chain @xmath243 is said to be a _
secondary folding structure _ if @xmath242 is legal and @xmath253 is isomorphic to @xmath63 the reader may enjoy the exercise of seeing that secondary foldings ( when folded ) form tree - like structures without any loops or knots .
this notion of secondary folding structure corresponds to the usage in molecular biology , and it is a nice application of the bra ket formalism .
this also shows the very rich combinatorial background in the bras and kets that occurs before the imposition of any combinatorial algebra .
here is the simplest non - secondary folding : @xmath256 note that @xmath257 is legal , but that @xmath258 is not isomorphic to @xmath259 @xmath260 is sometimes called a pseudo knot " in the literature of protein folding .
figure 4 should make clear this nomenclature .
the molecule is folded back on itself in a way that looks a bit knotted .
@xmath261 * figure 4 - a tertiary structure - @xmath262 * with these conventions it is convenient to abbreviate a chain by just giving its letter sequence and removing the ( reconstructible ) bras and kets .
thus @xmath243 above may be abbreviated by @xmath263 one may wonder whether at least theoretically there are foldings that would necessarily be knotted when embedded in three dimensional space . with open ends , this means that the structure folds into a graph such that there is a knotted arc in the graph for some traverse from one end to the other .
such a traverse can go along the chain or skip across the bonds joining the paired sites .
the answer to this question is yes , there are folding patterns that can force knottedness .
here is an example of such an intrinsically knotted folding .
@xmath264 it is easy to see that this string is not a secondary structure . to see that it is intrinsically knotted , we appeal to the conway - gordon theorem @xcite that tells us that the complete graph on seven vertices is intrinsically knotted . in closed circular form ( tie the ends of the folded string together ) , the folding that corresponds to the above string retracts to the complete graph on seven vertices .
consequently , that folding , however it is embedded , must contain a knot by the conway - gordon theorem .
we leave it as an exercise for the reader to draw an embedding corresponding to a folding of this string and to locate the knot ! the question of intrinsically knotted foldings that occur in nature remains to be investigated .
some examples from cellular automata clarify many of the issues about replication and the relationship of logic and biology .
here is an example due to maturana , uribe and varela @xcite .
see also @xcite for a global treatment of related issues .
the ambient space is two dimensional and in it there are molecules " consisting in dots " ( see figure 5 ) .
there is a minimum distance between the dots ( one can place them on a discrete lattice in the plane ) . and
bonds " can form with a probability of creation and a probability of decay between molecules with minimal spacing .
there are two types of molecules : substrate " and catalysts " . the catalysts are not susceptible to bonding , but their presence ( within say three minimal step lengths ) enhances the probability of bonding and decreases the probability of decay .
molecules that are not bonded move about the lattice ( one lattice link at a time ) with a probability of motion . in the beginning
there is a randomly placed soup of molecules with a high percentage of substrate and a smaller percentage of catalysts .
what will happen in the course of time ?
@xmath265 * figure 5 - proto - cells of maturana , uribe and varela * in the course of time the catalysts ( basically separate from one another due to lack of bonding ) become surrounded by circular forms of bonded or partially bonded substrate .
a distinction ( in the eyes of the observer ) between inside ( near the catalyst ) and outside ( far from a given catalyst ) has spontaneously arisen through the chemical rules " .
each catalyst has become surrounded by a proto - cell .
no higher organism has formed here , but there is a hint of the possibility of higher levels of organization arising from a simple set of rules of interaction . _
the system is not programmed to make the proto - cells .
_ they arise spontaneously in the evolution of the structure over time .
one might imagine that in this way , organisms could be induced to arise as the evolutionary behavior of formal systems .
there are difficulties , not the least of which is that there are nearly always structures in such systems whose probability of spontaneous emergence is vanishingly small .
a good example is given by another automaton
john h. conway s game of life " . in life "
the cells appear and disappear as marked squares in a rectangular planar grid .
a newly marked cell is said to be born " .
an unmarked cell is dead " .
a cell dies when it goes from the marked to the unmarked state . a marked cell survives if it does not become unmarked in a given time step . according to the rules of life ,
an unmarked cell is born if and only if it has three neighbors . a marked cell survives if it has either two or three neighbors .
all cells in the lattice are updated in a single time step .
the life automaton is one of many automata of this type and indeed it is a fascinating exercise to vary the rules and watch a panoply of different behaviors . for this discussion
we concentrate on some particular features .
there is a configuration in life called a glider " .
see figure 6 .
this illustrates a glider gun " ( discussed below ) that produces a series of gliders going diagonally from left to right down the life lattice .
the glider consists in five cells in one of two basic configurations .
each of these configurations produces the other ( with a change in orientation ) . after four steps the glider reproduces itself in form , but
shifted in space .
gliders appear as moving entities in the temporality of the life board .
the glider is a complex entity that arises naturally from a small random selection of marked cells on the life board .
thus the glider is a naturally occurring entity " just like the proto - cell in the maturana - uribe - varela automaton .
but life contains potentially much more complex phenomena .
for example , there is the glider gun " ( see figure 6 ) which perpetually creates new gliders .
the gun " was invented by a group of researchers at mit in the 1970 s ( the gosper group ) .
it is highly unlikely that a gun would appear spontaneously in the life board .
of course there is a tiny probability of this , but we would guess that the chances of the appearance of the glider gun by random selection or evolution from a random state is similar to the probability of all the air in the room collecting in one corner .
nervertheless , the gun is a natural design based on forms and patterns that do appear spontaneously on small life boards .
the glider gun emerged through the coupling of the power of human cognition and the automatic behavior of a mechanized formal system .
cognition is in fact an attribute of our biological system at an appropriately high level of organization .
but cognition itself looks as improbable as the glider gun !
do patterns as complex as cognition or the glider gun arise spontaneously in an appropriate biological context ?
@xmath266 * figure 6 - glider gun and gliders * there is a middle ground . if one examines cellular automata of a given type and varies the rule set randomly rather than varying the initial conditions for a given automaton , then a very wide variety of phenomena will present themselves . in the case of molecular biology at the level of the dna there is exactly this possibility of varying the rules in the sense of varying the sequences in the genetic code .
so it is possible at this level to produce a wide range of remarkable complex systems .
other forms of self - replication are quite revealing . for example , one might point out that a stick can be made to reproduce by breaking it into two pieces
. this may seem satisfactory on the first break , but the breaking can not be continued indefinitely . in mathematics on the other hand , we can divide an interval into two intervals and continue this process ad infinitum . for a self - replication to have meaning
in the physical or biological realm there must be a genuine repetition of structure from original to copy . at
the very least the interval should grow to twice its size before it divides ( or the parts should have the capacity to grow independently ) .
a clever automaton , due to chris langton , takes the initial form of a square in the plane .
the rectangle extrudes a edge that grows to one edge length and a little more , turns by ninety degrees , grows one edge length , turns by ninety degrees grows one edge length , turns by ninety degrees and when it grows enough to collide with the original extruded edge , cuts itself off to form a new adjacent square , thereby reproducing itself .
this scenario is then repeated as often as possible producing a growing cellular lattice .
see figure 7 .
@xmath267 * figure 7 - langton s automaton * the replications that happen in automata such as conway s life are all really instances of periodicity of a function under iteration .
the gilder is an example where the life game function @xmath260 applied to an initial condition @xmath62 yields @xmath268 where @xmath269 is a rigid motion of the plane . other intriguing examples of this phenomenon occur .
for example the initial condition @xmath270 for life shown in figure 8 has the property that @xmath271 where @xmath272 is a rigid motion of the plane and @xmath273 and the residue @xmath11 are disjoint sets of marked squares in the lattice of the game .
@xmath270 itself is a small configuration of eight marked squares fitting into a rectangle of size @xmath274 by @xmath275 thus @xmath270 has a probability of @xmath276 of being chosen at random as eight points from @xmath277 points .
@xmath278 * figure 8 - condition d with geometric period @xmath279 * should we regard self - replication as simply an instance of periodicity under iteration ? perhaps , but the details are more interesting in a direct view .
the glider gun in life is a structure @xmath280 such that @xmath281 further iterations move the disjoint glider away from the gun so that it can continue to operate as an initial condition for @xmath260 in the same way .
a closer look shows that the glider is fundamentally composed of two parts @xmath103 and @xmath282 such that @xmath283 is a version of @xmath103 and some residue and such that @xmath284 where @xmath11 is a rectangular block , and @xmath285 is a mirror image of @xmath103 , while @xmath286 where @xmath287 is a small non - rectangular residue .
see figure 9 for an illustration showing the parts @xmath103 and @xmath282 ( left and right ) flanked by small blocks that form the ends of the gun .
one also finds that @xmath288 this is the internal mechanism by which the glider gun produces the glider . the extra blocks at either end of the glider gun act to absorb the residues that are produced by the iterations .
thus the end blocks are catalysts that promote the action of the gun .
schematically the glider production goes as follows : @xmath289 @xmath290 whence @xmath291 the last equality symbolizes the fact that the glider is an autonomous entity no longer involved in the structure of @xmath103 and @xmath292 it is interesting that @xmath282 is a spatially and time shifted version of @xmath293 thus @xmath103 and @xmath282 are really copies " of each other in an analogy to the structural relationship of the watson and crick strands of the dna .
the remaining part of the analogy is the way the catalytic rectangles at the ends of the glider gun act to keep the residue productions from interfering with the production process .
this is analogous to the enzyme action of the topoisomerase in the dna .
@xmath294 * figure 9 - p(left ) and q(right ) compose the glider gun * the point about this symbolic or symbiological analysis is that it enables us to take an analytical look at the structure of different replication scenarios for comparison and for insight .
we began with the general question : what is the relationship of logic and biology .
certain fundamentals , common to both are handled quite differently .
these are certain fundamental distinctions : the distinction of symbol and object ( the name and the thing that is named ) .
the distinction of a form and a copy of that form . in logic
the symbol and its referent are normally taken to be distinct .
this leads to a host of related distinctions such as the distinction between a description or blueprint and the object described by that blueprint .
a related distinction is the dichotomy between software and hardware .
the software is analogous to a description .
hardware can be constructed with the aid of a blueprint or description . but software intermediates between these domains as it is an _ instruction .
_ an instruction is not a description of a thing , but a blueprint for a process .
software needs hardware in order to become an actual process .
hardware needs software as a directive force .
although mutually dependent , hardware and software are quite distinct .
in logic and computer science the boundary between hardware and software is first met at the machine level with the built - in capabilities of the hardware determining the type of software that can be written for it . even at the level of an individual gate , there is the contrast of the structure of that gate as a design and the implementation of that design that is used in the construction of the gate .
the structure of the gate is mathematical .
yet there is the physical implementation of these designs , a realm where the decomposition into parts is not easily mutable .
natural substances are used , wood , metal , particular compounds , atomic elements and so on .
these are subject to chemical or even nuclear analysis and production , but eventually one reaches a place where nature takes over the task of design . in biology
it is the reverse .
no human hand has created these designs .
the organism stands for itself , and even at the molecular level the codons of the dna are not symbols .
they do not stand for something other than themselves .
they cooperate in a process of production , but no one wrote their sequence as software .
there is no software .
there is no distinction between hardware and software in biology . in logic
a form arises via the syntax and alphabet of a given formal system . that formal system arises via the choices of the mathematicians who create it .
they create it through appropriate abstractions .
human understanding fuels the operation of a formal system .
understanding imaged into programming fuels the machine operation of a mechanical image of that formal system .
the fact that both humans and machines can operate a given formal system has lead to much confusion , for they operate it quite differently .
_ humans are always on the edge of breaking the rules either through error or inspiration .
machines are designed by humans to follow the rules , and are repaired when they do not do so .
humans are encouraged to operate through understanding , and to create new formal systems ( in the best of all possible worlds ) .
_ here is the ancient polarity of syntax ( for the machine ) and semantics ( for the person ) .
the person must mix syntax and semantics to come to understanding .
so far , we have only demanded an adherence to syntax from the machines .
the movement back and forth between syntax and semantics underlies all attempts to create logical or mathematical form .
this is the cognition behind a given formal system .
there are those who would like to create cognition on the basis of syntax alone .
but the cognition that we all know is a byproduct or an accompaniment to biology .
biological cognition comes from a domain where there is at base no distinction between syntax and semantics . to say that there is no distinction between syntax and semantics in biology is not to say that it is pure syntax .
syntax is born of the possibility of such a distinction . in biology an energetic chemical and quantum substrate
gives rise to a
syntax " of combinational forms ( dna , rna , the proteins , the cell itself , the organization of cells into the organism ) .
these combinational forms give rise to cognition in human organisms .
cognition gives rise to the distinction of syntax and semantics .
cognition gives rise to the possibility of design , measurement , communication , language , physics and technology . in this paper
we have covered a wide ground of ideas related to the foundations of mathematics and its relationship with biology and with physics .
there is much more to explore in these domains .
the result of our exploration has been the articulation of a mathematical region that lies in the crack between set theory and its notational foundations .
we have articulated the concepts of container @xmath0 and extainer @xmath1 and shown how the formal algebras generated by these forms encompass significant parts of the logic of dna replication , the dirac formalism for quantum mechanics , formalism for protein folding and the temperley lieb algebra at the foundations of topological invariants of knots and links .
it is the mathematician s duty to point out formal domains that apply to a multiplicity of contexts . in this case
we suggest that it is just possible that there are deeper connections among these apparently diverse contexts that are only hinted at in the steps taken so far .
the common formalism can act as compass and guide for further exploration .
e. l. zechiedrich , a. b. khodursky , s. bachellier , r. schneider , d. chen , d. m. j. lilley and n. r. cozzarelli , roles of topoisomerases in maintaining steady - state dna supercoiling in _ escherichia coli _ , _ j. biol
. chem . _ 275:8103 - 8113 ( 2000 ) .
s. lomonaco jr , a rosetta stone for quantum mechanics with an introduction to quantum computation , in quantum computation : a grand mathematical challenge for the twenty - first century and the millennium , " ams , providence , ri ( 2000 ) ( isbn 0 - 8218 - 2084 - 2 ) . | in this paper we explore the boundary between biology and the study of formal systems ( logic ) .
+ = # 1by#2(#3 ) |
wide - field surveys have now identified nearly a thousand ultracool dwarfs ( spectral types later than m6 ) in the solar neighborhood , spanning temperatures and ( inferred ) masses that range from the stellar / substellar boundary defined by the hydrogen - burning minimum mass ( @xmath575 ) , down to the proposed substellar / planetary boundary demarcated by the deuterium - burning limit ( @xmath513 ) .
this census has enabled extensive characterization of the spectrophotometric properties of these low - luminosity , low - temperature objects .
however , studies of the substellar field population are inevitably hindered by the unknown properties of individual objects , making it difficult to disentangle the effects of varying ages , masses , and compositions on the underlying physics . in principle
, a sample of star clusters with complete membership down to very low masses and encompassing a range of ages would be ideal .
however , such a sensitive census only exists for a handful of young ( @xmath6few myr ) clusters ( e.g. * ? ? ?
* ; * ? ? ?
* ) , where the lowest mass members are easiest to detect . with increasing age , dynamical evolution leads to depletion of the lowest mass members such that older clusters may be largely devoid of brown dwarfs .
one avenue for circumventing these limitations is through the study of binary brown dwarfs , as binaries constitute systems of common ( albeit unknown ) age and metallicity .
about 15% of field brown dwarfs are tight binary systems , resolved largely by _ hubble space telescope _ or ground - based adaptive optics ( ao ) imaging and composed mostly of nearly equal - luminosity ( and thus equal - mass ) components ( e.g. * ? ? ?
these `` mini - clusters '' have proven to be fertile ground for understanding the physical processes that govern the emergent spectra of brown dwarfs ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) and for testing theoretical models with dynamical mass determinations ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
as part of our ongoing program to study the multiplicity and physical properties of substellar binaries using laser guide star adaptive optics ( lgs ao ) , we present here the discovery of the binarity of 2mass j12095613@xmath11004008 , hereinafter .
this object was identified in 2mass data by @xcite and spectrally typed by them and by @xcite as a t3 dwarf based on integrated - light near - ir spectroscopy . @xcite
define as the primary t3 spectral type standard for their near - ir classification scheme for t dwarfs .
this object has not previously been targeted with high angular resolution imaging .
what distinguishes from most previously known substellar field binaries is the large ir brightness difference between its two components , indicating an atypical mass ratio compared to the plethora of nearly equal - mass binaries . as we describe below
, provides a new opportunity to test theoretical models of brown dwarfs by requiring the model - derived ages of the two components to be consistent with coevality of the system , an approach which we refer to here as the `` isochrone test . ''
we imaged on 22 april 2007 and 16 january 2008 ut using the sodium lgs ao system of the 10-meter keck ii telescope on mauna kea , hawaii @xcite .
we used the facility ir camera nirc2 with its narrow field - of - view camera , which produces a @xmath7 field of view .
conditions were photometric for both runs .
the lgs provided the wavefront reference source for ao correction , with the tip - tilt motion measured contemporaneously from the @xmath8 mag field star usno - b1.0 0799 - 0230529 @xcite located 68away from .
the lgs brightness , as measured by the flux incident on the ao wavefront sensor , was equivalent to a @xmath9 mag star .
in april 2007 , we obtained images with the mko @xmath10 ( 1.25 ) , @xmath11 ( 1.64 ) , and @xmath12 ( 2.20 ) filters . in january 2008 , we obtained images with the @xmath13 filter , which has a central wavelength of 1.592 and a width of 0.126 .
the images were analyzed in the same fashion as our previous lgs papers @xcite .
the raw images were reduced using standard methods of flat - fielding and sky - subtraction . the binary s flux ratios and relative astrometry
were derived by fitting an analytic model of the point spread function ( psf ) as the sum of three elliptical gaussians .
the measurements were made on the individual images , with outlier images of much poorer quality excluded .
the averages of the fitting results were adopted as the final measurements and the standard deviations as the errors .
the final flux ratios were used to correct the strehl ratio measurements of the images for the contamination from the light of the secondary component . we did not correct the relative astrometry for instrumental optical distortion , as the size of the effect as predicted from a distortion solution by b. cameron ( priv . comm . )
is insignificant compared to our measurements uncertainties . in order to validate our measurements
, we created myriad artificial binary stars from lgs images of single stars with comparable strehl and fwhm as the science data . for each filter ,
our fitting code was applied to artificial binaries with similar separations and flux ratios as over a range of pas .
the simulations showed that the random errors are reasonable and any systematic offsets are small . in cases where the measurement errors from the artificial binaries were larger than those from the measurements , we conservatively adopted the larger errors .
we adopted a pixel scale of @xmath14 mas / pixel and an orientation for the detector s
@xmath15 axis of @xmath16 for nirc2 as determined by @xcite .
we computed the expected shift in the relative astrometry of the two components due to differential chromatic refraction ( dcr ) in the same manner as in our previous work ( e.g. * ? ? ? * ) .
the dcr effects are small ( @xmath513 mas ) compared to the measurement errors , and therefore we did not account for them .
table [ table : keck ] presents our final keck lgs measurements , and figure [ fig : images ] shows our keck lgs images .
the near - ir broadband colors of b are extremely blue , which provides strong circumstantial evidence that it is a physically associated late - t dwarf companion and not a background object .
@xcite measure a proper motion for of @xmath17/yr at a pa of @xmath18 . considering this proper motion , examination of the digitized sky survey images shows no plausible optical counterpart if component b were a background object .
our two epochs of keck lgs imaging separated by 9 months are consistent with no change in the relative position of the two components .
if the companion were a background object , in jan 2008 the system should have had a separation of @xmath19 at a pa of @xmath20 , which is @xmath21 discrepant with the observed separation , neglecting the effect of parallax .
finally , our lgs photometry indicates the presence of @xmath11-band methane absorption in component b ( [ sec : phot ] ) , inconsistent with the background star hypothesis .
thus both the astrometry and photometry independently indicate that the system is a true physical binary .
we use our measured flux ratios and the published @xmath22 photometry from @xcite to derive resolved ir colors and magnitudes for on the mko system . for the @xmath13 data , we synthesize integrated - light photometry of @xmath23 mag from the near - ir spectrum of @xcite , flux - calibrated to the broadband mko photometry of @xmath24 mag .
we assume that the components of are themselves single , not unresolved binaries .
table [ table : resolved ] presents the resolved photometry . comparing to known ultracool dwarfs , figure [ fig : colorcolor ] shows that component a has ir colors most typical of t2t3 dwarfs .
the uncertainties are larger for component b , but its near - ir colors appear to be comparable or even bluer than the very latest t dwarfs .
the resolved @xmath25 colors provide more accurate estimates of the spectral types than the broadband colors , as the @xmath11-band methane absorption correlates well with overall near - ir spectral type ( e.g. , figure 2 of @xcite ) .
we convert these colors to near - ir spectral type using the polynomial fit from @xcite : @xmath26 where @xmath27 for t0 , = 21 for t1 , etc .
the rms scatter about the fits is 0.3 subclasses .
the fit was originally derived for objects from t0 to t8 , but as discussed below the more recently discovered t9 objects are still well - described by this relation .
the observed @xmath25 colors for give spectral types of t1.6 @xmath28 0.9 and t9 @xmath28 2 for components a and b , respectively , where the spectral type uncertainties come from propagation of the errors in the colors .
note that the @xmath25 color of component b ( @xmath29 mag ) is bluer than any of the objects used to define the polynomial fit , so the spectral type estimate represents a small extrapolation of the relation .
for comparison , the colors of the t8.5 dwarfs ulas 1238 + 0953 and ulas 2146@xmath10010 are @xmath10.65 and @xmath10.63 mags , respectively , and those of the t9 dwarfs ulas 0034@xmath10052 , cfbds 0059@xmath10114 , and ulas 1335 + 1130 are @xmath10.56 , @xmath10.63 , and @xmath10.66 mags , respectively , as synthesized from their published near - ir spectra .
( these very late - t dwarfs are not included in the liu et al .
fit , but their inclusion would make no significant difference in the polynomial relation . ) overall , component b shows @xmath11-band methane absorption comparable to the latest known t dwarfs .
we modeled the integrated - light spectrum of as the composite sum of two template t dwarfs , as an additional means to determine the resolved spectral types .
the templates were chosen from available near - ir spectra for dwarfs that have similar colors to those of 2mass j1209@xmath11004 a and b. we produced various fits using sdss j0758 + 3247 ( t2 ; @xcite ) , sdss j1254@xmath10122 ( t2 ; @xcite ) and sdss j1521 + 0131 ( t2 ; @xcite ) as templates for the component a and gl 570d ( t7.5 ; @xcite ) , 2mass j0415@xmath10935 ( t8 ; @xcite ) , ulas j1017 + 0118 ( t8p ; @xcite ) and ulas j1315 + 0826 ( t7.5 ; @xcite ) as templates for component b. the @xmath22 colors of the template dwarfs and the resolved binary are within 0.2 mag for the component a and 0.4 mag for the component b ( whose observed photometry is much more uncertain ) . before combining each template pair , we scaled each template spectrum to a distance of 10 pc .
trigonometric parallaxes are known for sdss j1254@xmath10122 and 2mass j0415@xmath10935 . for the other dwarfs distances
were estimated using the relationships between absolute magnitudes and spectral type from @xcite .
we explored fits using both the `` bright '' and `` faint '' relations , which differ significantly for early - type t dwarfs . for each pair , after shifting the flux to 10pc , the flux at each wavelength was summed , and the result compared to the observed unresolved spectrum for .
there is no parallax measurement for , and so we shifted the observed spectrum to that of the synthesized spectrum , determined the implied distance , and compared the resulting spectra .
in the end , model composite spectra composed of a t2 primary and t7.5 secondary showed good agreement with the observed integrated - light spectrum ( figure [ fig : blends ] ) .
any combination using the peculiar t8 dwarf ulas j1017 + 0118 failed to reproduce the observed strength of the feature at 1.6 .
the best fits were obtained using the t2 dwarf sdss j1254@xmath10122 in combination with either of the t7.5 dwarfs gl 570d or ulas j1315 + 0826 , or using the t2 dwarf sdss j1521 + 0131 in combination with the t8 dwarf 2mass j0415@xmath10935 . for the last combination to work , the distance for sdss j1521
+ 0131 must be derived using the `` faint '' liu et al.photometric relation .
the implied distance for is 23 pc from the sdss j1254@xmath10122 fits , or 18 pc from the sdss j1521 + 0131 fit . combining the results from the @xmath22 and @xmath25 colors and this spectral modeling , we adopt a spectral type of t2 @xmath28 0.5 for a , as all the available estimates are in good agreement .
the typing for component b is more challenging , given its larger photometric errors and its relative faintness ( and thus modest contribution to the integrated - light spectrum ) . in the spectral modeling , t7.5 templates provided the best matches .
our attempts to use t8 dwarfs failed to produce a good match , though this was possibly limited by the small number of templates with comparably blue near - ir colors as b ( 2mass j0415@xmath10935 and ulas j1017 + 0118 ) .
the polynomial fits for absolute magnitude as a function of spectral type from @xcite give @xmath30 ) = \{15.1 , 15.5 , 15.9 , 16.3 } mag , @xmath31 = \{15.5 , 15.9 , 16.3 , 16.7 } mag , and @xmath32 = \{15.6 , 16.0 , 16.4 , 16.7 } mag for near - ir spectral types of t6.5 , t7 , t7.5 , and t8 , respectively . based on the photometric distance to component
a ( [ sec : distance ] ) , the absolute magnitudes of component b are @xmath33 mag , in good agreement with t7.5 . given the concordance between this and the spectral modeling
, we adopt a spectral type estimate of t@xmath34 for component b. @xcite chose as the primary t3 spectral standard for near - ir classification of t dwarfs .
its binary nature now lessens its utility for this purpose , especially at the bluer wavelengths where the contribution of the secondary to the integrated - light spectrum is greater . the secondary standard proposed by burgasser et al .
was sdss j1021@xmath10304ab , which also has been resolved into a binary ( t1+t5 ; @xcite ) .
for both sources , their binary nature make them unappealing as the t3 spectral standard , since the secondary contributes significantly to the integrated near - infrared light ( e.g. , figure 3).10 . ]
furthermore the primary of neither system is of t3 type .
options for a different spectral standard are limited , due to the modest number of t3 dwarfs known . at the relatively brighter magnitudes , the choices are sdss j120602.51 + 281328.7 ( t3
; @xmath35 mag ) and sdss j141530.05 + 572428.7 ( t3@xmath281 ; @xmath36 ) , both of which were identified and classified by @xcite .
sdss j1206 + 2813 appears to be single in 0.06 fwhm images obtained by us with keck lgs ( liu et al .
, in prep . )
while no high angular resolution imaging is available yet for sdss j1415 + 5724 . as the individual spectral indices for sdss j1206 + 2813 are more consistent with a t3 typing and it appears to be single , we suggest that it be used as the t3 spectral standard . with an estimated spectral type of t@xmath37 for component
a , we use the @xcite polynomial fits of absolute magnitude as a function of spectral type to determine a photometric distance .
we use the @xmath10-band photometry ; the @xmath11 and @xmath12-band data produce nearly identical results but with somewhat larger uncertainties .
errors in the photometric distance include the uncertainties in the apparent magnitude ( 0.05 mag ) , the uncertainty in @xmath10-band absolute magnitude arising from the spectral type uncertainty ( 0.05 mag ) , and the intrinsic dispersion about the fit ( 0.4 mag ) .
liu et al .
provide both a `` bright '' and a `` faint '' relation , depending on the inclusion or exclusion of candidate overluminous objects of early / mid - t spectral type , respectively .
the two relations give consistent answers within the errors , @xmath38 and @xmath39 pc , respectively .
we average to arrive at a final photometric distance of @xmath40 pc .
pc based on its type of t@xmath34 .
this value is consistent with the distance derived component a , but it can not be considered truly an independent measurement , since we use the photometric distance in estimating the spectral type for component b in
[ sec : spectra ] . ] since the spectral templates used for modeling the composite spectra have distance determinations ( [ sec : spectra ] ) , the scaling factor used to match to ab also provides a distance estimate to the binary , with a value of 18 or 23 pc ( [ sec : spectra ] ) .
an error estimate from this approach is uncertain . the t2 dwarf sdss j1254@xmath10122 @xcite provides a well - matching template for component a and has parallax with an uncertainty of 4% @xcite , which represents the very smallest possible uncertainty . at the other extreme ,
the photometric relations of @xcite have a rms scatter about the fit of 1520% in distance . adopting a distance error of 10%
is a reasonable compromise ; for instance , liu et al .
demonstrate that fitting the composite spectrum of the t1+t6 binary @xmath41 ind bab produces a distance estimate that agrees to 10% with the measured parallax .
thus , the distance estimates from the spectral modeling agree with the photometric distance of @xmath40 pc .
we conservatively adopt the latter in our subsequent analysis.0122 is slightly brighter than the values from the adopted liu et al .
polynomial fits .
in fact , it has been suggested this system may be a binary based on this modest overluminosity @xcite , though it has not been resolved as such in high angular resolution imaging from _ hst _ or keck lgs ao ( @xcite ; liu et al . in prep . ) and modeling of its integrated - light spectra suggests the system is more likely be a relatively young , single object as opposed to a near - equal mass binary @xcite . ] there is no published value for the integrated - light bolometric luminosity of , so we computed it directly from its near - ir spectrum @xcite , flux calibrated with the @xcite photometry , and combined with mid - ir photometry . for the latter , irac photometry is available in the _ spitzer _ archive , obtained as part of the cycle 1 gto program 35 on ut 2005 june 13 .
exposure times were 30 s , and a 5-position small dither pattern was used .
the binary is unresolved in the irac images , and an 8 " aperture was used for the photometry .
we determined magnitudes of @xmath42=14.02\pm0.004 $ ] , @xmath43=13.49\pm0.004 $ ] , @xmath44=13.33\pm0.016 $ ] , and @xmath45=13.06\pm0.06 $ ] mags .
a 0.03 mag error should be added in quadrature to the quoted random errors to account for systematic effects . to derive the integrated - light , we numerically integrated the near - ir spectrum and the mid - ir photometry , interpolating between the gaps in the data , extrapolating at shorter wavelengths to zero flux at zero wavelength , and extrapolating beyond 4.5 assuming a blackbody .
we determined the luminosity error in a monte carlo fashion , by adding randomly drawn noise to our data over many trials and computing the rms of the resulting luminosities . in this process , we accounted for the noise in the spectrum , the errors in the mko photometry used to flux - calibrate it , and the errors in our mid - ir photometry estimates . before accounting for the error in the photometric distance
, we found a total bolometric luminosity of @xmath46 dex . with the 20% uncertainty in the photometric distance included , the uncertainty becomes 0.15 dex .
to compute the bolometric luminosities ( ) for each component , we apply bolometric corrections to the resolved near - ir photometry based on the compilation in appendix [ sec : bolcorr ] . despite the relatively larger rms about the @xmath47 fit at @xmath10-band ,
the photometry of the component b is most accurately determined in this bandpass , so we use it to to determine and associated quantities .
the results are given in table [ table : resolved ] .
the uncertainties are composed of the uncertainties from the apparent magnitudes , from the bolometric corrections arising from the spectral type uncertainty , the intrinsic dispersion about our polynomial fits , and the photometric distance .
the last item dominates the uncertainties in the values .
( however , the luminosity _ ratio _ of two components is known to much higher precision than the individual values , since the ratio is independent of the distance uncertainty . ) as a consistency check , we note that the sum of the individual component luminosities derived from @xmath48 ( @xmath49 dex ) agrees well with the direct computation from the integrated light spectrum ( @xmath50 dex ) . with the determinations for the individual components and an assumed age , evolutionary models
can then fully determine the physical properties of the two components ( e.g. * ? ? ?
figure [ fig : teff - logg ] shows the results graphically using the @xcite evolutionary models , where we have adopted a generic age range of 0.110 gyr for the system .
the figure indicates the complete possible set of \ { , } values for the two components , based on our determinations .
table [ table : evolmodels ] summarizes the physical properties derived from this approach , for a range of assumed ages .
note that for computing relative quantities between the components ( e.g. , luminosity ratio and mass ratio ) , we take care to account for the covariance of the measurements due to the fact that the uncertainty in the distance is common to both components , and thus uncertainties in these relative quantities are smaller than for the absolute value of these quantities for the individual components . to do this
, we draw values for all the relevant measurements in a monte carlo fashion , keeping track of all the calculations of the physical parameters in every trial , and then compute the final result and rms from the ensemble of trials . to compute the estimated orbital period of the system , following @xcite , we assume random viewing angles and a uniform eccentricity distribution from @xmath51 .
this gives a multiplicative correction factor of 1.10@xmath52 ( 68.3% confidence limits ) for converting the projected separation into a semi - major axis .
the estimated orbital periods range from 1626 years for ages of 0.5 , 1.0 , and 5.0 gyr , albeit with significant uncertainties .
dynamical masses for ultracool binaries are possible from orbital monitoring covering @xmath5330% of the orbital period .
thus , warrants continued high angular resolution imaging to determine its visual orbit , combined with a parallax measurement to compute its total mass .
the system is a rare example of an ultracool binary with a highly unequal mass ratio ( @xmath2 ) , as most ultracool binaries are composed of nearly equal components ( table [ table : binaries ] ) .
this novelty is highlighted in figure [ fig : massratio ] , which shows the estimated mass ratios of all known field brown dwarf binaries , where we choose field systems in which the primary has an estimated spectral type of l4 or later as a proxy for the stellar / substellar boundary @xcite . as explained below
, we specifically focus on ultracool binaries composed of two brown dwarfs , as opposed to those that contain one or two very low - mass stars . under the conservative assumption of coevality
, we can use to test current theoretical models of brown dwarfs , by assessing whether the two components lie along a single isochrone on the hertzsprung - russell ( h - r ) diagram .
we essentially treat the system as a star cluster composed of two members , with an unknown age but requiring that it is the same for components a and b. hereinafter , we refer to this approach as the `` isochrone test .
'' placement on the h - r diagram requires and .
the former can be well - determined for ultracool dwarfs ( appendix [ sec : bolcorr ] and references therein ) .
however , robust temperature determinations are still an active area of investigation ( e.g. * ? ? ?
atmospheric models for l and t dwarfs have not been confirmed by direct measurement of brown dwarf radii , only checked for consistency against evolutionary models using single brown dwarfs that are companions to stars of known age ( e.g. * ? ? ?
* ; * ? ? ?
* ) and the few brown dwarf binaries with dynamical mass determinations ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
modelling low - temperature atmospheres is challenging .
the molecular opacity line list for ch@xmath54 , a species important for both the l and t dwarfs , is incomplete or non - existent below 1.6 .
the feh line list is incomplete in the 1.21.3 region , which is important for l dwarfs .
nh@xmath55 is important for very late - type t dwarfs , but there are no opacities calculated below 1.4 . also , calculation of the very strongly pressure - broadened resonance doublet ( 0.78 ) is difficult , and hence models do not currently reproduce the far - red region of t dwarf spectra very well . finally , the treatment of the grains condensates is difficult , which impacts the 1.01.6 region of @xmath5l3 to @xmath5t2 types .
thus , our current ability to carry out the isochrone test represents a test of the _ joint _ accuracy of evolutionary and atmospheric models , not of either specific class of models . the fact that the isochrone test relies on components of very different masses also may make it a useful signpost for mass - dependent problems , which would not be revealed when studying the ( much more common ) nearly - equal mass binaries
of course , the coevality requirement is widely applied in studies of star clusters .
its particular characteristic in the case of substellar binaries arises from the fact that brown dwarfs steadily cool over their lifetime , and thus the isochrones for substellar objects are more dispersed on the h - r diagram compared to stellar isochrones .
we now subject to the isochrone test .
a parallax measurement is not available , so we use the photometric distance estimate ( [ sec : distance ] ) this is not a limiting factor in the analysis . for the effective temperatures of each component
, we consider previous analysis of the spectra for similarly typed t dwarfs .
three t2 dwarfs have been analyzed in such a fashion : sdss j0758 + 3247 ( 11001200 k ; @xcite ) , sdss j1254@xmath10122
( 11001200 k ; @xcite ) , and hn peg b ( 1115 k ; @xcite ) .
the model atmospheres were spaced by 100 k ( except for hn peg b ) , and the estimated systematic uncertainties are @xmath5100 k due to , e.g. , flux calibration , choice of wavelength region for fitting , and weighting scheme . thus we adopt @xmath56 k as the temperature for a. no t1 or t3 dwarfs have been subjected to such analysis , but the fitted for early / mid - t dwarfs are all relatively constant ( e.g. * ? ? ? * ) , in accord with the measurements derived from data that the l / t transition for field objects occurs at nearly constant temperature ( e.g. * ? ? ? * ) .
four t7.5 dwarfs have been analyzed to date : hd 3651b ( 820830 k ; @xcite ) , 2mass j1114@xmath12618 ( 725775 k ; @xcite ) , gl 570 d ( 800 - 821 k ; @xcite ) , and 2mass j1217@xmath10311 ( 850950 k ; @xcite ) .
to encompass the potential uncertainty in the spectral type of component b , we also consider the t8 dwarf 2mass j0415@xmath10935 ( 725775 k , @xmath57 k ; @xcite).0935 , the model atmosphere matching the derived from evolutionary models was shown to be consistent with the observed flux - calibrated spectrum , as opposed to directly fitting model atmospheres to the spectra .
] no t7 dwarfs have independently determined s .
taking the unweighted mean and rms of the results gives @xmath58 k as our adopted value for b. figure [ fig : hrd - both ] shows on the h - r diagram .
we compute the individual component ages based on their h - r diagram positions relative to the lyon / cond and tucson @xcite evolutionary models , using a monte carlo method to account for the uncertainties .
for each component , we draw trial values for and from a normal distribution , which are then used with finely interpolated tabulations of the models to derive a set of ages .
we then take the difference of the component ages and compile the distribution resulting from 10@xmath59 trials .
figure [ fig : hrd - agediff - both ] plots the result , which is well - described by a gaussian .
table [ table : ages ] gives the derived ages and age difference . at the 1@xmath60 level ,
the models successfully indicate that the two components are coeval , with a @xmath61 difference of @xmath62 and @xmath63 dex inferred from the lyon and tucson models , respectively .
we also apply the isochrone test to the bab system , which is a wide ( 1459 au ) separation binary companion to the k5 v star a .
the binary is composed of a t1 and a t6 component , with a @xmath12-band flux ratio of @xmath64 mag .
the system is quite similar to , but with more precisely determined physical parameters and therefore offers a more stringent application of the isochrone test . indicates that 2mass j22551861@xmath15713056ab ( @xmath65 ) , 2mass j05185995@xmath12828372ab ( @xmath66 ) , 2mass j12255432@xmath12739476ab ( @xmath65 ) have comparably small mass ratios as and bab . for these other binaries , the existing resolved measurements are insufficient to apply the isochrone test . ] for the of the two components , we use the results from of @xmath67 and @xmath68 .
their measurements come from integrating the resolved spectra of the two components from optical to mid - ir wavelengths .-
band photometry and our bolometric corrections in
[ sec : lbol ] . using the high quality parallax to a ( @xmath69 mas ; ) ,
the @xmath70-band absolute magnitudes for the two components are @xmath71 and @xmath72 mag , and @xmath12-band bolometric corrections are @xmath73 and @xmath74 mag ( including the intrinsic scatter and the @xmath280.5 subclasses uncertainty in the spectral types ) .
together , these give @xmath75 and @xmath76 dex for ba and bb , respectively . ]
the effective temperatures of the two components have been derived by fitting model atmospheres by two groups .
( 1 ) @xcite analyzed resolved near - ir ( @xmath22 ) spectroscopy of the two components using @xcite model atmospheres computed with @xmath2850 k spacing .
they identified the best 3 fitting models to the data .
we take the mean and rms of their results , @xmath77 k and @xmath78 k. the 25 k uncertainty is likely underestimated , as discussed in
[ sec : teff ] , especially as the fit to component ba does not appear as good as that for component bb .
( 2 ) obtained resolved 0.65.0 spectra of the two components and compared the full spectra and the 1.25 absorption doublet to bt - settl model atmospheres .
their comparisons applied an absolute flux calibration to the model atmospheres , based on the known luminosity and distance to the system .
they found good agreement for @xmath7913001340 k and 880940 k for the two components .
we adopt the mean values and use the quoted range as the rms .
using the temperatures , figure [ fig : hrd - both ] shows the position of the two components of bab on the h - r diagram , and figure [ fig : hrd - agediff - both ] shows the age difference between the two components , computed in the same fashion as
[ sec : isochrone-2m1209 ] . like for ,
the data are consistent with coevality for the binary , with a @xmath61 difference of @xmath80 and @xmath81 dex from the lyon and tucson models , respectively , though only at the 2@xmath60 level when using the lyon / cond models .
the uncertainties are smaller for compared to due to the much more precise distance and temperatures .
the @xcite temperatures give a similar result for the component ages , in fact slightly more consistent with coevality , though not significantly so given the uncertainties .
our conclusion of coevality for the system agrees with the ( somewhat different ) approach of @xcite , who determine and by fitting model atmospheres to the resolved spectra and then use the results to compute the ages from evolutionary models .
however , note that @xcite determine a hotter of @xmath82 k for ba based on high resolution near - ir spectroscopy ; if this temperature is adopted , then this component s h - r diagram position is off the model loci , with an age older than 10 gyr and inconsistent with coevality .
the age we derive from the h - r diagram is consistent with independent constraints on the age from the primary star a. the most recently used estimate is 0.82.0 gyr from , based on their semi - empirical relation between stellar rotational period and age .
appendix [ sec : epsind - age ] updates the age estimate , leading to a broader range of 0.57.0 gyr . both values for a agree well with the median ages of 0.62.5 gyr derived by us from the h - r diagram analysis . for both and bab , while the models do successfully indicate that the systems are coeval , the mass information derived from the h - r diagram reveals some interesting issues .
table [ table : ages ] gives the computed component masses and mass ratios , based on the same calculations used to determine the ages . at face value
, the hr diagram - derived mass ratio for appears to be inverted , where the much brighter component a is non - sensibly indicated to be the lower - mass component .
however , the formal uncertainties are very large and at the @xmath83 level , consistent with the @xmath2 value inferred from evolutionary models ( [ sec : properties ] and table [ table : evolmodels ] ) .
for bab , the mass ratio from the h - r diagram is in good agreement with the value of @xmath84 derived by , based on the component luminosities , evolutionary models , and an assumed age of 0.82.0 gyr ( where we have taken the average and rms of their computed values ) .
so within the accuracy afforded by the current data , the mass ratios derived from the h - r diagram seem reasonable .
however , a significant disagreement occurs when considering the total mass of .
the total mass derived from the h - r diagram is @xmath85 ( @xmath86 ) as derived from the lyon ( tucson ) models and the @xcite temperatures , with the temperatures giving a slightly higher total mass of @xmath87 ( @xmath88 ) .
( note that the calculation of the total mass from the h - r diagram is _ independent _ of the aforementioned uncertainty in a s age ) .
such values strongly conflict with the preliminary dynamical mass of @xmath89 measured by @xcite from the binary s relative orbital motion . and
thus , there is a problem in one or more of the input data or models involved in the analysis .
the observations of the component s and total dynamical mass seem secure .
the s are determined by using resolved spectroscopy across a very broad wavelength range , agree well with the values obtained using near - ir photometry with our well - determined bolometric corrections ( [ sec : isochrone - epsind ] ) , and the uncertainties are fully propagated through our analysis . likewise , the cardoso et al.dynamical mass also seems reasonable given the large angular separation between the two components , the large number of observing epochs , and the long duration of the monitoring ( @xmath540% of the derived orbital period , which should be sufficient to determine an accurate mass , e.g. , as demonstrated by and @xcite ) .
assuming for the moment that the loci of evolutionary models are correct , this discrepancy in total mass of must arise from errors in the temperatures assigned to the binary components from the model atmosphere fitting . generally speaking
, changes to the relative temperatures between the two components would alter the inferred mass ratio , whereas systematic shifts to the temperatures assigned to both components would alter the derived total mass .
figures [ fig : plot - teff - changes - epsind - trelative ] and [ fig : plot - teff - changes - epsind - tshift ] summarize the changes to the hr - diagram - derived properties of the system ( component ages , age difference , mass ratio , and total mass ) if the s for the two components are changed in these three ways : relative changes in the s , overall shifts in the s , and changes to the of only one component . these calculations place the following constraints on the estimates : * the _ relative _ difference between ba and bb appears to be consistent with most of the constraints , namely coevality , an estimated age of @xmath51 gyr , and a mass ratio of @xmath50.6 ( figure [ fig : plot - teff - changes - epsind - trelative ] ) .
however , the total mass of the system is under - predicted by the evolutionary models using these s . * the disagreement between the total mass from the h - r diagram and the orbital monitoring disappears if both components are @xmath550100 k hotter ( figure [ fig : plot - teff - changes - epsind - tshift ] ) , which still leaves the system coeval .
however , an older age of @xmath56 gyr then results . *
changing the for only one of the components can not simultaneously fulfill the requirements of coevality and agreeing with the dynamically measured total mass ( figure [ fig : plot - teff - changes - epsind - both ] ) .
altogether , this analysis points to a small , but non - zero , systematic error of @xmath550100 k in the published values from model atmosphere fitting .
this seems plausible given the known inadequacies of the model atmospheres , especially in the l / t transition region ( [ sec : isochrone ] ) , and thus invoking errors in the evolutionary models is not compelling .
the implied temperatures for the two components would then be @xmath513501400 k and @xmath5950 - 1000 k for components ba and bb , respectively . for component
ba , this hotter temperature is consistent with the @xmath82 k found by @xcite and noticeably hotter than previous fitting of early - t dwarfs ( @xmath511001200 k ; [ sec : isochrone-2m1209 ] ) . for component
bb , the implied hotter temperature agrees with that found expected for mid / late - t dwarfs based on model atmosphere fitting by @xcite , where a linear fit to their results gives 1030 k for the t6 subclass ( equation 7 of @xcite ) .
the systematic errors that we infer from the isochrone method are comparable to the estimated uncertainties in spectral fitting of field t dwarfs ( [ sec : isochrone-2m1209 ] and references therein ) , though provides a more stringent test of the models given this binary s abundance of observational constraints compared to ordinary field brown dwarfs .
also , note that the @xmath90-axes in figures [ fig : plot - teff - changes - epsind - trelative] [ fig : plot - teff - changes - epsind - both ] show the full possible range of changes ; larger changes would displace one or both binary components off the model loci on the h - r diagram , leading to ages that are either far too young ( @xmath9110 myr ) or too old ( @xmath5310 gyr ) . finally , we point out that given the measured s and dynamical mass for , figure [ fig : teff - logg ] shows that the evolutionary models would lead to an inferred age of about 5 gyr . in the same fashion ,
king et al .
( 2010 ) have inferred an age of 3.7 - 4.3 gyr from the evolutionary models ( see also cardoso et al .
this is at the older end of the age range inferred for the k5v primary star a ( appendix [ sec : epsind - age ] ) , though the age - dating methods are largely derived for @xmath911 gyr solar - type stars and are not well - calibrated for such old ages or for such a late - type primary .
in fact , it may be that the age of the abab system is most accurately determined from brown dwarf evolutionary models , though it will be difficult to test this hypothesis independently .
we have identified the t3 dwarf as a tight substellar binary , with an extreme mass ratio ( @xmath2 ) compared to previously known field brown dwarf binaries .
the near - ir photometry combined with modeling of the integrated - light spectrum indicates spectral types of [email protected] and [email protected] for the two components .
this newly discovered binarity lessens the utility of this object as the primary near - ir spectral - type standard for t3 dwarfs .
we suggest sdss j1206 + 2813 as a replacement .
the highly unequal mass ratio of the system allows us to test the joint accuracy of evolutionary and atmospheric models , by examining if the ages inferred from the h - r diagram position of the two components are consistent with coevality . using the photometric distance to the system and latest available s from model atmosphere fitting of t dwarfs
, we find that the models successfully indicate that the two components of are coeval , with a difference of @xmath92 ( @xmath63 ) dex based on the lyon ( tucson ) models . a similar analysis of the t1+t6 binary also supports the accuracy of the models , with the system being consistent with coevality at the 12@xmath60 level , depending on the choice of model atmosphere fits and evolutionary models .
( the temperatures for bab from in combination with the lyon / cond models produce the most non - coeval results , with an age difference of @xmath93 dex . ) while the models succeed in respect to the isochrones , for bab the total mass derived from the h - r diagram ( @xmath580 ) is discrepant with the measured dynamical mass of @xmath89 .
this is most simply resolved by assuming a systematic increase in for the two components by @xmath550100 k warmer , i.e. , an increase to @xmath51375 k and 950 k for components ba and bb , respectively .
such an error is not unexpected , given the known inadequacies of the model atmospheres in the l / t transition region , the quality of the spectral fit , and the general uncertainties in the procedures used to fit such models to observed spectra .
such a increase leads to an implied age of about 6 gyr for the system , somewhat older than previous estimates . while the @xmath550100 k errors are comparable to those typically cited in the atmospheric model fitting , it is worth emphasizing that indicates that such errors are present ( i.e .. non - zero ) , as the masses inferred from the h - r diagram are strongly discrepant with the dynamical mass , _ a result that is independent of the uncertainty in the age of the system .
_ however , assuming the evolutionary models are basically correct , errors much larger than @xmath5100 k are not warranted , as the inferred properties from the h - r diagram would then be inconsistent with the underlying constraints of coevality and a mass ratio @xmath941.0 .
also note that the very precise results from the t5+t5.5 mass - benchmark binary 2mass j1534@xmath12952ab and the m4+t8.5 age - benchmark system wolf 940ab point to @xmath5100 k overestimates of when model atmospheres are used to fit the near - ir spectrum of t dwarfs @xcite , similar in amplitude though opposite in sign to what we find here from . for wolf 940b ,
@xcite show that luminosity - constrained model fits to the near - ir spectrum of this t8.5 dwarf are poor , with the synthetic spectrum @xmath10 and @xmath11-band peaks too high and the @xmath12-band peak too low .
most likely this is due to the known incompleteness of the molecular opacity line lists in this region .
thus , substantial errors in evolutionary and/or atmospheric models for t dwarfs do not appear to be required , though it is not possible to fully reconcile all the model predictions with the observations , especially for bab where the measurements are exceptionally precise .
it is also worth noting that @xcite have found that the luminosities predicted by evolutionary models at @xmath50.8 gyr may be systematically underestimated by a factor of 23 , based on the luminosities and dynamical masses of the l4+l4 substellar binary hd 130948bc .
a simple factor of 3 boost to the model - predicted luminosities seems implausible the model tracks on the h - r diagrams ( figure [ fig : hrd - both ] ) would then give implausibly old ages ( @xmath9510 gyr ) .
such a luminosity error could be made compatible with the h - r diagram positions if the current effective temperatures for the t dwarfs were also systematically overestimated by @xmath53200 k , but such a large discrepancy seems implausible .
even stronger constraints on the models from the isochrone test will require smaller uncertainties in and . for , numerical tests using the measurements from
[ sec : isochrone-2m1209 ] indicate that a @xmath55% distance and @xmath525 k uncertainty in are needed to measure a 2@xmath60 discrepancy from coevality . in absence of a parallax ,
our analysis is consistent with coevality , but the results on absolute ages in table [ table : evolmodels ] can not be taken as a firm constraint .
the parallax measurement is within current capabilities ( e.g. * ? ? ?
* ) and is being pursued by us using the canada - france - hawaii telescope ( e.g. * ? ? ?
* ) , but the required accuracy is a challenge to atmospheric models . both and have plausibly short orbital periods for dynamical mass determinations , which will significantly strengthen the future tests of the models
this is especially true in the case of bab , given the additional independent constraint from the age estimate of the primary star .
the immediate prospects of finding similarly extreme systems to further test the models in this fashion are limited , given that hundreds of nearby ultracool dwarfs have already been imaged at high angular resolution but yielding only and bab .
as one alternative path , @xcite have pursued the possibility of finding previously unrecognized extreme mass - ratio ultracool binaries by modeling of integrated - light spectra , suitable for special cases where the primary and secondary spectral types are sufficiently discrepant ( though the resolved measurements of the individual components needed to place them on the h - r diagram may be challenging ) . in the near - future , the advent of powerful new wide - field surveys , such as the pan - starrs optical survey @xcite and the wise mid - ir survey satellite @xcite , will boost the prospects for expanding the census of field brown dwarfs and thereby enable discovery of new binaries .
a larger sample of extreme systems will allow the isochrone test to be applied over a wider range of temperatures and luminosities , thus challenging the models over a large span of brown dwarf masses and ages .
we gratefully acknowledge the keck lgs ao team for their exceptional efforts in bringing the lgs ao system to fruition . it is a pleasure to thank jim lyke , hien tran , heather hershley , gary punawai , al conrad , randy campbell , jason mcilroy , and the keck observatory staff for assistance with the observations . this work has benefited from discussions with michael cushing , markus kasper , and jon swift .
we thank david pinfield , nicolas lodieu , phillipe delorme , ben burningham , and steve warren for providing electronic version of published late - t dwarf spectra .
our research has employed the 2mass data products ; nasa s astrophysical data system ; the simbad database operated at cds , strasbourg , france ; the m , l , and t dwarf compendium housed at dwarfarchives.org and maintained by chris gelino , davy kirkpatrick , and adam burgasser @xcite ; the spex prism spectral libraries maintained by adam burgasser at http://www.browndwarfs.org/spexprism ; and the vlm binaries archive maintained by nick siegler at http://www.vlmbinaries.org .
mcl and tjd acknowledge support for this work from nsf grant ast-0507833 and an alfred p. sloan research fellowship .
skl s research is supported by the gemini observatory , which is operated by the association of universities for research in astronomy , inc .
, on behalf of the international gemini partnership of argentina , australia , brazil , canada , chile , the united kingdom , and the united states of america .
finally , 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 .
_ facilities : _ keck ii telescope ( lgs ao , nirc2 )
we improve the polynomial fit for the @xmath12-band bolometric correction as a function of near - ir spectral type from @xcite , by updating the spectral types for the t dwarfs in their sample to types from @xcite ; adding data for the t dwarfs hd 3651b and hn peg b @xcite ; and constructing fits for the @xmath10 and @xmath11-band bolometric corrections .
we use a 6th - order polynomial fit to accurately capture the behavior of the @xmath10 and @xmath11-band data and to avoid artificially steep changes in the polynomials at the earliest and latest spectral types .
the results of the new fits are shown in figure [ fig : bolcorr ] and table [ table : bolcorr ] .
note that the use of the revised spectral types for the t dwarfs leads to 3050% smaller rms about the fit compared to the original golimowski et al
. study .
we also plot the @xmath47 values for two of the latest type objects known , the t8.5 dwarf wolf 940b @xcite and the t9 dwarf ulas 0034@xmath10052 . from the published bolometric luminosities , mko photometry , and distances , we compute their bolometric corrections in @xmath22 , though these are not included in the polynomial fit .
the values are notably different from even the t7t8 objects , likely due to the sharply increasing fraction of emitted at mid - infrared wavelengths in the very latest known objects ( e.g. figure 1 of @xcite ) and perhaps improvements in the methods used to derive for late - t dwarfs using model atmospheres since the @xcite study , which assumed a rayleigh - jeans approximation beyond 5 .
we determined or compiled from the literature the spectral types , distances , and luminosities of all ultracool field binaries with integrated - light spectral types of l4 or later or those with ancilliary evidence for being substellar .
the available literature compilations are based on a broad mix of input data and calculation methods . to improve upon this
, we assembled a more homogenous compilation of binary properties , as described below .
the final results are in table [ table : binaries ] .
1ex * photometry : * for the integrated - light photometry , we preferentially used measurements on the mko photometric system where available and 2mass measurements for the remaining objects . to determine the resolved photometry , we compiled the highest precision @xmath22 flux ratio measurements available from the literature and supplemented them with measurements from public data archives ( _ hst _ , vlt , and gemini ) and our own unpublished keck lgs ao data . in cases where the integrated - light photometry and flux ratios were not on the same photometric system , we used conversions between mko and 2mass from @xcite .
infrared data were not available for 5 binaries originally discovered by _ hst _ using the @xmath96 bandpass ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) . for these cases , we estimated the @xmath12-band flux ratio from the optical results . to do this
, we assumed @xmath97f814w@xmath98 , which is accurate to @xmath99 mag according to @xcite . to estimate the secondary s spectral type , we derived the relation between spectral type and @xmath100 , based on the 11 objects that were used in the liu et al .
( 2006 ) faint relations and have @xmath101-band photometry in @xcite : @xmath102 where @xmath103 means l0 , etc , using optical spectral types for the l dwarfs .
the fit is valid from l0 to t7.5 , and the rms about the fit is 0.25 mag .
then using the adopted primary spectral type ( described below ) and the observed @xmath104 , we derived the secondary spectral type . with spectral types for both components , we use the @xcite spt absolute magnitude relations to estimate @xmath105 for the binary .
we accounted for the additional uncertainty introduced by this process in our analysis , and the affected objects are marked in table [ table : binaries ] as being correspondingly uncertain .
1ex * resolved spectral types and distances : * when available , we adopt the individual component spectral types previously determined from detailed analysis , e.g. , spectral decomposition . for the remaining binaries
, we assumed that the primary spectral types and uncertainties were identical to the integrated - light spectral types .
we adopted optical spectral types determined on the @xcite classification scheme for l dwarfs and infrared types on the @xcite classification scheme for t dwarfs , except in two cases .
one was denis - p j225210.7@xmath1173013ab , which classified as l7.5 based on a comparison of the @xmath11 and @xmath12-band spectra to those of optically typed l dwarfs .
the other was 2mass j09201223@xmath1063517429ab , which has highly discrepant integrated - light spectral types of l6.5 in the optical @xcite and t0p in the infrared .
we adopted the infrared type in this case . to derive the secondary spectral types , we used the average of the `` faint '' and `` bright '' spt absolute magnitude relations of @xcite along with the known or estimated absolute magnitudes for the secondaries .
( the liu et al .
relations were extended to objects of m6l1 using photometry from the @xcite literature compilation . ) for binaries with parallax measurements , we computed the absolute magnitudes of the secondary components directly . for the rest , we estimated photometric distances based on the primary components apparent magnitudes and the same averaged @xcite spt absolute magnitude relations . if distance estimates were available from multiple bandpasses , we preferred the @xmath12-band estimate , then @xmath11-band , and finally @xmath10-band , to minimize the potential non - monotonic behavior in the l / t transition region .
the uncertainties in the resulting distances account for the input photometry uncertainties and the intrinsic scatter in the empirical relations .
using the final assembled ir absolute magnitudes for the secondaries , we estimated spectral types using the method described in section 3.2 of @xcite . for a given bandpass , we computed the spectral type probability distribution for an object based on its absolute magnitude , the uncertainties in both the photometry and the distances , and the intrinsic scatter in the empirical spt - absolute magnitude relations .
we then combined the results from all available bandpasses to form the joint spectral type probability distribution and thus to estimate the final secondary spectral types and uncertainties .
we rounded all spectral types and their uncertainties to the nearest 0.5 subclass .
1ex * bolometric luminosities and luminosity ratios : * we derived individual luminosities for the binary components from each available bandpass ( @xmath22 ) using the resolved photometry , distances , and bolometric corrections .
bolometric corrections were computed based on the polynomial fits in appendix [ sec : bolcorr ] and the spectral types assigned to each component . for computing luminosities , we assume @xmath107 mag .
uncertainties for the resulting s were computed from the uncertainties in the resolved photometry , in the bolometric corrections due to the spectral type errors , and the intrinsic scatter in the bolometric correction relations .
we also computed luminosity ratios for each available bandpass ; this is typically known to higher precision than since the ratios are independent of the distance uncertainty .
the error in the luminosity ratio comes from the quadrature sum of the errors in the flux ratios and the bolometric corrections of the primaries and secondaries . for our final adopted luminosities and luminosity ratios
, we chose the bandpass that gave the smallest error in the luminosity ratio .
1ex * mass ratios : * we used the @xcite evolutionary models to derive the mass ratios for each system , using the derived s , their uncertainties , and an adopted age .
we take care to treat the distance uncertainty as an error common to each binary , as opposed to applying it independently to each individual component . for objects with specific age information in the literature , we use that ; otherwise we assume 1 gyr .
in both cases , we ignore the age uncertainty in calculating the mass ratio , since this is common to ( most of ) the objects .
four binaries have their total mass measured dynamically , and we use their published mass ratio measurements in these cases : 2mass j1534@xmath12952ab @xcite , hd 130948bc @xcite , lp 349@xmath125ab , and gj 569bab @xcite .
to summarize the mass ratio results , we fit the observed values with a power - law distribution.134826.7ab , from this ensemble analysis given its unusually large separation and mass ratio compared to all other field systems .
also , because of its large ratio , its mass ratio estimate is highly dependent on the assumed age , from 0.2 to 0.5 for ages of 1 to 10 gyr . ] since the sample is relatively sparse , instead of the usual process of fitting to a binned histogram , we adopt a maximum likelihood approach .
the probability distribution of the mass ratio @xmath108 is described by @xmath109 for @xmath110 , with the constant chosen so that the total probability is unity .
we define the log of the likelihood function as @xmath111 \\ & = & n \ln ( \alpha+1 ) + \alpha \sum_{i=1}^{n } \ln q_i \end{aligned}\ ] ] where @xmath112 is the number of objects in the sample and @xmath113 is the mass ratio for each object . to find the best fitting @xmath114
, we find the maximum of @xmath115 : @xmath116 which gives the final result @xmath117 .\end{aligned}\ ] ] for our sample , the distribution @xmath118 is well - fit by a gaussian , from which we extract 1@xmath60 confidence limits .
to incorporate the uncertainties in @xmath113 , we use a monte carlo approach , computing the best @xmath114 many times with the @xmath113 values drawn from normal distributions ; the resulting rms dispersion in @xmath114 from the ensemble of results is much smaller than the 68% confidence interval in any realization of @xmath115 . as a simple approximation
, we add the rms ( @xmath119 ) in quadrature to the confidence limits ( @xmath120 ) to obtain our final uncertainty on @xmath114 .
we find @xmath121 for the sample in table [ table : binaries ] .
for such an exponent , binaries with @xmath122 such as and amount to only 9% of the population .
our result is comparable to , though somewhat steeper , than the @xmath123 found by @xcite for a sample of 30 binaries with estimated primary masses of @xmath1240.075 .
there are some differences in the fitting method and the input data ; for the latter , the burgasser et al .
compilation included young binaries in open clusters and star - forming regions ( which tend to have a broader range of mass ratios ) and used the heterogenous set of mass ratios published in the literature .
burgasser et al . also corrected the observed mass ratio distribution to account for the bias towards equal - mass systems in magnitude - limited surveys , based on an approximate conversion between flux ratio and mass ratio ( see * ? ? ?
we choose not to apply any corrections , since the surveys involved in table [ table : binaries ] span a wide range of wavelengths and telescopes . given these differences ,
a more shallow power - law is expected from the burgasser et al . compilation .
similarly , using a full bayesian analysis to account for differing survey sensitivities , @xcite derived @xmath125 ; his sample included many late - m ( stellar ) primaries which likely led to a shallower exponent .
age determinations for the k5v star a are necessarily indirect and rely on empirical correlations between stellar activity , rotation , and age . estimated a mean ( 1.3 gyr ) with a range of 8.9 9.3 ( 0.8
2.0 gyr ) , based on their semi - empirical relation between stellar rotational period and age and the _ estimated _ rotational period of a from @xcite as derived from measurements and an empirical relation between chromospheric activity and rotation .
this age range has been used for analysis of bab in previous work .
additional age estimates come from the correlation between age and chromospheric activity traced by line emission , as measured by the index .
@xcite measured a on three occasions spanning 1.3 yr , with a mean and an rms of @xmath126 dex .
chromospheric activity is known to be variable on decade - long time scales , and @xcite found @xmath127 for roughly eight years later .
the available age calibrations of give consistent estimates .
( 1 ) using the @xcite calibration , we find ages of 1.2 gyr and 3.2 gyr from the @xcite and @xcite data , respectively .
( 2 ) the @xcite calibration results in ages of 1.2 gyr and 3.4 gyr . ( 3 ) note that with a @xmath128 color of 1.06 mag , a is beyond the most recent activity - age calibrations , those of @xcite which are only valid for @xmath129 mag stars . if we disregard this limitation , their calibration gives ages of 0.9 and 4.0 gyr for the two sets of measurements .
( 4 ) we can also use the `` preferred '' method of @xcite , namely converting to a rossby number , in order to estimate a rotation period that can then be used to derive an age from their gyrochronology relation .
the resulting ages are 1.6 gyr and 5.0 gyr for the higher and lower activity levels , respectively . the nominal uncertainty for such chromospheric ages is estimated to be around @xmath130% at a single epoch @xcite .
thus , the overall age range derived from chromospheric indicators is 0.57.0 gyr , with a mean of about 2 gyr .
an alternative estimate comes from @xcite .
they obtained high - contrast imaging of the system to attempt to directly image a close - in substellar companion , whose existence has been inferred from a long - term linear trend in the radial velocity of a . by combining their imaging non - detection combined with the radial velocity data , janson et al .
conclude that the star is likely to be older than than @xmath51 gyr , or else the companion would have been sufficiently self - luminous be have been directly detected . this age estimate may be improved in the future by measuring a s rotation period , as gyrochronology relations can provide more precise age estimates ( 1525% ; * ? ? ?
gyr for a. this age should be considered notional , since it relies on an estimated period , not a direct measurement . ]
based on @xcite , the rotation period of a is expected to be 1450 days given its level of chromospheric activity .
furthermore , because a is a very bright star ( @xmath131 mag ) , it is within reach of asteroseismological measurements with current facilities , which may be able to provide an even more precise age estimate .
t6 ) from @xcite , @xcite , @xcite , , and @xcite ; and individual components of resolved binaries from , @xcite , @xcite , @xcite , and @xcite .
( for the t8.5 dwarf cfbds 0059@xmath10114 , the @xmath12-band measurement is synthesized from the near - ir spectrum and -band photometry of . )
the photometry errors are comparable to or smaller than the size of the plotting symbols for most of the objects .
the numbers indicate the near - ir spectral subclass of the objects , with half subclasses being rounded down ( e.g. , t3.5 is labeled as `` 3 '' ) , and objects of the same subclass plotted in the same color .
the late - l dwarfs ( classified on the @xcite scheme ) are plotted as bare numbers . the t dwarfs ( on the @xcite scheme with extension by @xcite ) are plotted as circumscribed numbers , with squares for integer subclasses ( e.g. , t3 ) and diamonds for half subclasses ( e.g. , t3.5 ) .
[ fig : colorcolor],width=576 ] colors to the two components of .
the agreement between the observed spectrum and the modeled blends is good , meaning that the spectral types inferred for a and b from the near - ir photometry are consistent with the integrated - light spectrum .
[ fig : blends],width=432 ] and described in appendix [ sec : binaries ] .
the mass ratio for from our calculations is shown as the filled color bar , demonstrating the rarity of such an unequal mass ratio among the field population .
( sdss [email protected] is not plotted , given its very unusual properties relative to all other known systems and the large systematic uncertainty in its mass ratio ; see appendix [ sec : binaries ] . )
[ fig : massratio],width=432 ] 0.1 in yr ) .
the iso - mass tracks for 0.01 ( upper line ) and 0.04 ( lower line ) are shown as dashed black lines .
for , two sets of errors are shown for the @xmath132-axis : the longer , thinner grey error bars show the total error in the measurements , and the shorter , thicker black errors bars show the errors with the uncertainty in the distance removed , since this uncertainty is common to both components .
for , the plot uses the values from .
[ fig : hrd - both],title="fig:",width=288 ] 0.5 in yr ) .
the iso - mass tracks for 0.01 ( upper line ) and 0.04 ( lower line ) are shown as dashed black lines .
for , two sets of errors are shown for the @xmath132-axis : the longer , thinner grey error bars show the total error in the measurements , and the shorter , thicker black errors bars show the errors with the uncertainty in the distance removed , since this uncertainty is common to both components . for , the plot uses the values from .
[ fig : hrd - both],title="fig:",width=288 ] 3ex -0.65 in ) and the lyon ( cond ) and tucson evolutionary models .
the histogram shows the results from our monte carlo calculation , and the thin solid line shows the best fitting gaussian to the binned probability distribution .
the dashed vertical line denotes coevality of the two components .
[ fig : hrd - agediff - both],title="fig:",width=336 ] -1.5 in ) and the lyon ( cond ) and tucson evolutionary models .
the histogram shows the results from our monte carlo calculation , and the thin solid line shows the best fitting gaussian to the binned probability distribution .
the dashed vertical line denotes coevality of the two components .
[ fig : hrd - agediff - both],title="fig:",width=336 ] ) , the requirement of coevality ( @xmath133 ) , the mass ratio inferred from the luminosity ratio ( @xmath134 ) , and the total dynamical mass ( 121 ) .
the grey points are the results based on using the @xcite s and the green on the s .
the rightmost point is missing for the models as such a large temperature change moves the binary components off the loci of available models .
[ fig : plot - teff - changes - epsind - trelative],width=412 ] -0.1 in .
see figure [ fig : plot - teff - changes - epsind - trelative ] caption for more details .
* right : * changing the of component bb . [ fig : plot - teff - changes - epsind - both],title="fig:",width=307 ] 0.3 in .
see figure [ fig : plot - teff - changes - epsind - trelative ] caption for more details . *
right : * changing the of component bb . [ fig : plot - teff - changes - epsind - both],title="fig:",width=307 ] 2ex , @xmath11 , and @xmath12-band filters , based on data from @xcite , @xcite , @xcite , @xcite , , and @xcite .
points without visible error bars have uncertainties smaller than the plotting symbol .
the solid line overplots a 6th - order polynomial fit to the datasets for objects as cool as t8 , with coefficients given in table [ table : bolcorr ] .
the dotted line for the @xmath12-band data gives the fit from @xcite .
known and suspected binaries are plotted as open circles and as circumscribed small circles , respectively , and excluded from the fits .
suspected binaries are those flagged by @xcite based on their large overluminosity relative to objects of similar spectral type .
the l dwarfs are classified on the @xcite scheme and the t dwarfs on the @xcite scheme .
the two coolest objects ( t8.5 wolf 940b and t9 ulas 0034@xmath10052 ) are excluded from the fit.[fig : bolcorr],width=432 ] 2007-apr-22 & @xmath10 & 1.46 & 112 @xmath28 3 & 0.022 @xmath28 0.002 & 143 @xmath28 6 & 311 @xmath28 2 & 1.5 @xmath28 0.2 + & @xmath11 & 1.40 & 105 @xmath28 5 & 0.049 @xmath28 0.004 & 151 @xmath28 13 & 314 @xmath28 5 & 2.8 @xmath28 0.3 + & & 1.53 & 92 @xmath28 5 & 0.121 @xmath28 0.019 & 156 @xmath28 7 & 314 @xmath28 4 & 3.2 @xmath28 0.5 + 2008-jan-16 & @xmath13 & 1.22 & 77 @xmath28 5 & 0.083 @xmath28 0.012 & 142 @xmath28 7 & 313 @xmath28 3 & 2.07 @xmath28 0.17 + @xmath10 ( mag ) & 15.80 @xmath28 0.05 & 17.27 @xmath28 0.17 + @xmath13 ( mag ) & 15.20 @xmath28 0.04 & 17.27 @xmath28 0.15 + @xmath11 ( mag ) & 15.32 @xmath28 0.04 & 18.08 @xmath28 0.26 + @xmath135 ( mag ) & 15.12 @xmath28 0.14 & 18.29 @xmath28 0.53 + @xmath12 ( mag ) & 15.22 @xmath28 0.04 & 18.48 @xmath28 0.52 + @xmath136 ( mag ) & 0.48 @xmath28 0.06 & @xmath10.81 @xmath28 0.31 + @xmath137 ( mag ) & 0.58 @xmath28 0.07 & @xmath11.21 @xmath28 0.54 + @xmath138 ( mag ) & @xmath10.12 @xmath28 0.05 & @xmath10.81 @xmath28 0.30 + @xmath139 ( mag ) & 0.10 @xmath28 0.05 & @xmath10.40 @xmath28 0.58 + estimated spectral type & t2.0 @xmath28 0.5 & t7.5 @xmath28 0.5 + @xmath140 ( pc ) & + @xmath141 ( mag ) & 14.19 @xmath28 0.41 & 15.66 @xmath28 0.44 + @xmath142 ( mag ) & 13.59 @xmath28 0.41 & 15.66 @xmath28 0.44 + @xmath31 ( mag ) & 13.71 @xmath28 0.41 & 16.47 @xmath28 0.49 + @xmath143 ( mag ) & 13.51 @xmath28 0.43 & 16.68 @xmath28 0.67 + @xmath32 ( mag ) & 13.61 @xmath28 0.41 & 16.87 @xmath28 0.66 + @xmath144 & + @xmath145 ( from @xmath10-band ) & 16.21 @xmath28 0.45 & 18.24 @xmath28 0.47 + @xmath146 ( from @xmath10-band ) & @xmath14.58 @xmath28 0.18 &
@xmath15.39 @xmath28 0.19 + @xmath147 ( from @xmath10-band ) & + total mass ( @xmath148 ) & + mass ( @xmath148 ) & 36 @xmath28 6 & 17 @xmath28 4 + @xmath149 & + @xmath150 & 1281 @xmath28 162 & 771 @xmath28 103 + @xmath151 & + radius @xmath152 & 0.101 @xmath28 0.002 & 0.109 @xmath28 0.003 + @xmath153 ( cgs ) & 4.96 @xmath28 0.11 & 4.56 @xmath28 0.13 + orbital period ( yr ) & + total mass ( @xmath148 ) & + mass ( @xmath148 ) & 49 @xmath28 9 & 25 @xmath28 5 + @xmath149 & + @xmath154 & 1336 @xmath28 173 & 802 @xmath28 107 + @xmath151 & + radius @xmath152 & 0.092 @xmath28 0.002 & 0.101 @xmath28 0.003 + @xmath153 ( cgs ) & 5.18 @xmath28 0.11 & 4.80 @xmath28 0.12 + orbital period ( yr ) & + total mass ( @xmath148 ) & + mass ( @xmath148 ) & 77 @xmath28 5 & 55 @xmath28 8 + @xmath149 & + @xmath154 & 1418 @xmath28 163 & 881 @xmath28 120 + @xmath155 & + radius @xmath152 & 0.082 @xmath28 0.001 & 0.083 @xmath28 0.003 + @xmath153 ( cgs ) & 5.48 @xmath28 0.04 & 5.31 @xmath28 0.11 + orbital period ( yr ) & + total mass ( @xmath148 ) & + mass ( @xmath148 ) & 79 @xmath28 3 & 70 @xmath28 7 + @xmath149 & + @xmath154 & 1424 @xmath28 159 & 908 @xmath28 119 + @xmath151 & + radius @xmath152 & 0.081 @xmath28 0.002 & 0.078 @xmath28 0.002 + @xmath153 ( cgs ) & 5.50 @xmath28 0.03 & 5.47 @xmath28 0.09 + orbital period ( yr ) & + denis - p j004135.3@xmath1562112ab & m@xmath156 & m@xmath157 & m@xmath158 & @xmath159 & 0.01 ( 0.0050.2 ) & @xmath160 & @xmath161 & @xmath162 & [email protected] & 55 & 16 , 55 & 53 , 55 & 55 & 54 & + lp 349 - 25ab & m@xmath156 & m@xmath163 & m@xmath164 & @xmath165 & @xmath166 & @xmath167 & @xmath168 & @xmath169 & 0.872@xmath170 & 21 & 16 , 19 & 19 , 23 & 22 & 19 & a + gl 569bab & m@xmath171 & m@xmath171 & m@xmath172 & @xmath173 & @xmath174 & @xmath175 & @xmath176 & @xmath177 & 0.866@xmath178 & 46 & 19 , 36 & 28 , 36 & 61 & 19 & a + denis - p j035726.9@xmath1441730ab & l@xmath179 & m@xmath172 & l@xmath180 & @xmath181 & @xmath60.1 & @xmath182 & @xmath183 & @xmath184 & [email protected] & 3 & 4 , 16 & 15 , 47 & & 15 & b + kelu-1ab & l@xmath37 & l@xmath37 & l@xmath185 & @xmath186 & 0.5 ( 0.30.8 ) & @xmath187 & @xmath188 & @xmath189 & [email protected] & 39 & 37 , 39 & 31 & 17 & 39 & + 2mass j11463449@xmath1062230527ab & l@xmath190 & l@xmath190 & l@xmath191 & @xmath192 & 0.5 ( 0.30.8 ) & @xmath193 & @xmath194 & @xmath195 & [email protected] & 50 & 4 , 16 & 31 & 17 & 1 & b , c + sdss [email protected] & l@xmath191 & l@xmath190 & l@xmath196 & @xmath197 & 0.1 ( 0.020.3 ) & @xmath198 & @xmath199 & @xmath200 & [email protected] & 2 & 2 , 34 & 2 , 27 & 2 & 2 & + 2mass j00250365@xmath1064759191ab & l@xmath201 & l@xmath201 & l@xmath202 & @xmath203 & 0.3 ( 0.10.5 ) & @xmath204 & @xmath205 & @xmath206 & [email protected] & 51 & 9 , 16 , 51 & 14 & & 20 & d + hd 130948bc & l@xmath201 & l@xmath201 & l@xmath201 & @xmath207 & @xmath208 & @xmath187 & @xmath209 & @xmath210 & [email protected] & 49 & 18 & 26 & 61 & 18 & a + gl 417bc & l@xmath211 & l@xmath211 & l@xmath212 & @xmath213 & 0.74 ( 0.580.86 ) & @xmath214 & @xmath215 & @xmath216 & [email protected] & 3 & 16 , 42 & 32 & 61 & 2 & + gj 1001bc & l@xmath217 & l@xmath217 & l@xmath196 & @xmath218 & 3 ( 110 ) & @xmath219 & @xmath220 & @xmath221 & [email protected] & 25 & 16 , 25 & 33 & 29 & 38 & e + denis - p j1228.2@xmath11547ab & l@xmath217 & l@xmath217 & l@xmath196 & @xmath222 & & @xmath223 & @xmath224 & @xmath225 & [email protected] & 35 , 45 & 16 , 42 & 31 & 17 & & + 2mass j12392727@xmath1065515371ab & l@xmath217 & l@xmath217 & l@xmath226 & @xmath227 & & @xmath228 & @xmath229 & @xmath230 & [email protected] & 24 & 3 , 16 & 32 & & & b + 2mass j21321145@xmath1061341584ab & l@xmath231 & l@xmath217 & l@xmath34 & @xmath232 & & @xmath233 & @xmath234 & @xmath235 & [email protected] & 58 & 16 , 58 & 14 , 58 & & & + 2mass j17281150@xmath1063948593ab & l@xmath236 & l@xmath236 & l@xmath237 & @xmath238 & & @xmath239 & @xmath240 & @xmath189 & [email protected] & 24 & 16 , 42 & 32 & 62 & & + 2mass j08503593@xmath1061057156ab & l@xmath231 & l@xmath231 & l@xmath241 & @xmath242 & & @xmath243 & @xmath244 & @xmath245 & [email protected] & 50 & 37 , 42 & 31 & 17 & & + 2mass j22551861@xmath15713056ab & l@xmath246 & l@xmath196 & l@xmath247 & @xmath248 & & @xmath249 & @xmath250 & @xmath251 & [email protected] & 53 & 9 , 16 , 51 & 53 & & & + 2mass j21522609@xmath1060937575ab & l@xmath246 & l@xmath246 & l@xmath252 & @xmath253 & & @xmath254 & @xmath255 & @xmath256 & [email protected] & 51 & 16 , 51 & 53 & & & + sdss [email protected] & l@xmath231 & l@xmath231 & t@xmath34 & @xmath257 & & @xmath258 & @xmath259 & @xmath260 & [email protected] & 12 & 12 & 5 , 12 & 57 & & f + denis - p j225210.7@xmath1173013ab & l@xmath163 & l@xmath261 & t@xmath261 & @xmath262 & & @xmath263 & @xmath264 & @xmath265 & [email protected] & 52 & 9 , 16 , 42 , 51 & 11 , 30 & & & g + 2mass j05185995@xmath12828372ab & t@xmath266 & l@xmath34 & t@xmath217 & @xmath267 & & @xmath268 & @xmath269 & @xmath270 & [email protected] & 9 & 9 , 16 & 10 , 11 & & & + sdss [email protected] & t@xmath179 & l@xmath34 & t@xmath37 & @xmath271 & & @xmath272 & @xmath273 & @xmath274 & [email protected] & 7 & 9 , 37 & 10 , 11 & 62 & & + denis - p j020529.0@xmath1115925ab & l@xmath236 & l@xmath236 & l@xmath246 & @xmath275 & & @xmath276 & @xmath277 & @xmath278 & [email protected] & 35 & 16 , 42 & 31 & 17 & & + 2mass j09153413@xmath1060422045ab & l@xmath246 & l@xmath246 & l@xmath279 & @xmath280 & & @xmath281 & @xmath282 & @xmath225 & [email protected] & 51 & 9 , 16 , 42 , 51 & 53 & & & + 2mass j21011544@xmath1061756586ab & l@xmath34 & l@xmath34 & t@xmath283 & @xmath284 & & @xmath285 & @xmath286 & @xmath287 & [email protected] & 3 & 3 , 13 & 32 & 62 & & b + 2mass j03105986@xmath1061648155ab & l@xmath156 & l@xmath156 & l@xmath288 & @xmath289 & & @xmath290 & @xmath291 & @xmath292 & [email protected] & 59 & 37 , 42 & 32 & & & + gl 337cd & t@xmath179 & t@xmath179 & t@xmath293 & @xmath294 & 1.5 ( 0.63.4 ) & @xmath295 & @xmath296 & @xmath297 & [email protected] & 8 & 16 , 42 & 10 , 63 & 61 & 63 & + 2mass j09201223@xmath1063517429ab & t@xmath179 & t@xmath179 & t@xmath298 & @xmath299 & & @xmath300 & @xmath301 & @xmath302 & [email protected] & 50 & 16 , 42 & 10 , 32 & & & h + sdss [email protected] & t@xmath303 & t@xmath303 & t@xmath304 & @xmath305 & & @xmath306 & @xmath307 & @xmath308 & [email protected] & 42 & 13 , 42 & 13 & & & + sdss [email protected] & t@xmath309 & t@xmath309 & t@xmath310 & @xmath311 & & @xmath312 & @xmath313 & @xmath216 & [email protected] & 42 & 13 , 42 & 13 & & & + 2mass j14044948@xmath13159330ab & t@xmath314 & t@xmath179 & t@xmath217 & @xmath299 & & @xmath315 & @xmath316 & @xmath317 & [email protected] & 44 & 44 & 11 , 43 & & & + sdss [email protected] & t@xmath190 & t@xmath266 & t@xmath318 & @xmath319 & & @xmath320 & @xmath321 & @xmath322 & [email protected] & 9 & 9 , 37 , 42 & 10 , 11 & 60 & & + @xmath41 ind bab & t@xmath314 & t@xmath266 & t@xmath231 & @xmath323 & 2 ( 0.57.0 ) & @xmath324 & @xmath325 & @xmath326 & [email protected] & 48 & 16 , 48 & 10 , 48 , 56 & 61 & 1 & + sdss [email protected] & t@xmath185 & t@xmath180 & t@xmath217 & @xmath327 & & @xmath328 & @xmath329 & @xmath330 & [email protected] & 40 & 13 , 40 & 40 & & & + 2mass j12095613@xmath11004008ab & t@xmath190 & t@xmath37 & t@xmath34 & @xmath40 & & @xmath331 & @xmath332 & @xmath333 & [email protected] & 1 & 1 , 13 & 1 , 10 & & & + sdss [email protected] & t@xmath211 & t@xmath211 & t@xmath334 & @xmath305 & & @xmath335 & @xmath336 & @xmath337 & [email protected] & 9 & 9 , 37 & 10 & & & + 2mass j15344984@xmath12952274ab & t@xmath217 & t@xmath217 & t@xmath318 & @xmath338 & @xmath339 & @xmath340 & @xmath341 & @xmath342 & [email protected] & 6 & 34 , 41 & 10 , 41 & 60 & 41 & a + 2mass j12255432@xmath12739476ab & t@xmath231 & t@xmath231 & t@xmath34 & @xmath343 & & @xmath344 & @xmath345 & @xmath346 & [email protected] & 6 & 37 , 42 & 10 & 60 & & + 2mass j15530228@xmath1061532369ab & t@xmath236 & t@xmath236 & t@xmath34 & @xmath347 & & @xmath348 & @xmath349 & @xmath216 & [email protected] & 9 & 9 , 34 , 42 & 10 & & & + @xmath56 k , @xmath350 k ( this work ) & lyon / cond & 7.5@xmath351 & 8.5@xmath352 & @xmath10.8@xmath353 & 10@xmath354 & 18@xmath355 & 34@xmath356 & 1.7@xmath357 + & tucson & 7.6@xmath351 & 8.8@xmath358 & @xmath11.0@xmath359 & 11@xmath360 & 20@xmath361 & 36@xmath362 & 1.6@xmath363 + & & & & & & & & + + @xmath364 k , @xmath365 k ( kasper et al . 2009 ) & lyon / cond & [email protected] & [email protected] & [email protected] & 44@xmath366 & 27@xmath367 & 72@xmath368 & [email protected] + & tucson & [email protected] & [email protected] & [email protected] & 49@xmath288 & 28@xmath369 & 78@xmath370 & [email protected] + @xmath371 k , @xmath372 k ( king et al . 2010 ) & lyon / cond & [email protected] & [email protected] & 0.5@xmath373 & 57@xmath287 & 30@xmath288 & 86@xmath370 & 0.52@xmath374 + & tucson & [email protected] & [email protected] & 0.3@xmath375 & 63@xmath369 & 31@xmath376 & 95@xmath377 & 0.49@xmath378 + @xmath379 & 1.890448e+01 & @xmath18.053993e+00 & 1.491738e+00 & @xmath11.367605e@xmath101 & 6.540717e@xmath103 & @xmath11.558986e@xmath104 & 1.462266e@xmath106 & 0.14 + @xmath380 & 1.366709e+01 & @xmath15.426683e+00 & 1.027185e+00 & @xmath19.618535e@xmath102 & 4.733874e@xmath103 & @xmath11.171595e@xmath104 & 1.148133e@xmath106 & 0.07 + @xmath70 & 5.795845e+00 & @xmath11.471358e+00 & 2.868014e@xmath101 & @xmath12.647474e@xmath102 & 1.282657e@xmath103 & @xmath13.177629e@xmath105 & 3.159780e@xmath107 & 0.08 + | highly unequal - mass ratio binaries are rare among field brown dwarfs , with the mass ratio distribution of the known census described by @xmath0 .
however , such systems enable a unique test of the joint accuracy of evolutionary and atmospheric models , under the constraint of coevality for the individual components ( the `` isochrone test '' ) .
we carry out this test using two of the most extreme field substellar binaries currently known , the t1+t6 binary and a newly discovered 0.14 t2.0+t7.5 binary , 2mass j12095613@xmath11004008ab , identified with keck laser guide star adaptive optics .
the latter is the most extreme tight binary resolved to date ( @xmath2 ) .
based on the locations of the binary components on the h - r diagram , current models successfully indicate that these two systems are coeval , with internal age differences of @xmath3 ) dex and @xmath4 dex for and , respectively , as inferred from the lyon ( tucson ) models . however , the total mass of derived from the h - r diagram ( @xmath580 using the lyon models ) is strongly discrepant with the reported dynamical mass .
this problem , which is independent of the assumed age of the system , can be explained by a @xmath550100 k systematic error in the model atmosphere fitting , indicating slightly warmer temperatures for both components ; bringing the mass determinations from the h - r diagram and the visual orbit into consistency leads to an inferred age of @xmath56 gyr for bab , older than previously assumed .
overall , the two t dwarf binaries studied here , along with recent results from t dwarfs in age and mass benchmark systems , yield evidence for small ( @xmath5100 k ) errors in the evolutionary models and/or model atmospheres , but not significantly larger .
future parallax , resolved spectroscopy , and dynamical mass measurements for will enable a more stringent application of the isochrone test .
finally , the binary nature of this object reduces its utility as the primary t3 near - ir spectral typing standard ; we suggest sdss j1206 + 2813 as a replacement . |
in this appendix we outline the `` naive '' perturbative expansion of the loop with the gaussian contractions . we stress that since magnetic fields at @xmath69 are screened over distances @xmath70 @xcite , that this naive expansion can not be applied in the regime of interest in the present paper . to determine @xmath71 we need to determine the `` correction '' in @xmath23 from a pure gauge . from the baker - campbell - hausdorff relation , [ eq : exp_com ] w_m(a^i ) ( - [ x_1,x_2 ] ) , where terms of third order in the fields have been dropped from the exponent ; and @xmath72 is @xmath73 integrated along the loop
. we can express the square of the exponent on the r.h.s . of ( [ eq : exp_com ] ) as [ eq : h2 ] h^2 f^abc f^dec
x_1^a x_1^d x_2^b x_2^e , so that for two colors w_m(a^i ) _ _ 1,_2 1 - h^2 .
( [ eq : h2 ] ) contains four integrations over the periphery of the loop , two of which will be removed by the @xmath74 and @xmath75 contractions ; c.f . eq .
( [ eq : s2 ] ) .
hence @xmath76 is proportional to the square of the area of the loop .
[ fig : mfl_tau=0 ] shows that by matching to the lattice data at @xmath77 we estimate @xmath78 .
a. h. mueller , nucl .
b * 558 * , 285 ( 1999 ) .
l. d. mclerran and r. venugopalan , phys .
* d * * 49 * , 2233 ( 1994 ) , phys .
rev . * d * * 49 * , 3352 ( 1994 ) ; y. v. kovchegov , phys .
* d * * 54 * , 5463 ( 1996 ) .
j. jalilian - marian , a. kovner , l. d. mclerran and h. weigert , phys .
d * 55 * , 5414 ( 1997 ) .
a. kovner , l. d. mclerran and h. weigert , phys .
d * 52 * , 6231 ( 1995 ) ; phys .
d * 52 * , 3809 ( 1995 ) .
r. j. fries , j. i. kapusta and y. li , nucl - th/0604054 . t. lappi and l. mclerran , nucl .
a * 772 * , 200 ( 2006 ) .
j. d. stack , w. w. tucker and a. hart , hep - lat/0011057 .
a. krasnitz , y. nara and r. venugopalan , phys .
lett . * 87 * , 192302 ( 2001 ) ; t. lappi , phys .
c * 67 * , 054903 ( 2003 ) ; eur .
j. c * 55 * , 285 ( 2008 ) .
a. dumitru , h. fujii and y. nara , arxiv:1305.2780 [ hep - ph ] .
a. dumitru and e. petreska , nucl .
a * 879 * , 59 ( 2012 ) ; a. dumitru , j. jalilian - marian and e. petreska , phys .
d * 84 * , 014018 ( 2011 ) .
j. berges and s. schlichting , arxiv:1209.0817 [ hep - ph ] .
see , for example , j. preskill , `` lecture notes on quantum field theory '' , + ` http://www.theory.caltech.edu/~preskill/notes.html ` t. appelquist and r. d. pisarski , phys .
d * 23 * , 2305 ( 1981 ) ; e. manousakis and j. polonyi , phys . rev .
lett . * 58 * , 847 ( 1987 ) ; g. s. bali , j. fingberg , u. m. heller , f. karsch and k. schilling , phys .
lett . * 71 * , 3059 ( 1993 ) ; j. alanen , k. kajantie and v. suur - uski , phys .
d * 80 * , 075017 ( 2009 ) .
j. berges , s. scheffler and d. sexty , phys .
d * 77 * , 034504 ( 2008 ) .
k. fukushima , phys .
d * 77 * , 074005 ( 2008 ) . | we consider the expectation value of a chromo - magnetic flux loop in the immediate forward light cone of collisions of heavy nuclei at high energies .
such collisions are characterized by a non - linear scale @xmath0 where color fields become strong .
we find that loops of area greater than @xmath1 exhibit area law behavior , which determines the scale of elementary flux excitations ( `` vortices '' ) .
we also estimate the magnetic string tension , @xmath2 . by the time
@xmath3 even small loops satisfy area law scaling .
we describe corrections to the propagator of semi - hard particles at very early times in the background of fluctuating magnetic fields .
collisions of heavy ions at high energies provide opportunity to study non - linear dynamics of strong qcd color fields @xcite .
the field of a very dense system of color charges at rapidities far from the source is determined by the classical yang - mills equations with a recoilless current along the light cone @xcite .
it consists of gluons characterized by a transverse momentum @xmath4 on the order of the density of valence charges per unit transverse area @xmath5
@xcite ; this _ saturation momentum _ scale separates the regime of non - linear color field interactions at @xmath6 or distances @xmath7 from the perturbative regime at @xmath8 . near the center of a large nucleus
this scale is expected to exceed @xmath9 gev at bnl - rhic or cern - lhc collider energies , for a probe in the adjoint representation of the color gauge group .
the classical field solution provides the leading contribution to an expansion in terms of the coupling and of the inverse saturation momentum .
the soft field produced in a collision of two nuclei is then a solution of the yang - mills equations satisfying appropriate matching conditions on the light cone @xcite .
most interestingly , right after the impact strong longitudinal chromo - magnetic fields @xmath10 develop due to the fact that the individual projectile and target fields do not commute @xcite .
they fluctuate according to the random local color charge densities of the valence sources . in this letter
we show that magnetic loops @xmath11 exhibit area law behavior , and we compute the magnetic string tension .
furthermore , we argue that at length scales @xmath12 the field configurations might be viewed as uncorrelated z(n ) vortices . at finite times @xmath13 after the collision area law behavior is observed even for rather small wilson loops .
finally , we sketch how the background of magnetic fields affects propagation of semi - hard particles with transverse momenta somewhat above @xmath0 .
consider a spatial wilson loop with radius @xmath14 in the plane transverse to the beams , m(r ) & = & p ( ig _ -^d a^i ) + w_m(r ) & = & , [ eq : w_m ] where @xmath15 , and path ordering is with respect to the angle @xmath16 ; in numerical lattice simulations it is more convenient to employ a square loop .
we compare , also , to the expectation value of the z(@xmath17 ) part of the loop ; for a magnetic field configuration corresponding simply to a superposition of independent vortices the loop should equal @xmath18 , with @xmath19 the total vortex charge piercing the loop .
thus , for two colors we compute [ eq : w_m_z2 ] w_m^z(2)(r ) = where @xmath20 denotes the sign function . comparing ( [ eq : w_m ] ) to ( [ eq : w_m_z2 ] ) tests the interpretation that the drop - off of @xmath21 is due to z(@xmath17 ) vortices , without requiring gauge fixing of the su(@xmath17 ) links @xcite . the field in the forward light cone immediately after a collision @xcite , at proper time @xmath22 ,
is given in light cone gauge by @xmath23 . in turn , before the collision the individual fields of projectile and target are 2d pure gauges , [ eq : alphai ] ^i_m = u_m ^i u_m^ , ^i ^i_m = g _ m , where @xmath24 labels projectile and target , respectively , and @xmath25 are su(n ) gauge fields . eqs .
( [ eq : alphai ] ) can be solved either analytically in an expansion in the field strength @xcite or numerically on a lattice @xcite .
the large-@xmath26 valence charge density @xmath27 is a random variable this variable denotes the light - cone momentum of a parton relative to the beam hadron and should not be confused with a transverse coordinate . ] . for a large nucleus , the effective action describing color charge fluctuations is quadratic , [ eq : s2 ] s_eff[^a ] = , ^a(*x * ) ^b ( * y*)=^2 ^ab ( * x*-*y * ) , with @xmath28 proportional to the thickness of a given nucleus @xcite . the variance of color charge fluctuations determines the saturation scale @xmath29 @xcite .
the coarse - grained effective action ( [ eq : s2 ] ) applies to ( transverse ) area elements containing a large number of large-@xmath26 `` valence '' charges , @xmath30 .
the densities @xmath31 at two different points are independent so that their correlation length within the effective theory is zero . however , this is not so for the gauge fields @xmath32 which do exhibit a finite screening length @xcite . in fig .
[ fig : mfl_tau=0 ] we show numerical results for @xmath11 immediately after a collision .
it exhibits area law behavior for loops larger than @xmath33 .
the corresponding `` magnetic string tension '' is @xmath34 .
the area law indicates uncorrelated magnetic flux fluctuations through the wilson loop and that the area of magnetic vortices is rather small , their radius being on the order of @xmath35 . we do not observe a breakdown of the area law up to @xmath36 , implying that vortex correlations are small at such distance scales
also , restricting to the z(2 ) part reduces the magnetic flux through small loops but @xmath37 is comparable to the full su(2 ) result , if somewhat smaller .
the numerically small vortex size that we find is _ parametrically _ consistent with the classical gaussian approximation at weak coupling which , as already mentioned above , applies for areas @xmath38 .
corrections to @xmath39 of higher order in @xmath27 @xcite as well as due to quantum fluctuations @xcite of the fields should be investigated in the future . since the field @xmath40 of a single nucleus is a pure gauge it follows that @xmath41 . however , beyond linear order the field in the forward light cone , @xmath42 is not a pure gauge and so @xmath43 . as discussed in the appendix , for small loops the correction w_m(_1^i+_2^i ) - 1 ~- a^2 [ eq : w_m_pert ] is proportional to the square of the area of the loop .
there is no contribution at order @xmath44 , hence no area law behavior and no vortices , even though there is , of course , a non - zero longitudinal magnetic field even at the `` naive '' perturbative level : [ eq : bzcorr ] b_z(*r * ) b_z(*r*)= 4 q_s1 ^ 2 q_s2 ^ 2 ^2 .
eq .
( [ eq : w_m_pert ] ) applies for small @xmath45 while the non - perturbative lattice result exhibits area law behavior at @xmath46 for @xmath47 .
it indicates the presence of resummed screening corrections for magnetic fields @xcite . to see this more
explicitly it is useful to notice that @xmath48 is in fact @xmath49 , proportional to the product of single powers of the respective saturation scales of projectile and target .
we have verified this numerically in fig .
[ fig : mfl_asym ] .
naive perturbation theory can only produce even powers of the two - point function @xmath50 . to estimate the density of vortices one can consider a simple combinatorial model whereby the area @xmath51 of the loop is covered by patches of size @xmath52 containing a z(2 ) vortex with probability @xmath53 .
averaging over random , uncorrelated vortex fluctuations leads to @xcite w_m(a ) ~ ( - p(1-p ) a q_s^2 ) , or @xmath54 . from this relation
we estimate that the probability of finding a vortex within an area @xmath52 is @xmath55 . in fig .
[ fig : mfl_tau>0 ] we show the time evolution of the magnetic flux loop after a collision .
the magnetic field strength decreases due to longitudinal expansion and so @xmath11 approaches unity . on the other hand ,
the onset of area law behavior is pushed to smaller loops , implying that the size of elementary flux excitations or `` vortices '' decreases ; by the time @xmath56 area law behavior is satisfied even for rather small loops .
since long wavelength magnetic fields remain even at times @xmath13 , it will be important in the future to understand the transition of @xmath11 to behavior expected in _
thermal _ qcd where @xmath57 @xcite . in the context of late - time behavior
much beyond @xmath3 we refer to ref .
@xcite where area law scaling of spatial loops has been observed for classical field configurations emerging from unstable plasma evolution .
we have also investigated the dependence of the magnetic flux loop in the adjoint representation on its area , [ eq : adjloop ] w_m^adj = , and found behavior similar to fig . [
fig : mfl_tau>0 ] .
the adjoint magnetic string tension is about two times larger , as expected from ( [ eq : adjloop ] ) .
the third color component of the longitudinal magnetic field is shown in fig .
[ fig : bz ] , using a random residual gauge for @xmath32 .
domain - like structures where the magnetic field is either positive or negative are clearly visible ; they lead to the above - mentioned area law of the wilson loop .
also , one can see that in time the magnetic fields become weaker and smoother .
thus far we have not addressed the _ longitudinal _ structure of the initial fields .
our solution of eqs .
( [ eq : alphai ] ) is boost invariant and so , naively , the two - dimensional vortex structures mentioned above would form boost invariant strings . however , this simple picture could be modified by longitudinal smearing of the valence charge distributions @xcite and therefore requires more detailed consideration .
the magnetic fields modify the propagation of semi - hard modes with @xmath4 not too far above @xmath0 .
quantum mechanically , the transition amplitude from a state @xmath58 to @xmath59 is given by a feynman sum over paths , & & _
0^ds x^ ~ + & & _
0^ds x^ ( i _ 0^s d m^2 ) ( -_ma ) , [ eq : pathintarea ] where @xmath60 is a parametrization of the path with the given boundary conditions and length @xmath61 ; and @xmath62 . here
, the area @xmath51 is that enclosed by a quantum mechanical path from the initial to the final point returning to @xmath63 via the classical path ; see fig . [
fig : paths ] .
the classical path is obtained by extremizing the action but a single path is a set of measure zero .
semi - classical paths can dominate the integral only if there is constructive interference among neighboring paths from within a de broglie distance . on the other hand , destructive interference of such paths
leads to anderson localization of the wave function . hence , up to a numerical factor , the area in eq .
( [ eq : pathintarea ] ) should be given by @xmath64 .
integrating over the schwinger parameter then leads to the propagator , where @xmath65 from above and @xmath66 is the mass ( time - like virtuality ) of the particle .
this expression accounts for corrections to free propagation and could be useful for studies of the dynamics of the very early stage of a heavy - ion collision .
we obtain a rather interesting picture of the very early stage of ultrarelativistic heavy - ion collisions .
magnetic wilson loops of area @xmath67 effectively exhibit area law behavior which implies uncorrelated magnetic z(n ) vortex - like flux beyond the scale @xmath68 .
we do expect that corrections to this picture appear at much larger distance scales and we intend to study these in detail in the future .
the vortex structure of the longitudinal magnetic field modifies propagation of particle - like modes with de broglie wavelength somewhat larger than @xmath0 .
we thank a. kovner , l. mclerran , p. orland and r. pisarski for helpful comments .
a.d . and e.p .
gratefully acknowledge support by the doe office of nuclear physics through grant no .
de - fg02 - 09er41620 ; and from the city university of new york through the psc - cuny research grant 66514 - 00 44 . |
according to einstein s theory of general relativity , a generic gravitational field has , at a time , three main effects on light : the shapiro time delay , the bending of rays and the curvature of wavefronts .
the gravitational field acts respectively as a retarder delaying a light wavefront , as a prism tilting a light wavefront , and as a lens curving a light wavefront . in special relativity ,
time delays can be also induced by inertial motion of the observer or the source as in the doppler effect and uniform motion of a observer can result in aberration , i.e. , an apparent change in the position of the source .
hence , the curvature of a initially plane light wavefront by a gravity field is a purely general relativistic effect that has no special relativistic analogue . in order to obtain an experimental measurement of the curvature of a wave front ,
samuel @xcite proposed a method based on the relation between the differences of arrival time recorded at four points on the earth and the volume of a parallelepiped determined by four points in the curved wavefront surface .
we will establish a discretized model of the wavefront surface by means of a regular triangulation for the study of the curvature(s ) of this surface .
the main methods and results have been recently published in @xcite , in this work we will only expose a summary of them .
let us consider a spacetime @xmath0 corresponding to a weak gravitational field determined by a metric tensor given in a global coordinate system @xmath1 by @xmath2 where the coordinate components of the metric deviation @xmath3 are : @xmath4 here @xmath5 represents the gravitational constant of the sun , located at @xmath6 , and @xmath7 represents the vacuum light speed .
the null geodesics , @xmath8 satisfy the following equations ( see @xcite ) : @xmath9\dot{z}^k\nonumber\\ & & -(h_{44,k}\delta^a_l+h_{ak , l}-{{\textstyle\frac{1}{2}}}h_{kl , a})\dot{z}^k\dot{z}^l \nonumber\\ & & -(c^{-1}h_{4k , j}-{{\textstyle\frac{1}{2}}}c^{-2}h_{jk , t})\dot{z}^j\dot{z}^k\dot{z}^a,\label{pn1a}\\[.1 in ] 0 & = & g_{\alpha\beta}\dot{z}^\alpha\dot{z}^\beta.\label{pn1b}\end{aligned}\ ] ] where , in the first equation , the first and third terms in the right are of order @xmath10 , while the remaining terms are @xmath11 .
the second equation is the isotropy constraint satisfied by the null geodesics .
we consider a flat initial surface @xmath12 far from the sun formed by points @xmath13 ( with @xmath14 ) in an asymptotically cartesian coordinate system @xmath15 . for the discretization of @xmath12 a triangulation
is constructed in such a form that each vertex is represented by a complex number of the set : @xmath16 the complex plane and the plane @xmath12 may be identified by means of the mapping @xmath17 .
thus a regular triangulation @xmath18 of the initial wavefront @xmath12 is determined . at each vertex in @xmath18 a photon @xmath19 with velocity @xmath20 is located .
the null geodesics equation may be written as a first order differential system @xmath21 , in phase space @xmath22 , which determines in @xmath23 ( the quotient space of @xmath24 by the global timelike vector field @xmath25 associated to the global coordinate system used in the post - newtonian formalism ) a flow : @xmath26 . for each time @xmath27 the image of @xmath12 under the flow @xmath28 determines a curved wavefront @xmath29 .
the initial triangulation by @xmath18 induces a triangulation on the final wavefront @xmath29 , whose vertices we enumerate using the same labels used for the corresponding vertices in @xmath12 .
+ let @xmath30 be a _ normal coordinate system _ with pole at the point @xmath31 and associated normal reference frame @xmath32 . under the coordinate transformation @xmath33 , from post - newtonian to
normal coordinates , the metric tensor @xmath34 on the space @xmath23 is determined ( up to terms of first order in @xmath35 ) from @xmath36 by @xmath37 to compute differential magnitudes of the wavefront surface corresponding to the mesh @xmath18 at each inner vertex @xmath38 we consider the 1ring @xmath39_{j=1,\dots , j , k=0,\dots,5}$ ] formed by the six vertices @xmath40 closest to @xmath41 : for each 1ring @xmath42 $ ] one obtains on the mesh @xmath18 the image 1ring @xmath43 $ ] under the flow @xmath44 . in a neighbourhood of the image point @xmath45 the wavefront can be approximated by a least - squares fitting of the data @xmath43 $ ] as the quadric @xmath46 : @xmath47 using adapted normal coordinates @xmath48 .
using coordinates @xmath49 adapted to the quadric @xmath46 : the metric @xmath50 on @xmath23 induces a metric on @xmath51 of the form @xmath52 by means of a generalized gauss formula , the difference of sectional curvatures @xmath53 and @xmath54 associated with the plane generated by the tangent vectors @xmath55 , in @xmath56 and @xmath23 respectively , is the relative sectional curvature @xmath57 ( see @xcite ) : @xmath58 whereas the mean curvature @xmath59 is determined by half the trace of the second fundamental form @xmath60 : @xmath61
the ray tracing has been carried out using an integrator based on the classic taylor series method for ordinary differential equations ( see @xcite ) .
this integrator presents the following advantages : allows the control of both the order and the step size employed in the method . in this integrator
one may use extended arithmetic precision for the highly accurate computation required in this problem .
( precision of @xmath62 binary digits and a tolerance @xmath63 are used to solve this problem . ) at each step the null constraint equation is preserved .
+ hereafter , we consider the simplest gravitational model generated by a static sun , considered as a point . however , it can be applied to more complex and realistic gravitational fields of the solar system , as considered in @xcite .
we now apply the ray tracing method to a light wavefront region propagating along a tubular neighborhood around the @xmath64-axis and where the inner and outer hexagons have radii of lengths @xmath65 and @xmath66 , respectively .
the initial flat wave surface @xmath12 , perpendicular to the @xmath64-axis , is positioned at -100 a.u .. the wavefront @xmath67 is then determined after a trip of 101 a.u . ,
i.e. , at the position of the earth . remember that the focal length for a light ray grazing an opaque sun is located at 548.30 a.u . from the sun .
+ in figure [ figure ] , the surface @xmath67 at the time when the wavefront arrives at the earth is shown using a gray - scale to represent the relative sectional curvature ( note we have used a different scale on the @xmath64axis ) .
one sees in this figure that the absolute value of the relative sectional curvature function defined on @xmath67 increases as the distance between the photon and the @xmath64axis decreases .
suppose four receiving stations located at four points on an earth hemisphere .
the arrival time differences between these stations depend on the curvature of the wavefront . assumed known the measurements of arrival times corresponding to four points @xmath68 on the wavefront it is possible to determine an approximation of the wavefront curvature in a region far enough from the sun ( say the earth ) , without resorting to the ray tracing method .
an estimation of the gaussian curvature of the wavefront surface can be obtained using the notion of the wald curvature of a metric space established in distance geometry ( see @xcite ) that in the case of 2-dimensional manifolds agrees with the gaussian curvature .
the wald curvature is determined as the limit of the embedding curvatures of metric quadruples isometrically embedded in surfaces of constant curvature ( the euclidean plane @xmath69 , the 2sphere @xmath70 or the hyperbolic space @xmath71 ) .
+ in the hyperbolic plane @xmath72 of curvature @xmath73 , represented by the blumenthal model ( @xcite ) we consider the metric quadruple associated with the points @xmath74 assuming that the geometry of the 3space in the vicinity of the earth is euclidean .
the wald curvature associated to the chosen quadruple @xmath74 prove to be @xmath75 this result gives an approximation of the total curvature of the wavefront surface under the assumption that locally this surface may be identified with a hyperbolic plane in which the quadruple considered is isometrically embedded .
9 samuel j 2004 _ class .
quantum grav .
_ , * 21 * , l83l88 .
san miguel a , vicente f and pascual - snchez j f 2009 _ class .
quantum grav .
_ , * 26 * , 235004 .
brumberg v a 1991 _ essential relativistic celestial mechanics _ ,
( bristol : adam hilger ) .
do carmo m p 1992 _ riemannian geometry _ ( boston : birkhuser ) .
jorba and zou m 2005 _ experimental mathematics _ * 14 * , 99 - 117 .
klioner s a and peip m 2003 _ astron .
astrophys . _ * 410 * , 10631074 .
blumenthal l m 1970 _ theory and applications of distance geometry_. ( new york : chelsea publishing company , 2nd edition ) . | the geometry of a light wavefront , evolving from a initial flat wavefront in the 3-space associated with a post - newtonian relativistic spacetime , is studied numerically by means of the ray tracing method . for a discretization of the bidimensional light wavefront
, a surface fitting technique is used to determine the curvature of this surface .
the relationship between the intrinsic curvature of the wavefront and the change of the arrival time at different points on the earth is also numerically discussed . |
particle magnetic moments provide an important and valuable test of qed and of the standard model ( sm ) . in the case of the muon and the electron magnetic moments , both the experimental measurements and the sm predictions
are very precisely known . however , due to its larger mass , the muon magnetic moment is considered more sensitive to massive virtual particles and hence to new physics ( np ) . in the sm
, the total contributions to the muon @xmath4 ( @xmath5 ) can be divided into three parts : the qed , the electroweak ( ew ) and the hadronic contributions . while the qed @xcite and ew @xcite contributions are well understood , the main theoretical uncertainties lies with the hadronic part which are difficult to control @xcite .
the hadronic loop contributions can not be calculated from first principles , so that one relies on a dispersion relation approach @xcite . at present
the available @xmath6 data are used to calculate the leading - order ( lo ) and higher - order vacuum polarization contributions to @xmath5 ; the estimated contributions are given by @xcite a^had_lo = 6955 ( 40)(7)10 ^ -11 , a^had , disp_nlo = -98 ( 1)10 ^ -11 . on the other hand ,
the hadronic light - by - light contribution can not be calculated from data , hence , its evaluation relies on specific models .
the latest determination of this term is @xcite a^had_lbl = 116(39)10 ^ -11 . including all these corrections , the complete sm prediction is given by a_^sm = 116591834(2)(41)(26)10 ^ -11 [ amusm ] , whereas the current experimentally measured value is @xcite a_^exp = 116592089(54)(33)10 ^ -11 [ amuexp ] .
the sm prediction , therefore , differs from the the experimentally measured value by ( see also @xcite ) @xmath7 which allows some room for new physics . for the purpose of this work we are going to assume that the @xmath8 discrepancy in eq .
[ amm ] is due to np , although we are aware that the estimates of the hadronic contributions have appreciable uncertainties that may provide part of the discrepancy . in most extensions of the sm , new charged or neutral states@xmath9}$ ] , can contribute to the muon anomalous magnetic moment ( @xmath10amm ) at the one - loop ( lowest ) level .
for example , the @xmath10amm plays an important role in constraining the supersymmetric ( susy ) parameter space , where , as in the sm , the leading susy contribution to @xmath11 arises at one - loop , and is found to be enhanced for large @xmath12 . in particular , as was shown in @xcite , susy can address the observed muon @xmath4 discrepancy for @xmath13 and @xmath14 ( higgsino mass parameter ) , with typical susy masses , of the particles involved in the loops , in the range @xmath15 .
model independent analysis show that ( for details see @xcite ) , for small enough couplings , scalar exchange diagrams could account for the observed @xmath10amm with a scalar mass in the range @xmath16 , whereas pseudoscalar and axial - vector one - loop exchanges contribute with the wrong sign and the one - loop vector exchange contributions are too small . in this paper we will consider the @xmath10amm in a new 2hdm framework with a heavy 4th generation family .
indeed , we will show that the @xmath17 access ( with respect to the sm prediction ) shown in eq .
[ amm ] can be explained by one - loop exchanges of the heavy 4th generation neutrino ( @xmath18 ) in a model with two higgs doublets that we have constructed in @xcite and named the 4g2hdm .
these new class of two higgs doublet models were proposed in @xcite as viable low energy effective frameworks for models of 4th generation condensation . in particular ,
a theory with new heavy fermionic states is inevitably cutoff at the near by tev - scale , where one thus expects some form of strong dynamics and/or compositeness to occur .
thus , as was noted already 20 years ago @xcite , the low - energy ( i.e. , sub - tev ) dynamics of such a scenario may be more naturally embedded in multi - higgs theories , where the new composite scalars are viewed as manifestations of the several possible bound states of the fundamental heavy fermions . besides , our 4g2hdm can naturally ( albeit effectively ) accommodate the large ( ew - scale ) mass of the heavy 4th generation neutrino , which otherwise remains a cause of concern in theories with a 4th family of fermions .
we recall that an additional fourth generation of fermions can not be ruled out by any symmetry argument , and is not excluded by ew precision data @xcite .
it is also interesting to note , that already the simplest 4th generation extension of the sm , the so called sm4 , has the potential to address some of the current open questions in particle physics , such as the observed baryon asymmetry @xcite , the higgs naturalness problem @xcite , the fermion mass hierarchy problem @xcite etc .... @xmath19}$ ] the sm4 can also accommodate the emerging possible hints for new flavor physics @xcite .
however , the sm4 as such can not explain the observed muon g-2 discrepancy , see e.g. , @xcite .
in fact , even standard " 2hdm frameworks ( like the type ii 2hdm that underlies the minimal susy model ) with an additional 4th generation of heavy fermions , was shown to fail in explaining the measured @xmath10amm @xcite . in section [ sec2 ]
we calculate the @xmath10amm in the 4g2hdm framework . in sections
[ sec3 ] and [ sec4 ] we consider the constraints on @xmath10amm from the lepton flavor violating ( lfv ) decays @xmath1 and @xmath2 and from @xmath3 , respectively , and in section [ sec5 ] we summarize our results .
at the tree level the muon magnetic moment is predicted by the dirac equation to be @xmath20 with @xmath21 .
the effective vertex of a photon with a charged fermion can in general be written as @xmath22 u(p ) \
, , \ ] ] where , to lowest order , @xmath23 and @xmath24 . while @xmath25 remains unity at all orders due to charge conservation , quantum corrections yield @xmath26 .
thus , since @xmath27 , it follows that @xmath28 .
@xmath29 @xmath29 + [ 10 mm ] @xmath30 @xmath30
@xmath31 @xmath31 + with charged and neutral scalar exchanges.,width=340 ] @xmath32 @xmath33 @xmath34 @xmath32 @xmath35 @xmath34 in our 4g2hdm @xcite the one - loop contribution to the @xmath10amm can be subdivided as @xmath36^{sm4}_w + [ a_{\mu}]^{4g2hdm}_{\cal h},\ ] ] where @xmath37^{4g2hdm}_{\cal h}$ ] contains the charged and neutral higgs contributions coming from the one - loop diagrams in fig .
[ fig : one - loop ] ( see below ; the diagrams with @xmath38 and @xmath18 in the loop dominate ) , whereas the sm4-like contribution , @xmath37^{sm4}_w$ ] , comes from the one - loop diagram with @xmath39 in the loop and is given by @xcite = a(x_ ) , [ ag2sm4 ] where @xmath40 is the 24 element of the ckm - like pmns leptonic matrix , @xmath41 and the loop function @xmath42 is given by a(x_i ) = + .
for values of @xmath43 in the range @xmath44 one finds @xmath45^{sm4}_w/|u_{24}|^2 \lsim 3.0\times 10^{-9}$ ] , so that for @xmath46 ( as expected ) the simple sm4 can not accommodate the observed discrepancy in @xmath47 .
let us recapitulate the salient features of the 4g2hdm setups introduced in @xcite . in these models
one of the higgs fields ( the `` heavier '' field ) couples only to heavy fermionic states , while the second higgs field ( the `` lighter '' field ) is responsible for the mass generation of all other ( lighter ) fermions . applying this principle to the 4th generation leptonic sector we have @xmath48 where @xmath49
are left(right)-handed fermion fields , @xmath50 is the left - handed @xmath51 lepton doublet and @xmath52 are general @xmath53 yukawa matrices in flavor space . also , @xmath54 and @xmath55 are the two higgs doublets , @xmath56 is the identity matrix and @xmath57 .
the yukawa texture of ( [ eq : ly ] ) can be realized in terms of a @xmath58-symmetry under which the fields transform as follows : @xmath59 , @xmath60 , @xmath61 , @xmath62 ( for @xmath63 ) , @xmath64 ( for @xmath65 ) , and @xmath66 , @xmath67 . from the point of view of the leptonic sector ,
the yukawa interaction in ( [ eq : ly ] ) is the natural underlying setup that can effectively accommodate the heavy masses of the 4th generation leptons , by coupling them to the heavy higgs doublet .
this setup might also be an effective underlying description of more elaborate constructions in models of warped extra dimensions @xcite .
the yukawa interactions between the physical higgs bosons and the leptonic states are then given by ( see @xcite ) @xmath68 \right\ } \ell_j h \label{sff1}~ , \\
{ \cal l}(h \ell_i \ell_j ) & = & \frac{g}{2 m_w } \bar \ell_i \left\ { -m_{\ell_i } \frac{c_\alpha}{c_\beta } \delta_{ij } + f^h_\beta \cdot \left [ m_{\ell_i }
\sigma_{ij}^\ell r + m_{\ell_j } \sigma_{ji}^{\ell \ast } l \right ] \right\ } \ell_j h ~ , \\ { \cal l}(a \ell_i \ell_j ) & = & - i i_\ell \frac{g}{m_w } \bar \ell_i \left\ { m_{\ell_i } \tan\beta \gamma_5 \delta_{ij } - f_\beta \cdot \left [ m_{\ell_i } \sigma_{ij}^\ell r - m_{\ell_j } \sigma_{ji}^{\ell \ast } l \right ] \right\ } \ell_j a ~ , \\ { \cal l}(h^+ \nu_i e_j ) & = & \frac{g}{\sqrt{2 } m_w } \bar \nu_i \left\ { \left [ m_{e_j } \tan\beta \cdot u_{ji } - m_{e_k } f_\beta \cdot u_{ki } \sigma^{e}_{kj } \right ] r \right .
\nonumber \\ & & \left .
+ \left [ -m_{\nu_i } \tan\beta \cdot u_{ji } + m_{\nu_k } f_\beta \cdot \sigma^{\nu \ast}_{ki } u_{jk } \right ] l \right\ } e_j h^+ \label{sff2}~,\end{aligned}\ ] ] with @xmath69 and @xmath12 is the ratio between the two vevs .
also , @xmath70 is the charged higgs , @xmath71 are the physical neutral higgs states ( @xmath72 and @xmath73 are the lighter and heavier cp - even neutral states , respectively , and @xmath74 is the neutral cp - odd state ) , and @xmath75 or @xmath76 with weak isospin @xmath77 and @xmath78 , respectively .
also , @xmath79 and @xmath80 is the @xmath81 leptonic ckm - like pmns matrix .
finally , @xmath82 are new mixing matrices in the charged(neutral)-leptonic sectors , obtained after diagonalizing the lepton mass matrices @xmath83 where @xmath84 are the rotation ( unitary ) matrices of the right - handed charged and neutral leptons , respectively .
notice that @xmath85 and @xmath86 depend only on the elements of 4th rows of @xmath87 and @xmath88 , respectively , which we will treat as unknowns , i.e. , by expressing physical observables in terms of @xmath89 and @xmath90 or , equivalently in terms of @xmath91 and @xmath92.@xmath93}$ ] following @xcite , let us redefine the higgs yukawa interactions as @xmath94 \ell_j { \cal h } \label{generic}~,\end{aligned}\ ] ] with @xmath95 or @xmath76 and @xmath96 or @xmath74 .
then , neglecting terms of order @xmath97 for @xmath63 and terms of order @xmath98 for @xmath99 , the above scalar and pseudoscalar couplings , @xmath100 , @xmath101 and @xmath102 , @xmath103 ( @xmath104 or 3 ) , which mix the 4th generation leptons with the light leptons , are given in our 4g2hdm by @xmath105 ~ , \nonumber \\ & & p^{h^-}_{\ell_i \nu^\prime } = - \frac{g}{2 \sqrt{2 } } \frac{m_{\tau^\prime}}{m_w } f_\beta u_{44}^\ast \sigma_{44}^\nu \left [ \frac{m_{\nu^\prime}}{m_{\tau^\prime } } \left ( 1 - \frac{t_\beta}{f_\beta \sigma_{44}^\nu } \right ) \delta_{u_i } + \delta_{\sigma_i } \right ] \label{sppar}~,\end{aligned}\ ] ] where @xmath106 or @xmath74 , @xmath107 , @xmath108 , @xmath109 ( see eq . [ fpar ] ) and @xmath110 which are the small quantities that parameterize the amount of mixing between the 4th generation leptons and the light leptons of the 1st , 2nd and the 3rd generations .
in what follows we will take all quantities in eq .
[ sppar ] to be real and always set @xmath111 and @xmath112 ( for limits on @xmath12 in the 4g2hdm see @xcite ) .
we note that @xmath47 and the branching ratios for the lfv decays @xmath113 are proportional to @xmath114 ( see eqs . [ dominant ] and [ dominant2 ] ) , so that there is no enhancement for @xmath115 .
using eq .
[ sppar ] , the charged and neutral higgs contributions to @xmath11 [ with @xmath116 and @xmath117 in the loop ( @xmath106 or @xmath74 ) , respectively , see diagrams in fig .
[ fig : one - loop ] ] are given by ( see also @xcite ) ^4g2hdm_h^ ^1_0dx [ chargea ] , ^4g2hdm_h^0 ^1_0dx .
note that , for the neutral higgs case , the term proportional to @xmath118 vanishes since @xmath119 ( see eq .
[ sppar ] ) .
therefore , the dominant contribution , by far , to @xmath11 comes from the charged higgs exchange , in particular , from the second term ( proportional to @xmath118 ) in the numerator of eq .
[ chargea ] , where _ _ 2^ } , [ dominant ] so that @xmath47 is proportional to the product @xmath120 . in fig .
[ g-2 ] we plot @xmath47 as a function of the product @xmath120 ( assuming its real ) for several values of @xmath121 and @xmath122 and fixing @xmath123
( @xmath47 depends linearly on @xmath124 , see eq . [ dominant ] ) .
depending on the mass @xmath43 , we find that @xmath125 is typically required to accommodate the measured value of @xmath47 .
in what follows we will consider the constraints from the lepton flavor violating decays @xmath126 and from the decay @xmath127 , both of which are sensitive to the quantities @xmath128 and @xmath129 , as will be explained below .
lfv decays such as @xmath130 and @xmath1 , which are absent in the sm , are often found useful for constraining np models that can potentially contribute to the @xmath10amm , as such processes do not suffer from hadronic uncertainties .
the current experimental 90%cl upper bounds on these lfv decays are @xcite br ( ) < 4.4 10 ^ -8 , br(e ) < 2.410 ^ -12 [ lfvbounds ] .
let us define the amplitude for the transition @xmath113 as @xmath131 u_{\ell_i}(p ) \epsilon^{\mu\ast } \ , , \label{liljg}\ ] ] where @xmath132 is the photon polarization .
the decay width is then given by ( _ i _ j ) = ( 1 - ) . here again , the new 4g2hdm amplitude @xmath133 can be divided as ( _ i _ j ) ^4g2hdm ^sm4_w(_i _ j ) + m^4g2hdm_h^+(_i _ j ) + m^4g2hdm_h^0(_i _ j ) , where @xmath134 are the sm4-like w - exchange contribution which is obtained from the diagram ( right ) of fig .
[ fig : one - loop ] with @xmath31 replaced by @xmath135 plus the diagrams which contain the self - energy corrections to the external fermion line @xmath136 or @xmath137 . in particular , using the definition in eq . [ liljg ] and taking the limit @xmath138 , the net contribution to @xmath134 with internal @xmath18 in the loop is given by @xcite a^sm4_w = b^sm4_w = u_j4 u^_i4 f(x_^ ) , where @xmath139 and @xmath140 is given by f(x_i ) = . here also , we find that @xmath134 is much smaller than the charged and neutral higgs amplitudes , @xmath141 and @xmath142 ( calculated from the diagrams in fig . [
fig : one - loop ] ) , for which we obtain @xmath143 where the loop integrals
@xmath144 , @xmath145 , @xmath146 and @xmath147 are given by ( taking @xmath148 ) : i_1^h^+ & & ^1_0 dx , + i_2^h^+ & & ^1_0 dx , + i_1^h^0 & & ^1_0 dx , + i_2^h^0 & & ^1_0 dx .
+ the dominant terms in eqs .
[ 1stamp]-[4thamp ] are the ones proportional to @xmath121 from the charged higgs exchange contribution , @xmath149 since the terms proportional to @xmath124 in the neutral higgs exchanges vanish due to @xmath150 ( see eq . [ sppar ] ) .
we thus find that in our 4g2hdm , the decays @xmath1 and @xmath2 are sensitive to @xmath128 and @xmath129 through the products @xmath151 and @xmath152 , respectively , so that , in principle , one can avoid constraints on the quantities @xmath128 and @xmath129 if @xmath153 and @xmath154 are sufficiently small . in figs .
[ mueg ] we plot @xmath155 as a function of @xmath156 and @xmath157 , for @xmath158 gev , @xmath159 gev and fixing @xmath123 .
we see that for e.g. @xmath160 gev and for values of @xmath128 and @xmath129 of @xmath161 [ for which the product @xmath162 reproduces the measured @xmath47 ( see fig .
[ g-2 ] ) ] , @xmath163 and @xmath164 are required to be smaller than @xmath165 , implying that @xmath166 and @xmath167 . in fig .
[ taumug ] we plot @xmath168 as a function of @xmath169 and @xmath170 , and in fig . [ taumug2 ]
we give a scatter plot of the allowed values in the @xmath171 plane , for which @xmath172 ( i.e. , below its 90%cl bound ) . in both plots we use @xmath158 gev , @xmath159 gev and we fix @xmath123 .
the individual couplings @xmath128 and @xmath129 are randomly chosen to always be within values that reproduce the measured @xmath47 ( see fig .
[ g-2 ] ) .
we see that the products @xmath169 and @xmath170 are required to be at most @xmath173 , in order to be consistent with the current bounds on @xmath168 .
we can thus identify a typical benchmark texture for the 4th generation elements of the ckm - like pmns matrix , @xmath174 , and for the new mixing matrix @xmath175 that can explain the observed @xmath10amm and still be consistent with the current lfv constraints @xmath176 where e.g. , @xmath177 for @xmath160 gev .
admittedly , the above texture implies a hierarchical pattern which is different from the observed hierarchy in the quark s ckm matrix - usually termed as normal " . nonetheless , without a fundamental theory of flavor
, our insights for flavor should be data driven also in the leptonic sector . besides , the above texture is sensitive to the current precision in the measurement of the muon g-2 which can change e.g. , if more accurate calculations end up showing that part of the hadronic contributions can not be ignored . in table [ brlfv ] we list several representative values of the couplings @xmath128 , @xmath129 and the products @xmath178 , @xmath179 , @xmath180 and @xmath181 that are consistent with the measured @xmath10amm ( i.e. , @xmath182 given in eq . [ amuexp ] ) , and that give lfv branching fractions br@xmath183 and br@xmath184 at the level of @xmath185 and @xmath186 , respectively , that are accessible to near future experiments @xcite .
.the calculated @xmath47 and the branching fractions for the lfv decays @xmath187 and @xmath188 in the 4g2hdm , for several representative values of the couplings @xmath128 , @xmath129 and the products @xmath178 , @xmath179 , @xmath180 and @xmath181 that are consistent with the measured @xmath10amm and which give br@xmath183 and br@xmath184 at the level of @xmath185 and @xmath186 , respectively . [
cols="^,^,^,^,^,^,^,^,^ " , ]
in the sm , tree level @xmath190 fcnc transitions are forbidden , and also the purely leptonic @xmath191 decays , with @xmath192 , suffer from chiral suppression and are therefore very sensitive to new physics .
the sm predicted branching fractions for these decays are appreciably smaller than those of the semi - leptonic decays .
for example , for @xmath193 , the sm prediction is @xcite br(b_s ^+ ^- ) = ( 3.2 0.2 ) 10 ^ -9 . in the lhc era
the current limit on @xmath194 has been improved . a combined analysis by lhcb and cms , using 0.34@xmath195 and 1.14@xmath195 data sample , respectively , yields @xcite br(b_s ^+ ^- ) < 1.08 10 ^ -8 , ( lhcb+cms ) @ 95% cl whereas the same measurement by cdf - ii , using a 7@xmath195 data sample , gives @xcite br(b_s ^+ ^- ) < 4.0 10 ^ -8 cdf @ 95% cl . in fact , lhcb has the sensitivity to measure the @xmath194 down to @xmath196 , which is about @xmath197 smaller than the sm prediction . in general ,
the matrix element for the decay @xmath198 can be written as @xcite = , [ ampbsll ] where @xmath199 is the four momentum of the initial @xmath200 meson and @xmath201 s are functions of lorentz invariant quantities . squaring the matrix and summing over the lepton spins , we obtain the branching fraction br(|b_s^+ ^- ) = . in the sm , the dominant effect in @xmath202 arise from the diagrams shown in fig .
[ smdiags ] , which contribute to @xmath203 in eq .
[ ampbsll ] . at next - to - leading ( nlo ) qcd corrections , the net contribution in @xmath203
is given by @xcite f_a^sm = -if_b_sv_tb v_ts^ y(x_t ) = -if_b_sv_tb v_ts^ 0.997 ^1.55 .
in the sm4 it is again @xmath203 which receives a non - zero contribution : f_a^sm4 = ( f_a^sm4)^peng + ( f_a^sm4)^box , where @xmath204 is the contributions from the z - penguin diagrams in fig .
[ smdiags ] with the top quark replaced by a @xmath205 , and @xmath206 is the contribution from the box diagram in fig .
[ smdiags ] with the replacement @xmath207 and @xmath208 . at leading - order ( lo ) in qcd , they are given by ( f_a^sm4)^peng = - if_b_s v_tb v_ts^ , ( f_a^sm4)^box = - if_b_sv_tb v_ts^|u_24|^2 , with b_0(x_i , x_j ) & = & x_ix_j . in our 4g2hdm , we have additional contributions to @xmath203 coming from the charged higgs exchange penguin and box diagrams ( replacing @xmath209 in fig . [ smdiags ] ) and , in addition , there are new contributions to @xmath210 and @xmath211 . for our purpose ,
we are interested only in the diagrams that are sensitive to the @xmath212 vertex , which are , therefore , directly related to the muon g-2 .
the dominant diagrams that contribute to the @xmath202 decay with this vertex are the higgs - exchange box diagrams in fig .
[ smdiags ] , where one or two @xmath213-bosons are replaced by @xmath214 and @xmath215 are being replaced by both @xmath216 and @xmath217 .
thus , the net contributions to @xmath210 , @xmath211 and @xmath203 in eq . [ ampbsll ] can be written as f_s = f_s^h , f_p = f_p^h and f_a = f_a^sm + f_a^sm4 + f_a^h , where @xmath218 and @xmath219 are the contributions from the dominant new higgs - exchange box diagrams , which , for the @xmath217 exchange , are given by f_s^h & = & ( f_s)_hh^box + ( f_s)_wh^box , + & = & - [ fsh ] f_p^h & = & ( f_p)_hh^box + ( f_p)_wh^box , + & = & -[fph ] f_a^h & = & ( f_a)_hh^box + ( f_a)_wh^box , + & = & - , [ formfactors ] where @xmath220 , @xmath41 and the loop - functions are given by b_hh^s(y_i , y_j ) & = & b_hh^p(y_i , y_j ) = \ { - } - + b_wh^s(x_i , x_j , x_h ) & = & b_wh^p(x_i , x_j , x_h ) = + & & + + + b_hh^a(y_i , y_j ) & = & + + + b_wh^a(x_i , x_j , x_h ) & = & + + & & + .
also , @xmath221 , @xmath222 are defined in eq . [ sppar ] and s_tb^h^- & = & ( m_b - m_t ) t _
v_tb + f _ v_tb , + p_tb^h^- & = & ( m_b + m_t ) t _ v_tb - f _
v_tb , + s_ts^h^- & = & - p_ts^h^- = - m_t t _ v_ts+ f _ ( m_t v_t s _ 44^u
* + m_t v_ts _ 34^u * ) .
[ qcoup ] for the couplings @xmath223 we use the 4g2hdm yukawa terms in the quark sector as given in @xcite , where @xmath224 and @xmath225 are the corresponding new mixing matrices in the up and down - quark sectors , respectively , obtained after diagonalizing the quarks mass matrices .
in particular , adopting the type i 4g2hdm of @xcite , these matrices are given by @xmath226 in analogy with eq .
[ sigma ] , where @xmath227 are the rotation ( unitary ) matrices of the right - handed down and up - quarks , respectively .
they can be approximated by ( see @xcite ) @xmath228 so that @xmath229 if @xmath230 , and the @xmath231 blocks are parameterized by the quantities @xmath232 and @xmath233 .
as in @xcite , a natural choice that we will adopt below is @xmath234 .
thus , neglecting terms of @xmath235 ( and , therefore , neglecting also terms proportional to @xmath232 ) , the couplings @xmath223 can be approximated by s_tb^h^- & = & - p_tb^h^- v_tb + m_t f__t v_tb , + s_ts^h^- & = & - p_ts^h^- v_t s + m_t f__t v_ts , [ qcoup2 ] leading to @xmath236 , @xmath237 and f_a^h & & , [ formfactors2 ] notice that the term @xmath238 is proportional to @xmath239 and @xmath240 is proportional to both @xmath239 and @xmath241 .
thus , there is a net effect in @xmath3 from the charged - higgs box diagrams even when one of the small quantities that control the muon g-2 vanishes , i.e. , when either @xmath242 or @xmath243 , for which cases @xmath37^{4g2hdm}_{h^{\pm } } \to 0 $ ] ( see eq .
[ chargea ] ) .
for example , for @xmath242 we have @xmath37^{4g2hdm}_{h^{\pm } } \to 0 $ ] , while f_a^h ( _ u_2 0 ) b_hh^a(y_t,y_ ) f_^2 | u_44 |^2 | _
44^|^2 |__2 |^2 , the contributions from the charged higgs exchange diagrams with ( @xmath244 ) can be obtained directly from eqs .
[ fsh],[fph ] and [ formfactors ] by replacing @xmath245 with @xmath246 , which eventually requires the replacements : @xmath247 and @xmath248 in eq .
[ qcoup ] . in fig .
[ bsmumu ] we plot the contribution to @xmath249 in the 4g2hdm from the box diagrams with the @xmath214 and @xmath216 , @xmath217 exchanges , as a function of @xmath233 ( fixing @xmath250 ) and @xmath251 , for @xmath252 gev , @xmath123 and @xmath159 gev .
the allowed ranges of the key parameters @xmath128 and @xmath129 , which control the muon @xmath4 in our model , are randomly chosen in the range @xmath253 $ ] to be consistent with @xmath182 given in eq .
[ amuexp ] .
we also show the current experimental bound @xcite and the sm predicted value for @xmath249 .
we see that the contribution from the new ( i.e. , in the 4g2hdm ) box diagrams that involve the heavy 4th generation neutrino is consistent with the current experimental bound on @xmath249 for values of @xmath128 and @xmath129 that reproduce the observed muon @xmath4 .
it is also interesting to note that both in the sm4 and in the 4g2hdm , br@xmath254 can differ from the sm value by at - most a factor of ( 3 ) in either direction .
we have considered the effects of 1-loop exchanges of heavy 4th generation leptons on the muon @xmath4 , on the lepton flavor violating decays @xmath1 , @xmath2 and on @xmath127 , in the 4g2hdm which is a 2hdm where the higgs doublet with the heavier vev is coupled only to the 4th generation doublet while the lighter " higgs doublet is coupled to fermions of the 1st-3rd generations .
this model is particularly motivated for the leptonic sector , as it effectively addresses the heaviness of a 4th generation ew - scale neutrino .
the muon @xmath4 is sensitive in our model to the product @xmath162 , where @xmath255 , @xmath256 , @xmath257 is the leptonic ckm - like pmns matrix and @xmath258 are new mixing matrices in the charged and neutral leptonic sectors that are unique to the 4g2hdm .
we find that , depending on the mass @xmath43 , the experimentally measured muon magnetic moment can be accounted for if @xmath259 .
we also find that the decays @xmath1 and @xmath2 can have branching ratios which are not too far below the current bounds , i.e. , of @xmath260 and @xmath261 , respectively , if the products @xmath262 and @xmath263 , respectively .
we also considered the effects of one - loop exchanges of the 4th generation heavy neutrino @xmath264 on the decay @xmath127 and found that , in the four generations model considered here , @xmath265 can be larger or smaller than the sm predicted value by a factor of about three , for values of @xmath259 , which render the observed value of the muon @xmath4 to be consistent with the current upper limit on this decay .
t. kinoshita , m. nio , phys .
rev . * d70 * , 113001 ( 2004 ) ; t. kinoshita , m. nio , phys .
rev . * d73 * , 013003 ( 2006 ) ; t. kinoshita , m. nio , phys .
rev . * d73 * , 053007 ( 2006 ) ; g. gabrielse , d. hanneke , t. kinoshita , m. nio , b. c. odom , phys .
lett . * 97 * , 030802 ( 2006 ) ; t. aoyama , m. hayakawa , t. kinoshita , m. nio , phys .
lett . * 99 * , 110406 ( 2007 ) ; d. hanneke , s. fogwell , g. gabrielse , phys .
lett . * 100 * , 120801 ( 2008 ) ; g. degrassi , g. f. giudice , phys .
* d58 * , 053007 ( 1998 ) ; a. czarnecki , w. j. marciano , a. vainshtein , phys .
rev . * d67 * , 073006 ( 2003 ) ; s. heinemeyer , d. stockinger , g. weiglein , nucl .
* b699 * , 103 - 123 ( 2004 ) ; t. gribouk , a. czarnecki , phys .
* d72 * , 053016 ( 2005 ) ; j. prades , e. de rafael , a. vainshtein , in : roberts , lee b. , marciano , william j. ( eds.):lepton dipole moments " 303 - 317 , arxiv:0901.0306 [ hep - ph ] ; m. davier , a. hoecker , g. lopez castro , b. malaescu , x. h. mo , g. toledo sanchez , p. wang , c. z. yuan _ et al .
_ , eur .
j. * c66 * , 127 - 136 ( 2010 ) . c. bouchiat and l. michel , phys .
* 106 * , 170 ( 1957 ) ; m. gourdin and e. de rafael , nucl .
phys . b * 10 * , 667 ( 1969 ) .
m. davier , a. hoecker , b. malaescu , c. z. yuan and z. zhang , eur .
j. c * 66 * , 1 ( 2010 ) .
k. hagiwara , a. d. martin , d. nomura and t. teubner , phys .
b * 649 * , 173 ( 2007 ) .
f. jegerlehner and a. nyffeler , phys .
rept . * 477 * , 1 ( 2009 ) .
a. nyffeler , phys .
d * 79 * , 073012 ( 2009 ) .
j. l. lopez , d. v. nanopoulos , x. wang , phys .
* d49 * , 366 - 372 ( 1994 ) ; u. chattopadhyay , p. nath , phys .
rev . * d53 * , 1648 - 1657 ( 1996 ) ; s. p.
martin , j. d. wells , phys .
rev . * d64 * , 035003 ( 2001 ) .
s. bar - shalom , s. nandi and a. soni , phys . rev .
d * 84 * , 053009 ( 2011 ) m.a .
luty , phys . rev .
* d41 * , 2893 ( 1990 ) .
a. j. buras , b. duling , t. feldmann , t. heidsieck , c. promberger and s. recksiegel , jhep * 1009 * , 106 ( 2010 ) . w. s. hou and c. y. ma , phys .
rev . * d82 * , 036002 ( 2010 ) .
s. nandi and a. soni , phys .
d * 83 * , 114510 ( 2011 ) see also , m. bobrowski , a. lenz , j. riedl and j. rohrwild , phys . rev .
d * 79 * , 113006 ( 2009 ) ; o. eberhardt , a. lenz and j. rohrwild , phys .
d * 82 * , 095006 ( 2010 ) .
w. -s .
hou , f. -f .
lee , c. -y .
ma , phys .
* d79 * , 073002 ( 2009 ) .
leveille , nucl .
* b137 * , 63 ( 1978 ) .
j. adam _ et al . _
[ meg collaboration ] , phys .
lett . *
107 * , 171801 ( 2011 ) .
m. blanke , a. j. buras , b. duling , a. poschenrieder , c. tarantino , jhep * 0705 * , 013 ( 2007 ) . m.
bona _ et al . _
( superb collaboration ) , arxiv:0709.0451 [ hep - ex ] .
t. aaltonen _ et al . _
( cdf collaboration ) , phys .
lett . * 107 * , 239903 ( 2011 ) and phys .
* 107 * , 191801 ( 2011 ) .
h. e. logan and u. nierste , nucl .
phys . b * 586 * , 39 ( 2000 ) .
g. buchalla , a. j. buras , nucl .
* b400 * , 225 - 239 ( 1993 ) . | in the minimal standard model ( sm ) with four generations ( the so called sm4 ) and in standard " two higgs doublets model ( 2hdm ) setups , e.g. , the type ii 2hdm with four fermion generations , the contribution of the 4th family heavy leptons to the muon magnetic moment is suppressed and can not accommodate the measured @xmath0 access with respect to the sm prediction .
we show that in a 2hdm for the 4th generation ( the 4g2hdm ) , which we view as a low energy effective theory for dynamical electroweak symmetry breaking , with one of the higgs doublets coupling only to the 4th family leptons and quarks ( thus effectively addressing their large masses ) , the loop exchanges of the heavy 4th generation neutrino can account for the measured value of the muon anomalous magnetic moment .
we also discuss the sensitivity of the lepton flavor violating decays @xmath1 and @xmath2 and of the decay @xmath3 to the new couplings which control the muon g-2 in our model . |
fundamental progress on the knowledge of gamma - ray bursts ( grbs ) has been made possible by detection of their optical counterparts . of nearly 40 grbs
accurately and rapidly localized so far by bsax , batse / rxte , ipn , and promptly followed up in the optical , only about 50% exhibited optical afterglowsjcg / grbgen.html ] , suggesting that these sources are rapidly fading , or heavily obscured .
the best monitored afterglows ( grbs 970228 , 970508 , 980326 , 980519 , 990123 , 990510 ) exhibit a variety of behaviors , indicating that the shape of the optical decay must be determined not only by the intrinsic physics , but also by the nature , structure and composition of the surrounding medium . therefore , optical light curves of grb counterparts need to be frequently sampled for long time intervals , to follow the evolution of the afterglow and to allow mapping the characteristics of the medium .
grb000301c was detected by the ipn and by the rxte asm on 2000 march 1.4 ut with an error box of 50 arcmin@xmath3 ( smith et al .
its field was acquired starting @xmath41.5 days later by various optical , infrared and radio telescopes .
the optical afterglow was independently detected by fynbo et al .
( 2000 ) and by us ( bernabei et al .
2000a ) , and is among the brightest ever observed . near - infrared detection and monitoring of the afterglow
are reported in rhoads & fruchter ( 2000 ) .
observations of the counterpart at radio and millimetric wavelengths have been reported by berger & frail ( 2000 ) and bertoldi ( 2000 ) , respectively .
ultraviolet spectroscopy with the stis instrument onboard hst allowed the determination of the redshift ( smette et al .
2000 ) , then refined by optical ground - based spectroscopy ( @xmath5 = 2.03 , castro et al . 2000 ) .
the good sampling and the brightness of the grb000301c afterglow have allowed a detailed study of its evolution up to 15 days after the explosion . in this paper
we present the results of the optical monitoring conducted at loiano , calar alto , sierra nevada , nainital and canary islands .
rccccc exposure start & telescope & filter & exp .
time & seeing & magnitude@xmath6 + & & & ( minutes ) & ( arcsecs ) & + 2000 mar 2.906 & upso & r & 70 & 1.4 & 20.42 @xmath7 0.04@xmath3 + 3.144 & caha & r & 5 & 1.1 & 20.25 @xmath7 0.05 + 3.179 & caha & b & 15 & 1.1 & 21.07 @xmath7 0.05 + 3.185 & loiano & r & 16.7 & 2 & 20.16 @xmath7 0.05 + 3.205 & caha & r & 5 & 1.1 & 20.25 @xmath7 0.05 + 3.210 & caha & i & 10 & 1.1 & 19.94 @xmath7 0.07 + 3.219 & caha & v & 15 & 1.1 & 20.57 @xmath7 0.05 + 3.232 & caha & b & 15 & 1.1 & 21.10 @xmath7 0.12 + 3.913 & upso & r & 50 & 1.2 & 20.51 @xmath7 0.04 + 4.038 & caha & r & 15 & 1.6 & 20.53 @xmath7 0.06 + 4.149 & loiano & r & 36.7 & 3 & @xmath8 20.25@xmath9 + 4.165 & loiano & b & 20 & 3 & @xmath8 21.0 + 5.135 & sno & r & 20 & 2 & 20.47 @xmath7 0.07 + 5.152 & sno & b & 20 & 2 & 21.60 @xmath7 0.20 + 5.172 & sno & v & 20 & 2 & 21.04 @xmath7 0.20 + 5.930 & upso & r & 85 & 1.3 & 21.14 @xmath7 0.06 + 6.135 & loiano & r & 30 & 1.7 & 21.65 @xmath7 0.20 + 6.163 & loiano & b & 30 & 1.7 & 22.45 @xmath7 0.15 + 6.185 & loiano & i & 16.7 & 1.7 & 20.82 @xmath7 0.15 + 6.968 & upso & r & 35 & 1.6 & @xmath8 21.6 + 7.125 & loiano & r & 30 & 1.7 & 21.68 @xmath7 0.15 + 7.149 & loiano & b & 35 & 1.7 & 22.43 @xmath7 0.10 + 7.177 & loiano & i & 20 & 1.7 & 21.20 @xmath7 0.15 + 7.894 & upso & r & 105 & 1.6 & 22.00 @xmath7 0.15 + 8.146 & loiano & r & 30 & 1.6 & 21.68 @xmath7 0.10 + 8.170 & loiano & i & 30 & 1.6 & 21.61 @xmath7 0.10 + 8.924 & upso & r & 75 & 1.3 & 22.04 @xmath7 0.20 + apr 5.213 & tng & b & 20 & 0.5 & @xmath8 25.5 + + + +
optical @xmath0 images were collected soon after the notification of the grb000301c detection , starting @xmath41.5 days after the high - energy event .
observations were carried out with ccd cameras at the 1.52-meter
cassini " telescope of the bologna university in loiano , italy , at the 1.0-meter upso telescope in nainital , india , at the 1.2-meter caha telescope in calar alto , spain , at the 1.5-meter telescope of the sierra nevada observatory ( sno ) in granada , spain , and at the _ telescopio nazionale galileo _ ( tng ) in the canary islands , spain .
the complete log of these observations is reported in table 1 .
images were debiased and flat - fielded with the standard cleaning procedure . due to the proximity
( 7 arcsec ) of the target to a bright star , we chose to use standard psf - fitting as our photometric technique , and for this purpose we used the daophot ii image data analysis package psf - fitting algorithm ( stetson 1987 ) running within _ midas_. a two - dimensional gaussian with two free parameters ( the half width at half maxima along @xmath10 and @xmath11 coordinates of each frame ) was modeled on at least 5 non - saturated bright stars in each image .
for each filter , this procedure yields frames magnitude differences among the photometric references of less than 1% in all frames .
the errors associated with the measurements reported in table 1 represent statistical uncertainties ( at 1@xmath12 ) , obtained with the standard psf - fitting procedure .
calibration was done using the @xmath0 magnitudes of field stars as measured by henden ( 2000 ) .
in our images we detect a point - like source within the 50 square arcmin error box of the grb , at the position ra=
16@xmath13 20@xmath14 18@xmath155 , dec = 29@xmath16 26@xmath17 35@xmath18 ( j2000 ) , consistent with that given by fynbo et al .
the source variability ( see magnitude levels in table 1 ) suggests that this is the afterglow of grb000301c .
the light curves in @xmath0 bands are reported in fig .
1 , where our data are complemented with those published by other authors ( sagar et al . 2000 , jensen et al . 2000 , and the gcn circulars archive ) . note that some results presented in this paper supersede preliminary values reported in gcns .
our analysis of upso @xmath1 band data yielded results consistent with those reported by sagar et al .
no correction has been applied for galactic extinction , which is anyway small in the direction of the grb ( @xmath19 = 0.052 , schlegel et al .
1998 ) ; nor has been subtracted any host galaxy continuum emission , this being negligible ( fruchter et al . 2000 ) .
a 5% systematic error was added in quadrature to the errors reported in figs . 1 and 2 to take into account possible photometric discrepancies due to the use of different telescopes and instruments . the @xmath1 band light curve , the best sampled , exhibits in its early portion a flaring activity with hour time scale ( fig .
1 ) , with an initial increase ( confirmed by s.g .
bhargavi , priv . comm . ) .
the flux then shows a slow decline , lasting about 1.5 days and following approximately a power - law , with a slope @xmath20 0.7 ( @xmath21 , where @xmath22 is the grb trigger time ) .
subsequently the light curve flattens , and an approximately constant , or slightly increasing , behavior is seen till around 3.7 days after the grb .
the flux starts decreasing again thereafter .
this decline , which can be fitted by a very steep power - law ( index @xmath20 3.5 ) , levels off around 5 days after the grb trigger to a plateau " of two days duration .
the flux resumes then the decreasing trend , with a shallower power - law of @xmath20 3 , till the end of the monitoring . the late @xmath1 band epoch flux and upper limit reported by fruchter et al .
( 2000 ) and veillet ( 2000 ) , respectively , are consistent with this trend .
the @xmath23 band light curve appears well correlated with the @xmath1 band , though less well sampled .
the @xmath23 band points at 3.5 and 3.7 days after the grb suggest a variation opposite to that observed simultaneously in the @xmath1 band .
however , the @xmath23 band variation is not significant and determines only a marginally significant change in the @xmath24 color ( fig .
the @xmath23 band upper limit determined on april 5 with the tng is consistent with the power - law decline of the final portion of the light curve .
the fewer @xmath25 band points show a good correlation of the light curve with that in the @xmath1 band , with no measurable temporal lag ( the @xmath26 color is unchanged , fig .
2b ) . in particular , the @xmath25 band data around 3 - 4 days after the high - energy event also suggest a local flattening of the light curve .
the @xmath27 band data confirm the general steepening observed in the other bands , although the second plateau at 5 - 7 days after the high - energy event is less clearly seen than in the @xmath23 and @xmath1 light curves .
also , a rapid flux increase is apparent at the beginning of the @xmath27 band monitoring , delayed by @xmath47 hours with respect to that seen in the @xmath1 band light curve ( see figs . 1 and 2c ) .
our optical monitoring of the bright grb000301c afterglow has provided one of the best sampled afterglow datasets , especially in the @xmath1 filter .
the long term behavior of this optical afterglow is better described by a continuous steepening , rather than by a single power - law , as expected in afterglows developing in laterally spreading jets ( sari et al .
1999 ; rhoads 1999 ) or decelerating to non - relativistic regimes ( dai & lu 1999 ) , and seen in few other cases . among equally well monitored grb afterglows , gr000301c appears peculiar in that several shorter time scale variations are superimposed on the long term decrease .
the reality of two of these ( 3.1 - 3.7 days and 5 - 7 days after the event ) is supported by their appearance in more than one band .
the first two points of the @xmath1 and @xmath27 band light curves might suggest a rise and could be reminiscent of the early ( 1 - 2 days after the grb trigger ) light curve of grb970228 and grb970508 ( guarnieri et al . 1997 ; pedersen et al . 1998 ) , although in the latter the initial increase was more structured . in the present case
we can not exclude that the flux is declining since the start of the monitoring , and hour time scale flares modulate this decrease .
some isolated short term variability events are seen in grb980703 ( vreeswijk et al . 1999 ) and grb990123 ( castro - tirado et al .
1999 ) and are almost totally absent in grb990510 ( e.g. , stanek et al . 1999 ) .
recently , various scenarios have been developed in which intrinsic re - energization of the blast wave , or irregularities of the dense interstellar medium in which the blast is expanding can account for the observed behavior ( panaitescu et al .
1998 ; mszros et al . 1998 ; sari & mszros 2000 ; wang & loeb 2000 ; dai & lu 2000 ) .
in particular , a flattening of the afterglow light curve , similar to that exhibited by grb000301c in the @xmath1 band on days 3.1 - 3.7 and 5 - 7 days after the grb , is predicted by kumar & piran ( 2000 ) as a consequence of the collision of a slow shell ejected at a late time after the grb with an outer shell decelerated by its propagation in the circumburst medium ( see their fig . 5 ) .
we note that the temporal occurrence of the observed flattenings could be consistent with a colliding shells " interpretation , while it is incompatible with the time scale implied by an hypernova scenario ( see rhoads & fruchter 2000 )
. the lack of a clear correlation between the @xmath1 band light curve and the @xmath27 and @xmath28 band light curves ( see rhoads & fruchter 2000 for the latter ) might be due to non strict simultaneity of the data points .
in fact , the @xmath29 and @xmath30 colors as a function of time show only marginally significant deviations from constancy ( fig . 2c and 2d ) , and these are mainly exhibited by color values derived from pairs of measurements separated in time by more than 0.5 hours , the shortest variability time scale observed in this afterglow .
our findings underline the critical importance of intensive multiwavelength observations of afterglow sources .
we thank the staff of the loiano , calar alto , sierra nevada , upso and tng observatories .
cb , ag , and ap acknowledge the university of bologna ( funds for selected research topics ) .
gmb thanks the russian fund of fundamental researches for support ( grant 98 - 02 - 17570 ) . | we present @xmath0 light curves of the afterglow of grb000301c , one of the brightest ever detected at a day time scale interval after grb trigger .
the monitoring started 1.5 days after the grb and ended one month later .
inspection of the extremely well sampled @xmath1 band light curve and comparison with @xmath2 data has revealed complex behavior , with a long term flux decrease and various short time scale features superimposed .
these features are uncommon among other observed afterglows , and might trace either intrinsic variability within the relativistic shock ( re - acceleration and re - energization ) or inhomogeneities in the medium in which the shock propagates . |
the study of the quantum compass model ( cm ) for strongly correlated transition - metal compounds with orbital degrees of freedom and strong spin - orbit coupling is an active field of research ( for a review , see ref .
in particular , quantum and thermodynamic phase transitions in the two - dimensional ( 2d ) cm were studied in refs .
@xcite , where the directional - ordering transition of the 2d ising universality class was found for the symmetric cm . depending on the method of numerical computation
, the temperature of the phase transition for the quantum cm was found in the range @xmath5 ( in the units of the cm exchange interaction ) @xcite . in refs .
@xcite the compass - heisenberg ( ch ) model was introduced extending the cm by the 2d antiferromagnetic ( af ) heisenberg model , and the ground state and excited states of the ch model were analyzed . for 5d transition - metal compounds with a strong spin - orbit coupling , such as sr@xmath3iro@xmath4 and ba@xmath3iro@xmath4 with the nel temperatures @xmath6k and @xmath7k , respectively , ( see , e.g. , ref .
@xcite ) , an effective af heisenberg model for pseudospins @xmath8 with the cm anisotropy was derived in ref .
recently we have calculated the spin - wave excitation spectrum and @xmath1 for a layered ch model by means of the random phase approximation ( rpa ) @xcite ( see also refs .
an important issue is the description of magnetic long - range order ( lro ) and short - range order ( sro ) and of the thermodynamics at arbitrary temperatures by a theory going beyond rpa . in this paper
we employ a generalized mean - field approximation ( gmfa ) to the 2d ch model that is based on the equation - of - motion method for green functions . in the framework of a more general theory for the dynamic spin susceptibility including the self energy @xcite
, the neglect of the self - energy corresponds to the gmfa .
this approximation , for spin - rotation invariant ( sri ) systems also named sri green - function method ( rgm ) , has been successfully applied to several quantum spin systems ( see , e.g. , refs .
@xcite ) .
we start from the spin-@xmath0 ch model on the square lattice , @xmath9 where @xmath10 .
the nearest - neighbor ( nn ) exchange interaction parameters are @xmath11 , @xmath12 , @xmath13 , @xmath14 , @xmath15 , and @xmath16 .
we assume @xmath17 and @xmath18 .
the symmetric formulation of the model ( [ eq1 ] ) allows us to get expressions for quantities with indexes @xmath19 @xmath20 from those indicated by the index @xmath21 by cyclic permutation .
to evaluate the thermodynamic quantities in the ch model , we calculate the dynamic spin susceptibility @xmath22 ( @xmath23 denotes the retarded two - time commutator green function @xcite ) . using the equations of motion up to the second step , we obtain @xmath24 , where @xmath25 \rangle$ ] , @xmath26 $ ] , and @xmath27 , h]$ ] . for the model ( [ eq1 ] )
the moment @xmath28 is given by the exact expression @xmath29 . \label{eq2}\ ] ] here , @xmath30 with @xmath31 denote the spin correlation functions .
the second derivatives @xmath32 are approximated in the spirit of the scheme employed in refs .
that means , taking the site representation , in @xmath33 we decouple the products of three spin operators on different lattice sites along nn sequences @xmath34 as @xmath35 where the vertex renormalization parameters @xmath36 and @xmath37 are attached to nn and further - distant correlation functions , respectively . after some algebra
, we obtain @xmath38 and @xmath39 with the squared spin - excitation energy @xmath40 , \label{eq6}\ ] ] where @xmath41 , \label{eq7}\ ] ] @xmath42 the appearance of af lro at @xmath43 is reflected in our theory by the divergence of @xmath44 corresponding to the closure of the spectrum gap at the af ordering vector @xmath45 , @xmath46 . in the lro phase the correlation functions @xmath47 are written as @xcite @xmath48 with @xmath49 calculated from the green function ( [ eq5 ] ) by the spectral theorem , @xmath50 , \label{eq11}\ ] ] where @xmath51 .
the condensation part @xmath52 arising from @xmath53 determines the staggered magnetization @xmath54 that is defined by @xmath55 in the paramagnetic phase , we have @xmath56 and @xmath57 .
the nn correlation functions are related to the internal energy @xmath58 per site , @xmath59 , from which the specific heat @xmath60 may be obtained . to calculate the thermodynamic properties , the correlation functions
@xmath61 , the vertex parameters @xmath62 appearing in the spectrum @xmath63 , and the condensation term @xmath52 in the lro phase have to be determined .
besides eqs .
( [ eq10 ] ) and ( [ eq11 ] ) for calculating the correlators , we have the sum rules @xmath64 and the lro conditions @xmath53 ; that is , we have more parameters than equations . to obtain a closed system of self - consistency equations ,
we proceed as follows . as in our recent study of the model ( [ eq1 ] ) by means of the rpa and linear spin - wave theory ( lswt ) @xcite
, we consider an anisotropic cm interaction , @xmath65 , so that the lro phase is an easy - axis af with the magnetization along the x axis .
accordingly , we put @xmath66 , where we can also consider the limiting case @xmath67 .
then , at t = 0 , for determining the seven quantities @xmath62 and @xmath68 , besides the three sum rules and the three lro conditions , we need an additional condition . to this end , we adjust the ground - state magnetization @xmath69 to the expression obtained in lswt , @xmath70 , where the correlation function is given by eq .
( 17 ) of ref .
@xcite with the sublattice magnetization @xmath71 substituted by the spin @xmath8 . in the lro phase , @xmath72 , we have found that the ansatz @xmath73 , where @xmath74 with @xmath75 , is a reasonable approximation , as will be demonstrated in the appendix for the heisenberg limit . at @xmath76 , for calculating @xmath62 we use the sum rules and the condition @xmath77 .
because , as an input , we take the ground - state magnetization in lswt that describes the lro quite well for small enough cm interaction @xmath78 as compared to the exchange interaction @xmath79 , we present results in the region @xmath80 .
as described in sec .
ii , the thermodynamic quantities are calculated from the numerical solution of a coupled system of nonlinear algebraic self - consistency equations .
first we consider our results on the magnetic lro . in the heisenberg limit
we have no lro at finite temperature , in accord with the mermin - wagner theorem . in fig .
[ fig1 ] the nel temperature @xmath1 as function of @xmath81 for two ratios of @xmath82 is plotted .
our results for @xmath1 remarkably deviate from those found in rpa @xcite . for @xmath83 ,
@xmath1 exhibits qualitatively the same dependence on @xmath81 , but is appreciably reduced as compared to rpa . for the symmetric cm interaction @xmath84 , our theory yields a finite nel temperature that increases with @xmath85 , whereas in rpa , @xmath86 was found .
we ascribe these differences to a better description of strong spin fluctuations by our green - function theory going one step beyond rpa .
note that in the rpa @xcite the fluctuations of the @xmath87 component of spin are not taken into account . generally speaking , from our results
we conclude that the cm interaction added to the heisenberg model favors magnetic lro at finite @xmath88 in two dimensions . comparing our findings to experiment
, we consider the compound ba@xmath3iro@xmath4 , where quantum chemistry calculations @xcite yield the parameters @xmath89mev , @xmath90mev @xmath91 , and a very small interplane coupling @xmath92 that is not taken into account here . in fig .
[ fig2 ] the magnetization @xmath93 as function of @xmath88 is plotted . at @xmath94 , @xmath95
is enhanced by @xmath85 as compared to the value in the heisenberg limit ( @xmath96 ) .
the second - order phase transition occurs at @xmath97k which agrees rather well with the experimental value @xmath98k . at @xmath1 the specific heat ( inset to fig .
[ fig2 ] ) reveals a cusplike singularity .
such a structure at the transition temperature was also found in layered magnets treated by rgm in ref .
@xcite . from the analysis given there
we suggest that also here the height of the cusp may be underestimated .
it is desirable to compare our result for @xmath99 to experimental data which , however , are not yet available . finally , we discuss the uniform static spin susceptibility @xmath100 depicted in fig .
[ fig3 ] for the symmetric cm interaction , which reflects the behavior of magnetic sro . considering the heisenberg limit ,
the increase of @xmath2 with temperature is caused by the decrease of af sro , i.e. , of the spin stiffness against the orientation along a homogeneous magnetic field .
since the sro is less pronounced at higher @xmath88 , @xmath2 reveals a maximum near the exchange energy and a crossover to the high - temperature curie - weiss behavior . in the paramagnetic phase for @xmath101 ,
the susceptibilities @xmath102 are lowered as compared with the heisenberg limit due to the @xmath85-induced enhancement of af sro .
moreover , the susceptibility @xmath103 is reduced as compared with @xmath104 , because the sro of the x- and y - spin components is more pronounced due to the af cm interaction along the x- and y - axes . in the lro phase
, the susceptibility @xmath105 strongly decreases with decreasing @xmath88 , because the increasing easy - x - axis lro also enhances the sro . considering the susceptibility @xmath106 , with decreasing @xmath88 below @xmath1 the sro of the y - spin components
is reduced due to the increasing alignment of the spins in x - direction , so that @xmath106 increases with decreasing @xmath88 .
thus , @xmath106 exhibits a minimum at @xmath1 .
the susceptibility @xmath104 in the lro phase turns out to be nearly temperature independent .
in this paper we have calculated thermodynamic properties of the 2d antiferromagnetic ch model within a generalized mean - field approach for arbitrary temperatures . our main focus was the detailed investigation of magnetic lro .
we have found that the nel temperature @xmath1 is finite even in the case of a symmetric cm interaction .
we have explained the temperature dependence of the uniform static susceptibility in terms of magnetic sro .
our investigation forms the basis for forthcoming extended studies of the ch model .
especially , the generalization of the theory holding for arbitrary ch model parameters and allowing for the description of magnetic and directional - ordering transitions is of particular interest .
+ the authors would like to thank w. janke and j. richter for valuable discussions .
the financial support by the heisenberg - landau program of jinr is acknowledged .
crrrc + & @xmath107 & @xmath108 & average & sri + in the limit @xmath114 , the lro in the ground state of the heisenberg af may be described within the rgm by the sri forms of eqs .
( [ eq10 ] ) and ( [ eq12 ] ) , i.e. , by @xmath115 and @xmath116 as it was done , e.g. , in refs .
according to the consideration of the ch model with the easy - axis magnetization @xmath117 , let us outline an alternative possibility to describe the ground - state lro in the af heisenberg model . here , we may also break the rotational symmetry by putting @xmath66 , which is analogous to the introduction of a symmetry - breaking sublattice magnetization ( see , e.g. , our rpa approach @xcite ) .
performing the calculations at @xmath94 as described in sec .
ii , we obtain the correlation functions @xmath118 and the uniform static spin susceptibility @xmath102 listed in table [ table1 ] .
the spin - excitation spectrum at @xmath94 is found to be @xmath119 where @xmath120 and @xmath121 .
the amplitudes @xmath122 are given in table [ table1 ] .
note that eq .
( [ a1 ] ) leads to the sri result ( ref .
@xcite ) , if we put @xmath123 and @xmath124 , and has the same shape as in lswt and rpa . to relate the non - sri description of lro to the sri formulation ,
we calculate arithmetic averages , e.g. , @xmath125 .
as can be seen from table [ table1 ] , we obtain a very good agreement of the averaged ground - state properties with those resulting from the sri theory . moreover , the geometrical average @xmath126 used in our ansatz for the lro phase ( see sec .
ii ) given in the heisenberg limit by @xmath127 nearly agrees with the ratio @xmath128 obtained in the sri approach ( see also ref .
this gives some justification for formulating our ansatz in terms of the geometrical average . | a green - function theory for the dynamic spin susceptibility in the square - lattice spin-@xmath0 antiferromagnetic compass - heisenberg model employing a generalized mean - field approximation is presented .
the theory describes magnetic long - range order ( lro ) and short - range order ( sro ) at arbitrary temperatures .
the magnetization , nel temperature @xmath1 , specific heat , and uniform static spin susceptibility @xmath2 are calculated self - consistently . as the main result
, we obtain lro at finite temperatures in two dimensions , where the dependence of @xmath1 on the compass - model interaction is studied .
we find that @xmath1 is close to the experimental value for ba@xmath3iro@xmath4 .
the effects of sro are discussed in relation to the temperature dependence of @xmath2
. published in : eur .
phys .
j. * b 88 * , 148 ( 2015 ) . |
let @xmath0 be the interior of an orientable , compact , connected 3manifold with non - empty boundary consisting of a pairwise disjoint union of tori , and let @xmath1 be an ideal ( topological ) triangulation of @xmath0 with the property that all edges are essential . following an approach sketched by thurston @xcite , projective classes of transversely measured singular codimension
one foliations of @xmath0 are associated to degenerations of ideal hyperbolic triangulations of @xmath5 which are parameterised by an affine algebraic set @xmath3 called the _
deformation variety_. the set @xmath6 is also related to the study of representations of @xmath7 into @xmath8 and used to obtain an explicit understanding of the link between the topology of a cusped 3manifold and ideal points of varieties related to representations into @xmath9 and @xmath10 this link is known to exist from constructions by culler , morgan and shalen @xcite .
the new contribution of this paper lies in an analysis of the deformation variety using hyperbolic geometry and tropical geometry , as well as ideas of bestvina @xcite , thurston @xcite and yoshida @xcite .
the canonical morgan - shalen compactification of the character variety is infinite dimensional and relies on hironaka s theorem .
the approach taken here uses an ideal triangulation as a particular , finite - dimensional coordinate system .
the aforementioned link is explicitly described as a degeneration of the ideal hyperbolic triangulation , and the use of ideal triangulations and tropical geometry allows an algorithmic construction of surfaces ( or , more generally , transversely measured singular foliations ) dual to ideal points of the character variety .
the set of ideal points @xmath11 is defined to be bergman s _ logarithmic limit set _ of @xmath12 this tropical compactification turns out to have an explicit geometric interpretation : an ideal hyperbolic tetrahedron in @xmath13 is the convex hull of four distinct points on the sphere at infinity .
it can be positively or negatively oriented and is flat ( an ideal quadrilateral ) if its ideal vertices lie on a round circle .
it degenerates as one of its ideal vertices is moved to coincide with another .
the deformation variety describes the shapes of ideal hyperbolic tetrahedra , and has the property that an ideal point is approached if and only if some tetrahedron degenerates .
a degenerating tetrahedron is seen to become very long and thin .
it converges in the gromov
hausdorff sense to a dual spine if the hyperbolic metric is suitably rescaled , and it inherits a singular codimension
one foliation from the collapse as shown in figure [ fig : deg singular foliation ] .
applying the same rescaling process to a tetrahedron which does not degenerate gives the singular foliation shown in figure [ fig : non - deg singular foliation ] .
moreover , the foliation inherits a transverse measure from the relative growth rates at which the tetrahedra degenerate . the main result in section [ sec : ideal points and normal surfaces ] ( proposition [ comb : homeo ] ) formalises this interpretation by giving a canonical identification of @xmath11 with
a compact subset of the projective admissible solution space of spun - normal surface theory , denoted @xmath14 in @xcite , via a natural injection ; @xmath15 the set @xmath11 can be computed using work of bogart , jensen , speyer , sturmfels and thomas @xcite ; the algorithm is implemented in the software package by jensen @xcite .
the ideal points of @xmath16 are analysed via actions on @xmath17trees . the first part , found in section [ sec : the dual tree ] , applies to any admissible projective class .
every @xmath18 defines a transversely measured singular codimension
one foliation @xmath19 if @xmath20 has rational coordinate ratios , then there is also an associated spun - normal surface @xmath21 corresponding to a closed ( not necessarily connected ) leaf .
the foliation lifts to a foliation @xmath22 of the universal cover @xmath23 of @xmath5 and the leaf space @xmath24 is turned into an @xmath17tree @xmath25 on which @xmath7 acts by isometries with translation length function @xmath26 how can one decide whether @xmath27 is non - trivial ?
this question is addressed in section [ the geometric construction ] for the admissible projective classes corresponding to ideal points of @xmath12 so suppose that @xmath28 for some @xmath29 each ideal 3simplex in @xmath23 is imbued with a hyperbolic structure , and @xmath24 is suitably interpreted as a limit of @xmath23 as the ideal triangulation degenerates .
there is an associated degeneration of the hyperbolic space @xmath13 with a limiting action of @xmath7 with length function @xmath30 coming from a sequence of representations .
the function @xmath30 is not canonical , but it has the property that @xmath31 for all @xmath32 it is in general difficult to determine @xmath33 an exception are peripheral elements , where it is determined by the linear functional @xmath34 of @xcite , 3 .
the relationship of @xmath34 with the action on the tree is established in section [ holonomy variety and essential surfaces ] , where it is also shown that the boundary curves of the spun - normal surfaces in @xmath35 are the boundary curves of essential surfaces which are strongly detected by the character variety . before the next result
is stated , more terminology is introduced .
a surface in @xmath0 is _ non - trivial _ if it is essential or can be reduced to an essential surface by performing compressions and then possibly discarding some components . if @xmath36 has rational coordinate ratios , so does @xmath37 and there is an associated spun - normal surface @xmath38 which determines the foliation .
a compact spun - normal surface is an ordinary normal surface .
[ thm : main 1 ] let @xmath0 be the interior of a compact , connected , orientable , irreducible 3manifold with non - empty boundary consisting of a disjoint union of tori , and let @xmath1 be an ideal triangulation of @xmath0 with the property that all edges are essential .
let @xmath39 and @xmath40 the action on @xmath25 is non - trivial if @xmath41 for some peripheral element @xmath32 in particular , if @xmath36 has rational coordinate ratios and the associated surface @xmath42 is non - compact , then @xmath42 is non - trivial .
three further remarks to the above theorem should be added .
first , it can be viewed as a generalisation of yoshida s main result in @xcite .
second , it opens the door to an algorithmic approach to the character variety techniques of culler , morgan and shalen in general , and to computing boundary curves strongly detected by the character variety in particular .
third , it is proved without the use of culler - morgan - shalen theory , but merely with combinatorial and geometric arguments .
not all ideal points of @xmath16 give rise to non - trivial actions since there is a trivial , closed normal surface associated to an ideal point of the deformation variety of the whitehead link complement ( with its standard triangulation ) ; see @xcite .
it remains an open problem to determine a necessary and sufficient condition purely in terms of normal surface theory for the ideal points , where @xmath43 for all peripheral elements @xmath32 in the next main result , a sufficient condition using culler - morgan - shalen theory is given .
morgan and shalen @xcite compactified the character variety of a 3manifold by identifying ideal points with certain actions of @xmath7 on @xmath17trees ( given as points in an infinite dimensional space ) , and dual to these are codimension
one measured laminations in @xmath2 this paper has taken a different but related approach by compactifying the deformation variety with certain transversely measured singular codimension
one foliations ( a finite union of convex rational polytopes ) , and dual to these are ( possibly trivial ) actions on @xmath17trees .
the relationship between these compactifications is investigated in section [ the geometric construction ] , with focus on the original consideration of ideal points of curves and surfaces dual to bass serre trees due to culler and shalen @xcite . a surface which can be reduced to an essential surface that is detected by the character variety in the sense of @xcite
will be called _ weakly dual to an ideal point of a curve in the character variety_. [ thm : morgan - shalen ]
let @xmath0 be the interior of a compact , connected , orientable 3manifold with non - empty boundary consisting of a disjoint union of tori , and let @xmath1 be an ideal triangulation of @xmath0 with the property that all edges are essential .
let @xmath44 be a sequence approaching the ideal point @xmath45 and denote @xmath46 the associated sequence of characters .
then : 1 .
the action on @xmath25 is non - trivial if an ideal point of the character variety is approached by the sequence @xmath47 in the sense of morgan and shalen .
2 . if @xmath36 has rational cooordinate ratios and for some @xmath48 @xmath49 is unbounded , then @xmath42 is non - trivial and ( weakly ) dual to an ideal point of a curve in the character variety of @xmath0
. the reader may refer to the example in section [ sec : figure eight ] whilst reading this paper .
this example also illustrates the nicest setting : an ideal point is approached through positively oriented ideal hyperbolic triangulations .
the tetrahedra associated to the degeneration become very long and thin , and a surface develops in the thin part , along which the manifold will split apart .
this is illustrated in figure [ fig : degenerating triangulation ] .
the surface naturally inherits a cell decomposition as a spun - normal surface with respect to the ideal triangulation . in this case , one obtains a splitting of the limiting dehn
thurston surgered 3manifold along euclidean ( possibly cone ) 2manifolds into 3dimensional pieces , each of which is either a hyperbolic ( possibly cone ) 3manifold or a seifert fibered manifold ; the details are worked out in @xcite using angle structures and standard spines .
the definition of the deformation variety applies to any triangulation of a closed 3manifold @xcite or ideal triangulation of a topologically finite 3manifold , and many of the results proved in this paper go through verbatim .
examples show that the fake " ideal points of the deformation variety can bear interesting information ; e.g.if one applies the construction to the minimal triangulation of quaternionic space @xmath50 one obtains a curve with three ideal points corresponding to the three 1sided heegaard splitting surfaces even though the map to the character variety is constant .
the author thanks craig hodgson , walter neumann , steven boyer and feng luo for helpful discussions .
this research was supported by a crm / ism postdoctoral fellowship and a uq new staff research start - up fund .
throughout this paper , @xmath0 denotes the interior of a compact , connected , orientable 3manifold with non - empty boundary consisting of a pairwise disjoint union of tori .
let @xmath1 be any topological ideal triangulation of @xmath2 the _ deformation variety _
@xmath6 describes hyperbolic structures for the ideal 3simplices in @xmath0 subject to certain gluing equations .
it is birationally equivalent to thurston s _ parameter space _ , which appears for example in @xcite .
the latter requires the choice of an edge for each tetrahedron , whilst the former keeps the symmetries of the triangulation .
the ideal triangulation @xmath1 of @xmath0 consists of a pairwise disjoint union of standard euclidean 3simplices , @xmath51 together with a collection @xmath52 of euclidean isometries between the 2simplices in @xmath53 termed _
face pairings_. then @xmath54 and @xmath55 is the associated pseudo - manifold ( or end - compactification of @xmath0 ) with quotient map @xmath56 let @xmath57 be a @xmath58simplex in @xmath59 then @xmath60 may be a singular @xmath58simplex in @xmath61 and is termed a _
@xmath58singlex _ for short .
denote @xmath62 the set of all @xmath58singlices in @xmath63 an _
ideal @xmath58simplex _ is a @xmath58simplex with its vertices removed .
the vertices of the @xmath58simplex are termed the _
ideal vertices _ of the ideal @xmath58simplex .
similar for singlices .
the standard terminology of ( ideal ) edges , ( ideal ) faces and ( ideal ) tetrahedra will be used for the singlices in @xmath0 and @xmath63 by hypothesis on @xmath5 the link of each vertex in @xmath64 is a torus , so @xmath65 and @xmath66 these facts are irrelevant for much of the material below .
however , the case of spheres is treated in @xcite , and for the case of higher genus boundary components , it would be more natural to develop a theory allowing hyperideal vertices in order to capture representations whose restriction to the boundary is non - abelian .
let @xmath67 be the standard 3simplex with a chosen orientation .
suppose the edges from one vertex of @xmath67 are labeled by @xmath68 @xmath69 and @xmath70 so that the opposite edges have the same labeling .
then the cyclic order of @xmath68 @xmath69 and @xmath70 viewed from each vertex depends only on the orientation of the 3simplex .
it follows that , up to orientation preserving symmetries , there are two possible labelings , and we fix one of these labelings .
suppose @xmath71 since @xmath0 is orientable , the 3simplices in @xmath72 may be oriented coherently .
for each @xmath73 fix an orientation preserving simplicial map @xmath74 let @xmath75 and let @xmath76 be the number of edges in @xmath77 which have label @xmath78 for each @xmath79 define @xmath80 and for each @xmath81 let @xmath82 setting @xmath83 gives the _
parameter equations _ , and setting @xmath84 gives the _ hyperbolic gluing equations_. for a discussion and geometric interpretation of these equations , see @xcite .
the parameter equations imply that @xmath85 the _ deformation variety @xmath6 _ is the variety in @xmath86 defined by the parameter equations and the hyperbolic gluing equations .
four ordered points on @xmath87 determine a cross ratio by the following formula : @xmath88 if the triple @xmath89 of complex numbers satisfies the three equations @xmath90 then there is an oriented ideal hyperbolic tetrahedron @xmath91 $ ] in @xmath92 where @xmath93 and the order determines the orientation , such that @xmath94 the ideal hyperbolic tetrahedron is unique up to orientation preserving isometries , and hence congruent with @xmath95.$ ] it is termed _ positively oriented _ if @xmath96 _ flat _ if @xmath97 and _ negatively oriented _
if @xmath98 denote @xmath99 a compact core of @xmath100 this is obtained by removing a small open neighbourhood from each vertex in @xmath61 such that @xmath99 inherits a decomposition into truncated tetrahedra .
let @xmath101 be an ideal edge in @xmath2 the intersection @xmath102 is _ homotopic into @xmath103 _ if and only if there is ( 1 ) an arc @xmath104 such that @xmath105 and ( 2 ) a map @xmath106 such that @xmath107 and the restriction to the boundary is a local homeomorphism . the edge @xmath101 in @xmath0 is _ essential _ if @xmath108 is not homotopic into @xmath109 and it is _ not essential _ otherwise . for instance , if @xmath0 is the complement of a knot or link in @xmath110 then @xmath64 is simply connected . whence each edge is null - homotopic in @xmath64 but all edges may be essential in @xmath100 see figure [ fig : fig8_tets ] for an example .
denote @xmath111 the universal cover of @xmath5 and lift the ideal triangulation of @xmath0 to a @xmath7equivariant ideal triangulation of @xmath112 denote @xmath113 the compact core as above .
let @xmath114 be the space obtained from the universal cover @xmath115 of @xmath99 by attaching the cone over each connected boundary component to a point .
we then have natural inclusions @xmath116 and @xmath114 is termed the _ end - compactification of @xmath23 with respect to @xmath2 _ note that @xmath114 is also simply connected since adding cones over connected spaces does not increase fundamental group , and that there is a natural , simplicial map @xmath117 where the simplicial structure arises from lifting the decomposition of @xmath99 into truncated tetrahedra and completing them to tetrahedra when adding the cones .
the map @xmath118 is also denoted by @xmath119 since it restricts to @xmath120 it is hoped that the notation and terminology does not lead to confusion .
for instance , when @xmath0 is hyperbolic , then @xmath23 is an open 3ball and the natural compactification of this open 3ball ( without reference to @xmath0 ) is homeomorphic to the closed 3ball , whilst the end - compactification @xmath114 is @xmath23 with countably many points added .
also , as noted above , @xmath64 itself may already simply connected .
the following construction is described by yoshida @xcite , 5 .
a few details are repeated here because they are not explicitly stated in @xcite , and the reader should be able to fill the gaps in the outline below .
denote @xmath121 the @xmath122character variety of @xmath2 suppose @xmath0 is the interior of a compact , connected , orientable 3manifold with non - empty boundary consisting of a pairwise disjoint union of tori .
let @xmath1 be any topological ideal triangulation of @xmath0 with the property that all edges are essential .
then for each @xmath123 there exists a representation @xmath124 which is well - defined up to conjugacy .
moreover , the well - defined map @xmath125 is algebraic .
since each ideal edge in @xmath0 is essential , every 3singlex in @xmath114 is in fact a 3simplex . given @xmath126 every ideal 3simplex in @xmath23 inherits well defined edge invariants . as in @xcite , 5 , ( see also @xcite , 4.5 ) one can inductively define a continuous map @xmath127 with the property that @xmath128 and @xmath129 maps each ideal 3simplex in @xmath23 to a geodesic ideal tetrahedron preserving edge invariants . in a nutshell ,
one starts with an initial map of a 3simplex in @xmath114 to an ideal hyperbolic 3simplex in @xmath130 preserving edge invariants , and uses the gluing equations to see that this initial map has a unique , well - defined extension to all of @xmath114 with the desired properties .
there is a unique action of @xmath7 on @xmath13 which makes the restriction @xmath131 equivariant , and hence a unique representation @xmath132 the only choice in the construction of @xmath129 is the initial placement of a 3simplex , and a different choice results in a representation which is conjugate to @xmath133 by an orientation preserving isometry of @xmath134 so there is a well defined map @xmath135 to show that this map is algebraic , one can use a standard faithful representation of @xmath122 into @xmath136 ; see for instance @xcite . as
above , let @xmath99 be a compact core of @xmath2 each boundary torus @xmath137 @xmath138 of @xmath99 inherits a triangulation @xmath139 induced by @xmath140 let @xmath141 be a closed simplicial path on @xmath142 in @xcite , the _ holonomy _ @xmath143 is defined as @xmath144 times the product of the moduli of the triangle vertices touching @xmath141 on the right , where @xmath145 is the number of 1simplices of @xmath146 and the moduli asise from the corresponding edge labels .
at @xmath126 evaluating @xmath143 gives a complex number @xmath147 it is stated in @xcite , that @xmath148 is the square of an eigenvalue ; one has : @xmath149 this can be seen by putting a common fixed point of @xmath150 at infinity in the upper half space model , and writing @xmath151 as a product of mbius transformations , each of which fixes an edge with one endpoint at infinity and takes one face of a tetrahedron to another .
choose a basis @xmath152 for each boundary torus @xmath142 since @xmath153 is a homomorphism for each @xmath138 there is a well defined rational map : @xmath154 the closure of its image is contained in the _
@xmath122eigenvalue variety _
@xmath155 of @xcite . for cusped hyperbolic 3manifolds ,
there are a number of special facts that are well - known to follow from work in @xcite .
two are stated below to highlight the fact that the main results of this paper have interesting applications in the study of the dehn surgery components in the character variety .
let @xmath0 be an orientable , connected , cusped hyperbolic 3manifold .
let @xmath1 be an ideal triangulation of @xmath0 with the property that all edges are essential .
then there exists @xmath156 such that @xmath157 is a discrete and faithful representation .
moreover , the whole dehn surgery component @xmath158 containing the character of @xmath133 is in the image of @xmath159 as in @xcite , one can identify the discrete and faithful representation algorithmically .
one first imposes the _ completeness equation _ @xmath160 written @xmath161 in short - hand , in order to ensure that all peripheral subgroups have parabolic representations .
this gives a finite collection of points on @xmath162 the _ volume _ of @xmath163 can be defined using the lobachevsky milnor formula , and it follows from francaviglia @xcite that it suffices to identify a point of maximum volume .
there is a special case , where one can start with very little topological information ; this is thurston s method to construct hyperbolic structures using ideal triangulations .
denote @xmath164 ( resp.@xmath165 ) the set of all complex numbers with positive ( resp.non-negative ) imaginary part .
let @xmath0 be the interior of a compact , connected , orientable 3manifold with non - empty boundary consisting of a pairwise disjoint union of tori .
let @xmath1 be an arbitrary ideal triangulation of @xmath2 for each @xmath166 @xmath0 has a ( possibly incomplete ) hyperbolic structure , such that the topological ideal triangulation is isotopic to a hyperbolic ideal triangulation .
in particular , @xmath0 is irreducible and atoroidal , all edges in @xmath0 are essential , and the structure is complete if and only if the completeness equation @xmath161 is satisfied . the situation is more subtle if one considers solutions in @xmath167 the so - called _ partially flat _ ideal hyperbolic triangulations . see petronio and weeks @xcite for details .
since @xmath6 is a variety in @xmath168 bergman s construction in @xcite can be used to define its set of ideal points .
let @xmath169 the map @xmath170 defined by @xmath171 is continuous , and the _ logarithmic limit set _
@xmath11 is the set of limit points on @xmath172 of its image .
thus , for each @xmath173 there is a sequence @xmath174 in @xmath175 such that @xmath176 the sequence @xmath174 is said to _ converge _ to @xmath177 written @xmath178 and @xmath36 is called an ideal point of @xmath12 whenever an edge invariant of a tetrahedron converges to one , the other two edge invariants `` blow up '' .
thus , an ideal point of @xmath6 is approached if and only if a tetrahedron degenerates .
since the riemann sphere is compact , there is a subsequence , also denoted by @xmath179 with the property that each shape parameter converges in @xmath180 in this case @xmath174 is said to _ strongly converge _ to @xmath181 if @xmath36 has rational coordinate ratios , a strongly convergent sequence may be chosen on a curve in @xmath6 according to lemma 6 in @xcite .
let @xmath182 be a subvariety of @xmath183 defined by an ideal @xmath184 and let @xmath185=\c [ x_1^{\pm 1 } , \ldots , x_m^{\pm 1}].$ ] bergman @xcite noticed that the logarithmic limit set has the two following equivalent descriptions .
it is : \1 .
the set of @xmath186tuples @xmath187 as @xmath188 runs over all real valued valuations on @xmath185/j$ ] satisfying @xmath189 and \2 .
the intersection over all non zero elements of @xmath190 of the spherical duals of their newton polytopes ( see @xcite for details ) .
as noted in @xcite , the spherical dual of the convex hull of a finite subset @xmath191 of cardinality @xmath192 is a finite union of convex spherical polytopes .
it is the union over all @xmath193 of the set of @xmath36 satisfying the @xmath194 inequalities @xmath195 where @xmath196 ranges over @xmath197 this will be used in calculations below .
recall the description of the hyperbolic gluing equation ( [ eq : glue ] ) of @xmath198 : @xmath199 and the q matching equation @xcite , section 2.9 , of @xmath198 : @xmath200 to state the relationship between these sets of equations , let @xmath201 be the exponent matrix of the hyperbolic gluing equations , and @xmath202 be the coefficient matrix of the q matching equations .
let @xmath203 and let @xmath204 be the @xmath205 block diagonal matrix with @xmath206 copies of @xmath207 on its diagonal .
then @xmath208 and @xmath209 the deformation variety @xmath6 is not defined by a principal ideal , hence its logarithmic limit set @xmath35 is in general not directly determined by its defining equations ( see @xcite for details ) .
however , it is contained in the intersection of the spherical duals of its defining equations : @xmath210 the set @xmath211 is termed a _ tropical pre - variety_. in order to give a description of this set , consider first the intersection : @xmath212 a calculation using the equations ( [ eq : com log lim ] ) shows that @xmath213 is the set of all points which are made up of @xmath206 coordinate triples , each of the form @xmath214 @xmath215 or @xmath216 where @xmath217 ( if @xmath218 these correspond to the cases where @xmath219 @xmath220 or @xmath221 respectively . ) similarly , one obtains that each hyperbolic gluing equation gives rise to the intersection of @xmath213 with a hyperplane , leaving all @xmath222 such that @xmath223 thus , @xmath224 is the intersection of @xmath213 with the nullspace of @xmath225 [ comb : homeo ] let @xmath0 be the interior of an orientable , connected , compact 3manifold with non - empty boundary consisting of tori , and @xmath1 be an ideal triangulation of @xmath2 the set @xmath224 is homeomorphic with the projective admissible solution space @xmath14 of spun - normal surface theory . in particular , @xmath11 is homeomorphic with a closed subset of @xmath226 the set @xmath14 is the collection of all elements in the nullspace of @xmath227 with the property that at most one quadrilateral type has non zero coordinate for each tetrahedron , all coordinates are @xmath228 and their sum is equal to @xmath229 this set may be projected from the unit simplex onto the sphere of radius @xmath230 centered at the origin in @xmath231 and , for simplicity , this set is also denoted by @xmath232 the map @xmath233 takes @xmath14 to the unit sphere @xmath234 where @xmath224 is found , since the following correspondence between the @xmath235th coordinate triples holds : @xmath236 thus , if @xmath237 satisfies @xmath238 then @xmath239 furthermore , given @xmath240 one has @xmath241 the set containing @xmath242 now
@xmath243 implies that @xmath244 thus , there is a linear map @xmath245 the kernel of @xmath233 is generated by the vectors with @xmath246 in the @xmath235th triple and @xmath247 in the other positions .
it follows that different admissible solutions can not differ by an element in the kernel .
the linear map is therefore 11 .
since every element in @xmath213 has a unique inverse image under @xmath248 any @xmath249 can be taken to a normal @xmath250coordinate @xmath37 using ( [ comb : c ] ) .
thus , @xmath251 and hence @xmath252 this shows that the map is onto , and in fact , that there is a well defined inverse mapping .
the last claim follows since @xmath11 is a closed subset of @xmath242 the proof of proposition [ comb : homeo ] shows that @xmath11 is homeomorphic to a closed subset of @xmath253 in a canonical way . for @xmath254 let @xmath37 be the unique normal @xmath250coordinate such that @xmath255 if @xmath36 has rational coordinate ratios , one can associate a unique spun - normal surface @xmath42 to it as follows . by assumption
, there is @xmath256 such that @xmath257 is an integer solution , and hence corresponds to a unique spun - normal surface without vertex linking components ( see theorem 2.4 in @xcite ) .
one then requires @xmath192 to be minimal with respect to the condition that the surface is 2sided .
the properties studied below are independent of the choice of @xmath38 but it will be convenient to refer to a surface .
in this section , the derivative of the holonomy of @xcite is related to the linear functional @xmath258 of @xcite , and used to show that the boundary curves of the spun - normal surfaces in @xmath35 are the boundary curves of essential surfaces which are strongly detected by the character variety .
recall that one calls the boundary slope of an essential surface associated to an ideal point of the character variety _ strongly detected _ if no closed surface can be associated to that particular ideal point .
an _ ( embedded ) surface _ @xmath21 in the topologically finite 3manifold @xmath259 will always mean a 2dimensional pl submanifold of @xmath0 with the property that its closure @xmath260 in @xmath261 is _ properly embedded _ in @xmath262 that is , a closed subset of @xmath261 with @xmath263 a surface @xmath21 in @xmath0 is said to be _ essential _
if its closure @xmath260 is essential in @xmath261 as described in the following definition : @xcite a surface @xmath260 in a compact , irreducible , orientable 3manifold @xmath261 is said to be essential if it has the following five properties : 1 .
@xmath260 is bicollared ; 2 .
the inclusion @xmath264 is injective for every component @xmath265 of @xmath260 ; 3 .
no component of @xmath260 is a 2sphere ; 4 .
no component of @xmath260 is boundary parallel ; 5 .
@xmath260 is nonempty .
the boundary curves of @xmath266 @xmath267 are also called the _ boundary curves _ of @xmath268 it will be convenient to have a simplicial definition of the functional @xmath270 defined in @xcite , section 3.1 .
let @xmath271 be a triangle in the induced triangulation @xmath139 of @xmath142 if @xmath272 are the vertices of @xmath271 in clockwise ordering ( as viewed from the cusp ) , and @xmath273 is the @xmath250modulus of @xmath274,$ ] and @xmath275 is the @xmath250modulus of @xmath276,$ ] then define the _
@xmath250modulus _ of @xmath188 to be @xmath277 ( with respect to @xmath271 ) .
e.g. if @xmath271 is the triangle of figure [ fig : mu and nu](a ) , then the @xmath250modulus of the vertex with label @xmath278 is @xmath279 with respect to @xmath280 note that for a fixed vertex , the sum of moduli with respect to all triangles containing it is equal to the corresponding @xmath250matching equation .
if @xmath141 is an oriented simplicial path , let @xmath281 be the sum of @xmath250moduli of vertices of triangles touching @xmath141 to the right .
if @xmath282 is the right hand boundary component of a small regular neighbourhood of @xmath146 oriented in the same way as @xmath146 then @xmath283 ; hence put @xmath284 let @xmath20 be a solution to the @xmath250matching equations , and @xmath285 be the evaluation of @xmath286 at @xmath287 [ @xcite , lemma 3.1 ] [ boundary : homo ] let @xmath20 be a solution to the @xmath250matching equations .
the number @xmath288 depends only on the homotopy class of @xmath141 and defines a homomorphism @xmath289 let @xmath290 be the coefficient ( row ) vector of @xmath291 and @xmath292 be the exponent ( row ) vector of @xmath293 where @xmath294 and @xmath295 are the respective coordinate systems . then : [ lem : nu&mu ] @xmath296 it it sufficient to verify the relationship for vertex moduli , and hence for the vertex labels of the triangle in figure [ fig : mu and nu ] .
if @xmath141 touches the vertex with label @xmath278 to the right , then the contribution to @xmath143 is @xmath68 and the contribution to @xmath270 is @xmath297 similarly , @xmath69 corresponds to @xmath298 and @xmath70 to @xmath299 describing this relationship for coordinate triples using @xmath207 gives : @xmath300 this proves the lemma .
the following proposition only concerns the boundary curves of a spun - normal surface , and its proof _ uses _ the known results from @xcite .
section [ the geometric construction ] gives an independent proof of the fact that the non - compact spun - normal surfaces involved are indeed non - trivial .
[ comb : essential prop ] let @xmath0 be the interior of a compact , connected , irreducible , orientable 3manifold with non - empty boundary consisting of tori , and let @xmath1 be an ideal triangulation of @xmath0 with the property that all edges are essential .
let @xmath39 be a point with rational coordinate ratios , and assume that @xmath42 is not closed .
then there is an essential surface in @xmath0 dual to an ideal point of a curve in the character variety of @xmath0 which has ( up to projectivisation ) the same boundary curves as @xmath301 moreover , these boundary curves are strongly detected .
fix a basis @xmath152 for each boundary torus @xmath302 of @xmath5 and denote the resulting coordinates for @xmath155 by @xmath303 let @xmath39 be an ideal point with rational coordinate ratios .
lemma 6 of @xcite provides a curve @xmath99 in @xmath16 such that @xmath304 moreover , there is @xmath305 such that @xmath306 defines a normalised , discrete , rank 1 valuation @xmath307 on @xmath308 : @xmath309 indeed , one may choose @xmath310 such that the above is an integer valued vector whose entries have no common divisor .
then the normal @xmath250coordinate of @xmath42 is @xmath311 or @xmath312 therefore denote @xmath306 and @xmath313 by @xmath36 and @xmath314 respectively . for each boundary torus @xmath315 and each @xmath316
there are @xmath317 such that : @xmath318 recall that @xmath319 and @xmath320 using this and lemma [ lem : nu&mu ] , one has : @xmath321 thus , restricted to the boundary , @xmath322 and in particular , since @xmath42 is not closed , the eigenvalue of at least one peripheral element blows up .
this implies that the restriction @xmath323 is not constant , and its image is therefore a curve @xmath324 denote the ideal point of @xmath325 corresponding to @xmath36 by @xmath326 a corresponding normalised , discrete , rank 1 valuation @xmath327 of @xmath328 at @xmath329 is obtained as follows .
if @xmath330 contains a pair of coprime integers , define @xmath331 and @xmath332 otherwise , let @xmath333 denote the greatest common divisor of the entries in ( [ comb : val ] ) , and then define @xmath334 and @xmath335 in either case , one obtains a valuation of @xmath328 with the desired properties .
culler - shalen theory can be applied using @xmath329 and @xmath336 lemma 14 of @xcite yields that the projectivised boundary curves of an essential dual surface are given by : @xmath337\\ = & [ v_\xi ( \mu\l_1),-v_\xi(\mu\m_1),{\ldots } , v_\xi ( \mu\l_h ) , -v_\xi(\mu\m_h)]\\ = & [ -v_{n(\xi ) } ( \l_1 ) , v_{n(\xi)}(\m_1),{\ldots } , -v_{n(\xi ) } ( \l_h ) , v_{n(\xi)}(\m_h ) ] . \end{aligned}\ ] ] this proves the claim since the latter gives the projectivised boundary curves of @xmath42 according to equation ( 4.1 ) in @xcite .
throughout this section , let @xmath0 be the interior of a compact , connected 3manifold with non - empty boundary , and @xmath1 be an ideal triangulation of @xmath2 every element , @xmath338 of the projective admissible solution space @xmath14 of @xcite is interpreted as a transversely measured singular codimension
one foliation of @xmath2 this foliation is lifted to the universal cover @xmath23 and the leaf space is turned into an @xmath17tree , @xmath339 on which the fundamental group of @xmath0 acts by isometries ( theorem [ thm : general tree ] ) . the element @xmath18 determines a singular foliation of each ideal tetrahedron , @xmath340 as follows .
if all quadrilateral coordinates of @xmath20 supported by @xmath57 are zero , then the foliation consists of triangles and a single singular leaf _ ( butterfly ) .
_ otherwise there is a unique non - zero quadrilateral coordinate , and the foliation consists of quadrilaterals of that type , triangles and two singular leaves _
( wings)_. the two kinds of singular foliation are indicated in figure [ fig : foliation ] .
the pattern on each face is topologically the same , and determines a singular foliation thereof .
there also is a natural transverse measure , obtained as follows .
consider a _ dual spine _ in @xmath57 which is transverse to the foliation .
it naturally inherits the structure of a metric tree : it consists of two copies of @xmath341 attached at each end point of a ( possibly degenerate ) closed interval @xmath342,$ ] where @xmath58 is the maximum over all quadrilateral coordinates of @xmath20 supported by @xmath343 the metric tree then defines a unique transverse measure to the foliation .
there is a unique way to identify the foliations of adjacent tetrahedra such that the leaves and transverse measures match up across the faces .
the @xmath250matching equations ensure that the leaves close up around the edges of the triangulation , so the singular foliations of the tetrahedra glue up to give a singular foliation of @xmath0 with finitely many singular leaves .
if all coordinate ratios in @xmath20 are rational , then theorem 2.4 of @xcite implies that all leaves are closed and proper , giving a foliation by spun - normal surfaces together with finitely many singular leaves .
+ if all ends of @xmath0 are tori , the homomorphism @xmath34 determines the behaviour of the singular foliation of @xmath0 near the cusps as in @xcite , section 4.1 .
thus , if @xmath344 for some @xmath345 then the foliation near the @xmath235th cusp of @xmath0 is topologically of the form @xmath346 otherwise , the fraction @xmath347 determines the slope of the intersection of the leaves with the cusp , and hence the topology of the induced foliation of @xmath348 let @xmath20 be an admissible solution to the q matching equations , and denote by @xmath349 the transversely measured singular codimension
one foliation of @xmath0 defined by @xmath287 in each tetrahedron place the corresponding dual spine .
the universal cover @xmath23 is given the ideal triangulation induced by @xmath4 so that the covering map @xmath350 is simplicial and @xmath7 acts simplicially on @xmath112 then @xmath349 lifts to a transversely measured singular codimension one foliation @xmath22 of @xmath23 which is invariant under the group action .
the leaf spaces @xmath351 and @xmath24 can be obtained by introducing an equivalence relation on the disjoint union of all dual spines , which is generated as follows .
each point on a dual spine corresponds to a leaf , and the identification of two ideal 3simplices along a common face induces identifications of their dual spines according to how the leaves and the transverse measures match up across the face , as shown in figure [ fig : dual_cases ] .
indeed , the identification of points on adjacent dual spines coming from a face is uniquely determined by the way the ideal vertices match up : they only occur along the spines minus the interior of the intervals @xmath352 corresponding to the ideal vertices not belonging to the face .
the results are of the shape of a @xmath353 each of whose ends corresponds to an ideal vertex of the face . across the face
, the singularities of the corresponding @xmath354 s match up , and the ends are identified according to the face pairing and the transverse measure .
the foliation induces a pseudo
metric on @xmath355 which descends to a pseudo
metric on @xmath356 any path @xmath141 in @xmath23 inherits a measure @xmath357 which is the total mass of the transverse measure along it .
the measure of a path contained in a leaf is zero . call a path @xmath141 _ admissible _
if it is made up of pieces each of which is either contained in a leaf of @xmath22 or transverse to @xmath358 any path is homotopic to an admissible path by a homotopy which does not increase the measure . for @xmath359 define @xmath360 where
the infimum is taken over all admissible paths from @xmath361 to @xmath362 the pseudo
metric @xmath363 on @xmath23 descends to a pseudo
metric @xmath364 on @xmath356 two points @xmath365 are equivalent , written @xmath366 if @xmath367 denote by @xmath25 the space @xmath368 with the induced metric .
then @xmath24 is isomorphic to @xmath25 if and only if @xmath364 is a metric , and this is equivalent to @xmath24 being hausdorff .
since the foliation and the transverse measure are invariant under the action of @xmath7 by deck transformations on @xmath355 there is an induced isometric action of @xmath7 on @xmath369 this is the unique action which makes the quotient map @xmath370 @xmath7equivariant and continuous .
[ thm : general tree ] let @xmath0 be the interior of a compact , connected 3manifold with non - empty boundary , and @xmath1 be an ideal triangulation of @xmath2 for each admissible solution @xmath20 of the @xmath250matching equations , @xmath25 is an @xmath17tree on which @xmath7 acts by isometries .
it remains to show that @xmath25 is an @xmath17tree .
the argument given is due to frdric paulin . according to @xcite it is necessary and sufficient to show that @xmath25 is connected and @xmath247hyperbolic .
let @xmath371 then there are points @xmath372 corresponding to @xmath361 and @xmath362 since @xmath23 is connected , there is an admissible path from @xmath373 to @xmath374 this path descends to a path in @xmath339 showing that it is connected .
now @xmath25 is @xmath247hyperbolic if for all @xmath375 one has @xmath376 regard indices in the following as integers modulo four .
let @xmath377 and choose preimages @xmath378 by definition , given @xmath379 there are admissible paths @xmath380 connecting @xmath381 to @xmath382 with the property that @xmath383 since @xmath23 is simply connected , there is a pl map @xmath384 from the disc @xmath385 to @xmath23 such that the preimages @xmath386 of @xmath381 are ordered according to indices on the boundary of @xmath385 and such that the subarc @xmath387 on @xmath388 connecting @xmath386 to @xmath389 is mapped injectively onto @xmath390 without loss of generality , it may be assumed that @xmath384 is in general position , so that @xmath349 pulls back to a transversely measured singular foliation @xmath391 of @xmath392 the disc @xmath385 will be viewed as a square with sides @xmath393 the inequality ( [ eq:0-hyp ] ) compares the measures of opposite sides to the measure on the diagonals .
the poincar recurrence theorem assures that there are at most finitely many singularities in @xmath394 a leaf transverse to @xmath388 can not accumulate in the interior of @xmath395 and the measure of all points on @xmath387 which are contained on leaves having two endpoints on @xmath387 is bounded above by @xmath396 thus , @xmath387 can be divided into three subarcs with the property that ( apart from the set of leaves going back to @xmath387 ) each leaf meeting the first ( second , third ) subarc has an endpoint on @xmath397 ( @xmath398 @xmath399 ) .
note that if the measure of leaves connecting @xmath387 to @xmath400 is non zero , then the measure of leaves connecting @xmath397 to @xmath399 is zero .
the inequality ( [ eq:0-hyp ] ) now follows , since @xmath401 was arbitrary .
if @xmath21 is a 2sided surface in @xmath5 then there is a well defined _ dual simplicial graph _ @xmath402 associated with @xmath403 whose edges and vertices are in bijective correspondence with the components of @xmath21 and @xmath404 respectively ( see @xcite , section 1.4 ) .
there are a retraction @xmath405 and an inclusion @xmath406 implying that @xmath402 is connected and that @xmath407 is isomorphic to a subgroup and a quotient of @xmath408 let @xmath409 and denote the dual graph associated with @xmath410 by @xmath411 it follows that @xmath412 is trivial , so @xmath413 is a simplicial tree .
moreover , the action of @xmath7 on @xmath23 induces a simplicial action without inversions on @xmath414 this is the unique action which makes the map @xmath415 @xmath416equivariant .
let @xmath0 be the interior of a compact , connected 3manifold with non - empty boundary , and @xmath1 be an ideal triangulation of @xmath2 let @xmath20 be an admissible integer solution of the @xmath250matching equations with the property that the ( unique ) associated spun - normal surface @xmath21 is 2sided .
then @xmath25 is a simplicial tree on which @xmath7 acts simplicially without inversions .
moreover , it is isomorphic to @xmath24 and contains the dual graph of @xmath417 as a simplicial subtree .
replace the surface @xmath21 by a surface @xmath418 which may have infinitely many boundary parallel components such that it contains infinitely many triangle discs of each type .
the dual spine determined by the corresponding solution @xmath20 of the @xmath250matching equations inherits the structure of a simplicial graph : it consists of two copies of @xmath341 attached at each end point of a ( possibly degenerate ) closed interval @xmath342,$ ] where @xmath58 is the number of normal quadrilaterals of @xmath21 in @xmath343 assume that the normal triangles meet the half
open intervals in precisely the half integer places , and that the normal quadrilaterals meet the closed interval in precisely the half integer places .
the integer places are referred to as the vertices of the spine , and each interval between two vertices as an edge .
the dual graph @xmath419 is the dual graph @xmath402 together with a copy of @xmath341 attached to each vertex corresponding to a component of @xmath404 containing an ideal vertex of @xmath140 the dual tree of @xmath417 is then a subtree of the one of @xmath420 the normal discs divide each tetrahedron into regions , one for each vertex on its dual spine .
the components of @xmath421 are a partition of @xmath355 and there is a corresponding ( unique ) equivalence relation on the set of regions giving this partition ; it is generated by identifying two regions if they meet along a face of a tetrahedron .
similar considerations hold for the spun - normal surface and its cell decomposition by normal discs . thus ,
if vertices on spines are identified whenever the corresponding regions are glued to each other , and if edges on spines are identified whenever the corresponding normal discs are glued to each other , then the dual graph @xmath422 to @xmath423 is obtained .
this is exactly the equivalence relation described in section [ sec : equivalence relation ] ; hence @xmath422 is equivariantly isomorphic to @xmath424 and this implies that @xmath24 is isomorphic to @xmath369 the following consequence is obtained by rescaling : [ cor : trees ] let @xmath0 be the interior of a compact , connected 3manifold with non - empty boundary , and @xmath1 be an ideal triangulation of @xmath2 let @xmath20 be an admissible solution of the @xmath250matching equations with rational coordinate ratios . then @xmath25 is a simplicial tree with rational edge lengths , and isomorphic to the leaf space @xmath356
the main results stated in the introduction are proved in this section .
the statements of theorem [ thm : main 1 ] are contained in corollaries [ pro : non - trivial action via peripheral ] and [ cor : thm 1 part 2 ] ; part ( 1 ) of theorem [ thm : morgan - shalen ] is in corollary [ cor : non - trivial action 2 ] , and part ( 2 ) of theorem [ thm : morgan - shalen ] is in proposition [ pro : dual surfaces ] .
let @xmath425 and fix a sequence @xmath174 in @xmath6 which strongly converges to @xmath177 so @xmath426 in the notation of section [ sec : ideal points and normal surfaces ] , and each shape parameter converges in @xmath180 denote by @xmath427 the image under the natural homeomorphism . if @xmath36 has rational coordinate ratios , denote by @xmath428 the associated 2sided spun - normal surface without boundary parallel components .
for each @xmath156 there is a map @xmath429 as in section [ sec : developing ] .
let @xmath430 denote the map corresponding to @xmath431 and @xmath432 be the corresponding representation .
the ideal triangulation of @xmath433 is lifted equivariantly to an ideal triangulation of the universal cover , so that @xmath434 where @xmath435 denotes the face pairing scheme
. as in @xcite , let @xmath436 be the topological space which is obtained from @xmath23 by imbuing each 3simplex @xmath57 of @xmath23 with the hyperbolic structure determined by @xmath437 such that all faces are identified isometrically .
thus , if @xmath438 is an ideal hyperbolic tetrahedron isometric to @xmath439 then taking the disjoint union @xmath440 together with the isometric face pairings induced from @xmath435 gives : @xmath441 note that @xmath436 may not be separable , and that there is a @xmath7equivariant proper continuous map @xmath442 it is shown in section [ gh - limit of tet ] that the gromov - hausdorff limit of @xmath443 is a so called dual spine @xmath444 where @xmath333 denotes the hyperbolic metric . the leaf space of the singular foliation @xmath349 associated to @xmath36 is described in the previous section as the set of dual spines modulo an equivalence relation : @xmath445 this leads to an interpretation of @xmath24 as a limit of the sequence @xmath446 the leaf space may not be hausdorff ; identifying any two non separable points turns it into an @xmath17tree @xmath369 there is a map @xmath447 which allows the comparison between the action of @xmath7 on @xmath25 and the limiting behaviour of the sequence of actions @xmath448 on @xmath449 the image of each ideal ( topological ) tetrahedron in @xmath23 under @xmath430 is an ideal hyperbolic tetrahedron with shape parameters in @xmath450 for each @xmath451 to simplify notation , assume that the simplices and parameter triples are ordered such that as @xmath452 the simplices @xmath453 degenerate with @xmath454 and @xmath455 do not degenerate .
a geodesic dual spine @xmath456 is constructed in @xmath438 for each @xmath457 as follows .
let @xmath57 be an ideal ( topological ) tetrahedron in @xmath355 with shape parameters @xmath458 such that @xmath459 as @xmath460 regard @xmath438 as an abstract metric space ; first with the hyperbolic metric , then with this metric suitably rescaled .
the labels @xmath461 are used for the edges of @xmath438 without reference to their specific values at @xmath462 for each edge @xmath101 with parameter @xmath68 join the centres of mass of the two faces containing @xmath101 by a geodesic arc and consider the common perpendicular to the @xmath278edges .
the rotational symmetry about this perpendicular interchanges the centres of mass of the two faces meeting in @xmath463 and hence intersects the geodesic between them .
let the portion of the perpendicular between the two intersection points be the axis of the geometric spine , and add geodesic half lines going from the endpoints of the axis to the vertices as indicated in figure [ fig : geodesic_spine ] .
note that this spine is well defined for any value of @xmath68 and denote it by @xmath464 now assume that @xmath57 is an ideal simplex in @xmath23 which does not degenerate .
then let @xmath456 consist of the four geodesic arcs in @xmath438 going from its centroid to its ideal vertices .
the following notions can be found in @xcite , chapter i.5 .
a subset @xmath465 of a metric space @xmath466 is said to be @xmath401dense if every point of @xmath466 lies in the @xmath401neighbourhood of @xmath225 an @xmath401relation between two metric spaces @xmath467 and @xmath468 is a subset @xmath469 such that the projection of @xmath470 onto each factor is @xmath401dense , and such that if @xmath471 then @xmath472 the notation @xmath473 indicates that there is an @xmath401relation between them . if @xmath474 then there is a @xmath475relation whose projection to each factor is onto .
a sequence @xmath476 of metric spaces converges to a metric space @xmath466 in the gromov
hausdorff sense if there is a sequence @xmath477 of non negative real numbers such that @xmath478 and @xmath479 let @xmath188 be an ideal vertex of an ideal triangle in @xmath480 there is a ( geodesic ) half line perpendicular to the edge opposite @xmath188 which terminates in @xmath481 the three perpendiculars thus obtained meet in a point inside the triangle which is called its _ centre of mass_. the _ geometric spine _ of an ideal triangle is defined to be the union of the rays on the perpendiculars going from the centre of mass to the ideal vertices . using the fact that all ideal triangles are congruent and that every hyperbolic triangle is contained in an ideal triangle
, one can show : the distance between the centre of mass in an ideal triangle in @xmath13 and any of its edges is equal to @xmath482 ; the geometric spine of an ideal triangle is therefore @xmath483dense . in particular , every point on a side of a triangle is within distance @xmath484 of at least one point on the other two sides .
dense if @xmath485,width=264 ] [ lem : spine dense ] let @xmath57 be an ideal ( topological ) tetrahedron in @xmath355 with shape parameters @xmath458 such that @xmath459 as @xmath460 if @xmath486 then @xmath456 is @xmath487dense in @xmath488 since @xmath438 is the convex hull of its vertices , it suffices to show that its faces are within distance @xmath487 from @xmath464 a face is in the @xmath482neighbourhood of its spine , hence if the spines of all faces are in a @xmath489neighbourhood of @xmath490 then the whole tetrahedron is in a @xmath487neighbourhood of @xmath464 to simplify notation , @xmath57 will be written instead of @xmath488 consider two faces of @xmath57 meeting in an edge @xmath491 on @xmath101 are the endpoints of the perpendiculars on the faces going into the vertices not in @xmath491 the hyperbolic distance between them is precisely @xmath492 since @xmath493 one has @xmath494 put @xmath495 for @xmath278 sufficiently close to one , @xmath496 and @xmath497 since @xmath498 and @xmath499 assume without loss of generality that @xmath500 then @xmath501 and by assumption @xmath502 label the vertices of @xmath57 by @xmath503 such that @xmath504 and @xmath505 are the endpoints of an edge @xmath101 with parameter going to one .
consider the face opposite @xmath506 and denote its centre of mass by @xmath507 it needs to be shown that the geodesic rays @xmath508,$ ] @xmath509 are within distance @xmath489 from @xmath464 denote the singularities of @xmath456 by @xmath510 and @xmath511 such that @xmath510 is on the geodesic segment joining @xmath512 to its `` opposite '' centre of mass .
let @xmath119 be a point on @xmath513 $ ] which is at most distance @xmath482 from the endpoint of the perpendicular bisector through @xmath514 on a face containing the edge @xmath515.$ ] this point exists by the previous lemma .
the following inequalities can be read off from figure [ fig : spine dense ] : @xmath516 this completes the proof for the rays @xmath508,$ ] @xmath517 by considering the hyperbolic triangles containing them and the point @xmath518 and for @xmath519 $ ] by considering @xmath520.$ ] the endpoints of the geodesic segment @xmath521 $ ] are within distance @xmath489 from the geodesic segment @xmath522,$ ] and hence this must be true for the whole segment .
this completes the proof for the particular face , and , by symmetry , for all faces .
[ lem : global constant ] there is a constant @xmath523 and @xmath524 such that for each ideal tetrahedron @xmath525 @xmath456 is @xmath523dense in @xmath438 for all @xmath526 as @xmath178 we have @xmath527 for all @xmath528 , and hence @xmath529 thus , there exists @xmath530 such that @xmath531 for all @xmath528 and for all @xmath532 lemma [ lem : spine dense ] states that in this case @xmath456 is @xmath487dense in each @xmath57 which degenerates . if all tetrahedra degenerate , then @xmath533 and @xmath534 satisfy the requirements . hence assume that @xmath535 for each @xmath536 choose an ideal hyperbolic tetrahedron of the limiting shape @xmath537 and construct the dual spine .
let @xmath385 be the maximum over the distances of the centroids to the respective singularities on the faces .
then for each limiting tetrahedron , the dual spine is @xmath538dense .
let @xmath539 since the distances of the centroids to the singularities of faces vary continuously , there exists @xmath540 such that for all @xmath541 @xmath542 and all @xmath543 the dual spine is @xmath544dense . putting @xmath545 and
@xmath546 completes the proof .
recall the definition of the dual spine .
if the parameter of @xmath57 is determined by the coordinate @xmath547 then let @xmath548 be an abstract @xmath17tree consisting of a closed interval of length @xmath549 with two copies of @xmath352 attached at each of its endpoints .
, width=491 ] [ lem : spine is gh limit of tet ] there is a null sequence @xmath550 of positive real numbers such that for each @xmath551 there is a @xmath551relation @xmath552 for each ideal 3simplex @xmath553 thus , @xmath548 is a gromov - hausdorff limit of @xmath554 let @xmath555 and @xmath457 a simplex with corresponding parameter triple @xmath556 consider the sequence @xmath557 where @xmath558 equation ( [ eq : spines ] ) can be written as @xmath559 hence @xmath560 which by assumption is a positive real number .
after possibly passing to a subsequence of @xmath561 one may therefore choose a null sequence @xmath562 such that @xmath563 letting @xmath564 one has @xmath565 label the singularities of @xmath548 by @xmath566 and @xmath567 and its four ideal vertices by @xmath568 ... ,@xmath569 denote the distance function on @xmath548 by @xmath570 a relation @xmath571 is defined as follows .
label the ideal vertices and the singularities of the dual spine of @xmath438 with reference to @xmath340 so that they are chosen consistently for all @xmath451 then define @xmath470 between @xmath572 and @xmath548 by including @xmath573 and extending the relation isometrically on the infinite ends , and by scaling linearly on the intervals between the singularities .
the projection of @xmath470 to the second factor is onto , and the projection to the first factor is @xmath574dense , and in particular @xmath575dense . since the ends are identified isometrically , the maximal length distortion occurs on paths containing both singularities .
one has : @xmath576 whence @xmath470 is a @xmath575relation between @xmath577 and @xmath578 a similar argument applies to the tetrahedra which do not degenerate .
the sequence @xmath579 converges to the leaf
space @xmath24 of the transversely measured singular codimension one foliation in the sense that : @xmath580 to show that the proposition follows from lemma [ lem :
spine is gh limit of tet ] , one needs to show that the limiting `` face pairings '' amongst the dual spines give the equivalence relation described earlier .
this follows from the following two observations .
the limit of each face is of the shape of a @xmath353 and since the gluings of faces are by isometries , two @xmath354 shapes are identified isometrically along rays corresponding to the edges of tetrahedra .
an isometric gluing identifies the centres of mass of the faces , which correspond to the vertices of the @xmath354 shapes .
hence the singularities glue up , and the identification is isometrically along the three infinite rays . the above does not show that the sequence @xmath581 converges in the gromov
hausdorff sense to @xmath24 or @xmath25 ( though this is true for arbitrarily large simplicial subsets ) , but it gives a useful device for comparing the action on @xmath25 with the limiting action on @xmath480 this is similar to the situation in @xcite .
this section continues with the notation introduced in section [ sec : outline and definitions ] . for each @xmath582 consider the following two translation lengths : @xmath583 where @xmath363 denotes the distance function on @xmath369 [ pro : length functions ] for each @xmath582 the sequence @xmath584 contains a convergent subsequence , and its limit @xmath585 satisfies : @xmath586
let @xmath587 and @xmath588 be a path from @xmath361 to @xmath589 in @xmath112 then : @xmath590 the map @xmath591 restricts to an isometry on each ideal tetrahedron , so @xmath592 and one has @xmath593 recall from @xcite that a path is called admissible if it consists of finitely many sub
paths each of which is either contained in a leaf or is transverse to the foliation , and that the distance between two points in @xmath25 is the infimum over the measures of all paths connecting them ; the transverse measure is denoted by @xmath594 any path can be deformed into an admissible path without increasing the measure .
let @xmath595 and assume that @xmath596 is a preimage of @xmath597 then for any @xmath598 there is an admissible path @xmath588 in @xmath23 joining @xmath361 and @xmath599 such that @xmath600 the path can be subdivided into @xmath601 sub
paths for some @xmath602 such that each of them is contained in an ideal tetrahedron . according to lemma [
lem : spine is gh limit of tet ] , for some @xmath603 and all @xmath604 and all @xmath525 there is a @xmath551relation between @xmath438 and @xmath548 with @xmath605 this relation extends to a surjective @xmath606relation , and the properties of the gromov
hausdorff limit imply that : @xmath607 giving @xmath608 since @xmath598 is arbitrary , inequality ( [ eq : proof of transl length ] ) implies : @xmath609 where the infimum is taken over all paths from @xmath361 to @xmath610 hence the sequence @xmath584 is bounded , and since @xmath611 is arbitrary , the conclusion follows .
the above proposition gives a sufficient condition ( @xmath612 ) for the action on @xmath25 to be non - trivial , and a necessary condition ( @xmath613 ) for the action on @xmath25 to be trivial .
this is first translated into a condition only involving information from spun - normal surface theory , and then into a general statement linking the ideal points of the deformation variety to the compactification of the character variety due to morgan and shalen .
[ pro : non - trivial action via peripheral ] let @xmath39 and @xmath614 assume that for some peripheral element @xmath582 @xmath615 then the action of @xmath7 on @xmath25 is non - trivial .
let @xmath32 the hyperbolic translation length @xmath616 of @xmath617 satisfies @xmath618 the relationship between the growth rates of parameters and the valuation of squares of eigenvalues of peripheral elements described in the proof of proposition [ comb : essential prop ] gives that @xmath41 implies @xmath619 the result now follows from proposition [ pro : length functions ] .
[ cor : non - trivial action 2 ] let @xmath620 @xmath621 and @xmath622 be a sequence strongly converging to @xmath181 the action of @xmath7 on @xmath25 is non - trivial , if an ideal point of the character variety is approached by the sequence @xmath623 if the sequence @xmath623 approaches an ideal point of the character variety , i.e.some trace becomes unbounded , we need to show that the scaling factors @xmath624 are not too big as to render @xmath30 trivial .
this follows from the facts that @xmath623 is a rational function in the shape parameters , and that rescaling any shape parameter tending to @xmath247 or @xmath625 by @xmath624 gives a non - zero number proportional to its growth rate .
it follows from the construction that if the length function @xmath30 is non - trivial , it determines the point in the morgan - shalen compactification with coordinate @xmath626 \in \mathcal{c}},$ ] where @xmath627 is the set of conjugacy classes of elements in @xmath408 throughout this section , assume that @xmath36 has rational coordinate ratios , and recall that @xmath42 is a 2sided surface . in this case @xmath628 and it can be rescaled so that it contains the _ dual ( simplicial ) tree _
@xmath413 of the lift of @xmath629 to @xmath23 as a subtree .
a surface is _ non - trivial _ if it is essential or can be reduced to an essential surface by performing compressions and then possibly discarding some components .
each vertex stabiliser of the action on @xmath413 is conjugate to the image under the inclusion map @xmath630 for some component @xmath631 of @xmath632 and each edge stabiliser to the image of @xmath633 for some component @xmath634 of @xmath268 thus , @xmath21 is non - trivial if and only if none of the images @xmath630 is onto , and this is the case if and only if the action of @xmath416 on @xmath25 is non - trivial .
[ pro : dual surfaces ] let @xmath635 be a sequence strongly converging to @xmath177 and @xmath636 1 .
the sequence @xmath637 converges in @xmath638 for each connected component @xmath99 of @xmath639 2 .
the sequence @xmath640 converges in @xmath641 to a reducible character for each connected component @xmath634 of @xmath268 in particular , if for some @xmath48 @xmath49 is unbounded , then @xmath21 is non - trivial and ( weakly ) dual to an ideal point of a curve in the character variety of @xmath0 ; hence there is a @xmath7equivariant map from @xmath25 to the bass serre tree associated to the ideal point .
corollary 2 in @xcite , i6.2 , states that a finitely generated group acting on a simplicial tree fixes a vertex if and only if all generators and their double products have zero translation length .
this together with equation ( [ eq : length functions ineq ] ) implies that @xmath642 is contained in a compact subset of @xmath643 to prove that there is a unique accumulation point , some terminology and observations are needed . the spun - normal surface @xmath21 divides each tetrahedron into several types of regions .
a region which contains two normal isotopic discs is a trivial @xmath644bundle over a normal disc and will be called a _
slab_. a region which is not a slab is a _
vertex region _ if it is a corner cut off by a triangle , otherwise it is a _
thick region_. a thick region is a truncated tetrahedron or a truncated prism .
assume that @xmath99 contains some thick region .
this region meets some face of an ideal tetrahedron @xmath57 in @xmath0 in a truncated ideal triangle .
choose a lift of this face to @xmath23 and denote it by @xmath197 define the maps @xmath645 such that @xmath646 is sent to the ideal triangle @xmath647 $ ] for all @xmath451 each region is associated to a vertex on a dual spine as in section [ subsec : dual tree ] .
let @xmath119 be the vertex corresponding to the centre of mass of @xmath197 let @xmath141 be an element of the group @xmath648 in the conjugacy class of the image @xmath649 which stabilises @xmath650 then @xmath651 is contained in @xmath652 and hence within bounded hyperbolic distance of the vertex corresponding to another region of @xmath653 say @xmath654 there is a finite sequence of adjacent regions and their vertices @xmath655 such that the regions corresponding to @xmath656 and @xmath657 are identified along a face of some tetrahedron in @xmath112 as the tetrahedra degenerate , the images of @xmath656 and @xmath657 under @xmath645 stay within bounded hyperbolic distance for all @xmath345 and hence @xmath119 and @xmath658 stay within bounded hyperbolic distance .
thus , @xmath659 and the sequence @xmath660 converges since it only depends on the limiting shapes of the tetrahedra and the combinatorial gluing data .
since @xmath141 was chosen arbitrarily , the sequence @xmath661 converges algebraically , and the conclusion follows since choosing a different developing map changes the representation by conjugation .
this proves ( 1 ) for components that contain thick regions .
a normal disc in @xmath21 is called _ bad _ if it is contained in the boundary of some thick region .
if @xmath21 contains no normally isotopic components , then each connected component of @xmath21 meets some thick region in a bad disc .
thus , if @xmath634 does not contain a bad disc , one may replace it by a normally isotopic component for the purpose of proving ( 2 ) .
associated with the thick region met by @xmath634 is a centre of mass of some face in @xmath140 define the maps @xmath645 such that a fixed lift of this face is sent to the ideal triangle @xmath647 $ ] for all @xmath451 similar to the above , using the edges in dual spines associated to @xmath662 one sees that @xmath663 converges algebraically .
since the limiting representation of @xmath664 fixes a vertex of @xmath647,$ ] it is reducible .
whence ( 2 ) holds . to conclude the proof of ( 1 ) , suppose that the component @xmath99 contains no thick region . if it contains a vertex region , then it must be entirely made up of vertex regions , so its boundary is a vertex linking surface ; @xmath21 contains no such component .
if it contains a slab region , then it must be entirely made up of slabs , in which case it is either a trivial or a twisted @xmath644bundle over a spun - normal surface , and @xmath665 for some @xmath666 both cases follow from ( 2 ) ; the first trivially , the second from the fact that @xmath667 is uniquely determined by @xmath668 since @xmath669=2 $ ] and @xmath670 for each @xmath671 the last part of the proposition follows directly from @xcite , lemma 6 , and @xcite , section 3 .
the former shows that the sequence @xmath635 may be chosen on a curve in @xmath162 the sufficient condition ( @xmath612 ) for the action to be non - trivial , together with ( 1 ) and ( 2 ) , allows the construction in @xcite to be applied to yield the assertion .
note that the map to the bass serre tree may not be an isometry ( even up to scaling and restriction to maximal invariant subtrees ) .
if the above necessary condition for a trivial action ( @xmath672 ) is not sufficient , then there may be closed essential surfaces which are detected by the deformation variety , but not by the character variety .
the proofs of the main results are completed with the following immediate consequence of corollary [ pro : non - trivial action via peripheral ] and proposition [ pro : dual surfaces ] .
[ cor : thm 1 part 2 ] if @xmath42 is spun normal , then @xmath42 is non - trivial and ( weakly ) dual to an ideal point of a curve in the character variety .
let @xmath0 denote the complement of the figure eight knot .
an ideal triangulation of @xmath0 is shown in figure [ fig : fig8_tets ] .
the quadrilateral types dual to @xmath673 and @xmath674 will be denoted by @xmath675 and @xmath676 respectively .
one has the following hyperbolic gluing equations : @xmath677 this gives the matrix @xmath678 hence @xmath679 which determines a single @xmath250matching equation : @xmath680 this agrees with @xcite .
one can also work out the induced triangulation of the torus end and determine standard generators for the peripheral subgroup , giving @xmath681 and @xmath682 the set @xmath253 is @xmath247dimensional , its elements are scaled to integer solutions and listed in table [ tab : surfaces fig8 ] .
all spun - normal surfaces corresponding to these solutions are once punctured klein bottles . since @xmath683 its follows from the symmetries that @xmath684 a presentation of the fundamental group can be worked out from the triangulation , and simplifies to : @xmath685 furthermore , @xmath686 is a meridian , and the corresponding longitude is @xmath687 let a fundamental domain for @xmath0 be given by embedding the tetrahedra in @xmath13 as indicated by the vertex labels in figure [ fig : fig8_tets ] .
the face pairings associated to @xmath688 are defined by assignments of the following ordered triples : @xmath689 \to [ \infty , 1,z]\\ \m_z & : [ \infty , 0,zw ] \to [ 1,0,z]\\ g_z & : [ \infty , zw , z ] \to [ \infty , 0,1]\end{aligned}\ ] ] it follows that @xmath690 from above face pairings and the equation @xmath691 one obtains representations into @xmath122 by putting @xmath692 this is in fact a representaton into @xmath693 and the lower right entries in @xmath686 and @xmath694 correspond to square roots of the holonomies given by thurston in @xcite , where @xmath695 and @xmath696 the image of @xmath6 in the @xmath122eigenvalue variety is parameterised by : @xmath697 where @xmath698 and @xmath699 the map from @xmath6 to the @xmath122character variety can also be determined from the above .
one has @xmath700 and @xmath701 putting @xmath702 and @xmath703 the image of @xmath6 in the @xmath122character variety is parameterised by : @xmath704 giving a sphere in @xmath705 the smooth projective model of @xmath6 is a torus , and it can be verified that the map @xmath706 is generically 2to1 . given the symmetries of @xmath0 and its the triangulation , it suffices to consider a degeneration of the ideal triangulation to one of the four ideal points of @xmath12 the point whose associated spun - normal surface coordinates are @xmath707 will be chosen , and `` geometric '' degenerations , i.e. degenerations where both tetrahedra stay positively oriented and only in the limit become degenerate will be studied .
the deformation variety is birationally equivalent to the variety in @xmath708 defined by the single equation : @xmath709 solving the above equation in terms of @xmath710 gives @xmath711 at the complete structure , one has @xmath712 ; hence take the solution for @xmath278 with positive sign in front of the root .
the desired ideal point corresponds to the degeneration @xmath713 note that then @xmath714 at half the rate .
let @xmath715 and obtain a power series expansion of @xmath278 for @xmath192 around zero using equation ( [ fig8:solve ] ) .
this is of the form @xmath716 where @xmath717 all points on the path @xmath718 given by @xmath719 correspond to geometric solutions to the hyperbolic gluing equations ( see @xcite ) .
the face pairings can be used to determine the limiting representations . as @xmath720 one has the following limiting traces : @xmath721 whilst the eigenvalue of @xmath722 which is equal to @xmath723 approaches one .
thus , @xmath724 is a strongly detected boundary slope .
the limiting splitting of @xmath0 corresponds to a splitting of @xmath725 which is a graph manifold , along an essential torus ( see @xcite ) .
the limiting pieces admit seifert fibered structures , and are a twisted @xmath644bundle over the klein bottle , @xmath726 and a trefoil knot complement , @xmath727 note that in this case passing to an orientable surface introduces a certain redundancy ; splitting along the klein bottle gives a nicer decomposition .
the fundamental groups of the complementary pieces can be worked out from figure [ fig : fig8_klein ] .
let @xmath728 be the punctured klein bottle in @xmath0 shown in figure [ fig : fig8_klein ] , and let @xmath729 be the corresponding klein bottle in @xmath730 identify @xmath731 with a regular neighbourhood of @xmath729 in @xmath730 standard generators for @xmath732 are @xmath733 and @xmath734 one has @xmath735 standard generators for the boundary torus of @xmath731 are @xmath736 and @xmath737 generators for the complement of @xmath731 in @xmath738 are given by @xmath739 and @xmath740 one has @xmath741 generators for the boundary torus of the trefoil knot complement are @xmath742 and @xmath743 ( the meridian is standard , but the longitude is not ) . the limiting representation on the trefoil knot complement is determined by : @xmath748 this representation corresponds to a 2dimensional hyperbolic structure on the base orbifold of @xmath749 which is a @xmath750turnover . the limiting representation of the boundary torus is : @xmath751 in order to obtain a ( finite ) limiting representation of @xmath726 one needs to conjugate the face pairings by a diagonal matrix with eigenvalues @xmath752 the limiting representation is then : @xmath753 giving a cyclic group of order two in @xmath10 the limiting image of @xmath754 is @xmath755 and the limiting image of @xmath756 is @xmath757 the topology of the leaves in the singular foliation @xmath349 associated to the above ideal point can be determined from the spun - normal surface data .
one leaf is the once punctured klein bottle made up of three quadrilaterals and infinitely many triangles ; the single singular leaf is made up of the four singular pieces plus one quadrilateral and infinitely many triangles removing the triple curve yields a once punctured mbius band ; all other leaves are twice punctured tori made up of six quadrilaterals and infinitely many triangles . an explicit description of @xmath349 was found by thurston @xcite , and is given in the following commentary to figure [ fig : fol fig8 ] .
thickening the surface shown in figure [ fig : fig8_klein ] yields a genus two handlebody @xmath758 in @xmath759 with the figure eight knot lying on its boundary , and @xmath758 minus the knot is foliated by one once punctured klein bottle and parallel twice punctured tori .
it remains to describe the foliation of the complementary handlebody @xmath760 minus the knot .
a small neighbourhood of the singular curve on the singular leaf in @xmath349 is shown in figure [ fig : fol fig8 sing curve ] .
the singular curve loops through the two holes of @xmath761 near these holes , the foliation is as shown in figures [ fig : fol fig8 b_1 ] and [ fig : fol fig8 1-handle ] . so attaching a small 1handle to each hole of @xmath758 which meets @xmath349 in the shown product gives
a ball @xmath762 in @xmath110 with the figure eight knot contained partly on its boundary , partly in its interior .
the foliation of the complementary ball @xmath763 minus the knot is shown in figures [ fig : fol fig8 b_2 sing ] and [ fig : fol fig8 b_2 non sing ] .
there is one singular leaf consisting of two boats attached along their keels , and the remaining leaves are discs .
any non singular leaf contained in @xmath760 meets @xmath763 in exactly one discs in each of the four components of @xmath763 minus the singular leaf , and @xmath762 in six rectangles ; one in each component of the 1handles minus the singular leaf .
the foliation of @xmath760 minus the knot by one singular leaf and parallel twice punctured tori is now obtained by appropriately attaching two foliated 1handles to @xmath764 roger alperin , hyman bass : _ length functions of group actions on @xmath765-trees _ , in combinatorial group theory and topology ( alta , utah , 1984 ) , 265378 , ann . of math .
, 111 , princeton univ . press , princeton , nj , 1987 .
walter d.neumann : _ combinatorics of triangulations and the chern - simons invariant for hyperbolic @xmath766-manifolds _ , topology 90 ( columbus , oh , 1990 ) , 243271 , ohio state univ .
publ . , 1 ,
de gruyter , berlin , 1992 . | let @xmath0 be a cusped 3manifold , and let @xmath1 be an ideal triangulation of @xmath2 the deformation variety @xmath3 a subset of which parameterises ( incomplete ) hyperbolic structures obtained on @xmath0 using @xmath4 is defined and compactified by adding certain projective classes of transversely measured singular codimension
one foliations of @xmath2 this leads to a combinatorial and geometric variant of well known constructions by culler , morgan and shalen concerning the character variety of a 3manifold . |
solid water is a very abundant material .
indeed , it is believed to be the most abundant condensed material in the universe , thanks to its propensity for remaining as a deposit on interstellar dust particles in dense clouds .
amorphous solid water ice , together with a wealth of other condensed species , is widely observed spectroscopically via infra - red absorption lines present in the spectrum of a field star or of the embedded object in star forming regions @xcite .
icy mantles covering dust grains are reputed to form after direct freeze out of gas phase species and after surface reactions of atoms and radicals .
although water ice is by far the most abundant species in icy mantles , the chemical origin of water molecules in interstellar icy mantles is still uncertain .
water molecule formation in the gas phase is not efficient enough to reproduce the observed abundances in dark clouds , especially in its solid form @xcite .
therefore , water ice is likely to form directly on the cold interstellar grains and not as a condensate after formation in the gas phase .
a reaction scheme including water formation on grain surfaces was proposed some years ago by @xcite .
they suggested that h@xmath2o formation would be initiated by h - atoms reacting with o , o@xmath2 and o@xmath3 , although the o@xmath3 + h pathway was considered to be the most effective and o@xmath2 would play a mere catalytic role .
nonetheless , o@xmath1 and o@xmath0 molecules have not so far been detected in interstellar ices on dust grains , although their presence is presumed .
ozone non - detection may be partly explained because its strongest @xmath4 ( o - o ) vibrational transition band at 1040 @xmath5 ( 9.6 @xmath6 m ) is swamped by the strong infrared absorption band of silicates at 10 @xmath6 m which has been observed by iso ( infrared space observatory ) in the 2.4 - 25 @xmath6 m region ( e.g. , spectrum toward the protostar w33a , resolution @xmath7 500 - 1000 , * ? ? ?
astrochemical models of dark clouds predict that condensed oxygen is likely to be a major component of apolar dust grains within interstellar clouds at 10 k. recently , the o + h pathway @xcite and the o@xmath1 + h pathway @xcite have been explored experimentally .
all these experiments show that these reactions lead to water formation , even though the efficiencies of intermediate reactions are still under debate . besides o and o@xmath1 , ozone
too can be present in icy mantles chemically processed by irradiation from galactic cosmic - ray particles and internal secondary uv field @xcite .
several laboratory works confirm the efficient production of o@xmath0 following energetic electron irradiation of pure oxygen ices @xcite , and uv , proton and ion irradiation of ice mixtures simulating realistic analogues of interstellar ices @xcite . in solid phase , o@xmath1 and co@xmath1
are easily dissociated by uv radiations ( @xmath8 nm ) and cosmic rays ( c.r . ) producing atomic o in the electronic ground state ( @xmath9p ) and in the first electronically excited ( @xmath10d ) state @xcite : @xmath11 @xmath12 in the interstellar grain mantles , the o(@xmath9p ) atoms in the ground state , which are in thermal equilibrium with the ice surface , react with o@xmath1 molecules to form o@xmath0 via the following reaction pathway : @xmath13 the formation of ozone via equation ( 3 ) is exothermic ( @xmath14 ev ) .
while the main ozone destruction reactions are : @xmath15 @xmath16 in addition , in dark clouds hydrogen is mainly in its molecular form so h - atoms are a rather rare reactant with h / h@xmath17 @xmath18 @xcite . the number density of h - atoms is mostly governed by the destruction of h@xmath17 due to cosmic rays .
this value , whatever the density of the cloud , is about 1 @xmath19 .
the o / h@xmath17 ratio remains approximately constant ( 10@xmath20 ) , thus the number of atomic o , unlike h , is proportional to the density of the cloud .
see , for example , table 1 of @xcite ; for a cloud density of 10@xmath21 @xmath19 the h / o ratio is @xmath22 while for a denser cloud with a density of 10@xmath23 @xmath19 the h / o ratio is @xmath24 .
therefore , for very dense clouds , o is the most abundant species in atomic form and it can accrete on grains and subsequently form o@xmath1 and o@xmath0 . in their model , using a monte carlo approach , @xcite showed that under dense cloud conditions the ( o@xmath0 + h ) pathway is the most efficient route leading to the formation of h@xmath1o .
finally , it should be noted that , in the diverse models , some barriers can be raised in order to increase one of the competitive reactions @xcite . yet
, such barriers can only be experimentally probed , because extrapolations from the gaseous data are quite uncertain .
this letter is the first report of water formation via ozone hydrogenation in the series of recent experiments that aim to investigate water synthesis in the ism .
we present the evidence for the formation of water molecules from the reaction of o@xmath0 with d - atoms on non - porous amorphous solid water ice under conditions relevant to the dense interstellar clouds .
our experiments were performed using the formolism set - up developed to study the reaction and the interaction of atoms and molecules on surfaces simulating dust grains under interstellar conditions .
the apparatus is composed of an ultrahigh - vacuum chamber with a base operating pressure of around 10@xmath25 mbar .
the sample holder , a copper cylinder block , is attached to the cold finger of a closed - cycle he cryostat .
it can be cooled to 8 k. the temperature of the sample is measured using a calibrated silicon diode and a thermocouple ( cromel - aufe ) clamped on the back of the sample holder .
the copper surface is covered with 100 layers of non - porous amorphous solid water ( np - asw ) by condensation of water vapor on the cold sample maintained at 120 k. water is sprayed through a microchannel array doser located 2 cm in front of the surface .
this ice is amorphous but non porous @xcite .
the water ice substrate is then cooled to 10 k. more details can be obtained in @xcite .
ozone can be synthesized by two methods .
one is an ex - situ method : o@xmath0 is prepared from gaseous o@xmath1 introduced in a glass bottle at a pressure of about 30 mbar .
the o@xmath1 gas is then excited by a radiofrequency electric discharge ( @xmath7 2 mhz ) through a copper coil placed around the o@xmath1 container .
the blue ozone gas produced by the recombination of o@xmath1 and o atoms during the discharge is condensed on the bottom wall plunged in liquid nitrogen ( 77 k ) .
residual o@xmath1 is then removed by primary pumping .
the glass bottle is then mounted onto the main set - up .
ozone is introduced into the vacuum chamber via a triply differentially pumped beam line aimed at the cold surface covered with h@xmath1o ice . due to the metallic parts of the inlet system
on which o@xmath0 in part dissociates , the beam is composed of 70 % ozone and 30 % o@xmath1 . however , o@xmath1 is easily removed from the surface when the sample is held at 50 k. by varying the exposure time of the ozone beam , it is possible to vary the initial coverage of the surface .
the o@xmath0 flux was previously calibrated using temperature programmed desorption ( tpd ) by determining the exposure time of a pure o@xmath1 beam required to saturate the o@xmath1 monolayer ( 1 monolayer = 10@xmath26 molecules @xmath27 ) on np - asw ice . in the experiments described below we have deposited a maximum of half a monolayer of ozone under the same beam conditions used for calibrating the o@xmath1 flux .
owing to the non - pure o@xmath0 beam , however , ozone exposures had a relative accuracy of @xmath7 20 % and an absolute accuracy of @xmath7 30 % .
nevertheless , we did not see any significant effect due to the initial coverage ( between 0.1 and 0.5 monolayers @xmath28 ) .
an alternative technique of ozone production was also used , namely by depositing an atomic o on top of an o@xmath1 layer pre - adsorbed on the surface and then producing ozone in - situ .
the disadvantage of this technique is that the control of the initial amount of ozone have to be checked after each exposure .
whichever technique was used to deposit ozone onto the surface , we obtained the same final results . after depositing ozone ,
an atomic d - beam aimed at the surface of the sample irradiates the o@xmath0-h@xmath1o film maintained at 10 k. the d - atoms are produced by the dissociation of d@xmath1 molecules in a quartz tube using a surfatron microwave discharge at 2.45 ghz .
the rate of dissociation is @xmath7 60 % .
after the irradiation of @xmath29 d - atoms @xmath27 , a tpd is performed up to 180 k so that all the water ice desorbs .
the tpd heating rate is 10 k / min from 10 to 180 k. the desorbed species are detected using a quadruple mass spectrometer ( qms ) .
the spectrum ( a ) in figure [ fig1 ] shows the tpd curve of mass 32 obtained after the exposure of half a layer of o@xmath0 on a np - asw substrate held at 10 k ( ex - situ synthesis ) .
we observe two peaks : the desorption peak at low temperature ( @xmath7 30 k ) represents the desorption of o@xmath1 molecules .
these molecules have been deposited by the o@xmath1 fraction in the o@xmath0 beam , probably due to ozone destruction on the metallic walls of the gas inlet system .
the high - temperature peak at @xmath7 60 k is the desorption of ozone .
mass 48 of o@xmath0 is also simultaneously detected in smaller amount and is always proportional to the signal at mass 32 .
in fact , for the most part o@xmath30 fragments are detected , as a result of the o@xmath0 cracking in the head of the qms .
also , the desorption of ozone from asw surfaces is known to occur at around 60 k @xcite .
although this may be seen as an indirect measure , mass 32 gives information about o@xmath1 ( below 50 k ) and o@xmath0 ( above 50 k ) desorption .
as shown in the spectrum ( a ) of fig . 1 , o@xmath0 desorption is about three times larger than that of o@xmath1 .
this indicates that the ozone fraction deposited with the ex - situ method is @xmath7 70 % , the rest being o@xmath1 molecules . in order to deposit only ozone molecules on the cold surface
, we performed the exposure step with the water ice substrate held at 50 k. thus o@xmath1 molecules desorbing at 30 k were eliminated .
then we cooled the sample to 10 k before proceeding with the tpd .
as shown in the spectrum ( b ) of fig . 1 , the tpd peak observed comes from o@xmath0 desorbing at 60 k. by using these procedures , we are sure that at 10 k only reactions involving ozone and d - atoms will occur . for this work , we performed a set of experiments with various d - atom exposures . in the first one , d - atoms were sent onto the ozone - free water ice surface to check that no reaction occurred between d and the h@xmath1o substrate to produce deuterated water molecules .
this result is displayed in the mass 19 spectrum ( b ) of fig .
2 . there is neither hdo nor d@xmath1o desorption peak when d - atoms alone are deposited on the asw ice substrate at 10 k as was previously noticed by @xcite and @xcite . in a second experiment we exposed the np - asw sample held at 10 k to d - atoms before ozone deposition . in this experiment ozone
desorbs as if no atoms had been sent to the surface .
this is because d - atoms form promptly d@xmath1 on non - porous water ice ( e.g. , * ? ? ?
* ) and apparently because ozone does not react with d@xmath1 ( see below ) .
this proves that hydrogenation reactions take place at around 10 k and not at higher temperatures since d - atoms react promptly at 10 k and furthermore they will have completely desorbed above 13 k @xcite .
in another experiment , d - atoms were sent onto the np - asw surface previously exposed to 0.5 monolayers of o@xmath0 . in the spectrum ( c ) of figure 1
we firstly observe the disappearance of the o@xmath0 peak .
this suggests that all the ozone molecules have reacted .
we also note that the reaction d + o@xmath0 is very efficient , because the number of d - atoms sent to the surface is only 10 times the number of o@xmath0 on the surface .
consequently , we can conclude that the d + o@xmath0 reaction which is in competition with the d + d reaction in our experiment proceeds on the ice surface with a small energy barrier .
an estimation of the reaction rate can be made , for example , with the formula used by @xcite to simulate the surface reactions in their model : @xmath31 where @xmath32 is the attempt frequency for reaction , @xmath33 the width of the barrier , @xmath6 the reduced mass and @xmath34 the activation barrier . by taking @xmath34/k@xmath35 = 450 k for the d + o@xmath0 reaction @xcite
, we found that this reaction is very rapid ( @xmath7 1 ps ) if compared with the diffusion time of d - atoms on an amorphous water ice ( @xmath36 = @xmath7 10 ms on porous asw ice , * ? ? ? * ) . in passing , we would like to point out that , over the last years , we have performed experiments on water formation in space starting from o , o@xmath1 and now o@xmath0 .
even though we can not infer absolute reaction rates which are very hard to estimate we can attempt to make relative comparisons .
it should be noted that the destruction rate of ozone following d exposures is as important as the one for o + d and d + o@xmath0 reactions within the error bar of our experimental data .
hence we conclude that the d + o@xmath0 reaction proceeds without a measurable activation barrier in the solid phase . obtained subsequently to the d + o@xmath0 experiment , the spectrum ( a ) of figure [ fig2 ] shows a desorption peak of mass 19 at 160 k attributed to the singly deuterated water ( hdo ) .
this clearly indicates that the reaction of d - atoms with o@xmath0 molecules on the ice surface is efficient and produces deuterated water molecules detected upon desorption of the whole ice film .
we also notice that nearly no desorption peak of d@xmath1o ( mass 20 ) is detected .
this is due to thermally activated h / d exchanges during the phase transition of the water ice @xcite .
therefore , even if d@xmath1o is actually formed on the surface , mostly water molecules in the form of hdo are likely to desorb and be detected .
the formation of deuterated water molecules at t@xmath37 = 10 k from d - atoms and o@xmath0 molecules may then follow the successive reaction pathways : @xmath38 @xmath39 @xmath40 @xmath41 @xmath42 whereas at t@xmath37 greater than 130 k , as mentioned above , the following isotope exchange reaction is active : @xmath43 of course , the reversal of reaction ( 12 ) is also possible , but in our case it is statistically negligible . as for what concerns the intermediate products of the reaction scheme , we performed again the d + o@xmath0 experiment but with a low exposure of d - atoms ( 0.01 monolayers ) on 0.15 monolayers of o@xmath0 , thus with a dose of d insufficient to go on with the hydrogenation reactions leading to water molecules .
this experiment shows that o@xmath0 molecules are not thoroughly consumed and also that o@xmath1 molecules appear on the surface indicating that reaction ( 7 ) proceeds at 10 k and that the formation of d@xmath1o can proceed as well from the intermediate reaction ( 9 ) .
the studied mechanism for water formation from the hydrogenation of ozone might be an explanation for the non- detection of o@xmath0 in the interstellar ices .
in fact , ozone reacts rapidly with hydrogen atoms to form water molecules on a water ice surface at 10 k and under conditions similar to those found in dense interstellar clouds .
likewise , the lack of a clear detection of o@xmath1 in interstellar ices may be explained by the high reactivity between o@xmath1 and h - atoms reported in miyauchi et al .
2008 , ioppolo et al . 2008 , and matar et al . 2008 .
in addition , the oh + h@xmath1 reaction could also be a key surface reaction in dark clouds , because h@xmath1 ( and d@xmath1 ) is present on the surface of the grains in large amounts @xcite . in our experiments too , there is a large amount of molecular deuterium because of the incomplete dissociation of d@xmath1 in the d beam . at the moment
it is not possible to us to probe explicitly the oh + h@xmath1 reaction , but we can investigate the role of d@xmath1 , by testing the o@xmath0 + d@xmath1 reaction . in order to do that ,
we have deposited a very large amount of d@xmath1 ( @xmath7 10@xmath44 @xmath27 ) on np - asw ice held at 10 k and previously dosed with o@xmath0 . in this experiment
, we did not detect any hdo , even when the surface was held at 50 k during d@xmath1 deposition in order to overcome a possible activation barrier .
therefore , we can fairly conclude that the d@xmath1 + o@xmath0 reaction does not proceed under our experimental conditions .
unlike the o@xmath0 + d@xmath1 reaction , o@xmath0 + d shows no barrier under our experimental conditions and this result shows clearly that ozone is a likely precursor to water formation on a water ice layer at 10 k as proposed by @xcite .
this pathway is to be preferred in environments where the number of o atoms exceeds the number of h atoms , which is supposed to be the case in dark clouds with a density higher than 10@xmath21 @xmath19 @xcite .
unfortunately , our present data do not allow us to disentangle the different intermediate reactions that lead to the formation of water . in order to investigate in more detail the kinetics of water formation at 10 k via ozone
, we are developing an additional ir diagnostic on formolism .
this will allow us to see the appearance of some of the intermediate reactants and to provide accurate quantitative results .
in this paper we have shown that it is possible to form efficiently water molecules from the reaction of o@xmath0 with hydrogen atoms on non - porous water ice .
this reaction exhibits no activation barrier at 10 k , and is about as efficient as the pathways having o or o@xmath1 as precursors .
this experimental result corroborates theoretical studies that consider ozone as one of the major actors of water formation in dark clouds .
we acknowledge the support of the french pcmi program funded by the cnrs , as well as the strong financial support from the conseil rgional dile de france ( sesame program i-07 - 597/r ) , the agence nationale de la recherche ( ref . 07 - 0129 ) and the conseil gnral du val doise .
amiaud l. , fillion j. h. , baouche s. , dulieu f. , momeni a. , lemaire j. l. 2006 , j. chem .
phys , 124 , 94702 amiaud , l. , dulieu , f. , fillion .
h . , momeni , a. , lemaire , j. l. 2007 , j. chem .
phys , 127 , 144709 bennett , c. j. , & kaiser , r. i. 2005 , apj , 635 , 1362 boogert , a. c. a. , pontoppidan , k. m. , lahuis , f. , jrgensen , j. k. , augereau , j .- c . , blake , g. a. , brooke , t. y. , brown , j. , dullemond , c. p. , evans , ii , n. j. , geers , v. , hogerheijde , m. r. , kessler - silacci , j. , knez , c. , morris , p. , noriega - crespo , a. , schier , f. l. , van dishoeck , e. f. , allen , l. e. , harvey , p. m. , koerner , d. w. , mundy , l. g. , myers , p. c. , padgett , d. l. , sargent , a. i. & stapelfeldt , k. r. 2004 , apjs , 154 , 359 borget , f. , chiavassa , t. , allouche , a. , aycard , j. p. 2001 , j. phys . chem .
b , 105 , 449 caselli , p. , stantcheva , t. , shalabiea , o. m. , shematovich , v. i. , & herbst , e. 2002 , planetary and space science , 50 , 1257 ceccarelli , c. , caselli , p. , herbst , e. , tielens , a. g. g. m. , & caux , e. 2007 , in protostars and planets , ed . v. b. reipurth , d. jewitt , & k. keil ( u. arizona press ) , 47 chaabouni , h. , schriver - mazzuoli , l. , & schriver , a. 2000 , j. phys .
chem . a , 104 , 6962 cuppen , h.m . , & herbst , e. 2007 apj , 668 , 294 dartois , e. 2005 , space science reviews , 119 , 293 dulieu , f. amiaud , l. , congiu , e. , fillion , j. h. , matar , e. , momeni , a. , pirronello v. , lemaire j. l. , submitted to a&a .
2009:arxiv 0903.3120v1 ehrenfreund , p. , breukers .
, r. and dhendecourt , l. & greenberg , j. m. 1992 , a&a , 260 , 431 gibb , e. l. , whittet , d. c. b. , schutte , w. a. , boogert , a. c. a. , chiar , j. e. , ehrenfreund , p. , gerakines , p. a. , keane , j. v. , tielens , a. g. g. m. , van dishoeck , e. f. & kerkhof , o. 2000 , apj , 536 , 347 greenberg , j. m. 1984 , sci .
am , 250 , 124 ioppolo , s. , cuppen , h. m. , romanzin , c. , van dishoeck , e. f. , & linnartz , h. 2008 , astrophysical journal , 686 , 1474 kimmel , g. a. , stevenson , k. p. , dohnalek , z. , smith , r. s. , & kay , b. d. 2001 , j. chem .
phys , 114 , 5284 kristensen , l. e. , amiaud , l. , dulieu , f. , lemaire , j. l. 2009 submitted to a&a lee , j. h. , michael , j. v. , payne , w. a. , & stief , l. j. 1978 , j. chem .
phys , 69 , 350 li , d. , & goldsmith , p. f.2003 , apj , 585 , 823 matar , e. , congiu , e. , dulieu , f. , momeni , a. , & lemaire , j. l. 2008 , a&a lett , 492 , l17 mathis , j. s. , mezger , p. , & panagia , n. 1983 , a&a , 128 , 212 miyauchi , n. , hidaka , h. , chigai , t. , nagaoka , a. , watanabe , n. , & kouchi , a. 2008 , chemical physics letters , 456 , 27 nagaoka , a. , watanabe , n. , & kouchi , a. 2005 , apj , 624 , l29 parise , b. , ceccarelli , c. , & maret , s. 2005 , a&a , 441 , 171 parise , b. 2004 , phd thesis , cesr ( toulouse , france ) pontopiddan , k. m. , van dishoeck , e.f . , dartois , e. 2004 , a&a , 426 , 925 schmitt , b. , de bergh , c. , festou , m ( eds . ) .
1998 , solar system ices , kluwer , dordrecht sivaraman , b. , jamieson , c. s. , mason , n. j. , & kaiser , r. i. 2007 , apj , 669 , 1414 smith , s. r. , huang , c. , & kay , b. d. , 1997 , j. phys .
b 101 , 6123 strazzulla , g. , & johnson , r. e. 1991 , in comets in the post - halley era , ed .
r. l. newburn , jr . ,
m. neugebauer , & j. rahe ( dordrecht : kluwer ) , 243 strazzulla , g. , brucato , j. r. , palumbo , m. e. , & satorre , m. a. 1997 , a&a , 321 , 618 tielens , a. g. g. m. , & hagen , w. 1982 , a&a , 114 , 245 molecules @xmath27 of o@xmath0 on np - asw : ( a ) o@xmath0 deposited at 10 k , ( b ) o@xmath0 deposited at 50 k and cooled to 10 k , ( c ) o@xmath0 deposited at 50 k , cooled to 10 k and irradiated with @xmath45 @xmath27 of d - atoms.,title="fig:",width=302 ] + @xmath27 d - atoms on the np - asw substrate at 10 k previously exposed to @xmath46 molecules @xmath27 of o@xmath0 , ( b ) deposition of @xmath47 @xmath27 d - atoms on the ozone - free water ice surface at 10 k .,title="fig:",width=302 ] + | the formation of water molecules from the reaction between ozone ( o@xmath0 ) and d - atoms is studied experimentally for the first time .
ozone is deposited on non - porous amorphous solid water ice ( h@xmath1o ) , and d - atoms are then sent onto the sample held at 10 k. hdo molecules are detected during the desorption of the whole substrate where isotope mixing takes place , indicating that water synthesis has occurred .
the efficiency of water formation via hydrogenation of ozone is of the same order of magnitude of that found for reactions involving o atoms or o@xmath1 molecules and exhibits no apparent activation barrier .
these experiments validate the assumption made by models using ozone as one of the precursors of water formation via solid - state chemistry on interstellar dust grains . |
the discovery of three x - ray transient ultracompact accreting millisecond pulsars ( msps ) , xte j1751 - 305 @xcite , xte j0929 - 314 @xcite , and xte j1807 - 294 @xcite have demonstrated the existence of binary pulsar systems with low mass , @xmath4 , donors .
these three ultracompact systems ( here defined as binaries with orbital periods , @xmath5 min ) are remarkably homogeneous , with measured @xmath6 min respectively , well below the minimum period for a system with a donor composed primarily of hydrogen @xcite . since the nature of the donors in these systems today depends on the prior evolution of the system , it is useful to discuss the potential formation mechanisms for these systems .
binary systems with @xmath7 min can form through at least two channels .
stable mass transfer from an evolved main - sequence star @xcite or a he burning star @xcite onto a neutron star ( ns ) is one mechanism . in this scenario ,
the main - sequence star is brought into roche lobe contact due to orbital angular momentum losses from magnetic braking at a time when the core has nearly completed h burning .
such a system will evolve to orbital periods comparable to the ultracompact msps and can reach @xmath8 min @xcite . @xcite and @xcite
show that the resulting ultracompact binaries have donor masses @xmath9 as they pass through @xmath10 min on their way towards a shorter period .
these masses are significantly greater than those measured in the ultracompact msps @xcite .
however , systems evolving through 40 min on the way out from the period minimum have masses more in line with the measurements ( @xmath11 ) and by this time the donors have become partially degenerate with core temperatures @xmath12 k @xcite .
the second scenario that may form ultracompact systems involves triggering a common envelope phase during an unstable mass transfer episode from the donor onto the ns .
the core of the donor , either a he or c / o white dwarf ( wd ) , and the ns spiral - in to shorter orbital periods until the envelope is expelled @xcite .
several authors have proposed binary evolution scenarios in which the system , after emerging from the common envelope phase , then suffers in - spiral due to gravitational wave ( gw ) emission and eventually re - establishes contact @xcite . during this long gw in - spiral , the wd will have had time to cool , setting the entropy of the donor at the onset of the second mass transfer phase @xcite .
@xcite finds that a large fraction of the ns - wd binaries that undergo a common envelope phase will reach contact within 1 gyr .
even considering a longer 4 gyr delay between the formation of the wd secondary and the onset of mass transfer , a he wd will have @xmath13 k @xcite , while a c / o wd will have @xmath14 k @xcite .
the mass transfer time - scale at contact is much shorter than the wd cooling time - scale for these wds so that the initial entropy of these objects is the minimum attainable . as noted by @xcite ,
if these objects adiabatically expand under mass loss , their @xmath3 will have been reduced by a factor of @xmath15 by the time they have reached @xmath11 .
in addition to the evolutionary arguments that donors in the ultracompact msps have not reached a @xmath16 configuration , the system xte j1751 - 305 provides observational evidence for a hot donor since , as noted by @xcite , a fully degenerate companion composed of he or c can not fill its roche lobe ( rl ) in this system .
hence , in examining the donors in the ultracompact accreting msps , we need to consider them to be arbitrarily degenerate , low mass objects of evolved ( he or c / o ) composition . to further constrain the nature of these donors ( for example , their @xmath3 ) requires knowledge of their mass - radius , @xmath0-@xmath1 , relation . for the compositions ( he , c / o ) , mass ( @xmath17 ) and @xmath3 ( @xmath18 k ) ranges of relevance to these objects , the corresponding central densities , @xmath19 ( @xmath20 ) are such that coulomb and thermal contributions to the equation of state ( eos ) provide non - negligible corrections to the degenerate electron pressure , impacting their @xmath0-@xmath1 relations . in this paper
, we detail the @xmath0-@xmath1 relation for low - mass stellar objects of finite @xmath3 , extending the @xmath0-@xmath1 relations of @xcite for @xmath16 objects . in particular , we make clear that there is a _ continuous connection _ between fully degenerate objects ( i.e. wds ) , fully convective low - mass stars ( i.e. @xmath21 polytropes ) , and coulomb dominated objects .
we begin in [ sec : toymod ] by constructing a simple model of these objects using an approximate eos .
although crude , this model describes adequately the relevant physics and yields an analytic description of the qualitative behavior of the @xmath0-@xmath1 relations and how they are affected by coulomb and thermal contributions to the eos .
we find that at finite @xmath3 , arbitrarily degenerate sequences exhibit a two branch solution , a fact noted previously [ e.g.,@xcite ] .
further , for sufficiently high @xmath3 , the sequence of solutions on these two branches exhibit a mutual end point at a _ non - zero mass _ , @xmath22 , below which equilibrium solutions do not exist .
when coulomb contributions are small and electrons are non - relativistic , fully convective stellar models of arbitrary degeneracy are well represented by @xmath21 polytropes @xcite . in [ sec : polytropes ] , we review the role played by degeneracy in determining the @xmath0-@xmath1 relation for @xmath21 polytropes and the existence of a two branch solution for the polytrope @xmath0-@xmath1 relation .
other authors have noted that for a given @xmath0 , there is a maximum @xmath3 that such polytrope models may have @xcite .
we connect the existence of a maximum @xmath3 with that of @xmath22 . in
[ sec : adiabats ] , we construct realistic @xmath0-@xmath1 relations using an eos for fully ionized plasmas . there
we exhibit explicitly the impact of coulomb physics on the @xmath0-@xmath1 relations . like the simplified and polytrope models , we find that the @xmath0-@xmath1 relations of this model exhibit a two branch solution and a non - zero @xmath22 at high @xmath3 .
in [ sec : binapps ] we apply our stellar models to the ultracompact msp systems .
for each of these systems , there is a donor ( of some composition and @xmath3 ) which will fill the rl at the required @xmath23 for any given orbital inclination . and , for a given composition , there is , in each system , a relation between orbital inclination and @xmath3 .
we examine what constraints this places on the composition and @xmath3 of the donors in these systems .
for example , in xte j1807 - 294 , c and he donors can have any @xmath3 , while an o donor s @xmath3 has a minimum value . in xte j1751 - 305 , all he or c / o donors must be hot .
we also examine how the future evolution of these systems depends on donor composition and @xmath3 and highlight the fact that a multiple - valued @xmath0-@xmath1 relation leads to a multiple - valued relation between @xmath23 and the system s mass transfer rate , @xmath24 . finally ,
in the context of adiabatic evolution , we discuss the expected number distribution of ultracompact systems as a function of @xmath23 . in a steady state
, this distribution depends almost solely on the response of the donor radius , @xmath25 , to mass loss through the quantity @xmath26 .
the increased importance of coulomb physics in c / o donors alters @xmath27 as compared to he donors and the expected distribution for the two donor types differs dramatically .
depending on donor type , the relative number of systems at @xmath10 min as compared to those at @xmath8 min is @xmath28 for he donors and @xmath29 for c / o donors .
we conclude in [ sec : conc ] by discussing future applications of and refinements to our models .
in the mass range of interest ( @xmath30 ) , @xmath0-@xmath1 relations for various objects have previously been constructed . for h
rich objects , @xcite summarize the work that has been done on the structure and evolution of brown dwarfs and related objects . for these objects ,
the models include detailed treatment of the equation of state ( eos ) and atmospheric physics . on the other hand , for objects with more evolved composition ( i.e. he or c / o )
the theory is not so mature .
@xcite calculated the @xmath0-@xmath1 relations for @xmath16 objects using the eos they derived from the thomas - fermi - dirac equation @xcite .
this eos treats in an approximate manner the corrections due to coulomb interactions and exchange effects in a fully - degenerate plasma .
additional @xmath0-@xmath1 relations produced by several different @xmath16 eos are presented in @xcite , but do not differ appreciably from the @xcite results . for the partially degenerate case
, there is a large body of literature for he and c / o core wds more massive than @xmath31 @xcite .
but , only recently have @xmath0-@xmath1 relations for arbitrarily degenerate he and c / o wds with masses @xmath32 been calculated @xcite and only for a limited number of cases . here
we fill this gap by constructing low - mass wd models utilizing a realistic eos .
degenerate stellar objects with @xmath33 have central densities , @xmath34@xmath35 . in this density range
, the relative energy contributions to the plasma from an ideal fermi gas and coulomb interactions can be comparable .
also , for these densities , at a temperature @xmath36 k , the thermal contributions to the pressure are about 10% of those of the electrons . at lower densities , coulomb and thermal effects
become even more significant in calculating the eos . to examine the interplay between these contributions and how they impact the structure of low - mass stars , we start with a simple eos for a plasma composed of ions with charge number @xmath37 and atomic mass number @xmath38 , @xmath39 where @xmath40 is the pressure of a fully degenerate non - interacting fermi gas of electrons @xmath41 , @xmath42 is the pressure of an ideal gas of ions and electrons , @xmath43 , and @xmath44 is the negative pressure contribution due to coulomb interactions in the wigner - seitz approximation , @xmath45 . here
@xmath46 ( @xmath47 is the mean molecular weight per electron and equals @xmath48 in a single composition plasma ) , @xmath49 , and @xmath50 .
consider a one - zone stellar model , i.e. a spherical system characterized by a single pressure and density , @xmath51 and @xmath52 . from dimensional analysis , @xmath53 and @xmath54 where @xmath0 and @xmath1 are
the mass and radius of the star and @xmath55 is the newtonian constant of gravitation . with the pressure given by equation ( [ eq : psimp ] ) , @xmath56 in the @xmath57 limit
, the term involving @xmath58 vanishes and solving for @xmath1 gives @xmath59 showing that the gravitational and coulomb attraction act to collapse the star , which is supported by the degenerate electron pressure . as @xmath60 ,
coulomb forces dominate gravity and @xmath61 . as @xmath62 , coulomb forces become negligible and @xmath63 .
the transition from gravitational to coulomb dominated regime occurs where @xmath64 .
the relation @xmath61 in the low @xmath0 limit implies a constant density , @xmath65 , fixed by the balance between the electron pressure and coulomb attraction . exhibiting @xmath52 explicitly as a function of @xmath0 , @xmath66 and @xmath67 . for @xmath68 ,
equation ( [ eq : mrsmast ] ) has solutions given by @xmath69 figure [ fig : mrsimp ] exhibits several isotherms of the @xmath0-@xmath1 relation of equation ( [ eq : mrsimp ] ) .
the two - branch nature of this relation is apparent , as is the clear separation into three classes of stellar objects .
the large @xmath0 lower - branch is made up of gravitationally bound objects supported by degeneracy pressure .
on the lower - branch at small @xmath0 , we have coulomb dominated objects .
the upper branch consists of objects supported by thermal pressure . for @xmath68
, there exists a minimum mass , @xmath22 , found by setting the discriminant in equation ( [ eq : mrsimp ] ) to zero , @xmath70 this expression is only positive if @xmath71 k. for he(c ) this temperature is @xmath72 k. above this critical temperature , the two branches meet at a mutual end point and no solution with @xmath73 exists .
when coulomb physics is negligible , @xmath74 occurs at the point where @xmath75 along the solution curve .
the existence of @xmath22 results from the fact that for @xmath76 , the pressure provided by @xmath77 at any @xmath52 is greater than that needed to support the star ( this is not the case for either an ideal gas or a fermi gas independent of the other in either of these cases it is well known that equilibrium solutions down to @xmath78 exist ) . alternately , for a given @xmath0 , the isotherm on which @xmath79 gives the maximum @xmath80 for which solutions exist with this mass .
for an object starting out on an upper branch solution , e.g. a recently expired star , as it loses entropy via radiation , it contracts . for a fixed mass ,
a star supported by thermal pressure has @xmath81 , and it heats up , as expected .
the thermal pressure in this case goes as @xmath82 , but @xmath83 and eventually @xmath40 dominates , halting substantial contraction .
further entropy loss leads to a reduction in @xmath80 . for a given @xmath80 and @xmath0 ,
the two possible solutions are physically distinguished by their entropy . ignoring non - ideal effects and assuming that the electrons are non - relativistic
, a fully convective object has an eos obeying @xmath84 and is modeled by an @xmath21 polytrope .
the specific entropy , @xmath85 , throughout such a model is a constant . as @xmath85 only depends on the degeneracy parameter @xmath86 , defined as the ratio of the chemical potential of the electrons to @xmath87 ( @xmath88 being the boltzmann constant ) , @xmath86 is also a constant of the model @xcite .
following @xcite , this allows us to write the pressure of a non - interacting gas of electrons and ions as @xmath89 where @xmath90 is the mass of a proton and @xmath91 is defined as @xmath92 thus varying as @xmath86 changes . here
@xmath93 is the ion mean molecular weight and @xmath94 is the fermi - dirac function of order @xmath88 from @xcite .
we then utilize the @xmath21 polytrope relations , equation ( [ eq : polyp ] ) , and the fact that @xmath95 to determine @xmath0 and @xmath1 as a function of @xmath86 , @xmath3 .
these results can be expressed as @xmath96^{-1}\label{eq : polyr}\,,\\ & t_c = { \ensuremath{3.42 \times 10^{5}}}\ , \mathrm{k } \left(\frac{m}{0.01 { { m}_{\odot}}}\right)^{4/3 } \left[\mu_e^{1/3 } \mu_{\mathrm{eff } } f^{1/3}_{1/2}\right]^2\label{eq : polyt}\,,\end{aligned}\ ] ] which is a minor rewrite of the results in @xcite . equation ( [ eq : polyt ] ) shows the relation between @xmath3 and @xmath0 is a function of @xmath86 through the combination @xmath97 , which has a single maximum at @xmath98 for expected wd interior compositions .
just as in the simple models of [ sec : toymod ] , this shows explicitly the connection between a maximum @xmath3 for a given @xmath0 and the existence of an @xmath22 for a fixed @xmath3 . here again
, there are 2 equilibrium radii for a given @xmath0 and @xmath3 differentiated by their degree of degeneracy or equivalently their entropy .
the transition from a thermal pressure dominated to degeneracy pressure dominated state in these models approximately determines @xmath22 .
this transition occurs near where @xmath99 @xmath100 . for increasing @xmath93 ,
this occurs at lower @xmath86 , i.e. at lower @xmath19 if @xmath47 and @xmath3 are held fixed . with @xmath84 , dimensional analysis shows @xmath101 ; a lower density at the transition between degenerate and non - degenerate states gives a smaller @xmath22 .
for a fixed @xmath3 , a pure c wd will has a smaller @xmath22 than a pure he wd .
we now calculate the @xmath0-@xmath1 relations derived from stellar models that include coulomb physics . in these calculations , we assume that the interior profile of our stellar models are adiabatic . in reality , the actual entropy profile in a given donor will depend on its evolutionary history . many factors such as whether the system formed through a stable or unstable mass transfer channel , whether or not h burning is still ongoing at the point of contact , and how the mass transfer rate changes with time can impact either the initial entropy profile of the donor or its subsequent evolution . in calculating models for donors in ultracompact binaries , without choosing their evolutionary history , a reasonable approximation of their internal entropy profile is the best that can be done . since we aim to construct models that will enable analysis of the donor s properties today , irrespective of their past histories , we must assume some internal profile in calculating them . a system that initiates mass transfer at a @xmath102 min has a donor mass @xmath103 .
the mass transfer time - scale for such a system with a ns primary is roughly @xmath104 myr , depending on @xmath105 .
consideration of the flux through the half - mass point in our models due to electron conduction ( calculated using the conductive opacities of @xcite ) compared to the heat content of the interior half of the model shows roughly that the time required to transport this heat out of the interior is @xmath106 myr-1 gyr , again depending on @xmath105 .
thus during the mass transfer episode that would lead up to the creation of systems at @xmath10 min , the internal evolution is to first order an adiabatic expansion and , in the absence of tidal heating , the initial entropy profile should be more or less preserved with some corrections due to heat transport . to what ever degree
this occurs , the critical point is that the interior will not be able to maintain an isothermal profile .
we chose to use an adiabatic profile , instead of another arbitrary choice , since it provides several convenient features
it is a limiting case for the possible thermal profiles of the donor and produces the most compact configuration for a given @xmath3 and @xmath105 , it is completely determined by the utilized eos , and it allows parameterization of a set of models by one quantity , the specific entropy which is constant throughout the model .
in addition , the calculations of @xcite show that he donors tend to become adiabatic as they lose mass .
these latter calculations also highlight that throughout the mass loss episode , donors remain far from isothermal except in the deep interior of the star , and then only at masses near 0.1 @xmath107 @xcite .
for the equation of state , we use the results of @xcite and @xcite .
their work provides accurate prescriptions for calculating the eos for a fully ionized plasma in either liquid or solid states and which includes the ideal contributions from non - degenerate ions , degenerate electrons and the non - ideal contributions due to coulomb interactions .
we also include the radiation contribution , a small effect . in the calculations below
, we assume that the plasma is in the solid state when the quantity @xmath108 @xcite . here
@xmath109 is the inter - ion spacing given by @xmath110 , @xmath111 being the number density of ions .
the quantity @xmath112 is a measure of the strength of the ionic coulomb interactions .
we display a set of isotherms for a pure he plasma calculated with this eos in figure [ fig : eos_ad ] with the solid and dotted lines .
the dotted lines indicate regions in parameter space where @xmath80 and @xmath52 are such that the plasma will likely become only partially ionized . in these regimes
this eos is not strictly valid .
the dashed lines in figure [ fig : eos_ad ] show representative adiabats for this eos , some of which cross into regions of partial ionization .
since we utilize an adiabatic internal profile in calculating our models in [ sec : ad : calc ] , in some of these models there can be a point in their outer layers where our eos becomes invalid .
we use our eos in these regimes despite this problem for two reasons .
first , in the models of most interest only the very most outer layers of the models lie in regimes where ionization state transitions become an issue and the calculated @xmath0-@xmath1 relation is not affected .
second , this eos provides a simple method of calculating the eos for various compositions over a wide range of @xmath52 and @xmath80 and this ease of use would be sacrificed by constructing composition specific extensions to the eos .
we will highlight in our results models in which these eos concerns may cause more than a few percent uncertainty in the calculated @xmath0-@xmath1 relations .
we constructed models for arbitrarily degenerate objects by integrating mass conservation and hydrostatic balance while presuming an adiabatic temperature profile .
the calculation of our models proceeds as follows . in the interior , where degeneracy pressure dominates , we integrate @xmath113 where @xmath114 is the adiabatic exponent @xmath115 found from the eos . the density at each integration step is then solved for from @xmath51 and @xmath80 using the eos . at low @xmath52 , where coulomb effects on the pressure become significant , we switch to integrating @xmath116 where @xmath117 is another adiabatic exponent , and solve for @xmath80 at each integration step from @xmath51 and @xmath52 .
each model integration was terminated once the following criteria were met(1 ) between two integration steps in which @xmath51 differed by more than a factor of @xmath118 , @xmath119 and @xmath120 changed by no more than 1 part in @xmath121 , and ( 2 ) the current pressure is such that @xmath122 .
we change the integrated variable because as @xmath123 along an adiabat , @xmath52 becomes very insensitive to @xmath51 and it is numerically intractable to determine @xmath52 by root finding from @xmath51 and @xmath80 . on the other hand , in the low @xmath51 limit , determining @xmath80 accurately from @xmath51 and @xmath52 is possible , something that was not true in the highly degenerate regime .
we switch from integrating @xmath80 to integrating @xmath52 when a rough measure of the degeneracy , @xmath124 , is satisfied ( where @xmath52 and @xmath80 are in cgs units ) . in the solid state ,
when @xmath125 , @xmath126 due to the rapid decline in the plasma s specific heat once crystalline .
we use an adaptive step - size explicit runge - kutta algorithm to integrate our equations with the step size chosen to maintain a fixed fractional accuracy . a dramatic increase in @xmath114
causes a sharp decline in the speed of the integration as the algorithm tries to maintain this accuracy in all three integrated quantities , @xmath119 , @xmath51 , and @xmath80 . to deal with this problem ,
once @xmath127 k , we set @xmath128 .
this causes no issues in the @xmath0-@xmath1 relations because by this point we are already well in the @xmath129 limit as far as the @xmath51-@xmath52 relation is concerned in any of our models . finally in these models
the ion coupling parameter , @xmath112 , increases with @xmath120 .
this is due to the fact that along an adiabat , @xmath130 varies from @xmath131 to @xmath132 in a coulomb plasma . since @xmath133 , and @xmath134
, @xmath112 goes as @xmath52 to a negative power .
this is not just true for adiabatic profiles any object with a profile @xmath135 for @xmath136 will have @xmath112 increasing with @xmath120 ( assuming @xmath37 is fixed)because of this some models transition from liquid to solid in their outer layers . for these cases we do not attempt to match adiabats in the solid and liquid phase (
i.e. we do not account for the latent heat ) .
instead , the integration assumes continuity of @xmath51 and @xmath80 ; the entropy in these models have a small discontinuity .
if crystallization of the object were to actually occur from the surface inward , this could have significant impact on the mass loss rate , since the primary s gravitational field would have to overcome the coulomb binding of the crystal to effect mass transfer .
more realistic calculations are obviously needed to consider this further and consideration of this potential effect in evolutionary calculations are encouraged .
typical results for our model calculations are shown in figures [ fig : polycomp ] and [ fig : headmr ] .
figure [ fig : polycomp ] shows for pure he models how the results utilizing the eos of [ sec : ad : eos ] differ from the polytropes that neglect coulomb physics . in this , and in figure [ fig : headmr ] , portions of the @xmath0-@xmath1 curves shown in dotted lines indicate models where more than 5% of the mass lies in regions where our eos is not strictly valid .
the impact of coulomb interactions on the structure of low mass wds is clear from the comparison of the he polytrope models ( short dashed lines ) to the realistic eos he models ( solid lines ) .
figure [ fig : headmr ] displays a set of @xmath0-@xmath1 isotherms for our he ( solid lines ) and c ( dash lines ) models along with lines of constant @xmath23 for a donor filling its roche lobe overlaid ( dash - dot lines with @xmath23 indicated ) .
again , the dotted portions of the @xmath0-@xmath1 relations indicate models in which more than 5% of the model s mass lies in a regime where the eos is not strictly valid . from figures [ fig : polycomp ] and [ fig :
headmr ] , it is clear that the realistic treatment of coulomb physics in the eos is necessary to calculate accurately the structure of low mass wds .
we now apply the adiabatic models of [ sec : adiabats ] to the three known ultracompact accreting millisecond pulsars and the high - field x - ray pulsar 4u 1626 - 67 . in figure
[ fig : amsprl ] we display ( short dashed lines ) the @xmath0-@xmath1 relations of roche lobe ( rl ) filling donors in xte j0929 - 314 , xte j1751 - 305 , and xte j1807 - 294 @xcite and @xmath0-@xmath1 relations for our adiabatic models .
we show two isotherms each for he , c , and o models , an approximate @xmath16 @xmath0-@xmath1 relation and one for hot ( @xmath137 k ) models .
if the donors in all three systems are he wds , then @xmath16 objects are allowed in xte j0929 - 314 and xte j1807 - 294 , but xte j1751 - 305 requires a hot donor @xcite . for c / o donors , whose @xmath0-@xmath1 relations will lie between the c and o models shown , only xte j1807 - 294 permits a rl filling cold donor .
the other two systems both require hot c / o donors .
the curves showing the rl filling @xmath0-@xmath1 relations are parameterized by the orbital inclination , @xmath138 ( where @xmath139 is a face on system ) and there is a correspondence between @xmath3 and @xmath138 for each donor composition , as shown in figure [ fig : mtsini ] . from figure
[ fig : mtsini ] , the donor in xte j1751 - 305 must have @xmath140 k. from orbital inclination constraints , the probability that xte j1807 - 294 can accommodate a he donor is 15% ( for xte j0929 - 304 , it is @xmath141 35% ) . the other ultracompact accreting pulsar , 4u 1626 - 67 , at @xmath142 @xcite is also shown in figure [ fig : amsprl ] . though the orbit has not yet been detected by timing the pulsar , the current upper limit of @xmath143 lt - ms @xcite allows us , in conjunction with our theoretical work , to constrain the nature of the donor star .
ever since the discovery of a neon emission line @xcite from this system , there have been active discussions of the nature of the donor .
the @xcite measurement of a high ne to o ratio ( further inferred in other ultracompacts by @xcite ) , led them to suggest that the donors in these binaries are the cores of previously crystallized c / o wds . @xcite
have since also seen strong c and o lines , but no evidence for helium .
hence , this system seems a likely one to use for probes of c / o donors . for any @xmath144 , a star that fills the rl at the measured @xmath23 for each of the ultracompacts
can always be found by some combination of entropy , composition , and mass and the current values of each impact the orbital evolution of the system . for @xmath145 , the roche radius , @xmath146 ,
can be approximated by @xcite @xmath147 where @xmath109 is the separation between @xmath148 and @xmath105 .
combined with kepler s third law , this leads to the so - called period - mean density relationship , @xmath149 assuming conservative mass transfer , the mass transfer rate ( a _ positive _ quantity ) is given by @xcite @xmath150 where @xmath151 is the orbital angular momentum , @xmath152 the angular momentum loss rate set by gw emission @xcite and @xmath153 denotes how the donor s radius changes with mass loss . for a given system , @xmath152 will depend on the inclination through @xmath105 and @xmath109 .
the rate at which the orbit evolves will vary accordingly as will the @xmath154 relation over the course of the evolution . to illustrate this , we calculated the forward evolution of xte j0929 - 304 assuming a he donor and four different @xmath144 values .
we assumed that the ns has @xmath155 ( and ignore the change in @xmath148 due to accretion ) and set @xmath156 so that the donor evolves adiabatically , ignoring any heating mechanisms ( such as irradiation or tidal heating ) and cooling .
we show the results in figure [ fig : hej0929 t ] , displaying @xmath105 and @xmath24 as a function of @xmath23 .
these relations are not single valued , but parameterized by orbital inclination , or equivalently , by the donor s entropy .
a smaller @xmath144 requires a more massive donor ( which must be hotter than a lower mass donor if it is to have the fixed mean density implied by the system s @xmath23 ; see also figure [ fig : mtsini ] ) .
this gives a higher @xmath24 for a fixed @xmath23 , as seen in figure [ fig : hej0929
this also impacts the rate at which the orbit evolves . in figure
[ fig : hej0929 t ] , the age of the system from today is indicated along each curve by symbols and it can be seen that the smaller @xmath144 the faster the system will evolve in @xmath23 . at a given orbital inclination in a specified system , a c / o donor must have a higher @xmath3 to fills the roche lobe than a he donor .
this is due to the stronger coulomb physics in the c / o object , which also causes @xmath157 to differ between the donor types and impacts the binary s evolution .
the difference in @xmath157 between composition is shown in figure [ fig : nad ] for two representative adiabatic tracks for each composition .
the solid dots show the zapolsky - salpeter @xmath27 for the same compositions .
the effect of different @xmath157 on orbital evolution is evident in figure [ fig : aj0929 t ] , which compares the evolution of xte j0929 - 304 for @xmath158 and he , c , and o donors .
the difference in the tracks comes about due to the difference in @xmath157 due to composition ; higher @xmath37 donors evolve fastest in mass , but slowest in @xmath23 because they remain more dense than lower @xmath37 donors at a given mass .
these evolution calculations highlight the dependence of observables ( @xmath23 and @xmath24 ) on the donor s entropy and composition .
we now emphasize the impact of the donor composition on the ultracompact population , especially the resulting orbital period distribution . in the scenario of cooling wds reaching contact via in - spiral ( e.g. @xcite )
, the relative number of he vs c / o wd s that come into contact and stably reach longer orbital periods is hard to know .
however , what we will show here is the ability to constrain the relative populations of say , he donors , at one orbital period versus another .
a more complete picture of this dependence is shown in figure [ fig : ns_ucsys ] , where we display for he ( solid lines ) , c ( dashed lines ) , and o ( dash - dot lines ) donors , the @xmath24-@xmath23 relation along @xmath0-@xmath1 isotherms with @xmath159 , and @xmath160 k , assuming @xmath161 and @xmath162 .
these are instantaneous @xmath24 along an adiabatic track at a given @xmath23 and @xmath3 .
one can see immediately that for a given @xmath23 , @xmath24 can constrain the combination of donor @xmath3 and composition .
in particular , a sufficiently strong upper limit on @xmath24 can _ rule out _ a he donor for a given @xmath23 . above the minimum @xmath24 for a he donor , further information about the donor composition
is difficult to infer without constraints on @xmath3 .
now consider adiabatic evolution with initial @xmath163 and donors of varying composition and @xmath3 which fill their rl at @xmath164 min .
we evolve these systems assuming the donor responds adiabatically to mass loss .
the resulting tracks in the @xmath24-@xmath23 diagram are shown in figure [ fig : adiabmdotp ] , along with the measured periods of the known ultracompact binaries with a ns primary ( vertical dotted lines ) and the critical @xmath24 below which the accretion disk in these systems is subject to thermal instabilities for both irradiated ( hatched region , @xcite ) and non - irradiated disks ( nearly horizontal lines , @xcite ) . as compared with figure [ fig : ns_ucsys ] , it can be seen that for systems at @xmath165 min to have _ time - averaged _ mass transfer rates @xmath166 yr@xmath167 , the donor can not have evolved adiabatically from systems coming into contact at @xmath8 min .
there are potentially two examples of such systems : 4u 1626 - 67 ( @xmath168 min and @xmath169 @xcite ) and 4u 1916 - 05 ( @xmath170 min , @xmath171 @xcite ) . if these measured @xmath24s reflect the long - term average @xmath24 , then the donors must be extremely hot ( @xmath172 k , see figure [ fig : ns_ucsys ] ) . however , the typical @xmath24 at these orbital periods are below where a he or c / o disk is thermally stable ( see figure [ fig : adiabmdotp ] ; @xcite ) . in that case
we would explain the present @xmath24 as a higher rate indicative of a system in outburst .
indeed , the luminosity from 4u 1626 - 67 is observed to be in a steady slow decline @xcite .
if the ultracompacts evolve adiabatically , we can determine their relative numbers as a function of @xmath23 .
defining @xmath173 such that @xmath174 is the number of systems with orbital period between @xmath23 and @xmath175 and demanding continuity gives @xmath176 leading to the expected relation between @xmath177 and @xmath178 : @xmath179 where @xmath180 and @xmath181 are the respective quantities at some reference orbital period , @xmath182 .
for @xmath157 fixed , @xmath183 and with @xmath184 , equations ( [ eq : rl ] ) , ( [ eq : pdrel ] ) , and ( [ eq : nofpdot ] ) lead to the simple relation @xmath185 in this case @xmath186 and from this it is clear that _
@xmath157 alone _ determines the number distribution . as @xmath187 increases with @xmath157 , systems with c / o donors will have a stronger increase in @xmath178 with @xmath23 than those with he donors due to the difference in @xmath157 shown in figure [ fig : nad ] . in the more general case
, @xmath157 is variable and equation ( [ eq : dlnndlnp ] ) becomes , up to terms of order @xmath188 , @xmath189 and in general the distribution is almost solely a function of @xmath157 . while equations ( [ eq : dlnndlnp ] ) and ( [ eq : gen_dlnndlnp ] ) highlight the centrality of @xmath157 in determining @xmath178 , to calculate @xmath178 for each of the adiabats in figure [ fig : adiabmdotp ] it is more straightforward , and slightly more accurate due to the small dependence of @xmath178 on @xmath190 , to calculate @xmath177 numerically and evaluate equation ( [ eq : nofpdot ] ) .
we calculate @xmath177 from @xmath191 where @xmath192 .
we display the resulting @xmath178 distribution for each adiabatic track , normalized to the value of @xmath180 at @xmath193 min , in figure [ fig : dndp ] .
it is clear that the distribution is a strong function of composition , through the differing @xmath194 , but only a weakly depends on the entropy of the donor .
the entropy dependence of @xmath178 derives from the fact that @xmath195 [ equations ( [ eq : rl ] ) , ( [ eq : mdot ] ) , and ( [ eq : pdot ] ) ] and donors with a higher entropy at a fixed mean density have larger @xmath105 ( see figure [ fig : amsprl ] ) .
figure [ fig : dndp ] also shows that depending on donor type and entropy , we expect to see roughly 60 - 160 ( for he and c / o donors , respectively ) times as many systems at @xmath196 min than at 10 min .
while consideration of non - adiabatic evolution will change the value of this ratio , the fact that it will be larger for c / o donors than for he donors is expected to be a robust result , regardless of the evolutionary scenario considered .
we have presented models for arbitrarily degenerate stellar objects including coulomb physics with masses @xmath2 . at these low masses , the well known @xmath0-@xmath1 relations for @xmath21 polytropes ( @xmath197 ) , wds ( @xmath63 ) , and coulomb dominated objects ( @xmath61 ) merge and transition from one to another .
the connection between @xmath16 degenerate and coulomb dominated objects has been known from the @xcite @xmath0-@xmath1 relations .
the connection between polytropes and degenerate objects , neglecting coulomb physics , is seen in the @xmath21 polytropes .
our models make the final connection between the three classes of objects , filling in the gap occupied by @xmath198 objects in which coulomb physics can not be neglected . as discussed in
[ sec : models ] , a ubiquitous feature of stellar @xmath0-@xmath1 relations at sufficiently high @xmath3 is the existence of a minimum mass , @xmath22 , below which equilibrium solutions do not exist .
the cause of this is the transition from ideal gas to degenerate electrons providing the pressure support and we showed in [ sec : polytropes ] that the well known maximum @xmath3 in @xmath21 polytropes for a fixed @xmath0 @xcite is the same physics .
the value of @xmath22 depends on both @xmath3 ( in fact , @xmath199 at sufficiently low @xmath3 ) and the strength of coulomb physics . in general , the lower @xmath3 and the stronger the coulomb interactions , the smaller @xmath22 .
the existence of an @xmath22 may have a profound impact on the evolution of a donor undergoing mass loss .
for our he wds , @xmath200 for models with @xmath201 k. as discussed in @xcite , the existence of an @xmath202 leads to the possibility of disrupting the donor through mass loss .
this could be accomplished through stable mass loss down to @xmath203 or via a mass transfer instability .
the latter will occur if the expansion of the donor under mass loss exceeds that of the rl .
the entropy input needed to cause this instability and the fate of the donor are the subject of future work .
we have applied our model to the accreting ultracompact msps . in xte j1751 - 305 ,
the donor must be hot regardless of its composition ; in xte j0929 - 314 and xte j1807 - 294 , fully degenerate donors are possible , depending on the composition . from orbital inclination constraints ,
the probability that xte j1807 - 294 can accommodate a he donor is 15% while for xte j0929 - 304 , this probability is @xmath141 35% , providing support to the notion that some of these donors are likely c / o wds @xcite .
the evolution of each system will differ depending on the orbital inclination .
in particular , how far the system can evolve in @xmath23 in a specified time and the expected @xmath24-@xmath23 relation depends on @xmath144 through both the mass , core temperature , and composition of the donor .
in general , the @xmath24-@xmath23 relation is additionally parameterized by the donor @xmath3 and composition .
we find that the number distribution of systems as a function of @xmath23 , @xmath173 , is determined by the donor s @xmath27 .
the distribution for systems with c / o donors thus varies significantly from those with he donors . in the case of adiabatic evolution ,
the relative number of systems at 40 min to that at 10 min is @xmath29 for c / o donors and @xmath28 for he donors .
in addition to the accreting msp systems , our models are applicable to the am cvn binary systems , which are believed to be double wd binaries with a he donor @xcite .
the application of our models to these systems is the subject of current work .
the constant entropy models we have calculated give a lower limit on @xmath1 for a given @xmath0 , @xmath3 , and composition .
the actual thermal profile of a donor will depend on its prior evolution : how entropy is deposited into the star , entropy losses , and how quickly heat transport occurs as compared to mass loss . to determine the entropy profile of a donor in a specific system requires consideration of the coupled evolution of the binary and the donor .
while our models do not address this uncertainty , they do provide limiting @xmath0-@xmath1 relations based on the _ total _ entropy content of the model in a consistent and systematic treatment without consideration of past evolution .
as such they will be useful in consideration of binary systems with low - mass wd companions where time evolution of the donor s structure coupled to that of the binary itself is not computationally feasible or necessary .
tracks of our @xmath0-@xmath1 relations and @xmath157 as a function of @xmath19 and @xmath3 for he , c , and o donors are available along with the electronic version of the paper .
we thank deepto chakrabarty , craig markwardt , gijs nelemans , and ira wasserman for insightful and productive comments and discussions over the course of this work .
this work was supported by the national science foundation under grants phy99 - 07949 and ast02 - 05956 and by nasa through grant nag 5 - 8658 .
l. b. is a cottrell scholar of the research corporation . 99 althaus , l. g. & benvenuto , o. g. 1997 , apj , 477,313 angelini , l. , white , n. e. , nagase , f. , kallman , t. r. , yoshida , a. , takeshima , t. , becker , c. , & paerels , f. 1995 , , 449 , l41 bildsten , l. 2002 , apj , 577 , l27 burrows , a. , hubbard , w. b. , lunine , j. i. , & liebert , j. 2001 rvmp , 73 , 719 burrows , a. & liebert , j. 1993 , rvmp , 65 , 301 chabrier , g. 1993 , apj , 414 , 695 chabrier , g. & potekhin , a. y. 1998 , pre , 58 , 4941 chakrabarty , d. 1998 , , 492 , 342 chakrabarty , d. , bildsten , l. , grunsfeld , j. m. , koh , d. t. , prince t. a. , vaughan , b. a. , finger , m. h. , scott , d. m. , & wilson , r. b. , 1997 , , 474 , 414 cox , j. p. & guili , r. t. 1968 , _ principles of stellar structure _
, new york : gordon & breach cox , j. p. & salpeter , e. e. 1964 ,
apj , 140 , 485 dewi , j. d. m. , pols , o. r. , savonije , g. j. , & van den heuvel , e. p. j. 2002 , mnras , 331 , 1027 driebe , t. , bl " ocker , t. , schnberner , d. , & herwig , f. 1999 , , 350 , 89 dubus , g. , lasota , j. , hameury , j. , & charles , p. 1999
, , 303 , 139 fedorova , a. v. & ergma , e. v. 1989 , ap&ss , 151 , 125 farouki , r. t. & hamaguchi , s. 1993 , pre , 47 , 4330 fontaine , g. , brassard , p. , & bergeron ,
p. 2001 , pasp , 113 , 409 fontaine , g. , graboske , h. c. , & van horn , h. m. 1977 , , 35 , 293 galloway , d. k. , chakrabarty , d. , morgan , e. h. , remillard , r. a. 2002 , apj , 576 , l137 hansen , c. j. & sprangenberg , w. h. 1971 , apj , 168 , 71 hayashi , c. & nakano , t. 1963 , prog .
, 30 , 460 homer , l. , anderson , s. f. , wachter , s. , & margon , b. 2002 , , 124 , 3348 iben , i. jr .
& tutukov , a. v. 1985 , apjs , 58 , 661 ichimaru , s. , iyetomi , h. , & tanaka , s. 1987 , phys .
, 149 , 91 juett , a. m. , psaltis , d. , & chakrabarty , d. 2001 , , 560 , l59 kerlet , g. i. & chhabildas , l. , 2001 , sandia report 2001 - 2619 , sandia national laboratories lai , d. , abrahams , a. m. , shapiro , s. l. , 1991 apj , 377 , 612 landau , l. d. & lifshitz , e. m. 1962 , _ the classical theory of fields _ , oxford : pergamon , 2nd . ed .
levine , a. , ma , c. p. , mcclintock , j. , rappaport , s. , van der klis , m. , & verbunt , f. 1988 , , 327 , 732 markwardt , c. b. , swank , j. h. , strohymayer , t. e. , in t zand , j. j. m. , & marshall , f. e. 2002 , apj , 575 , l21 markwardt , c. b. , smith , e. , & swank , j. h. , 2003 , iauc circ .
8080 markwardt , c. b. , juda , m. , & swank , j. h. , 2003 , iauc circ .
8095 mavromatakis , f. 1994 , , 285 , 503 menou , k. , perna , r. , & hernquist , l. 2002 , apj 564 , l81 middleditch , j. , mason , k. o. , nelson , j. e. , & white , n. e. 1981 , , 244 , 1001 nelemans , g. , portegies zwart , s. f. , verbunt , f. , & yungelson , l. r. 2001 , a&a , 368 , 939 .
nelson , l. a. , rappaport , s. a. , joss , p. c. 1986 , apj , 304 , 231 nelson , l. a. & rappaport , s. a. , 2003 , apj , submitted , astro - ph/0304374 paczynski , b. 1967 acta astron . , 17 , 287 paczynski , b. 1976 , in _ the structure and evolution of close binary systems _ , eggleton , p. , mitton , s. , & whelan , j. ( eds . ) , dordrecht , reidel , p. 75
panei , j. a. , althaus , l. g. , & benvenuto , o. g. 2000 , a&a , 353 , 970 podsiadlowski , ph . , rappaport , s. , & pfahl , e. d. 2002 , apj , 565 , 1107 potekhin , a. y. , baiko , d. a. , haensel , p. , & yakovlev , d. g. 1999 , , 346 , 345 potekhin , a. y. & chabrier , g. 2000 , pre , 62 , 8554 rasio , f. a. , pfahl , e. d. , & rappaport , s. , 2000 , apj , 532 , l47 rappaport , s. & joss , p. c. 1984 , apj , 283,232 rappaport , s. , joss , p.c .
, & webbink , r. f. 1982 , apj , 254,616 rogers , f. j. , swenson , f. j. , & iglesias , c. a. 1996 , apj , 456 , 902 ruderman , m. & shaham , j. 1983 , nature , 304 , 425 salpeter , e. e. & zapolsky , h. s. 1967 , phys .
, 158 , 876 salaris , m. , garca - berro , e. , hernanz , m. , isern , j. , & saumon , d. 2000 , , 544 , 1036 saumon , d. , chabrier , g. , & van horn , h. m. 1995 , apjs , 99 , 713 savonije , g. j. , de kool , m. , & van den heuvel , e. p. j. 1986 , a&a , 155 , 51 schulz , n. s. , chakrabarty , d. , marshall , h. l. , canizares , c. r. , lee , j. c. , & houck , j. 2001 , , 563 , 941 serenelli , a. m. , althaus , l. g. , rohrmann , r. d. , & benvenuto , o. g. 2001 , , 325 , 607 shaviv , g. & kovetz , a. 1972 , a&a , 16 , 72 stevenson , d. 1991 , ara&a , 29 , 163 swank , j. h. , taam , r. e. , & white , n. e. 1984 , , 277 , 274 tauris , t. m. 1996 , , 315 , 453 ushomirsky , g. , matzner , c. d. , brown , e. f. , bildsten , l. , hilliard , v. g. , & schroeder , p. c. 1998 , apj , 497 , 253 verbunt , f. 1993 , araa , 31 , 93 warner , b. 1995 , , 225 , 249 yungelson , l. r. , nelemans , g. , & van den heuvel , e. p. , 2002 ,
a&a , 388 , 546 zapolsky , h. s. & salpeter , e. e. 1969 , apj , 158 , 809 | we discuss the mass - radius ( @xmath0-@xmath1 ) relations for low - mass ( @xmath2 ) white dwarfs ( wds ) of arbitrary degeneracy and evolved ( he , c , o ) composition .
we do so with both a simple analytical model and models calculated by integration of hydrostatic balance using a modern equation of state valid for fully ionized plasmas .
the @xmath0-@xmath1 plane is divided into three regions where either coulomb physics , degenerate electrons or a classical gas dominate the wd structure . for a given @xmath0 and central temperature , @xmath3 , the @xmath0-@xmath1 relation has two branches differentiated by the model s entropy content .
we present the @xmath0-@xmath1 relations for a sequence of constant entropy wds of arbitrary degeneracy parameterized by @xmath0 and @xmath3 for pure he , c , and o. we discuss the applications of these models to the recently discovered accreting millisecond pulsars .
we show the relationship between the orbital inclination for these binaries and the donor s composition and @xmath3 .
in particular we find from orbital inclination constraints that the probability xte j1807 - 294 can accommodate a he donor is approximately 15% while for xte j0929 - 304 , it is approximately 35% .
we argue that if the donors in ultracompact systems evolve adiabatically , there should be 60 - 160 more systems at orbital periods of 40 min than at orbital periods of 10 min , depending on the donor s composition .
tracks of our mass - radius relations for he , c , and o objects are available through the electronic version of this paper . |
given a vector field on a smooth closed manifold @xmath0 there is a corresponding dynamical system and one can investigate the closed orbits of this flow .
it is desirable to collect all closed orbits in one power series and study the algebraic topology and @xmath1-theory of this object . to do this
observe that closed orbits represent elements in @xmath2 and also in the set of conjugacy classes of @xmath3 .
we set @xmath4 .
+ in @xcite , fried defines a commutative zeta function for certain nonsingular flows as a formal power series and relates it to a reidemeister torsion invariant of the manifold .
+ the first noncommutative invariant for flows was introduced in geoghegan and nicas @xcite for suspension flows .
their analogue of a zeta function is what they call the lefschetz - nielsen series which lives in an infinite product of @xmath5-dimensional hochschild homology groups .
+ in the case of vector fields with singularities the first papers to obtain relations between zeta functions and torsion are hutchings and lee @xcite and pajitnov @xcite , both dealing with gradients of circle - valued morse functions and with commutative invariants .
in that situation the torsion invariant no longer depends only on the topology of @xmath0 but the critical points enter via the novikov complex .
both papers have been generalized , hutchings @xcite discusses closed 1-forms , still in a commutative setting , while pajitnov @xcite gets a noncommutative result for circle - valued morse functions .
+ circle - valued morse functions correspond to closed rational morse 1-forms .
this paper discusses the noncommutative case for arbitrary closed morse 1-forms .
the geometric methods largely follow pajitnov @xcite .
in fact , the geometry in @xcite is mainly contained in his earlier paper @xcite .
the main difficulty is that the algebra required to keep track of the invariants is more complicated than in the commutative case .
so instead of looking at a zeta function , pajitnov @xcite and we look at an eta function ( or pre - zeta function , compare fried @xcite ) , which generalizes the logarithm of the zeta function of the commutative case . since the conjugacy classes of @xmath6 do not form a group , we can not take the exponential function of this eta function . to compare this eta function with a certain torsion one needs a logarithm - like homomorphism @xmath7 from @xmath8 of the novikov ring to the object containing the eta function .
we depart somewhat in the definition of @xmath7 from pajitnov @xcite in that we take a detour through hochschild homology using the dennis trace , compare geoghegan and nicas @xcite .
the main theorem we get is let @xmath9 be a closed morse 1-form on a smooth connected closed manifold @xmath10 .
let @xmath11 be induced by @xmath9 and let @xmath12 be the simplicial @xmath13 complex coming from a smooth triangulation of @xmath0 .
for every @xmath14 there is a natural chain homotopy equivalence @xmath15 whose torsion @xmath16 lies in @xmath17 and satisfies @xmath18 [ intro ] this theorem was obtained by pajitnov in @xcite in the rational case . here
@xmath19 is the vector field whose eta function we look at , @xmath20 is the novikov complex , a complex over the novikov ring @xmath21 and @xmath17 a particular subgroup of @xmath22 .
the set @xmath23 is a set of @xmath24-generic vector fields which are gradient with respect to @xmath9 , see section 4 .
the chain homotopy equivalence can be described as follows : given a smooth triangulation of @xmath0 , we can adjust this triangulation so that each simplex is transverse to the unstable manifolds of the critical points of @xmath9 . then for a @xmath25-simplex @xmath26 we define @xmath27\,p\ ] ] where @xmath28 is the set of critical points of @xmath9 having index @xmath25 and @xmath29\in\widehat{\mathbb{z}g}_\xi$ ] is the intersection number of a lifting of @xmath26 to @xmath30 with translates of the unstable manifold of a lifting of the critical point @xmath31 .
this chain homotopy equivalence is basically described in hutchings and lee @xcite .
the restriction that @xmath19 lie in @xmath23 , a geometric condition due to pajitnov @xcite , then allows us to identify the torsion of @xmath32 .
this is achieved using the work of farber and ranicki @xcite and ranicki @xcite .
we choose a triangulation of @xmath0 such that @xmath32 factors through a complex @xmath33 , where @xmath34 is a @xmath13 complex which comes from a handlebody decomposition on a codimension 1 submanifold @xmath35 that separates @xmath0 and a handlebody decomposition on the cobordism obtained by splitting along @xmath35 .
it turns out that the complex @xmath34 is the mapping cone of an injective @xmath13 homomorphism which depends on the vector field @xmath19 .
after tensoring with the novikov ring the natural projection to the cokernel is a chain homotopy equivalence .
but for @xmath14 the cokernel can be identified with the novikov complex .
+ if the novikov complex is not acyclic the torsion of a chain homotopy equivalence is not determined by the complexes and we will give an example of two @xmath9-gradients @xmath36 with @xmath37 , but @xmath38 , see remark [ rm54 ] .
+ as mentioned before this paper is closely related to pajitnov @xcite .
the work of hutchings @xcite and hutchings and lee @xcite is in the same spirit , but with quite different methods .
in particular , hutchings @xcite contains a theorem ( see theorem [ htheo ] for the precise statement ) which might be considered a commutative version of theorem [ intro ] .
the role of @xmath16 is played by two reidemeister torsions .
we show in section 5 how to recover hutchings theorem for vector fields in @xmath23 as a corollary of theorem [ intro ] .
in fact we obtain a stronger `` commutative theorem '' ; see example [ examp ] .
+ the author would like to thank the referee for several suggestions , in particular for formula ( [ formula ] ) .
this paper will form a portion of the author s doctoral dissertation written at the state university of new york at binghamton under the direction of ross geoghegan .
let @xmath6 be a group and @xmath11 be a homomorphism . for a ring @xmath39
we denote by @xmath40 the abelian group of all functions @xmath41 . for @xmath42 let + supp @xmath43
. then we define @xmath44 for @xmath45 we set @xmath46 , then @xmath47 is a well defined element of @xmath48 and turns @xmath48 into a ring , the _ novikov ring__. it contains the usual group ring @xmath49 as a subring and we have @xmath50 if and only if @xmath51 is the zero homomorphism .
_ _ _ _ the _ norm _ of @xmath52 is defined to be @xmath53)\}\ ] ] _ _ _ it has the following nice properties : 1 .
@xmath54 and @xmath55 if and only if @xmath56 .
2 . @xmath57 .
3 . @xmath58 .
4 . @xmath59 . if @xmath35 is a normal subgroup of @xmath6 that is contained in @xmath60 we get a well defined homomorphism @xmath61 and a well defined ring epimorphism @xmath62 given by + @xmath63 . +
now let @xmath64 be the set of conjugacy classes of @xmath6 . again
the homomorphism @xmath51 induces a well defined map @xmath65 which we also denote by @xmath51 . in analogy with above
we define @xmath66 , but since there is no well defined multiplication in @xmath64 , this object is just an abelian group .
again there is an epimorphism @xmath67 of abelian groups .
we can think of @xmath66 as lying between @xmath48 and @xmath68 .
if @xmath69 , we denote the conjugacy class of @xmath70 by @xmath71 .
+ now we will turn our attention to @xmath72 . for the definition of @xmath8 we refer the reader to cohen @xcite or milnor @xcite .
first we disregard units of the form @xmath73 , hence look at @xmath74\rangle$ ] .
there is another type of `` elementary unit '' in @xmath21 , namely , let @xmath75 satisfy @xmath76 . then @xmath77 is a well defined element of @xmath21 and the inverse of @xmath78 .
these form a subgroup of the units in @xmath21 .
we denote the image of this subgroup in @xmath22 by @xmath17 .
let @xmath10 be a closed connected smooth manifold .
by de rham s theorem @xmath79closed @xmath80-forms on @xmath81exact @xmath80-forms on @xmath82hom@xmath83 , so a closed @xmath80-form @xmath9 induces a homomorphism @xmath84 which can be explicitely stated by the formula @xmath85 , where @xmath86 is a smooth loop representing @xmath87 .
set @xmath4 .
then @xmath6 is finitely presented , so the image of @xmath88 is a finitely generated subgroup of @xmath89 , hence isomorphic to @xmath90 for some integer @xmath25 . if @xmath91 @xmath9 is said to be _ rational _ _ , if @xmath92 it is _
irrational__. + rational @xmath80-forms can be described by circle valued functions @xmath93 in the following way : let @xmath94 be the usual covering projection , let @xmath95 be the closed @xmath80-form on @xmath96 such that @xmath97 ; then @xmath98 is a closed @xmath80-form and i m @xmath99 . to obtain other infinite cyclic subgroups of @xmath89 as images of @xmath51
one uses circles of different size .
+ now , given a rational @xmath80-form @xmath9 there is an infinite cyclic covering space @xmath100 such that @xmath101 , namely the one corresponding to @xmath102 .
let @xmath103 be the generator of the covering transformation group of @xmath104 with @xmath105 for @xmath106 .
then @xmath107 defines a map @xmath108 which induces a surjection on fundamental group .
+ notice that for irrational closed @xmath80-forms @xmath9 there is a @xmath90-covering space @xmath100 such that @xmath101 .
+ locally a closed @xmath80-form is exact .
we will call a closed @xmath80-form a _ morse form _
if @xmath9 is locally represented by the differential of real valued functions whose critical points are nondegenerate .
so if @xmath9 is a morse form , then @xmath9 has only finitely many critical points and every critical point has a well defined index .
_ _ _ _ _ _ _ _ _ let @xmath9 be a closed @xmath80-form .
a vector field @xmath19 is called an _ @xmath9-gradient _ , if there exists a riemannian metric @xmath70 such that @xmath109 for every @xmath110 and @xmath111 .
_ _ _ the next lemma allows us to forget about the riemannian metric and will be useful in using vector fields as gradients of different morse forms .
[ smalo]let @xmath9 be a morse form and @xmath19 a vector field .
then @xmath19 is an @xmath9-gradient if and only if 1 . for every critical point @xmath31 of @xmath9 there exists a neighborhood @xmath112 of @xmath31 and a riemannian metric @xmath70 on @xmath112 such that @xmath109 for every @xmath113 and @xmath114 .
2 . if @xmath115 , then @xmath116 .
the `` only if '' direction is clear . for the `` if '' direction choose disjoint neighborhoods @xmath117 , each with a riemannian metric coming from 1 . for every critical point of @xmath9
now choose finitely many contractible open sets @xmath118 with @xmath119critical points@xmath120 that together with the @xmath121 s cover @xmath0 .
using 2 . , it is easy to find a riemannian metric on each @xmath122 that turns @xmath123 into a gradient of @xmath124 . now
the required riemannian metric is obtained by using a partition of unity .
_ some authors ( e.g. milnor @xcite , pajitnov @xcite ) use a more restricted version for an @xmath9-gradient , namely , a sharper version of 1 . in the lemma .
for an even more general definition of @xmath9-gradient we refer the reader to pajitnov @xcite , which contains most of the modifications on a vector field that we will need .
_ _ _ given a morse form @xmath9 and an @xmath9-gradient @xmath19 we denote for a critical point @xmath31 of @xmath9 the unstable , resp .
stable , manifold of @xmath31 by @xmath125 , resp .
so if @xmath127 denotes the flow of @xmath19 , then @xmath128 for @xmath129 and @xmath130 for @xmath131 .
if the index of @xmath31 is @xmath132 , then @xmath125 is an immersed open disk of dimension @xmath133 and @xmath126 of dimension @xmath132 .
we say @xmath19 satisfies the _ transversality condition _ if all discs @xmath126 and @xmath134 intersect transversely for all critical points @xmath135 of @xmath9 .
+ given a morse form @xmath9 and an @xmath9-gradient @xmath19 satisfying the transversality condition we can define the _ novikov complex _ @xmath20 which is in each dimension @xmath132 a free @xmath21 complex with one generator for every critical point of index @xmath132 . here
@xmath51 is the homomorphism induced by @xmath9 .
the boundary homomorphism of @xmath20 is based on the number of trajectories between critical points of adjacent indices . for more details
see pajitnov @xcite or latour @xcite .
this chain complex is chain homotopy equivalent to @xmath136 , where @xmath137 is the simplicial chain complex of the universal cover @xmath30 of @xmath0 with respect to a smooth triangulation of @xmath0 lifted to @xmath30 .
but we are mainly interested in the universal covering .
] furthermore there is a chain homotopy equivalence whose torsion is in @xmath138 . in the rational case
this is proven in pajitnov @xcite , for the general case see latour @xcite .
the map described in ( [ formula ] ) in the introduction can be used for this .
we will show this in [ sb4.1 ] at least for an @xmath9-gradient @xmath19 satisfying a `` cellularity condition '' .
+ let us discuss this map .
a smooth triangulation @xmath139 of @xmath0 is called _ adjusted to _
@xmath19 , if every @xmath25-simplex @xmath26 intersects the unstable manifolds @xmath125 transversely for all critical points @xmath31 of index @xmath140 . to see the existence , assume @xmath9 is rational , the general case follows by approximation , see [ sb4.2 ] .
a triangulation @xmath139 lifts to a triangulation of @xmath104 , an infinite cyclic covering space , compare [ chtytp ] . for a diffeomorphism @xmath141 of @xmath0
we denote by @xmath142 the triangulation of @xmath0 where simplices are composed with @xmath141 .
if we change the triangulation of @xmath0 by an isotopy , we can get transverse intersections in @xmath104 of lifted simplices with finitely many unstable manifolds by the results of [ app1 ] in the appendix .
since the results there give openness and density among diffeomorphisms we get a generic set of diffeomorphisms @xmath141 of @xmath0 isotopic to the identity such that @xmath142 is adjusted to @xmath19 . + given an adjusted triangulation
@xmath139 we get a chain map @xmath143 by formula ( [ formula ] ) . that @xmath32 is
indeed a chain map follows from the exact case , which is described in the appendix .
lemma [ alem2 ] also carries over so that different adjustments to a triangulation lead to chain homotopic maps . finally , if @xmath144 is a subdivision of @xmath139 such that @xmath145 is adjusted to @xmath19 ,
so is @xmath142 and the diagram @xmath146 \varphi(v)\searrow & & \swarrow\varphi(v)\\[0.2 cm ] \multicolumn{3}{c}{c_\ast(\omega , v)}\end{array}\ ] ] commutes .
so once we show that @xmath32 is a chain homotopy equivalence , its torsion does not depend on the triangulation .
_ _ _ _ _ _ the following is a construction of farber and ranicki @xcite written to fit our purposes , compare also pajitnov @xcite . + given a circle valued morse map @xmath93 which induces a surjection on fundamental group we get a lifting @xmath148 where @xmath104 is an infinite cyclic covering space .
assuming that @xmath149 is a regular value , set @xmath150 , @xmath151)$ ] .
we get a handle decomposition of the cobordism @xmath152 from the morse function @xmath153 $ ] . by choosing a cell decomposition of @xmath35 , farber and ranicki @xcite
construct a finitely generated free @xmath13 complex @xmath34 homotopy equivalent to @xmath137 .
let us recall the construction from @xcite .
let @xmath154 be the universal covering projection and let @xmath155 be @xmath156 .
the cell decomposition of @xmath35 leads to a cell decomposition of @xmath157 and the resulting cell complex , denoted by @xmath158 , is a finitely generated free @xmath13 complex .
let @xmath159 be the number of @xmath132-cells in @xmath35 .
now @xmath160 is a finitely generated free @xmath13 complex containing @xmath158 as a subcomplex and the remaining generators correspond to critical points of @xmath161 .
+ the inclusions @xmath162 and @xmath163 induce @xmath13 chain maps @xmath164 and @xmath165 and since @xmath158 is a subcomplex , @xmath164 is a split injection .
then @xmath166 is the mapping cone of @xmath167 .
in particular rank @xmath168 critical points of @xmath169 having index @xmath132 . in [ sb4.1 ] we will use the geometry of pajitnov @xcite to get a more detailed version of this chain complex .
+ let @xmath39 be a ring and @xmath170 a ring homomorphism such that + id@xmath171 proj@xmath172 is an automorphism .
then by the deformation lemma of farber and ranicki @xcite , see also ranicki ( * ? ? ?
1.9 ) , the chain complex + @xmath173 is chain homotopy equivalent to coker@xmath174id@xmath175 , a finitely generated free @xmath39 complex with rank @xmath176 critical points of @xmath169 having index @xmath132 .
in fact the chain equivalence is identified in @xcite to be the natural projection + @xmath177 coker@xmath174id@xmath178 .
+ so to use the deformation lemma one has to turn a certain square matrix @xmath179 representing proj@xmath180 over @xmath13 into an invertible matrix over a ring @xmath39 .
the matrix @xmath181 can be chosen to satisfy @xmath182 for every entry of @xmath181 , where @xmath51 is induced by @xmath169 .
obvious candidates for @xmath39 are the noncommutative localization used by farber and ranicki @xcite and the novikov ring @xmath21 .
a not so obvious candidate is a novikov ring @xmath183 where @xmath184 is `` close '' to @xmath51 ; this will be discussed in section 4 . _
farber @xcite has extended the deformation lemma of @xcite to the case of closed @xmath80-forms using a certain noncommutative localization .
_ _ _ furthermore , ranicki ( * ? ? ?
1.9 ) contains the calculation of the torsion of the chain homotopy equivalence @xmath185 .
it is given by @xmath186 let @xmath19 be a vector field . by a closed orbit of @xmath19 we mean a nonconstant map @xmath187 with @xmath188 .
the _ multiplicity _
@xmath189 is the largest positive integer @xmath190 such that @xmath86 factors through an @xmath190-fold covering @xmath191 .
we say two closed orbits are _ equivalent _ if they only differ by a rotation of @xmath96 .
we denote the set of equivalence classes by @xmath192 . notice that @xmath193 gives a well defined element @xmath194 .
a closed orbit @xmath86 is called _ nondegenerate _ if @xmath195 , where @xmath196 is a poincar map corresponding to @xmath86 . in that case
we define @xmath197 to be the sign of @xmath198 .
is the fixed point index of @xmath196 at the isolated fixed point coming from the closed orbit . ] + now let @xmath9 be a morse form .
we denote by @xmath199 the set of all @xmath9-gradients that satisfy the transversality condition and whose closed orbits are nondegenerate .
for @xmath200 we define the _ eta - function _ of @xmath201 to be the element of @xmath202 defined by @xmath203 again @xmath51 is induced by @xmath9 . for the proof that @xmath204 is a well defined element of @xmath202
we refer the reader to hutchings @xcite .
_ _ _ _ _ _ _ _
let @xmath39 be a ring and @xmath205 an @xmath39-algebra . for an @xmath206 bimodule @xmath0
we define the _ hochschild chain complex _ @xmath207 by @xmath208 , where the product contains @xmath209 copies of @xmath205 and the tensor products are taken over @xmath39 .
the boundary operator is given by @xmath210 the @xmath209-th hochschild homology group of @xmath205 with coefficients in @xmath0 is denoted by @xmath211 . if @xmath212 and the bimodule structure is given by ordinary multiplication we write @xmath213 instead of @xmath214 .
we will mainly be interested in the case where @xmath215 , @xmath216 is a novikov ring and @xmath217 .
a useful observation is that @xmath218 and hence classes represented by @xmath219 are automatically @xmath5 in @xmath220 .
+ given an @xmath221 matrix @xmath222 over @xmath205 and an @xmath223 matrix @xmath224 over @xmath0 we define an @xmath225 matrix @xmath226 with entries in @xmath227 by setting @xmath228 .
the trace of this matrix , trace @xmath226 , is given by @xmath229 and is an element of @xmath230 , it is a cycle if and only if trace@xmath231 trace@xmath232 . for more on hochschild homology see geoghegan and nicas @xcite or igusa @xcite . _ _
for a ring @xmath39 with unit , which , in view of subsection 3.1 , we can think of as a @xmath233-algebra , there is the dennis trace homomorphism @xmath234 defined as follows : if @xmath235 is represented by the matrix @xmath181 , then + @xmath236trace @xmath237\in h\!h_1(r)$ ] , see igusa @xcite .
it is easy to see that the dennis trace factors through @xmath238\rangle$ ] .
+ we want to define a homomorphism @xmath239 . to do this define + @xmath240 by @xmath241 we can think of @xmath190 as a weighted combination of multiplication in @xmath21 and the augmentation @xmath242 .
the homomorphism @xmath190 induces a homomorphism @xmath243 .
it is to be shown that @xmath190 vanishes on boundaries .
so let @xmath244 satisfy @xmath245 , then @xmath246 @xmath247 since @xmath248 and @xmath51 is a homomorphism .
for @xmath69 we have @xmath249 , therefore the composition @xmath250 factors through @xmath22 , call this homomorphism @xmath251 .
we want to examine how this homomorphism behaves on @xmath17 . for future purposes it will be useful not just to look at @xmath252 matrices . _ an @xmath225 matrix @xmath181 over @xmath21 is called _
@xmath51-regular _ , if there exists @xmath253 such that @xmath254 for all @xmath255 , @xmath256 , @xmath257 .
_ _ _ for example , a matrix @xmath181 which satisfies @xmath258 for every entry is @xmath51-regular , but @xmath51-regular matrices can have entries @xmath259 with @xmath260 .
the motivation for @xmath51-regular matrices comes from our approximation arguments , see remark [ apprem ] .
let @xmath181 be a @xmath51-regular matrix .
then 1 .
the matrix @xmath179 is invertible over @xmath21 and the inverse is given by @xmath261 .
2 . denote the image of @xmath179 in @xmath22 by @xmath262 ; then @xmath263 .
we need to show that @xmath261 is a well defined matrix over @xmath21 .
note that @xmath264 .
we will look at terms of the form + @xmath265 and get an estimate for @xmath266 .
the idea is to write @xmath267 as a word @xmath268 where the length of the word @xmath269 is smaller than @xmath209 and the words @xmath270 are of the form @xmath271 .
+ so assume that @xmath265 where @xmath272 .
let @xmath273 be the first index whose value appears more than once .
since these numbers are between @xmath80 and @xmath209 we have @xmath274 .
let @xmath25 be the largest number such that @xmath275 , then @xmath276 where @xmath277 , @xmath278 , @xmath279 , @xmath280 and + @xmath281 .
notice that @xmath282 .
now look at @xmath283 ; among the indices @xmath284,@xmath285 are at most @xmath286 numbers ; the numbers @xmath287,@xmath273 , which are all different , do not appear .
if @xmath288 , one of these numbers will appear more than once .
let @xmath289 be the first such index and @xmath290 the last index equal to @xmath289 .
again we get @xmath291 , hence @xmath292 . as above
we can write @xmath293 we continue this process until we get @xmath294 with @xmath295 .
since @xmath267 consists of @xmath296 letters we get @xmath297 .
let @xmath298 be a number such that @xmath299 for all @xmath300 .
then @xmath301 .
since @xmath302 as @xmath303 , @xmath304 is a well defined element of @xmath21 .
the argument is the same as in pajitnov ( * ? ? ?
1.2 ) . using elementary row reductions we obtain a matrix of the form @xmath305 where @xmath306 is an @xmath307 matrix which is again @xmath51-regular with the same @xmath1 .
induction gives the result .
[ logar]let @xmath181 be a @xmath51-regular matrix over @xmath21 , then @xmath308 in particular , @xmath309 induces a homomorphism @xmath239 . as before denote the image of @xmath179 in @xmath310 by @xmath262 .
then @xmath311\\ & = & -\mu\,[\mbox{trace } ( a\otimes\sum_{m=0}^\infty a^m)]\hspace{2cm}\mbox{since } 1\otimes x\mbox { is a boundary}\\ & = & -\sum_{m=0}^\infty\sum_{i , k=1}^n\mu\,[a_{ik}\otimes a^m_{ki}].\end{aligned}\ ] ] it is sufficient to show that @xmath312\,=\,\varepsilon\left(\frac{\mbox{trace } a^{m+1}}{m+1}\right)\ ] ] both sides are clearly @xmath5 for @xmath244 with @xmath313 .
call the left side of ( [ lside ] ) @xmath314 and let @xmath244 with @xmath245
. then @xmath315 let @xmath316 act on @xmath317 by the cycle @xmath318 and on @xmath319 by rotation . for @xmath320 denote by @xmath321 $ ] the orbit of @xmath322 and by @xmath205 the orbit set .
we get @xmath323\in s}\sum_{(i_1,\ldots\hspace{-1pt},i_{m+1})\in [ x]}\sum_{h_1,\ldots\hspace{-1pt},h_{m+1}\in g \atop \{h_1\cdots h_{m+1}\}=\gamma}\sum_{t\in\mathbb{z}_{m+1}}\xi(h_{t1})a_{i_1i_2}(h_{t1})\cdots a_{i_{m+1}i_1}(h_{t(m+1)})\\ & = & \frac{1}{m+1}\frac{1}{\xi(\gamma)}\sum_{[x]\in s}\sum_{h_1,\ldots\hspace{-1pt},h_{m+1}\in g \atop \{h_1\cdots h_{m+1}\}=\gamma}\sum_{t\in\mathbb{z}_{m+1}}\xi(h_{t1})\sum_{(i_1,\ldots\hspace{-1pt},i_{m+1})\in [ x]}a_{i_1i_2}(h_{t1})\cdots a_{i_{m+1}i_1}(h_{t(m+1)})\end{aligned}\ ] ] now @xmath324}a_{i_1i_2}(h_{t1})\cdots a_{i_{m+1}i_1}(h_{t(m+1)})\,=\,\sum_{(i_1,\ldots\hspace{-1pt},i_{m+1})\in [ x]}a_{i_1i_2}(h_1)\cdots a_{i_{m+1}i_1}(h_{m+1}),\ ] ] since the orbit is obtained by shifting @xmath325 , so @xmath326\in s}\sum_{h_1,\ldots\hspace{-1pt},h_{m+1}\in g \atop \{h_1\cdots h_{m+1}\}=\gamma}\sum_{(i_1,\ldots\hspace{-1pt},i_{m+1})\in [ x]}a_{i_1i_2}(h_1)\cdots a_{i_{m+1}i_1}(h_{m+1})\sum_{t\in\mathbb{z}_{m+1}}\xi(h_{t1})\\ & = & \frac{1}{m+1}\frac{\xi(\gamma)}{\xi(\gamma)}\sum_{[x]\in s}\sum_{(i_1,\ldots\hspace{-1pt},i_{m+1})\in [ x]}\sum_{h_1,\ldots\hspace{-1pt},h_{m+1}\in g
\atop \{h_1\cdots h_{m+1}\}=\gamma}a_{i_1i_2}(h_1)\cdots a_{i_{m+1}i_1}(h_{m+1})\\ & = & \frac{1}{m+1}\sum_{i_1,\ldots\hspace{-1pt},i_{m+1}=1}^n\sum_{h_1,\ldots\hspace{-1pt},h_{m+1}\in g \atop \{h_1\cdots h_{m+1}\}=\gamma}a_{i_1i_2}(h_1)\cdots a_{i_{m+1}i_1}(h_{m+1})\\ & = & \frac{1}{m+1}\,\varepsilon(\mbox{trace } a^{m+1})(\gamma).\end{aligned}\ ] ] in the case where @xmath51 is a homomorphism to the integers it is now easily seen that @xmath7 agrees with pajitnov s @xmath7 from @xcite on @xmath17 once the correct identifications are made .
in @xcite , pajitnov defines a condition @xmath327 , in @xcite denoted by @xmath328 , for an @xmath169-gradient @xmath19 , where @xmath93 is a morse function that induces a surjection @xmath51 on fundamental group . for the full condition we refer the reader to these papers , but
informally , it can be described as follows : just as in [ chtytp ] we get a cobordism @xmath329 and a morse map @xmath330,\{0\},\{b\})$ ] .
here we use an arbitrary @xmath331 instead of just @xmath332 to indicate that @xmath169 might come from a rational morse form . now the condition @xmath327 requires a morse map @xmath141 on @xmath35 and a @xmath141-gradient @xmath333 which gives a handle decomposition on @xmath35 and @xmath334 . the vector field @xmath19 which lifts to a vector field @xmath335 on @xmath336 now has to satisfy a `` cellularity condition '' : whenever @xmath31 is a critical point of @xmath107 of index @xmath132 , it should be the case that some thickening of @xmath126 is attached to the union of the @xmath337-handles of @xmath35 and of @xmath336 .
also a thickening of an @xmath132-handle in @xmath334 has to flow under @xmath338 into the @xmath132-skeleton of @xmath35 and @xmath336 .
a symmetric condition holds for right hand discs and handles of @xmath35 .
+ by pajitnov ( * ? ? ?
* prop.5.4 ) , the set of @xmath169-gradients satisfying @xmath339 and the transversality condition is @xmath24-open and dense in the set of @xmath169-gradients that satisfy the transversality condition .
such gradients should be thought of as cellular approximations of arbitrary @xmath169-gradients . +
now let @xmath340 be the universal cover , @xmath155 the lifting of @xmath169 and @xmath341)$ ] .
as in [ app2 ] we get a filtration of @xmath336 which we denote by @xmath342 so that @xmath343 gives a free @xmath344 complex calculating @xmath345 , where @xmath346 .
we also have @xmath347 .
denote @xmath348 and similarly for other @xmath344 complexes .
+ the flow of @xmath201 on @xmath336 induces a chain map + @xmath349 by starting in @xmath350 and flowing into @xmath351 .
see appendix [ app2 ] for the sets @xmath352 and details on this flowing .
notice that @xmath353 is given through condition @xmath327 .
the map @xmath354 is the homological gradient descent of pajitnov @xcite .
choose a basis of @xmath355 by lifting critical points of @xmath141 and let the matrix @xmath356 represent the homomorphism @xmath354 in dimension @xmath132 .
since we can choose the liftings within @xmath357 , we get @xmath358 for the entries of every matrix @xmath356 .
[ ptor ] let @xmath19 be an @xmath169-gradient satisfying the transversality condition and @xmath327 .
then @xmath32 is a chain homotopy equivalence and @xmath359 let @xmath139 be a triangulation of @xmath0 which has @xmath35 as a subcomplex @xmath144 .
this induces a triangulation @xmath360 of @xmath336 which has two copies of @xmath144 as subcomplexes .
denote the one corresponding to @xmath35 by @xmath361 and the one corresponding to @xmath334 by @xmath362 .
assume @xmath139 has the following properties : 1 .
@xmath139 is adjusted to @xmath19 .
@xmath144 is adjusted to @xmath333 .
3 . there is an @xmath363 such that if @xmath26 is a @xmath25-simplex in @xmath144 , then @xmath364 .
there is an @xmath363 such that if @xmath365 and the trajectory of @xmath201 starting at @xmath322 ends in @xmath366 , then @xmath367 .
the existence of such a triangulation is shown in [ lastap ] .
+ [ app2 ] gives a simple chain homotopy equivalence @xmath368 let us define a chain map @xmath369 .
if @xmath26 is a simplex in @xmath144 , look at the liftings @xmath370 and @xmath371 used for the basis of @xmath372 .
there is exactly one @xmath69 such that @xmath373 in @xmath30 .
set @xmath374 .
look at the diagram @xmath375 because of property 4 . above , this diagram commutes .
therefore the map @xmath376 is a simple homotopy equivalence between the mapping cones @xmath377 and @xmath378 , where @xmath132 represents inclusion .
but by the deformation lemma of farber and ranicki @xcite , @xmath377 is chain homotopy equivalent to @xmath379 , in fact simple homotopy equivalent by ranicki ( * ? ? ?
* prop.1.9 ) ( the corresponding matrix term is just @xmath380 ) . but
@xmath379 is easily seen to be @xmath381 .
+ after tensoring with the novikov ring we have the following sequence of chain homotopy equivalences @xmath382 and all except the last one are simple .
the first map is described in the proof of ranicki ( * ? ? ?
* prop.1.7 ) .
because of the special form of the vector field @xmath19 the novikov complex @xmath383 can be identified with @xmath384 , see ranicki ( * ? ? ?
* remark 4.8 ) and pajitnov ( * ? ? ?
* remark 7.3 ) .
we claim that this composition is exactly @xmath32 .
denote the composition by @xmath385 .
+ we denote @xmath386 $ ] .
let @xmath387 , lift it to @xmath388 ( if @xmath26 is a cell in @xmath357 , lift it to @xmath389 ) .
then @xmath390=[\varphi_n(\bar{\sigma}),\varphi_r(\bar{\sigma})]\in { \rm coker}(i - k)=c_\ast(\omega , v),\ ] ] where @xmath391 and @xmath392 are defined to give + @xmath393 .
+ now @xmath394=[\varphi_n(\bar{\sigma}),0]+[0,\varphi_r ( \bar{\sigma})]$ ] .
@xmath395 represents the part of @xmath396 that flows into critical points of index @xmath25 in @xmath397 under @xmath201 while @xmath398 represents the part that flows into @xmath357 .
now @xmath399 with @xmath400 we thus get @xmath401=[0,k_2(\sum_{m=0}^\infty k_1^m(\varphi_n(\bar{\sigma}))].\ ] ] but @xmath402 represents the part of @xmath396 that flows into critical points of index @xmath25 in @xmath403 under @xmath201 . therefore @xmath404 and @xmath32 is a chain homotopy equivalence whose torsion is given by ( [ torsion ] ) . _
pajitnov @xcite obtains a chain homotopy equivalence + @xmath405 by including the novikov complex into a complex @xmath406 simple homotopy equivalent to @xmath407 , in fact the map from @xmath406 to @xmath408 is just a simple change of basis , compare @xcite . the composition of this equivalence with ranicki s equivalence @xmath409 is readily seen to be the identity on @xmath383 .
_ in this subsection we describe a useful method due to pajitnov @xcite . given a morse form @xmath9 and an @xmath9-gradient @xmath19 , the induced homomorphism @xmath410 splits @xmath2 as @xmath411 .
choose @xmath412 so that the images @xmath413 generate the @xmath90 part .
now let @xmath414 be closed @xmath80-forms with @xmath415 and @xmath416 vanishes on @xmath417 .
then @xmath418 vanishes on @xmath102 and satisfies @xmath419 .
furthermore the closed @xmath80-forms can be chosen to vanish in a neighborhood of the critical points of @xmath9 . + for @xmath420 we can now define @xmath421 .
by choosing the @xmath422 small we can make sure that the @xmath9-gradient @xmath19 is also an @xmath423-gradient . to see this notice that in a neighborhood of the critical points of @xmath9 the new form agrees with @xmath9 .
denote the complement of this neighborhood by @xmath424 .
because of the compactness of @xmath424 and lemma [ smalo ] there is a @xmath425 such that @xmath426 for all @xmath427 .
now the @xmath422 have to be chosen so small that @xmath428 for all @xmath427 which is possible again by compactness .
lemma [ smalo ] now gives that @xmath19 is an @xmath423-gradient .
+ we have @xmath429 , so we can also choose the @xmath422 to have @xmath430 factor through @xmath431 . hence we get [ approx]for a morse form @xmath9 and an @xmath9-gradient @xmath19 there exists a rational morse form @xmath432 with the same set of critical points and that agrees with @xmath9 in a neighborhood of these critical points such that @xmath19 is also an @xmath432-gradient .
let us compare the novikov complexes we obtain for a morse form @xmath9 and a rational approximation @xmath432 that both use the same vector field @xmath19 .
the complexes are taken over different rings , @xmath433 and @xmath434 respectively .
but for two critical points @xmath135 of adjacent index the elements @xmath435 and @xmath436 agree when viewed as elements of @xmath437 since both count the number of flowlines between @xmath438 and translates of @xmath439 , and these only depend on @xmath19 .
so we can compare chain complexes even though they are over different rings .
this is an important observation and will remain useful in the next subsection .
again let @xmath9 be a morse form .
an @xmath9-gradient @xmath19 satisfies the condition @xmath440 , if there exists a rational morse form @xmath432 such that @xmath19 is an @xmath432-gradient and as such it satisfies @xmath327 .
we think of this condition as `` approximately @xmath327 '' .
+ we want to carry over the density results of pajitnov @xcite .
then @xmath24-density in @xmath199 can be seen as follows : given an @xmath9-gradient @xmath335 there is by lemma [ approx ] a rational morse form @xmath432 that agrees with @xmath9 near the critical points and such that @xmath335 is also an @xmath432-gradient .
now the density of @xmath432-gradients satisfying @xmath327 allows us to choose a vector field @xmath19 as close as we like to @xmath335 . to see that we can find an @xmath9-gradient this way observe that 1 . of lemma [ smalo ]
is trivially fulfilled and since @xmath441 away from a neighborhood of the critical points we get @xmath442 by choosing @xmath19 close enough to @xmath335 .
therefore @xmath19 is an @xmath9 gradient satisfying @xmath440 .
the @xmath24-openness now follows from pajitnov ( * ? ? ?
* prop.5.4 ) . + now if an @xmath9-gradient @xmath19 satisfies @xmath440 , let @xmath432 be the rational morse form as in the definition and denote by @xmath11 and @xmath443 the homomorphisms induced by @xmath9 and @xmath432 . for the rational form @xmath432 we can form @xmath357 , the @xmath13 complex @xmath378 , the homomorphism @xmath444 and the matrix @xmath356 just as in [ sb4.1 ] and the proof of proposition [ ptor ] . if we can show that @xmath445 is invertible over @xmath21 we get the chain homotopy equivalence between @xmath446 and coker@xmath174id@xmath447 . we know from 4.1 that @xmath445 is invertible over @xmath183 and that the cokernel over this novikov ring is exactly the novikov complex @xmath448 .
in order to keep the notation simple denote the matrix @xmath356 by @xmath449 .
[ apprem]_that @xmath450 is invertible over @xmath183 is easily seen since a basis can be chosen so that @xmath451 for every entry @xmath452 .
if we could choose a basis of @xmath158 such that @xmath453 we would also immediately get that @xmath450 is invertible over @xmath21 . a similar argument
is used in latour @xcite .
but since we have to choose liftings of cells instead of critical points it is not clear that a nice basis can be chosen .
so instead of trying to find a nice basis we use the notion of @xmath51-regular matrices .
_ the matrix @xmath449 over @xmath13 is @xmath51-regular.[regul ] we have chosen a basis of @xmath454 by choosing @xmath132-cells in @xmath357 , call these cells @xmath455 . if @xmath456 supp @xmath457 , then there exist negative flowlines from @xmath458 to @xmath459 by the construction of @xmath449 . +
we need to show that there exists a @xmath253 with the property that given @xmath255 and indices @xmath460 and @xmath461 supp @xmath462 , @xmath463 supp @xmath464 supp @xmath465 we have @xmath466 .
+ now we have to recall the proof of the main theorem in pajitnov @xcite .
every cell @xmath455 defines a thickened sphere in @xmath467}/\tilde{v}_l^{(i-1)}$ ] that we denote by @xmath468 , also let @xmath469 .
the spaces @xmath470}$ ] and @xmath471 are defined in pajitnov ( * ? ? ?
* , 4.5 ) .
since @xmath461 supp @xmath462 we have that @xmath472 induces a homologically nontrivial map from @xmath473 to @xmath474 .
similarly every @xmath475 supp @xmath476 gives rise to a homologically nontrivial map from @xmath477 to @xmath478 .
the composition of all these maps plus @xmath479 is homologically nontrivial and hence has a fixed point other than the base point , compare the proof of lemma 5.1 in @xcite .
notice that the existence of this fixed point does not require the condition that closed orbits of @xmath19 are nondegenerate .
this fixed point corresponds to a flow line @xmath480\to\tilde{m}$ ] of @xmath472 with @xmath481 and @xmath482 which passes through the cells @xmath483 .
+ we need the following there exists a @xmath253 such that for every flowline @xmath86 of @xmath472 that starts in @xmath484 and ends in @xmath485 we have @xmath486 .
[ easyl ] we have @xmath487 and @xmath488 with @xmath489 as in 4.1 .
since @xmath432 is rational and @xmath490 has no critical points in @xmath484 there is a @xmath491 such that @xmath492)$ ] also contains no critical points .
so if @xmath493 is a flowline of @xmath472 with @xmath494 , there is a @xmath495 which depends smoothly on @xmath31 such that @xmath496 .
now @xmath497}\omega=\int\limits_0^{t_p}\omega_{\rho\circ\gamma_p(s)}(-v(\rho\circ\gamma_p(s)))=-\int\limits_0^{t_p}\omega_{\rho\circ\gamma_p(s)}(v(\rho\circ\gamma_p(s)))<0\ ] ] by lemma [ smalo ] .
since @xmath498 acts cocompactly on @xmath484 and the value of the integral depends smoothly on @xmath499 there is @xmath253 such that @xmath497}\omega\leq k.\ ] ] this @xmath1 now works for the lemma , since integrating over a longer flowline will just make the integral smaller .
conclusion of the proof of [ regul ] : our flowline @xmath86 is the concatenation of flowlines @xmath500 to which lemma [ easyl ] applies .
let @xmath155 satisfy @xmath501 .
then we get @xmath502 therefore @xmath503 for all @xmath461 supp @xmath504 supp @xmath465 which implies that @xmath449 is @xmath51-regular .
define @xmath505 satisfies @xmath506 . by the remarks above ,
@xmath23 is @xmath24-dense in @xmath199 .
[ maint ] let @xmath9 be a morse form on a smooth connected closed manifold @xmath10 .
let @xmath11 be induced by @xmath9 and let @xmath12 be the simplicial @xmath13 complex coming from a smooth triangulation of @xmath0 .
for every @xmath14 there is a natural chain homotopy equivalence @xmath507 given by ( [ formula ] ) whose torsion @xmath16 lies in @xmath17 and satisfies @xmath18 since @xmath14 we can form the @xmath13 complex @xmath378 from the proof of proposition [ ptor ] which is simple homotopy equivalent to @xmath12 .
the matrices @xmath356 are @xmath51-regular by proposition [ regul ] , so the projection of @xmath508 coker@xmath174id@xmath509 is a chain homotopy equivalence .
we have already seen that the boundary homomorphisms of @xmath20 and @xmath448 are the same when viewed as matrices over @xmath437 .
the same holds for coker@xmath174id@xmath509 and coker@xmath174id@xmath510 .
but since we identified coker@xmath174id@xmath510 with @xmath448 we now get that coker@xmath174id@xmath509 is the same complex as @xmath20 .
also , if we use the triangulation from the proof of proposition [ ptor ] we get that @xmath32 factors as @xmath511 , where @xmath512 is simple
. therefore @xmath513 by proposition [ logar ] we have @xmath514 by the proof of the main theorem in 5 of @xcite the right hand side is exactly @xmath204 .
of course , @xcite only shows this in the rational case , but @xmath204 is independent of the homomorphism @xmath51 when viewed as an element of @xmath515 .
as mentioned in the introduction , for singular vector fields one of the first formulas to relate the torsion of the novikov complex to a zeta function appeared in hutchings and lee @xcite , a generalization appeared in hutchings @xcite looking similar to theorem [ maint ] , but using quite different methods .
in this section we will relate these results , in fact we will show that theorem [ maint ] implies ( * ? ? ?
* theorem b ) , at least for gradients satisfying condition @xmath440 .
+ all these papers deal with commutative invariants only , so let @xmath516 be the universal abelian cover of @xmath0 and @xmath517 the covering transformation group .
let @xmath9 be a morse form and @xmath19 an @xmath9-gradient satisfying the transversality condition .
the novikov complex in hutchings @xcite is given by @xmath518 .
similarly @xmath519 and therefore a chain equivalence @xmath520 induces a chain equivalence @xmath521 id@xmath522 with @xmath523 .
+ let @xmath524 be the localization of @xmath525 along non - zero divisors .
it is known that @xmath524 is a finite direct product of fields , @xmath526 , see hutchings ( * ? ? ? *
lemma a.4 ) or geoghegan and nicas ( * ? ? ?
* lemma 7.8 ) .
denote @xmath527 for the projection .
+ to define torsion in the sense of hutchings @xcite , we need one more construction .
let @xmath39 be a commutative ring with unit and @xmath528 a subgroup of @xmath529 , the group of units of @xmath39 .
we say that two elements of @xmath39 are _ equivalent _ _ , @xmath530 , if there exists a @xmath531 such that @xmath532 .
we denote by @xmath533 the set of equivalence classes .
the multiplication on @xmath39 turns @xmath533 into a semigroup which contains @xmath534 as a subgroup . _ _ _ _ ( * ? ? ?
* def.a.1 ) let @xmath535 be a field and @xmath536 a finite complex over @xmath535 with a fixed basis and @xmath528 a subgroup of @xmath537 .
then the _ reidemeister torsion _ of @xmath536 is defined to be @xmath538 here @xmath539 is whitehead torsion . _
_ _ we take the inverse of the determinant , because hutchings @xcite uses a different sign convention for torsion . when the group of units is clear
, we will suppress it in the notation of the torsion .
+ now @xmath540 is a subgroup of @xmath541 .
denote @xmath542 for @xmath543
. then + @xmath544 .
hutchings @xcite defines @xmath545 and @xmath546 notice that @xmath547 , and hence @xmath548 , can only be nonzero if @xmath549 .
equals the euler characteristic of @xmath0 and since @xmath550 is a field it can be calculated from the homology of that complex . ]
these torsions are related by the zeta function of @xmath201 which is defined as follows .
let + @xmath551 .
similarly we get @xmath552 and @xmath553 .
define @xmath554 and @xmath555 by @xmath556 it is readily seen that @xmath557 and @xmath558 are well defined and mutually inverse to each other .
+ if @xmath559 , then @xmath204 is an element of @xmath560 and we define the _ zeta function _ of @xmath201 to be @xmath561 hutchings relates @xmath548 and @xmath547 by _ _
* theorem b)[htheo ] let @xmath9 be a morse form on the closed connected smooth manifold @xmath0 , let @xmath559 and let @xmath562 be given by inclusion and projection
. then @xmath563 .
we will show how to derive theorem [ htheo ] from theorem [ maint ] for @xmath14
. it would be desirable to extend theorem [ maint ] to @xmath559 , possibly by the methods of @xcite .
+ so let @xmath15 be the chain homotopy equivalence from theorem [ maint ] . as seen above , this induces a chain homotopy equivalence + @xmath564 .
the homomorphism @xmath7 actually gives a map @xmath565 .
it is easy to see that the following diagram commutes .
\big\downarrow&\multicolumn{5}{c } { } & \big\downarrow\\[0.3 cm ] \overline{k}_1^g(\widehat{\mathbb{z}g}_\xi)&\stackrel{\varepsilon_\ast}{\longrightarrow}&\multicolumn{3}{c}{\overline{k}_1^h(\widehat{\mathbb{z}h}_{\bar{\xi}})}&\stackrel{\det}{\longrightarrow}&\widehat{\mathbb{q}h}^\ast_{\bar{\xi}}/\pm h\end{array}\ ] ] the last vertical arrow is inclusion of @xmath567 into @xmath568 followed by projection . hence we get @xmath569 .
look at the commutative diagram @xmath570 \big\downarrow\det & & \big\downarrow\det & & \big\downarrow\oplus\det\\[0.3 cm ] \widehat{\mathbb{q}h}^\ast_{\bar{\xi}}/\pm h&\stackrel{\bar{\iota}}{\longrightarrow}&q\widehat{\mathbb{z}h}^\ast_{\bar{\xi}}/\pm h&\stackrel{\simeq}{\longrightarrow}&\bigoplus\limits_{j=1}^k f_j^\ast/\pm h_j \end{array}\ ] ] we will show that @xmath571 we have to compare @xmath572 with @xmath573 and @xmath574 .
+ if @xmath575 is acyclic , then so is @xmath576 , because id@xmath577 is an equivalence of these complexes .
furthermore @xmath578 so @xmath579id@xmath580 .
this gives by ( [ commd ] ) and theorem [ maint ] @xmath581 if @xmath575 is not acyclic , neither is @xmath576 and ( [ formu ] ) reduces to @xmath582 .
hence we obtain the desired formula @xmath563 .
+ since the complexes @xmath575 do not always have to be acyclic , theorem [ htheo ] can not recover the zeta function in general just from the torsion information .
in particular , theorem [ htheo ] contains no information for @xmath583 . to see that we can get reasonable results for @xmath583 we have the following _
let @xmath0 be a closed surface of genus 2 and @xmath93 a morse map indicated by figure [ first ] , _
i.e. we take a projection @xmath584 and add a @xmath80-handle which only gets mapped to one half of @xmath96 .
this @xmath169 has 2 critical points , both of index 1 .
with the loops in figure [ first ] we have @xmath585[a_2,b_2]=1\rangle$ ] .
the homomorphism @xmath586 is then given by @xmath587 and all other generators are send to @xmath5 .
+ with the construction of [ chtytp ] we get @xmath588 and @xmath336 as in figure [ secon ] .
we can choose @xmath35 so that it is the image of the loop @xmath589 , the basepoint being on the bottom .
put an @xmath169-gradient @xmath19 on @xmath0 so that the trajectories of @xmath201 starting and ending at critical points are as in figure [ secon ] .
+ we need @xmath19 to satisfy condition @xmath327 .
the morse map @xmath590 is chosen as the height function , so we have a minimum and a maximum .
if the thickenings of the critical points on @xmath35 are chosen to fill about half of the circle it is clear that we can find a @xmath19 that satisfies @xmath327 .
now we can also get a @xmath591 with trajectories as in figure [ secon ] .
+ let @xmath592 be the chain equivalence from theorem [ maint ] . to calculate @xmath16 we have to look at the @xmath252 matrices @xmath593 and
@xmath594 that come from the negative gradient descent .
all trajectories that start in @xmath334 and are not drawn in figure [ secon ] flow to @xmath35 and can not cross each other . to calculate @xmath593 notice that trajectories starting in the lower half of @xmath334 follow the loop that represents @xmath595 up to conjugacy .
therefore @xmath596 .
the trajectories starting in the upper half of @xmath334 and ending in the upper half of @xmath35 follow the loop @xmath597 up to conjugacy , so @xmath598
. therefore @xmath599 by theorem [ maint ] we get @xmath600 and @xmath601)(1-[a_2a_1])^{-1}\in\widehat{\mathbb{q}h}_{\bar{\xi}}.\ ] ] [ examp ] [ rm54]_notice that in figure [ secon ] the unstable manifolds of @xmath19 intersect @xmath334 in the upper half of @xmath334 while the stable manifolds intersect @xmath35 in the lower half of @xmath35 .
this allows @xmath19 to satisfy @xmath327 with the morse map @xmath141 on @xmath35 .
if we push the unstable manifolds down and the stable manifolds up in figure [ secon ] , we get a different vector field @xmath602 which also satisfies @xmath327 , but with the morse map @xmath603 .
so if we want to calculate @xmath604 we have to interchange the roles of @xmath593 and @xmath594 which gives @xmath605 and @xmath606)^{-1}(1-[a_2a_1])\in\widehat{\mathbb{q}h}_{\bar{\xi}}.\ ] ] we can interpret this the following way : by looking at figure [ secon ] we can expect two closed orbits , one on top of the cobordism , call it @xmath607 , and one on the bottom , call it @xmath608 .
the conjugacy class represented by @xmath607 is the class of @xmath597 while @xmath608 represents the conjugacy class of @xmath595 .
now @xmath609 and @xmath610 for the vector field @xmath19 , but by passing to @xmath602 the unstable and stable manifolds move and the signs change . _
the chain homotopy equivalence given by formula ( [ formula ] ) has its counterpart in the exact case .
the purpose of this appendix is to describe the properties in that case .
an alternative approach can be found in hutchings and lee @xcite , see also schwarz @xcite , but since we need the torsion of the equivalence , we give full proofs .
let @xmath611 be a compact cobordism , @xmath612 $ ] a morse function and @xmath19 an @xmath169-gradient satisfying the transversality condition .
a smooth triangulation @xmath139 of @xmath613 is said to be _ adjusted to _
@xmath19 , if every @xmath25-simplex @xmath26 intersects the unstable manifolds @xmath125 transversely for all critical points @xmath31 of index @xmath140 . in particular ,
if @xmath31 is a critical point of index @xmath25 , a @xmath25-simplex @xmath26 intersects @xmath125 in finitely many points . using the orientations we can assign to every such point a sign .
given a regular covering space @xmath614 we can use the covering transformation group @xmath6 and liftings of critical points and simplices to assign an element @xmath615\in\mathbb{z}g$ ] to the intersection and define a map + _ _ here @xmath618 is the morse - smale complex generated by the critical points of @xmath169 .
for @xmath619 we denote @xmath620 . before we show the existence of adjusted triangulations
let us show that @xmath621 is indeed a chain map .
there exists a filtration @xmath622 of @xmath613 such that @xmath623 is a compact cobordism containing all critical points of index @xmath624 and such that @xmath625 and the boundary homomorphism comes from the long exact sequence of the triple @xmath626 , see milnor @xcite .
also , @xmath627 , where @xmath628 denotes the @xmath25-skeleton of the triangulation .
a simplex @xmath629 is represented by a map @xmath630 .
let @xmath631 be induced by the flow of @xmath201 , where a flowline is supposed to stop once it hits the boundary .
for @xmath632 let @xmath633 . since @xmath139 is adjusted to @xmath19 there is a @xmath634 such that @xmath635 maps @xmath636 to @xmath637 and @xmath638 to @xmath639 .
it follows from intersection theory that @xmath640\in h_k(\tilde{w}_k,\tilde{w}_{k-1}).\ ] ] furthermore this does not depend on @xmath103 as long as @xmath103 is large enough .
a diagram chase gives that @xmath621 is a chain map .
now we want to show the existence of adjusted triangulations .
let @xmath641 be a diffeomorphism homotopic to the identity and @xmath139 a smooth triangulation of @xmath613
. then @xmath142 is the triangulation of @xmath613 where simplices are composed with @xmath141 .
the corresponding chain complexes can be identified by choosing a lifting @xmath642 .
+ so let @xmath139 be any smooth triangulation and @xmath643
. we can adjust @xmath644 near the @xmath5-skeleton so that @xmath5-simplices intersect all unstable manifolds transversely . since the boundary of @xmath613 is transverse to the flow ,
we can leave it invariant . this way we get a diffeomorphism @xmath645 isotopic to the identity .
now assume @xmath646 is isotopic to the identity and every @xmath647-simplex of @xmath648 with @xmath649 intersects the unstable manifolds transversely for critical points with index @xmath650 .
we modify @xmath646 on the @xmath25-skeleton so that @xmath25-simplices intersect @xmath125 transversely for all @xmath31 with index @xmath140 .
notice that for a @xmath25-simplex of @xmath648 this is already true for @xmath646 near the boundary so we can leave the @xmath651-skeleton fixed . this way we obtain @xmath652 isotopic to the identity and we can proceed by induction .
+ then @xmath653 is adjusted to @xmath19 .
furthermore we can find an adjusted triangulation @xmath142 with @xmath141 as close as we like to the identity .
moreover , compactness gives that if @xmath139 is adjusted to @xmath19 , so is @xmath142 for every @xmath141 close enough to the identity . on the other hand ,
given a triangulation @xmath139 and two diffeomorphisms @xmath654 homotopic to the identity such that @xmath655 and @xmath656 are adjusted to @xmath19 , we get two chain maps @xmath657 and @xmath658 which can be different .
let @xmath659 be a homotopy between @xmath657 and @xmath658 . as above we can change @xmath660 to a homotopy @xmath661 between @xmath657 and @xmath658 such that @xmath662 intersects @xmath125 transversely for all critical points @xmath31 with @xmath663 , where @xmath26 is a @xmath25-simplex . describe the morse - smale complex as in the proof of lemma [ alem1 ] .
then we define @xmath664 by @xmath665\in h_{k+1}(\tilde{w}_{k+1},\tilde{w}_k)$ ] . here
@xmath634 is so large that @xmath666 , @xmath667 and @xmath668 is a lifting of @xmath669 .
use @xmath670 and @xmath671 to identify the triangulated chain complexes .
then @xmath672 is the desired chain homotopy .
notice that for a @xmath25-simplex @xmath26 and a disc @xmath125 where @xmath673 @xmath674 is a finite set .
so together with liftings and orientations we can write the chain homotopy as @xmath675\,p\hspace{1cm}\mbox{where } [ \sigma : p]\in \mathbb{z}g,\ ] ] which is independent of the filtration and only involves intersection numbers . [ sheqth ] let @xmath676 $ ] be a morse function , @xmath19 an @xmath169-gradient satisfying the transversality condition , @xmath139 a triangulation adjusted to @xmath19 and @xmath614 a regular covering space
. then @xmath677 given by ( [ chheq ] ) is a simple homotopy equivalence .
let @xmath144 be a subdivision of @xmath139 .
if @xmath145 is adjusted to @xmath19 , so is @xmath142 .
moreover , the diagram @xmath678 \varphi\searrow & & \swarrow\varphi\\[0.2 cm ] \multicolumn{3}{c}{c_\ast^{ms}(\tilde{w},\tilde{m}_0)}\end{array}\ ] ] commutes , where sd is subdivision , a simple homotopy equivalence . by munkres
@xcite it is good enough to show the theorem for a special smooth triangulation .
+ as in the proof of lemma [ alem1 ] we have the filtration @xmath679 where @xmath623 is a compact cobordism containing the critical points of index @xmath624 . choose a triangulation such that each @xmath623 is a subcomplex for all @xmath680 and so that for each critical point @xmath31 of index @xmath132 the disc @xmath681 is a subcomplex .
we set for @xmath682 @xmath683 .
the complex @xmath684 is given by @xmath685 the chain map @xmath621 induces maps @xmath686 and @xmath687 .
since the diagram @xmath688 & & \big\downarrow \varphi^{(k-1 ) } & & \big\downarrow\varphi^{(k ) } & & \big\downarrow\varphi^{(k , k-1)}\\[0.3 cm ] 0&\longrightarrow&d_\ast^{(k-1)}&\longrightarrow & d_\ast^{(k)}&\longrightarrow & d_\ast^{(k)}/ d_\ast^{(k-1)}&\longrightarrow & 0\end{array}\ ] ] commutes , it suffices to show that each @xmath689 is a simple homotopy equivalence to finish the proof .
+ clearly @xmath689 induces an isomorphism in homology , so it remains to show that it is simple .
we set @xmath690 .
then the inclusion @xmath691 is the inclusion of the core of the handles into the handles , hence a simple homotopy equivalence .
now @xmath692 is a simple homotopy equivalence by cohen @xcite , since we can choose the lifts of @xmath454 so that the matrices representing @xmath692 and the boundary operators have only integer values .
therefore @xmath689 is a simple homotopy equivalence .
_ pajitnov ( * ? ? ?
* appendix a ) describes a simple homotopy equivalence + @xmath693 , where the triangulation is given by ( * ? ? ?
* lm.a.8 ) .
the liftings can be chosen so that @xmath694 is the identity on the morse - smale complex , so @xmath621 and @xmath141 are mutually inverse equivalences . _ to calculate the absolute homology @xmath695 for a cobordism with nonempty boundary we need more technicalities .
so if @xmath612 $ ] is a morse function on a compact cobordism and @xmath19 an @xmath169-gradient , pajitnov @xcite defines sets @xmath696 and @xmath697 for @xmath353 and @xmath698 which form filtrations of the cobordism . to simplify the notation
we denote them by @xmath699 and @xmath700 .
pajitnov ( * ? ? ?
* df.4.2 ) now defines a condition @xmath701 on the vector field which also involves morse functions @xmath702 on @xmath342 , @xmath702-gradients @xmath703 for @xmath704 and a @xmath353 .
+ now we define a filtration of @xmath613 by @xmath705 .
it follows from the methods of pajitnov @xcite that @xmath706 gives a free @xmath13 complex calculating @xmath695 with @xmath707 .
we want to explicitely describe a chain homotopy equivalence between @xmath708 and @xmath709 based on [ app1 ] .
+ we need a triangulation @xmath139 of @xmath613 with subtriangulations @xmath710 of @xmath342 for @xmath704 with the properties 1.-4 . described in lemma [ trialem ] below .
notice that for @xmath711 we have + @xmath712 .
+ so if @xmath26 is a @xmath25-simplex of @xmath139 we can flow it along @xmath201 as in the proof of lemma [ alem1 ] into @xmath713 .
this induces a chain map @xmath714 .
also @xmath715 and @xmath716 . in this decomposition
we have @xmath717 , where @xmath718 and @xmath719 are the simple homotopy equivalences of [ app1 ] .
therefore @xmath621 is also a simple homotopy equivalence .
let @xmath611 be a compact cobordism , + @xmath720 $ ] an ordered morse function and @xmath19 an @xmath169-gradient satisfying the transversality condition and condition @xmath701 of pajitnov @xcite .
let @xmath721 , @xmath703 be the @xmath702-gradients and @xmath353 given through condition @xmath701 . 1 .
@xmath139 is adjusted to @xmath19 .
2 . for @xmath704 @xmath710
is adjusted to @xmath703 .
3 . there is an @xmath363 such that if @xmath26 is a @xmath25-simplex in @xmath710 , then @xmath724 .
4 . there is an @xmath363 such that if @xmath725 and the trajectory of @xmath201 starting at @xmath322 ends in @xmath726 , then @xmath727 .
_ notice that all conditions in lemma [ trialem ] are open in the sense that if @xmath139 satisfies 1.-4 .
, so does @xmath142 , provided @xmath141 is close enough to the identity in the smooth topology .
therefore a triangulation as needed in the proof of proposition [ ptor ] also exists . _
for @xmath704 choose triangulations @xmath710 of @xmath342 adjusted to @xmath703 and let + @xmath728 be the flow of @xmath729 .
then there is a @xmath634 such that 3 . is satisfied for @xmath730 .
extend @xmath731 to a triangulation @xmath139 of @xmath613 . choose a diffeomorphism @xmath141
so close to the identity that @xmath142 is adjusted to @xmath19 and 2 . and 3 .
still hold .
modify this triangulation again so that if @xmath26 is a @xmath25-simplex and @xmath732 flows to @xmath726 , then @xmath733 for critical points @xmath734 of @xmath735 with @xmath736 .
notice that @xmath737 is at most @xmath738 dimensional . by making only very small changes
we now have a triangulation satisfying 1.-3 . and by compactness the condition of 4 . for some @xmath739 , but not necessarily for @xmath363
rename this triangulation @xmath139 and the two subcomplexes of the boundary @xmath361 and @xmath362 .
+ notice that condition 4
. already holds for @xmath361 and @xmath362 .
this is trivial for @xmath361 and for @xmath362 it follows from condition @xmath701 . by continuity it also holds in a small collar neighborhood of the boundary .
we can think of this collar as @xmath740)$ ] for some @xmath741 and assume that there are no 0-simplices in @xmath742 .
let @xmath743 $ ] be a smooth function which is 1 outside of the collar and 0 in a smaller collar .
let @xmath744 be the flow of the vector field @xmath745 .
there is a @xmath746 such that if @xmath26 is a @xmath25-simplex in @xmath139 which does not meet @xmath747 , then @xmath748)$ ] .
furthermore @xmath749 satisfies 1.-3 . and the same form of 4 . as @xmath139 does .
+ let @xmath750 for @xmath751 and @xmath752 be diffeomorphic to @xmath753 $ ] with @xmath754 corresponding to @xmath322 and @xmath755 lying on the same trajectory of @xmath19 .
let + @xmath756 , then @xmath757 is a compact cobordism . by changing @xmath169
if necessary we can assume that @xmath758 is a morse function on this cobordism .
+ the morse function @xmath759 is ordered so we can assume that @xmath760)$ ] gives a filtration with @xmath761 for @xmath762 .
let @xmath763 $ ] be a smooth function with @xmath764 and @xmath765 and define @xmath766 and let @xmath767 be the flow of @xmath768 .
+ define @xmath769 as follows .
let @xmath770 . if @xmath771 , let @xmath772 .
if not , the trajectory of @xmath201 reaches @xmath726 .
let @xmath773 .
now @xmath774 flows along @xmath19 all the way back to @xmath775 . for if not , we have @xmath776 .
by @xmath701 we have @xmath777 . but then @xmath778 because @xmath779 and then @xmath774 flows all the way back to @xmath322 .
this and the implicit function theorem give that @xmath780 is smooth on @xmath781 .
similarly , a smooth homotopy @xmath782 between @xmath783 and @xmath784 with @xmath785 } = 1 $ ] and @xmath786}=0 $ ] defines a smooth map @xmath787 such that all @xmath788 and @xmath789 define an isotopy @xmath790 with the following property : if @xmath791 with @xmath792 , then there is a @xmath793 such that @xmath794 and @xmath795 for all @xmath796 . by compactness
there is a @xmath797 such that @xmath798 satisfies the lemma .
99 r. abraham and j. robbin , transversal mappings and flows , w.a .
benjamin , 1967 .
m. cohen , introduction to simple homotopy theory , springer 1973 .
m. farber , morse - novikov critical point theory , cohn localization and dirichlet units , commun .
math 1 ( 1999 ) , 467 - 495 .
m. farber and a. ranicki , the morse - novikov theory of circle - valued functions and noncommutative localization , tr . mat .
inst . steklova 225 ( 1999 ) 381 - 388 .
d. fried , homological identities for closed orbits , inv .
71 ( 1983 ) , 419 - 442 .
r. geoghegan and a. nicas , trace and torsion in the theory of flows , topology 33 ( 1994 ) , 683 - 719 .
m. hutchings , reidemeister torsion in generalized morse theory , harvard university ph.d .
thesis , 1998 .
m. hutchings , reidemeister torsion in generalized morse theory , preprint , math.dg/9907066 .
m. hutchings and y - j .
lee , circle - valued morse theory , reidemeister torsion , and seiberg - witten invariants of three manifolds , topology 38 ( 1999 ) , 861 - 888 .
m. hutchings and y - j .
lee , circle - valued morse theory and reidemeister torsion , geom .
topol . 3 ( 1999 ) , 369 - 396 . k. igusa , what happens to hatcher and wagoner s formula for @xmath799 when the first postnikov invariant is nontrivial ? , algebraic k - theory , number theory , geometry and analysis , lecture notes in math .
vol 1046 , springer 1984 , 104 - 172 . f. latour , existence de 1-formes fermes non singulires dans une classe de cohomologie de de rham , publ .
ihes no.80 ( 1994 ) , 135 - 194 .
j. milnor , lectures on the h - cobordism theorem , princeton university press , 1965 .
j. milnor , whitehead torsion , bull .
72 ( 1966 ) , 358 - 426 .
j. munkres , elementary differential topology , princeton university press , 1963 .
a. pazhitnov , on the novikov complex for rational morse forms , ann .
toulouse 4 ( 1995 ) , 297 - 338 .
a. pajitnov , incidence coefficients in the novikov complex for morse forms : rationality and exponential growth properties , preprint , math.dg/9604004 .
a. pajitnov , simple homotopy type of the novikov complex and lefschetz @xmath800-functions of the gradient flow , russ .
surveys 54 ( 1999 ) , 119 - 169 .
a. pajitnov , @xmath24-generic properties of boundary operators in the novikov complex , pseudoperiodic topology , amer .
ser . 2 , 197 ( 1999 ) ,
a. pajitnov , closed orbits of gradient flows and logarithms of non - abelian witt vectors , preprint , math.dg/9908010 .
a. ranicki , the algebraic construction of the novikov complex of a circle - valued morse function , preprint , math.dg/9903090 m. schwarz , equivalences for morse homology , geometry and topology in dynamics ( winston - salem , nc , 1998/san antonio , tx , 1999 ) , contemp .
( 1999 ) , 197 - 216 . | in @xcite , pajitnov considers the closed orbit structure of generic gradient flows of circle - valued morse functions .
it turns out that the torsion of a chain homotopy equivalence between the novikov complex and the completed simplicial chain complex of the universal cover detects the eta function of the flow .
this eta function counts the closed orbits and reduces to the logarithm of the zeta function after abelianizing .
we extend this result to the case of closed 1-forms which are morse .
to relate the torsion to the eta function we use the dennis trace . |
in the hr diagram , many families of pulsators appear as couple . in the hot part of the main sequence , we encounter the @xmath0 cep stars with low radial order modes ( @xmath3 d@xmath4 ) having a mixed p - g mode nature ( p - mode type in most of the envelope and g - mode type in the region just above the convective core ) .
and just at their cool side the slowly pulsating b ( spb ) stars with high radial order gravity modes ( @xmath5 d@xmath4 , @xmath6 ) .
the same occurs at lower temperature : we find the @xmath1 sct stars with low radial order p - modes and just at their right the @xmath2 dor stars with high radial order g - modes ; and at later evolution stages also we find the couple of sdb and betsy stars .
we concentrate here on the case of @xmath0 cep and spbs .
their modes are driven by a @xmath7-mechansism operating in the iron opacity bump ( iob ) at @xmath8 . at first sight ,
the different pulsation periods and locations in the hr diagram of these two families seem to take their origin in the location of this iob .
as @xmath9 decreases , it goes deeper into the star , entering progressively the domain of g - modes with larger periods .
but recent observations complicate this simple scenario : hybrid stars with both types of oscillations at the same time are observed .
the goal of this paper is to explain their existence and show some aspects of their high potential for asteroseismology .
frequencies of unstable @xmath10 ( left ) and @xmath11 ( right ) modes of @xmath0 cep - type ( blue ) and spb - type ( red ) as a function of @xmath9 , along the evolution sequence of 10 m@xmath12 from zams to tams with op opacities . ] frequencies of unstable @xmath10 ( left ) and @xmath11 ( right ) modes of @xmath0 cep - type ( blue ) and spb - type ( red ) as a function of @xmath9 , along the evolution sequence of 10 m@xmath12 from zams to tams with op opacities . ]
recent observations show the existence of hybrid b stars with at the same time oscillations of @xmath0 cep type ( @xmath13 ) and high order g - modes of spb type ( @xmath14 ) .
the clearest cases are 19 mon @xcite , @xmath15 eri @xcite and 12 lac @xcite .
another very promising star is hd 180642 ( a primary target of corot ) .
several other cases exist such as @xmath2 peg @xcite , but they must be considered with caution because of binarity or because of the low s / n and the risks of aliases , asking for more precise observations . from a theoretical point of view , @xcite , @xcite and @xcite showed that with op opacities ( @xcite , @xcite ) , the instability strips of @xmath0 cep and spbs intersect , so that theoretical models with both unstable p - modes and high order g - modes exist .
an illustration is given in fig .
[ evoll1l2 ] for 10 m@xmath12 models .
work integrals @xmath16 for different @xmath10 modes ( left ) and @xmath11 modes ( right ) of a 10 m@xmath12 model with @xmath17 k , @xmath18 and op opacities . ]
work integrals @xmath16 for different @xmath10 modes ( left ) and @xmath11 modes ( right ) of a 10 m@xmath12 model with @xmath17 k , @xmath18 and op opacities . ]
work integrals @xmath16 for different @xmath10 modes ( left ) and @xmath11 modes ( right ) of a 10 m@xmath12 model with @xmath17 k , @xmath18 and op opacities . ]
work integrals @xmath16 for different @xmath10 modes ( left ) and @xmath11 modes ( right ) of a 10 m@xmath12 model with @xmath17 k , @xmath18 and op opacities . ] for the structure models of this study , we used the stellar evolution code cles @xcite , with by default the following physical prescriptions : the new op opacities @xcite , the opal2001 equation of state @xcite , the mixture of elements by @xcite , with the enhancement of ne proposed by @xcite .
to identify the driving and damping regions , we give in fig .
[ work ] the work integrals for different modes of a well chosen 10 m@xmath12 model .
left panels correspond to @xmath10 modes and right panels to @xmath11 modes .
regions where @xmath19 increases ( resp .
decreases ) outwards have a driving ( resp . damping )
we analize first the @xmath10 modes ( left ) . unstable modes of @xmath0 cep type ( low radial order mixed modes ) are given in blue in the top left panel and unstable modes of spb type ( high radial order g - modes ) are given in red in the bottom left panel .
for these two types of unstable modes , we see clearly the significant driving around the iob at @xmath8 .
but g - modes of intermediate radial order ( green curves ) are stable because of a significant damping in the very deep layers just above the convective core .
this damping is a classical radiative damping mechanism .
the short - wavelength oscillations in the g - mode cavity due to the large values of the brunt - visl frequency lead to very large values of the derivatives of the eigenfunctions ( @xmath20 with @xmath21 ) . in particular , the temperature gradient variations due to the oscillations are very large in this region ; it is easily seen that they lead to significant loss of heat during the hot phase ( see eq .
7 of @xcite ) , which always damps the modes . but
why this significant damping occurs for some g - modes and not for others ?
square of the dimensionless brunt - visl frequency @xmath22 ( red ) , lamb frequency @xmath23 ( black ) and angular pulsation frequencies @xmath24 ( horizontal blue lines ) , for @xmath10 modes ( left ) and @xmath11 modes ( right ) . ]
square of the dimensionless brunt - visl frequency @xmath22 ( red ) , lamb frequency @xmath23 ( black ) and angular pulsation frequencies @xmath24 ( horizontal blue lines ) , for @xmath10 modes ( left ) and @xmath11 modes ( right ) . ] real parts of the transversal displacement @xmath25 ( blue ) and + @xmath26 ( red ) , for the modes @xmath10 , g@xmath27 ( left ) and @xmath10 , g@xmath28 ( right ) . ] real parts of the transversal displacement @xmath25 ( blue ) and + @xmath26 ( red ) , for the modes @xmath10 , g@xmath27 ( left ) and @xmath10 , g@xmath28 ( right ) . ]
changes in the respective weights of the driving and damping regions originate in the shape of the eigenfunctions . to understand this , classical propagation diagrams are very useful , as shown in fig .
[ vaisb ] for the 10 m@xmath12 model considered previously . according to the asymptotic theory , the transversal component of the displacement ( dominating for g - modes ) behaves as : @xmath29 where the function @xmath30 is a cosine of constant amplitude in the propagation regions ( green in fig .
[ vaisb ] ) , and an exponential decreasing outwards in the evanescent regions ( magenta in fig .
[ vaisb ] ) .
[ xih ] ( left panel ) shows that this formula applies well for the mode @xmath10 , g@xmath27 .
the propagation region is very large for this mode going up to the iob , the change of amplitude of @xmath25 is dominated there by the factor @xmath31 which leads to values much smaller near the center compared to the surface .
hence , the deep radiative damping is small compared to the @xmath7-driving in the iob and this mode is unstable .
the right panel of fig .
[ xih ] shows the same eigenfunctions for the mode @xmath10 , g@xmath32 .
now the propagation region is smaller with a larger evanescent region above it ( see fig .
[ vaisb ] ) .
the corresponding exponential decrease leads to amplitudes much smaller around the iob compared to the deep layers .
hence the deep radiative damping overwhelms the @xmath7-driving and this mode is stable .
in the right panels of fig . [ work ] the work integrals for the same model as before are given , but now for @xmath11 modes .
we see in this case that the @xmath0 cep - type mixed mode @xmath11 , g@xmath34 is predicted to be unstable ( blue , top right panel ) , but all @xmath11 high radial order g - modes of spb type are predicted to be stable , contrary to the @xmath10 case .
this can be understood as follows . at high frequencies ,
the eigenfunctions do not depend significantly on @xmath33 .
hence the range of excited modes of @xmath0 cep type ( as a function of the frequency ) remains essentially the same for different @xmath33 ( blue diamonds in fig .
[ evoll1l2 ] and blue curves in the top panels of fig .
[ work ] ) . at lower frequencies
, we have seen that the respective weight of the damping and driving regions is determined by the size of the propagation and evanescent regions .
as @xmath35 , g - modes having the same values of @xmath36 have the same propagation regions .
this corresponds to a lower frequency for @xmath11 compared to @xmath10 modes ( compare the left and right panels of fig .
[ vaisb ] ) .
hence the critical frequency at which the upper boundary of the propagation region reachs the iob is smaller for @xmath11 modes . at this smaller frequency ,
the transition region where the pulsation period is of the same order as the thermal relaxation time is deeper in the star , around the hot side of the iob ; so the damping effect of this region is larger than the driving of the more superficial layers ( red curves in the right bottom panel of fig .
[ work ] ) and all high order g - modes are stable .
@xcite showed that many more hybrid @xmath0 cep - spb are predicted when op opacity tables ( @xcite ) are used compared with opal ( @xcite ) .
this can be seen in fig .
[ evolopopal ] where we compare the range of excited modes obtained in the two cases for 10 m@xmath12 models .
the comparison of the op and opal opacities and their derivative @xmath37 presented in fig .
[ kappa ] shows significant differences around the hot wing of the iob .
[ workopopal ] compares the work integrals obtained with op and opal for a typical @xmath0 cep - type mode ( left ) and spb - type mode ( right ) .
the results are very close for @xmath0 cep - type modes because the opacities are essentially the same with op and opal in the superficial layers ( @xmath38 ) where p - modes have significant amplitudes .
but the results are different for the high order g - modes : with op they are excited while most are stable with opal . to understand this
, i recall that two things are required for an efficient driving of the g - modes .
first , the transition region where the pulsation period is of the same order as the thermal relaxation time must coincides with the iob . and
second , this period must be large enough so that the corresponding evanescent region is negligible .
these two requirements are fulfilled with op ; but with opal the period at which the transition region and the iob coincide is too small , leading to a too large evanescent region .
this explains why the driving of high order g - modes ( probing these deep layers ) is more efficient with op than with opal , and more hybrid models are predicted with op .
opacities ( top ) and @xmath37 ( bottom ) obtained with op ( blue ) and opal ( red ) opacity tables , 10 m@xmath12 models with @xmath17 k and @xmath18 . ] opacities ( top ) and @xmath37 ( bottom ) obtained with op ( blue ) and opal ( red ) opacity tables , 10 m@xmath12 models with @xmath17 k and @xmath18 . ] opacities ( top ) and @xmath37 ( bottom ) obtained with op ( blue ) and opal ( red ) opacity tables , 10 m@xmath12 models with @xmath17 k and @xmath18 . ] work integrals for the mode @xmath10 , g@xmath39 ( left ) and @xmath10 , g@xmath40 ( right ) of the same models as fig .
[ kappa ] , using op ( blue ) and opal ( red ) opacities . ]
work integrals for the mode @xmath10 , g@xmath39 ( left ) and @xmath10 , g@xmath40 ( right ) of the same models as fig .
[ kappa ] , using op ( blue ) and opal ( red ) opacities . ]
frequencies of unstable @xmath41 modes of @xmath2 dor type ( red ) and @xmath1 sct type ( blue ) , for the same model as fig .
[ workdsgd ] . ] frequencies of unstable @xmath41 modes of @xmath2 dor type ( red ) and @xmath1 sct type ( blue ) , for the same model as fig .
[ workdsgd ] . ] frequencies of unstable @xmath41 modes of @xmath2 dor type ( red ) and @xmath1 sct type ( blue ) , for the same model as fig .
[ workdsgd ] . ]
lower along the main sequence , hybrid models with both unstable @xmath1 sct - type mixed modes and @xmath2 dor - type g - modes are also predicted ( @xcite , @xcite ) the observational search of such hybrid models was less conclusive than in b stars but some cases were found : the am star hd 8801 @xcite and the binary hd 209295 @xcite .
the driving of these stars is very different from the one of b stars , being complicated by the time - dependent coherent interaction with convection for cool models @xcite .
however , we point out here that some aspects of the damping mechanisms remain very similar . in fig .
[ workdsgd ] , we give a propagation diagram ( top , dimensionless frequencies ) and the work integrals ( bottom ) for the modes @xmath11 , p@xmath42 , g@xmath43 and g@xmath44 of a young 1.55 m@xmath12 @xmath1 sct-@xmath2 dor hybrid model with @xmath45 k , @xmath46 and @xmath47 . this model is computed with opal opacities and @xmath48 .
we used for the non - adiabatic computation the time - dependent convection ( tdc ) treatment of @xcite .
the mode @xmath11 , p@xmath42 ( blue , @xmath49 c / d ) is a typical unstable mode of @xmath1 sct type ; tdc plays a central role in its driving and damping .
the mode @xmath11 , g@xmath44 ( red , @xmath50 c / d ) is a typical unstable mode of @xmath2 dor type ; the small driving at @xmath51 comes from a @xmath7-mechanism in the iob and the significant driving at @xmath52 is a flux blocking mechanism operating a the base of the convective envelope .
the mode @xmath11 , g@xmath43 ( green , @xmath53 c / d ) is a stable mode which is significantly damped in the deep radiative g - mode cavity
. we note in the top panel of fig .
[ workdsgd ] that the sizes of the propagation and evanescent regions for these modes are similar to the case of b stars considered above .
hence , the explanation of the stable gap between the @xmath1 sct and @xmath2 dor unstable modes is the same as before : these g - modes of intermediate radial order have a large evanescent region ; hence the amplitudes are large in the deep g - mode cavity , and the radiative damping occurring there is much larger than the driving of the superficial layers .
we note that the model considered here is very young , without region of variable molecular weight above the convective core .
hence , no bump of the brunt - visl frequency is present there . in fig .
[ dsgdis ] , all the frequencies of unstable @xmath41 modes are given for the same model as before .
we note that the stable gap between the @xmath1 sct and @xmath2 dor unstable modes decreases as @xmath33 increases .
this is simply due to the fact that @xmath54 , so that the size of the propagation cavity of g - modes increases with @xmath33 for a given frequency .
theoretical models and observations indicate the existence of hybrid models having at the same time unstable pressure and gravity modes . concerning first the driving mechanism of hybrid @xmath0 cep - spb models we have shown , in addition to its already known sensitivity to the iron abundance , that the characteristics of the propagation and evanescent regions also play a major role in this context , determining the extension of the stable gap in the frequency domain between the unstable low order pressure and high order gravity modes .
the same phenomenon occurs for the hybrid @xmath1 sct-@xmath2 dor pulsators .
hence , the comparison between the observed and theoretical ranges of unstable modes allows us to constrain the brunt - visl and lamb frequencies in the very deep interior of these stars .
13 asplund m , grevesse , n and sauval a j 2005 _ aspc _ * 336 * 25 balona l a , james d j , motsoasele p , et al .
2002 _ mnras _ * 333 * 952 chapellier e , le contel d , le contel j m , et al . 2006
_ a&a _ * 448 * 697 dupret m a , grigahcne a , garrido r , et al . 2005 _ a&a _ * 435 * 927 grigahcne a , dupret m a , gabriel m , et al . 2005 _ a&a _ * 434 * 1055 grigahcne a , martin - ruiz s , dupret m a , et al .
2006 _ mem .
* 77 * 1 559 cunha k , hubeny i and lanz t 2006 _ apj _ * 647 * l143 handler g , balona l a , shobbrook r r , et al .
2002 _ mnras _ * 333 * 262 handler g , shobbrook r r , jerzykiewicz m , et al .
2004 _ mnras _ * 347 * 454 handler g , jerzykiewicz m , rodrguez e , et al .
2006 _ mnras _ * 365 * 327 henry g w and fekel f c 2005 _ aj _ * 129 * 2026 iglesias c a and rogers f j 1996 _ apj _ * 464 * 943 miglio a , montalbn j and dupret m - a 2007 _ mnras _ * 375 * l21 miglio a , montalbn j and dupret m - a 2007 _ coast _ * 151 * 48 pamyatnykh a a 1999 _ acta astron . _ * 49 * 119 rogers f j and nayfonov a 2002 _ apj _ * 576 * 1064 scuflaire r , thado s , et al .
2007 _ apss _ in press seaton m j 1996 _ mnras _ * 279 * 5 seaton m j 2005 _ mnras _ * 362 * l1 | we study the energetic aspects of hybrid pressure - gravity modes pulsations .
the case of hybrid @xmath0 cephei - spb pulsators is considered with special attention .
in addition to the already known sensitivity of the driving mechanism to the heavy elements mixture ( mainly the iron abundance ) , we show that the characteristics of the propagation and evanescent regions play also a major role , determining the extension of the stable gap in the frequency domain between the unstable low order pressure and high order gravity modes .
finally , we consider the case of hybrid @xmath1 sct-@xmath2 dor pulsators . |
in hierarchical models of galaxy formation , the origin of galactic angular momentum is ascribed to tidal torques operating early on the material destined to form a galaxy . over the years , starting from the ideas of stromberg ( 1934 ) and hoyle ( 1949 ) , and quantified by the work of peebles ( 1969 ) , doroshkevich ( 1970 ) , and white ( 1984 ) , a number of important properties inherent to the acquisition of angular momentum through tidal torques ( hereafter referred to , for short , as `` tidal torque theory '' , or ttt ) have been identified . the efficiency of the torquing is low , typically endowing galactic material with a rather small amount of net rotation .
this implies that large collapse factors within massive dark matter halos are needed to explain the centrifugally - supported nature of galactic disks ( fall & efstathiou 1980 ) . to leading order
, galaxy spins result from the misalignment between the principal axes of the inertia momentum tensor ( @xmath0 ) of the material being torqued and of the `` shear '' tensor ( @xmath1 ) generated by external material . in the cartesian principal axis frame of the protogalaxy
, the leading term of the torque is given by @xmath2 ( here @xmath3 , @xmath4 , and @xmath5 are cyclic permutations of @xmath6 to @xmath7 , and @xmath8 are the cartesian components of the angular momentum ) .
the inertia term is maximal along the direction that maximizes the difference between @xmath9 and @xmath10 ; i.e. , the _ intermediate _ axis of inertia .
angular momentum growth is typically linear with time at early times ( since @xmath11 then ) and effectively ends at turnaround , when @xmath12 is maximal . in general , then , the direction of the angular momentum will be determined by the _ shape _ of the protogalactic material at turnaround , and is expected to align with the intermediate axis of inertia if @xmath13 and @xmath12 are uncorrelated . many of these expectations have been confirmed by a number of studies of the angular momentum properties of dark matter halos formed in cosmological simulations ( see , e.g. , barnes & efstathiou 1987 , as well as the more recent work of porciani , dekel & hoffman 2002a , b , and references therein ) .
the latter authors point out that @xmath14 and @xmath12 are , in fact , highly correlated , and conclude that , relative to the principal axes of @xmath14 , the most significant trend is for the spin to be perpendicular to the direction of maximum compression ( i.e. , that of the minimum eigenvector of @xmath15 ) .
because @xmath14 and @xmath12 are correlated , the direction of maximum compresion roughly coincides with the direction of the minor axis of inertia at the time of turnaround .
a solid prediction of ttt is , therefore , that galaxy spins should be nearly perpendicular to the minor axis of the collapsing material at late times .
the main shortcoming of this work is that it applies mainly to dark halos , rather than to the luminous ( observable ) component of galaxies .
simulations that include the presence of a dissipative gaseous component have highlighted the possibility that large losses of angular momentum may accompany the collapse of the baryonic component ( navarro & benz 1991 , navarro & white 1994 , navarro & steinmetz 1997 ) .
the rather indirect mapping between the angular momentum of dark halos and that of their baryonic components makes it difficult to assess the success of ttt in accounting for the spin of spiral galaxies .
the discussion of the previous paragraphs , however , suggests that ttt may be tested by looking for residual correlations originating from the coupling between the shear and inertia momentum tensors that dominates the angular momentum growth in protogalaxies .
one example of this is the recent work of lee & pen ( 2000 , 2001 , 2002 , see also lee , pen & seljak 2000 ) , who attempt to reconstruct the shear field from the observed galaxy spin fields and to uncover correlations between the spin field and the large - scale distribution of matter .
we investigate here a related signature expected from ttt ; i.e. , the correlation between the `` shape '' of protogalactic matter distribution at turnaround and the direction of the resulting angular momentum .
as explained above , ttt suggests a trend for spin axes to be perpendicular to the minor axis of the sheet , so that disk galaxies should be highly inclined relative to the plane defined by their surrounding structure . in this _
letter _ we explore whether disk galaxies assembled hierarchically in @xmath16cdm numerical simulations indeed follow these expectations .
in addition , we search for a statistical signature of the predicted trend in catalogs of nearby galaxies .
we have analyzed four n - body / gasdynamical simulations of the formation of disk galaxies in the `` concordance '' @xmath16cdm cosmogony ( @xmath17 , @xmath18 , @xmath19 , @xmath20 , @xmath21 ) .
these simulations are similar to those described in detail by abadi et al ( 2003a , b ) and meza et al ( 2003 ) , where we refer the reader for full details .
the simulations follow self - consistently the evolution of a small region surrounding a target galaxy , excised from a large periodic box and resimulated at higher resolution preserving the tidal fields from the whole box . for the purposes of the discussion here all four simulations give consistent results .
we have chosen to illustrate these using the simulation presented by abadi et al ( 2003a , b ) , but will use the others to verify the general applicability of our results .
figure 1a shows , at late times ( @xmath22 ) , the baryons within @xmath23 mpc of the target galaxy .
the three different projections in the figure have been chosen so as to coincide with the principal axes of the inertia momentum tensor .
this figure shows the high degree of anisotropy that characterizes the non - linear evolution of structures in a @xmath16cdm universe .
what started off as a roughly spherical region @xmath24 comoving mpc across develops into a coherent triaxial structure that surrounds the target galaxy ( located at the center of the panels in figure 1a ) .
the minor axis of the structure , in particular , is very stable across the structure , which may thus be described as a `` sheet '' criss - crossed by filaments
. the ratios of the inertia tensor eigenvectors are given in figure 1a , and are not dissimilar to the mean of all four systems : @xmath25 ( with dispersion @xmath26 ) , @xmath27 ( @xmath28 ) and @xmath29 ( @xmath30 ) .
the arrows in the panels of figure 1a indicate the relative magnitude of the cartesian components of the angular momentum of the sheet , and illustrate the ttt predicted trend ( see 1 ) for spins to be approximately perpendicular to the direction of maximum compression or , roughly , to that of the minor axis ( @xmath31 ) during the non - linear stages of the evolution .
indeed , the whole sheet of material is slowly spinning around the _ intermediate _ principal axis of inertia ( i.e. , @xmath32 , @xmath33 ) .
this is a generic result of our four simulations ; quantitatively , we find that the angular momentum of the high - resolution region surrounding the target galaxies to be oriented such that @xmath34 ( @xmath35 ) , @xmath36 ( @xmath37 ) .
this two - dimensional structure is present since early times and , at @xmath38 , its minor axis is approximately parallel to that of the material that will collapse to form the target galaxy at @xmath39 ( i.e. , the `` protogalaxy '' , see figure 1b , as well as sugerman , summers & kamionkowski 2000 for a similar analysis ) .
the shape of the protogalactic material at turnaround is indicated in the the panels of figure 1b ; for the four simulations we find @xmath40 ( with dispersion @xmath41 ) , @xmath42 ( @xmath43 ) and @xmath44 ( @xmath45 ) . the angular momentum of the protogalaxy at @xmath38 shows similar alignment properties to that of its surrounding material at @xmath22 ;
in particular , the close alignment with the intermediate inertia axis at turnaround is well defined in all four simulations : @xmath46 ( @xmath47 ) , latexmath:[$\langle latexmath:[$\langle the angular momentum direction of the baryons at turnaround is approximately preserved by the disk component at @xmath39 and , as a result , the plane of the disk is highly inclined relative to the plane defined by its surrounding large - scale structure .
this is illustrated in the bottom right panels of figures 1a and 1b , where the disk at @xmath39 is seen approximately face - on in projections that contain either the minor axis of the surrounding structure at @xmath22 or that of the protogalaxy at turnaround .
this indicates that the shape of the protogalactic material during the expansion phase effectively determines the orientation of the spin axis of the resulting galaxy .
since these anisotropies correlate across a wide range of scales in @xmath16cdm , a signature of this process might be preserved as a residual alignment between the disk and its surroundings ; the rotation axis of disk galaxies may be expected to lie _ on _ the plane defined by its surrounding structure .
we explore next whether such alignments are indeed present in samples of disk galaxies in the nearby universe .
as recognized by de vaucouleurs ( 1953 ) in the shapley - ames catalog , galaxies in the vicinity of the milky way are arranged in a two - dimensional slab usually referred to as the `` super galactic plane '' ( sgp ) .
this structure stands out clearly in whole - sky catalogs of nearby galaxies ; extends out to several tens of mpc from the galaxy ; and includes many of the major features of the local large scale distribution of matter , such as the virgo cluster , the perseus - pisces supercluster , and the great attractor ( lahav et al 2000 ) .
the plane of the galaxy is highly inclined relative to the sgp ; indeed , the two planes are approximately perpendicular to each other , as indicated by the low supergalactic latitude of the north galactic pole ( @xmath49 ) .
this situation is similar to that of the simulated disk galaxy in relation to its surrounding structure , as discussed in [ sec : numexp ] .
it is therefore tempting to regard the rather peculiar orientation of the galactic plane relative to the sgp as a result of early torques acting during the protogalactic stage .
if this interpretation is correct , we would expect an excess of nearby galaxies whose rotation axes are approximately perpendicular to the normal to the sgp .
the existence of such alignment has proven controversial ; prior work has led to claims of statistical evidence for such `` anti - alignment '' of galactic planes relative to the sgp ( see , e.g. , flin & godlowski 1989 and references therein ) , but also to suggestions that there is no convincing evidence for deviations from a random distribution of orientations ( see , e.g. , dekel 1985 ) a number of factors complicate this analysis and may explain this disagreement .
for example , the statistics of the position angles of galaxy samples that do not discriminate properly between elliptical and disk galaxies may be inconclusive , as the projected major axis of early type galaxies in contrast to those of disks may bear little relation to the angular momentum of the galaxy .
in addition , determining the direction of the rotation axis of a spiral galaxy requires knowledge not only of the position angle and inclination ( readily available in most catalogs ) , but also of _ how _ the galaxy is inclined on the sky , i.e. , which side of the galaxy is closer to the observer .
this is known for a few well - studied spirals , such as m31 degrees . ] , but it is generally unavailable for most galaxies .
the uncertainty introduced by this ambiguity depends on inclination : it is maximal for galaxies inclined by 45 degrees , but is negligible for galaxies seen either face - on or edge - on . in this _
letter _ we restrict our analysis to edge - on spirals and search for a possible correlation between the orientation of galaxy disks and the sgp .
we explore ways of circumventing the aforementioned ambiguity and extend this analysis to all nearby spirals in a forthcoming paper . figure [ figs : fig2]a shows the distribution of position angles ( in the supergalactic reference frame ) of all edge - on .
] nearby spiral galaxies in the principal galaxy catalog ( pgc ) with recession velocities within various limits .
since we are only interested in the direction of the rotation axis relative to the sgp , we have reduced all position angles to between @xmath50 and @xmath51 . as shown in figure [ figs : fig2]a , there is a clear excess of nearby edge - on galaxies highly inclined relative to the sgp .
the significance of the excess decreases the larger the volume considered around the milky way . within @xmath52 km s@xmath53 there are @xmath54 spirals in the pgc with measured inclinations and position angles . of these @xmath55 are edge - on .
figure [ figs : fig2]a shows that roughly three times as many edge - on galaxies with position angles between @xmath50 and @xmath56 as in the range ( @xmath57 , @xmath51 ) ; a simple ks test shows that this excess is significant to the @xmath58 level .
the magnitude of the excess ( and its significance ) decreases to @xmath59 for spirals with recession velocities under @xmath60 km s@xmath53 , and is essentially negligible when galaxies within @xmath61 km s@xmath53 ( or greater ) are considered .
this result supports the tendency of spirals to be highly inclined relative to the sgp claimed by flin & godlowski ( 1989 ) .
our study provides a compelling physical interpretation for the origin of such alignment and supports the view that the angular momentum of spiral galaxies originates in tidal torques tightly coupled to anisotropies during the expansion , turnaround , and collapse of protogalactic material .
we have used cosmological n - body / gasdynamical simulations to study the orientation of disk galaxies relative to their surrounding structure .
we find that the angular momentum axis of simulated galaxies is perpendicular to the minor axis ( and very well aligned with the _ intermediate _ axis ) of the material destined to form the galaxy at turnaround , which is usually arranged on a two - dimensional slab criss - crossed by filaments .
these sheet - like structures may extend out to several mpc , in a way reminiscent of the `` supergalactic plane '' that surrounds the milky way .
this coherence over a wide range of scales in the anisotropies present during the expansion and turnaround of protogalactic material generally results in galaxy disks highly inclined relative to the large - scale two - dimensional structure where they are embedded .
this provides a natural explanation for the high inclination of the milky way relative to the supergalactic plane , as well as for the excess of nearby edge - on spirals whose rotation axes lie approximately _ on _ the supergalactic plane .
it may also offer a physical interpretation for the tendency of satellite galaxies to align along the minor axis of bright spirals ( the `` holmberg effect '' ) .
our interpretation offers as well a number of `` natural '' predictions that may be used to falsify it .
in particular , we expect an excess of highly inclined spirals in _ any _ two - dimensional large - scale distribution of galaxies , a result that may have already been detected in the perseus supercluster ( flin 1988 ) , and that could be verified in large galaxy surveys such as the 2dfgrs and the sdss .
the detection of these and other non - trivial correlations between the spin and matter fields would serve to establish beyond doubt the validity of tidal torque theory as the origin of the angular momentum of spiral galaxies . | we use numerical simulations to investigate the orientation of the angular momentum axis of disk galaxies relative to their surrounding large scale structure .
we find that this is closely related to the spatial configuration at turnaround of the material destined to form the galaxy , which is often part of a coherent two - dimensional slab criss - crossed by filaments .
the rotation axis is found to align very well with the intermediate principal axis of the inertia momentum tensor at this time .
this orientation is approximately preserved during the ensuing collapse , so that the rotation axis of the resulting disk ends up lying on the plane traced by the protogalactic material at turnaround .
this suggests a tendency for disks to align themselves so that their rotation axis is perpendicular to the minor axis of the structure defined by surrounding matter .
one example of this trend is provided by our own galaxy , where the galactic plane is almost at right angles with the supergalactic plane ( sgp ) drawn by nearby galaxies ; indeed , the sgp latitude of the north galactic pole is just 6 degrees .
we have searched for a similar signature in catalogs of nearby disk galaxies , and find a significant excess of edge - on spirals ( for which the orientation of the disk rotation axis may be determined unambiguously ) highly inclined relative to the sgp .
this result supports the view that disk galaxies acquire their angular momentum as a consequence of early tidal torques acting during the expansion phase of the protogalactic material . |
this is one of my favorite jokes : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ three logicians walk into a bar .
the waitress asks , `` do you all want beer ? ''
+ the first logician answers , `` i do not know . ''
+ the second logician answers , `` i do not know . ''
+ the third logician answers , `` yes . ''
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ this joke reminds me of hat puzzles . in the joke
each logician knows whether or not s / he wants a beer , but does nt know what the others want to drink . in hat puzzles
logicians know the colors of the hats on others heads , but not the color of their own hats . here is a hat puzzle where the logicians provide the same answers as in the beer joke .
three logicians wearing hats walk into a bar . they know that the hats were placed on their heads from the set of hats in figure [ fig : hatset ] .
the total number of available red hats was three , and the total number of available blue hats was two . [ cols="^,^,^,^,^ " , ] [ fig : distinctcolors ] it is natural to try to reuse the solution we know for the previous puzzle .
let us say that the number of hats is @xmath0 and the number of logicians is @xmath1 .
suppose the last person sums the colors modulo @xmath0 and announces his / her own color so that the total sum is zero .
the problem is that s / he might announce the color of one of the logicians in front of him / her .
suppose the last logician says `` red . '' when it s time for the logician who is actually wearing a red hat to speak , s / he will realize that his / her hat is red , and oops , s / he ca nt repeat this color . if s / he says a different color randomly , s / he will throw off the strategy and mislead everyone in front .
is there a way to rescue this strategy so that only a small number of logicians might be mistaken ?
it appears that there is .
suppose the last logician says `` red , '' and the logician in the red hat says `` blue . ''
the blue color was nt announced yet so the logician in the blue hat is most probably way in front of the red - hatted logician .
that means that all the logicians between the red - hatted and the blue - hatted logicians will know that something wrong happened .
they will not be misled .
they will realize that the logician who said `` blue '' said so because s / he were not allowed to say his / her own color . so that color must be red .
that means all the logicians in between can still figure out their colors .
here is the strategy where not more than three people will name a wrong color : the last person announces the color so that his / her color and the sum of the colors s / he sees modulo @xmath0 is zero .
if there is a logician with this hat color , s / he says the hat color of the first person .
this way the three people who might be mistaken are the last one , the logician with the hat color announced by the last person and the first person in line .
everyone else is guaranteed to be correct .
this is quite a good solution , but there is a completely different solution that guarantees that not more than * one * person is mistaken .
i invite you to try and solve it yourself .
if you ca nt do it , you can find the solution in my paper devoted to this puzzle @xcite .
convert the above puzzles into jokes ! | this paper serves as the announcement of my program a joke version of the langlands program . in connection with this program ,
i discuss an old hat puzzle , introduce a new hat puzzle , and offer a puzzle for the reader . |
the start - up of the cern large hadron collider ( lhc ) opens a new discovery era of high energy particle physics . with proton beams colliding at unprecedented energy ,
it provides us with great opportunities to discover the higgs boson and new physics beyond the standard model ( sm ) , with typical signals involving multiple high @xmath0 @xmath1 s , jets , @xmath2 s and @xmath3 s .
reliable searches for these signatures require a good understanding of all sm background processes , which is usually done by simulations performed by monte carlo ( mc ) event generators , such as madgraph @xcite .
however , the complicated event topology expected for some new physics signals makes their background simulation time consuming , and it is important to increase the computational speed for simulations in the lhc data analysis .
in previous studies @xcite , the gpu ( graphics processing unit ) has been used to realize economical and powerful parallel computations of cross sections by introducing a c - language version of the helas @xcite codes , heget ( * h*elas * e*valuation with * g*pu * e*nhanced * t*echnology ) .
heget is based on the software development system cuda @xcite introduced by nvidia @xcite . for pure qed processes , @xmath4 , with @xmath5
to 8 , the calculations ran 40 - 150 times faster on the gpu than on the cpu @xcite .
for pure qcd processes , @xmath6 with @xmath7 up to 4 , @xmath8 and @xmath9 with @xmath7 up to 5 , 60 - 100 times better performance was achieved on the gpu @xcite . in this paper
, we extend these exploratory studies to cover general sm processes , opening the way to perform the complete matrix element computation of madgraph on the gpu .
the complexity of the calculations is increased due to new interaction types and complicated event topologies expected in background simulations for various new physics scenarios .
we introduce additional heget functions to cover all of the sm particles and their interactions , and a phase space parameterization suited for gpu computations . in order to test all of the new functions and the efficiency of the gpu computation in semi - realistic background simulations , we systematically study multi - jet processes associated with the production of sm heavy particles(s ) , followed by its ( their ) decay into final states including light quarks and leptons , including full spin correlations . in particular , we report numerical results on the following processes : [ eq : process ] @xmath10 for the processes ( [ eq : process - v ] ) to ( [ eq : process - httbar ] ) ,
we examine all of the major subprocesses at the lhc , while for the multiple higgs production ( [ eq : process - hwbf ] ) , we study only the weak - boson fusion ( wbf ) subprocesses to test the higgs self interactions . we present numerical results for the cross sections of processes in eqs .
( [ eq : process - v])-([eq : process - hwbf ] ) computed by using the gpu version of the monte carlo integration program , gbases @xcite , with the new heget functions in the amplitude calculations , and compare the results with those obtained running two different programs on the cpu , the fortran version bases @xcite programs with helas subroutines and the latest version of madgraph ( ver.5 ) @xcite .
we also compare the performance of two versions of the bases program , one on the gpu and the other on the cpu .
the paper is structured as follows . in sect .
[ sec : physics_processes ] , we present the cross section formulae for general sm production processes at the lhc , and list all of the subprocesses we study in this report . in sect .
[ sec : algorithm ] , we briefly describe a new phase space parameterization for efficient gpu computation . in sect .
[ sec : heget ] , we introduce new heget functions for all sm particles and their interactions . in sect .
[ sec : mg2cuda ] , we introduce the software used to generate cuda codes with heget functions from fortran amplitude programs with helas subroutines obtained by madgraph . in sect . [
sec : computation ] , we review the computing environment , basic parameters of the gpu and cpu machines used in this analysis .
section [ sec : results ] gives numerical results of computations of cross sections and comparisons of performance of gpu and cpu programs .
section [ sec : summary ] summarizes our findings .
[ sec : ps ] explains in more detail our phase space parameterization introduced in sect .
[ sec : algorithm ] .
[ sec : rng ] explains our method for generating random numbers on gpu .
[ sec : new - heget - codes ] lists all the new heget codes introduced in this paper .
in order to test not only the validity and efficiency of our gpu computation but also its robustness , we examine a series of multi - jet production processes in association with the sm heavy particle(s ) ( @xmath2 , @xmath3 , @xmath11 and @xmath12 ) , followed by their decays into light quarks and leptons , that can be backgrounds for discoveries in many new physics scenarios . in this section , we list all the subprocesses we study in this paper and give the definition of multi - jet cross sections that are calculated both on the cpu and on the gpu at later sections . at the lhc with a collision energy of @xmath13 ,
the cross section for general production processes in the sm can be expressed as @xmath14 where @xmath15 and @xmath16 are the parton distribution functions ( pdf s ) , @xmath17 is the factorization scale , @xmath18 and @xmath19 are the momentum fractions of the partons @xmath20 and @xmath21 , respectively , in the right- and left - moving protons , @xmath22 is the subprocess center of mass energy , and @xmath23 gives the differential cross section for the @xmath24 subprocess @xmath25 the subprocesses cross section can be computed in the leading order as @xmath26 where @xmath27 is the invariant @xmath7-body phase space , @xmath28 are the helicities of the initial and final partons , @xmath29 and @xmath30 are the color degree of freedom of the initial partons , @xmath20 and @xmath21 , respectively .
@xmath31 are the helicity amplitudes for the process ( [ eq : hprocess ] ) , which can be generated automatically by madgraph and expressed as @xmath32 where the summation is over all the feynman diagrams .
the subscripts @xmath28 stand for a given combination of helicities ( @xmath33 for higgs bosons , @xmath34 for quarks and leptons , @xmath35 for photons and gluons ) , and the subscripts @xmath36 correspond to a set of color indices ( none for colorless particles , 1 , 2 , 3 for flowing - in quarks , @xmath37 for flowing - out quarks , and 1 to 8 for gluons ) .
details on color summation in madgraph can be found in ref .
@xcite . in this paper
, we are interested in general sm processes , typically involving the production of heavy resonances with decays , associated with multiple jet production , which often appear as major sm background for new physics signals . the limitation in the number of extra jets , @xmath7 or @xmath38 in eqs .
( [ eq : process - v])-([eq : process - hwbf ] ) , is primarily due to limitations in the amount of memory available to the gpu as reported in previous studies @xcite .
correlated decays of the heavy particles into the following channels are calculated : [ eq : decay ] @xmath39 we do not consider @xmath40 decays in this report for brevity .
the same selection cuts are imposed for @xmath41 , so that the higgs boson production cross sections listed in this report can be used as a starting point for realistic simulations .
a higgs boson mass of 125 and its branching fraction into , 0.0405 , are used throughout this report decay , based on the @xmath40 decay helicity amplitude is available in the framework of madgraph5 @xcite . ] . for definiteness
, we impose the following final state cuts at the parton level . for jets
, we require the same conditions as in refs .
@xcite : @xmath42 [ eq : jet - cuts ] where @xmath43 and @xmath44 are the pseudo - rapidity and the transverse momentum of the @xmath45-th jet , respectively , in the @xmath46 collisions rest frame along the right - moving ( @xmath47 ) proton momentum direction , and @xmath48 is the relative transverse momentum @xcite between the jets @xmath45 and @xmath49 defined by @xmath50 [ eq : isolation - cuts ] here @xmath51 measures the boost - invariant angular separation between the @xmath45 and @xmath49 jets , and @xmath52 and @xmath53 are defined as differences of pseudorapidities and azimuthal angles of the two jets . for @xmath21 jets from @xmath11 decay ,
we require @xmath54 [ eq : bjet - cuts ] note that we do not require them to be isolated from other jets via ( [ eq : jet - cuts - ptjj ] ) . for charged leptons , we require @xmath55 [ eq : lepton - cuts ] and for simplicity we treat @xmath40 the same as @xmath56 and @xmath57 . like @xmath21 jets from @xmath11
decays , we do not impose isolation cuts for leptons , since performing realistic simulations is not the purpose of this study .
we use the set cteq6l1 @xcite parton distribution functions ( pdf ) for all processes .
the following four types of @xmath2 production subprocesses are studied in this paper : @xmath58 [ eq : w ] the subprocess ( [ eq1:w ] ) starts with the leading order @xmath59 , the subprocess ( [ eq2:w ] ) starts with the next order @xmath60 , and those of ( [ eq3:w ] ) and ( [ eq4:w ] ) start with the @xmath61 order . these subprocesses give the dominant contributions in @xmath46 collisions .
the corresponding production cross sections are smaller in @xmath46 collisions since an incoming @xmath62-quark in the subprocesses ( [ eq1:w ] ) to ( [ eq3:w ] ) should be replaced by a @xmath63-quark , whose pdf is significantly softer than the up - quark pdf in the proton .
similarly , the following subprocesses are studied for @xmath3 production : @xmath64 [ eq : z ] as in the case of @xmath65+@xmath7-jets , we examine all of the dominant contributions up to 4-jets .
it should be noted that the down quark contribution to the @xmath3+jets cross section is less suppressed than the @xmath66+jets case , since the @xmath3-boson couples stronger to the down - quarks than the up - quarks ( cf .
@xmath67 ) . for the @xmath2 boson pair production
, we study the following subprocesses : @xmath69 [ eq : ww ] the subprocess ( [ eq1:ww ] ) starts with the leading order , the subprocess ( [ eq2:ww ] ) starts with the next order , and those of ( [ eq3:ww ] ) to ( [ eq5:ww ] ) start with the order . the subprocesses ( [ eq4:ww ] ) are included as the dominant same sign @xmath2-pair production mechanism in @xmath46 collisions .
similarly , the following subprocesses are studied for production : @xmath70 [ eq : wz ] as in the @xmath71+@xmath7-jets case , we consider all of the dominant @xmath72 production subprocesses up to 3 associated jets . note again that the down quark contribution to the @xmath73+jets cross section can be significant because of the large @xmath3 coupling to the @xmath63-quarks .
the following @xmath74 production subprocesses are also studied : @xmath75 [ eq : zz ] all of the dominant @xmath74 production subprocesses up to 3 associated jets are studied .
note , however , that the down - quark contribution to the @xmath76 collision subprocess ( [ eq2:zz ] ) has the coupling enhancement factor of @xmath77 .
although we study only 4 lepton final states , @xmath74+jets processes can be backgrounds for new physics signals with a @xmath3 boson plus jets and large missing . for @xmath79 production ,
we consider the following subprocesses : @xmath80 [ eq : ttbar ] the subprocess ( [ eq1:ttbar ] ) starts with the leading order , the subprocess ( [ eq2:ttbar ] ) starts with the next order , and those of ( [ eq3:ttbar ] ) and ( [ eq4:ttbar ] ) start with the order .
again , only those subprocesses which give dominant contributions in @xmath46 collisions are studied in each order . as for the associate production of the higgs boson and the @xmath2
, we consider the following subprocesses : @xmath81 [ eq : wh ] all of the subprocesses ( [ eq1:wh ] ) to ( [ eq4:wh ] ) are obtained by replacing one gluon in the @xmath82jets subprocesses ( [ eq1:w ] ) to ( [ eq4:w ] ) by @xmath12 , respectively .
the @xmath83+jets subprocesses corresponding to ( [ eq1:wh ] ) to ( [ eq3:wh ] ) are suppressed since an incoming @xmath62-quark would be replaced by a softer @xmath63-quark in the proton .
likewise , the following subprocesses are studied for @xmath84 production : @xmath85 [ eq : zh ] all of the four subprocesses are obtained from the corresponding @xmath3+jets subprocesses ( [ eq1:z ] ) to ( [ eq4:z ] ) , by replacing one final gluon by @xmath12 .
again , the down quark contributions to the subprocesses ( [ eq1:hz ] ) to ( [ eq3:hz ] ) are less suppressed than those to the @xmath86 production processes ( [ eq1:hw ] ) to ( [ eq3:hw ] ) due to the large @xmath3 coupling to the @xmath63-quark . for the @xmath87 production process , the following subprocesses are examined in this paper : @xmath88 [ eq : th ] all of the subprocesses ( [ eq1:th ] ) to ( [ eq4:th ] ) are obtained , respectively , from the @xmath89jets subprocesses ( [ eq1:ttbar ] ) to ( [ eq4:ttbar ] ) by replacing one gluon in the final state by a higgs boson .
in all of the above subprocesses we consider only qcd interactions for production of jets and top quarks , while the weak interactions contribute only in @xmath2 , @xmath3 , @xmath12 productions and in decays . for higgs@xmath90jets processes , we study weak - boson fusion ( wbf ) subprocesses which can be identified at the lhc for various decay modes @xcite .
@xmath68 fusion contributes to the subprocesses ( [ eq1:hw ] ) , @xmath74 fusion contributes to the subprocesses ( [ eq2:hw ] ) , both contribute to the subprocesses ( [ eq3:hw ] ) and ( [ eq4:hw ] ) : @xmath91 [ eq : hw ] the subprocesses ( [ eq1:hw ] ) and ( [ eq2:hw ] ) start with the leading order , those in ( [ eq3:hw ] ) start with the next order , the subprocess ( [ eq4:hw ] ) occurs at the order .
we do not consider @xmath92 fusion process , because its simulation requires non - renormalizable vertices @xcite which are absent in the minimal set of helas codes @xcite .
the following multiple higgs boson production processes are studied in this paper as a test of cubic and quartic vertex functions among scalar and vector bosons : @xmath93 [ eq : hh ] the quartic scalar boson vertex appears only in the subprocesses ( [ eq3:hh ] ) and ( [ eq4:hh ] ) .
in this section , we briefly introduce our phase space parame - trization , designed for efficient gpu computing .
details are given in [ sec : ps ] .
optimizing code to run efficiently on the gpu requires several special considerations .
in addition to the careful use of memory mentioned earlier , one needs to consider that each `` batch '' of calculations processed on the gpu must undergo identical operations .
that has particular consequences when one considers generating momentum in phase space that satisfy the appropriate cuts .
the efficiency of generating momenta that pass cuts is not particularly important on the cpu , since one can simply repeatedly generate momenta until a set is found that pass the cuts .
the structure of the gpu however does not allow such flexibility .
if the generation program is running on multiple processors , each processor has one opportunity to generate a valid phase space point before moving forward to calculate the amplitude .
so if the efficiency for generating a point in phase space that passes cuts is only 10% then you loose a factor of 10 in computing speed . in previous studies @xcite for pure qed and qcd processes at hadron colliders : @xmath94 the following phase space parameterization has been used including the integration over the initial parton momentum fractions ( see [ sec : ps ] for details ) : @xmath95 with this parameterization , when all the final particles are massless partons , we can generate phase space points that satisfy the rapidity constraints , @xmath96 , for all partons ( @xmath97 to @xmath7 ) and the constraints , @xmath98 , for @xmath99 partons ( @xmath97 to @xmath99 ) . only those phase space points which violate the conditions , @xmath100 , @xmath101 or @xmath102 , will be rejected .
studies in refs .
@xcite find that , for example , 78% , 58% , 42% and 29% of generated phase space points satisfy the final state cuts , @xmath103 , for @xmath7-photons or @xmath7-gluons .
it is then only @xmath104 which can go up to @xmath105 .
] , for 2 - 5 photon productions at the lhc , when we choose @xmath106 ( @xmath97 to @xmath7 ) , @xmath107 and @xmath108 ( @xmath97 to @xmath99 ) as integration variables .
we adopt this parameterization of the generalized phase space in this report , which is extended to account for the production and decay of several breit - wigner resonances .
in particular , for process with @xmath7-jet plus @xmath109-resonance production @xmath110 when the resonance @xmath111 decays into a final state @xmath112 , we parameterize the generalized phase space as follows : @xmath113 here @xmath114 is the invariant phase space for the @xmath115 decay when the invariant mass of the final state @xmath112 is @xmath116 , and @xmath117 is the @xmath7-body generalized phase space for @xmath118 massless particles and @xmath109 massive particles of masses @xmath116 ( @xmath119 = 1 to m ) .
integration over the invariant mass squared @xmath120 is made efficient by using the standard breit - wigner formalism @xmath121 and the transverse momenta are generated as @xmath122 finally , we note that s - channel splitting of massless partons can be accommodated by using the same parameterization eq .
( [ ps2-resonance ] ) , where @xmath111 in eq .
( [ ps2-resonance ] ) is a virtual parton of mass @xmath116 and @xmath112 is a set of partons . instead of the breit - wigner parameterization ( [ ps4-resonance ] )
, we simply use @xmath123 as integration variables .
in this section , we explain the new heget functions for computing the helicity amplitudes for arbitrary sm processes . all of the heget functions that appear in this report are listed in [ sec : new - heget - codes ] as list [ list : ixxxxx ] to list [ list : hsssxx ] . in the previous works @xcite , the wave functions for massless particles , which are named ` ixxxx`@xmath119 , ` oxxxx`@xmath119 and ` vxxxx`@xmath119 ( @xmath124 ) , have been introduced for fermions and vector bosons . in this report , we introduce wave functions for massive spin 0 , 1/2 and 1 particles , which can also be used for massless particles by setting the mass parameters to zero .
the naming scheme for heget functions follows that of helas subroutines : the heget ( helas ) function names start with ` i(i ) ` and ` o(o ) ` for flow - in and flow - out fermions , respectively , ` v(v ) ` for vector boson wave functions , and ` s(s ) ` for scalar boson wave functions .
all of the heget functions for external lines are summarized in table [ tab : wavefc ] .
the function ` vvsxxx ` ( list [ list : vvsxxx ] ) computes an amplitude of the ` vvs ` coupling . the arguments of this function as : @xmath126 where the inputs and the outputs are : @xmath127 } & \textrm{first vector wavefunction}\\ \texttt{cmplx v2[6 ] } & \textrm{second vector wavefunction}\\ \texttt{cmplx sc[3 ] } & \textrm{input scalar}\\ \texttt{cmplx gc } & \textrm{coupling constant of \texttt{vvs}}\\ & \\ \textsc{outputs : } & \\
\texttt{cmplx\ & vertex } & \textrm{amplitude}\\ & \gamma(\texttt{v1},\texttt{v2},\texttt{s } ) \label{eq : vvsxxx } \end{array}\ ] ] the function ` jvsxxx ` ( list [ list : jvsxxx ] ) computes an off - shell vector current from the ` vvs ` coupling .
the vector propagator is given in feynman gauge for a massless vector and in unitary gauge for a massive vector .
the arguments of this function as : @xmath128 where the inputs and the outputs are : @xmath129 } & \textrm{input vector wavefunction}\\ \texttt{cmplx sc[3 ] } & \textrm{input scalar wavefunction}\\ \texttt{cmplx gc } & \textrm{coupling constant of \texttt{vvs}}\\ \texttt{double vmass } & \textrm{mass of output vector } \\ \texttt{double vwidth } & \textrm{width of output vector } \\ & \\ \textsc{outputs : } &
\\ \texttt{cmplx jvs[6 ] } & \textrm{vector current}\\ & j^\mu(\texttt{v$^\prime$ } : \texttt{vc } , \texttt{sc } ) \label{eq : jvsxxx } \end{array}\ ] ] the function ` hvvxxx ` ( list [ list : hvvxxx ] ) computes an off - shell scalar current from the ` vvs ` coupling .
the arguments of this function as : @xmath130 where the inputs and the outputs are : @xmath127 } & \textrm{input first vector wavefunction}\\ \texttt{cmplx v2[6 ] } & \textrm{input second vector wavefunction}\\ \texttt{cmplx gc } & \textrm{coupling constant of \texttt{vvs}}\\ \texttt{double smass } & \textrm{mass of output scalar } \\
\texttt{double swidth } & \textrm{width of output scalar } \\ & \\ \textsc{outputs : } &
\\ \texttt{cmplx hvv[3 ] } & \textrm{scalar current}\\ & j(\texttt{s$^\prime$ } : \texttt{v1 } , \texttt{v2 } ) \label{eq : hvvxxx } \end{array}\ ] ] the function ` vvssxx ` ( list [ list : vvssxx ] ) computes an amplitude of the ` vvss ` coupling . the arguments of this function as : @xmath132 where the inputs and the outputs are : @xmath133 } & \textrm{first vector wavefunction } \\ \texttt{cmplx v2[6 ] } & \textrm{second vector wavefunction } \\ \texttt{cmplx s1[3 ] } & \textrm{first scalar wavefunction } \\ \texttt{cmplx s2[3 ] } & \textrm{second scalar wavefunction } \\ \texttt{cmplx gc } & \textrm{coupling constant of \texttt{vvss}}\\ & \\ \textsc{outputs : } & \\
\texttt{cmplx\ & vertex } & \textrm{amplitude of the \texttt{vvss}}\\ & \gamma(\texttt{v1},\texttt{v2},\texttt{s1},\texttt{s2 } ) \label{eq : vvss } \end{array}\ ] ] the function ` jvssxx ` ( list [ list : jvssxx ] ) computes an off - shell massive vector current from the ` vvss ` coupling .
the vector propagator is given in unitary gauge for a massive vector .
the arguments of this functions as : @xmath134 where the inputs and the outputs are : @xmath135 } & \textrm{input vector wavefunction } \\ \texttt{cmplx s1[3 ] } & \textrm{first scalar wavefunction } \\ \texttt{cmplx s2[3 ] } & \textrm{second scalar wavefunction } \\
\texttt{cmplx gc } & \textrm{coupling constants of \texttt{vvss } vertex } \\
\texttt{double vmass } & \textrm{mass of output vector}\\
\texttt{double vwidth } & \textrm{width of output vector}\\\\ \textsc{outputs : } & \\
\texttt{cmplx jvss[6 ] } & \textrm{vector current}\\ & j^\mu ( \texttt{v}^\prime:\texttt{vc},\texttt{s1},\texttt{s2 } ) \\
\label{eq : jvss } \end{array}\ ] ] the function ` hvvsxx ` ( list [ list : hvvsxx ] ) computes an off - shell scalar current of ` vvss ` coupling .
the arguments of this function as : @xmath136 where the inputs and the outputs are : @xmath133 } & \textrm{first vector wavefunction } \\
\texttt{cmplx v2[6 ] } & \textrm{second vector wavefunction } \\
\texttt{cmplx sc[3 ] } & \textrm{input scalar wavefunction } \\
\texttt{cmplx gc } & \textrm{coupling constant of \texttt{vvss } }
\\ \texttt{double smass } & \textrm{mass of output scalar}\\ \texttt{double swidth } & \textrm{mass of output scalar } \\ \\
\textsc{outputs : } & \\ \texttt{cmplx hvvs[3 ] } & \textrm{scalar current}\\ & j(\texttt{s}^\prime:\texttt{v1 } , \texttt{v2},\texttt{sc } ) \\
\label{eq : hvvsxx } \end{array}\ ] ] the function ` sssxxx ` ( list [ list : sssxxx ] ) computes an amplitude of the three - scalar coupling , ` sss ` .
the arguments of this function as : @xmath138 where the inputs and the outputs are : @xmath139 } & \textrm{first scalar wavefunction}\\ \texttt{cmplx s2[3 ] } & \textrm{second scalar wavefunction}\\ \texttt{cmplx s3[3 ] } & \textrm{third scalar wavefunction}\\ \texttt{cmplx gc } & \textrm{coupling constant of \texttt{sss}}\\ & \\ \textsc{outputs : } & \\
\texttt{cmplx\ & vertex } & \textrm{amplitude}\\ & \gamma(\texttt{s1 } , \texttt{s2 } , \texttt{s3 } ) \label{eq : sssxxx } \end{array}\ ] ] the function ` hssxxx ` ( list [ list : hssxxx ] ) computes an off - shell scalar current from the three - scalar coupling , ` sss ` .
the arguments of this function as : @xmath140 where the inputs and the outputs are : @xmath139 } & \textrm{first scalar wavefunction}\\ \texttt{cmplx s2[3 ] } & \textrm{second scalar wavefunction}\\ \texttt{cmplx gc } & \textrm{coupling constant of \texttt{sss}}\\ \texttt{double smass } & \textrm{mass of output scalar}\\
\texttt{double swidth } & \textrm{width of output scalar}\\ & \\
\textsc{outputs : } & \\ \texttt{cmplx hss[3 ] } & \textrm{scalar current}\\ & j(s^\prime:\texttt{s1 } , \texttt{s2 } ) \label{eq : hssxxx } \end{array}\ ] ] the function ` ssssxx ` ( list [ list : ssssxx ] ) computes an amplitude of the four - scalar , ` ssss ` , coupling .
the arguments of ` ssssxx ` as : @xmath142 where the inputs and the outputs are : @xmath143 } & \textrm{first scalar wavefunction } \\ \texttt{cmplx s2[3 ] } & \textrm{second scalar wavefunction } \\ \texttt{cmplx s3[3 ] } & \textrm{third scalar wavefunction } \\ \texttt{cmplx s4[3 ] } & \textrm{fourth scalar wavefunction } \\ \texttt{cmplx gc } & \textrm{coupling constants of \texttt{ssss } vertex } \\ \\ \textsc{outputs : } & \\ \texttt{cmplx\ & vertex } & \textrm{amplitude of four scalar coupling}\\ & \gamma(\texttt{s1},\texttt{s2},\texttt{s3},\texttt{s4 } ) \\
\label{eq : ssssxx } \end{array}\ ] ] the function ` hsssxx ` ( list [ list : hsssxx ] ) computes an off - shell scalar current from ` ssss ` coupling .
this function has the argument as : @xmath144 where the inputs and the outputs are : @xmath143 } & \textrm{first scalar wavefunction } \\
\texttt{cmplx s2[3 ] } & \textrm{second scalar wavefunction } \\ \texttt{cmplx s3[3 ] } & \textrm{third scalar wavefunction } \\ \texttt{cmplx gc } & \textrm{coupling constant \texttt{ssss } } \\
\texttt{double smass } & \textrm{mass of output scalar}\\
\texttt{double swidth } & \textrm{width of output scalar}\\ & \\
\textsc{outputs : } & \\ \texttt{cmplx hsss[3 ] } & \textrm{scalar current}\\ & j(\texttt{s}^\prime:\texttt{s1},\texttt{s2},\texttt{s3 } ) \\
\label{eq : hsssxx } \end{array}\ ] ] j. alwall , m. herquet , f. maltoni , o. mattelaer and t. stelzer , jhep * 1106 * ( 2011 ) 128 [ arxiv:1106.0522 [ hep - ph ] ] .
j. alwall , p. demin , s. de visscher , r. frederix , m. herquet , f. maltoni , t. plehn and d. l. rainwater _
et al . _ ,
jhep * 0709 * ( 2007 ) 028 [ arxiv:0706.2334 [ hep - ph ] ] .
t. stelzer and w. f. long , comput .
commun .
* 81 * ( 1994 ) 357 [ hep - ph/9401258 ] .
f. maltoni and t. stelzer , jhep * 0302 * ( 2003 ) 027 [ hep - ph/0208156 ]
. k. hagiwara , j. kanzaki , n. okamura , d. rainwater and t. stelzer , eur .
j. c * 66 * ( 2010 ) 477 [ arxiv:0908.4403 [ physics.comp-ph ] ] .
k. hagiwara , j. kanzaki , n. okamura , d. rainwater and t. stelzer , eur .
j. c * 70 * ( 2010 ) 513 [ arxiv:0909.5257 [ hep - ph ] ] .
h. murayama , i. watanabe and k. hagiwara , kek-91 - 11 . ` http://www.nvidia.com/content/global/global.php ` ` http://www.nvidia.com/object/cuda\_home\_new.html ` j. kanzaki , eur .
j. c * 71 * ( 2011 ) 1559 [ arxiv:1010.2107 [ physics.comp-ph ] ] .
s. kawabata , comput .
commun . * 41 * ( 1986 ) 127 .
k. hagiwara , t. li , k. mawatari and j. nakamura , eur .
j. c * 73 * ( 2013 ) 2489 [ arxiv:1212.6247 [ hep - ph ] ] .
s. catani , y. l. dokshitzer , m. h. seymour and b. r. webber , nucl .
b * 406 * ( 1993 ) 187 .
j. pumplin , d. r. stump , j. huston , h. l. lai , p. m. nadolsky and w. k. tung , jhep * 0207 * ( 2002 ) 012 [ hep - ph/0201195 ] .
d. l. rainwater and d. zeppenfeld , jhep * 9712 * ( 1997 ) 005 [ hep - ph/9712271 ] .
d. l. rainwater , d. zeppenfeld and k. hagiwara , phys .
d * 59 * ( 1998 ) 014037 [ hep - ph/9808468 ] .
n. kauer , t. plehn , d. l. rainwater and d. zeppenfeld , phys .
b * 503 * ( 2001 ) 113 [ hep - ph/0012351 ] .
w. giele , g. stavenga and j. -c .
winter , eur .
j. c * 71 * ( 2011 ) 1703 [ arxiv:1002.3446 [ hep - ph ] ] .
j. kanzaki , in preparation .
g. marsaglia , journal of statistical software vol.8 , issue 14 ( july 2003 ) 1 .
m. matsumoto and y. kurita , acm trans . on modeling and computer simulation , 2(1992),179194 doi:10.1145/146382.146383 ; m. matsumoto and y. kurita , acm trans .
on modeling and computer simulation , 4(1994),254266 doi:10.1145/189443.189445 ; m. matsumoto and t. nishimura , acm trans . on modeling and computer simulation vol .
1 , january pp.3 - 30 ( 1998 ) doi:10.1145/272991.272995 ; see also ` http://www.math.sci.hiroshima - u.ac.jp/~ m - mat / mt/ ` ` emt.html ` . | continuing our previous studies on qed and qcd processes , we use the graphics processing unit ( gpu ) for fast calculations of helicity amplitudes for general standard model ( sm ) processes .
additional heget codes to handle all sm interactions are introduced , as well as the program mg2cuda that converts arbitrary madgraph generated helas amplitudes ( fortran ) into heget codes in cuda .
we test all the codes by comparing amplitudes and cross sections for multi - jet processes at the lhc associated with production of single and double weak bosons , a top - quark pair , higgs boson plus a weak boson or a top - quark pair , and multiple higgs bosons via weak - boson fusion , where all the heavy particles are allowed to decay into light quarks and leptons with full spin correlations .
all the helicity amplitudes computed by heget are found to agree with those computed by helas within the expected numerical accuracy , and the cross sections obtained by gbases , a gpu version of the monte carlo integration program , agree with those obtained by bases ( fortran ) , as well as those obtained by madgraph .
the performance of gpu was over a factor of 10 faster than cpu for all processes except those with the highest number of jets . |
we assume throughout the paper that @xmath2 is a field of characteristic distinct from @xmath3 .
let @xmath5 be a lie algebra over @xmath2 .
a non - zero element @xmath6 is called _ extremal _ if @xmath7 \in k x$ ] for all @xmath8 . if @xmath9 is extremal , the existence of a linear map @xmath10 such that @xmath7 = f_x(y)x$ ] for all @xmath8 immediately follows from linearity of @xmath11 $ ] . if for some extremal element @xmath6 this linear map @xmath12 is identically @xmath13 , we call @xmath9 a _ sandwich_. extremal elements were originally introduced by chernousov @xcite in his proof of the hasse principle for @xmath14 . zelmanov and kostrikin proved that , for every @xmath15 , the universal lie algebra @xmath16 generated by a finite number of sandwich elements @xmath17 is finite - dimensional @xcite
. cohen , steinbach , ushirobira , and wales generalized this result and proved that a lie algebra generated by a finite number of extremal elements is finite dimensional .
moreover , they give an explicit lower bound on the number of extremal elements required to generate each of the classical lie algebras @xcite .
recently , in t panhuis , postma , and the author gave explicit presentations for lie algebras of type @xmath18 , @xmath19 , @xmath20 , and @xmath21 , by means of minimal sets of extremal generators @xcite . moreover , draisma and in t panhuis considered finite graphs and corresponding algebraic varieties whose points parametrize lie algebras generated by extremal elements .
they proved in particular that if the graph is a simply laced dynkin diagram of affine type , all points in an open dense subset of the affine variety parametrize lie algebras isomorphic to the simple chevalley lie algebra corresponding to the associated dynkin diagram of finite type @xcite .
looking at these lie algebras from a different point of view , cohen , ivanyos , and the author proved that if @xmath5 is a lie algebra over a field @xmath2 ( of characteristic distinct from @xmath3 and @xmath22 ) that has an extremal element that is not a sandwich , then @xmath5 is generated by extremal elements , with one exception in characteristic @xmath0 @xcite .
the strong connection between extremal elements and geometries is further investigated in two papers by cohen and ivanyos @xcite , and in the ph.d .
theses by postma @xcite and in t panhuis @xcite .
we follow the setup of @xcite and @xcite .
assume that @xmath1 is a connected undirected finite graph with @xmath15 vertices , without loops or multiple bonds , and that @xmath2 is a field of characteristic distinct from @xmath3 .
we let @xmath23 be the vertex set of @xmath1 and denote adjacency of two vertices @xmath24 by @xmath25 .
we denote by @xmath26 ( often abbreviated to @xmath27 ) the quotient of the free lie algebra over @xmath2 generated by @xmath23 modulo the relations @xmath28 = 0 \mbox { for all } x , y\in \pi\mbox { with } x \not\sim y.\ ] ] often , we write elements of @xmath27 as linear combinations of _ monomials _ @xmath29]}$ ] , where @xmath30 .
the _ length _ of such a monomial is said to be @xmath31 , and we often abbreviate such a monomial to @xmath32}$ ] .
note that @xmath27 inherits the natural @xmath33-grading from the free lie algebra generated by @xmath23 ; homogeneous elements of @xmath27 are linear combinations of monomials of equal length .
we write @xmath34 for the space of all @xmath2-linear functions @xmath35 . for every @xmath36 ,
also written @xmath37 , we denote by @xmath38 ( often abbreviated to @xmath39 ) the quotient of @xmath26 by the ideal @xmath40 generated by the infinitely many elements @xmath41 - f_x(y)x \mbox { for } x \in \pi \mbox { and } y \in { \mathcal f}.\ ] ] by construction @xmath39 is a lie algebra generated by @xmath42 extremal elements , the extremal generators corresponding to the vertices of @xmath1 and commuting whenever they are not adjacent .
the element @xmath43 is a parameter expressing the extremality of @xmath44 . in the lie algebra @xmath45
the elements of @xmath23 map to sandwich elements . by @xcite
this lie algebra is finite - dimensional ; for general @xmath36 we have @xmath46 by ( * ? ? ?
* lemma 4.3 ) .
it is therefore natural to focus on the lie algebras @xmath39 of maximal possible dimension , i.e. , those of dimension @xmath47 .
we define the set @xmath48 the parameter space for all maximal - dimensional lie algebras of the form @xmath39 .
[ ex_twogen ] consider the case where @xmath1 consists of two vertices @xmath9 , @xmath49 , connected by an edge .
then for every @xmath50 the lie algebra @xmath39 is spanned by @xmath51 \}$ ] since @xmath7 = f_x(y ) x$ ] ( where @xmath52 ) , and similarly @xmath53 = -f_y(x ) y$ ] .
consequently , @xmath22-dimensional lie algebras generated by a distinguished pair of extremal generators are parametrized by the single value @xmath57 . as mentioned before , if @xmath60 then @xmath39 is the heisenberg algebra .
it is straightforward to verify that those lie algebras where @xmath61 are mutually isomorphic and isomorphic to the split simple lie algebra of type @xmath62 . the parameter space @xmath4 defined above , then , is the affine line , and all lie algebras corresponding to the nonzero points on the line are mutually isomorphic .
the following theorem asserts that the two generator case is exemplary : @xmath4 always carries the structure of an affine algebraic variety .
[ thm_kvariety_struct ] the set @xmath4 is naturally the set of @xmath2-rational points of an affine variety of finite type defined over @xmath2 .
this variety can be described as follows .
fix any finite - dimensional subspace @xmath63 of @xmath27 consisting of homogeneous elements such that @xmath64 .
then the restriction map @xmath65 maps @xmath4 injectively onto the set of @xmath2-rational points of a closed subvariety of @xmath66 .
this yields a @xmath2-variety structure of @xmath4 which is independent of the choice of @xmath63 .
we give an example of an embedding of @xmath4 as mentioned in theorem [ thm_kvariety_struct ] .
recall that we showed @xmath4 to be the affine line ; we may take @xmath67\}$ ] . then
, for @xmath68 we have @xmath69 ) ) , ( f_y(x ) , f_y(y ) , f_y([x , y])),\ ] ] and if we again identify elements @xmath68 with @xmath52 we find @xmath70 .
we have now introduced enough background to be able to state our main results .
[ thm_537_xtriv ] let @xmath1 be the complete graph on @xmath0 vertices , @xmath2 a field of characteristic not @xmath3 , and , as above , the set @xmath4 the parameter space for all maximal - dimensional lie algebras of the form @xmath38 .
then @xmath71 .
the following corollary immediately follows from this theorem and ( * ? ? ?
* lemma 4.2 ) ( and section [ sec_comp_results ] with regards to the dimensions mentioned ) . [ cor_537_nilp ]
suppose @xmath5 is a lie algebra generated by @xmath0 extremal elements over the field @xmath2 , where @xmath72 . if @xmath5 is of maximal dimension among such lie algebras ( i.e. , @xmath73 if @xmath74 and @xmath75 if @xmath76 ) then @xmath5 is nilpotent .
looking at results for other cases shows why this result is interesting .
for example , if @xmath1 is the complete graph on on @xmath3 , @xmath22 , or @xmath77 vertices , then @xmath4 is nontrivial .
moreover , for almost all @xmath68 , @xmath39 is a simple lie algebra of type @xmath62 , @xmath78 , and @xmath79 , respectively ( see @xcite and section [ sec_comp_results ] of this paper ) .
moreover , there are infinite families of graphs @xmath1 , for example the affine dynkin diagrams , where @xmath39 is almost always a simple lie algebra of classical type ( see @xcite ) . in short
: this is the first nontrivial case encountered where @xmath4 is a point .
note that the fact that @xmath4 is a point means that the extremality of the generators and the maximal - dimensionality assumption force all generators to be sandwiches , and therefore force @xmath5 to be nilpotent .
we mention one more result .
[ thm_all_x_affine ] let @xmath1 be a connected undirected finite graph with at most @xmath0 vertices , without loops or multiple bonds , @xmath2 a field of characteristic not @xmath3 , and , as above , the set @xmath4 the parameter space for all maximal - dimensional lie algebras of the form @xmath38 .
then @xmath4 is isomorphic to a finite - dimensional affine space over @xmath2 .
the dimension of @xmath4 is given in tables [ tab_comp_results23 ] [ tab_comp_results5 ] .
this theorem provides a partial answer to one of the questions posed in ( * ? ? ?
* section 5.2 ) , namely , `` is @xmath4 always an affine space ? '' .
similarly , the results discussed in section [ sec_comp_results ] suggest that in the cases we considered one of the other questions posed , `` is there always a generic lie algebra ? '' can be answered affirmatively . in section [ sec_extr_props ]
we present a number of known results on these lie algebras and their elements that we need for our proofs and calculations . in section [ sec_fsets ]
we introduce @xmath55-sets , a concept we need in order to be able to calculate with the affine algebraic variety @xmath4 that parametrizes the lie algebras we are interested in . in section [ sec_algs ]
we discuss the algorithms we developed and implemented in magma to obtain the computation results presented in section [ sec_comp_results ] .
finally , in section [ sec_conclusion ] we briefly reflect on our results and possible future research .
there is a number of properties of the functions @xmath12 introduced earlier that we should discuss .
the following result is an important tool in our computations .
[ thm_csuw_bil_f ] suppose that @xmath5 is a lie algebra over @xmath2 generated by extremal elements .
there is a unique bilinear symmetric form @xmath82 such that , for each extremal element @xmath6 , the linear form @xmath12 coincides with @xmath83 .
this form is associative , in the sense that @xmath84 ) = f([x , y],z)$ ] for all @xmath85 .
this theorem implies a number of identities in the values of @xmath55 , as presented in the following lemma .
we use these relations extensively in our computer calculations .
[ lem_ttr_f ] let @xmath86 , let @xmath87 , and let @xmath88 . then the following equalities hold : 1 .
[ itm_lem_ttr_fx_cx][itm_lem_ttr_f_first ] @xmath60 whenever @xmath89 or @xmath90 , 2 .
[ itm_lem_ttr_fx_y ] @xmath91 , 3 .
[ itm_lem_ttr_fx_cxm ] @xmath92 ) = 0 $ ] whenever @xmath89 or @xmath90 , 4 .
[ itm_lem_ttr_fx_ym ] @xmath92 ) = -f_y(x , m)$ ] , 5 .
[ itm_lem_ttr_fx_yzx ] @xmath93 ) = f_x(y)f_x(z)$ ] , 6 .
[ itm_lem_ttr_fx_yxm ] @xmath94 ) = -f_x(y)f_x(m)$ ] , 7 .
[ itm_lem_ttr_fx_yzm ] @xmath95 ) = f_z([y,[x , m ] ] ) - f_z([x,[y , m]])$ ] , 8 .
[ itm_lem_ttr_f_rev][itm_lem_ttr_f_last ] @xmath96 ] } ) = ( -1)^{l-1 } f_{y_l}({[y_{l-1},[y_{l-2},\ldots,[y_2,[y_1,x]]]]})$ ] .
[ itm_lem_ttr_fx_cx ] follows from the observation that @xmath7 = 0 $ ] whenever @xmath89 or @xmath90 .
[ itm_lem_ttr_fx_y ] follows immediately from the symmetry of @xmath55 in theorem [ thm_csuw_bil_f ] : @xmath97 .
[ itm_lem_ttr_fx_cxm ] and [ itm_lem_ttr_fx_ym ] follow similarly from its associativity : @xmath92 ) = f(x , [ y , m ] ) = f([x , y ] , m)$ ] , which is equal to @xmath13 for case [ itm_lem_ttr_fx_cxm ] and reduces to @xmath98 , m ) = -f(y , [ x , m ] ) = -f_y([x , m])$ ] in the case of [ itm_lem_ttr_fx_ym ] . for [ itm_lem_ttr_fx_yzx ] , observe @xmath93 ) = f(x,[y,[z , x ] ] ) = -f(y,[x,[z , x ] ] ) = f(y , f_x(z)x ) = f_x(z)f(y , x ) = f_x(z)f_x(y)$ ] by associativity and bilinearity of @xmath55 ; [ itm_lem_ttr_fx_yxm ] follows similarly .
[ itm_lem_ttr_fx_yzm ] follows from the jacobi identity and associativity : @xmath95 ) = f(x,[y,[z , m ] ] ) = -f([x , y],[m , z ] ) = -f([[x , y ] , m ] , z ) = -f_z([[x , y],m ] ) = -f_z([x,[y , m ] ] ) + f_z([y,[x , m]])$ ] .
finally , to see [ itm_lem_ttr_f_rev ] simply apply associativity of @xmath55 and anti - symmetry of lie algebras @xmath99 times .
our goal in the research described here was to find more information on the structure of the variety @xmath4 describing the parameter space , and of @xmath39 for various @xmath68 . in order to achieve this goal
, we have to compute a multiplication table for @xmath39 .
there are two ways to approach this problem : either pick an @xmath68 in advance and compute a multiplication table for @xmath39 , or compute a general multiplication table , i.e. , one that has entries in the coordinate ring of @xmath4 .
the latter approach , which we have chosen to pursue , has two important advantages : firstly , we only need to compute the multiplication table once , and can afterwards easily instantiate it for particular @xmath68 . secondly , by insisting that @xmath39 be a lie algebra and of maximal dimension we obtain information about the structure of @xmath4 . in this example
we would automatically recover the relation @xmath91 via the jacobi identity on @xmath49 , @xmath9 , and @xmath59 $ ] : @xmath100 ] + [ x,[[x , y],y ] ] + [ [ x , y],[y , x ] ] = f_x(y)[y , x ] + f_y(x)[x , y ] + 0,\ ] ] and @xmath101 ) = 0 $ ] via evaluating @xmath102]$ ] from the inside out : @xmath103)x = [ x,[x,[x , y ] ] ] = [ x , f_x(y ) x ] = f_x(y ) [ x , x ] = 0,\ ] ] so that @xmath101 ) = 0 $ ] since the fact that @xmath39 is of maximal dimension implies that @xmath104 . to aid in the description of our algorithms we introduce the concept of _
@xmath55-sets_. for the remainder of this section fix a graph @xmath1 and a field @xmath2 , and let @xmath105 be a set of monomials of @xmath26 that project to a monomial @xmath2-basis of @xmath45 ( which exists by ( * ? ?
* lemma 3.1 ) ) .
furthermore , let @xmath106 be some subset of @xmath107 , and let @xmath108 be the rank @xmath109 multivariate polynomial ring @xmath110 $ ] with variables @xmath111 .
furthermore , let @xmath112 , @xmath113 be a map from @xmath107 into this polynomial ring .
we call such a pair @xmath114 an
_ @xmath55-set of size @xmath109_. moreover , if @xmath115 is such that for all @xmath68 we have , under the evaluation @xmath116 for @xmath117 , @xmath118 then we call @xmath114 a _ sufficient @xmath55-set_. finally , suppose @xmath114 is an @xmath55-set where we require , to exclude trivialities , that @xmath119 for all @xmath117 .
we construct , for an arbitrary @xmath120 , an element @xmath121 of @xmath122 as follows .
take , for every @xmath123 , the value @xmath124 to be @xmath125 evaluated in @xmath126 , that is , @xmath127 .
this defines a map @xmath128 .
note that @xmath129 is always a lie algebra , but not always one of maximal dimension . if , however , @xmath129 is a lie algebra of dimension @xmath47 for all @xmath120 , then we call @xmath114 a _
free @xmath55-set_. for example , @xmath130 is trivially a sufficient @xmath55-set .
it is , however , in general not a free @xmath55-set . on the other hand
, @xmath131 forms a free @xmath55-set , but in general not a sufficient one .
recall @xmath132 and @xmath51 \}$ ] , so that @xmath133 ) , ( y , x ) , ( y , y ) , ( y,[x , y ] ) \}$ ] . then @xmath134 is a ( trivial ) sufficient @xmath55-set of size @xmath135 . in order for @xmath39 to be a lie algebra of maximal dimension @xmath22 we must have @xmath136 ) = f_y(y ) = f_y([x , y ] ) = 0 $ ] , and we have seen before that @xmath58 must be equal to @xmath57 .
this means that @xmath114 is not a free @xmath55-set , since @xmath137 is not a lie algebra if @xmath138 .
however , if we take @xmath139 , @xmath140 , and @xmath141 ) = r_y(y ) = r_y(x , y ) = 0 $ ] , then @xmath114 is a sufficient @xmath55-set of size @xmath142 , and by checking the jacobi identity for @xmath39 we find that it is a free sufficient @xmath55-set . [ lem_free_fset_implies_affine_space ] if @xmath106 is a free sufficient @xmath55-set then @xmath143
. recall from above the map @xmath144 . by the assumption that @xmath106 is free , we have @xmath145 .
the assumption that @xmath106 is sufficient guarantees that @xmath4 is not a proper subset of @xmath146 , so that @xmath147 is a bijection between @xmath148 and @xmath4 .
we remark that a priori it is possible for @xmath4 to be an affine space while no free sufficient @xmath55-set exists .
however , we have not yet encountered such a case ( see also the discussion in section [ sec_conclusion ] ) .
in this section we describe the algorithms used to obtain the results in section [ sec_comp_results ] . for the computational results in this paper
we have developed an algorithm that , given a graph @xmath1 and a field @xmath2 , computes the variety @xmath4 and the multiplication table of @xmath39 for any @xmath68 .
these algorithms were implemented in the magma computer algebra system @xcite .
the algorithm may be broken up into four distinct parts , of which the third ( multiplicationtable ) and the fourth ( minimizefset ) are the most time - consuming . computebasiscomputing a basiscomp_basis * in : * a field @xmath2 and a graph @xmath1 , with @xmath149 . + * out : * a set of monomial elements of @xmath27 . + * begin * + * let * @xmath150 be the free associative algebra with generators @xmath23 , + * let * @xmath151 be the linear map defined by @xmath152 for @xmath44 and + @xmath153 \mapsto
\mu(b)\mu(c ) - \mu(c)\mu(b)$ ] for @xmath154 , + * let * @xmath155 and @xmath156 , + * let * @xmath157 ) \mid x , y \in \pi , x \not\sim y \rangle$ ] , + * while * @xmath158 * do * + _ //update the ideal representing extremality of elements of @xmath23 _
+ * let * @xmath159 ) \mid x \in \pi , b \in b ' \rangle$ ] , + _ //find
new elements that are not linear combinations of existing basis elements _ + * let * @xmath160 , @xmath161 , @xmath162 , + * for each * @xmath163 * do * + * let * @xmath164)$ ] , + * if * @xmath165 * then * + * let * @xmath166 $ ] , @xmath167 $ ] , + * let * @xmath168 , @xmath169 + * end if * , + * end for * , + * end while * , + * return * @xmath105 . + * end * in this section we describe how we compute , strictly speaking , a set of monomial elements of @xmath26 that projects to a basis of @xmath39 for any @xmath68 . however , because in particular such a basis projects to a basis @xmath45 that consists of monomials , we will often call such a basis `` a monomial basis of @xmath45 '' in the remainder . a sketch of the algorithm computebasis is given as algorithm [ alg_comp_basis ] .
we initialize @xmath105 with the set @xmath23 of generators of @xmath26 , so that @xmath105 contains a basis for all monomials of length @xmath142 .
we then iteratively extend a basis for monomials up to length @xmath31 to a basis for monomials up to length @xmath170 by forming all products @xmath171 $ ] , where @xmath44 and @xmath172 a monomial of length @xmath31 , and testing whether this product can be written in previously found basis elements .
once we have arrived at a length for which no new basis elements are found , we are finished .
note that while we return elements of @xmath27 , all nontrivial computations take place inside the free associative algebra @xmath150 on the sandwich generators of a lie algebra @xmath173 ( which will be identical to @xmath45 ) . in order to connect the two algebras we introduce the map @xmath174 in line @xmath3 , that maps elements of @xmath27 to elements of @xmath150 . in the course of the algorithm , we are also constructing an ideal @xmath175 of @xmath150 such that the adjoint representation @xmath176 factors through @xmath177 . here
@xmath175 describes the fact that the generators are sandwich elements ; @xmath175 should be viewed as the analog in @xmath150 of the ideal @xmath178 of @xmath27 ( cf .
section [ sec_prelim_mainres ] ) .
this construction is what allows us to include @xmath179 in @xmath175 in line @xmath77 of the algorithm , since @xmath180 if @xmath9 is a sandwich .
note finally that in order to be able to compute in @xmath181 ( as required in line @xmath182 of algorithm [ alg_comp_basis ] ) we have to repeatedly compute a grbner basis of @xmath175 but , fortunately , since the elements of @xmath175 are all homogeneous a truncated grbner basis is sufficient ( and sufficiently efficient in the cases we are interested in ) .
[ lem_alg_compute_basis ] the algorithm computebasis returns a set @xmath105 of monomials in @xmath27 , and @xmath105 projects to a basis of @xmath39 for any @xmath68 .
moreover , all elements of @xmath105 are either of the form @xmath9 ( @xmath44 ) or of the form @xmath183 $ ] ( @xmath44 , @xmath123 ) .
the fact that @xmath105 consists of monomials , and that these monomials are of the stated form , is immediate from the algorithm ( note the construction of new elements of @xmath105 in line @xmath184 ) . moreover , because our @xmath68 are such that @xmath185 , a homogeneous basis for @xmath45 is a homogeneous basis for @xmath39 ( cf .
* lemma 3.1 ) . ) . initialfsetfinding an initial @xmath55-setinit_fset * in : * a field @xmath2 and a graph @xmath1 , with @xmath149 , and a monomial basis @xmath105 of @xmath45 .
+ * out : * a sufficient @xmath55-set @xmath114 . +
* begin * + * let * @xmath186 , + * for * @xmath187 , @xmath188 * do * + * if * @xmath189 can be expressed in @xmath190 using lemma [ lem_ttr_f][itm_lem_ttr_f_first ] [ itm_lem_ttr_f_last ] * then * + * set * @xmath191 to be the corresponding expression , where @xmath192 is replaced by @xmath193 + * else * + * set * @xmath194 , + * set * @xmath119 + * end if * + * end for * + * return * @xmath114 . +
* end * the computation of the multiplication table , described in section [ sec_alg_comp_mt ] , will take place over the ring @xmath108 for some sufficient @xmath55-set @xmath114 . in this section
we describe initialfset , the procedure for initialisation of @xmath108 , as presented in algorithm [ alg_init_fset ] .
note that the @xmath55-set returned by this algorithm is a sufficient @xmath55-set , but in general not a free @xmath55-set .
we fix in advance an ( arbitrary ) total ordering on the elements of @xmath23 , and a total ordering on @xmath105 respecting the natural order by monomial length , i.e. , @xmath195 whenever @xmath196 .
these extend to a total ordering of @xmath107 by @xmath197 whenever @xmath195 , or @xmath198 and @xmath199 .
the main for loop in algorithm [ alg_init_fset ] traverses @xmath107 in ascending order ; this is the ordering meant by `` @xmath200 '' in line @xmath22 as well .
initialfset uses the relations in lemma [ lem_ttr_f ] as reduction steps : @xmath201 is interpreted as `` @xmath202 may be reduced to @xmath105 '' .
[ lem_alg_initial_fset ] the algorithm initialfset returns a sufficient @xmath55-set . observe that the relations in lemma [ lem_ttr_f ] all must hold if @xmath39 is to be a lie algebra of dimension @xmath45 ; this immediately implies sufficiency . finally , observe that all reductions thus obtained are polynomial , so that indeed @xmath203 defined in line @xmath77 of algorithm [ alg_init_fset ] is an element of @xmath108 .
monomialtobasisexpressing a monomial as a linear combination of basis elementsmonomial_to_basis * in : * a field @xmath2 and a graph @xmath1 , with @xmath149 , a monomial basis @xmath105 of @xmath45 , + a sufficient @xmath55-set @xmath114 , a partially defined multiplication table @xmath204 , + and a monomial @xmath205}$ ] . + * out : * @xmath206 such that @xmath207 , or * fail*. + * begin * + * if * @xmath208 can be written as a linear combination @xmath209 of basis elements , + by iteratively computing @xmath210}$ ] for @xmath211 using @xmath204 * then * + * return * @xmath126 .
+ * else if * @xmath212 in @xmath1 * then * + * return * @xmath213 .
+ * else if * @xmath214 for some @xmath215 * then * + _ / * use @xmath7 = f_x(y ) x$ ] * / _ + * let * @xmath216 , @xmath217}$ ] , + * let * @xmath218 , + * let * @xmath126 such that @xmath219}$ ] ( using @xmath204 ) , + * return * @xmath220 .
+ * else if * @xmath221 for some @xmath222 * then * + _ / * use @xmath223 ] = f_x([y , z ] ) x - f_x(z)[x , y ] - f_x(y)[x , z]$ ] * / _ + * let * @xmath216 , @xmath224 , @xmath225}$ ] , + * let * @xmath226 ) \in r_{\mathbf{f}}$ ] , + * let * @xmath227 , + * let * @xmath228 , + * let * @xmath229 such that @xmath230}$ ] ( using @xmath204 ) , + * let * @xmath231 such that @xmath232]]}$ ] ( using @xmath204 ) , + * let * @xmath233 such that @xmath234]]}$ ] ( using @xmath204 ) , + * return * @xmath235 .
+ * else * + * return fail * + * end if*. + * end * monomialrelationsfinding monomial relationsmonomial_relations * in : * a field @xmath2 and a graph @xmath1 , with @xmath149 , a monomial basis @xmath105 of @xmath45 , + a sufficient @xmath55-set @xmath114 , a partially defined multiplication table @xmath204 , + a set @xmath236 of monomials of length @xmath31 currently under consideration , + an integer @xmath237 , and a monomial @xmath205 } \in { \mathcal m}$ ] . + * out : * a set @xmath238 of elements @xmath239 such that @xmath240 . + * begin * + * let * @xmath241 , + _ / * by the jacobi identity * / _ + * write * @xmath242}$ ] as a sum @xmath243 of @xmath244 monomials using equation ( [ eqn_jac_from_start ] ) , + * find * @xmath239 such that @xmath245 ] } = \sum_{b \in b } a_b b + \sum_{u \in
{ \mathcal m } } a_u u$ ] , + using monomialtobasis to obtain the entries indexed by @xmath105 , + * let * @xmath246 , + _ / * using lemma [ lem_eqn_sxnxm ] * / _ + * for * @xmath247 such that @xmath248 , @xmath249 , and @xmath250 * do * + * write * @xmath208 as a sum @xmath243 of @xmath251 monomials using lemma [ lem_eqn_sxnxm ] , + * find * @xmath239 such that @xmath252 , using + for the entries indexed by @xmath105 , + * let * @xmath246 + * end for * , + * return * @xmath238 .
+ * end * computemultiplicationtablecomputing a multiplication tablecompute_mt * in : * a field @xmath2 and a graph @xmath1 , with @xmath149 , a monomial basis @xmath105 of @xmath45 , + and a sufficient @xmath55-set @xmath114 . + * out : * a partial multiplication table @xmath204 , where @xmath253 . + * begin * + * let * @xmath254 , + * for * @xmath255 * do * + _ / * consider monomials of length @xmath31 * / _ + * let * @xmath256 \mid x \in \pi , b \in b \mbox { such that } |b|=l\}$ ] , + * let * @xmath257 be the empty @xmath258 matrix , + * let * @xmath259 be the empty @xmath260 matrix , + * let * @xmath261 , + * while * @xmath262 * do * + * for * @xmath263 * do * + _ / * find relations for @xmath208 * / _ + * if * @xmath264 monomialtobasis(@xmath208 ) @xmath265 * fail * * then * + * append * @xmath266 to @xmath257 , * append * @xmath126 to @xmath259 + * else * + * let * @xmath267 monomialrelations@xmath268 , + * for each * @xmath269 * append * @xmath270 to @xmath257 , * append * @xmath271 to @xmath259 + * end if * , + * end for * , + * let * @xmath272 , + * end while * , + _
/ * read the result * / _ + * for * @xmath273 \in { \mathcal m}$ ] * do * + * let * @xmath274 , where @xmath275 is such that @xmath276 , + * let * @xmath277 * for * @xmath187 + * end for * + * end for * , + * return * @xmath204 . + * end * the third part of the algorithm is multiplicationtable , where part of a multiplication table for @xmath39 is computed .
let @xmath114 be a sufficient @xmath55-set ( such as the one produced by initialfset ) , and let @xmath23 be the set of generating extremal elements and @xmath105 be a monomial basis of @xmath45 as before .
we seek to compute @xmath278 such that , for @xmath279 , for specialisation to any @xmath68 , we have @xmath280 = \sum_{c \in b } { a^{{\mathbf{f}}}_{xbc}}c.\ ] ] we will call such a set @xmath204 a _ multiplication table of @xmath39_. note that this is a multiplication table over @xmath108 ; for @xmath68 it specializes to a multiplication table for the lie algebra @xmath39 over @xmath2 upon evaluation @xmath116 for @xmath117 . first , observe that once we have computed @xmath281 for all @xmath44 and all @xmath123 every monomial @xmath282 $ ] can easily be written as a linear combination of basis elements by using @xmath204 to multiply `` from the inside out '' . secondly , observe that then similarly for any @xmath283 we can use the jacobi identity to write @xmath153 $ ] as a linear combination of monomials , and then apply the same trick ( it would of course in general yield an exponential number of terms ) .
therefore we restrict to computing the @xmath204 for @xmath44 , @xmath123 .
we present sketches of the algorithms used as algorithms [ alg_monomial_to_basis ] , [ alg_monomial_relations ] , and [ alg_compute_mt ] , where [ alg_compute_mt ] is the main function .
we clarify some of the notation : firstly , we consider elements of @xmath284 as vectors indexed by elements of @xmath105 , as for example in lines @xmath285 , @xmath286@xmath287 of algorithm [ alg_monomial_to_basis ] .
these vectors may of course be multiplied by elements of @xmath108 , as in lines @xmath182 ,
@xmath288@xmath289 of that algorithm .
the same notation is used for elements of @xmath290 and @xmath291 .
secondly , we write @xmath266 for the vector with @xmath13 everywhere except for the position indexed by @xmath208 ( cf .
line @xmath292 of algorithm [ alg_compute_mt ] ) , and @xmath293 for the restriction of the vector @xmath294 to a subset @xmath295 of its index set . to clarify the monomial rewriting in line @xmath135 of algorithm [ alg_monomial_relations ] , we state the following lemma .
[ lem_eqn_sxnxm ] let @xmath205}$ ] be a monomial of length @xmath31 , such that @xmath248 for some @xmath296 such that @xmath297 .
then @xmath208 may be rewritten as a sum of @xmath22 monomials of length less than @xmath31 and @xmath298 monomials of length equal to @xmath31 . for brevity and ease of reading
we will prove the following easier lemma , of which the above is a straightforward generalisation .
let @xmath205}$ ] be a monomial of length @xmath31 , such that @xmath299 for some @xmath300 such that @xmath301 .
then @xmath208 may be rewritten as a sum of @xmath22 monomials of length less than @xmath31 and @xmath302 monomials of length equal to @xmath31 .
moreover , it follows immediately from the jacobi identity that @xmath305 ] = - [ [ x_2 , x_1 ] , [ x_3 , \ldots , x_l ] ] + [ x_2 , [ x_1 , [ x_3 , \ldots , x_l]]]\end{aligned}\ ] ]
[ lem_alg_compute_mt ] if multiplicationtable returns a multiplication table @xmath204 then , for any @xmath68 , these @xmath204 specialize to a multiplication table for the lie algebra @xmath39 upon evaluation @xmath116 for @xmath117 . recall that , as described at the start of section [ sec_alg_comp_mt ] , the structure constants @xmath204 for @xmath315 follow immediately from those where @xmath44 and @xmath316 .
we will consequently assume that @xmath204 is known for @xmath315 .
observe that in algorithm [ alg_compute_mt ] the products are computed by ascending length , and for each length @xmath31 two matrices @xmath257 and @xmath259 are computed . throughout the algorithm the rows of @xmath257 and @xmath259 contain linear relations between elements of @xmath236 and elements of @xmath105 , i.e. , the @xmath317-th row of @xmath257 must be equal to the @xmath317-th row of @xmath259 .
to see that these relations are true , observe that those gained from monomialtobasis ( algorithm [ alg_monomial_to_basis ] ) are true by induction on the length of @xmath208 ( if returned in line @xmath3 ) , by definition of extremality ( if returned in line @xmath182 ) , or by the first premet identity equation [ eqn_premet1 ] ( if returned in line @xmath318 ) . similarly , those gained from monomialrelations ( algorithm [ alg_monomial_relations ] ) are true by the jacobi identity ( those produced in line @xmath22 ) or by lemma [ lem_eqn_sxnxm ] ( those produced in line @xmath285 ) . also , because we traverse the products of basis elements by ascending length ,
the elements of @xmath284 required in lines @xmath285 and @xmath286@xmath287 of algorithm [ alg_monomial_to_basis ] are guaranteed to be easily deduced from the entries of @xmath204 computed earlier .
finally , correctness of the @xmath281 computed by solving the system of linear equations represented by @xmath257 and @xmath259 in line @xmath318@xmath319 of algorithm [ alg_compute_mt ] follows from bilinearity of the lie algebra multiplication .
we remark that we can not give an estimate of the number of relations we have to produce ( i.e. , the number of times we call monomialrelations and the value of @xmath320 ) before @xmath257 is of full rank . in practice , however , we find that @xmath320 does generally not exceed @xmath3 .
in particular , this implies that termination of computemultiplicationtable is not guaranteed but by lemma [ lem_alg_compute_mt ] correctness upon termination is . minimizefsetfinding a minimal sufficient @xmath55-setminimize_fset * in : * a field @xmath2 and a graph @xmath1 , with @xmath149 , a monomial basis @xmath105 of @xmath45 , + a sufficient @xmath55-set @xmath114 , and a multiplication table @xmath204 . + * out : * an updated @xmath55-set @xmath114 of smaller size , and updated multiplication table . + * begin * + _ / * collect relations between elements of @xmath106 * / _ + * let * @xmath321 - f_y(c)y \mid ( y , c ) \in { \mathbf{f}}\}$ ] , + * let * @xmath322 + [ b,[c , a ] ] + [ a,[b , c ] ] \mid a , b , c \in b \}$ ] , + _ / * reduce size of @xmath106 using @xmath323 * / _ + * do * + * for each * non - zero coefficient @xmath324 of each element @xmath325 * do * + * find * @xmath326 , and @xmath327 such that @xmath111 occurs only in the linear term @xmath328 of @xmath329 + * or continue * if no such element exists , + * set * @xmath330 , + * replace * @xmath111 by @xmath203 in @xmath323 , + * replace * @xmath111 by @xmath203 in @xmath204 , + * set * @xmath331 + * end for * + * until * @xmath106 is unchanged , + * return * @xmath332 . + * end * after the multiplication table has been computed , we try to minimize the size of the @xmath55-set .
this procedure is called minimizefset and presented here as algorithm [ alg_minimize_fset ] .
the procedure is straightforward : we find a set of relations @xmath323 between the @xmath111 by requiring that the jacobi identity holds and that for all @xmath188 , @xmath187 we have @xmath333 = f_y(c ) y$ ] , and we subsequently use those relations to find nontrivial @xmath203 and remove @xmath327 from @xmath106 .
once the algorithm finishes , we apparently could not find any new linear relations .
this does not necessarily mean that the @xmath55-set @xmath114 returned is free .
however , if it happens to be the case that @xmath334 upon exiting , then @xmath114 is indeed a free sufficient @xmath55-set since then the jacobi identity apparently holds .
for all the cases we have tested , the @xmath55-set returned was in fact free .
observe that in this algorithm only division by elements of @xmath335 is taking place ( cf .
lines @xmath0@xmath135 of algorithm [ alg_minimize_fset ] ) and not by other elements of @xmath108 , so that all these calculations take place in multivariate polynomial rings @xmath108 over @xmath2 .
while it is a priori conceivable that division by elements of @xmath336 would be necessary to reduce the size of the @xmath55-set ( and thus that our algorithm would return a non - free @xmath55-set ) , we have not observed such a case in practice .
we remark that a rather more sophisticated alternative to algorithm [ alg_minimize_fset ] has been implemented .
for example , we execute a similar algorithm in the course of the computation of the multiplication table ( rather than only after the fact ) , we alternate between collecting a relatively small number relations and reducing the size of @xmath106 ( rather than do them in succession ) , and in line @xmath0 of algorithm [ alg_minimize_fset ] we give preference to @xmath337 such that the pair @xmath327 is large in the ordering on @xmath107 introduced in section [ sec_alg_init_fset ] .
[ lem_alg_minimize_fset ] if @xmath334 upon exiting of the algorithm minimizefset , then @xmath4 is isomorphic to the affine space @xmath338 .
this follows immediately from lemma [ lem_free_fset_implies_affine_space ] .
cclccl \(a ) + & .computational results ( 2 or 3 generators ) [ cols="^ " , ] & ll @xmath4 : & @xmath339 + @xmath340 : & @xmath341 + @xmath342 : & + runtime : & 2s + we present the results of applying the algorithms described in section [ sec_algs ] in tables [ tab_comp_results23 ] [ tab_comp_results5 ] . for each connected
undirected finite graph @xmath1 with @xmath15 vertices , @xmath343 , without loops or multiple bonds , we state the dimension of the variety @xmath4 , the dimension of @xmath45 , the generic dimension and isomorphism type of @xmath342 , and the cpu time taken ( on a @xmath3 ghz amd opteron ) to execute the steps described in section [ sec_algs ] .
recall that we claim the results presented in tables [ tab_comp_results23 ] [ tab_comp_results5 ] to be valid for any field @xmath2 of characteristic distinct from @xmath3 . to establish this ,
we have performed the following calculations for every graph @xmath1 occurring in the results .
first we calculate a basis of the equivalent lie ring over @xmath344 using a procedure similar to the one described in ( * ? ? ?
* section 3 ) .
this gives us a set @xmath345 of primes for which the monomial basis computed in algorithm [ alg_comp_basis ] would differ from the characteristic @xmath13 case .
then we execute the algorithms described in section [ sec_algs ] for @xmath346 , and in algorithm [ alg_minimize_fset ] we store in the set @xmath347 the prime components of the denominator of the @xmath348 we divide by .
subsequently , we perform the computation with @xmath349 for @xmath350 ; fortunately this last task is easily done in parallel . only in one case
did we find a different result : for @xmath351 the case where @xmath352 is of dimension @xmath353 ( rather than @xmath354 ) .
the result that @xmath71 remains valid , however .
we emphasize that in each of the cases considered we found a free @xmath55-set , and consequently prove that @xmath4 is an affine space ( cf .
lemma [ lem_free_fset_implies_affine_space ] ) .
moreover , we keep track of the reductions and relations used to turn the trivial sufficient @xmath55-set into a free sufficient @xmath55-set , so that we obtain a certificate for the correctness of these results . unfortunately , however , such certificates are far too long to reproduce here . before turning to some of the results ,
let us elaborate on a particular piece of data , namely the isomorphism type of @xmath342 .
for a number of special @xmath1 ( namely the dynkin diagrams of affine type ) it has been shown that the lie algebra @xmath39 is isomorphic to a fixed lie algebra @xmath5 for @xmath55 in an open dense subset of @xmath4 ( * ? ? ?
* theorem 22 ) .
a similar result is proved in ( * ? ? ?
* section 7 ) for the four infinite families there considered .
we investigate this property in our cases as follows .
firstly , we carry out all the computations described in the previous section over @xmath355 . recall that in all cases we found a free @xmath55-set @xmath114 ; we let @xmath356 , so that @xmath357 . now take some finite field @xmath358 , whose characteristic @xmath359 is distinct from @xmath3 and such that no multiples of @xmath359 occur in denominators in the multiplication table @xmath204 ( in our examples the multiplication table always turns out to be integral , so any finite field of characteristic distinct from @xmath3 works )
. then take @xmath126 uniformly random in @xmath360 and let @xmath361 , where we interpret the multiplication table @xmath204 over @xmath362 and @xmath147 is evaluation as described in section [ sec_fsets ] .
this gives us a lie algebra @xmath5 defined over @xmath358 , for which it is straightforward for magma to compute the isomorphism type of @xmath342 .
we repeated this process a number of times over various finite fields ( e.g. , @xmath363 , @xmath364 ) and stored the results in each case finding the same dimension and isomorphism type of @xmath342 in the vast majority of cases .
we remark that we have not distinguished between split and twisted forms of lie algebras of the same cartan type .
the following results are particularly worth pointing out ( we refer to the case presented in table i(x ) simply as case [ ix ] ) . 1
. [ 1c ] , [ 2c ] , [ 3a ] , and [ 3h ] agree with cases proved in @xcite ; 2 .
[ 2b ] , [ 2d ] , [ 3f ] , and [ 3j ] agree with cases proved in @xcite ; 3 .
[ 3n][3q ] are the four cases with @xmath0 vertices and @xmath22 commuting edges .
where one might expect a structural similarity , [ 3n ] is the odd one out both in terms of dimension ( @xmath365 vs @xmath366 ) and type ( a large radical vs simple of type @xmath367 ) ; 4 .
similarly , [ 3r ] and [ 3s ] are the two cases with @xmath0 vertices and @xmath3 commuting edges . again , despite the difference in @xmath47 being only @xmath142 , their isomorphism types are very different ; 5 . in our computations
, we have not encountered a lie algebra of type @xmath14 . even though case
[ 3 t ] is of dimension @xmath368 , the biggest simple component we found there was of type @xmath367 .
it has been proved that @xmath14 can be generated by @xmath0 extremal elements ( * ? ? ?
* theorem 8.2 ) , but apparently not in the manner considered in this paper : @xmath14 is not a subquotient of any maximal - dimensional lie algebra generated by @xmath0 extremal elements plus commutation relations , for then it would have to occur either in [ 3 t ] , where it does not , or in [ 3u ] , where it can not for such a lie algebra is always nilpotent . on the other hand ,
@xmath14 can be generated in this manner by @xmath182 extremal elements ( * ? ? ?
* theorem 22 )
[ 3u ] is the case of largest dimension among the @xmath0 generators ; it is also the only case where @xmath4 is a point and consequently only one lie algebra occurs , and that it is a nilpotent lie algebra ( cf .
* lemma 4.2 ) ) .
these theorems follow immediately from the correctness of our algorithm , as proved in lemmas [ lem_alg_compute_basis ] , [ lem_alg_initial_fset ] , [ lem_alg_compute_mt ] , and [ lem_alg_minimize_fset ] , and the computational results presented in section [ sec_comp_results ] . in theorem [ thm_537_xtriv ] and corollary [ cor_537_nilp ] we prove that if @xmath1 is the complete graph on @xmath0 vertices , then @xmath71 and the only lie algebra that occurs is nilpotent .
this of course leads to the question whether this pattern continues : `` for @xmath1 the complete graph on @xmath15 vertices ( @xmath369 ) , is @xmath4 a point ? '' .
unfortunately , this question is not feasible using our software ( experiments show that for @xmath370 already @xmath371 ) .
another approach would be to consider relations between the variety for @xmath1 and the varieties for subgraphs @xmath372 .
this problem is the subject of ongoing research .
secondly , as mentioned earlier , two questions were posed in ( * ? ? ?
* section 5.2 ) : `` is @xmath4 always an affine space ? '' , and `` is there always a generic lie algebra ? '' .
although the authors of that paper expect the answers to both questions to be negative , we have found no counterexamples . moreover , since these questions can be answered affirmatively for both a number of infinite series of sparse graphs ( as studied in @xcite ) and a number of dense graphs ( as studied here ) , we would not dispose of these conjectures too easily .
the author would like to thank arjeh cohen and jan draisma for numerous fruitful discussions on this topic , willem de graaf for his help in computing several lie rings , and the anonymous reviewers for their very thorough evaluation and their invaluable comments .
zelmanov and a.i .
kostrikin . a theorem on sandwich algebras . , 183:106111 , 225 , 1990 .
translated in proc .
steklov inst .
1991 * , no . 4 , 121126 , galois theory , rings , algebraic groups and their applications ( russian ) | we give an overview of some properties of lie algebras generated by at most @xmath0 extremal elements .
in particular , for any finite graph @xmath1 and any field @xmath2 of characteristic not @xmath3 , we consider an algebraic variety @xmath4 over @xmath2 whose @xmath2-points parametrize lie algebras generated by extremal elements . here
the generators correspond to the vertices of the graph , and we prescribe commutation relations corresponding to the nonedges of @xmath1 . we show that , for all connected undirected finite graphs on at most @xmath0 vertices , @xmath4 is a finite - dimensional affine space .
furthermore , we show that for maximal - dimensional lie algebras generated by @xmath0 extremal elements , @xmath4 is a single point .
the latter result implies that the bilinear map describing extremality must be identically zero , so that all extremal elements are sandwich elements and the only lie algebra of this dimension that occurs is nilpotent .
these results were obtained by extensive computations with the magma computational algebra system .
the algorithms developed can be applied to arbitrary @xmath1 ( i.e. , without restriction on the number of vertices ) , and may be of independent interest . |
approximately 50 planetary nebulae ( pns ) are presently known to have small scale " heterogeneities located inside or outside the main ionized nebulae @xcite .
cometary knots are a subcategory of small scale structures found commonly in nearby , evolved pnspns @xcite . because of its closest proximity ( 213 parsecs @xcite ) , the helix nebula ( ngc 7293 ) is the best case to study the structure and excitation conditions of cometary knots .
the nature of the cometary knots in the helix was first established by @xcite .
the detailed structure of the cometary knots has been resolved in ionized gas lines in the optical by @xcite with further detailed analysis by @xcite , and @xcite .
the emerging optical picture of the cometary knots reveals that they are neutral gas condensations that appear as comet like structures with rims bright in h@xmath10 and tails that appear as shadows in [ oiii ] and that point away from the central star .
the rim of low - excitation ionize gas has a steep temperature gradient indicating that the knots are photo - evaporating and that ionization fronts are advancing into the knots @xcite .
a recent anlaysis of knots over the whole helix nebula by @xcite revealed a new 3-d picture for the main ring of the helix : it is composed of a disk structure and an outer ring tilted almost perpendicularly with respect to the disk . within each of these components
, they observed a similar , progressive evolution in the structure of the knots .
the knots closest to the central star and clearly inside of the ionization front were elegantly carved with the brightest rims .
the knots furthest from the central star appeared slightly more amorphous in their structure with less well defined rims . the culmination of these optical observations appear to support the theory that these knots were initially formed earlier by instabilities at the ionization front or perhaps by the interaction of the fast stellar wind and then have been sculpted by interaction with the harsh radiation field of the central star @xcite .
in contrast to the high angular resolution ( @xmath30.01 ) optical studies of the ionized gas lines in the cometary knots , the molecular gas observations have had lower angular resolution ( 441 ) and sensitivity making it difficult to determine the detailed structure and excitation of the main gas component of the cometary knots .
these low resolution studies have revealed that the helix has retained a significant amount of molecular gas @xcite and that the molecular gas appears to be very clumpy and is probably confined to cometary knot structures @xcite .
the only detailed study of an isolated cometary knot , which is close to the central star , shows no evidence for large velocities in the molecular gas , ruling out a stellar wind shaping the knot , and reveals a stratified structure for the ionized and molecular gas emissions that is expected in a photodissociation region ( pdr ) @xcite .
however , since recent optical studies show an evolution of the knot structure with radial distance from the central star @xcite , it is not clear that this single knot study is representative of all the knots in the nebula . in order to determine the structure and excitation of the h@xmath0 emission in the cometary knots at comparable resolution to optical images across the helix
, we pursued high angular resolution ( @xmath30.2 ) nicmos / nic3 f212n h@xmath0 images at several locations in the nebula , in parallel with the hst / acs program recently published by @xcite .
the remainder of this paper is organized as follows . in section 2 ,
we report the observation and data processing procedures . in section 3
, we discuss the major observational results and how these relate to the optical ionized gas line emissions imaged by @xcite . in section 4 ,
we interpret the observations in the context of current understanding of the helix s 3-d structure and discuss the number density , mass , evolution and excitation of the knots as revealed by our h@xmath0 images .
we summarize our conclusions in section 5 .
the hubble helix project ( go program 9700 ; pi : m. meixner ) imaged the helix nebula during the 2002 leonids meteor shower that presented a risk to the hst .
the imaging involved a 9-panel mosaic of the helix using the acs wfc instrument in the f658n filter ( transmitting equally well both the h@xmath10 6563 and [ n ii ] 6584 lines ) and the f502n filter ( dominated by the [ o iii ] 5007 line ) . in parallel with the acs imaging , we used nicmos @xcite to image 7 of the possible 9 field positions , 5 of which landed on the nebula ( positions 1 , 2 , 3 , 4 and 5 ) and 2 of which were off the nebula ( positions 7 and 9 ) and used for background measurements for the 5 fields on the nebula .
figure [ helixmap ] shows the location of these fields on the helix and the ra and dec of the field centers for field positions 1 , 2 , 3 , 4 , and 5 are listed in table [ loctab ] .
these parallel nicmos field positions had insignificant overlap with the the acs images . because we wanted maximum field of view and our target was a diffuse nebula , we used the nic3 camera , @xmath11 pixel@xmath7 , with the f212n filter to image the h@xmath0 2.12 @xmath1 m line emission in the nebula . for field positions 1 and 2 ,
half the time was spent in the pa@xmath10 filter f187n that is sufficiently low signal - to - noise as to be useless and is not discussed further . for each field position
, the two dither positions for acs resulted in two slightly overlapping nicmos / nic3 images .
the nic3 multiaccum , fast readout mode was used . the hubble helix project and its results @xcite immediately went into the public domain .
the acs images were analyzed in combination with ground based ctio images in similar filters and have been published by @xcite . in this work
we analyze and discuss the nicmos h@xmath0 2.12 @xmath1 m emission and its relation to the ionized gas at high spatial resolution . the nicmos / nic3 images were reduced and calibrated using the standard set of nicmos calibration programs provided in the latest version ( version 3.1 ) of iraf / stsdas .
the calnica calibration routines in stsdas perform zero - read signal correction , bias subtraction , dark subtraction , detector non - linearity correction , flat - field correction , and flux calibration .
the pedestal effect was removed by first manually inserting the stsdas task _ biaseq _ in the middle of the calnica processes ( before flat - fielding ) and then employing the stadas task _ pedsub _ after the calnica processes .
cosmic rays were identified and replaced by the median filtered pixel value .
the four dither positions for field positions 7 and 9 were combined to make a sky " image that is completely attributed to the telescope emission .
this sky " was subtracted from each of the dither frames for field positions 15 resulting in a h@xmath0 dominated emission frame .
the continuum emission from the helix nebula in the f212n filter is negligible in the sky - subtracted images as demonstrated by @xcite .
the two dither positions for each field position were combined using drizzle which magnifies the images by a factor of 2 .
the final drizzled images have a plate scale of @xmath12 pixel@xmath7 and have been rotated so that north is up and east is to the left .
total integration times for the final , drizzled f212n images ranged from 768 seconds for field positions 1 and 2 to 1792 seconds for field position 3 .
figures [ nicpos1 ] , [ nicpos2 ] , [ nicpos3 ] , [ nicpos4 ] and [ nicpos5 ] show the final nicmos images for field positions 1 , 2 , 3 , 4 and 5 , respectively , in comparison with the optical emission line images from @xcite .
these final nicmos images , in units of count rate per pixel , are available at the mast web site under the hubble helix project . in order to discern the relative distribution of the h@xmath0 line emission with the ionized gas line tracers , we compare our results with comparable resolution optical images .
the overlap between the nicmos fields and the acs fields is insignificant .
fortunately , @xcite presented ground based ctio images of h@xmath10/[nii ] , [ oiii ] , h@xmath13 and [ sii ] of the entire helix nebula at comparable resolution to our nicmos images .
we registered these ctio images to the wcs of the nicmos images .
the initial comparison , using just the absolute coordinates of the nicmos and ctio images , permitted a close enough alignment to identify at least 1 star in common between the nicmos and ctio images that was used for translational alignment .
the four ctio images had 67 stars in common with each other and our detailed comparison revealed small rotational errors between them up to 0.08 relative angular rotations . using the star in common with the nicmos image as the origin ,
" we improved the relative rotational alignment of the ctio images to better than 0.005 .
the ctio images were then translated to the nicmos wcs position , by aligning the star in common using the task `` register '' in iraf . only the part of the ctio images that overlap with the nicmos field positions are shown .
for each field position , we selected a prominent h@xmath0 knot , labeled with a cross in the figures , for surface brightness measurements of the h@xmath0 emission and made a cross - cut along the radial direction from the central star in order to quantify the relation between h@xmath0 emission and distance from the central star . these positions and cross - cuts are labeled on figs .
[ nicpos1][nicpos5 ] and the locations of the h@xmath0 knots with respect to the field centers are listed in table [ loctab ] .
table [ surbrtab ] lists the surface brightness measurements for h@xmath0 and the optical line tracers at the knot position .
the average surface brightness of the h@xmath0 knot emission ( dn / pixel ) was determined for a circular aperture enclosing the brightest part of the h@xmath0 emission ( @xmath33 pixel radius centered on the + ) .
we then converted this average into physical units by multiplying by @xmath14 erg s@xmath7 @xmath8 @xmath7 dn@xmath7 ( dn = counts per second ) , which is the photometry conversion keyword derived for the nicmos , nic3 camera , filter f212n for the 77.1 k detector , appropriate for observations taken after january 2002 . finally , to arrive at the units in table [ surbrtab ]
, we mulitplied by 212.1 which is the filter bandwidth of the f212n filter , and divided by @xmath15 sr which is the solid angle of the @xmath12 pseudo - pixel . for each of the measured h@xmath0 knots in the nicmos fields , we measured the surface brightness of the ionized gas emission lines , h@xmath10[nii ] , [ oiii ] , h@xmath13 and [ sii ] , in the same way and list the results in table [ surbrtab ] .
a conversion factor of @xmath16 was used to convert adus / pixel to photons s@xmath7 @xmath8 sr@xmath7 for all the ctio images and then each image was multiplied by its photon energy , @xmath17 to determine the surface brightness in units of erg s@xmath7 @xmath8 sr@xmath7 , the same as the nicmos h@xmath0 line measurements . for the cross - cuts shown in figure [ nicprof ] , we applied the same conversion factors .
figures [ nicpos1 ] , [ nicpos2 ] , [ nicpos3 ] , [ nicpos4 ] and [ nicpos5 ] compare the nicmos h@xmath0 2.12 @xmath1 m line emission images to the ionized gas tracers of [ oiii ] , h@xmath13 , h@xmath10/[nii ] and [ sii ] from @xcite for the respective field positions 1 , 2 , 3 , 4 , and 5 .
our nic3 images have better sensitivity and angular resolution than the image in the @xcite and we observe h@xmath0 line emission at much larger distances than seen in their large scale , ground based mosaic .
the nicmos field positions 1 and 2 lie closer to the central star at approximately the same projected distance .
the field position 3 follows next in radial distance with positions 5 and 4 overlapping at the farthest radial distances ( fig .
[ nicprof ] ) . in all of the nicmos field positions the h@xmath0 emission
is highly structured revealing arcs and pillars of emission that point towards the central star ( figures [ nicpos1][nicpos5 ] ) .
this highly structured appearance contrasts with the more smooth , and less structured appearance of the ionized lines .
this difference indicates that the h@xmath0 line emission is confined to the high density neutral gas of the cometary knots @xcite . on the other hand ,
the ionized gas emission arises from both the more diffuse nebula ( 50 @xmath18 ) and the cometary knots .
closer inspection of all 5 positions reveals that the h@xmath13 emission structures correlates very well with the structures observed in h@xmath0 .
the [ sii ] emission appears to correlate with the h@xmath0 emission in positions 1 and 2 , but does not show the structure as well as the h@xmath13 .
this suggests that the [ sii ] emission is more extended , diffuse or less defined by the cometary knots .
the h@xmath10[nii ] , which appears like a combination of the h@xmath13 and [ sii ] emission , follows the h@xmath0 emission but is much more diffuse in appearance than the h@xmath0 or h@xmath13 . in positions 1 and 2 ,
the uniformity of the [ oiii ] emission is punctuated by dark shadows of cometary knots that appear clearly in h@xmath0 emission .
the h@xmath0 knots must lie in front of some highly ionized gas emitting in order to cause the extinction of the [ oiii ] emission .
the fact that most of the h@xmath0 knots do not appear as [ oiii ] shadows indicates that most of the h@xmath0 knots are confined within the neutral disk and below the highly ionized , diffuse gas apparent in [ oiii ] emission .
the [ oiii ] emission trails off in intensity in positions 3 and 4 having a diffuse and streaked appearance with no cometary knot emission .
the [ oiii ] emission at these more distant positions probably arises from above the disk .
the bright streams of emission in the [ oiii ] line may be examples of the crepuscular rays found on the large scale [ oiii ] emission by @xcite . the combined cross - cuts ( fig . [ nicprof ] ) dramatically show the difference between the h@xmath0 emission and optical line ionized gas tracers .
the ionized gas tracers appear to smoothly drop with increasing radius .
differences in the intensity vs. radial distance for these ionized gas tracers are due to the photo - ionization structure of the helix @xcite . of all the ionized gas tracers shown , the h@xmath13 profiles reveal some clumped structure at the 510% level on its basically smooth profile . in contrast
, the h@xmath0 emission appears to almost randomly fluctuate with radial distance because of the highly clumped , knot structure of the h@xmath0 emission .
the h@xmath0 intensity does not appear to decrease significantly with radial distance because the individual knots have a small range of intensities .
positions that overlap in radial distance , 1 & 2 and 4 & 5 , are offset in their optical line emissions because they are located at very different azimuthal positions and the nebula has a distinct variation with azimuthal angle .
interestingly , the h@xmath0 emission does not appear to have offsets in emission in the overlap regions further supporting that the h@xmath0 emission appears to have an almost constant level with radial distance .
detailed comparison of line intensities of the brighter knots in each field position also support this contrast between optical ionize gas emission and h@xmath0 emission .
table [ surbrtab ] lists the brightnesses of a small region in each field position shown as a white cross in the figures [ nicpos1][nicpos5 ] .
these regions were selected as bright rims of the h@xmath0 knots and hence measure the brightest h@xmath0 emission in the fields .
our nicmos images were taken farther out in the helix nebula into regions that were below the detection limit of @xcite , @xmath19 erg s@xmath7 @xmath8 sr@xmath7 , and hence complement their picture .
the h@xmath0 emission from the bright h@xmath0 clumps decreases by a factor of two between the nearest ( pos1 ) and most distant ( pos5 ) clumps . in comparison ,
the ionized line emission drops more steeply with the h@xmath13 emission dropping by a factor of 4 and the [ oiii ] line dropping by a factor of 5 .
@xcite imaged the entire helix nebula in h@xmath0 showing that it followed the rest of the gas mass tracers and interpreted the h@xmath0 as arising in knots in the main disk of the nebula .
however , in their recent analysis of the acs and ctio data in combination with velocity data from the literature , @xcite revealed an entirely new 3-d structure of the helix nebula in which the main ring is broken into an inner disk and an outer ring that are almost perpendicular with respect to each other ( @xmath378 ) .
it is the superposition of these two rings that gives the helix its helical appearance .
similar multiple axis structures have been observed easily in edge - on poly - polar planetary nebulae such as ngc 2440 @xcite ; however , the almost pole - on view of the helix has made it more difficult to define its geometry .
how does the full - nebula h@xmath0 image of @xcite fit into this new paradigm for the helix ?
figure [ fullnebh2 ] shows the inner disk and outer ring structures discussed by @xcite superposed on the near - ir h@xmath0 image of @xcite .
the h@xmath0 near - ir emission of the inner disk is more cleanly separated from the outer ring structure than is found in the optical line tracers and supports the conclusion that the inner disk is a separate structure than the outer ring @xcite .
this cleaner separation occurs because the h@xmath0 emission arises only from the knots and thus there is better contrast than for the optical line tracers that are diluted by the diffuse emission .
the h@xmath0 emission arises at the outer edge of the inner disk that is filled with higher excitation ionized gas , as traced by [ oiii ] , in the center @xcite .
the h@xmath0 emission arising in the inner disk is also much fainter than the h@xmath0 emission in the outer ring .
the opposite is true for the ionized gas tracers : they are much brighter in the inner disk than the outer ring .
this reversal suggests that the molecular gas in the inner disk has been largely photo - dissociated in comparison to the outer ring and , hence , the cometary knots are more evolved in the inner disk .
the outer ring does not appear very ring - like in the h@xmath0 emission , but more like a ring with southern and northern arc - extensions at the eastern and western edges that surround the northwest and southeast plumes " mentioned by @xcite . in fact , the h@xmath0 emission appears brightest in these plumes .
the co kinematics suggest that these plumes are coherent structures @xcite .
one possible explanation is that the outer ring defined by @xcite is really part of an outer bipolar structure and the southern and northern arcs are the limb - brightened edges of the bicones in the outer structure .
such an interpretation is in line with the helix nebula being a poly - polar planetary nebula being viewed pole - on .
the edge - on poly - polar pn , ngc 2440 , has an inner torus of h@xmath0 emission and an outer bipolar nebula of h@xmath0 emission with axes of symmetry that are tilted with respect to one another @xcite .
our nicmos field positions appear to lie in some of the fainter or non - existent regions of h@xmath0 emission as observed by @xcite which had less sensitivity than our study .
positions 1 and 2 are located outside the southern edge of the inner disk in what appears to be a gap between the inner disk and outer structures .
position 3 lies to the southwest of the outer structure where no apparent h@xmath0 emission appears in the h@xmath0 image by @xcite . positions 4 and 5 are located even further away from previously detectable h@xmath0 emission south of the nebula .
positions 3 , 4 and 5 also lie in regions where no hi 21 cm line emission was detected by @xcite nor co emission detected by @xcite . the number density of h@xmath0 knots and area filling factor of h@xmath0 emission decreases with increasing radius and is the lowest in positions 3 , 4 and 5 ( figs . [ nicpos3][nicpos5 ] ) . in the large beams of these radio line observations ( @xmath331 fwhm for co and @xmath342 fwhm for hi 21 cm ) ,
the intensity of the neutral gas emission from these knots is beam diluted and falls below the detection limit of the radio observations .
our observations clearly show that the morphology of the molecular hydrogen emission is highly clumped in comparison to the ionized gas tracers .
figure [ nicpos1_color ] shows a multi - color image comparing the [ oiii ] , the h@xmath10/[nii ] and the h@xmath0 emission for position 1 .
the structure observed in this image is primarily due to the h@xmath0 emission clumps .
in fact , the h@xmath0 emission images are striking by their lack of diffuse h@xmath0 emission .
close inspection of the more intense regions shows they are composed of overlapping knots of h@xmath0 emission .
hence , we confirm the conclusion of @xcite that the molecular hydrogen is confined to the high density knots such as seen in the optical by @xcite .
a similar conclusion was reached by @xcite for the ring nebula based on comparison of high resolution ground - based h@xmath0 emission images to the optical hst images .
thus in two evolved planetary nebulae , the helix and the ring nebulae , the h@xmath0 line emission is highly structured and confined to knots .
the near - ir h@xmath0 emission provides us with the highest angular resolution map of the neutral gas knots in the helix nebula .
previous work has suggested that the knots contain all the neutral gas detected at substantially lower angular resolution in the co emission ( @xmath331 fwhm ) by @xcite , in the ci emission ( @xmath315 fwhm ) by @xcite , the hi emission ( @xmath342 fwhm ) by @xcite and the h@xmath0 line emission ( @xmath34 fwhm ) by @xcite .
however , all previous neutral gas studies have had insufficient resolution and sensitivity to separate and determine the structure and number density of these neutral gas knots .
the optical study of the knots by @xcite provided an initial , lower limit for the total number of cometary knots to be 3500 in the entire nebula .
they base this estimate by extrapolating the number density of knots they can identify in their wfpc2 optical images vs. radius to the entire nebula .
our h@xmath0 images reveal that many more molecular knots exist as defined by the h@xmath0 emission arcs than can be identified in the optical images ( figures [ nicpos1][nicpos5 ] ) . for example , in the nicmos field position 1 the number of knots that appear as [ oiii ] shadows are less than 10 ; however , the number of arc - shaped h@xmath0 emission structures is @xmath3150 . in table
[ numdentab ] , we list the number of knots , which we identify by arcs of h@xmath0 emission , and the fov of the image .
the number density of knots is simply the total divided by the fov .
the area filling factor of h@xmath0 emission is the percentage of the fov that contains h@xmath0 emission structures above a 1@xmath20 threshold intensity ( @xmath31@xmath21 erg s@xmath7 @xmath8 sr@xmath7 ) .
the total number of knots , the number density of knots and the area filling factors are the highest for positions 1 and 2 , and decrease with larger radial distance from the star as seen in positions 4 and 5 .
interestingly , the peak surface brightness of the knots does not decrease significantly with radial distance as we see in figure [ nicprof ] and table [ surbrtab ] .
we estimate the total number of molecular knots in the helix by scaling the number of knots we observe in our nicmos images to the total angular size of the helix .
if we look at the h@xmath0 image of @xcite we find that our nicmos field positions 1 and 2 land in a region of average or slightly below average h@xmath0 intensity .
so , we base a conservative estimate of the total number of knots by using the average knot number density of positions 1 and 2 , 0.041 knots / arcsec@xmath2 .
the h@xmath0 emission region is an annulus with an inner radius of @xmath3170 and an outer radius of 450 covering a total angular area of @xmath22 arcsec@xmath2 . multiplying the average knot number density by the total angular area equals @xmath323,000 molecular hydrogen knots in the helix nebula or a factor of 6.5 larger than previous estimates based on optical images @xcite .
the estimated mass of a single helix knot is @xmath3@xmath23 m@xmath4 @xcite or 10@xmath24 m@xmath4 @xcite .
however , the @xcite ci study defined a knot " to be the size of 30 @xmath2510 in size , which would include @xmath310 h@xmath0 knots if we assume the knot densities of postion 1 .
because our h@xmath0 knots appear similar in shape and size to the optical knots observed by @xcite , we adopt their mass esimate for individual knots . if , for simplicity , we assume that all h@xmath0 knots have similar properties with an average mass of @xmath3@xmath5 m@xmath4 @xcite , then the total neutral gas mass of the helix nebula is @xmath30.35 m@xmath4 .
our neutral gas mass estimate is substantially larger than the 0.01 m@xmath4 estimated by @xcite who underestimated the total number of knots .
it is also higher than the 0.18 m@xmath4 estimated from the co observations that were corrected for the atomic ci emitting gas by @xcite who have underestimated the mass because the co observations do not detect all of the molecular gas .
our estimate of the neutral gas mass is comparable to the ionized gas mass of 0.3 m@xmath4 @xcite , and the total gas mass in the nebula is @xmath260.65 m@xmath4 in the main part of the disk .
the mass loss rate , that created the main disk over @xmath328,000 years is m @xmath27 m@xmath4 y@xmath28 .
such a large mass loss rate supports the independent conclusion that the helix s progenitor star was massive , 6.5 m@xmath4 , resulting in a present day core mass of 0.93 m@xmath4 @xcite .
figure [ knot1 ] shows a close up of the bright knot in nicmos field position 1 in three tracers , [ oiii ] , h@xmath0 and h@xmath13 .
the knot extinguishes the [ oiii ] emission and its long shadow runs off to the bottom left corner .
the arc - shaped head of the knot is clearly observed in the h@xmath0 emission and is apparent , at lower contrast , in the h@xmath13 emission .
close comparison of the molecular hydrogen emission shows that it almost coincides with the h@xmath13 and is displaced towards the star from the shadowed regions of the the [ oiii ] absorption of the knot ( fig .
[ knot1 ] ) .
this structure suggests that the h@xmath0 emission arises in mini - pdrs on the clump surfaces that point toward the central star .
the intensities measured for the h@xmath0 emission are consistent with the pn - pdr models of @xcite .
comparison of this knot structure ( fig .
[ knot1 ] ) with the detailed study of an optically bright knot in the inner most regions of the helix @xcite reveals a change in the knot s structure .
the most inward knot studied in h@xmath0 by @xcite is pillar - like with a well developed crown structure .
the h@xmath10 emission in this knot lies closer to the central star compared to the h@xmath0 emission and the h@xmath0 emission lies closer to the central star than the co emission of the knot .
this stratified structure is what we expect for a pdr .
however , we do not see a separated stratification of the h@xmath0 and h@xmath13 emission from our example knot from position
this difference in layered vs. not layered structure of the knots pdrs suggest that the inner most knot has a more evolved pdr front than the position 1 knot .
a comparison of the different field positions shows a progression in the morphology of the h@xmath0 knots with radial distance from the central star .
the positions closest to the central star ( positions 1 and 2 ) have numerous knots , with cometary structures , i.e. arcs at the top spires .
the farthest positions ( 4 and 5 ) have substantially less knots and less structure to the knots .
position 3 , which lies in between the two extremes , has a middle density of knots that are lined in continuous rows with fewer spire structures .
a similar trend from highly structured to almost amorphous was observed in the h@xmath10 structure of the knots by @xcite .
the h@xmath0 emission traces this morphological change even further out into the nebula showing that the neutral gas clumps have even less structure pns at the outer edge of the nebula .
this progression supports the idea that the initial stages of the structure formation is caused by instabilities in the interacting winds front or by the ionization front that are later refined by the photo - excitation .
when molecular hydrogen was first imaged in pns it was found that their brightnesses were too high for the molecular emission to have originated in photo - dissociation regions ( pdrs ) and thus attributed to shock excitation @xcite .
however , the pdr models used for these comparisons were designed for interstellar molecular clouds , rather than circumstellar nebulae around rapidly evolving stars .
as such these models only included far - uv photons and failed to include the soft x - ray emission inherent from very hot ( @xmath26100,000 k ) white dwarfs .
the temperature of the white dwarf also changes as the star evolves during the lifetime of the pn .
furthermore , the original models assumed that the cloud was homogeneous , while the gas around pns is clearly very clumpy in structure .
in addition , the interstellar clouds are not expanding , whereas this is the case for the gas around pns , causing a change in the optical depth of the gas and therefore allowing the photons to penetrate the gas more easily .
the models of @xcite included three of these factors ( evolving central star , expanding gas , x - ray photons ) into their pdr model for pns , and showed that the molecular hydrogen emission was approximately consistent with excitation of h@xmath0 in pdrs in these environments .
the @xcite model has been applied to three pns , ngc 2346 ( vicini et al .
1999 ) , the helix nebula @xcite , and the ring nebula @xcite .
our new observations of the helix molecular knots confirms that the pdr gas in the helix resides in a multitude of mini - pdrs , not in a diffuse molecular component @xcite .
the other , perhaps more suprising fact is that the intensity of individual knots remains fairly constant with distance from the central star in the range 5@xmath29 erg s@xmath7 @xmath8 sr@xmath7 ( table [ surbrtab ] ; fig .
[ nicprof ] ) .
the apparent , almost random variation in the h@xmath0 line intensity with respect to radius ( fig .
[ nicprof ] ) occurs because the number density of knots varies along the cross cut not the intensity of the knots themselves .
the relative consistency of the h@xmath0 line intensity of the individual knots with respect to distance from the central star further supports that the pdrs are distributed as mini - pdrs and there is not one pdr front for the helix but a multitude of them
. the higher rotational lines of h@xmath0 , that were observed with isocam by @xcite , indicate a thermalized temperature of the gas to be @xmath3900 k that appears to be independent of distance from the central star .
the gas density within the knots , @xmath310@xmath3010@xmath31 @xmath18 is high enough to thermalize the flourescently excited h@xmath0 gas .
the observed h@xmath0 line intensities of the helix and the derived molecular gas temperature are approximately consistent with the pn / pdr theory of @xcite . for evolution of the most massive stars , e.g. core mass 0.836 m@xmath4 , at the age of the helix , @xmath316,000 years , @xcite predict h@xmath0 gas temperatures of @xmath31000 k and h@xmath0 2.12 @xmath1 m line intensities of @xmath3@xmath32 erg s@xmath7 @xmath8 sr@xmath7 in good agreement with the observed temperature and the individual h@xmath0 line intensities of the majority of knots . at this stage of the pn pdr evolution ,
the h@xmath0 line intensity decreases only gradually with time and the heating of the molecular gas is dominated by the soft x - ray emission of its 123,000 k central star @xcite . despite this approximate success of the pn / pdr models of @xcite
, these models fall short of complete success .
the brightest h@xmath0 intensity is almost a factor of 10 larger than the @xcite prediction .
recent model calculations of the h@xmath0 intensity of knots in radiative equilibrium with the stellar radiation field by @xcite also under - predict the h@xmath0 line intensities .
the solution to this underprediction of both models , may be to combine the time evolution aspects of the @xcite models with the inclusion of knots , as modelled by @xcite .
for example , the time - dependent process of photo - evaporation causes an advection of the h@xmath0 from the surface of the knot and a propagation of the pdr front into the knot which may boost the h@xmath0 emission because of the constant photodissociation of fresh molecular gas @xcite .
evidence for photoevaporation has been found in the ionized gas studies of the inner most knots @xcite , which are knots that are directly exposed to the central star light .
thus , the higher h@xmath0 knots could be those that are directly exposed to the central starlight and experiencing photoevaporation .
however , most knots experience a softer starlight that has been attenuated by intervening knots of molecular gas and dust .
secondly , the brightest h@xmath0 intensity detected by @xcite , @xmath3@xmath33 erg s@xmath7 @xmath8 sr@xmath7 , may be the result of multiple knots along the line of sight , i.e. filling factors greater than 1 that have a multiplying effect on the intensity .
this raises an interesting point that the spatial distribution of h@xmath0 over the entire nebula varies not because of substantial h@xmath0 intensity variation , as one might expect in a pdr front .
rather , the apparent h@xmath0 surface brightness is proportional to the number density of h@xmath0 knots .
new observations of h@xmath0 2.12 @xmath1 m line reveal several new aspects to the molecular knots of the helix nebula .
the h@xmath0 images reveal that the knots have arcuate structures with the apex pointing towards the central star .
these molecular hydrogen knots are most highly structured in the field positions closest to the central star and become increasingly less structured with increasing radius .
all of the h@xmath0 emission is confined to knots .
in contrast the ionized gas tracers have a significant component of diffuse ionized gas emission . using the number density of molecular hydrogen knots in the 5 nicmos field positions
, we estimate the total number of knots to be @xmath323,000 , a factor of 6.5 more than previous estimates based on optical images .
the total neutral gas mass in the helix based on these new knots estimates is 0.35 m@xmath4 assuming an average mass of @xmath3@xmath5 m@xmath4 for the individual knots based on previous work by @xcite .
the h@xmath0 emission structure of the entire helix nebula supports the recent interpretation of the helix as a nearly pole - on poly - polar planetary nebula .
the average intensity is 5@xmath29 erg s@xmath7 @xmath8 sr@xmath7 remains relatively constant with projected distance from the central star .
the temperature and h@xmath0 2.12 @xmath1 m intensity of the knots suggest an origin in the photodissociation regions ( pdrs ) of the individual knots ; however , theoretical models for the pdrs in planetary nebulae do not adequately reproduce the h@xmath0 intensity .
the brightest knots appear in regions of more numerous knots and may be exposed to direct starlight that may cause rapid photoevaporation in comparison to the more embedded knots of the disk .
we gratefully acknowledge the work of many stsci colleagues who contributed to this observational setup of this project .
zoltan levay who superposed the nicmos fields on the combined ctio and acs image .
this work was supported in part by an stsci grant go 01041 and by the internal stsci funds , ddrf d0001.82319 .
lcccc positions & ra(2000 ) of ( 0,0 ) & dec(2000 ) of ( 0,0 ) & @xmath34ra to h@xmath0 knot & @xmath34dec to h@xmath0 knot + & ( h m s ) & ( ) & ( ) & ( ) + central star & 22 29 38.5&@xmath3520 50 13.5 + pos1 & 22 29 53.03 & @xmath3520 53 27.87 & @xmath3528.32 & 31.57 + pos2 & 22 29 40.23 & @xmath3520 54 44.07 & @xmath3534.51 & 18.68 + pos3 & 22 29 27.44 & @xmath3520 56 0.26 & @xmath354.97 & 22.53 + pos4 & 22 29 59.67 & @xmath3520 56 21.79 & 7.11 & 41.01 + pos5 & 22 29 44.98 & @xmath3520 58 10.38 & @xmath357.01 & 35.72 + lccccccc positions & radius & h@xmath0 & h@xmath10/[nii ] & h@xmath13 & [ oiii ] ( 5007 ) & [ sii ] ( 6725 ) + of h@xmath0 knot & ( ) & + pos1 & 253 & 1.27365e-04 & 7.95596e-02 & 9.33381e-03 & 4.31994e-02 & 1.42868e-03 + pos2 & 257 & 9.38642e-05 & 8.19865e-02 & 1.05037e-02 & 6.45470e-02 & 1.51075e-03 + pos3 & 360 & 8.55222e-05 & 4.51389e-02 & 9.29548e-03 & 9.49064e-03 & 1.09471e-03 + pos4 & 448 & 8.61214e-05 & 2.47975e-02 & 2.90198e-03 & 6.68796e-03 & 6.28393e-04 + pos5 & 436 & 5.99320e-05 & 2.42383e-02 & 2.36069e-03 & 3.43582e-03 & 6.77509e-04 + lcccc positions & total & fov area & number density & area filling factor + & & & & of h@xmath0 emission + pos1 & 150 & @xmath36 & 135 knots / arcmin@xmath2 & 80% + pos2 & 180 & @xmath36 & 162 knots / arcmin@xmath2 & 70% + pos3 & 90 & @xmath36 & 81 knots / arcmin@xmath2 & 60% + pos4 & 90 & @xmath37 & 65 knots / arcmin@xmath2 & 50% + pos5 & 20 & @xmath36 & 18 knots / arcmin@xmath2 & 25% + | we present hst / nicmos imaging of the h@xmath0 2.12 @xmath1 m emission in 5 fields in the helix nebula ranging in radial distance from 250450 from the central star .
the images reveal arcuate structures with their apexes pointing towards the central star .
these molecular hydrogen knots are most highly structured in the fields closest to the central star and become increasingly less structured with increasing radius .
comparison of these images with comparable resolution ground based images reveals that the molecular gas is more highly clumped than the ionized gas line tracers . from our images ,
we determine an average number density of knots in the molecular gas ranging from 162 knots / arcmin@xmath2 in the denser regions to 18 knots / arcmin@xmath2 in the lower density outer regions .
the decreasing number density of h@xmath0 knots in the outer regions creates a lower filling factor of neutral and molecular gas emission in the radio observations of co and hi and may explain why these outer regions , where we clearly detect h@xmath0 2.12 @xmath1 m , fall below the detection limit of the radio observations . using this new number density
, we estimate that the total number of knots in the helix to be @xmath323,000 which is a factor of 6.5 larger than previous estimates .
the total neutral gas mass in the helix is 0.35 m@xmath4 assuming a mass of @xmath3@xmath5 m@xmath4 for the individual knots .
the h@xmath0 emission structure of the entire helix nebula supports the recent interpretation of the helix as a nearly pole - on poly - polar planetary nebula .
the h@xmath0 intensity , 59@xmath6 erg s@xmath7 @xmath8 sr@xmath7 , remains relatively constant with projected distance from the central star suggesting a heating mechanism for the molecular gas that is distributed almost uniformly in the knots throughout the nebula .
the temperature and h@xmath0 2.12 @xmath1 m intensity of the knots can be approximately explained by photodissociation regions ( pdrs ) in the individual knots ; however , theoretical pdr models of pn under - predict the intensities of some knots by a factor of 10 . the brightest h@xmath0 emission ( @xmath9 erg s@xmath7 @xmath8 sr@xmath7 ) may be enhanced by a larger than unity area
filling factor of h@xmath0 knots or may be an individual h@xmath0 knot exposed to direct starlight causing rapid photoevaporation compared with the more embedded knots of the disk . |
it is taken for granted that the nucleon consists of quarks and gluons and the underlying theory describing their interaction is qcd . however , solving the qcd equations for the many - body system at large distances is at the present time still beyond our practical abilities .
deep - inelastic scattering ( dis ) of leptons is known to be a very good tool to study the structure of the proton . here in the bjorken limit one
directly tests the quark distributions in the nucleon .
although the perturbative regime seems to be well under control and the qcd evolution equations turned out to be very successful in relating the quark distributions at different momentum scales , the nonperturbative effects are poorly known .
recent experiments on deep - inelastic scattering of leptons from nucleons @xcite have shown the incompleteness of our understanding of the proton structure .
the violation of the gottfried sum rule observed by nmc @xcite @xmath4{dx\over x } = 0.24 \pm 0.016\ ] ] indicates that the nucleon sea is not flavour symmetric .
recent fits of quark distributions to the world data for deep - inelastic and drell - yan processes confirm the asymmetry @xcite .
also the new results of the cern dedicated na51 experiment @xcite on the dilepton production in proton - proton and proton - deuteron scattering gives evidence for the asymmetry .
more detailed information can be expected from the experiments planned at fermilab @xcite .
perturbative effects @xcite can not explain the observed asymmetry .
the dressing of the nucleon with virtual mesons provides a natural explanation for the excess of @xmath5 over @xmath6 quarks in the proton @xcite and explains the result of the na51 experiment @xcite .
another intriguing result , involving the spin structure of the proton , was found by the emc collaboration .
their measured value of the ellis - jaffe sum rule @xcite caused great excitement .
after a smooth regge extrapolation of the data they have found @xmath7 this value is more then two standard deviations away from the original ellis - jaffe prediction ( @xmath8 ) @xcite based on the assumption of vanishing polarized strange sea .
newer experiments of the smc @xcite and slac @xcite collaborations give a somewhat larger value for @xmath9 .
the polarized deep inelastic scattering experiments at cern have shown that , when supplemented with some information from semileptonic decays , only a small fraction of the proton spin is carried by valence quarks . in view of the importance of the meson cloud for the gottfried sum rule violation
, one might expect that it also plays an important role for the nucleon spin .
indeed , simple estimates within the one pion exchange model seem to indicate that the meson cloud effects could contribute to the so - called spin crisis .
moreover , due to the close connection of the ellis - jaffe sum rule with semileptonic decays of octet baryons ( see section vi ) we also have to evaluate the consequences of the meson cloud for semileptonic decays of octet baryons .
modifications of the axial - vector coupling constants due to the meson cloud can be quite important @xcite since the corresponding axial - currents are not protected against renormalization due to the mesonic cloud .
the analysis of the emc experiment under the assumption of @xmath10 symmetry ( ! ) indicates that there may be a relation between the spin and strangeness content of the nucleon .
the value for the polarized strange quark content of the nucleon obtained from such an analysis is @xmath11 ( newer smc@xcite and slac@xcite experiments indicate a somewhat smaller amount ) , which suggests a revision of the simple view of the nucleon . in meson cloud models
the strangeness content of the nucleon is generated by virtual fluctuations into strange mesons ( @xmath12 or @xmath13 ) and strange baryons ( @xmath14 , @xmath15 or @xmath16 ) .
we stress the fact that the experimental result obtained for the polarized strange quark content has been obtained under the assumption of @xmath10 symmetry , which is approximately realized in semileptonic decays @xcite .
on the other hand , @xmath10 symmetry is violated as far as the masses of hadrons are concerned and it is also violated in the sea quark distributions of the proton , which we discuss in the present paper . in models which include @xmath10 symmetry breaking effects the conclusion about the polarized strangeness content can be modified . on the other hand
, the approximate validity of @xmath10 symmetry for the semileptonic decays of baryons has to be an important check for any baryonic model .
the discussion above shows that in order to understand the structure of the nucleon ( the same is true for other baryons ) one has to treat consistently the flavour and spin content .
because of the surprising success of the simple cabibbo model , the semileptonic decays have to be treated simultaneously .
the pion , being the lightest meson was the first included in the deep - inelastic scattering @xcite . it was suggested in ref.@xcite to include also other mesons which had proven to be important in the low - energy nucleon - nucleon and nucleon - hyperon scattering .
their effects in deep inelastic scattering are discussed in more detail in ref.@xcite .
it has been suggested recently to study the virtual pion component in the nucleon in deep inelastic scattering experiments at hera @xcite .
the semi - inclusive reaction @xmath17 is presently being studied by the zeus collaboration , which has installed a test forward neutron calorimeter to complement its leading proton spectrometer @xcite .
it is also worth mentioning that a recent lattice qcd calculations gave evidence for the importance of pion loop effects for nucleon properties @xcite .
the renewed interest in the virtual pions ( mesons ) in the nucleon , especially in connection to deep inelastic scattering , requires a better understanding of the methods used up to now in the literature on that subject . in the meson cloud model
the nucleon is viewed as a quark core , termed a bare nucleon , surrounded by the mesonic cloud .
the convolution model seems to be the best tool to understand the structure of such a composed object .
the internal consistency of the convolution model has been recently a subject of discussion @xcite .
a special emphasis has been put on satisfying sum rules @xcite .
when standard @xmath18-dependent form factors @xcite are used , both number and momentum sum rules can not be satisfied automatically .
while number sum rules can be satisfied by independent adjustment of cut - off parameters of the vertex form factors for the diagrams with off - shell mesons and baryons , momentum conservation is violated to a degree dependent on the functional form of the form factor ( see also @xcite ) .
it has been pointed out recently that special care is required , as far as form factors are concerned , in order to conserve both sum rules in a natural way @xcite .
for the first time in the literature we present useful complete set of formulas for the helicity dependent splitting functions of the nucleon ( octet baryon ) into pseudoscalar / vector meson octet / decuplet baryon states , i.e. we include all particles which turned out to be crucial in modern meson exchange models for nucleon - nucleon and nucleon - hyperon scattering at low energies @xcite .
in addition to unpolarized deep - inelastic scattering we are also interested in possible effects of the meson cloud for the polarized deep - inelastic scattering .
the proton helicity flow to the pionic cloud has been calculated recently in the relativistic light - cone perturbation theory including @xmath19 and @xmath20 vertices @xcite .
the analysis of unpolarized deep - inelastic scattering suggests that also vector mesons ( @xmath21 , @xmath22 ) may play an important role @xcite , provided rather hard form factors are used .
the importance of vector mesons ( @xmath23 ) for deep inelastic scattering of polarized particles can be even greater .
therefore we extend the formalism of ref.@xcite to vector mesons .
we derive useful formulas for both unpolarized and polarized deep - inelastic scattering . in comparison to our earlier analysis @xcite ,
where only unpolarized deep - inelastic scattering has been considered , in the present paper we analyze possible effects of the meson cloud also for the polarized deep - inelastic scattering as well as for the semileptonic decays of the octet baryons .
there are a few technical details in which the present analysis differs from the previous one @xcite . in ref.@xcite
we have assumed for simplicity a dipole form of the vertex form factors with an universal cut - off parameter for all mesons .
the value of the cut - off parameters of the vertex form factors had been adjusted in ref.@xcite to reproduce the experimental data for @xmath24 .
since the universality of the dipole form factor is not obvious and the fit to the @xmath25 data is not very sensitive to the value of the cut - off parameter in the present paper we take slightly different attitude .
following @xcite , we estimate the free parameters of our model by analyzing high - energy @xmath26 and @xmath27 production processes in a one - boson - exchange model . in distinction to @xcite we include also the effects of the vector meson fock components
. then we first discuss various effects of the meson
baryon fock components on the flavour structure of the nucleon and calculate the gottfried sum rule .
next we present the results of calculations for the axial - vector coupling constants @xmath28 for all possible semileptonic decays of the octet baryons .
finally , we calculate the ellis - jaffe sum rule for the proton and neutron as well as the bjorken sum rule .
we expand the nucleon wave function in terms of a few principle fock components . in this model
the nucleon can be viewed as a bare nucleon ( core ) , surrounded by a mesonic cloud .
the wave function of the nucleon with helicity @xmath29 can be schematically written as @xmath30 , \label{fock}\ ] ] where @xmath31 is the probability amplitude that a physical nucleon with helicity @xmath29 is in a state consisting of a virtual baryon @xmath32 with longitudinal momentum fraction @xmath33 , transverse momentum @xmath34 , and helicity @xmath35 , and a virtual meson @xmath36 with momentum fraction @xmath37 , transverse momentum @xmath38 , and helicity @xmath39 .
note , that the helicities @xmath35 and @xmath39 need not to add up to @xmath29 , since an additional relative angular momentum between the particles is possible . in eq.([fock ] ) @xmath40 is the standard wave function renormalization constant , which can be interpreted as the probability of the bare nucleon @xcite
. it can be expected , that the structure of the core is rather simple .
presumably , it can be described as a three quark system in the static limit .
of course , in the deep inelastic regime at higher @xmath41 additional sea of perturbative nature is created unavoidably by the standard qcd evolution .
the main idea of the convolution approach is that there are no interactions among the particles in a multi - particle fock state during the interaction with the hard photon in deep inelastic scattering ( for a detailed discussion see @xcite ) .
this enables one to relate the contribution of a fock state @xmath42 to the nucleon structure function @xmath43 , to the structure functions of either the struck meson @xmath36 or the struck baryon @xmath32 ( see fig .
1a , b ) @xmath44 the convolution formulas can be written in an equivalent way in terms of quark distributions : @xmath45 .
\label{vconv}\ ] ] the main ingredients in the formulas above are the _ splitting functions _ @xmath46 and @xmath47 , which are related to the probability amplitudes @xmath48 @xmath49 because the description of the nucleon as a sum of @xmath50 fock states is independent of the reaction mechanism , the relation @xmath51 must hold .
it simply expresses the fact , that if a meson @xmath36 carries a longitudinal momentum fraction @xmath33 of the nucleon momentum , the remaining part of the nucleon is a baryon with the remaining longitudinal momentum fraction @xmath37 .
moreover this relation automatically ensures global charge conservation @xmath52 and momentum conservation @xmath53 where @xmath54 and @xmath55 are the first and second moments of the splitting functions . in this notation ,
the wave function renormalization constant is given as @xmath56^{-1}.\ ] ] both the ellis - jaffe sum rule and the axial - vector coupling constants for semileptonic decays can be expressed with the help of matrix elements of the axial - currents @xmath57 ( see section vi ) , where @xmath58 are the gell - mann matrices .
the expansion ( [ fock ] ) implies ( see fig .
2 ) @xmath59 where @xmath60 the essential ingredients in our model are the amplitudes @xmath61 . to calculate these quantities
we employ the time - ordered perturbation theory ( topt ) @xcite which has the advantage that the intermediate fock states can be written down explicitly : @xmath62 ( compare @xcite ) .
this formula gives the amplitude of finding a nucleon with momentum @xmath63 and helicity @xmath29 in a fock state where the baryon @xmath32 has the momentum @xmath64 , helicity @xmath35 and the meson @xmath36 the momentum @xmath65 and helicity @xmath39 .
the factors @xmath66 ( @xmath67 ) are the usual fermion wave function normalization factors @xmath68 , @xmath69 ; @xmath70 is a bosonic normalization factor @xmath71 .
the important feature of topt is , in contrast to the covariant perturbation theory , that the intermediate particles are on their mass - shell .
therefore , the vertex function @xmath72 in eq.([phi ] ) can be calculated by using on - mass shell spinors .
@xmath72 depends on a particular model , i.e. on the form of the lagrangian used . in general @xmath72
can be written as @xmath73 where summing and averaging over all possible spin - states is implicitly assumed .
@xmath74 , @xmath75 and @xmath76 are bi - spinor and/or vector indices dependent on the representation used for particles of a given type .
@xmath77 and @xmath78 are the wave functions ( field operators ) of the intermediate meson and baryon , respectively .
it has been shown @xcite for the @xmath79 case , that in the infinite momentum frame ( imf ) contributions of fock states with anti - particles vanish and only contributions with forward moving particles survive .
this statement is also true for other fock states we are dealing with . in the imf - limit
the momenta of the particles involved can be parameterized in terms of @xmath33 and @xmath80 : @xmath81 in the limit @xmath82 only states with @xmath83 $ ] do not vanish @xcite , and the amplitudes @xmath48 can be expressed as @xmath84 with @xmath85 being the invariant mass squared of the @xmath42 fock state @xmath86 @xmath87 is the vertex - function in the imf . in the formula above an extra factor
@xmath88 has been taken out .
it would cancel when going to probability densities by an appropriate factor of the jacobian of the transformation ( [ kinimf ] ) .
if the vertex function used does not contain a derivative of the meson field , we could have equally well used covariant perturbation theory instead of topt .
but if the vertex function used contains a derivative of the meson field this is no longer true . to illustrate this point
let us consider the example of the pseudovector @xmath19 vertex , given by @xmath89 where for simplicity the coupling constant and isospin structure have been suppressed .
the standard covariant technique @xcite leads to the following splitting function of the meson @xmath90 \over [ m_n^2-m_{bm}^2(1-y , k^2_\perp)]^2}. \label{pvmb}\end{aligned}\ ] ] the result for the baryon is : @xmath91 ^ 2 } .
\label{pvbm}\end{aligned}\ ] ] here the two results are not related by eq.([sym ] ) , which leads to violation of the charge and momentum conservation .
the reason for this puzzle is , that by using a derivative coupling , an additional off - shell dependence is introduced into the vertex function , which can not be suppressed in the imf - limit .
a way out is to use topt . here
, however , the problem arises how to choose the meson energy in the vertex . in principle
, there are two possible prescriptions : * one uses in the vertex the meson four - momentum @xmath92 : @xmath93 , i.e. the meson energy in the vertex is @xmath94 . with this form of the vertex
one reproduces the baryon splitting function given by eq.([pvbm ] ) .
the meson splitting function is related to this result by eq.([sym ] ) .
* instead of @xmath92 one uses the difference of the baryon four - momenta @xmath95 : @xmath96 , i.e. the meson energy in the vertex is @xmath97 . with this prescription one
gets the splitting function given by eq.([pvmb ] ) . also here
the resulting splitting functions for the baryon and meson fulfill eq.([sym ] ) .
thus topt , in contrast to a covariant calculation , is consistent with the convolution approach .
the remaining point to clarify is which of the two prescription one should use .
we will always use prescription b , because in this prescription the splitting functions for the pseudovector case are identical to those of the pseudoscalar case , if the coupling constants are identified properly .
moreover , in this prescription the structure of the vertex is due to the baryonic current only @xcite .
the most general structure of the hadronic tensor must include the dependence on off - shell effects @xcite .
because a realistic calculation of such off - shell effects is not possible at present , we shall neglect such effects in this paper . because of the extended nature of the hadrons involved one has to introduce phenomenological vertex form factors , which parameterize complicated unknown microscopic effects .
the form factors are included in the calculation by replacing the vertex function @xmath98 by @xmath99 .
now relation ( [ sym ] ) , which can be shown to hold exactly in the case of point - like particles ( in either covariant and topt calculations ) , imposes a severe restriction to these form factors : @xmath100 the form factors often used in meson exchange models and convolution models are functions of @xmath18 only , the four - momentum squared of the meson .
it is to be expected that form factors which depend only on the kinematical variables of only one of the two particles of the two - body system , like the often used dipole form factor @xmath101 will not satisfy eq.([sym ] ) .
it can be shown @xcite that these form factors do not conserve basic quantities like charge and momentum simultaneously .
one simple method to obtain form factors with the right symmetry is to multiply a @xmath18-dependent form factor by a @xmath102-dependent one with the same functional form with @xmath103 replaced by @xmath104 @xmath105 in terms of the imf variables @xmath33 and @xmath106 , @xmath18 and @xmath102 , the four momentum squared of the intermediate baryon , is given by @xmath107 the importance of using such symmetric form factors was noticed only recently @xcite .
another possible approach , to fix the cut - off parameters to assure number sum rules ( global charge conservation ) ( see @xcite ) is in our opinion somewhat arbitrary , and does not guarantee momentum conservation . in numerical calculations we use vertex form factors in the exponential form
@xmath108 , \label{form}\ ] ] where @xmath109 is the invariant mass squared of the baryon - meson @xmath42 fock state @xmath110 a form factor of this type , first introduced in ref.@xcite , fulfills the necessary symmetry conditions eq .
( [ sym ] ) .
parameterization of the vertex form factors in terms of the invariant mass of the two - body system is very natural in the light - cone approach .
furthermore , the light - cone approach is the best suited for applications to deep inelastic scattering . in order to fix the cut - off parameters @xmath14 of the form factors we use high - energy baryon production data .
the neutron production data in the @xmath26 scattering seems to be best tailored for extracting the cut - off parameter for the @xmath111 and @xmath112 fock states .
because we shall limit ourselves to data for relatively low exchanged four - momenta , it is reasonable to assume that the neutron is produced by a simple one - boson - exchange mechanism ( obe ) ( see fig.3 ) . in this model
the differential cross section can be described as a product of the probability for a proton being in a @xmath113 or @xmath114 fock state and the total cross section of @xmath115 scattering . in general , the invariant cross section for @xmath116 production in the obe model is @xmath117 here @xmath33 is the longitudinal momentum fraction of the baryon with respect to the momentum of the incoming proton ; @xmath34 the corresponding transverse momentum .
specialized to neutron production with @xmath118-exchange this formula reads : @xmath119 ^ 2 } { |g_{nn\pi}(y , k^2_\perp)|^2\over y(1-y ) } \\ & & \cdot\sigma^{\pi p}_{tot}(s(1-y ) ) .
\nonumber\end{aligned}\ ] ] for @xmath120-exchange the formula is much longer and will be not presented here .
the necessary ingredients can be found in appendix b. these formulas are strictly valid only in the imf - limit , but for a sufficiently high energy they are a good approximation . in the following we neglect the slow energy dependence of the total @xmath79 cross section and
use a value of @xmath121 @xcite .
in addition , due to the lack of experimental data we also assume @xmath122 . in fig.4
we show a fit by the obe model to the experimental data ( the @xmath123 data are taken from ref .
@xcite with @xmath124 ; the other data are taken from ref .
@xcite with @xmath125 ) . in this calculation
we take @xmath126 @xcite and @xmath127 , @xmath128 @xcite . as a criterion for the fit
we have assumed that the calculated result must not exceed the experimental data . for low @xmath129 @xmath130-exchange
is the dominant contribution , the @xmath21-exchange plays a rather marginal role . for higher
@xmath129 @xmath21-exchange becomes the dominant mechanism .
thus the cut - off parameter for the @xmath131 and @xmath132 vertices can be fixed almost unambiguously .
the fit to the experimental data yields @xmath133 .
we obtain a good description of the experimental data at intermediate @xmath134@xmath135 . outside this region
the fit fails .
this is to be expected , because for low @xmath33 one could expect additional contributions from e.g. multi - meson exchange processes . for our purpose
( later we shall concentrate on moments of @xmath1-distributions rather than the distributions themselves ) the assumption of a obe mechanism is sufficient , as the maximum cross section lies at intermediate @xmath33 , where the mesons can be treated as not reggeized .
the correct @xmath136-dependence obtained from the fit is a further argument in favor of the simple obe model . despite the success of the fit one should bear in mind
that we obtain only an upper bound of the @xmath21-exchange contribution . in order to extract information about the @xmath137 vertex
, one can examine high - energy @xmath14 production in the @xmath138 reaction in the same way as neutron production . also here
we assume that the dominant mechanism is a obe mechanism , @xmath12 or @xmath13-exchange .
the differential cross section for @xmath12-exchange is given by @xmath139 ^ 2 } { |g_{n\lambda k}(y , k^2_\perp)|^2\over y(1-y)\\ & & \cdot\sigma^{kp}_{tot}(s(1-y ) ) .
\nonumber\end{aligned}\ ] ] for @xmath13-exchange we refer to eq .
( [ cross ] ) and appendix b. we have taken @xmath140 , which is related by @xmath3 symmetry to the @xmath141 given before . the coupling constant obtained from experiment is @xmath142 @xcite which supports the assumption of @xmath3 symmetry .
also for the coupling constants involving vector mesons we assume @xmath3 symmetry ( for details see @xcite ) . the total cross section @xmath143 is @xmath144 @xcite .
for @xmath145 we assume the same value .
the fit of the cut - off parameters @xmath146 and @xmath147 is shown in fig.4 .
the general features of the fit are similar to those found for neutron production . also , the cut - off parameters @xmath148 obtained from the fit are quite similar to those found for @xmath130- and @xmath21-exchange .
thus , it is possible to assume a kind of universality and use a cut - off parameter @xmath149 for all vertices involving octet baryons and pseudoscalar or vector mesons .
this is rather attractive , that the same cut - off parameters can be used for pseudoscalar and vector mesons . in the literature , instead of the symmetric form factor given by eq.([form ] ) , a dipole form factor eq .
( [ dipole ] ) is often used . as discussed above
, the use of such form factors leads to a violation of basic symmetries . in fig.5
we show , for example , the differential cross section for neutron production calculated using the dipole form factor ( @xmath150 @xcite ) .
as seen from the figure , one gets a satisfactory description for small @xmath151 .
one can not , however , describe the data at large @xmath129 .
comparison of fig.4 and fig.5 clearly indicates a preference for the symmetric form factor ( eq . [ form ] ) .
the violation of the gottfried sum rule is an interplay between the @xmath79 and @xmath152 components and crucially depends on vertex form factors , i.e. its functional form and cut - off parameter .
for instance , taking the dipole form and assuming the same cut - off for both components leads to strong cancellations . as a consequence only about half of the experimentally observed violation
can be explained @xcite . in order to shed light on the @xmath20 vertex we follow @xcite and examine @xmath153 production in @xmath154 and @xmath155 collisions .
the differential cross section for one - pion - exchange ( fig.6 ) is given by : @xmath156 ^ 2 [ ( y m_n - m_\delta)^2+k^2_\perp]\over 6y^2 m_\delta^2 [ m_n^2-m^2_{\delta\pi}(y , k^2_\perp)]^2 } \label{delta } \\ & & \cdot\sigma^{\pi h}_{tot}(s(1-y ) ) . \nonumber\end{aligned}\ ] ] the differential cross section for @xmath21-exchange can be calculated analogously ( see eq . [ cross ] and appendix b ) . as coupling constants we take @xmath157 and @xmath158 @xcite .
the available @xmath153-production spectra @xcite are biased by an ambiguity in extraction of background contributions .
thus , in fig.7 two scales are shown ( different background assumptions ) .
the experimental data shown in fig.7 have been taken at three different energies @xmath159 , @xmath160 and @xmath161 .
we do not distinguish the data taken at different energies , as no energy dependence is visible .
this fact is in agreement with the simple assumption of an obe mechanism and supports the imf - approach .
as far as the absolute normalization is concerned some extra experimental information is needed .
we use more precise @xmath33-integrated - spectra dependent only on @xmath106 @xcite for this purpose .
the result of the fit shown in fig.7 and fig.8 yields @xmath162 .
in contrast to the @xmath163 and @xmath14 production an unambiguous distinction between the @xmath130 and @xmath21 exchange contributions is not possible . for simplicity
, we have assumed the same cut - off parameters for both , in analogy to the octet baryon production .
we assume that the cut - off parameters for the whole decuplet to be the same as those obtained from the @xmath27 production process .
this universality within an @xmath10 multiplet greatly reduces the number of , in principle unknown , parameters . to reduce the number of parameters further
, we employ @xmath3 symmetry as mentioned above to relate coupling constants .
thus only the coupling of the @xmath22 to the nucleon remains to be fixed .
it has to be treated differently , because the @xmath22 is an almost ideal mixture of an octet and a singlet @xcite @xmath164 the @xmath22-couplings were found to be : @xmath165 and @xmath166 @xcite .
it is interesting to ask the question of how sensitive is our model to the functional form of the form factor . in the following
we restrict ourselves to form factors which have the correct symmetry .
the simplest are the symmetric monopole ( @xmath167 ) or dipole ( @xmath168 ) form factors : @xmath169 in principle , one might expect that they produce results different from the exponential form ( eq . [ form ] ) for the first moments of the splitting functions , because they cut off the integrands much more softly .
surprisingly , this is not the case . in table 1
we show values of cut - off parameters of the exponential , monopole and dipole form factors fitted to the neutron production data and the corresponding moments of the splitting functions .
almost the same first moments are obtained with the different functional forms .
our results are almost insensitive to the form of the form factors used ; it is rather the symmetry ( eq .
[ sym ] ) which allows one to describe the data in a broad range of @xmath33 and @xmath136 .
the renewed interest @xcite in the meson cloud of the nucleon comes from the fact that it provides a natural explanation of the @xmath5-@xmath6 asymmetry and , as a consequence , the violation of the gottfried sum rule .
the gottfried sum rule can be expressed as @xmath4{dx\over x } = { 1\over3}\int [ u(x)+\bar u(x)-d(x)-\bar d(x ) ] \ ; dx = { 1\over 3}+{2\over 3}a_{\bar u-\bar d } , \label{gottfried}\ ] ] where @xmath170 ( @xmath171 ) are the light quark ( antiquark ) distributions in the proton and @xmath172\,dx$ ] . in obtaining eq.([gottfried ] ) isospin symmetry within the nucleon doublet has been assumed . in accordance with the simple structure of the bare nucleon , it is natural to expect that the only sea in the bare nucleon is of perturbative origin .
such a sea , being @xmath173 symmetric , does not contribute to @xmath174 . in our model
the asymmetry is caused by the valence quarks in the virtual mesons .
only @xmath130 and @xmath21 mesons contribute to the asymmetry @xcite , which can be expressed in terms of the first moments of the corresponding splitting functions @xmath175.\ ] ] contributions of a perturbative nature as well as any symmetric sea in mesons and/or baryons would not change this result , because @xmath176 is sensitive only to the difference @xmath177 ( eq . [ gottfried ] ) . in table 2
we show various contributions to the @xmath178 asymmetry and to the gottfried sum rule .
the inclusion of the @xmath79 fock component leads to a reduction of @xmath176 from @xmath179 to @xmath180 . in comparison to earlier works where
an universal cut - off parameter of the dipole form factor has been assumed for simplicity @xcite , the procedure proposed in section iii leads to important modifications .
we find a much smaller probability of the @xmath181 ( 6.2 % in comparison to 16.1 % in ref.@xcite ) fock component and a substantially larger probability of the @xmath182 component ( 10.9 % in comparison to only 1.6 % in ref.@xcite ) .
summarizing , including all components in the nucleon wave function expansion ( [ fock ] ) leads to a good agreement with the nmc result @xcite of the gottfried sum rule @xmath183 ( see table 2 ) .
it should be noted here , that the true experimental value of @xmath176 might be even somewhat smaller than the nmc result due to shadowing effects in the deuteron @xcite . in calculating the @xmath184 asymmetry in the present paper we have neglected the interference diagrams like @xmath185 or @xmath186 .
while the contributions of such diagrams to the number and momentum sum rule cancel , their contribution to @xmath187 , @xmath188 as well as to the @xmath184 asymmetry is not protected by any conserved current and is in principle possible .
a reliable calculation of the @xmath1-dependence of the @xmath189 difference is rather difficult and requires microscopic models for the mesons involved .
in contrast to the @xmath1-dependence the bulk effect of @xmath190 can be easily estimated assuming that the discussed mesons are members of the @xmath10 multiplets .
the contribution of the interference diagrams of two mesons to the @xmath189 difference can be calculated in a full analogy to the diagonal terms : @xmath191 where the matrix element @xmath192 can be calculated using the @xmath3 wave functions of the mesons involved .
we find @xmath193 and @xmath194 for @xmath195 and @xmath186 , respectively .
these contributions have to be compared to @xmath196 ( @xmath130 ) and @xmath197 ( @xmath21 ) obtained from the diagonal mesonic terms .
there are different reasons for the smallness of these contributions . while the @xmath185 is small due to a small coupling constant @xmath198 , the @xmath186 contribution is small due to the vanishing tensor coupling constant @xmath199 .
thus the peculiarities of the nucleon - meson - nucleon couplings allows one to safely neglect the interference terms .
as has been emphasized in ref.@xcite , an attractive feature of the meson cloud is that a large fraction of the nucleon sea can be attributed to the virtual mesons ( see also ref.@xcite ) .
let us consider this issue in more detail in the case of the present model .
for this purpose the global ccfr experimental data @xcite @xmath200 obtained at an average value of @xmath201 are useful . in the formulas above @xmath202 denote second moments of the quark distributions @xmath203 . obtaining the quantities given by eq.([ratio ] ) in our model is not straightforward .
first of all , the experimental data has been measured at relatively large @xmath41 .
therefore , the model has to be extended by the inclusion of extra sea of a perturbative nature .
secondly , the quantities in eq.([ratio ] ) are sensitive to the valence quark distributions which can not be calculated fully consistently @xcite . in spite of these difficulties it seems instructive to make an estimate of those quantities .
it is consistent within our approximation scheme to assume a simple structure of bare baryons and take into account fully dressed mesons .
calculation of the ccfr quantities implies knowledge of the @xmath1-dependences of the quark distributions . for the mesons we take quark distribution functions known from the analysis of the drell - yan processes in the pion - nucleus collisions @xcite . for the bare baryons the situation is somewhat more complicated , and some approximations are unavoidable . for the purpose of comparison with the global quantities like those given by eq.([ratio ] ) it is sufficient to approximate valence quark distributions by those of the physical baryons . a similar approximation for the sea quarks is , however , not realistic as within our model a large fraction of the nucleon s sea is included explicitly via the pionic ( mesonic ) cloud . as discussed in ref.@xcite
, this effect can be approximately included by reducing the total nucleon s sea , known from dis , by a factor @xmath204 .
our model , as discussed later , predicts both asymmetry between @xmath205 and @xmath206 quark distributions and a suppression of the strange sea in comparison to the non - strange sea . on the other hand ,
perturbative qcd predicts @xmath207 - @xmath206 symmetry and allows for relative suppression of the strange component .
it is an empirical fact that at @xmath41 = 4 gev@xmath208 the strange sea is suppressed by a factor 2 .
it is interesting how much of this effect is of perturbative / nonperturbative nature .
one could use the ccfr parameter @xmath209 ( see eq.([ratio ] ) ) to fix the strangeness suppression factor @xmath210 for the sea quark distributions in mesons , baryons and the bare nucleon ( not dressed with mesons ) , @xmath211 in the present paper we adjust the two unknown parameters @xmath204 and @xmath210 to the ccfr quark distributions @xmath212 @xcite and to the ccfr parameter @xmath209 @xcite . in principle
both @xmath213 and @xmath209 depend explicitly on @xmath204 , @xmath210 . in practice
the situation is simpler because @xmath213 depends mainly on @xmath204 and is almost independent of @xmath210 while @xmath209 depends mainly on @xmath210 and is almost independent of @xmath204 .
this allows us to extract the two parameters .
if we take the @xmath214 msr @xcite sea parameterization for the bare nucleon ( we assume that the valence quark distributions in baryons are related via @xmath10 symmetry ) , we find @xmath215 and @xmath216 . the corresponding @xmath217 is compared with the `` experimental '' sea quark distribution of the ccfr collaboration @xcite in fig.9 and the resulting parameter @xmath209 = 0.40 .
the other ccfr parameters are then : @xmath218 , @xmath219 , slightly above their experimental counterparts .
the small overestimation can be easily understood because these two quantities are sensitive to the valence quark distributions which can not be calculated fully consistently @xcite .
we expect the values of @xmath220 and @xmath221 to be biased by the approximations made up to 15% .
while @xmath204 depends to some extend on the parameterization of the sea quark distributions used , @xmath210 is fairly stable
. therefore the result of the fit suggests that the empirical suppression of the perturbative strange sea component at @xmath41 values of a few gev@xmath208 is about 0.75 in comparison to the total strange sea suppression of 0.5 .
the difference is due to nonperturbative effects of the meson cloud .
not only global quantities like the gottfried sum rule or the ccfr parameters are of interest .
also the @xmath1-dependence of various quantities is helpful in testing models .
it is known that sufficiently far above the strangeness production threshold the qcd evolution equations @xcite lead to the production of a perturbative @xmath10 symmetric sea . in this context
it is useful to study differences of quark distributions instead of the distributions themselves , especially those describing the @xmath1-dependence of the @xmath173 and @xmath10 symmetry violation of the sea quarks @xcite .
the @xmath222 is the quantity which describes the @xmath1-dependence of the @xmath173 symmetry breaking of the nucleon sea .
this quantity can not be easily obtained from the integrand of @xmath176 which , in addition to @xmath223 , involves the @xmath1-dependence of valence quark distributions ( see eq.[gottfried ] ) , making the extraction of the interesting quantity rather difficult .
the analysis of dilepton production in @xmath154 and @xmath224 collisions seems to be much better in this respect @xcite .
@xmath225 is of special importance since it can be used to verify our model as it does not require further parameters .
having fixed the cut - off parameters of the vertex form factors from hadronic reactions it can be obtained parameter free . if one assumes that the sea of the bare nucleon is @xmath173 symmetric , then only pions ( mesons ) contribute to @xmath225 . if , in addition , one assumes that the pion sea is also @xmath173 symmetric ( there are no reasons for an asymmetry ) , then the whole effect comes from the valence quarks in the pion which are relatively well known from the drell - yan processes . in fig .
10a we show the predictions of our model for three different pion structure functions : the leading order ( lo ) parameterization of the na3 collaboration @xcite ( dotted line ) , the next - to - leading order ( nlo ) parameterization of ref.@xcite ( solid line ) and structure function of the pion calculated in a model of the radiatively generated sea @xcite ( dashed line ) . as seen from the figure the results for different parameterizations are rather similar . in fig .
10b we compare our result ( solid line ) with parameterizations of the quark distributions which have been fitted to the world data on dis and drell - yan processes .
it is worth noting that our result is almost identical for @xmath226 with a recent nlo parameterization msr(a ) of the msr group @xcite which includes both the new hera data @xcite and the experimental result of the drell - yan na51 experiment @xcite .
the drell - yan experiment planned at fermilab @xcite will be a severe test of our model and should shed new light on the origin of the gottfried sum rule violation .
let us consider now the @xmath1-dependence of @xmath227 which has been extracted by kumano @xcite from the experimental data of the e615 collaboration @xcite . in fig .
11 we show the prediction of our model . in analogy to the @xmath228 difference we show separately the effect of the valence quarks in mesons ( dashed line ) .
it explains already a significant fraction of the total strange quark suppression .
the dotted line shows in addition the effect when the extra perturbative contributions , as described in this section , are included .
the slight overestimation of this experimental data may suggest even smaller suppression of the perturbative component with respect to that obtained based on the ccfr data .
the net strangeness of the nucleon is zero , which requires : @xmath229 .
it has been customarily assumed that @xmath230 .
there is no basic principle that forces one to this assumption other than the fact that it appears as a consequence of a perturbative approach . in this context
nonperturbative effects can be of crucial importance .
recently @xcite , a possible charge @xmath1-asymmetry of the strange sea has been studied in the chiral gross - neveu model .
although a large asymmetry has been found , no absolute normalization of the effect was possible .
a similar analysis has been performed somewhat earlier by signal and thomas @xcite . in their approach
the result strongly depends on the bag radius . at present ,
experimental proposals to investigate this effect are under discussion @xcite .
the presence of ( strange meson)-(strange baryon ) fock components in the nucleon wave function suggests the asymmetry in a natural way , as @xmath231 quarks are constituents of light mesons ( @xmath232 ) and @xmath233 quarks are constituents of heavy baryons ( @xmath234 ) .
let us estimate the effect in our model . in distinction
to ref.@xcite , we can estimate the absolute magnitude of the effect as our free parameters have been determined ( see the preceding section ) by fitting to the experimental data on @xmath235 production at high energy .
the contributions from the valence strange quarks in baryons and valence anti - strange quarks in mesons are shown in fig .
12a for two different sets of quark distributions : ( a ) lo quark distributions in the pion @xcite and in the nucleon @xcite ( dashed line ) and ( b ) nlo quark distributions in the pion @xcite and in the nucleon @xcite ( solid line ) . making the plausible assumption that the effects discussed in the present paper are the only source of the @xmath236 asymmetry
, the @xmath236 difference can be obtained without additional free parameters .
naively one could expect @xmath237 in the large-@xmath1 region and @xmath238 in the small @xmath1 region ( see also @xcite ) .
we get a rather opposite effect to the naive expectation as seen in fig .
12b , where we show predictions of our model using the same structure functions as in fig . 12a .
these contributions are rather small in comparison to similar contributions to @xmath207 and @xmath206 quark distributions ( see for instance ref.@xcite ) . as seen from the figure , in comparison to the @xmath239 difference , @xmath240 depends not only on the parton distributions in mesons but also on the parton distributions in baryons .
this difference can be expressed in terms of the convolution integrals @xmath241 \ ; . \label{sas}\end{aligned}\ ] ] eq .
( [ sas ] ) shows that the shape of the @xmath240 difference depends both on the splitting functions @xmath242 and valence quark distributions both in @xmath32 ( @xmath243 ) and @xmath36 ( @xmath244 ) .
therefore , the final result depends on two competing effects : ( a ) @xmath47 peaks at @xmath245 ( for the @xmath246 case , for instance , it is , however , close to @xmath247 ) ; ( b ) the valence quarks in mesons are concentrated at larger @xmath1 in comparison to the valence quarks in baryons ( 2 vs. 3 valence quarks , respectively ) .
it is the second effect which wins in our case .
because the @xmath240 difference depends on these two competing effects it is somewhat less reliably determined in comparison to the @xmath239 difference . the total ( anti)strange quark distributions obtained by applying the procedure described above are shown in fig.12c .
in addition we present a decomposition into contributions from the meson s sea and that from the bare nucleon and bare baryons .
can the strange quark density be obtained in a reliable way more directly from experimental data ?
it was proposed that the opposite sign dimuon production in the charge current ( anti)neutrino dis can be used as an unique probe of charm particle production and the strange sea content of the nucleon @xcite .
the mass of the charm quark introduces a threshold suppression into the dimuon production rate which has been overcome in ref.@xcite by applying a simple slow - rescaling model @xcite . for comparison in fig .
12c we show also the result obtained by the ccfr collaboration in the e744 and e770 experiments at fermilab @xcite ( dashed line ) . by parameterizing the sea quark distributions by the simple @xmath248 form
, the ccfr collaboration has found a quantitative indication that the strange sea in the nucleon is softer than the non - strange sea with @xmath74 = 9.45 vs. 6.95 @xcite for the non - strange distributions .
this observation is consistent with the prediction of our model . in the large-@xmath1 ( @xmath2490.25 ) region
the non - strange sea is completely dominated by the mesonic ( mainly pionic ) effects which exhaust practically the total strength obtained from the analysis of dis data .
in contrast in the strange sector the mesonic effects are strongly suppressed by the meson / baryon mass effects .
thus the model predicts less strength in the large-@xmath1 region in comparison to the non - strange sea quark distributions . although we get a rather reasonable agreement with the ccfr dimuon data we do not consider it as an ultimate test of our model .
first of all , as already discussed , the mesonic effects are not the only source of the strange quarks at @xmath41 of a few gev@xmath208 .
secondly , we want to stress that the interpretation of the ccfr charm production data should be taken with some grain of salt . it has been demonstrated recently @xcite that a clear interpretation of the ccfr result in terms of the strange quark distributions is questionable at least in the case of the ( @xmath250 gluon ) fusion component of the strange sea and the experimentally measured cross sections may include also contributions of different origin . furthermore , it is not fully clear whether the slow - rescaling procedure includes the mass effects in a correct way .
our effect of the @xmath240 asymmetry is very small .
recently , this asymmetry has been studied by the ccfr collaboration . they have found a very little effect @xcite , almost consistent with zero , in agreement with our predictions .
in terms of quark distributions the ellis - jaffe sum rule can be written as @xmath251 with @xmath252 being the polarized quark distributions @xmath253\ > dx,\ ] ] where @xmath254 @xmath255 $ ] is the quark distribution with flavour @xmath170 having spin ( helicity ) parallel ( anti parallel ) to the nucleon spin .
it is well known that semileptonic decays of octet baryons can be well described in the cabibbo model , where one assumes that the axial currents responsible for the semileptonic decays belong to an @xmath10 octet .
the diagonal matrix elements of these axial currents in this model give the well - known connection to the ellis - jaffe sum rule .
they can be expressed as @xmath256 with @xmath257 being the spin - vector ( @xmath258 , @xmath259 ) , @xmath260 is an axial current defined by @xmath261 , with @xmath262 being @xmath10 generators ( @xmath58 are the gell - mann matrices ) , and @xmath263 being the identity matrix and @xmath264 the axial coupling constants @xmath265 and @xmath266 can be fixed by fitting to semileptonic decay data . the non - singlet quantities @xmath267 and @xmath268 are multiplied by a known qcd correction factors @xmath269 @xcite ( the correction factor for the singlet quantity @xmath270 is different ) which for the sake of simplicity we shall suppress throughout this paper . taking the emc result ( eq.[emc ] ) at face value and assuming @xmath10 symmetry ( ! ) , in combination with experimental results for the neutron beta decay and the semileptonic decay data for the octet baryons @xcite , led to an unexpected result @xmath271 which was consistent with zero .
this can be interpreted such that only a small fraction of the proton spin is carried by quarks .
newer experimental data taken at cern @xcite and slac @xcite are to some degree controversial . the experimental data for the polarized muon scattering on a deuterium target @xcite agrees with the earlier emc data , with the ellis - jaffe integral smaller than the prediction of the ellis - jaffe sum rule , the fraction of the nucleon spin carried by quarks very small and appreciable negative polarization of the strange quarks .
on the other hand , the result obtained in polarized electron scattering from the @xmath272 target @xcite in the slac e142 experiment agrees with the prediction of the ellis - jaffe sum rule , finding that a relatively large fraction of the nucleon spin is carried by quarks and giving a @xmath273 consistent with zero .
the experimental results will be summarized later when comparing with results obtained from our model . in the naive @xmath3 quark model ( nqm )
@xmath274 and @xmath275 , and as a consequence @xmath276 , @xmath277 . the experimental result ( eq . [ g0 ] )
is in marked disagreement with the nqm .
the nqm neglects the fact that the proton constituents are highly relativistic @xcite and interacting @xcite objects .
furthermore , not only quarks contribute to the axial - vector singlet but also gluons .
unlike the flavour octet currents , the singlet current @xmath278 has an anomalous divergence ( for a review see e.g. ref.@xcite ) @xmath279 where @xmath280 is the number of active flavours and @xmath281 is the gluonic field .
the triangle axial anomaly gives rise to an independent gluonic contribution to the flavour singlet axial current @xcite .
as proposed by efremov - teryaev @xcite and altarelli - ross @xcite the parton model ( pm ) values @xmath282 , @xmath283 and @xmath273 should be modified @xmath284 where @xcite @xmath285 is a correction induced by the axial anomaly .
@xmath286 is the probability to find a quark of longitudinal momentum fraction @xmath287 in a gluon and appropriate polarization .
the total gluonic contribution to @xmath288 is @xmath289 @xmath290 in the last formula is the fraction of the proton spin carried by gluons . as a consequence
@xmath291 should be replaced by @xmath292 .
the coefficients @xmath293 have been calculated @xcite .
on the other hand , @xmath290 can not be calculated and most of the authors tried to adjust @xmath290 to describe experimental data ( see for instance @xcite ) .
there has been also much discussion over the uniqueness of the separation into @xmath294 and @xmath295 ( see for instance @xcite ) .
it has been argued @xcite that the anomaly induces also a non - perturbative gluonic contribution to @xmath296 .
although it is very important in understanding the ellis - jaffe sum rule violation , up to now no definite treatment of the axial anomaly has been worked out .
we shall come back to the problem of the anomaly when discussing the mesonic effects .
according to our present understanding , the weak semileptonic decays of the octet baryons can be classified into two groups : either a @xmath297-quark is transformed into a @xmath102-quark , or an @xmath233-quark is transformed into a @xmath102-quark .
the matrix elements of the current operators responsible for the semileptonic decays of the baryons belonging to the octet can be parameterized in terms of @xmath298-dependent form factors @xmath299{rllll } c\bar u_{b_1}\big [ & f_1(q^2)\gamma_\mu & \displaystyle + i{f_2(q^2)\over m_1+m_2}\sigma_{\mu\nu}q^\nu & \displaystyle + { f_3(q^2)\over m_1+m_2 } q_\mu & \\ + & g_1(q^2)\gamma_\mu\gamma^5 & \displaystyle + i{g_2(q^2)\over m_1+m_2 } \sigma_{\mu\nu } q^\nu\gamma^5 & \displaystyle + { g_3(q^2)\over m_1+m_2 } q_\mu\gamma^5 & \big ] u_{b_2}. \\ \end{array}\ ] ] the factor @xmath300 here is the cabibbo factor ( @xmath301 or @xmath302 ) . at low momentum transfer only two terms , @xmath303 ( vector ) and @xmath304 ( axial vector ) ,
are important .
it is customary to extract from experiments the ratio @xmath305 .
the semileptonic decays can be well described assuming the so - called @xmath10 ( cabibbo ) model@xcite . within this model
the operators for the @xmath306 and @xmath307 transitions can be expressed in terms of @xmath10 group generators @xmath308 which are related to the familiar gell - mann matrices @xmath309 .
mesonic corrections lead to the renormalization of the axial - vector coupling constants .
the vector coupling constants are protected against renormalization by vector current conservation .
mesonic corrections to the axial - vector coupling constants have been taken into account by calculating the loop corrections to the tree level approximation according to eq.([axial ] ) .
the corresponding diagrams are shown in fig.2 .
preliminary results with the inclusion of intermediate pseudoscalar mesons and associated octet and decuplet baryons have been already presented elsewhere @xcite . to perform numerical calculations within our model
requires the knowledge of the axial coupling constants for the bare octet and decuplet baryons , vector mesons and transitions octet @xmath310 decuplet .
the transitions within the baryonic octet are traditionally parameterized by the so - called anti - symmetric @xmath265 and symmetric @xmath266 coupling constants .
the axial coupling constant for the transition within the decuplet ( @xmath311 ) can be fixed by the relation @xmath312 . in analogy
we define the coupling constant for the interference diagram octet @xmath313 decuplet ( @xmath314 ) as @xmath315 .
the matrix elements of axial - vector currents between pseudoscalar mesons vanish .
they are , however , finite for vector mesons . here
the structure is analogous to that of the baryonic octet .
we denote the corresponding constants as @xmath316 and @xmath317 . due to parity - conservation the axial coupling constant @xmath316 vanishes .
a special role is played here by the @xmath22 meson which consists of an octet and a singlet part @xcite . for simplicity
we neglect couplings to the singlet part , although they are in principle present .
the vector meson @xmath310 pseudoscalar meson interference terms have an octet structure analogous to the other cases , with coupling constants called @xmath318 and @xmath319 . in the @xmath3 model , i.e. in the model in which all particles are described by their @xmath3 wave functions @xcite , the axial coupling constants can easily be calculated @xmath320 in table 3 we present a list of all measured semileptonic decays of the octet baryons .
the experimental @xmath321 ratios are taken from refs.@xcite .
in addition , we present the @xmath10 values for @xmath28 s expressed in terms of the symmetric and asymmetric coupling constants and @xmath322 s
. vector current conservation allows one to extract the experimental @xmath28 s , which are presented in the last column . in table 4
we present the result of our calculations for all possible semileptonic decays . in the column named @xmath323
we show the @xmath28 s calculated within our model with the @xmath3 axial - vector coupling constants ( eq . [ su6coup ] ) for the bare hadrons . in the column labeled @xmath324 similar results
are shown with @xmath265 and @xmath266 fitted to the measured values of the axial - vector coupling constants . for comparison
, we show the results for the pure ( no mesonic corrections ) @xmath3 model and pure @xmath10 model ( @xmath265 and @xmath266 fitted to the experimental data from table 3 ) .
the @xmath325 values presented in the last row for each model give an idea of the fit quality .
it is well known that the naive @xmath3 model gives a very poor description of the experimental semileptonic decay data . on the other hand , when fitting the @xmath265 and @xmath266 parameters an extremely good description of the existing data can be achieved .
it is commonly believed that any correction to the @xmath10 model may only destroy the nice agreement .
inclusion of mesonic corrections with @xmath3 axial coupling constants improves the description of the data dramatically ( @xmath326 ) .
an additional variation of the @xmath265 and @xmath266 parameters improves the fit further .
we can not allow for variation of the remaining parameters ( @xmath311 , @xmath314 , @xmath317 , @xmath318 ) , because a completely unrestricted fit could result in unphysical values of parameters since the number of experimental data points is limited to only 5 . in order to demonstrate the effect of different fock components we present in table 5 the axial - vector coupling constants @xmath28 calculated with the inclusion of intermediate pseudoscalar mesons and octet baryons @xmath327 , with the additional inclusion of decuplet baryons @xmath328 and with the additional inclusion of vector mesons @xmath329 . in the case of the @xmath3 axial coupling constants ( eq . [ su6coup ] ) the inclusion of pseudoscalar mesons
octet baryons fock components improves the quality of the fit to the experimental @xmath28 s in comparison to nqm ( see the @xmath325 values in the last row of the table ) tremendously .
the additional inclusion of intermediate decuplet baryons deteriorates the fit , increasing the @xmath325 value . in this context
it is worth mentioning that the contribution of the octet - decuplet interference terms to the axial - vector coupling constant for neutron beta - decay @xmath330 obtained in our calculation is much smaller than in the cloudy bag model@xcite . adding vector mesons
again improves the fit to the experimental axial - vector coupling constants .
summarizing , taking into account both pseudoscalar and vector meson corrections does not lead to any significant deviations from the experimental data for semileptonic decays .
the quality of the fit is comparable to that of the original cabibbo model .
therefore , the experimental data for semileptonic decays does not contradict the meson cloud model of the nucleon , which one might naively expect .
we do not predict any essential difference between the @xmath10 cabibbo model and our model for any so far unmeasured transitions .
the meson radiative corrections were recently calculated in the framework of baryon chiral perturbation theory including pseudoscalar mesons with both intermediate baryon octet @xcite and decuplet @xcite components . in order to simplify the calculation
the baryon fields were treated there as heavy static fermions . although the details of the calculation in refs.@xcite differ from ours , similar conclusions have been drawn . in comparison
to refs.@xcite , where the experimental error bars where increased in calculating their @xmath325 values , our agreement with the data is much better .
if we increase the error bars as in refs.@xcite we would get @xmath331 of the order of 0.1 - 0.4 , compared to @xmath332 in refs.@xcite .
in section ii we discussed how the matrix elements of axial currents of the dressed nucleon are related to the corresponding matrix elements of the constituents ( intermediate baryons and mesons ) in the convolution approach . eq.([currents ] ) allows one to relate axial current matrix elements to quark polarizations .
then the ellis - jaffe sum rule can be expressed in terms of axial - vector matrix elements @xmath270 , @xmath333 and @xmath334 ( eq . [ currents ] ) @xmath335 in the same way the bjorken sum rule becomes @xmath336 dx = { 1\over3}\delta q_3\ ] ] which can be expressed in terms of the axial - vector coupling constant for neutron beta decay @xmath337 practical calculation of axial - vector matrix elements ( see eq .
[ axial ] ) requires additional assumptions about the axial properties of the bare particles .
we will assume @xmath338 for all non - strange bare particles , which is an extension of the original ellis - jaffe ansatz @xmath339 @xcite .
in analogy to the previous section we will consider two models : the @xmath3 model , with axial coupling constants given by eq.([su6coup ] ) and the @xmath10 model with @xmath265 and @xmath266 fitted to the semileptonic data .
in table 6 we present our results for the ellis - jaffe sum rule for the proton @xmath9 , the ellis - jaffe sum rule for the neutron @xmath340 , the bjorken sum rule @xmath341 and @xmath270 in both models with different fock states included : ( a ) tree level , ( b ) octet baryons and pseudoscalar mesons @xmath327 , ( c ) octet and decuplet baryons and pseudoscalar mesons ( @xmath342 ) and ( d ) the same with an additional inclusion of vector mesons ( @xmath343 ) . in the case of the @xmath3 model inclusion of the @xmath327 fock states
brings the bare @xmath3 value of @xmath344 for the proton ellis - jaffe sum rule almost half way down to the experimental result @xcite .
inclusion of the intermediate decuplet baryons enhances @xmath9 a little bit , and conversely the addition of vector mesons decreases @xmath9 slightly .
even with the inclusion of all fock states we are not able to describe the experimental emc result for @xmath9 .
it would be , however , too naive to expect that our model alone can account for all spin problems ; other effects , like the famous axial anomaly ( see section v ) , should be present as well .
currently it is , however , very difficult to give a quantitative estimate of the axial anomaly effect . on the purely phenomenological side one could try to map the @xmath1 dependence of the axial anomaly contribution to @xmath345 by ascribing all the deficiencies of the existing models with respect to the experimental data to the axial anomaly
. then one can write @xmath346 where @xmath347 is the contribution of the axial anomaly .
in ref.@xcite the unknown contribution was ascribed to the deficiency of the mit bag model .
such an estimate of the axial anomaly must rely on the model calculation of @xmath348 and @xmath349 . in the present paper we have concentrated on the role of mesonic corrections ,
therefore considered only very simple models of the bare nucleons ( baryons ) .
it has been shown very recently @xcite that by combining the mesonic effects calculated according to the present paper together with one version of the adelaide group bag model @xcite an impressive agreement with the emc @xcite and the slac @xcite data can be obtained at @xmath350 .
the meson cloud scenario leaves therefore much less phenomenological room for the axial anomaly contribution , at least at @xmath350 .
there is still some room left at @xmath351 , i.e. precisely in the region where the gluonic contribution is expected to play an important role @xcite .
the bjorken sum rule is independent of the anomaly .
here we get a good agreement with the classical value eq.([bjorken ] ) ( without qcd corrections ) when including only intermediate octet baryons with pseudoscalar mesons as well as in the full model . in this context
the similarity to the gottfried sum rule is worth noting . in order to compare @xmath341 with the experimental results of the slac and smc experiments ( see lower panel of table 6 ) , higher order perturbative corrections have to be included @xcite .
perturbative qcd corrections to the bjorken sum rule have been calculated up to @xmath352 @xcite @xmath353.\ ] ] in the case of the slac experiment also higher twist effects probably play an important role @xcite .
once higher - order perturbative corrections are taken into account we get a good agreement with the result of the smc analysis which include all available proton and deuteron data @xcite .
the classical @xmath3 model is attractive due to its simplicity . in practice , in order to describe different data it is necessary to break this symmetry and allow @xmath265 and @xmath266 to be fitted to the semileptonic decay data .
we also follow this line here .
then we get @xmath9 closer to the experimental emc result with the best result when only intermediate octet baryons and pseudoscalar mesons are included . in this case
@xmath340 becomes negative , which seems to be the case experimentally @xcite .
of course the bjorken sum rule is fulfilled here in all cases , which is due to the relation between @xmath341 and neutron beta decay constant @xmath354 ( eq . [ bjorken ] ) .
also the calculated ellis - jaffe sum rule has to be corrected for perturbative qcd effects .
it has been shown @xcite that the singlet ( @xmath355 ) and the non - singlet ( @xmath356 ) contributions to the integral of the ellis - jaffe sum rule get different corrections @xmath357 for example for @xmath358 @xmath359 calculated within our model ( see table 6 ) is reduced : @xmath360 @xmath361 and @xmath362 @xmath363 . here
we have taken @xmath364 @xcite .
although substantial , the pqcd corrections does not allow to bring our results down to the experimental @xmath359 . in all models considered ,
the singlet axial matrix element @xmath365 and @xmath273 are far from the one obtained from the emc analysis ( ! ) @xcite .
we have stressed the word analysis , as the emc result is biased by the assumption of the @xmath10 symmetry .
since our model violates @xmath10 symmetry , @xmath366 does not need to coincide with the value extracted by the emc . in our model
@xmath270 is reduced mainly due to the @xmath79 fock state admixture , which causes a partial depolarization of the spin as discussed in refs.@xcite .
the depolarization is caused by a cancellation of the spin - preserving and spin - flipping contributions to the @xmath79 fock state .
please note that our model predicts @xmath367 in contrast to the negative value found from the emc analysis . in the constituent quark models the nucleon does not contain any internal strangeness and as a consequence @xmath368 because @xmath369 is known to be an almost pure @xmath370 state .
this is also the case of the so - called @xmath3 related coupling constants @xcite which we have used throughout this paper .
however , @xmath369 could couple to the nucleon through @xmath371 very much the same as the @xmath21 meson couples to the nucleon via @xmath372 @xcite . due to its structure
the @xmath369 meson could be a good candidate to understand the problem of the `` missing '' polarized strangeness . to estimate a maximal possible effect of @xmath369 we have taken @xmath373 from the vector meson dominance model @xcite .
then we get extra @xmath374 , which is rather small and of opposite sign to the `` experimental '' @xmath273 .
in the light of recent experiments on the deep - inelastic lepton scattering by nucleons , the understanding of the nucleon structure has become a hot topic in particle and nuclear physics .
the recent observation of gottfried sum rule breaking strongly suggests a flavour asymmetry of light sea quarks in the nucleon .
the asymmetry occurs in a natural way within the meson cloud model .
the meson cloud model , as in any phenomenological model , requires specification of some external parameters which can not be calculated at the present time from a more microscopic theory . analysis of the model over the past few years has shown that the results are rather sensitive to these parameters . at the present stage
, it seems to be impossible to derive these parameters from the underlying qcd . in the present paper
we have tried to set limits on these parameters by applying the model to high - energy baryon production data .
the application of the mcm to these reactions is consistent with its application to the nucleon structure . in practical calculations
one has to make an ansatz for the vertex form factors .
it was argued @xcite that one should use form factors which guarantee certain symmetries of the longitudinal momentum distribution functions of virtual mesons and baryons . in the present work we followed this approach .
in contrast to form factors often used in traditional nuclear physics calculations , the symmetric form factors give a good description of the experimental data .
we find an universal cut - off parameter for all fock states involving octet baryons .
this leads to important differences in comparison to other calculations , where universality has been assumed for the @xmath18-dependent form factors .
after we had fixed the parameters of the vertex form factors in the hadronic sector , we have looked at the consequences in unpolarized and polarized deep - inelastic scattering and for the semileptonic decays .
we find a few interesting results .
the value of the gottfried sum rule obtained from our model ( @xmath375 ) is in good agreement with the experimental result obtained by nmc ( @xmath183 ) .
the meson cloud model predicts a @xmath376 asymmetry concentrated at rather small @xmath1 in impressive agreement with a recent global fit to the world data on the drell - yan and dis @xcite . since in our model the probability of the strange meson strange baryon fock components
is substantially reduced in comparison to the non - strange counterparts , we obtain a significant reduction of the strange sea quark distributions .
we find @xmath377 in contrast to the customary assumption @xmath378 .
the effect found is small and differs from naive expectations .
the mesonic corrections lead to renormalization of axial - vector current matrix elements .
large one - loop corrections , which explicitly violate @xmath10 symmetry , are in surprisingly good agreement with semileptonic data even with @xmath3 axial coupling constants for bare particles . we find that a significant fraction ( 20% ) of the nucleon spin is carried by the angular momentum of the mesonic cloud .
although the meson cloud model does not reproduce completely the emc result for the ellis - jaffe sum rule , the contribution of the meson cloud can not be neglected in the total balance of the proton spin , especially in the context of the success of the meson cloud in the explanation of the gottfried sum rule violation and @xmath379 asymmetry .
the presence of higher fock components seems , however , not sufficient to resolve the spin crisis . here
, probably other effects , to mention only the axial anomaly , play an important role . summarizing , in our view
the meson cloud model has many attractive features and can account for the description of many experimental data . here , we have discussed only some aspects .
it should be mentioned that this picture of the nucleon provides , in addition , a good description of nucleon electric polarizabilities @xcite and that it has a close connection to very successful models of low - energy hadron - hadron scattering @xcite .
finally we want to point out that the same model ( with the same set of parameters which has been determined from quite different phenomena ) turned out to be very successful @xcite in the description of the new na51 cern data for the dilepton production in the proton - proton and proton - deuteron collisions @xcite .
the model provides also a good description of the production of slow protons on the `` neutron '' ( extracted from the deuteron target @xcite ) in the charge - current neutrino and antineutrino dis reactions @xcite .
the model discussed here gives also unique predictions for the semi - inclusive dis @xmath380 which will be tested soon at hera @xcite .
such a measurement is being prepared by the zeus collaboration which is installing the forward neutron calorimeter @xcite . a pilot experiment on detecting neutrons from the beam - gas interactions
has already been performed and the result is in good agreement with the pion - exchange predictions @xcite . like any model , the meson cloud model requires further testing . in this respect
the planned fermilab experiment measuring the relative dilepton yield in proton - proton and proton - deuteron interactions @xcite seems to be very important and promising . very useful discussion with s.d .
bass , n.n .
nikolaev , a.w .
thomas and v.r .
zoller is gratefully acknowledged .
we wish to thank d. von harrach for drawing our attention to the problem of the @xmath233-@xmath381 asymmetry .
we are indebted to b. gibson and ch .
goodman for carefully reading the manuscript and very useful discussion .
this work was supported in part by the polish kbn grant 2 2409 9102 .
here we present the interaction langragians we employ in our calculations .
they are usually used in meson exchange models @xcite .
@xmath369 denotes a spin-@xmath382 field ( @xmath383 ) , @xmath78 a spin-@xmath384 field ( @xmath385 ) of rarita - schwinger form ; @xmath130 denotes a pseudoscalar field and @xmath386 a vector field ( @xmath387 ) : @xmath388 the anti - symmetric tensor @xmath389 here is defined as @xmath390 $ ] .
here we collect our results for the helicity dependent vertex functions @xmath391 .
@xmath33 here denotes the longitudinal momentum fraction of the baryon in the nucleon ; @xmath392 the transverse momentum of the baryon with respect to the nucleon momentum .
the contributions are listed according to particle helicities @xmath393 , with @xmath35 and @xmath39 being the baryon and meson helicities , respectively .
@xmath397=@xmath417= @xmath418 + @xmath408@xmath401 @xmath419 + @xmath420 + @xmath408@xmath421 @xmath422 + @xmath400@xmath423 @xmath424 + @xmath425 + @xmath400@xmath401 @xmath426 + @xmath427 + @xmath400@xmath421 @xmath428 + @xmath406=@xmath417 = @xmath432 + @xmath433@xmath401 @xmath434 + @xmath433@xmath421 @xmath435 + @xmath408@xmath423 @xmath436 \over y(1-y)m_b{\sqrt{\vphantom{i } y m_nm_b}}}$ ] + @xmath408@xmath401 @xmath437 \over(1-y)m_b{\sqrt{\vphantom{i } y m_nm_b}}}$ ] + @xmath408@xmath421 @xmath438 \over ( 1-y)m_b{\sqrt{\vphantom{i } y m_nm_b}}}$ ] + @xmath400@xmath423 @xmath439 + @xmath400@xmath401 @xmath440 + @xmath400@xmath421 @xmath441 + @xmath411@xmath423 @xmath442 + @xmath411@xmath401 @xmath443 + @xmath411@xmath421@xmath443 + lddd & exponential & monopole & dipole + @xmath444 & 1.08 & 1.9 & 2.2 + @xmath445 & 1.08 & 1.4 & 2.0 + @xmath446 & 0.204 & 0.210 & 0.219 + @xmath447 & 0.017 & 0.015 & 0.014 + @xmath448 & 0.062 & 0.075 & 0.060 + @xmath449&@xmath4500.038&@xmath4500.042&@xmath4500.034 + lddd model & @xmath40 & @xmath174&@xmath176 + @xmath451 & 1 . & 0 . &
0.333 + @xmath452 & 0.755 & @xmath4500.155 & 0.230 + @xmath453 & 0.697 & @xmath4500.119 & 0.254 + @xmath454 & 0.580 & @xmath4500.163 & 0.224 + .list of all measured semileptonic decays of baryons and the corresponding coupling constants in the @xmath10 model .
the data were taken from @xcite and @xcite ( for @xmath455 ) . [ cols="<,^,>,^,^,^,^",options="header " , ] | we present a complete set of formulas for longitudinal momentum distribution functions ( splitting functions ) of mesons in the nucleon .
it can be applied in the framework of convolution formalism to the deep - inelastic structure functions ( quark distributions ) of the nucleon viewed as a system composed of virtual mesons and baryons. pseudoscalar and vector mesons as well as octet and decuplet baryons are included .
in contrast to many approaches in the literature the present approach ensures charge and momentum conservation by the construction .
we present not only spin averaged splitting functions but also helicity dependent ones , which can be used to study the spin content of the nucleon .
the cut - off parameters of the underlying form factors for different vertices are determined from high - energy particle production data .
we find an universal cut - off parameter for processes involving octet baryons .
this information allows one to calculate the flavour and spin content of the nucleon .
the value of the gottfried sum rule obtained from our model ( @xmath0 ) nicely agrees with that obtained by the nmc .
in addition , we calculate the @xmath1-dependence of the @xmath2 asymmetry and get an impressive agreement with a recent fit of martin - stirling - roberts .
the calculated axial coupling constants for semileptonic decays of the octet baryons agree with the experimental data already with @xmath3 wave function for the bare nucleon . as a consequence
the bjorken sum rule is nicely reproduced .
although we get improvements for the ellis - jaffe sum rules for the proton and neutron in comparison to the naive quark model , the mcm is not sufficient to reproduce the experimental data . |
some of the most popular extensions of the standard model ( sm ) include additional higgs bosons .
when sufficiently heavy a neutral higgs boson @xmath13 and a pseudoscalar higgs boson @xmath14 may have a dominant @xmath15 final state .
the masses of the @xmath16 and @xmath11 need not be degenerate , and given a sufficient mass difference , the @xmath17 decay may compete with the @xmath15 decay . when the lighter higgs boson ( @xmath16 or @xmath11 ) is above the @xmath10 threshold then it can be expected to have a dominant decay to @xmath1 , leading to the final state @xmath12 .
this final state has a much smaller background than @xmath1 .
also the @xmath18 coupling depends only on the gauge coupling times a factor that is unity in the alignment limit of the two higgs doublet model ( 2hdm ) . here
we determine the first limits on the higgs cascade decay to the @xmath12 final state from the @xmath19 tev data .
in the case that the lighter of @xmath16 or @xmath11 is below the @xmath10 threshold , the limits from the resulting cascade decay to @xmath20 have been considered in @xcite .
first we calculate the model independent limits on the @xmath21 process from the atlas @xmath12 cross section measurement @xcite .
second we use this process to constrain the masses in the 2hdm in the alignment limit .
finally we consider the implications from the @xmath22 and @xmath7 decay modes when the pseudoscalar is consistent with the @xmath5 gev diphoton excess reported by atlas @xcite and cms @xcite . both atlas @xcite and cms @xcite have multileptons searches which measure the sm @xmath12 production cross section at @xmath19 tev and also at @xmath23 tev @xcite .
neither experiment sees a significant deviation from their sm expectations .
the total theoretical sm @xmath12 cross section is @xmath24 fb at @xmath25 tev . due to the small size of this cross section
, the @xmath19 tev data should provide useful limits on the process @xmath26 .
the strongest limit is given by a particular region of the atlas @xcite search that has the following selection criteria : * four anti-@xmath27 jets with @xmath28 gev , * three leptons with @xmath29 gev , * on @xmath30 selection within @xmath31 gev of the @xmath30 mass , * one @xmath32 tagged jet .
we use checkmate @xcite to recast this signal region .
we also include the jet - lepton overlap removal and an approximation to the lepton isolation criteria .
we then validated our checkmate analysis against a sample of sm @xmath12 events generated with madgraph @xcite and showered with pythia 6 @xcite .
we use a modified 2hdm feynrules model @xcite with herwig++ @xcite to generate our signal events . for the @xmath11 production cross section we use the sm heavy higgs gluon - gluon fusion cross section @xmath33 from @xcite .
the @xmath16 production cross section @xmath34 is larger than @xmath33 by a mass dependent scale factor that we deduce from the results in @xcite . in fig .
[ ratio ] we show our resulting upper limits on the product of the branching ratios @xmath35 and @xmath36 as a function of @xmath37 and @xmath38 .
these results can easily be scaled to account for different values of the production cross sections .
( left ) and @xmath39 ( right ) as a function of @xmath37 ( @xmath38 ) ( left and right respectively ) .
the masses of the heavier of @xmath40 are from bottom to top [ 500 ( black),550 ( blue),600 ( red),650 ( green),700 ( magenta ) and 750 ( orange ) ] gev .
, title="fig : " ] ( left ) and @xmath39 ( right ) as a function of @xmath37 ( @xmath38 ) ( left and right respectively ) .
the masses of the heavier of @xmath40 are from bottom to top [ 500 ( black),550 ( blue),600 ( red),650 ( green),700 ( magenta ) and 750 ( orange ) ] gev . , title="fig : " ] our signal has similar characteristics to the sm @xmath2 production and if anything our signal has a higher acceptance times efficiency .
thus we can compare our signal with the @xmath12 background for any given search and quickly estimate additional possible limits .
we have not found more stringent limits from other available lhc searches . the 2hdm is a useful benchmark model in which to discuss these decays . in order to ensure that our results are consistent with a sm higgs boson we work in the alignment limit
where @xmath41 , and we set @xmath42 @xcite as well . at tree level the @xmath43 decay is proportional to @xmath44 while the decays @xmath45 and @xmath46 are proportional to @xmath47 and are thus absent in the alignment limit .
the @xmath48 couplings do not vanish in this limit and @xmath1 becomes an important decay mode when @xmath49 .
as long as this decay dominates the other fermion decays our results are not very dependent on the type of the 2hdm .
we use the two higgs doublet model calculator @xcite to calculate the branching ratios for different values of @xmath50 and @xmath37 .
we satisfy constraints from stability of the potential , unitarity and oblique parameters , in particular by allowing @xmath51 to be freely varying .
perturbativity is satisfied for most of the mass pairs .
we consider both of the cases @xmath52 and @xmath53 for some low values of @xmath3 .
we set the charged higgs mass to be the same as either the heavier or the lighter of @xmath11 and @xmath16 . in the latter case
there is only a limit at @xmath54 since the competing decay mode @xmath55 is available with a somewhat larger branching ratio .
the excluded mass regions with @xmath56 are shown in fig .
[ ratio3 ] .
constraints on the large mass are as high as @xmath57 gev while constraints on the mass differences are as small as @xmath58 gev . ( @xmath6 ) .
the excluded region is enclosed by the solid curves for @xmath54 , the dot - dashed curves for @xmath59 , and the dotted curve for @xmath60 ( only in the @xmath6 case ) .
the dashed curves are the limits when @xmath61 is the lighter of @xmath40 with @xmath54 .
, title="fig : " ] ( @xmath6 ) .
the excluded region is enclosed by the solid curves for @xmath54 , the dot - dashed curves for @xmath59 , and the dotted curve for @xmath60 ( only in the @xmath6 case ) .
the dashed curves are the limits when @xmath61 is the lighter of @xmath40 with @xmath54 .
, title="fig : " ] for the @xmath62 decay mode we consider @xmath63 which has a large background from sm @xmath1 production .
we choose masses that are allowed from fig .
[ ratio3 ] , @xmath64 gev and @xmath65 gev .
we then scan the relevant searches available in checkmate .
we find that the signal contributes at most a relatively small number of events to some of the signal regions in @xcite .
we may use the atlas @xmath12 measurement at 13 tev @xcite to help estimate future 13 tev limits . assuming the future observations are consistent with the sm
, 10 fb@xmath66 of data will provide comparable limits to our @xmath19 tev analysis .
the improvement from @xmath67 fb@xmath66 of data is such that the dashed and dotted curves on the corresponding plots would move out to roughly the location of the solid curves in fig .
[ ratio3 ] .
the solid curves would extend out as far as @xmath68 and @xmath69 gev in the vertical and horizontal directions respectively .
the constraints on the mass differences would be as small as @xmath70 gev . during the lhc run 2 other types of searches
will further constrain heavy higgs masses .
heavy higgs boson production in association with top quarks , @xmath71 and @xmath72 , with @xmath73 is another process of interest for the alignment limit at low @xmath74 . with @xmath67 fb@xmath66 of data at 13 tev
this process can rule out a heavy higgs mass in the 500 to 700 gev range at @xmath54 ( see e.g. @xcite ) .
this process does not yield a corresponding constraint from the 8 tev data .
recently atlas @xcite and cms @xcite have observed an excess of events in the diphoton spectrum at 750 gev .
we consider a 2hdm - like scenario where the possible @xmath5 gev state is a pseudoscalar ( @xmath16 ) that can decay to @xmath76 and @xmath77 .
we again ignore the @xmath78 and @xmath79 decay modes that are not present at tree level in the alignment limit .
we assume that the 2hdm is supplemented by additional heavy states so that the @xmath8 and @xmath80 rates are sufficiently enhanced .
some choices that we make , such as setting @xmath81 , are motivated by a benchmark model that we outline below . within this set of assumptions
we consider constraints coming from our @xmath12 analysis and the cms @xmath1 resonance limits @xcite .
we consider @xmath82 gev and the @xmath83 partial width in the range @xmath8440 gev .
the upper end of this range is the natural width in the 2hdm at @xmath85 .
the partial widths to @xmath76 and @xmath77 at the masses that we consider are shown in table [ br ] , as are our limits on the cross section for the production and decay of @xmath16 ending with @xmath12 .
.the partial widths for @xmath6 ( @xmath86 ) and @xmath7 ( @xmath87 ) . [ cols="^,^,^,^,^,^",options="header " , ] in order to provide some background for the choices we have made we briefly describe a benchmark model . here
the new heavy states are a chiral quark doublet @xmath88 and a chiral lepton doublet @xmath89 with sm charges .
how such a extension of the sm is consistent with properties of the light higgs boson is discussed in @xcite
. a four higgs doublet model is thought to emerge as an effective description of the fluctuations around the chiral condensates of the four flavors @xmath90 .
the lighter of these bosons have a flavor content dominated by @xmath91 and @xmath92 and they form a set that is 2hdm - like . for these states
the @xmath93 form an isotriplet and the @xmath16 is an isosinglet relative to an approximate custodial symmetry in the @xmath94 sector .
this implies that @xmath95 .
it also implies that while @xmath16 has an enhanced production from gluon fusion , @xmath11 has a suppressed production .
we set @xmath96 gev and choose @xmath97 to avoid the @xmath98 decay . for the @xmath16 widths of most interest to the @xmath99 signal
we find @xmath100 gev and @xmath1010.004 gev , corresponding to @xmath102375 gev . from our results in fig .
[ masterplot1 ] we see that @xmath38 around 650 gev can be consistent with these widths , namely the lower right corner of the allowed region in the @xmath103 gev plot .
but then in addition we see that the top width @xmath104 needs to be reduced from its naive 2hdm value of @xmath105 gev . the @xmath106 coupling , along with other couplings of the bosons and their mass spectrum ,
is ultimately determined by the multi - higgs potential .
we find that parameters of this potential , partially already constrained by the light higgs properties , can be constrained further such that @xmath104 falls in the allowed range , e.g. @xmath107 gev .
coupling need not be so suppressed and it tends to enhance @xmath108 .
the @xmath109 coupling is suppressed , but still the @xmath10 decay of @xmath11 dominates .
] from the plot we also see how the @xmath10 resonance searches of cms and atlas differ on determining what this allowed range is . for larger @xmath38
, the @xmath110 gev plot shows that the model parameters would have to be more tuned to obtain a sufficiently small @xmath104 .
we turn now to the question of whether there can be direct limits due to the decay @xmath7 .
we consider the @xmath111 decay and find an approximate upper limit from @xcite using checkmate . at @xmath112 gev
we estimate that the cross section limit is @xmath113 pb for this final state , to be compared with the predicted value of @xmath114 fb when @xmath115 .
gev , @xmath116 . ]
we note that if the width of the @xmath117 is large ( for example @xmath118 gev ) , the branching ratio to @xmath119 where the @xmath117 is virtual can remain quite unsuppressed beyond @xmath120 gev . when this is the case then for such @xmath11 masses the corresponding plot of the type in fig .
[ masterplot1 ] can continue to look like the @xmath103 gev plot ( but without the limit from @xmath121 ) . in the benchmark model
another possible decay mode is @xmath122 , and when kinematically open it can be a dominant decay mode that results in multilepton final states .
we take the dominant heavy lepton decays to be @xmath123 ( rather than @xmath124 ) and @xmath125 ( @xmath126 would be even more striking ) .
then there is a quite striking final state , @xmath127 the cross section limits for this process from the atlas four or more lepton search @xcite are approximately @xmath12890 fb when @xmath129 gev and @xmath130200 gev . given the size of @xmath131 and any sizable @xmath122 branching ratio , these data imply that the @xmath132 and @xmath133 masses are such that this cascade decay is either kinematically forbidden or at least kinematically suppressed .
when the masses of additional higgs bosons are not degenerate , the @xmath134 final state offers a relatively model independent and clean probe of the heavy higgs sector . in the 2hdm ( at low @xmath74 and
close to the alignment limit ) we found that it eliminates quite a large region of the @xmath38@xmath37 plane .
in general any future search that is sensitive to the @xmath12 standard model background may also be sensitive to this signal .
assuming that a 750 gev mass @xmath16 is the source of the diphoton excess observed by atlas and cms @xcite , we considered the impact of a lighter @xmath11 that leads to the decay @xmath135 .
basically we find that the lighter the @xmath11 , the larger the @xmath99 width of the @xmath16 needs to be . in a benchmark model where the new fermions are @xmath136 and the @xmath99 width can not be arbitrarily large , we find that @xmath38 is constrained to a narrow range around @xmath137 gev . in this case
the model predictions nearly saturate the bounds both from the @xmath138 sm search and the @xmath10 resonance search .
the @xmath139 signal from the charged higgs decay may also be of interest .
thus this picture should be quite testable as new lhc data emerge .
99 b. coleppa , f. kling and s. su , jhep * 1409 * , 161 ( 2014 ) [ arxiv:1404.1922 [ hep - ph ] ] .
g. c. dorsch , s. j. huber , k. mimasu and j. m. no , phys .
* 113 * , 211802 ( 2014 ) [ arxiv:1405.5537 [ hep - ph ] ] . b. holdom and m. ratzlaff , phys .
d * 91 * , 035031 ( 2015 ) [ arxiv:1412.1513 [ hep - ph ] ] . g. c. dorsch , s. j. huber , k. mimasu and j. m. no , phys .
d * 93 * , 115033 ( 2016 ) , arxiv:1601.04545 [ hep - ph ]
. g. aad _ et al . _
( atlas collaboration ) , jhep * 1511 * , 172 ( 2015 ) [ arxiv:1509.05276 [ hep - ex ] ] .
the atlas collaboration , atlas - conf-2015 - 081 .
the atlas collaboration , atlas - conf-2016 - 018 .
the cms collaboration , cms - pas - exo-15 - 004 .
the cms collaboration , cms - pas - exo-16 - 018 .
v. khachatryan _ et al . _
( cms collaboration ) , eur .
j. c * 74 * , 3060 ( 2014 ) [ arxiv:1406.7830 [ hep - ex ] ] .
v. khachatryan _ et al . _
( cms collaboration ) , jhep * 1601 * , 096 ( 2016 ) [ arxiv:1510.01131 [ hep - ex ] ] .
the atlas collaboration , atlas - conf-2016 - 003 .
the cms collaboration , cms - pas - top-16 - 009 .
m. drees , h. dreiner , d. schmeier , j. tattersall and j. s. kim , comput .
commun . * 187 * , 227 ( 2015 ) [ arxiv:1312.2591 [ hep - ph ] ] .
j. de favereau _ et al .
_ [ delphes 3 collaboration ] , jhep * 1402 * , 057 ( 2014 ) [ arxiv:1307.6346 [ hep - ex ] ] .
m. cacciari , g. p. salam and g. soyez , eur .
j. c * 72 * , 1896 ( 2012 ) [ arxiv:1111.6097 [ hep - ph ] ] .
m. cacciari and g. p. salam , phys .
b * 641 * , 57 ( 2006 ) [ hep - ph/0512210 ] .
m. cacciari , g. p. salam and g. soyez , jhep * 0804 * , 063 ( 2008 ) [ arxiv:0802.1189 [ hep - ph ] ] . c. g. lester and d. j. summers , phys .
b * 463 * , 99 ( 1999 ) [ hep - ph/9906349 ] .
a. barr , c. lester and p. stephens , j. phys .
g * 29 * , 2343 ( 2003 ) [ hep - ph/0304226 ] .
y. bai , h. c. cheng , j. gallicchio and j. gu , jhep * 1207 * , 110 ( 2012 ) [ arxiv:1203.4813 [ hep - ph ] ] .
j. alwall _ et al .
_ , jhep * 1407 * , 079 ( 2014 ) [ arxiv:1405.0301 [ hep - ph ] ] . t. sjostrand , s. mrenna and p. z. skands , jhep * 0605 * , 026 ( 2006 ) [ hep - ph/0603175 ] .
a. alloul , n. d. christensen , c. degrande , c. duhr and b. fuks , comput .
commun . * 185 * , 2250 ( 2014 ) [ arxiv:1310.1921 [ hep - ph ] ] .
m. bahr , s. gieseke , m. a. gigg , d. grellscheid , k. hamilton , o. latunde - dada , s. platzer and p. richardson _ et al .
_ , eur .
j. c * 58 * , 639 ( 2008 ) [ arxiv:0803.0883 [ hep - ph ] ] .
s. heinemeyer _ et al . _
[ lhc higgs cross section working group collaboration ] , arxiv:1307.1347 [ hep - ph ] .
j. f. gunion , h. e. haber , g. l. kane and s. dawson , `` the higgs hunter s guide '' , ( brookhaven national laboratory , new york , 1989 ) , pp .
1404 ( 2000 ) . j. f. gunion and h. e. haber , phys .
d * 67 * , 075019 ( 2003 ) [ hep - ph/0207010 ] .
d. eriksson , j. rathsman and o. stal , comput .
commun . * 181 * , 189 ( 2010 ) [ arxiv:0902.0851 [ hep - ph ] ] .
g. aad _ et al .
_ ( atlas collaboration ) , eur .
j. c * 74 * , 2883 ( 2014 ) [ arxiv:1403.5222 [ hep - ex ] ] .
g. aad _ et al . _
( atlas collaboration ) , jhep * 1411 * , 118 ( 2014 ) [ arxiv:1407.0583 [ hep - ex ] ] .
g. aad _ et al . _
( atlas collaboration ) , eur .
j. c * 75 * , 510 ( 2015 ) [ arxiv:1506.08616 [ hep - ex ] ] . the atlas collaboration , atlas - conf-2013 - 037 .
n. craig , j. hajer , y. y. li , t. liu and h. zhang , arxiv:1605.08744 [ hep - ph ] .
s. chatrchyan _ et al . _
( cms collaboration ) , phys .
lett . * 111 * , 211804 ( 2013 ) [ phys .
lett . * 112 * , 119903 ( 2014 ) ] [ arxiv:1309.2030 [ hep - ex ] ] . g. aad _ et al . _
( atlas collaboration ) , jhep * 1508 * , 148 ( 2015 ) [ arxiv:1505.07018 [ hep - ex ] ] .
m. r. buckley , eur .
j. c * 76 * , 345 ( 2016 ) , arxiv:1601.04751 [ hep - ph ]
. v. khachatryan _ et al .
_ ( cms collaboration ) , phys .
b * 750 * , 494 ( 2015 ) [ arxiv:1506.02301 [ hep - ex ] ] .
g. aad _ et al . _
( atlas collaboration ) , phys .
d * 92 * , 032004 ( 2015 ) [ arxiv:1504.05511 [ hep - ex ] ] .
b. holdom , phys .
d * 90 * , 015004 ( 2014 ) [ arxiv:1404.6229 [ hep - ph ] ] .
the atlas collaboration , atlas - conf-2013 - 036 . | in models with multiple nondegenerate higgs bosons , the decay chain @xmath0 may have a partial width comparable to the @xmath1 decay mode .
we recast the atlas standard model @xmath2 measurement to put limits on the rate for this process .
limits are also set on the two higgs doublet model at low @xmath3 that are sensitive to a heavy higgs mass as high as @xmath4 gev .
we then discuss the @xmath5 gev diphoton excess in terms of a pseudoscalar that also has the decays @xmath6 and @xmath7 .
these decays strongly constrain the partial widths for @xmath8 and @xmath9 when combined with the @xmath10 resonance search limits . in a benchmark model the mass of @xmath11 should be close to 650 gev . |
the properties of the cluster galaxy luminosity function ( lf ) are still a matter of active debate . on the one hand considerable stress
has been laid on a single analytic expression for the lf , as determined by schechter ( 1976 ) , which it is argued holds true for all clusters and the field ( e.g. abell 1975 ) . on the other hand ,
much work has been concerned to establish differences in the lf as a function of cluster morphology .
indeed , a universal lf is hard to explain theoretically .
for example , according to the cannibalism line of argument ( e.g. hausman and ostriker 1978 ) , cd galaxies grow at the expense of other fainter galaxies - thus such a model predicts a deficit of bright galaxies in the most dynamically evolved clusters .
the brightness of the 1st ranked cluster galaxy is thus related to the dynamical state of the cluster and the lf is expected to correlate with bautz - morgan ( b - m ) type , as this measures the contrast of the 1st ranked galaxy compared to the other bright galaxies .
such a deficit in bright galaxies is also expected if tidal stripping in the outer halos of galaxies , due to interactions , is taking place ( richstone 1975 , 1976 ; merritt 1983 ; malumuth & richstone 1984 ) . only in the special case in which the luminosity function is determined during the formation period of the cluster , with little further dynamical evolution taking place , are no correlations of the lf with other cluster properties expected ( merritt 1984 , 1985 ) .
the most popular approach taken to the study of the lf is to fit schechter s ( 1976 ) parameterized form , given by @xmath10 where @xmath11 is a normalization factor , @xmath6 is the power - law slope at the faint end , and @xmath12 ( or equivalently @xmath13 ) is the characteristic luminosity ( magnitude ) .
schechter(1976 ) determined the best - fit parameter values to be @xmath14 ( j band ) and @xmath15 .
in subsequent studies there has been claim and counter claim regarding the properties of this function . in a study of 12 very rich clusters , dressler ( 1978 ) concluded that @xmath5 does vary significantly from cluster to cluster and that it is correlated with the absolute magnitude of the first - ranked galaxy , in the sense consistent with the trend expected in the cannibalism model .
lugger ( 1986 ) studied a sample of 9 abell clusters and concluded that while the mean values of @xmath13 and @xmath6 are in good agreement with those of dressler(1978 ) and schechter(1976 ) , there was no evidence for a correlation of @xmath13 with cluster morphology , central density or magnitude of the first - ranked galaxy .
the examination of a further 14 rich clusters by colless ( 1989 ) , draws very much the same conclusions as lugger(1986 ) - with no variation in the cluster lfs with b - m type or richness seen .
simulations suggest that variations of more than 0.4 mag in @xmath13 or 0.15 in @xmath6 are ruled out by these data .
colless did find a marginally significant result such that the lf of the high velocity dispersion clusters has fainter @xmath13 compared with the low velocity cases .
there are potential criticisms of these previous studies : first , as pointed out by lugger(1986 ) , the early studies such as that of dressler ( 1978 ) did not use a consistent cluster radius or limiting absolute magnitude in comparing different cluster lfs .
secondly , the best previous studies ( those of dressler 1978 , lugger 1986 and colless 1989 ) confine their selection to abell clusters , and typically only the richer examples of these systems . since there are potential systematic biases in the way in which abell clusters were selected ( for example , cd dominated clusters are much easier to detect by eye ) ,
these studies can not necessarily be held to be a true reflection of the properties of all clusters .
last , previous studies have relied on global number counts to correct for the non - cluster background present , and have used counts often taken in different passbands and from different surveys .
this can lead to systematic errors due to differences in the calibrations of the data the lf is derived from and the data the counts are taken from .
here we have re - examined the lf afresh using a sample of 46 clusters , selected automatically from digitized data . as a result
this study has certain advantages over previous work in the field .
the clusters are selected from photometrically calibrated digitized scans of photographic plates using an automatic search algorithm .
hence our data should be more uniform than the abell catalogue .
the sample covers a wide range in richness and many of the clusters have well determined velocity dispersions , allowing us to test for variations in the lf as a function of intrinsic cluster properties .
the sample of clusters used in this paper is drawn from the edinburgh / durham cluster catalogue ( hereafter edcc : lumsden et al .
1992 ) , which in its turn is derived from the edinburgh / durham southern galaxy catalogue ( hereafter edsgc : eg heydon - dumbleton et al . 1989 ) .
the edsgc consists of digitized cosmos scans of 60 uk schmidt telescope iii - aj survey plates centred on the south galactic pole , and is nominally 95% complete to @xmath16 with less than 10% contamination .
a cluster sample was derived from the galaxy data in the edsgc as outlined in lumsden et al .
there are over 700 potential clusters in the edcc , and this sample is the first published machine based cluster sample .
the objective nature of the construction of the edcc implies that it does not suffer from the systematic subjective biases that appear in the southern extension of the abell catalogue covering the same area ( abell , corwin & olowin , 1989 ) .
the edcc is complete within its selection criteria to a limiting magnitude for the tenth ranked cluster member of @xmath17 . for a smaller sub - sample of the edcc clusters
, we obtained redshifts with the multi - slit spectrograph efosc on the eso 3.6 m and with the multi - fibre spectrograph autofib at the 3.9 m anglo - australian telescope .
the nature of this sample is discussed in nichol et al .
( 1992 ) and guzzo et al .
( 1992 ) and the redshifts are published in collins et al .
approximately 10% of those clusters observed proved to have significant contamination from other groups along the line of sight at significantly different redshifts .
we have excluded those clusters with ` projection effects ' as defined in collins et al .
( 1995 ) from our sample .
we note that the remaining clusters do not have a well defined cut - off in cluster richness ( unlike the sample of nichol et al ) , since we have included all the clusters for which we actually obtained redshifts . from this dataset
, we have excluded those clusters where the cluster background radius ( as defined in section 3.3 ) overlaps the boundaries of the edsgc for ease of computation , and those clusters where their respective abell radii are entirely superposed , since there is then no well defined cluster magnitude distribution .
we also individually checked all the other clusters where the two clusters abell radii overlapped since the deblending procedure described in lumsden et al .
can occasionally cause the derived cluster centroid to move away from the ` true ' position : we removed those clusters where the catalogue positions differed from both the centre derived from a visual inspection of the plate and the centroid of those galaxies for which we have redshifts by a significant fraction ( @xmath1830% ) of the abell radius . for blended clusters that pass this test , we assign galaxy membership to the individual clusters using the method described in lumsden et al .
lastly , to be confident that we have removed all clusters that may have contaminating projection effects , we choose to use only those with a minimum of 4 cluster redshifts , and where the number of galaxies with redshifts outside of the cluster is less than the actual number of _ bona - fide _ cluster members .
the parameters of the remaining 46 clusters ( as taken from lumsden et al .
1992 and collins et al .
1995 ) are given in table 1 .
for all of the clusters , we have considered cluster members to be those galaxies within one abell radius ( where we have taken @xmath19mpc and we use the value for the hubble constant @xmath20 throughouth@xmath21 , and h@xmath22 is the hubble constant in units of km@xmath23s@xmath23mpc@xmath23 ] ) .
our sample is the largest single database of clusters which has been used to study the lf to date .
the edsgc was calibrated from photometric ccd sequences assuming that the plate and ccd magnitudes were linearly related ( heydon - dumbleton et al .
however , for bright objects whose profiles are either strongly peaked ( such as elliptical galaxies ) or stellar , there is a regime in which the measured cosmos magnitude is less than the true magnitude , caused by lack of dynamic range within the measuring machine itself .
this ` saturation ' occurs for elliptical galaxies at @xmath25 ( note that the plates themselves only saturate at much brighter magnitudes ) .
fainter than this magnitude , and for most late type galaxies , the original calibration is correct , so that previously published results from the edsgc are unchanged by these corrections .
however , since most clusters are dominated by elliptical galaxies at the bright end of the lf it is vital that we correct for this effect or else we will induce errors in the measured value of the ` break ' in the lf for the nearby clusters . we have calculated a correction for this effect , by comparing the catalogued edsgc magnitude against the observed magnitude for 451 galaxies present in the ccd sequences used to calibrate the edsgc initially ( heydon - dumbleton et al ) . since for faint magnitudes
the slope should be linear , but at bright magnitudes it should diverge from this , we used a quadratic fit to the data . from this
we derive a fit with the form ( see fig .
1 ) @xmath26 @xmath27 this fit should be reliable in the range @xmath28 . for most of this range
, the linear component present is actually the most important .
this slope has been commented on previously by heydon - dumbleton ( 1989 ) , and gives rise to the observed slope in the fit at fainter magnitudes , and is a result of the definition of measured photographic magnitude used in the edsgc . as can be seen
the effect is small ( the overall effect on the composite lf described in section 5.1 for example is 0.2 magnitudes ) .
however , for the purposes of the lf , we include all terms , so that our magnitude scale is more closely matched to the ccd based magnitude scale .
it should be noted that this correction is an average , and may not represent every cluster absolutely .
since the saturation effect is a function of the photographic density , anything else that modifies this ( such as vignetting , emulsion flaws , plate fogging or emulsion desensitisation near the plate corners ) will consequently affect the amount of saturation present .
we do not have sufficient ccd sequences to map these effects fully however , although we note that the overall photographic intensity
magnitude relation has been corrected for these effects ( see heydon - dumbleton 1989 ) .
these effects may in part give rise to the observed scatter in fig . 1 .
we expect the magnitude of any error caused by using this average calibration to be small , since the deviations from the fit are small .
as is common with studies of the cluster galaxy luminosity function we use the standard schechter function ( schechter 1976 ) as the analytical model to our data .
this function has the following form when expressed in terms of absolute magnitude : @xmath29 @xmath30 where @xmath31 . here
@xmath32 is the absolute magnitude , derived from the measured apparent magnitude @xmath33 using , @xmath34 where @xmath35 is the k - correction suitable for the @xmath36 passband for the mix of galaxies typically observed in low redshift clusters ( ellis 1983 ) , and @xmath37 is the effective extinction in this passband .
this latter quantity is derived using techniques described in nichol & collins ( 1993 ) for each cluster ( see also appendix a ) . in common with other studies of the luminosity function
we assume that @xmath38 in deriving the distance modulus used , and we have adopted @xmath20 for simplicity as well .
the lf normalisation @xmath39 is constrained using the actual observed number of cluster galaxies , and we carry out either a two parameter fit to @xmath5 ( the absolute magnitude of the ` break ' in the lf ) and @xmath6 ( the exponent of the power - law slope at fainter magnitudes ) , or a one parameter fit to @xmath5 with @xmath6 fixed where required .
there are too few galaxies per bin in the individual lfs to fit both @xmath5 and @xmath6 simultaneously .
since we wish to search for deviations from some ` mean ' lf , it is valuable to consider coadding all the lf data into an average or composite luminosity function , in a way that allows us to estimate both parameters simultaneously .
the easiest manner in which to do this is to consider the composite luminosity function defined by colless ( 1989 ) to be @xmath40 where @xmath41 is the ` richness ' of the @xmath42th cluster , @xmath43 is the number of clusters that actually contribute to the @xmath44th bin of the composite lf , @xmath45 is the background corrected number of galaxies in the @xmath44th bin of the @xmath42th cluster s lf and @xmath46 .
we use a different definition of @xmath41 to that of colless , since he used the total number of galaxies brighter than @xmath47 and we use the background corrected number of cluster galaxies between @xmath48 and @xmath49 .
this is because our sample contains more distant clusters on average than the colless sample , and we would otherwise discard a significant fraction of the data as being objects where @xmath47 was beyond our chosen completeness limit .
we also impose an upper limit in absolute magnitude when calculating the richness to avoid the problems with very bright galaxies discussed above . for typical values for the lf ( @xmath50 and @xmath15 ) , the relationship between our definition of richness and that of colless is @xmath41(colless)@xmath51(here ) .
similarly , the errors for the composite lf are calculated according to @xmath52^{1/2},\ ] ] where the @xmath53 are derived according to the error given in section 3.4 . in order to derive the lf
, we must be able to define the expected contribution from background galaxies , @xmath54 , accurately .
we have considered three approaches to this problem .
first , we used the counts in an annulus , centred on the cluster , with an inner radius of one degree and a width of one degree .
this fixed radius was adopted for simplicity using a fixed metric radius as opposed to a fixed angular radius gives essentially the same results .
this method has the advantage that the correction is strictly local , and hence the variation in field counts seen in different regions of the edsgc is correctly accounted for . for the other two methods we used the the global counts as derived from the whole of the edsgc , and the average of the counts from the annuli from the first method described above .
the first of these is similar to the procedure used in previous studies , but since our global counts actually come from the same data the lfs themselves are constructed from , we remove the possibility of errors arising from discrepancies in the calibration of the lf and the global counts .
the second has the potential advantage that it excludes the rich clusters from the counts and hence is a better measure of the ` true ' background . in fig .
2 we show the derived counts from the catalogue as a whole together with the average of the counts derived from the annuli around the clusters .
there is essentially little difference between these _ on average_. the difference that does exist is at the level expected from the presence of real large - scale structure .
there is little sign that excluding the rich clusters cores makes the background as calculated from the annuli lower than the global counts .
previous estimates of the background count for use in the lf have been derived from either separate ` eye - ball ' galaxy catalogues ( e.g. dressler 1978 , lugger 1986 ) or have been based on data from only very small areas ( colless 1989 ) .
we note in passing that the slope of the number counts that we find is larger than that assumed in previous studies of the lf .
this is however consistent with maddox et al .
( 1990 ) , the only other measure of the counts over significantly large areas of the sky that is derived using a measuring machine .
the fit given by colless ( 1989 ) appears to overcount the true background by @xmath55% .
lastly , we note that the derived number counts shown in fig .
2 drop off more rapidly than @xmath56 at @xmath57 .
this disagrees with other observations ( eg the deep ccd counts presented in metcalfe et al .
1991 ) , and indicates that the overall completeness limit of the edsgc is @xmath58 .
however , inspection of the number counts on individual schmidt plates also shows that this limit varies from plate to plate .
we have allowed for a possible variable completeness with plate in our work by examining the number counts on each plate separately and assigning completeness limits relative to the 7 deepest fields in the survey . from this
we find that the best plates are complete to @xmath59 and the worst to @xmath60 .
we do not however use data fainter than @xmath16 from any field . for those fields where the adopted completeness limit is brighter than this
, we do not use data fainter than that limit .
in calculating the background correction required when using the global number counts we have interpolated the actual measured data , rather than fitting a single power law .
the latter is a poor representation to the actual data , since the true number counts ` turn - over ' near @xmath61 ( eg metcalfe et al . ) .
this regime is important for our work since for clusters with @xmath62 , @xmath16 is equivalent to an absolute magnitude of @xmath63 .
we also have to derive suitable error estimates for these background counts .
there are two sources of error , which are present in all the methods given above .
the first is simply the poisson uncertainty in the counts and can be represented simply as @xmath64 .
the second is systematic and represents the presence of large - scale clustering within the sample . for the global correction method , although we would assume that the slope of the background number counts is constant over the survey region , the normalisation of these counts changes with position .
colless ( 1989 ) and lugger ( 1986 ) follow geller & beers ( 1982 ) in assuming that the background is only known to 50% . for the background derived from the local annuli we can estimate the error rather more accurately however .
if we take all of these annuli , and find the scatter about the mean of the number counts , then we can derive the systematic error for this method .
since all the annuli have the same area any number density deviations are purely due to the presence of large - scale structure .
the derived variance is also shown on figure 2 .
comparison of the expected contribution from poisson error and the total error observed indicates that the mean field - to - field variance is 25% for @xmath65 .
the largest excursions seen are @xmath6650% over the same magnitude range .
therefore , using the 50% variation in what follows , although conservative , does account for all the observed variations , whereas using the 25% mean variation is a better representation of the data as a whole but will underestimate the error in the background for a few clusters .
we demonstrate in section 5.1 the effect of using this smaller error on the field counts compared to the value favoured by geller & beers .
we note that dressler ( 1978 ) also found that a mean variation of 25% was a good fit to his data , and this result is compatible with the known angular clustering in the edsgc ( collins , nichol & lumsden 1992 ) .
we have considered only a @xmath9 squared fitting method in deriving the parameters @xmath5 and @xmath6 .
if we let @xmath67 be the difference between the total counts in a bin and the expected background in the bin ( however the latter is calculated ) , so that @xmath68 then we seek to minimise the function @xmath69 where , @xmath70 and @xmath71^{1/2}.\ ] ] throughout we use a bin width , @xmath72 , of 0.5 magnitudes .
the presence of the second derivative in the form of @xmath73 takes account of the finite bin width used , since the number counts change extremely rapidly ( schechter 1976 ) .
the first term in the error represents the poissonian error in the total observed counts .
therefore , @xmath74 .
the second term must take account of both poissonian error and systematic errors in the background as noted in section 3.3 .
this has the form @xmath75 , where @xmath76 ( except where explicitly noted that @xmath77 ) .
although for many of the clusters data exists with @xmath32 brighter than @xmath1 and fainter than @xmath0 , there are good reasons for not using such data in deriving values of @xmath5 and @xmath6 .
first , for the fainter galaxies , the correction for the background becomes more difficult since the background fraction per bin rises steadily .
for any given cluster we impose two limits on this : that any data with @xmath78 be discarded , for reasons outlined in section 3.3 , and that any bin for which the number of background galaxies exceeds the cluster galaxies also be discarded . in practice
, the former limit essentially ensures the latter for faint galaxies in any event . for bright galaxies
there are also problems we wish to avoid .
it is desirable , as has been noted previously by , e.g. lugger , to exclude the brightest cluster galaxy when deriving the lf .
we automatically do this since we discard all galaxies brighter than @xmath79 ( typically of order five galaxies in the richer clusters ) .
there are more specific reasons for doing this in our case as well .
first , the saturation correction is poorly defined brighter than @xmath80 , and hence the magnitudes of the galaxies brighter than this are likely to have larger errors . at @xmath81 ,
this magnitude corresponds to @xmath82 , and at @xmath83 @xmath84 .
therefore , it is clear that galaxies brighter than @xmath79 should not be used because their photometry is suspect .
secondly , cosmos data is deblended where object mergers occur ( cf the discussion in lumsden et al .
1992 ) , but the accuracy of the magnitudes of the deblended data is less reliable than for single isolated objects . the brightest galaxies in clusters are almost always blended with fainter satellite systems in the edsgc data , so again the photometry of these objects is not reliable . discarding the brightest bins of the lf guards against problems of this kind .
it is possible that similar problems may also be present in previous derivations of the lf , since the work described in the introduction also relied extensively on photographic plate material .
in any event we recommend that results which rely on accurate magnitudes for the brightest galaxies in any lf derived from photographic plates should be treated with caution .
we have also placed constraints on the values of @xmath5 and @xmath6 that the fitting routine can accept , to prevent any possibility of the fitting process ` running away . '
these limits are that @xmath5 should lie between @xmath0 and @xmath85 and that @xmath6 should lie between 0 and @xmath86 . therefore quoted values for the fits that are near
these limits are indicative of a very shallow minimum in the fit , and hence that the reliability of the fit is likely to be low ( whatever the quoted formal probability that the fit is good ) .
lastly , we note that we sum bins to ensure that there are @xmath87 galaxies per bin in the data being fitted , as is standard for @xmath9 minimisation techniques .
before we can consider the actual data , we need some estimate of the reliability of our fitting procedure and the true error on these fits . to achieve this , we conducted extensive monte - carlo simulations .
in particular , we carried out three sets of simulations : first , to estimate the effect that errors in the assumed background had on the fit parameters as a function of redshift ; secondly , to derive a measure of the true error distribution for a one parameter fit to a model lf as a function of richness , for @xmath41 typical of the single cluster values we found ; lastly , to derive the error distribution for the two parameter fits to the composite sets .
these latter two will then allow us to estimate the true likelihood that observed departures from the composite lf are in fact truly significant .
in carrying out these simulations , we have assumed that @xmath88 and @xmath15 for the model of the lf ( in accordance with the results derived in section 5.1 ) . in order to check the effect of the background subtraction alone
, we initially assumed that the form of the lf was known perfectly , and only the background was allowed to vary .
we used the global background as derived above , and rescaled it at random so that the variance was 25% as was found to be the case in practise in section 3.3 .
thus after normal background correction , the random lf would have either too many or too few background galaxies subtracted . by deriving fit parameters for many such datasets
we can derive the average error induced in those parameters as a function of cluster richness and redshift .
in general , for @xmath89 , the effect of errors in the background on the fit parameters is negligible , even for poor clusters . at higher redshifts , @xmath5 can be measured to an accuracy of 0.05 magnitudes if @xmath90 ( ie the average of the measured @xmath5 from the simulations is different by this amount from the assumed input @xmath5 ) . at @xmath91 and @xmath92 , the scatter in the derived value of @xmath5 becomes larger ( @xmath93 magnitudes ) .
the values derived here should be the dominant source of error when the richness is very large ( since then the lf is known almost perfectly ) .
since the composite luminosity functions given all have effective values of the richness that are much greater than 100 , it is clear that the empirically derived errors for the effect of different background corrections given in section 5.1 are in good agreement with the results of these monte - carlo simulations .
more generally , we need to map out true estimates of the likely scatter in the fits to @xmath5 when the lf was drawn at random from the same model lf used above . in this case , since we draw individual galaxies at random from the parent lf , we also include random fluctuations in the lf itself .
for low to moderate richness clusters , it is these fluctuations that dominate the scatter in the fits to @xmath5 rather than the background correction method .
although the resultant values are therefore independent of redshift , we have adopted @xmath83 as typical of our clusters , in order to include a suitable background contribution .
this background is derived in the same manner as described above .
table 2 gives the limits outside of which 90% , 31.5% , 10% , 5% and 1% of the data lie .
one of the most noticeable features of these simulations is that the distribution of fitted values for @xmath5 is skew , with a strong tail towards brighter magnitudes .
an example of this is shown in figure 3 , where we present the distribution of fits to @xmath5 for two richnesses , @xmath94 and @xmath95 .
this trend is due to the restricted range of absolute magnitudes that are actually used to fit against .
since we set m@xmath96 in accordance with the result we find in section 5.1 , there are only two bins in the fitted lf that actually sample the break . at low richness values
there are often insufficient counts per bin to truly define this break , and hence , since there is no constraint from the actual data , the fitted value of @xmath5 can take quite large negative values without changing the @xmath9 of the fit significantly . in practice ,
the limit for @xmath5 at the bright end is set by our constraint that @xmath97 .
clearly , for the richer sample shown in figure 3 the trend towards a strongly skewed distribution is much less evident as expected .
this skewness also shows in the mean @xmath5 derived from the simulations , with this value being the input value of @xmath98 at large richness but diverging to @xmath99 at @xmath94 .
this same behaviour is also demonstrated ( in a slightly different fashion ) by the simulations of colless ( 1989 ) , where he shows that the probability that a two - sample @xmath9 test can reject the possibility that two clusters are drawn from the same underlying population is also skew with respect to @xmath5 .
we can further confirm the reason for this skewed distribution , by looking at the distribution of @xmath5 when the absolute magnitude range being fitted is @xmath100 to @xmath0 .
the results for @xmath101 are given in table 2 and also shown in figure 3 .
it is easy to see the effect that including the brighter galaxies has on fixing the location of the break in the lf .
the slight skewness left in the distribution now reflects the fact that the total number of galaxies with @xmath102 in these low richness systems is rather few .
the derived error bounds including these bright galaxies are most similar to those for a much higher value of @xmath41 ( @xmath103 ) when fitted over the normal range in @xmath32 .
the key result of these simulations therefore is that it is difficult to rule against overly bright values of @xmath5 as being from the same distribution as the composite .
however , those clusters for which @xmath5 is rather fainter than the composite are strongly discriminated against .
we note that the results of these simulations can also be applied to the data of colless ( 1989 ) , and that the simulations of dressler ( 1978 ) are not applicable to the case of fitting over a restricted magnitude range ( or rather that the dressler results would be closer to those achieved in practice but for a much larger true richness ) .
it is worth also considering how sensitive these results are to our assumed value of @xmath6 .
we tested this using @xmath104 for @xmath105 , typical of our sample .
we found that the changes are small ( @xmath106 in the faint limit on @xmath5 , and @xmath107 in the faint bright limit ) and always act to push these limits _
brighter_. since the cluster lfs we can exclude are those which have faint @xmath5 , it is clear that adopting a lower value of @xmath6 will only make this difference larger .
the formal average of our composites using all methods of calculating them give @xmath108 ( table 3 ) .
this implies that the probability that any lf agrees with the derived composite as given by the simulations is slightly conservative . since we have carried out two parameter fits in section 5.1 and 5.3 , we also need to know the likely joint distribution of @xmath6 and @xmath5 for the case of rich systems alone ( ie composite lfs ) .
we therefore carried out simulations as described above but in this case applied a two parameter fit to the lf . in this instance
the best way to present the data is graphically , and the results of the simulations are shown in figure 4 .
the contour levels plotted correspond to the same probabilities given for the one parameter case in table 2 , except that we exclude the contour that contains 10% of the data in this instance .
the simulations have been binned into 0.05 in @xmath6 and 0.1 in @xmath5 for presentation .
it is clear from figure 4 that @xmath5 and @xmath6 are highly correlated . in
what follows we will present formal error estimates on each fit , derived by determining the value(s ) at which @xmath109 . as will be seen
, there is a considerable difference between these formal errors and both the error distributions given in table 2 for the one parameter fits and the distribution shown in figure 4 for the two parameter fits .
this difference largely stems from the correlation between @xmath5 and @xmath6 , since the errors in these are also therefore clearly correlated .
these formal errors are included for completeness only ( since they do provide a measure of ` goodness - of - fit ' ) .
we will only use the monte - carlo simulations and two - sample @xmath9 tests when assessing the true deviations that may be present between any given lf and a universal composite .
the first lf we consider is the composite of all the clusters which have @xmath110 .
there are 23 clusters in this sample , and the richness of the composite was found to be @xmath111 ( @xmath112 ) .
this allows us to compare our results with previous samples which generally dealt with rich abell clusters . for this composite
we have considered the effect of using the different background correction methods outlined in section 3.3 , as well as for the different assumed errors in that background subtraction .
the results of the fitting process for these different background corrections are given in table 3 and shown in figure 5 .
the first three rows were derived by fitting both @xmath5 and @xmath6 to this sample over the magnitude range ( @xmath113 ) with an assumed field - to - field error in the background counts of 50% .
the first entry gives the results for the local correction method , the second for the global correction using all of the counts and the third the global correction but using only the average of the annuli around the clusters .
clearly the two global background corrections we have used give essentially the same result . directly comparing the data derived from the local and global background correction methods using a 2-sample @xmath9 test also shows that these composites are in good agreement for the absolute magnitude range given above ( @xmath11499% probability they are same ) .
moreover , since we have not renormalised the lfs before comparing using this test , this also indicates that differences that do exist between the background subtraction methods for individual clusters ( as will be shown in section 5.2 ) tend to average out when combined in the composite .
the last three rows in table 3 give the equivalent results when the field - to - field background error is assumed to be 25% as noted in section 3.3 .
this change therefore gives the same results as before within the derived errors on @xmath5 and @xmath6 for the global correction method , and only very slightly different in the formal sense for the local correction method .
this shows that the dominant error in deriving estimates of @xmath5 or @xmath6 is actually the poissonian contribution ( cf the discussion after equation 4 ) , and that the variation due to assuming a different error on the background term alone is small .
indeed , the magnitude of these changes found here is negligible compared to the expected measurement errors as shown in section 4 .
henceforth , we consider only the 50% field - to - field variation , since it adequately represents all of the clusters in the present sample .
lastly , it is worth noting the effect that changing the magnitude range over which the fits are made has on these conclusions . including the brighter galaxies pushes @xmath5 brighter . including the fainter galaxies makes a slight difference to the slope but little else .
it should be noted however that only one cluster actually contributes for @xmath32 fainter than @xmath0 , whereas all 23 clusters contribute at @xmath32 brighter than @xmath1 .
the result derived when the fit is over the range @xmath115 , using the local background correction , is also given in table 3 .
the low probability that the fit is good is due to the fact that a schechter function is now a poor fit to the data at both bright _ and _ faint magnitudes .
we note that a similar ( though larger ) increase in brightness of @xmath5 was found by lugger when she considered mean luminosity functions with the brightest cluster galaxies included .
however , we repeat our caution of section 2.2 that data on such bright galaxies ( @xmath116 ) can not be relied upon absolutely because of problems of saturation and image blending ( cf the discussion in colless 1989 ) .
it is also not clear therefore whether or not the greater variance from the mean seen in the brightest bins , as compared to those in the more restricted absolute magnitude range over which we normally determine the lf , is actually real or merely an artifact of the cosmos measurements .
we also fitted each cluster lf separately to the schechter function , but this time holding @xmath6 constant ( since there are insufficient points in many of the clusters to enable more than a one parameter fit to be made successfully ) .
we adopted @xmath15 as an appropriate compromise , based both on our own results and those previously published .
table 4 shows the results of these tests for all 46 clusters that were found to have positive values of @xmath41 and figure 6 shows the actual data for those with @xmath117 .
we quote results derived using both the global and local background correction methods .
we also tested for differences arising from the choice of background subtraction method by comparing the lf distributions directly .
we used a two - sample @xmath9 test to check for differences between the two sets of lfs , without rescaling the lfs for the possibility that the two background subtraction methods may have resulted in a different overall normalisation .
we have allowed for the fact that the adopted magnitude completion limit may be brighter than @xmath118 , so that we compare the lfs in the range @xmath1 to the brighter of @xmath0 or the plate completion limit .
these limits are given in table 4 .
the results of the two - sample @xmath9 test are given in column 10 of table 4 .
as can be seen the agreement is good , and even for the worst case ( e748 ) , we can not rule out the possibility that the two samples represent the same underlying distribution .
we also compared the single cluster lfs with a suitably normalised model of the composite ( assuming the average values of @xmath5 and @xmath6 found above ) , using a simple @xmath9 test ( given in table 4 ) , and with the results of the monte - carlo simulations from table 2 .
it can be seen that there are several clusters where the lf derived using global background subtraction is a poor match to the model , but the lf derived using the local background is a much better fit .
we take this as evidence for the fact that assuming an overall global background will not be a good match to every cluster .
generally the two tests give similar results , though we note that the two - sample @xmath9 test is considerably more conservative at ruling out individual lfs , as expected .
however , all those lfs for which the two - sample @xmath9 test gives a probability below 20% , are also found to be a poor match to the composite from the simulations .
there are two exceptions to the corollary however : both e438 and e462 have unexceptional probabilities from the two - sample @xmath9 test , but are excluded at the 5% level from the simulations .
if we include only those clusters where the derived @xmath5 lies outside the 1% confidence level from the simulations ( as taken from table 2 ) for one or other of the background subtraction methods , then only clusters e400 , e429 and e524 show strong evidence for an lf that differs from the composite .
all three of these clusters also disagree at the 5% level when considering both background subtraction methods . at the 5% level
, there are several additional clusters that disagree .
leaving aside those where only one of the background subtraction methods gives divergent results ( e499 , e726 ) , we are left with e438 , e462 and e742 .
all of these , except e400 and e429 , have a value of @xmath119 , so we may reasonably exclude these as being different on the grounds that accurate background subtraction is crucial for these clusters , and , as found in section 4 , errors of 0.5 magnitudes in the fitted value of @xmath5 are easily accounted for by an inappropriate background correction .
this leaves only e400 and e429 .
it can be seen from table 4 that even the direct @xmath9 comparison of their lfs with the composite might lead to them being thought discrepant .
the lfs of these clusters are apparently normal ( see figure 6 ) . however , before concluding that these clusters do show a significant difference , we first consider the possibility that there may be some problem with the data .
examination of the data in collins et al .
( 1995 ) shows that all of these clusters have good reliable redshifts , with the positions of the galaxies for which redshifts were obtained agreeing well with the cluster centres given in table 1 .
further , we note that for the cases of e400 and e429 , other clusters appear on the same photographic plate and do not have discrepant @xmath5 , hence the photometry of these systems should be good .
therefore , we find that there are at least two clusters that do not have the same lf as the derived composite .
this is in keeping with earlier studies by dressler ( 1978 ) , lugger ( 1986 ) and colless ( 1989 ) who found that @xmath1810% of their clusters similarly did not agree with the mean schechter lf they derived .
the major advantage of our sample over previous studies is that it is large enough that we can test the dependence of the lf on intrinsic cluster properties .
in particular , we will consider the possible variations with cluster richness ( as defined by @xmath41 ) , bautz - morgan class , which we have taken from the abell , corwin & olowin catalogue , and velocity dispersion ( taken from collins et al .
1995 ) .
for each case we split our total sample into two , and used the local background correction method only . when comparing different richness samples we used the cuts @xmath120 and @xmath110 ( which is of course just the composite lf derived previously ) . for bautz - morgan class ,
we considered only those of type i and iii , and counted intermediate types ( e.g. type i - ii ) as type ii which we did not use . in this way we sought to maximise the differences between the clusters to test for any effect on the lf .
we split the sample into two in velocity dispersion using a cut at 700kms@xmath23 .
unfortunately , most of the very high velocity dispersion clusters ( those with @xmath121kms@xmath23 ) are also the most distant , making it difficult to constrain @xmath6 , and leaving few bins in the data to compare the 2 sets using a @xmath9 test .
we therefore were forced to use a lower velocity cut - off for our high velocity dispersion sample than colless ( 1989 ) did , so do not sample the same part of parameter space as he did .
since we do not have well defined velocity dispersions for all clusters , and those which are not in the abell , corwin & olowin catalogue do not have bautz - morgan classes , the two subsamples do not comprise all of the observed clusters .
the number of clusters comprising each subsample is given in table 5 .
in addition , for the velocity dispersion and bautz - morgan samples we have included two richness cuts , at @xmath110 and @xmath122 .
this provides some indication of the stability of the result on the clusters making up the actual subsamples .
we give both one and two parameter fits to the schechter lf in table 5 . for the one parameter
fits we again used @xmath123 .
the results from this analysis are also shown in figure 7 .
we have also compared the lfs for these subsamples against the composite derived in section 5.1 , and against each other , using a two - sample @xmath9 test , as well as comparing the subsamples directly with the results of our simulations ( figure 4 ) .
the results of the two - sample @xmath9 test are also given in table 5 .
since @xmath39 can vary between these different sets of clusters , we rescaled the binned data before the comparison , so that the derived @xmath124 values from the composites agreed
. there are noticeable differences between the richness cuts for both the bautz - morgan class iii clusters and the high velocity dispersion clusters , with the richer subsamples having a notably poorer fit to the standard schechter lf , and being discrepant compared to the composite .
however , only one cluster contributes to the faintest bin in the richer subsample , and only one other is added to that when the poorer limit is considered .
if we exclude this bin , then very similar fits are found for both richness limits and the discrepancy is removed .
it is clear from this that variations in the lf due to the last bin alone are not significant .
we do not consider these two subsamples further .
once these discrepant sets are removed , it is clear that neither bautz - morgan class shows any evidence for deviation from a universal lf , nor do they differ from each other , according to any of the tests applied . the same is true for the split according to richness alone , where the fits to both samples are very similar within the derived errors .
the actual ` goodness ' of the fit for the lowest richness sample is poor , but this most likely reflects difficulties in background correction at these low richness values as noted previously . for the @xmath122 velocity dispersion samples , there is weak evidence that the two samples differ from each other ( significance @xmath125% ) . when compared with the composite , neither sample shows any evidence for difference when using the two - sample @xmath9 test .
however , when compared with the monte - carlo simulations , we find that the higher velocity dispersion clusters are clearly unlikely to have been drawn from the same basic distribution as the composite ( at much better than the 1% level ) .
we therefore find that there may be evidence for some dependence of the lf on velocity dispersion .
it is worth considering the effect of including the brighter galaxies as well in this analysis .
as is readily evident from figure 7 , there is little difference between the rich and poor samples across the entire range of @xmath32 , and the differences for the high and low velocity dispersion samples is largely restricted to the fainter galaxies .
considering the @xmath122 bautz - morgan samples however , there are clearly fewer bright galaxies in the bautz - morgan class i sample than in the bautz - morgan class iii sample .
this is not a normalisation effect since the two samples have essentially the same lfs at fainter magnitudes .
again however , we caution against overinterpretation of this difference .
lastly , we note that all of the velocity dispersion subsets , and the low richness subset can formally be excluded as being drawn from a model of the composite lf with @xmath6 fixed at @xmath7 on the basis of the one parameter fits
. small deviations in both @xmath5 and @xmath6 mean that the one parameter fits are distinct from the composite derived in section 5.1 .
it is clearly dangerous to compare such subsamples ( with their relatively small intrinsic errors ) using such methods .
first , we compare our derived estimates for the composite lf , for all clusters with @xmath110 , with those derived previously .
colless ( 1989 ) gives the best fits from previous data in his table 2 , scaled into the colour and value of @xmath126 that we are using .
excluding colless own data this shows that the quoted values of @xmath5 and @xmath6 are @xmath127 and @xmath128 respectively , with typical errors in @xmath5 of 0.5 , when the brightest cluster galaxies are excluded from the fit .
colless himself finds @xmath129 and @xmath130 for a two parameter fit similar to ours .
these values are in good agreement within the errors with those derived by us .
it is clear that the luminosity functions of moderately rich clusters , when considered across this relatively narrow range in absolute magnitude , are indeed very similar on average .
in addition we can compare our one parameter fits for 3 clusters with those given by colless ( 1989 ) .
edcc clusters 394 , 400 and 124 correspond to his clusters c02 , c03 and c52 . from table 4
we can see that when a comparison is made of similar fitting procedures ( ie , @xmath9 fitting and using a global background correction ) , the agreement is excellent .
colless quotes values for @xmath5 of @xmath131 , @xmath132 and @xmath133 for c02 , c03 and c52 respectively .
although the original photographic plates used in both studies are the same , a different measuring machine was used ( the apm ) , with consequent differences in the reduction from measuring machine magnitudes to final derived astronomical magnitudes . therefore , this agreement for the specific clusters that are common to the two studies is highly encouraging .
it is worth making some specific comments on these clusters however .
e394 was found by colless to be discrepant with both the fit by a schechter lf and with a comparison with his composite lf .
we note that table 4 shows that whilst the global background correction gives a moderate probability for both the fit and the comparison with the composite , the local background correction method gives better agreement .
this difference shows in the direct comparison of the magnitude distributions for the two background correction methods ( table 4 ) .
this is clearly an example where global background correction gives rise to these deviations .
we note that e394 is one of the poorest clusters in both our sample and that of colless .
colless found that e124 was marginally discrepant when compared against the composite lf .
we find no evidence for any difference .
lastly , e400 , which we found in section 5.2 to be potentially significantly different from the composite , colless finds to be a reasonable match .
however , since this cluster gives a difference for both background subtraction methods , we find our result to be reliable .
we also note that the smaller errors derived for by us for both e400 and e124 allow us to make more definitive statements than colless can for these clusters .
finally we note that unlike colless , we find that all the subsets of the data we took are well fit by a one parameter schechter lf , though they are not necessarily in agreement with the fit to the composite .
there have been few detailed n - body simulations of the formation of clusters that have suitable dynamic range to map out the whole of the expected lf as a function of the cosmological model assumed .
for example , van kampen ( 1995 ) makes simple predictions for the lf based on extensive simulations , but is led to the conclusion that his modelling can not correctly account for all of the merging that takes place within the cluster .
his fit to the bright end of the luminosity function ( which we might expect to be easier to model ) is similar to the one observed .
there have been studies that lead to simpler predictions as to the expected form of the lf according to the dominant processes that occur during cluster formation .
the mode of cannabalism for the growth of cd galaxies as suggested by hausman & ostriker ( 1978 ) predicts that there should be differences between the luminosity function when broken down by bautz - morgan class .
we see no real evidence for such a difference when we consider the range @xmath134 .
there is an indication for such a trend in the galaxies brighter than @xmath79 , in the sense that the bautz - morgan class i sample has fewer bright galaxies .
however , better photometry of the bright galaxies would be required to confirm this result .
merritt ( 1983,1984 ) argued that most of the properties of clusters are essentially ` frozen in ' when they form and that merging and tidal stripping processes are ineffective in rich cluster cores .
this leads to the prediction that the lf should appear to be much the same from cluster to cluster . however , hst imaging data of moderate redshift clusters ( e.g. dressler et al .
1994 ) tends to rule against such a model , since it provides convincing evidence that mergers are important at least in the early stages of cluster evolution .
one possibility is that the bulk of the change in the lf between clusters occurs only amongst the very faintest or brightest galaxies , the regime to which our data is insensitive .
for example , it is known that some clusters show an upturn in the lf at faint magnitudes ( e.g. driver et al .
1994 , bernstein et al . 1995 ) , and there has always been marginal evidence for greater deviations in the lf when the brightest cluster galaxies are included ( e.g. lugger 1986 and the comments in section 5.1 ) , though we again caution against over - interpretation of results on the brightest cluster members .
however , our data clearly shows that in the region of the ` break ' in the lf , the cluster - to - cluster variance is extremely small .
we have considered the cluster galaxy luminosity function as derived from data in the edinburgh - durham southern galaxy catalogue and the edinburgh - milano redshift survey .
our dataset is the largest multi - redshift cluster sample to be considered to date for this purpose . from this
we find : 1 .
our data shows that there is strong evidence that _ most _ clusters have very similar lfs when both the brightest and faintest galaxies are discarded .
the average of the composite lfs derived using the 3 background correction methods in section 5.1 give @xmath135 and @xmath4 , when the lf is determined in the range @xmath134 .
we find a very good agreement between our results and those derived by colless ( 1989 ) from similar input data , and with the mean of the data presented by dressler ( 1978 ) and lugger ( 1986 ) .
we have tested the stability of the derived schechter luminosity function parameters to the type of background correction made .
we find excellent agreement between a method based on a global background correction and that based on a strictly local correction .
however , the largest systematic error remaining in our derivation of the best fit parameters for the composite lfs is the background correction , with an uncertainty of up to 0.05 in @xmath5 and 0.05 in @xmath6 possible .
we also tested the universality of the luminosity function , by comparing individual cluster lfs with the composite function derived in section 5.1 , and by testing each cluster lf separately against a schechter function ( section 5.2 ) . from this
we find that at most @xmath1810% of the clusters may have lfs that are significantly different from the composite .
we also broke our sample into sub - samples defined by richness , bautz - morgan class and velocity dispersion ( section 5.3 ) . from these tests we found weak evidence that the high and low velocity dispersion samples have different composite lfs , and that the higher velocity dispersion clusters may have a different lf to the global composite .
this agrees with the marginal detection of a difference between similar samples by colless ( 1989 ) .
we also found evidence for differences between the bautz - morgan class samples , but only for the brightest galaxies ( @xmath102 ) that we otherwise excluded for reasons of photometric reliability ( section 3.4 ) .
the trend evident is in line with the predictions of the cannabalism model of hausman & ostriker ( 1978 ) , but we caution against over - interpretation of this result without more reliable photometry for the bright galaxies .
there is no evidence for any convincing difference between the other subsets of the data . our results , and our simulations ,
do however show that it is important to consider more than one method of testing the difference between subsets of the data . on the basis of a one parameter fit alone
we would have concluded that there were significant differences between the low and high velocity dispersion samples and the composite , and between the low and high richness clusters .
the other tests show no evidence for any difference however , with the exception of the high velocity dispersion clusters where the two parameter fits , and monte - carlo simulations show weak evidence for a difference from the composite .
it is likely that to make any further progress on the nature of the cluster galaxy luminosity function we will need to overcome the limitations of data such as ours .
first , there is clearly a need to derive cluster membership free of any correction for backgrounds , by obtaining sufficient redshifts that we can actually map the galaxy distribution in three dimensions around the cluster .
secondly , better photometry is required to tie down the bright and faint ends of the lf . given these it should be possible to not only discriminate between differences in the data broken down by richness , velocity dispersion or whatever else is desired , but also to test for the differences that our data suggest may exist in the lf when comparing the very brightest and very faintest galaxies .
sll would like to acknowledge the support of an serc postdoctoral fellowship during the early part of his work on the edcc .
cac acknowledges the support of pparc through an advanced fellowship .
we would like to thank the referee , steve maddox , for his comments .
= 0.5 in = 1 abell , g.o . , 1975 , in sandage , a. , sandage , m. , kristian , j. , eds , stars and stellar systems ix : galaxies and the universe , univeristy of chicago , chicago , p. 601
.the full cluster sample .
the edcc number refers to the identification in lumsden et al .
the field is the original schmidt j survey field number .
n@xmath136 is the actual number of redshifts obtained for that cluster , and n@xmath137 is the number of those redshifts adjudged to lie in the cluster .
bm class is the bautz - morgan class , where 1 represents class i , 3 represents class iii , 2 represents any other defined class and 0 implies that the cluster concerned is not in the abell et al .
( 1989 ) catalogue and therefore has not been classified .
the velocity dispersion is taken directly from collins et al .
[ cols=">,>,>,>,>,>,>,>,<,>,>,^ , > , > , > " , ] * figure captions * + * figure 1 : * the relationship between the catalogued edsgc @xmath36 magnitude and the derived @xmath36 magnitude from the ccd sequences .
the solid line represents the best fitting quadratic given in the text .
the dashed line is given as a guide to the eye to show the deviation between the ccd magnitude scale and the edsgc scale .
* figure 2 : * the number counts from the edsgc derived from the whole catalogue ( the solid line ) , and from the average of all the annuli used for background correction of the lf ( @xmath138 ) .
the difference between the two never exceeds 10% .
the error bars shown represent the variance of the counts derived from the annuli .
* figure 3 : * derived values of @xmath5 from one parameter fits to model schechter luminosity functions for clusters with ( a ) @xmath94 and ( b ) @xmath95 , when the fit is over the range @xmath134 .
the model lf has @xmath88 and @xmath15 .
when this range is extended to @xmath139 , the distribution of @xmath5 for @xmath101 is given in ( c ) .
this is clearly similar to the higher richness distribution when fitted over the narrower range in absolute magnitude .
* figure 4 : * contours of the derived distribution in @xmath5 and @xmath6 for 10000 simulations of a two parameter fit to the standard schechter function .
the initial model lf has @xmath88 and @xmath15 .
the contours contain 10% , 68.5% , 90% , 95% and 99% of the data . the initial richness given to the model lf
is @xmath140 , close to the measured value for most of our subsamples in section 5.3 .
this is therefore the correct error map for comparison with the observed values of @xmath5 and @xmath6 from those composite sub - samples .
* figure 5 : * measured luminosity functions for ( a ) the local background correction method , ( b ) global background correction and ( c ) global correction but using the average of the annuli from ( a ) .
the data are fitted to a schechter function in the range @xmath141 .
the best fitting schechter function is shown as a solid line . only the data derived assuming a 50% background variance is shown . * figure 6 : * measured luminosity functions for all those cluster with measured richness values , @xmath41 , of 15 or more using the local background correction method .
the best one parameter fit to the schechter function is shown by a solid line .
the fit is over the range @xmath142 .
* figure 7 : * luminosity functions for ( a ) bautz - morgan class 1 clusters with @xmath110 , ( b ) bautz - morgan class 3 clusters with @xmath110 , ( c ) bautz - morgan class 1 clusters with @xmath122 , ( d ) bautz - morgan class 3 clusters with @xmath122 , ( e ) @xmath110 clusters , ( f ) @xmath120 clusters , ( g ) clusters with velocity dispersion @xmath143kms@xmath23 and @xmath110 , ( h ) clusters with velocity dispersion @xmath144kms@xmath23 and @xmath110 , ( i ) clusters with velocity dispersion @xmath143kms@xmath23 and @xmath122 , ( j ) clusters with velocity dispersion @xmath144kms@xmath23 and @xmath122 .
the best fitting two parameter fits to a schechter function over the range @xmath134 are shown as solid lines .
the best fitting one parameter fits ( with @xmath15 ) are shown as dashed lines .
as can be seen , there is often a marked difference between the best fitting one and two parameter fits . | we have re - examined the nature of the cluster galaxy luminosity function using the data from the edinburgh - durham southern galaxy catalogue and the edinburgh - milano redshift survey .
we derive a best fit luminosity function over the range @xmath0 to @xmath1 in @xmath2 , for a composite sample of 22 of the richer clusters that has @xmath3 and @xmath4 . the dominant error in these values results from the choice of background subtraction method . from extensive simulations we can show that when the lf is fitted over this narrow range , it is difficult to discriminate against bright values of @xmath5 in the single cluster fits , but that faint values provide a strong test of the universality of the luminosity function .
we find that all the individual cluster data are well fit by a schechter function with @xmath6 fixed at @xmath7 , and that @xmath810% of these have fitted values of @xmath5 that disagree from the average at the 99% confidence level .
we further show that fitting only a single parameter schechter function to composite subsets of the data can give erroneous results for the derived @xmath5 , as might be expected from the known tight correlation between @xmath5 and @xmath6 . by considering two parameter fits
, the results of monte - carlo simulations and direct two - sample @xmath9 tests we conclude that there is only weak evidence for differences between the data when broken down into subsets based on physical properties ( bautz - morgan class , richness , velocity dispersion ) : from our simulations , only the evidence for a difference between subsets based on velocity dispersion may in fact be significant .
however , we find no evidence at all that a schechter function is not a good model for the intrinsic cluster luminosity function over this absolute magnitude range . models that invoke strong evolution of galaxy luminosity of _ all _ galaxies within clusters are inconsistent with our results . [ firstpage ]
galaxies : clusters of galaxies : luminosity function |
it is well - known that loop effects in the minimal supersymmetric extension of the standard model ( mssm ) can become large when the ratio @xmath6 of the two higgs vacuum expectation values is sizeable .
an important example is the higgs coupling to bottom quarks , @xmath7 @xcite . in the limit of heavy superpartner particle masses ,
the relevant couplings are @xcite @xmath8 this includes a coupling @xmath9 to the `` wrong '' higgs doublet , which is forbidden by supersymmetry at tree level .
the quantity @xmath9 is an ordinary one - loop effect and not enhanced in any way .
however , due to the large ratio of vacuum expectation values ( vevs ) the bottom quark mass @xmath10 receives large @xmath0-enhanced corrections to its tree - level value @xmath11 @xcite .
so does the @xmath7 coupling , when @xmath12 is eliminated in favour of @xmath13 as is usually done .
in general , while the loop effect is not large by itself , it provides large corrections to couplings or observables , which are suppressed by the small vacuum expectation value @xmath14 or which vanish at tree level .
this type of relative enhancement is clearly a one - loop effect , which does not repeat itself in higher loop orders .
here we investigate other effects of this type from the higgs sector , where loop effects can lead to a mixing of the @xmath15 and @xmath16 fields through two- and four - point interactions , which indirectly modify the yukawa couplings to the physical higgs bosons .
we also address the issue of renormalization conventions for @xmath0 and their stability against radiative corrections , as investigated in a numerical approach in @xcite .
the indirect effects from the higgs sector have not been considered explicitly in the literature , although our results are implicit in one - loop calculations of mssm higgs decay rates @xcite . however , these calculations are usually available only in numerical form . having simple analytic expressions at hand is often useful to acquire a better understanding of parameter dependences , especially for the mssm with its large parameter space .
higgs - sector effects from two - point interactions have been considered recently in @xcite , while their consequences for @xmath17 mixing are discussed in @xcite .
the theoretical issues can be exposed most clearly in the decoupling limit , assuming that all superpartner particles are heavy , of order @xmath18 , while the standard model ( sm ) particles together with all five physical higgs bosons are light , at most @xmath19 , the electroweak scale . in practice
, the decoupling limit may be a good approximation for @xmath18 as small as a few hundred gev .
the assumed hierarchy implies a certain amount of fine - tuning , since the soft susy breaking parameters and the @xmath20 parameter must be @xmath21 , while @xmath22 are @xmath19 to make the higgs bosons light .
the higgs - mixing parameter @xmath23 must even be much smaller than @xmath24 to achieve large @xmath25 .
the first fine - tuning is the well - known `` little hierarchy problem '' , here built in by hand by the assumption @xmath26 , but the second is an extra fine - tuning of @xmath23 specific to the large-@xmath0 scenario , which has received less attention up to now . to simplify the discussion we consider only the third family of fermions and
assume cp conservation in the mssm .
including the ckm matrix and cp - violating mssm phases would be straightforward . under these assumptions
we match the mssm to an effective general two - higgs doublet model ( 2hdm ) , and determine the physical higgs fields , the yukawa couplings , and trilinear higgs couplings from the one - loop corrected effective 2hdm .
the higgs and yukawa terms of the mssm lagrangian are given by @xmath27 the conventions are as follows : the antisymmetric @xmath28 matrix @xmath29 is defined with @xmath30 .
the hypercharge u(1)@xmath31 and su(2 ) couplings are @xmath32 and @xmath33 , respectively .
the quark fields are @xmath34 , @xmath35 , @xmath36 , the leptons @xmath37 , @xmath38 , where @xmath39 denotes a weyl spinor index . for the higgs doublets we also use the notation @xmath40 , @xmath41 , such that @xmath42 and @xmath43 both have hypercharge @xmath44 .
our mssm conventions follow @xcite except for @xmath45_{\rm rosiek}$ ] , @xmath46_{\rm rosiek}$ ] , @xmath47_{\rm rosiek}$ ] and @xmath48_{\rm rosiek}$ ] . in this way our definitions of the yukawa couplings and @xmath49 parameters are consistent with the spa convention @xcite .
the relevant parameters of the mssm higgs sector are @xmath32 , @xmath33 , @xmath50 , @xmath23 , @xmath51 and the two vevs @xmath52 .
it appears to be contrary to the spirit of interpreting the mssm as the high - scale theory and the 2hdm as the low - energy effective theory to introduce the vevs as mssm parameters , since the vevs should be found _ a posteriori _ by minimizing the 2hdm effective potential .
however , since we wish to use standard mssm renormalization conventions , and moreover introduce @xmath0 as one of the mssm parameters , we substitute @xmath53 in the lagrangian ( [ mssm ] ) . in this way , after further fixing the renormalization convention , the parameter @xmath6 is well - defined , contrary to the general 2hdm , where @xmath0 is unphysical , since it can assume any value by performing a rotation in the @xmath54 fields . in the mssm such rotations are excluded by holomorphy of the superpotential .
an equivalent set of parameters consists of @xmath55 related to the previous set by @xmath56 with @xmath57 .
the quark masses are given by @xmath58 at tree level .
for vanishing tadpole couplings @xmath59 , @xmath60 is the mass of the cp - odd neutral higgs boson .
the parameters so far are bare parameters .
we introduce renormalized parameters by substituting @xmath61 , where now @xmath62 ( @xmath63 ) is the bare ( renormalized ) parameter and @xmath64 the corresponding counterterm .
the renormalization conventions are specified by imposing conditions on the parameters of the second set : @xmath65 are defined by the conventional on - shell renormalization conditions for the electric charge and the particle pole masses , the tadpole couplings @xmath59 by the condition that the tadpoles ( one - point functions ) vanish .
the renormalization condition for @xmath0 and the yukawa couplings , and the definition of @xmath66 beyond tree level will be discussed later .
the counterterms @xmath64 for the parameters of the first set are then defined by requiring that ( [ parameterrel ] ) holds for the bare as well as the renormalized parameters .
specifically , for @xmath0 this implies @xmath67_0 = \frac{[v_u]_0}{[v_d]_0 } = \frac{v_u+\delta v_u}{v_d+\delta v_d } = \frac{v_u}{v_d } + \frac{v_u}{v_d } \left(\frac{\delta v_u}{v_u } - \frac{\delta v_d}{v_d}\right ) + \ldots.\ ] ] employing @xmath6 , the one - loop relation @xmath68 follows .
similarly , one obtains a set of equations relating the shifts of all parameters of the second set to those of the first . we do not need to discuss the renormalization of superpartner masses and the @xmath49 and @xmath20 terms , since these occur only in loops . in a one - loop calculation , it is sufficient to know these parameters at tree - level .
other sm parameters like the strong coupling or ckm matrix play no role here .
for the subsequent discussion of @xmath0-enhanced terms field renormalization in the mssm can also be ignored .
thus , we shall not introduce renormalization factors @xmath69 for the su(2)-doublet fields .
the divergent parts of the vev counterterms @xmath70 are identical to those in @xmath69 , since the theory in the broken phase can be rendered finite by the same counterterms as the unbroken theory before the shift ( [ introducevevs ] ) .
thus , when we ignore field renormalization , we should formally consider @xmath70 and consequently @xmath71 as finite quantities , which indeed they are in the large-@xmath25 approximation , whose values are nevertheless fixed by the imposed renormalization conventions .
we integrate out the heavy susy particles and expand the mssm effective action in local operators .
the result is a general 2hdm , where the higgs - yukawa terms read @xmath72 \label{twohdm}\\ & & + y_b \phi_1\epsilon q b + y_t \phi_2^\dagger q t + y_\tau \phi_1\epsilon
l \tau - \delta \tilde y_b \phi_2\epsilon q b + \delta \tilde y_t \phi_1^\dagger q t - \delta \tilde y_\tau \phi_2\epsilon
l \tau + { \rm h.c.}\quad \nonumber\end{aligned}\ ] ] neglecting operators with dimension higher than four .
the last line equals @xmath73 in terms of the @xmath74 fields , so the yukawa couplings take their standard form . since we assumed cp conservation in the mssm , the matrices @xmath75 , @xmath76 ( @xmath77 ) and the couplings @xmath78 are real .
note that the kinetic terms of the higgs fields are not yet canonical , because we match 1pi green functions .
the kinetic terms of the fermions can be disregarded , since their renormalization is an ordinary one - loop correction to an unsuppressed tree term . at tree - level @xmath79 in the following the one - loop corrections to the couplings that vanish or are small at tree - level
are of particular interest : @xmath80 , @xmath81 , @xmath82 , and @xmath83 . the result of the calculation is summarized in the next section . and @xmath84.,scaledwidth=90.0% ]
all matching calculations are performed in the su(2)@xmath85u(1)@xmath86 symmetric `` phase '' of the mssm .
the couplings @xmath84 and @xmath87 follow from an expansion of the off - diagonal @xmath88 two - point function in its external momentum . the relevant one - loop diagrams with heavy superpartner particles in the loop are shown in figure [ fig1 ] .
for the kinetic - mixing term we find @xmath89 for the off - diagonal mass term , we write @xmath90_{\rm heavy}$ ] and obtain for @xmath91 the expression ( [ z121loop ] ) with the replacement @xmath92 , where @xmath93 and @xmath94 .
the kinetic - mixing correction @xmath84 is finite , as it should be , while the correction to @xmath87 is ultraviolet divergent and requires regularization ( here dimensional regularization , space - time dimension @xmath95 and renormalization scale @xmath96 ) .
the @xmath97 pole is cancelled by the counterterm contribution @xmath98_{\rm heavy}$ ] to @xmath87 .
the counterterm contribution is the difference between the mssm and the 2hdm counterterm , which survives the matching relation , and cancels the singularity in the heavy - particle loops . in the following
we shall drop the label `` heavy '' in the counterterm contributions .
we remark that we have kept complex conjugation on the @xmath20 and @xmath49 parameters , where required , if they were complex .
however , by our assumption of cp - conservation , @xmath20 and @xmath99 are real . here and below the masses that enter the arguments of the loop functions
are the gaugino and sfermion mass parameters in the soft supersymmetry - breaking part of the mssm lagrangian and not the physical mass parameters .
the physical masses receive additional contributions at tree level proportional to the vevs @xmath52 from electroweak symmetry breaking , which are of higher order in the @xmath100 expansion . for the four - point higgs couplings @xmath101 , which vanish at tree level
, we obtain the ultraviolet - finite one - loop expressions @xmath102 & & - \ , ( \mu^ * a_\tau^*)^2 y_\tau^2 k_2(m_{\tilde l},m_{\tilde \tau_r } ) \nonumber\\[0.15 cm ] & & + \ , ( \mu^*)^2 m_1 ^ 2 \ , g_1 ^ 4 \ , k_2(|\mu|,m_1 ) + \ , 3 ( \mu^*)^2 m_2 ^ 2 \ , g_2 ^ 4 \ , k_2(|\mu|,m_2 ) \nonumber\\[0.1 cm ] & & + \,2 ( \mu^*)^2 m_1 m_2 \ , g_1 ^ 2 g_2 ^ 2 \ , k_3(|\mu|,m_1,m_2 ) \ , , \\[0.4 cm ] \delta \lambda_6 & = & - \ , 3\mu^ * a_b^ * \,y_b^3\left\ { j_2(m_{\tilde q},m_{\tilde b_r } ) + j_2(m_{\tilde b_r},m_{\tilde q } ) \right\ } \nonumber\\ & & - \ , \mu^ * a_\tau^ * \,y_\tau^3\left\ { j_2(m_{\tilde l},m_{\tilde \tau_r } ) + j_2(m_{\tilde \tau_r},m_{\tilde l } ) \right\ } \nonumber\\ & & + \,3\mu^ * y_t a_t^ * \left(-\frac{1}{4}\right ) \left\ { \left[g_2 ^ 2-\frac{g_1 ^ 2}{3}\right ] j_2(m_{\tilde q},m_{\tilde t_r } ) + \left[\frac{4g_1 ^ 2}{3}\right ] j_2(m_{\tilde t_r},m_{\tilde q } ) \right\ } \nonumber\\ & & + \,3\mu^ * y_b a_b^ * \left(-\frac{1}{4}\right ) \left\ { \left[-g_2 ^ 2-\frac{g_1 ^ 2}{3}\right ] j_2(m_{\tilde q},m_{\tilde b_r } ) + \left[-\frac{2 g_1 ^ 2}{3}\right ] j_2(m_{\tilde b_r},m_{\tilde q } ) \right\ } \nonumber\\ & & + \,\mu^ * y_\tau a_\tau^ * \left(-\frac{1}{4}\right ) \left\ { \left[-g_2 ^ 2+g_1 ^ 2\right ] j_2(m_{\tilde l},m_{\tilde \tau_r } ) + \left[-2 g_1 ^ 2\right ] j_2(m_{\tilde \tau_r},m_{\tilde l } ) \right\ } \nonumber\\[0.1 cm ] & & -\ , 3 \mu^ * \,|\mu|^2 a_t^ * y_t^3 k_2(m_{\tilde q},m_{\tilde t_r } ) - 3 \mu^ * \,|a_b|^2 a_b^ * y_b k_2(m_{\tilde q},m_{\tilde b_r } ) \nonumber\\[0.2 cm ] & & - \ , \mu^ * \,|a_\tau|^2 a_\tau^ * y_\tau k_2(m_{\tilde l},m_{\tilde \tau_r } ) \nonumber\\[0.2 cm ] & & -\ , 3 \mu^ * \,m_2 g_2 ^ 4 \,l_2(|\mu|,m_2 ) - \mu^ * \,m_1 g_1 ^ 4 \,l_2(|\mu|,m_1 ) \nonumber\\[0.2 cm ] & & - \,\mu^ * \,(m_2+m_1 ) \,g_2 ^ 2 g_1 ^ 2 \,l_3(|\mu|,m_2,m_1 ) \ , , \\[0.4 cm ] \delta \lambda_7 & = & -\ , 3\mu^ * y_t a_t^ * \,y_t^2\left\ { j_2(m_{\tilde q},m_{\tilde t_r } ) + j_2(m_{\tilde t_r},m_{\tilde q } ) \right\ } \nonumber\\ & & -\,3\mu^ * y_t a_t^ * \left(-\frac{1}{4}\right ) \left\ { \left[g_2 ^ 2-\frac{g_1 ^ 2}{3}\right ] j_2(m_{\tilde q},m_{\tilde t_r } ) + \left[\frac{4g_1 ^ 2}{3}\right ] j_2(m_{\tilde t_r},m_{\tilde q } ) \right\ } \nonumber\\ & & -\,3\mu^ * y_b a_b^ * \left(-\frac{1}{4}\right ) \left\ { \left[-g_2 ^ 2-\frac{g_1 ^ 2}{3}\right ] j_2(m_{\tilde q},m_{\tilde b_r } ) + \left[-\frac{2 g_1 ^ 2}{3}\right ] j_2(m_{\tilde b_r},m_{\tilde q } ) \right\ } \nonumber\\ & & -\,\mu^ * y_\tau a_\tau^ * \left(-\frac{1}{4}\right ) \left\ { \left[-g_2 ^ 2+g_1 ^ 2\right ] j_2(m_{\tilde l},m_{\tilde \tau_r } ) + \left[-2 g_1 ^ 2\right ] j_2(m_{\tilde \tau_r},m_{\tilde l } ) \right\ } \nonumber\\[0.1 cm ] & & -\ , 3 \mu^ * \,|a_t|^2 a_t^ * y_t k_2(m_{\tilde q},m_{\tilde t_r } ) - 3 \mu^ * \,|y_b \mu^*|^2 a_b^ * y_b k_2(m_{\tilde q},m_{\tilde b_r } ) \nonumber\\[0.2 cm ] & & - \ , \mu^
* \,|y_\tau \mu^*|^2 a_\tau^ * y_\tau k_2(m_{\tilde l},m_{\tilde \tau_r } ) \nonumber\\[0.2 cm ] & & -\ , 3 \mu^ * \,m_2 g_2 ^ 4 \,l_2(|\mu|,m_2 ) - \mu^ * \,m_1 g_1 ^ 4 \,l_2(|\mu|,m_1 ) \nonumber\\[0.2 cm ] & & - \,\mu^ * \,(m_2+m_1 ) \,g_2 ^ 2 g_1 ^ 2 \,l_3(|\mu|,m_2,m_1 ) \ , .
\label{eq : lambda7}\end{aligned}\ ] ] the diagrams relevant for the computation of @xmath103 are shown in figure [ fig2 ] .
the loop functions read @xmath104 with @xmath94 , @xmath105 .
.,scaledwidth=95.0% ] the loop - induced yukawa couplings to the `` wrong '' higgs field are given by @xmath106 \delta \tilde y_{t } & = & \frac{8}{3 } \mu^ * y_t m_3 \
, g_3 ^ 2 j_3(m_3,m_{\tilde q},m_{\tilde t_r } ) + \ , \mu^ * y_b y_t a_b^ * j_3(|\mu|,m_{\tilde q},m_{\tilde b_r } ) \nonumber\\ & & - \ , \frac{3}{2 } \mu^ * y_t m_2 \ , g_2 ^ 2 j_3(|\mu|,m_2,m_{\tilde q } ) + \ , \frac{2}{9 } \mu^ * y_t m_1 \ , g_1 ^ 2 j_3(m_1,m_{\tilde q},m_{\tilde t_r } ) \nonumber\\ & & - \ , \frac{2}{3 } \mu^ * y_t m_1 \ , g_1 ^ 2 j_3(|\mu|,m_1,m_{\tilde t_r } ) + \ , \frac{1}{6 } \mu^ * y_t m_1 \ , g_1 ^ 2 j_3(|\mu|,m_1,m_{\tilde q } ) \ , , \label{eq : deltayt } \\[0.2 cm ] \delta \tilde y_{\tau } & = & \frac{3}{2 } \mu^ * y_\tau m_2 \
, g_2 ^ 2 j_3(|\mu|,m_2,m_{\tilde l } ) - \ , \mu^ * y_\tau m_1 \
, g_1 ^ 2 j_3(m_1,m_{\tilde l},m_{\tilde \tau_r } ) \nonumber\\ & & + \ ,
\mu^ * y_\tau m_1 \ , g_1 ^ 2 j_3(|\mu|,m_1,m_{\tilde \tau_r } ) - \ , \frac{1}{2 } \mu^ * y_\tau m_1 \
, g_1 ^ 2 j_3(|\mu|,m_1,m_{\tilde l } ) \ , , \label{eq : deltaytau}\end{aligned}\ ] ] where @xmath107 the expression for @xmath108 agrees with the well - known result for the @xmath0-enhanced correction to the bottom quark mass @xcite , when the bino contribution and terms of order @xmath109 are neglected .
our full results for @xmath110 and @xmath111 agree with the expressions for the correspondent quantities @xmath112 and @xmath113 $ ] given in @xcite , which were written in terms of the rescaled @xmath49-parameters , @xmath114 .
our result for @xmath115 agrees with @xcite , where the bino contributions have been neglected . it is no surprise that the one - loop effects of interest are all proportional to the @xmath20-parameter , since the off - diagonal two - point terms and the @xmath101 couplings are protected by a u(1)@xmath116 symmetry that is broken in the mssm only by the @xmath20 and @xmath23 term @xcite .
note that the quantum correction to @xmath87 is of @xmath117 , while the @xmath23 ( or rather @xmath87 ) parameter was assumed to be of order @xmath118 .
thus , the large-@xmath0 scenario requires more severe fine - tuning ( by a factor of @xmath0 ) than the one to maintain the `` little hierarchy '' @xmath119 .
it therefore appears less natural in this respect than the generic mssm .
we are now ready to derive the one - loop corrected higgs couplings , which are @xmath0-enhanced relative to their tree - level expressions . for any quantity @xmath120 we can write its nlo mssm expression as @xmath121 where @xmath122 is the contribution from the scale @xmath18 , in practice from loops containing susy particles including the corresponding counterterms , and @xmath123 the remaining contribution from the electroweak scale ( standard model particles and higgses ) .
we will be interested only in effective interactions that vanish or are suppressed at tree - level by a factor of @xmath0 relative to their natural size , and we shall compute them at one loop only at leading order in @xmath124 . in this case
@xmath123 is suppressed relative to @xmath122 and can be set to zero .
this can be understood from the fact that these interactions are protected by the pq - symmetry mentioned above .
this symmetry is broken strongly by the @xmath20 parameter that enters the high - scale contributions @xmath125 , but only weakly in the low - energy contributions by the small @xmath23 ( or rather @xmath87 , but in one - loop diagrams there is no distinction ) parameter , @xmath126 .
the low - energy contributions @xmath123 all involve an insertion of @xmath23 , or an explicit coupling to the small vev @xmath127 , and are therefore negligible . on the other hand , the contribution @xmath122 from the superpartner particle loops is already encoded in the effective couplings of the 2hdm ( [ twohdm ] ) .
we therefore conclude that the leading one - loop contributions to the tree - suppressed higgs couplings in the mssm are simply obtained by computing the physical higgs fields with the general tree - level 2hdm lagrangian ( [ twohdm ] ) .
subsequently , the specific expressions for the couplings @xmath84 , @xmath87 , @xmath128 , and @xmath83 given in the previous section are inserted .
we first derive a relation between the shift ( counterterm ) of the vev @xmath14 and @xmath0 and the mssm parameter @xmath23 , which feeds into all other higgs couplings .
the tadpole counterterm is determined by requiring that the one - loop @xmath129 tadpole ( 1pi one - point function ) vanishes : @xmath130_{0,\rm heavy } + [ \gamma_{h_d}]_{0,\rm light } \stackrel{!}{=}0 .
\label{vevcondition}\ ] ] the tadpole coupling @xmath131 is given by ( [ parameterrel ] ) and set to zero by our renormalization convention .
the tadpole counterterm in terms of the shifts of the vevs is also obtained from ( [ parameterrel ] ) and given by @xmath132 the ( unrenormalized ) susy loop contribution to @xmath133 follows from the linear term in the 2hdm lagrangian ( [ twohdm ] ) after shifting the neutral higgs fields by the vevs , which leads to @xmath134_{0,\rm heavy}= \delta m_{11}^2 v_d -\delta b v_u+\delta \lambda_1 v_d^3 - 3 \delta \lambda_6 v_d^2 v_u+ ( \delta \lambda_3+\delta\lambda_4+\delta \lambda_5 ) v_d v_u^2 - \delta \lambda_7 v_u^3 , \label{dtd}\ ] ] where @xmath135 , @xmath136 denote the susy loop contributions to the corresponding 2hdm parameters , which , however , will not be needed later on .
the low - energy contributions can be neglected , as explained above , so the desired relation for @xmath137 follows from @xmath138_{0,\rm heavy}\stackrel{!}{=}0 $ ] .
this condition simplifies in the large-@xmath0 limit . applying @xmath139 , @xmath140 , ( [ dtd ] ) reads @xmath141
the second approximation uses @xmath142 which follow from inverting ( [ parameterrel ] ) for the renormalized parameters together with @xmath143 .
similarly , @xmath134_{0,\rm heavy } \approx -\delta b \,v - \delta \lambda_7 v^3 . \label{dtdsimp}\ ] ] solving for @xmath137 results in @xmath144 we see that unless the counterterm @xmath145 is fine - tuned the shift of the vev is of order @xmath146 , which is very large relative to its assumed value of order @xmath147 . as discussed above , the shift of the vev is a finite quantity at leading order in large @xmath0 , since @xmath145 cancels the divergences in @xmath91 .
similar statements now apply to the shift of @xmath0 . inserting ( [ vevresult ] ) into ( [ tanbcounter ] ) , we find @xmath148 the higgs masses themselves receive no @xmath0-enhanced one - loop corrections , so @xmath60 corresponds to the physical mass of the cp - odd higgs boson even at the one - loop level in our approximations .
contrary to the other parameters of the second set , there is no natural renormalization convention for the parameter @xmath0 .
gauge - independence and numerical stability of various schemes have been investigated in @xcite .
quite generally , we expect problems with stability of observables against adding perturbative corrections , if the finite part of a counterterm is large , since this leads to a large shift between the tree - level and one - loop corrected values of parameters and large counterterm contributions to observables . from ( [ tanbcounter2 ] )
we conclude that @xmath0 receives large relative corrections , i.e. @xmath149 , unless @xmath150 some of the stability features at large @xmath0 discussed in @xcite can be understood on the basis of this equation .
the @xmath151 scheme for @xmath0 has @xmath152 by definition and fulfills ( [ tbren ] ) as does the dcpr definition @xcite , for which @xmath153 . indeed
, these two schemes were found to exhibit good stability . on the other hand ,
the @xmath154-scheme ( @xmath151 for the @xmath23 parameter ) , implies @xmath155_{\rm fin}+\delta\lambda_7 v^2 } { m_a^2},\ ] ] which is very large , of order @xmath156 .
the higgs mass and tadpole scheme @xcite also violate ( [ tbren ] ) , and are unstable .
good perturbative renormalization schemes should thus have @xmath157 in the large-@xmath0 limit .
since @xmath158 but @xmath159 , this implies fine - tuning of the counterterm for the @xmath23 parameter as expected . to determine the physical higgs fields , we first bring the kinetic terms of
the 2hdm ( [ twohdm ] ) into canonical form by a field redefinition .
this is not necessary to determine the higgs masses , which receive no @xmath0-enhanced corrections , since there are no tree - level suppressions , or observables that involve only highly virtual internal higgs lines , as for instance in low - energy flavour physics .
the latter statement can be illustrated with the help of figure [ fig3 ] , representing part of a diagram with an internal higgs propagator .
the expression for this diagram using the lagrangian ( [ twohdm ] ) is @xmath160_{ba } v_a$ ] , where the propagator is a @xmath28 matrix in higgs doublet space . in low - energy scattering
the momentum transfer satisfies @xmath161 , so the term involving the kinetic - mixing matrix @xmath75 in the 2hdm lagrangian drops out at leading order in the low - energy expansion , and can not affect physical observables . if one performs a field rotation to diagonalize @xmath75 , the rotation affects the vertices @xmath162 and mass matrix @xmath76 in such a way , that once again the effect disappears at leading order in @xmath163 .
however , in this paper our aim is to calculate the couplings of on - shell higgs bosons ( @xmath164 ) to quarks and the higgs fields themselves , relevant to higgs production and decay , so we need the fields corresponding to the physical states .
the field redefinition that diagonalizes the kinetic terms ( and includes the shift of the fields by the vevs ( [ introducevevs ] ) ) is @xmath165_0 + \left(1-\frac{z_{11}}{2}\right ) h_d^0 + \frac{1+a}{2 } z_{12 } h_u^{0 * } , \nonumber\\ \phi_1 ^ 2 = h_d^2 & \to & \left(1-\frac{z_{11}}{2}\right ) h_d^-- \frac{1+a}{2 } z_{12 } h_u^{+ * } , \nonumber\\ -\phi_2^{1 * } = h_u^2 & \to & [ v_u]_0 + \left(1-\frac{z_{22}}{2}\right ) h_u^0 + \frac{1-a}{2 } z_{12 } h_d^{0 * } , \nonumber\\ \phi_2^{2 * } = h_u^1 & \to & \left(1-\frac{z_{22}}{2}\right ) h_u^+- \frac{1-a}{2 } z_{12 } h_d^{- * } , \label{kineticredef}\end{aligned}\ ] ] where @xmath166 is the one - loop correction to @xmath75 , and we used that the hermitian matrix @xmath75 is also symmetric due to cp - conservation .
the diagonal coefficients @xmath167 , @xmath168 can be set to zero , since they are ordinary one - loop corrections to a non - vanishing tree term .
the interesting terms are those that mix @xmath129 with the complex conjugate of @xmath15 . the arbitrary quantity @xmath169 parameterizes a real field rotation in @xmath54 space , which preserves the diagonal form of the kinetic term .
we could set @xmath170 , but prefer to keep it to demonstrate explicitly the independence of physical quantities on @xmath169 below .
note that we do not rotate the fields and then shift them by the vevs , since the vevs ( and @xmath0 ) have been defined as parameters of the mssm lagrangian before matching to the 2hdm . after substituting ( [ kineticredef ] ) into ( [ twohdm ] )
, we perform a unitary ( in fact orthogonal , on account of cp - conservation ) field rotation to diagonalize the higgs mass matrix . the transformation to the physical higgs fields @xmath171 , including the pseudo - goldstone fields @xmath172 , is @xmath173 & s_\beta+\delta s_\beta
\end{array}\right ) \ , \left(\begin{array}{c } g^0\\ a^0 \end{array}\right ) , \nonumber\\[0.2 cm ] \left(\begin{array}{c } h_u^+\\ h_d^{- * } \end{array}\right ) & = & \left(\begin{array}{cc } \phantom{-}s_\beta+\delta s_\beta & c_\beta+\delta c_\beta\\ -[c_\beta+\delta c_\beta ] & s_\beta+\delta s_\beta \end{array}\right ) \
, \left(\begin{array}{c } g^+\\ h^+ \end{array}\right ) , \nonumber\\[0.2 cm ] \left(\begin{array}{c }
\mbox{re}\,h_u^0\\ \mbox{re}\,h_d^0 \end{array}\right ) & = & \frac{1}{\sqrt{2 } } \left(\begin{array}{cc }
\phantom{-}c_\alpha+\delta c_\alpha & s_\alpha+\delta s_\alpha\\ -[s_\alpha+\delta s_\alpha ] & c_\alpha+\delta c_\alpha\\ \end{array}\right ) \ , \left(\begin{array}{c } h^0\\ h^0 \end{array}\right),\end{aligned}\ ] ] where @xmath174 parameterize the correction to the corresponding mssm tree - level rotation , and we use the conventional notation @xmath175 , @xmath176 .
we already incorporated here that the correction @xmath177 to the tree - level mixing matrix turns out to be the same for the cp - odd and the charged higgs fields .
the mixing angle @xmath178 is given by @xmath179 the correction terms @xmath180 , @xmath177 are of the size of an ordinary loop correction , and hence relevant only if the corresponding tree contribution is suppressed .
this is the case for the off - diagonal elements , since @xmath181 .
we therefore neglect the @xmath180 terms relative to @xmath182 . for the off - diagonal correction
we obtain @xmath183 the second term vanishes in `` good '' renormalization schemes .
in determining the correction to @xmath178 , the cases @xmath184 and @xmath185 should be distinguished .
in the following we discuss explicitly only the case @xmath184 .
the other case follows roughly ( that is , up to some signs ) from interchanging @xmath186 and @xmath187 .
for large @xmath0 we have @xmath188 , so ( [ alphatree ] ) implies that either @xmath189 or @xmath190 is small , unless @xmath60 is very close to @xmath191 to compensate the @xmath0 suppression .
for @xmath184 the relevant case is @xmath190 small and negative and @xmath189 near 1 .
the correction to @xmath192 is ( valid for both cases @xmath184 , and @xmath193 ) @xmath194 + \frac{2 m_a^2}{m_a^2-m_z^2}\,\delta\lambda_7 v^2\right ) .
\label{dt2al}\end{aligned}\ ] ] for @xmath184 we have @xmath195 , @xmath196 ( relative to @xmath197 ) .
the field redefinitions that lead to the mass eigenstates therefore read @xmath198 \,h^0 + i [ c_\beta+\delta c_\beta ] \,a^0+i s_\beta g^0\right ) , \nonumber\\ h_d^0 & = & \frac{1}{\sqrt{2}}\left(- [ s_\alpha+\delta s_\alpha ] \,h^0 + c_\alpha h^0 + i s_\beta a^0-i [ c_\beta+\delta c_\beta ] \,g^0\right ) , \nonumber\\ h_d^- & = & s_\beta h^- - [ c_\beta+\delta c_\beta]\ , g^- , \nonumber\\[0.2 cm ] h_u^+ & = & [ c_\beta+\delta c_\beta]\ , h^+ + s_\beta g^+ .
\label{physfields}\end{aligned}\ ] ] we briefly comment on the case @xmath199 , since ( [ dt2al ] ) appears to be singular for this parameter input .
the masses used here are tree - level masses , since we keep loop corrections only when they are enhanced by @xmath0 relative to the tree term . when @xmath199 the cp - even higgs bosons @xmath186 , @xmath187 are mass - degenerate and in fact physically indistinguishable .
( [ dt2al ] ) is derived under the assumption that the difference between the diagonal elements of the neutral higgs boson mass matrix is not small .
otherwise one must keep the one - loop correction to the diagonal elements in the calculation of @xmath200 , and the approximation discussed here can not be used .
thus the resonant enhancement in ( [ dt2al ] ) is a real effect as long as @xmath201 is large compared to @xmath202 .
however , if the widths of the higgs bosons are so large that the @xmath186 and @xmath187 intermediate states are not separated by some measurement , nothing distinguishes the large tree - unsuppressed @xmath203 component from the smaller @xmath204 h^0 $ ] component in ( [ physfields ] ) and there is no observable @xmath0-enhancement in this measurement . in the following
, we do not consider this degenerate case further .
the yukawa couplings and quark mass terms follow from inserting ( [ physfields ] ) into ( [ twohdm ] ) .
the bottom mass term is of special interest .
the relevant bilinear terms are @xmath205 \,\bar b_r b_l + \mbox{h.c . }
\label{massb}\ ] ] the physical ( pole ) quark mass @xmath13 is obtained from this mass term plus low - energy self - energy contributions calculated in the 2hdm . as before we keep one - loop corrections only when they are enhanced relative to the tree expressions .
this allows us to neglect the low - energy contributions ( see @xcite for the standard model electroweak loops ) as well as @xmath206 , and @xmath207 in ( [ massb ] ) .
@xmath208 is the counterterm that cancels the divergences from superpartner loops in @xmath207 .
the finite part of @xmath208 defines the relation between the physical bottom quark mass and the renormalized yukawa coupling @xmath12 .
there is no reason to artificially define this finite part to be parametrically larger than the one of @xmath207 , so we may drop @xmath208 as well . introducing the definitions
@xmath209 we obtain the mass terms @xmath210 with @xmath211 note that we do not expand in @xmath71 , since according to ( [ tanbcounter2 ] ) it counts as a @xmath0-enhanced loop correction relative to @xmath0 , just as @xmath212 does relative to 1 .
no such enhancements are present for the top quark mass , which retains its tree expression .
the expression for @xmath213 is obtained from the one for @xmath13 by replacing @xmath214 . solving ( [ mb ] ) for @xmath12 and eliminating @xmath51 by @xmath66 ,
as it is conventionally done , we obtain the higgs - yukawa interactions @xmath215 } { \sqrt{2 } v \,(1+\epsilon_b \,[\tan\beta+\delta t_\beta ] ) } \bigg\{i\,\bar b_r b_l a^0-\sqrt{2 } \bar b_r t_l h^- \nonumber\\ & & \hspace*{1.7cm}-\,\frac{s_\alpha + \delta s_\alpha - \epsilon_b-\frac{1+a}{2 } z_{12}}{s_\beta } \,\bar b_r b_l h^0 + \frac{c_\alpha}{s_\beta } \,\bar b_r b_l h^0 \nonumber\\ & & \hspace*{1.7 cm } - \ , \left(\frac{1}{\tan\beta } + \epsilon_b + \delta c_\beta + \frac{1+a}{2 } z_{12}\right ) \left(i\,\bar b_r b_l g^0-\sqrt{2 }
\bar b_r t_l g^-\right ) \bigg\ } \nonumber\\ & & + \ , { \rm h.c . } , \label{eq : yukawa}\end{aligned}\ ] ] where as before we neglect ordinary one - loop corrections relative to unsuppressed tree terms , and the @xmath216 yukawa couplings are given by the bottom ones after replacing all bottom parameters and fields by the corresponding expressions for tau .
several comments can be made .
first , the explicit dependence on the arbitrary parameter @xmath169 cancels with the implicit dependence in ( [ dcbe ] ) , ( [ dt2al ] ) .
second , our result is independent on the renormalization convention for @xmath0 . a change of renormalization prescription is effected by a change in the finite part of @xmath71 , or ( [ tanbcounter2 ] ) @xmath137 , or ( [ vevresult ] ) @xmath145 . from ( [ dcbe ] ) , ( [ dt2al ] )
, we therefore find that @xmath217 , \left[\frac{1}{\tan\beta } + \delta c_\beta\right ] , [ s_\alpha + \delta s_\alpha ] , s_\beta , c_\alpha\ ] ] are invariant under a change of renormalization prescription that contains only @xmath0-enhanced one - loop terms .
( [ eq : yukawa ] ) shows that a resummation of @xmath0-enhanced terms is necessary only in the overall factor multiplying the bottom ( and tau ) yukawa couplings @xcite , where the need for resummation arises from using the quark ( lepton ) mass as a parameter rather than the mssm yukawa coupling .
all of the yukawa couplings which are @xmath0-suppressed at tree - level receive unsuppressed loop corrections , which are incorporated in ( [ eq : yukawa ] ) through the `` wrong - higgs '' yukawa couplings @xmath218 , and the quantities @xmath177 , @xmath219 , @xmath84 from the one - loop effects in the higgs sector .
the loop - induced couplings may well exceed the tree terms , if @xmath0 is large . to compare our result with previous results in the literature , see eq .
( 21 ) in @xcite , we consider the bottom quark coupling to the lightest higgs boson @xmath186 , which is the product of two factors @xmath220 } { \sqrt{2 } v \,(1+\epsilon_b \,[\tan\beta+\delta t_\beta ] ) } \times \frac{s_\alpha + \delta s_\alpha - \epsilon_b-\frac{1+a}{2 } z_{12}}{s_\beta}.\ ] ] with respect to the first factor , note that we do not discuss logarithmic one - loop effects , and hence do not distinguish the physical quark mass @xmath13 from the running @xmath221 quark mass @xmath222 used in @xcite . in reality ,
renormalization group evolution from @xmath13 to the electroweak scale is an important effect that should be taken into account as done in @xcite .
the main difference in the first factor constitutes the presence of the ( finite ) counterterm @xmath71 , which must not be neglected in general , and which renders the resummed result manifestly scheme - independent .
this clarifies that the previously known result where @xmath152 , is only valid in `` good '' renormalization schemes ( [ tbren ] ) where the one - loop shift of @xmath0 is not itself @xmath0-enhanced , as also observed recently in the related context of the muon yukawa coupling @xcite . the resummation formula from @xcite without the counterterm contribution
should not be used in other schemes , for which @xmath71 is an important correction in the first factor . using @xmath223 in the one - loop corrections , and ( [ tanbcounter2 ] ) , ( [ dt2al ] )
, the second factor can be rewritten as @xmath224 the first two terms reproduce the corresponding result in @xcite , but the higgs - sector effects embodied in the last three terms have not been included there .
higgs - sector effects related to kinetic mixing have recently been computed in @xcite . the @xmath151 scheme for @xmath0 is assumed there , hence the last term involving @xmath225 vanishes .
we find several discrepancies in the comparison with @xcite .
the third term proportional to @xmath84 would be consistent with the result of @xcite if @xmath226 , but our result for @xmath84 in ( [ z121loop ] ) gives @xmath227 .
we also find that the corrections to the @xmath228 coupling are not included in @xcite .
the most important difference is that the fourth term involving @xmath229 is missed , since it is assumed that higgs - sector effects arise only from kinetic mixing , and not ( also ) from the higgs self - couplings .
the large quantum correction to the tree - level suppressed charged higgs @xmath230 coupling is present in the discussion of the @xmath231 transition in @xcite .
these papers account for the gluino contribution to @xmath232 .
the higgs - sector terms involving @xmath177 and @xmath84 are again not considered .
more precisely , in the expression for the third term proportional to @xmath84 ref .
@xcite has @xmath233 , where we have @xmath234 .
the two expressions are consistent with each other , since in the one - loop correction we may use the large-@xmath0 limits of the tree - level mass relations , which give @xmath235 , and @xmath236 . in reality
, the higgs masses must receive large quantum corrections to satisfy experimental limits , which are not included in our equations , since they are not @xmath0-enhanced . since the apparent singularity for @xmath199 arises from the degeneracy of the two cp - even higgs bosons , as discussed above , it is plausible that @xmath233 provides a better approximation , even though formally the difference to @xmath234 is a higher - order effect .
in section [ numestimate ] we perform some representative numerical estimates of the new higgs - sector corrections to the yukawa couplings and compare them to the previously known term @xmath237 .
the higgs self - couplings are obtained in the same straightforward fashion . here
we focus on the trilinear couplings , since the quartic self - interactions with coefficients induced by quantum effects will be too small to be measured in any foreseeable experiment .
the detection of @xmath0- or loop - suppressed trilinear couplings provides a challenge even to the ilc ( see , e.g. , @xcite ) .
the lagrangian for the trilinear couplings @xmath238 consists of a piece @xmath239 involving interactions that do not vanish at tree level when @xmath240 , and another piece @xmath241 with interactions that do .
the tree - level couplings in the mssm are standard and will not be repeated here .
we find a compact expression for the @xmath0-enhanced one - loop corrections relative to the small tree - level interactions in @xmath241 , applicable to the case @xmath184 : @xmath242 & & \hspace*{-1 cm } + \,(m_a^2 \,c_h -\delta\lambda_6-\delta\lambda_7 ) \,({h^{0}}^2+{a^{0}}^2)h^{0 } \nonumber\\[0.1 cm ] & & \hspace*{-1 cm } + \,2\,\big[(m_a^2 - 2 m_w^2 ) \,c_h -(\delta\lambda_6+\delta\lambda_7)\big ] \,h^0 h^+ h^- \nonumber\\[0.1 cm ] & & \hspace*{-1 cm } + \ , 2 \,(m_a^2-m_z^2)\,c_h \,h^0 a^0 g^0 - 2 \delta\lambda_5 \,h^0 a^0 g^0 - m_a^2\,c_h\,h^0 ( { g^0}^2 + 2\,g^+ g^- ) \nonumber\\[0.1 cm ] & & \hspace*{-1 cm } + \,2\,(m_a^2-m_z^2+m_w^2 ) \,c_h\,h^0 ( h^- g^++h^+ g^- ) \bigg\}\ , , \label{eq : tri}\end{aligned}\ ] ] where @xmath243 + \delta\lambda_7 \right).\ ] ]
in this section we present numerical estimates of the higgs - sector corrections by evaluating them for degenerate susy masses , and by performing a random parameter space scan .
we are especially interested in comparing the new higgs - sector effects to the known one - loop induced `` wrong - higgs '' yukawa couplings .
we begin the analysis of the size of the new terms in the yukawa couplings relative to @xmath244 by assuming that all susy parameters are equal to a common mass scale @xmath18 and by allowing for arbitrary signs in the @xmath20 and @xmath49 parameters , but not the gaugino masses . in the following we give analytic results for @xmath245 and the quantities @xmath84 and @xmath246 , which determine the higgs - sector corrections to the yukawa couplings and the higgs self - couplings , and numerical results to judge the importance of individual terms . to make contact with the literature , we introduce the rescaled @xmath49 parameters @xmath247 by @xmath248 . in units of @xmath249 , we find @xmath250 ^ 2 } \
, , \nonumber\\ \epsilon_\tau & = & \,- 18\pi \alpha_{\rm em } \left(\frac{1}{s^2_w } - \frac{1}{3 c^2_w } \right ) = -1.80 \label{eq : analytic1}\end{aligned}\ ] ] for the `` wrong - higgs '' yukawa couplings , and @xmath251 ^ 2 } \,+\ , { \rm sgn}(\widetilde{a}_\tau ) \ , \frac{0.26}{[50\,c_\beta]^2 } \ , , \nonumber\\[0.2 cm ] \delta\lambda_5 & = & { \rm sgn}(\mu ) \,\bigg\ { - 3 y_t^4 - 3 y_b^4 - y_\tau^4 + 16\pi^2 \alpha_{\rm em}^2 \left ( \frac{3}{s^4_w } + \frac{1}{c^4_w } + \frac{2}{c^2_w s^2_w } \right ) \bigg\ } \nonumber\\ & = & { \rm sgn}(\mu ) \,\bigg\ { 0.712 \,-\ , \frac{3.04}{s_\beta^4 } \,-\ , \frac{1.59}{[50\,c_\beta]^4 } \bigg\ } \ , , \nonumber\\[0.2 cm ] \delta\lambda_6 & = & 15 \ , y_b^4\ , { \rm sgn } ( \widetilde{a}_b ) - 3 y_t^4 \ , { \rm sgn } ( \widetilde{a}_t ) + 5 y_\tau^4 \ , { \rm sgn } ( a_\tau ) \nonumber\\[0.1 cm ] & & - \ , \frac{3\pi \alpha_{\rm em}}{c^2_w \ , s^2_w } \left ( -\,3 y_t^2 \ , { \rm sgn } ( \widetilde{a}_t ) + 3 y_b^2 \ , { \rm sgn } ( \widetilde{a}_b ) + y_\tau^2 \ , { \rm sgn } ( \widetilde{a}_\tau ) \right ) \nonumber\\ & & + \ , 32 \pi^2 \alpha_{\rm em}^2 \left ( \frac{3}{s^4_w } + \frac{1}{c^4_w } + \frac{2}{s^2_w c^2_w } \right ) \nonumber\\ & = & 1.42 \,-\ , { \rm sgn}(\widetilde{a}_t ) \left(\frac{3.04}{s^4_\beta } \,-\,\frac{1.29}{s_\beta^2}\right ) \,-\ , { \rm sgn}(\widetilde{a}_b ) \left(\frac{0.91}{[50\,c_\beta]^2 } \,-\ ,
\frac{7.62}{[50\,c_\beta]^4}\right ) \nonumber\\ & & \,-\ , { \rm sgn}(\widetilde{a}_\tau ) \left(\frac{0.11}{[50\,c_\beta]^2 } \,-\ , \frac{0.34}{[50\,c_\beta]^4}\right ) \ , , \nonumber\\[0.2 cm ] \delta\lambda_7 & = & 15 \ , y_t^4 \ , { \rm sgn } ( \widetilde{a}_t ) - 3 y_b^4 \ , { \rm sgn } ( \widetilde{a}_b ) - y_\tau^4 \ , { \rm sgn } ( \widetilde{a}_\tau ) \nonumber\\[0.1 cm ] & & + \ , \frac{3\pi \alpha_{\rm em}}{c^2_w \ , s^2_w } \left ( -3 y_t^2 \ , { \rm sgn } ( \widetilde{a}_t ) + 3 y_b^2 \ , { \rm sgn } ( \widetilde{a}_b ) + y_\tau^2 \ , { \rm sgn } ( \widetilde{a}_\tau ) \right ) \nonumber\\ & & + \ , 32 \pi^2 \alpha_{\rm em}^2 \left ( \frac{3}{s^4_w } + \frac{1}{c^4_w } + \frac{2}{s^2_w c^2_w } \right ) \nonumber\\ & = & 1.42 \,+\ , { \rm sgn}(\widetilde{a}_t ) \left(\frac{15.2}{s^4_\beta } \,-\ , \frac{1.29}{s_\beta^2}\right ) \,+\ , { \rm sgn}(\widetilde{a}_b ) \left ( \frac{0.91}{[50\,c_\beta]^2 } \,-\ ,
\frac{1.52}{[50\,c_\beta]^4}\right ) \nonumber\\ & & \,+\ , { \rm sgn}(\widetilde{a}_\tau ) \left ( \frac{0.11}{[50\,c_\beta]^2 } \,-\ , \frac{0.07}{[50\,c_\beta]^4}\right ) \label{eq : analytic}\end{aligned}\ ] ] for the higgs - sector couplings . our results for @xmath246 agree with the special case given in @xcite , which assumed all squark masses equal to @xmath18 and neglected the gaugino and slepton contributions .
the numerical values in ( [ eq : analytic ] ) have been evaluated with sm parameter input as follows : for the gauge boson masses we choose @xmath252gev and @xmath253gev , the gauge couplings are fixed to @xmath254 and @xmath255 .
the cosine of the weak mixing angle is given by @xmath256 .
for the top yukawa coupling we use @xmath257 , where @xmath258gev and @xmath259gev .
for the down - type and lepton yukawa couplings we here use the tree - level relations @xmath260 and @xmath261 , where @xmath262gev and @xmath263mev .
all fermion masses correspond to the renormalized masses in the @xmath264 scheme at @xmath265 @xcite rather than the pole masses . , on @xmath0 . for the ratios @xmath266 ,
@xmath267 the plots correspond to the sign choices @xmath268 . in the plots for @xmath269 we adopt @xmath270.,scaledwidth=100.0% ] since @xmath237 is nearly independent of @xmath0 for @xmath271 , we show the @xmath25 dependence of the ratios @xmath272 and @xmath273 from ( [ eq : analytic1 ] ) , ( [ eq : analytic ] ) in figure [ fig4 ] for different choices of the signs of the @xmath247-parameters . the slepton terms give a negligible contribution except for @xmath84 at very large values of @xmath0 , so we always set the sign of @xmath274 to @xmath275 . for @xmath0
smaller than 20 30 the value of @xmath84 , @xmath246 is determined by the @xmath276 terms and the gaugino contributions . for @xmath229 , in particular , which feeds into the yukawa couplings , the terms proportional to @xmath276 give by far the largest contributions , while the effect of @xmath277 becomes relevant for very large @xmath0 values , reaching up to a 20% contribution . on the other hand we observe from the plots that the @xmath277 terms and a change in the sign of @xmath277 has a large effect on @xmath278 , which is driven by the @xmath279 term in ( [ eq : analytic ] ) with its sizeable coefficient . in general , under the assumption of a common mass scale for all susy parameters , we thus conclude that the quantities @xmath280 and particularly @xmath229 are comparable in size with the `` wrong - higgs '' couplings @xmath244 even though they do not receive gluino contributions . note that @xmath281 is larger than @xmath282 for all choices of signs of the @xmath283 .
we now investigate in more detail the size of the higgs one - loop corrections to the yukawa interactions relative to @xmath244 for different susy parameter input .
we consider good renormalization schemes for @xmath0 , _
i.e. _ schemes satisfying @xmath284 in the large-@xmath0 limit , in which case the two relevant corrections from ( [ eq : yukawa ] ) are @xmath285 in terms of the quantities @xmath286 , the one - loop improved couplings of the higgs fields to the bottom and top quark are rewritten as , omitting the global factors outside the curly brackets in eq .
( [ eq : yukawa ] ) , @xmath287 the function @xmath288 , which appears in the one - loop corrections to the small tree - level trilinear higgs couplings ( [ eq : tri ] ) , also has a simple expression in terms of @xmath289 , given by @xmath290 we perform a random parameter sampling for the susy parameters in the following ranges : 500 gev@xmath291tev , with @xmath292 being the squark , slepton or gaugino masses , and 500 gevlatexmath:[$\,\le |\mu| , | \widetilde{a}_t| , | \widetilde{a}_b| , the parameters @xmath20 , @xmath276 and @xmath277 are explored separately , but we fix @xmath294 .
since the @xmath60 dependence of @xmath289 comes in only through the prefactor @xmath295 , we shall fix it to the reference value @xmath296gev . here
contrary to the analytic expressions in the previous subsection , we correctly include the @xmath0-enhanced loop correction to the down - type quark and lepton yukawa couplings , _
i.e. _ @xmath297 is used .
we remark that not all points allowed in the parameter scan are physical , since vacuum stability requirements do not allow large @xmath49 terms relative to the scale of the susy particle masses . and @xmath298 versus @xmath232 and @xmath237 respectively , for @xmath296gev and positive @xmath20 .
the plots in the first row have @xmath299 and those in the second row have @xmath300 ( @xmath301 in all plots ) . ]
let us first study the case of positive @xmath20 , where @xmath237 is preferentially positive , so that @xmath12 is reduced by the one - loop correction . in figure [ fig5 ]
we show @xmath302 ( left ) and @xmath298 ( right ) versus @xmath232 and @xmath237 , respectively , for both signs of @xmath303 ( positive in the upper plots , negative in the lower ones ) . for positive @xmath20 , the value of @xmath0 does not significantly change the shape of the scatter plot , and
hence has been fixed to @xmath304 .
the typical size of the effective couplings is a few percent .
values of @xmath305 range up to 0.1 , which implies that the one - loop induced coupling may exceed the @xmath0-suppressed tree coupling for values of @xmath0 as small as 10 . in comparison
, @xmath302 , which originates only from kinetic mixing , tends to be significantly smaller .
thus , the higgs sector correction @xmath298 competes in size with the `` wrong - higgs '' yukawa coupling @xmath237 even for large values of @xmath60 , and is in fact the dominant one - loop correction to the @xmath306 yukawa interaction .
some general features that can be observed in figure [ fig5 ] , or extracted from the analytic expressions are : * in the assumed range of susy parameters , the sign of @xmath237 ( @xmath232 ) is correlated ( anti - correlated ) with the sign of @xmath20 , since the gluino contribution ( first term in ( [ eq : deltayb ] ) ) dominates and the function @xmath307 is always negative . the term proportional to @xmath276 can overcome the gluino contribution and change the sign of @xmath237 only for negative @xmath276 in the parameter - space region where @xmath308 is much larger than @xmath309 .
similarly , large negative @xmath277 can make @xmath232 positive , when @xmath310 . *
since @xmath237 can not reach large negative values , there is no significant enhancement of the bottom yukawa coupling from the large-@xmath0 resummation .
the numerical effect of the terms proportional to @xmath311 is therefore small compared to those proportional to @xmath312 .
we verified that the scatter plots for negative @xmath277 are similar to the ones with positive @xmath277 shown in figure [ fig5 ] . * larger values of @xmath313 ( @xmath314 )
tend to be correlated with larger values of @xmath315 ( @xmath305 ) .
however , there is no strict relation , since the dominant contribution to the `` wrong - higgs '' yukawa couplings is proportional to the gluino mass , while those to the higgs - sector couplings involve the @xmath49 terms . * the largest contribution to @xmath302 is given by the term proportional to @xmath276 in ( [ z121loop ] ) .
however , the sign of @xmath302 is not determined uniquely by the sign of @xmath276 due to the gaugino term proportional to @xmath316 , which is always positive ( the function @xmath317 is positive ) and can become comparable to the @xmath276 term in some regions of the parameter space . * in most of the explored parameter space
the sign of @xmath298 is opposite to the sign of @xmath276 .
since the @xmath229 term in the definition ( [ eq : h1h2 ] ) of @xmath298 is the larger of the two , we have @xmath318 .
the relation @xmath319 can be understood as follows : neglecting the smaller terms involving gauge couplings , eq .
( [ eq : lambda7 ] ) for @xmath229 contains the terms @xmath320 proportional to powers of @xmath276 .
the first term has the same sign as @xmath276 , since @xmath321 is always negative , while the second term has opposite sign , since @xmath322 is positive . defining the ratio @xmath323 with @xmath324
, we find that @xmath325 is monotonically increasing , satisfying @xmath326 when @xmath327 as allowed by our parameter - space sampling . the third term in ( [ eq : at ] )
can compete with the first only if @xmath328 .
for @xmath329 this relation requires large @xmath49 terms , @xmath330 , which are disfavoured by vacuum stability arguments .
we also note that the largest gaugino term in @xmath229 ( the one proportional to @xmath331 ) gives a positive contribution because the function @xmath332 is negative . versus @xmath237 , for @xmath296gev and negative @xmath20 .
the plots in the first row correspond to @xmath333 and @xmath334 , while those in the second row have @xmath335 .
black , lightgray and darkgray points ( blue , yellow and red in colored plots ) correspond to @xmath336 and @xmath337 , respectively . ]
let us now turn to the case @xmath338 .
the most important difference is that the negative sign of the @xmath20 parameter flips the sign of @xmath237 to negative values ( and the one of @xmath232 to positive values ) , which leads to a strong increase of the bottom and @xmath216 yukawa coupling , when @xmath237 cancels the @xmath124 term in ( [ mb ] ) .
the validity of a perturbative expansion is in doubt when the yukawa couplings become too large . in the parameter space sampling for negative @xmath20
, we therefore only keep points that satisfy @xmath339 .
the dependence of the higgs sector correction @xmath298 versus @xmath237 , for three representative values of @xmath340 , @xmath296gev , and all other susy parameters scanned randomly in the above intervals is shown figure [ fig6 ] .
the perturbativity cut on @xmath12 has a strong effect , since it eliminates points with large negative values @xmath237 given a value of @xmath0 .
the value of @xmath237 at which this happens can be estimated by @xmath341 , which gives @xmath342 for @xmath340 , respectively , in agreement with the figures . at these points
we observe a rapid increase of @xmath305 driven by the term @xmath343 in ( [ eq : lambda7 ] ) , which gives a contribution to @xmath344 with sign opposite to @xmath277 .
point i/ discussed above also applies to the case of negative @xmath20 taking into the account the reversed signs of @xmath244 .
finally , figure [ fig7 ] shows the the higgs correction @xmath302 versus @xmath232 for @xmath340 .
compared to the case of positive @xmath20 , the sign of @xmath277 has now a relevant effect , especially for larger values of @xmath0 , due to the contribution of the @xmath277 term in @xmath84 .
however , the growth with @xmath12 is less pronounced than in case of @xmath298 , since @xmath302 does not contain terms proportional to @xmath345 . versus @xmath232 , for @xmath296gev and negative @xmath20 .
plots in the first row have @xmath346 and @xmath334 , and @xmath347 in the second row .
black , lightgray and darkgray points ( blue , yellow and red in colored plots ) correspond to @xmath336 and @xmath337 , respectively . ]
this paper has been motivated by previous work @xcite that systematically investigated the resummation of susy qcd large-@xmath0 effects in the decoupling limit , where the standard model particles and higgs scalars remain light , but which did not consider electroweak effects .
we performed a complete one - loop matching of the mssm to a two - higgs doublet model for all couplings that are absent or suppressed in the mssm with large @xmath0 , keeping all contributions that are @xmath0-enhanced relative to the suppressed tree terms .
this includes complete expressions for the kinetic mixing and higgs self couplings @xmath246 , which have not been given previously .
our result confirms , as expected , that a resummation of large-@xmath0 effects to arbitrary loop order is necessary only for the bottom and tau yukawa couplings , if the bottom and tau mass are used as parameters of the mssm .
it clarifies a point left open in @xcite , namely how the renormalization of @xmath0 affects this resummation .
we find that the standard expression for the resummed yukawa coupling is valid in renormalization schemes where the counterterms ( shifts ) of @xmath0 and the vacuum expectation value @xmath14 do not receive large-@xmath0 contributions .
in other schemes , a finite counterterm must be explicitly included in the resummation formula .
besides the known `` wrong - higgs '' yukawa couplings , all other higgs - sector effects that feed into the yukawa couplings by modifying the definition of the physical higgs fields at one loop can be parameterized in terms of two couplings @xmath302 , @xmath298 dominated by kinetic mixing and the higgs self - coupling @xmath229 , respectively .
the same @xmath298 also enters the effective trilinear higgs couplings .
our numerical study suggests that , where present , the higgs - sector effect @xmath298 is more important than the `` wrong - higgs '' yukawa couplings , and we identified the dominant susy parameter dependences .
our result extends and corrects a previous result @xcite that included the kinetic - mixing effect , but not the one from the one - loop induced higgs couplings .
although obtained in the decoupling limit , the effective one - loop couplings derived in this paper should be useful to obtain simple analytic estimates of the leading quantum corrections to those yukawa and higgs trilinear interactions , that are suppressed at tree - level , such as the @xmath348 coupling .
we thank u. nierste and s. jger for useful discussions .
this work is supported in part by the dfg sonderforschungsbereich / transregio 9 `` computergesttzte theoretische teilchenphysik '' and the swiss national science foundation ( snf ) .
m.b . acknowledges hospitality from the university of zrich and the cern theory group , where part of this work was performed .
feynman diagrams have been drawn with the packages axodraw @xcite and jaxodraw @xcite .
a. dabelstein , nucl .
b * 456 * ( 1995 ) 25 [ hep - ph/9503443 ] .
j. a. coarasa , r. a. jimenez and j. sola , phys .
b * 389 * ( 1996 ) 312 [ hep - ph/9511402 ] .
m. s. carena , d. garcia , u. nierste and c. e. m. wagner , nucl .
b * 577 * ( 2000 ) 88 [ hep - ph/9912516 ] .
l. j. hall , r. rattazzi and u. sarid , phys .
d * 50 * ( 1994 ) 7048 [ hep - ph/9306309 ] .
a. freitas and d. stckinger , phys .
d * 66 * ( 2002 ) 095014 [ hep - ph/0205281 ] .
n. baro , f. boudjema and a. semenov , arxiv:0807.4668 [ hep - ph ] .
a. freitas , e. gasser and u. haisch , phys .
d * 76 * ( 2007 ) 014016 [ hep - ph/0702267 ] .
s. trine , to appear in the proceedings of susy07 , arxiv:0710.4955 [ hep - ph ] .
j. rosiek , phys .
d * 41 * ( 1990 ) 3464 [ erratum : hep - ph/9511250 ] j. a. aguilar - saavedra _ et al .
_ , eur .
j. c * 46 * , 43 ( 2006 ) [ hep - ph/0511344 ] .
a. dabelstein , z. phys .
c * 67 * ( 1995 ) 495 [ hep - ph/9409375 ] .
h. chankowski , s. pokorski and j. rosiek , nucl .
b * 423 * , 437 ( 1994 ) [ hep - ph/9303309 ] .
r. hempfling and b. a. kniehl , phys . rev .
d * 51 * ( 1995 ) 1386 [ hep - ph/9408313 ] .
b. a. kniehl , nucl .
b * 376 * ( 1992 ) 3 .
s. marchetti , s. mertens , u. nierste and d. stckinger , arxiv:0808.1530 [ hep - ph ] .
m. s. carena , d. garcia , u. nierste and c. e. m. wagner , phys .
b * 499 * ( 2001 ) 141 [ hep - ph/0010003 ] .
m. ciuchini , g. degrassi , p. gambino and g. f. giudice , nucl .
b * 534 * ( 1998 ) 3 [ hep - ph/9806308 ] .
a. djouadi , w. kilian , m. mhlleitner and p. m. zerwas , eur .
j. c * 10 * ( 1999 ) 27 [ hep - ph/9903229 ] .
z. z. xing , h. zhang and s. zhou , phys .
d * 77 * ( 2008 ) 113016 , arxiv:0712.1419 [ hep - ph ] .
h. e. haber and r. hempfling , phys .
d * 48 * ( 1993 ) 4280 [ hep - ph/9307201 ] .
vermaseren , comput .
phys . commun .
83 ( 1994 ) 45 .
d. binosi and l. theussl , comput .
commun . 161
( 2004 ) 76 [ hep - ph/0309015 ] . | we consider @xmath0-enhanced quantum effects in the minimal supersymmetric standard model ( mssm ) including those from the higgs sector . to this end
, we match the mssm to an effective two - higgs doublet model ( 2hdm ) , assuming that all susy particles are heavy , and calculate the coefficients of the operators that vanish or are suppressed in the mssm at tree - level .
our result clarifies the dependence of the large-@xmath0 resummation on the renormalization convention for @xmath0 , and provides analytic expressions for the yukawa and trilinear higgs interactions .
the numerical effect is analyzed by means of a parameter scan , and we find that the higgs - sector effects , where present , are typically larger than those from the `` wrong - higgs '' yukawa couplings in the 2hdm .
pitha 08/27 + sfb / cpp-08 - 83 + mpp-2008 - 132 + october 14 , 2008 1.5 cm * higgs couplings in the mssm at large @xmath0 * m. beneke@xmath1 , p. ruiz - femena@xmath2 , and m. spinrath@xmath3 + _ @xmath2 institut fr theoretische physik e , rwth aachen university , + d52056 aachen , germany
_ + @xmath4 _ institut fr theoretische physik , universitt zrich , + ch 8057 zrich , switzerland _ + _ @xmath5 max - planck institut fr physik , fhringer ring 6 , + d80805 mnchen , germany _ |
generalized parabolic vector bundles ( gpbs ) on a smooth curve @xmath3 are vector bundles on @xmath3 together with parabolic structures on finitely many disjoint divisors @xmath7 @xcite .
there is an open subscheme @xmath8 of the moduli space @xmath0 of gpbs on which one can define a determinant morphism into the moduli space of generalized parabolic line bundles @xmath9 , the map does not extend to @xmath0 .
let @xmath10 be its locally closed subset consisting of gpbs with a fixed determinant @xmath5 . in this note
, we define a moduli functor and construct a coarse moduli scheme @xmath4 for it .
the moduli scheme contains @xmath10 as an open dense subscheme .
let @xmath6 be an irreducible projective nodal curve with nodes @xmath11 and @xmath12 its desingularization with @xmath13 the inverse image of @xmath14 .
denote by @xmath15 the moduli variety of torsion - free sheaves of rank @xmath1 , degree @xmath2 on @xmath6 .
let @xmath16 be the open subvariety of @xmath15 corresponding to vector bundles on @xmath6 .
there is a surjective morphism @xmath17 from @xmath0 onto @xmath15 @xcite .
the restriction of the morphism @xmath17 to @xmath18 is an isomorphism onto the open subvariety @xmath16 of @xmath15 .
a gpb @xmath5 gives a torsion - free sheaf @xmath19 on @xmath6 .
if @xmath19 is locally free , let @xmath20 be the closed subset of @xmath16 consisting of vector bundles with fixed determinant @xmath19 .
using @xmath21 may be identified with @xmath22 and @xmath4 can be regarded as compactified moduli variety of vector bundles on @xmath6 with determinant @xmath19 .
we show that @xmath23 , the closure of @xmath20 in @xmath15 , thus giving @xmath24 the scheme structure of an image subscheme of @xmath15 .
let @xmath25 denote the ideal sheaf at the node @xmath14 .
for a torsion - free sheaf @xmath26 of rank @xmath1 on @xmath6 , let @xmath27 where ( torsion ) denotes the torsion subsheaf .
then we show that for any @xmath5 , the image @xmath28 can be described ( as a set ) by @xmath29 this gives a proof of a conjecture by seshadri and nagaraj ( conjecture ( a ) , p. 136
of @xcite ) .
proving seshadri nagaraj conjecture was not the aim of this note .
the conjecture was proved by sun @xcite by degeneration methods .
however he does not get a scheme structure on @xmath30 or a moduli functor ( except in some low rank cases ) .
our aim is to give a moduli functor and an explicit construction of a projective moduli space for it which contains an open subvariety isomorphic to @xmath20 if @xmath19 is a line bundle
. we also deal with the case when @xmath19 is torsion - free but not locally free .
the construction is much simpler than that of schmitt @xcite and hence the moduli space is easier to study .
for example , properties like reduced , irreducible , cohen macaulay follow immediately for our moduli space .
normality is true in rank @xmath31 and is expected to be true in general .
these properties have been used in computation of picard groups in the rank two case .
[ [ section ] ] let @xmath3 be a nonsingular projective curve over an algebraically closed base field @xmath32 .
let @xmath33 be disjoint divisors on @xmath3 with @xmath34 , where @xmath35 are distinct closed points .
we recall here some basics on generalized parabolic bundles ( gpbs ) , details may be found in @xcite .
@xmath36 a generalized parabolic bundle ( gpb , in short ) of rank @xmath1 and degree @xmath2 on @xmath3 is a vector bundle @xmath37 of rank @xmath1 and degree @xmath2 on @xmath3 together with @xmath1-dimensional vector subspaces @xmath38 of @xmath39 . for a subbundle @xmath40 of @xmath37 ,
define @xmath41 and @xmath42 .
@xmath36 fix a rational number @xmath43 $ ] .
a gpb @xmath44 is @xmath45-stable ( resp .
@xmath45-semistable ) if for every proper subbundle @xmath40 of @xmath37 , one has @xmath46
@xmath47 @xmath36 let @xmath48 be the projections .
assume that for each @xmath49 , at least one of @xmath50 is an isomorphism .
the subspace @xmath38 determines an element @xmath38 of @xmath51 .
one has a ( rational ) morphism @xmath52 .
let @xmath53 denote the one - dimensional subspace of @xmath54 determined by @xmath55 .
define the determinant of @xmath56 to be the generalized parabolic line bundle @xmath57 .
@xmath36 a family of gpbs of rank @xmath1 , degree @xmath2 parametrized by a scheme @xmath58 is a tuple @xmath59 where @xmath60 is a family of vector bundles of rank @xmath1 , degree @xmath2 on @xmath3 which is flat over @xmath58 and @xmath61 is a rank @xmath1 subbundle of @xmath62 .
the notion of equivalence of families is the obvious one .
we fix a generalized parabolic line bundle @xmath63 .
fix isomorphisms @xmath64 .
then @xmath65 can be identified to a point @xmath65 of @xmath66 of the form @xmath67 or @xmath68 . for simplicity ,
let us assume that there is only one divisor @xmath69 .
let @xmath70 be a family of gpbs of rank @xmath1 , degree @xmath2 on @xmath3 with @xmath71 of fixed determinant @xmath72 .
for @xmath73 we have vector bundles @xmath74 let @xmath75 denote the grassmannian bundle of rank @xmath1 subbundles of @xmath76 .
it is embedded as a closed subvariety in @xmath77 by plcker embedding .
note that @xmath78 defines a section of @xmath79 .
since @xmath80 , it follows that @xmath81 for some line bundle @xmath40 on @xmath58 .
hence for @xmath82 one has @xmath83 , using the isomorphism @xmath84 .
let @xmath85 be the projections , @xmath86 .
define a hyperplane subbundle @xmath87 of @xmath88 by @xmath89 if @xmath90 , by @xmath91 if @xmath92 and by @xmath93 if @xmath94 .
let @xmath95 .
it is a closed reduced subscheme of @xmath96 ) ) .
note that @xmath97 is independent of the choice of @xmath98 .
more generally , if we consider parabolic structures over finitely many disjoint divisors @xmath99 , for each @xmath49 one constructs the hyperplane bundle @xmath100 and grassmannian bundle @xmath101 over @xmath58 .
let @xmath79 be the fibre product of @xmath101 over @xmath58 and @xmath97 the fibre product of @xmath102 over @xmath58 .
@xmath36 let @xmath103 be the functor @xmath104 which associates to a scheme @xmath58 the set of equivalence classes of families @xmath105 of @xmath45-semistable gpbs of rank @xmath1 and degree @xmath2 with @xmath106 for all @xmath107 such that the section of @xmath79 defined by @xmath108 maps into @xmath97 .
one similarly defines a full subfunctor @xmath109 of @xmath103 with semistable replaced by stable .
let @xmath110 denote the set of semistable gpbs @xmath111 , where @xmath37 is a vector bundle of rank @xmath1 , degree @xmath2 with fixed determinant @xmath72 and @xmath112 is a subspace of @xmath113 of dimension @xmath1 with fixed weights @xmath114 . for @xmath115 , all gpbs in @xmath110 satisfy the condition
@xmath116 is surjective .
let @xmath117 denote the quot scheme of coherent quotients of @xmath118 with fixed hilbert polynomial determined by @xmath119 .
let @xmath120 be the nonsingular subvariety of @xmath117 corresponding to quotient vector bundles @xmath37 satisfying condition @xmath121 .
denote by @xmath122 the nonsingular closed subvariety in @xmath120 corresponding to @xmath37 with @xmath123 .
let @xmath124 be the universal quotient bundle . over @xmath120 , we have the fibre bundle @xmath79 with each fibre an @xmath125-fold product of @xmath126 as in 2.2 .
over @xmath122 , we have the fibre bundle @xmath127 whose fibres are @xmath125-fold products of hyperplane sections of @xmath126 .
then @xmath128 is a closed subvariety of @xmath79 .
let @xmath129 ( resp .
@xmath130 ) be the open set of stable ( resp .
semistable ) points in @xmath128 .
the git quotient @xmath0 of @xmath79 by pgl@xmath131 for a suitable polarization ( depending on @xmath45 ) is the coarse moduli space for gpbs @xcite .
it is a normal projective variety .
since @xmath128 is a pgl@xmath131-invariant closed subscheme ( subvariety ) of @xmath79 , the git quotient @xmath4 of @xmath132 by pgl@xmath131 is a closed subvariety of @xmath0 ( with a natural reduced subscheme structure ) .
the git quotient @xmath133 is an open subscheme of @xmath4 .
it is easy to see that @xmath4 ( resp .
@xmath134 ) is the coarse moduli space for the functor @xmath103 ( resp .
@xmath109 ) . let @xmath135 . then there is a coarse moduli space @xmath4 ( resp .
@xmath134 ) for the functor @xmath103 ( resp .
@xmath136 ) .
the moduli space @xmath4 is a projective ( irreducible ) variety , containing @xmath134 as an open subvariety .
let @xmath137 ( resp .
@xmath22 ) be the open subscheme of @xmath0 ( resp .
@xmath4 ) consisting of gpbs @xmath44 such that the projections @xmath50 are isomorphisms for all @xmath49 .
then @xmath22 corresponds to gpbs in @xmath0 with fixed determinant @xmath5 with @xmath138 and @xmath4 is the closure of @xmath22 .
[ [ section-1 ] ] let @xmath6 be an irreducible projective nodal curve with nodes @xmath11 and @xmath3 its desingularization with @xmath139 , the inverse image of @xmath14 .
then there is a correspondence from gpbs on @xmath3 of rank @xmath1 , degree @xmath2 to torsion - free sheaves on @xmath6 of the same rank and degree @xcite .
the correspondence induces a surjective morphism @xmath17 from @xmath0 onto @xmath15 , where @xmath15 is the moduli space of torsion - free sheaves of rank @xmath1 , degree @xmath2 on @xmath6 .
the restriction of the morphism @xmath17 to @xmath137 is an isomorphism onto the open subvariety @xmath16 of @xmath15 corresponding to vector bundles on @xmath6 .
one has @xmath140 . for @xmath141
, a gpb @xmath5 corresponds to a torsion - free sheaf @xmath19 on @xmath6 .
then @xmath19 is a line bundle if and only if @xmath142 .
suppose that @xmath19 is a line bundle and let @xmath20 be the closed subset of @xmath16 corresponding to vector bundles with fixed determinant @xmath19 .
then @xmath143 and the morphism @xmath17 maps @xmath22 isomorphically onto @xmath20 .
hence , if @xmath19 is a line bundle , then @xmath28 contains @xmath20 as an open subset . since @xmath4 is an irreducible , closed subscheme of @xmath0 , the image @xmath28 is a closed
, irreducible subscheme of @xmath15 and @xmath20 , being open , is dense in it .
it follows that @xmath28 is the closure of @xmath20 .
the projective scheme @xmath4 can be regarded as the compactified moduli space of vector bundles of rank @xmath1 with determinant @xmath19 on the nodal curve @xmath6 .
in fact , for any torsion - free sheaf @xmath19 of rank @xmath144 , the image @xmath28 is a closed , irreducible subscheme of @xmath15 containing the subset of @xmath15 consisting of torsion - free sheaves with fixed determinant @xmath19 as an open dense set . for a torsion - free sheaf @xmath26 of rank @xmath1 on @xmath6 , let @xmath145 , where ( torsion ) denotes the maximum subsheaf with proper support .
denote by @xmath25 the ideal sheaf of the node @xmath14 .
define @xmath146 as the set @xmath147 seshadri and nagaraj had defined this set for @xmath6 with one node and conjectured that if @xmath19 is a line bundle , then @xmath148 is the closure of @xmath20 ( conjecture ( a ) , page 136 of @xcite ) .
we prove this conjecture .
let @xmath149 denote the subset consisting of @xmath144-semistable points , then @xmath150 .
let @xmath151 be the moduli space of @xmath144-semistable gpbs [ 5 ] .
one has morphisms @xmath152 inducing @xmath153 and @xmath154 inducing @xmath155 .
@xmath36 let @xmath5 be any @xmath156 of rank @xmath144 on @xmath3 and @xmath19 the corresponding torsion - free sheaf of rank @xmath144 on @xmath6 . 1 . if @xmath157 , then @xmath158 and hence @xmath159 .
2 . the morphism @xmath160 surjects onto @xmath146 .
3 . @xmath161 for @xmath45 sufficiently close to @xmath144 .
we may assume that @xmath6 has only one node @xmath162 .
it is easy to see ( from the proof ) that the general case follows exactly on same lines .
consider a gpb @xmath163 .
let @xmath164 be the projections and @xmath165 .
let @xmath166 , then @xmath167 .
since @xmath168 , one has @xmath169 . hence to check that @xmath170 , we have only to check it locally at the node @xmath162 .
let @xmath171 be the local ring at @xmath162 .
its normalization @xmath172 is a semilocal ring with two maximal ideals @xmath173 .
the inclusion @xmath174 may be identified with the inclusion @xmath175 ( proposition 4.2 of @xcite ) . 1 .
we consider the following cases separately .
+ suppose that @xmath176 are both isomorphisms
. then @xmath177 corresponds to a torsion - free sheaf @xmath178 which is locally free at @xmath162 . in this case
, @xmath26 is locally free at @xmath162 with @xmath179 so that @xmath180 .
+ assume that @xmath181 is an isomorphism , @xmath182 is not an isomorphism ( the opposite case can be dealt with similarly ) .
then @xmath183 corresponds to @xmath19 which is not locally free at @xmath162 .
one always has @xmath184 and @xmath185 if and only if @xmath186 .
locally , @xmath187 so that @xmath188 .
then @xmath189 so that @xmath190 .
hence @xmath191 .
since @xmath192 , it follows that @xmath193 .
+ if both @xmath181 and @xmath182 are not isomorphisms , one has @xmath194 . then locally ,
it follows that @xmath196 maps to zero in @xmath197 so that @xmath198 . since @xmath199 , one has @xmath200 .
+ note that any @xmath201 with @xmath202 ( i.e. @xmath19 locally free at @xmath162 ) occurs only in cases
( i ) or ( iii ) . for @xmath203 with @xmath204 ( or @xmath92 ) only cases ( ii ) and ( iii ) occur .
part ( 1 ) now follows .
note that @xmath44 is @xmath144-semistable if and only if @xmath26 is semistable @xcite .
2 . let @xmath205 .
since @xmath206 it follows that @xmath207 where @xmath208 . since @xmath209 outside @xmath210 , one has @xmath211 outside @xmath210
it follows that @xmath212 .
let @xmath37 be given by an extension @xmath213 the composite @xmath214 induces a linear map @xmath215 .
let @xmath38 be the image of this linear map .
then @xmath44 maps to @xmath26 and it is @xmath144-semistable as @xmath26 is semistable @xcite . by construction , @xmath216 .
it follows that @xmath217 .
3 . let @xmath218 denote the closure of @xmath219 in @xmath151 . for @xmath45 close to @xmath144
, there is a surjective birational morphism @xmath220 with @xmath221 .
it maps @xmath4 birationally into @xmath218 .
since both these spaces are irreducible and of the same dimension , it follows that @xmath222 .
since @xmath223 surjects onto @xmath146 , it follows that @xmath224 and hence @xmath225 . from ( 1 ) , it follows that @xmath226 . @xmath227 @xmath36 if @xmath19 is a line bundle , then @xmath146 is the closure of @xmath20 . from proposition 3.3 ,
one has ( as sets ) @xmath228 . since @xmath28 is the closure of @xmath20 if @xmath19 is a line bundle , the result follows.@xmath227 the proof of proposition 3.3(1 ) easily generalizes to @xmath229 replaced by a family @xmath230 .
sun @xcite had proved the conjecture ( a ) of seshadri
nagaraj by considering a smooth curve @xmath3 degenerating to an irreducible nodal curve @xmath6 .
however he does not get a moduli functor or a scheme structure on @xmath146 except in some cases ( e.g. one node , rank @xmath31 , degree @xmath144 ) .
schmitt @xcite has constructed a moduli space @xmath231 of @xmath45-semistable descending singular sl@xmath232-bundles @xmath233 where @xmath234 is a gpb on @xmath3 and @xmath235 sym@xmath236 a nontrivial homomorphism .
it is shown that @xmath237 and for @xmath238 , there is a forgetful morphism @xmath239 ( 5.1 of @xcite ) . for @xmath45 close to @xmath144 ,
one has a forgetful morphism @xmath240 whose set theoretic image is @xmath146 ( proposition 5.1.1 of @xcite ) .
then , since @xmath241 , it follows that @xmath242 ( assets ) .
i would like to thank i biswas and a schmitt for useful comments and careful reading of the previous version . | let @xmath0 be the moduli space of generalized parabolic bundles ( gpbs ) of rank @xmath1 and degree @xmath2 on a smooth curve @xmath3 .
let @xmath4 be the closure of its subset consisting of gpbs with fixed determinant @xmath5 .
we define a moduli functor for which @xmath4 is the coarse moduli scheme . using the correspondence between gpbs on @xmath3 and torsion - free sheaves on a nodal curve @xmath6 of which
@xmath3 is a desingularization , we show that @xmath4 can be regarded as the compactified moduli scheme of vector bundles on @xmath6 with fixed determinant .
we get a natural scheme structure on the closure of the subset consisting of torsion - free sheaves with a fixed determinant in the moduli space of torsion - free sheaves on @xmath6 .
the relation to seshadri
nagaraj conjecture is studied .
[ theore]remark [ theore]proposition [ theore]lemma [ theore]definition [ theore]corollary [ theore]example |
we say a foliation @xmath3 of a compact @xmath0manifold @xmath4 is _
@xmath1covered _ if the pulled back foliation @xmath5 of the universal cover @xmath6 of @xmath4 is topologically the standard foliation of @xmath7 by horizontal @xmath8 s .
the first step in our construction is to produce a manifold @xmath4 with an @xmath1covered foliation @xmath3 which is not uniform .
the condition that a foliation be @xmath1covered is a somewhat elusive one , and in particular it does not seem to be sufficient to find a cover @xmath9 of @xmath4 so that the leaf space of @xmath10 is @xmath1 .
this is related to the question of when an infinite braid in @xmath7 is trivial .
a taut foliation @xmath3 of @xmath4 is @xmath1covered iff @xmath3 has no spherical or projective plane leaves , and for every arc @xmath11 between two points @xmath12 there is an arc @xmath13 homotopic to @xmath11 rel .
endpoints which is either contained in a leaf of @xmath3 or which is everywhere transverse to @xmath3 . if @xmath3 is @xmath1covered , lift @xmath11 to some @xmath14 in @xmath6 and make it transverse there .
if @xmath3 is not @xmath1covered , either @xmath4 is covered by @xmath15 or the leaf space @xmath16 of @xmath5 is a non - hausdorff simply connected @xmath17manifold .
this follows from a theorem of palmeira , which says that the foliation of @xmath7 by the universal cover of a taut foliation is topologically equivalent to a product of a foliation on @xmath8 and @xmath1 , and therefore such a foliation of @xmath7 by planes whose leaf space is @xmath1 is a product foliation ( see @xcite ) .
in particular , there are distinct leaves @xmath18 which can not be joined by an immersed path in @xmath16 .
that is , the topology of @xmath16 provides an obstruction to finding such a @xmath13 as above .
note that one does not have an analogue of palmeira s theorem for arbitrary open @xmath0manifolds that a foliation with leaf space @xmath1 should be a product and in fact this is not true .
for example , remove from @xmath7 foliated by horizontal @xmath8 s a properly embedded bi - infinite transverse curve which does not intersect every leaf .
this is where the difficulty resides in showing that a foliation is @xmath1covered by investigating an intermediate cover .
many ( most ? ) taut foliations of @xmath0manifolds are not @xmath1covered . in particular , by a theorem of fenley , an @xmath1covered foliation of a hyperbolic @xmath0manifold has the property that in the universal cover , every leaf limits to all of @xmath19 .
however , any compact leaf of a taut foliation which is not a fiber of a fibration over @xmath20 lifts to a quasi - isometrically embedded plane in the universal cover , and its limit set is a quasicircle ( see @xcite for a fuller discussion ) .
a taut foliation @xmath3 of a compact @xmath0manifold @xmath4 is _ uniform _ if in the pullback foliation @xmath5 of the universal cover @xmath6 , every two leaves @xmath21 are a bounded distance apart .
that is , there is some @xmath22 depending on @xmath21 so that @xmath23 is contained in the @xmath22neighborhood of @xmath24 , and vice versa .
a foliation @xmath3 is obtained from a _ slithering over @xmath20 _ if there is a fibration @xmath25 such that @xmath26 acts as bundle maps of this fibration , and such that the foliation of @xmath6 by components of the fibers of @xmath27 agrees with @xmath5 .
we will also refer to such a foliation , perhaps ungrammatically , as a slithering . for additional details and definitions , see @xcite .
it is shown in @xcite that a uniform foliation with every leaf dense is obtained from a slithering .
it is almost tautological from the definition of a slithering over @xmath20 that the action of @xmath26 on the leaf space @xmath16 of @xmath5 is conjugate to a representation in @xmath28 , the universal central extension of @xmath29 .
@xmath0manifold topologists will be familiar with the short exact sequence @xmath30 this sits inside the short exact sequence @xmath31 informally , @xmath32 is the group of homeomorphisms of @xmath1 which are periodic with period @xmath17
. let @xmath33 be the generator of the center of @xmath28 .
then @xmath33 acts on @xmath16 by translations , and by choosing an invariant metric on @xmath16 for this action , the action of every element of @xmath26 on @xmath16 is periodic with some period equal to the translation length of @xmath33 .
@xmath33 is known as _ the slithering map_. the condition that a foliation be uniform is reflected in the action of @xmath26 on @xmath16 in the following way : since leaves do not converge or diverge too much at infinity , holonomy can not expand or contract the leaf space too much .
every compact interval in @xmath16 can be included in a larger compact interval which is `` incompressible '' : no element of @xmath26 takes it to a proper subset or superset of itself .
if every leaf is dense , @xmath16 can be tiled with a countable collection of these incompressible intervals laid end to end , so that the slithering map @xmath33 acts on this tiling as a permutation .
note that a foliation @xmath3 may come from a slithering in many different ways .
for instance , if @xmath3 admits a nonsingular transverse measure then the leaf space of @xmath5 inherits an invariant measure making it isometric to @xmath1 .
then for any real @xmath34 there is a slithering @xmath35 defined by the composition @xmath36 where the last map is reduction mod @xmath37 .
since the action of @xmath26 on @xmath16 preserves the property of points being integral multiples of @xmath37 apart , this map is a slithering .
however _ generic _ foliations come from a slithering in essentially at most one way . the slithering map @xmath33 commutes with the action of every element of @xmath26
so , for instance , if an element @xmath11 acts on @xmath16 with isolated fixed points , the map @xmath33 must permute this fixed point set .
if , further , the action of @xmath26 on @xmath16 is minimal ( ie every leaf of @xmath3 is dense ) , the map @xmath33 is determined uniquely up to taking iterates .
that is , there is a _ minimal _
slithering @xmath25 with the property that for every other slithering @xmath38 determining the same foliation , there is a finite cover @xmath39 so that @xmath40 .
this theory is developed in @xcite .
following @xcite we define some auxiliary structure that will be used to show that certain foliations are uniform or @xmath1covered .
let @xmath41 transverse to @xmath3 be a vector field .
then @xmath41 is _ regulating _ if the lifts of the integral curves of @xmath41 to @xmath6 intersect every leaf of @xmath5 . these lifts determine a one dimensional foliation of @xmath6 . a leaf in this foliation and a leaf in @xmath5 intersect in exactly one point , and
consequently one can identify the leaf space of the one dimensional foliation with any fixed leaf of @xmath5that is , with @xmath8 .
such one dimensional foliations are called _ product covered _ in @xcite . the main point for our purposes of this structure is the following theorem : [ excise_loop ] suppose @xmath3 is uniform ( respectively @xmath1covered ) and @xmath41 is a transverse regulating vector field with a closed trajectory @xmath11 .
then the restriction of @xmath3 to @xmath42 is also uniform ( respectively @xmath1covered ) .
the universal cover @xmath6 of @xmath4 is foliated as a product by @xmath5 and the integral curves of @xmath41 give this the structure of a product @xmath43 in such a way that @xmath26 acts by elements of @xmath44 .
let @xmath45 be obtained from @xmath6 by removing the lifts of @xmath11 .
then @xmath45 is the cover of @xmath42 corresponding to the subgroup of @xmath46 normally generated by the meridian of @xmath11 .
one sees from the structure of @xmath41 that @xmath45 is foliated as a product by infinitely punctured disks and therefore that @xmath47 is @xmath7 foliated by horizontal @xmath8 s . if @xmath3 was uniform , any two leaves in @xmath45 would be a finite distance apart . but
leaves in @xmath47 correspond bijectively with leaves in @xmath45 under the covering projection and therefore the same is true in @xmath47 ; that is , the restriction of @xmath3 to @xmath42 is uniform .
notice from the construction that if @xmath3 came from a slithering , then the restriction of @xmath3 to @xmath42 comes from a slithering which agrees with the restriction of the slithering map on @xmath6 to the complement of the lifts of @xmath11 .
let @xmath48 denote the surface of genus @xmath49 with @xmath50 punctures .
then @xmath51 is the punctured torus , and @xmath52 .
let @xmath53 be standard generators for @xmath54 .
then we can choose a representation @xmath55 of @xmath56 by sending @xmath57 to translation through length @xmath37 and @xmath58 to some monotone element perhaps with a periodic collection of fixed points , each distance @xmath17 apart .
let @xmath59 be the trivial circle bundle @xmath60 over the punctured torus , and pick a flat @xmath29 connection on this bundle whose holonomy realizes the representation @xmath55 . note that after fixing a trivialization of the product , the representation is well - defined in @xmath28 and not just @xmath29 .
the distribution determined by this connection is integrable , by flatness , and integrates to give a foliation @xmath61 .
geometrically , there is a foliation of @xmath62 by leaves @xmath63 .
@xmath54 acts on this space by @xmath64 which preserves the foliation .
this foliation therefore descends to a foliation on @xmath65 transverse to the @xmath20 fibers .
this foliation of @xmath59 comes from an obvious slithering @xmath66 which is just projection onto the second factor followed by the covering map @xmath67 corresponding to the circles in @xmath59 .
the action of @xmath68 on the leaf space is exactly given by the representation @xmath55 in @xmath69 thought of as sitting in @xmath70 .
then the curve @xmath71 sits in @xmath59 transverse to @xmath61 , and there is also a foliation on @xmath72 which we also denote by @xmath61 .
it is easy to see that for irrational choice of @xmath37 the foliation @xmath61 has every leaf dense .
furthermore we have the following lemma : [ regulates ] the foliation @xmath61 of @xmath72 comes from a slithering .
furthermore , this slithering can be taken to be the restriction of @xmath73 to the complement of the lifts of @xmath57 in @xmath74 .
moreover , by choosing @xmath75 suitably generic , this slithering is minimal as defined above . by theorem [ excise_loop ] it suffices to show
there is a regulating vector field of @xmath59 which agrees with @xmath76 when restricted to @xmath11 . since @xmath59 is topologically just @xmath77 it has a projection map to @xmath51 .
let @xmath78 be the torus foliated by circles that is the preimage of the curve @xmath11 on @xmath51 under this projection .
then @xmath79 has an obvious codimension @xmath80 foliation by @xmath20 fibers .
we extend this foliation over @xmath81 , which can be parameterized as @xmath82 $ ] , by foliating each @xmath83 with parallel lines which rotate continuously from vertical ( parallel to the @xmath84 direction in @xmath59 ) on the boundary to horizontal ( parallel to @xmath11 ) on @xmath78 , always staying transverse to @xmath61 .
it is obvious that this is a foliation by regulating curves , and we denote its associated unit tangent vector field by @xmath85 .
if we choose @xmath75 to be generic and close to the identity with isolated fixed points , the slithering defined in the statement of the theorem is minimal . on another punctured torus with basis for @xmath86 given by @xmath87
we pick another representation @xmath88 in @xmath32 so that @xmath89 is translation through @xmath90 , where again @xmath90 is irrational and incommensurable with @xmath37 , and @xmath91 is some random element which commutes with @xmath33 but not with @xmath92
. then we can form @xmath93 foliated as above , and remove @xmath92 from @xmath94 to produce another foliated manifold with a slithering .
let @xmath4 be obtained by gluing @xmath95 and @xmath96 along the torus boundaries of neighborhoods of @xmath57 and @xmath92 respectively .
denote this torus in the sequel by @xmath97 .
each piece @xmath72 , @xmath96 admits a regulating vector field @xmath98 as constructed in lemma [ regulates ] .
we perform this gluing in such a way that the foliations of the boundary tori by meridional circles agree .
if we like , we can perform the gluing so that the restriction of the slithering maps for the left and the right foliations commute , when restricted to their action on the leaf space of the universal cover of @xmath99 .
then @xmath100 acts on @xmath1 by the amalgamated action of each piece on the leaf space of its respective universal cover , each canonically identified with the leaf space of the foliation of @xmath101 .
topologically , @xmath4 is a graph manifold obtained from four copies of @xmath102 . to see this ,
observe that @xmath79 ( with notation as above ) is exactly @xmath103 which is @xmath102 .
also , observe that @xmath104 $ ] in many different ways , including a way in which @xmath11 is @xmath105 .
it follows that @xmath106 is topologically also @xmath102 .
however , these foliations by circles can not be made to agree on the boundary tori of different pieces , and @xmath4 is not a seifert fibered space .
let @xmath3 denote the induced foliation of @xmath4 .
is @xmath3 @xmath1covered ? to establish that it is indeed @xmath1covered , it will suffice to show that each piece is uniform and admits a regulating transverse vector field @xmath98 which agree on the gluing torus to make a global regulating vector field @xmath41 .
@xmath3 is an @xmath1covered foliation of @xmath4 .
let @xmath9 foliated by @xmath10 be the cover of @xmath4 obtained by taking copies of the universal covers of @xmath59 and @xmath94 , drilling out countably many copies of the lifts of @xmath57 and @xmath92 , then gluing along the boundary components . then the regulating vector fields @xmath98 lift to @xmath9 to give a global trivialization of this manifold as a product of an infinite genus riemann surface with @xmath1 .
this implies that the universal cover of @xmath9 is the standard @xmath7 foliated by @xmath8 s , and we see therefore that @xmath3 is @xmath1covered .
since every leaf of @xmath3 is dense in @xmath4 , if @xmath3 were uniform it would come from a slithering by @xcite .
however , we have seen that the action of @xmath26 on the leaf space of the universal cover is the amalgamation of the actions of @xmath107 and @xmath108 along their gluing @xmath109 .
it follows that there is no single translation @xmath110 ( ie an element without fixed points ) which commutes with both @xmath58 and @xmath111 for sufficiently generic choice of @xmath91 and @xmath75 , since the periods of the left and right slithering maps are incommensurable .
more explicitly , let us fix a lift @xmath101 of the torus @xmath99 to @xmath6 which divides a piece of @xmath6 which is a lift of @xmath72 from a piece which is a lift of @xmath112 .
let us identify the leaf space @xmath16 with the leaf space of @xmath101 .
let @xmath113 corresponding to the longitude of @xmath99 preserve @xmath101 and act on @xmath16 as a translation . fix a metric on @xmath16 such that @xmath11 acts as translation through a unit length
let @xmath114 and @xmath115 be the translations in @xmath116 corresponding to the slithering map of @xmath72 and @xmath112 thought of as acting on the leaf space of @xmath101 .
then @xmath114 acts as translation through length @xmath117 and @xmath115 acts as translation through length @xmath118 . by minimality ,
the only translations in @xmath116 that commute with @xmath75 are multiples of @xmath114 , whereas the only translations that commute with @xmath91 are multiples of @xmath115 .
it follows that no translation commutes with both elements , and @xmath3 does not come from a slithering .
we have therefore proved the following theorem : @xmath3 as above is @xmath1covered but not uniform .
the following definition is from @xcite : on any manifold @xmath4 , a _ ( lorentz ) cone field _ @xmath119
transverse to a codimension one foliation @xmath3 is the field of timelike vectors ( ie with positive norm ) for a ( continuously varying ) form on @xmath120 of signature @xmath121 such that @xmath122 are spacelike . a cone field is _ regulating _ if every complete curve @xmath41 with @xmath123 is regulating for an @xmath1covered @xmath3 . regulating cone fields
are discussed in @xcite , and shown to exist for all uniform foliations .
we show that the example constructed in the previous section admits a regulating cone field .
each piece @xmath72 , @xmath124 admits a regulating cone field @xmath125 which is degenerate along the boundary torus , coming from the restriction of the regulating cone fields on @xmath126 which are tangent to @xmath57 and @xmath92 .
let @xmath119 denote the cone field on @xmath4 which agrees with @xmath127 and @xmath128 away from a collar of the separating torus , and which near this separating torus is sufficiently thin so that every curve which crosses this collar must wind a distance at least @xmath129 , as measured in periods of the longitude , transverse to the foliation .
@xmath119 as defined above is a regulating cone field for @xmath3 .
we will do our calculations in @xmath9 , using the fact that @xmath9 admits a regulating vector field coming from @xmath98 which preserves each uniform piece .
observe that @xmath9 has a decomposition into a countable collection of covering spaces of @xmath95 , @xmath96 which we denote @xmath130 for some particular choice of indices @xmath131 . also label the separating cylinders , all of them lifts of the gluing torus in @xmath4 , as @xmath132 for some index @xmath131 .
let @xmath133 denote the metric on @xmath16 , the leaf space of @xmath10 , given by the transverse measure on @xmath134 scaled so that the curves @xmath135 have period @xmath17 and so that the left and right slithering maps act by translations of this measure , as measured in the cylinder @xmath132 .
then we claim the following lemma : if @xmath132 and @xmath136 are separated by @xmath50 cylinders @xmath137 then for any two leaves @xmath18 , @xmath138 the proof follows immediately by induction once we show the result for @xmath139 . if @xmath132 and @xmath136 bound a single @xmath140 or @xmath141 , then the fact that the pieces @xmath140 and @xmath141 slither over @xmath20 implies that @xmath142 and @xmath143 differ by at most one period of the slithering , which in terms of the measure on the boundary torus , is either @xmath117 or @xmath118 depending on whether we are in an @xmath140 or an @xmath141 .
let @xmath144 .
now we will show that any curve @xmath145 supported by @xmath119 makes definite progress relative to any given @xmath133 , and therefore relative to all of them .
in particular , it is regulating .
after re - ordering indices , any such infinite curve in @xmath9 , starting on some leaf @xmath146 , can be broken up into segments @xmath147 where each @xmath148 is contained in @xmath140 or @xmath141 ( according to sign ) .
it is clear that if there are only finitely many @xmath148 ( ie @xmath145 crosses only finitely many separating annuli ) then eventually @xmath145 can be seen to be regulating , since it is supported by some lift of @xmath128 or @xmath127 .
so we suppose there are infinitely many @xmath148 .
let @xmath149 be the union of the first @xmath131 segments of @xmath145 .
let @xmath150 be the _ shadow _ of @xmath149 on @xmath23 ; that is , the curve on @xmath23 obtained by projecting @xmath9 to @xmath23 along the integral curves of @xmath41 .
let @xmath151 be the integral curve of @xmath41 interpolating between the endpoint of @xmath150 and the endpoint of @xmath149 .
then @xmath152 for all @xmath153 since the curves @xmath154 and @xmath155 have endpoints on the same pair of leaves .
we can estimate @xmath156 by hypothesis on @xmath119 .
it follows that @xmath157 .
but then @xmath158 and so @xmath159 . continuing in this way and by induction , we get @xmath160 .
but then @xmath161 and one can see that by choosing @xmath162 the curve @xmath145 makes arbitrary progress relative to some _ fixed _ @xmath133 , and therefore is regulating .
an instructive analogy is given by the comparison between imperial and metric weights and measures : suppose i have a small object which i can measure to the nearest inch or to the nearest centimeter
. then i get estimates which vary greatly compared to the length of the object .
if the object is much bigger , the estimates are comparatively better .
my regulating curve above makes definite progress , even though its progress is translated into inches , then centimeters , then inches , then centimeters rounding down every time .
[ t : lorentz ] suppose @xmath4,@xmath3 is any compact oriented @xmath0manifold with a co - orientable @xmath1covered foliation , and suppose that @xmath3 admits a transverse regulating lorentz cone field @xmath119 .
let @xmath145 be any simple closed curve supported by @xmath119 .
if @xmath163 is obtained by taking an @xmath50fold branched cover over @xmath145 , and @xmath164 denotes the pullback foliation , then @xmath164 is @xmath1covered .
moreover , @xmath164 is uniform iff @xmath3 is .
the point of having a regulating cone field @xmath119 is that for any @xmath145 supported as above , there is a regulating vector field @xmath41 of @xmath4 so that @xmath165 .
this follows immediately from obstruction theory , once one notices that sections of @xmath119 are contractible ; eg use a partition of unity . now in @xmath6 , the universal cover of @xmath4 , @xmath145 lifts to a collection of bi - infinite regulating curves , and @xmath166 is the universal orbifold cover of @xmath6 where we declare that there are order @xmath50 cone angles along the lifts of @xmath145 .
let @xmath167 be the pullback foliation in that universal orbifold cover and let @xmath21 be two leaves there .
they can be joined by some arc @xmath11 in the complement of the lifts of the cone locus , which projects to an arc @xmath168 in @xmath6 . by homotoping @xmath168 rel .
endpoints along integral curves of @xmath41 , we can make it transverse to @xmath5 _ without crossing any lift of @xmath145_. then this perturbed @xmath168 lifts to a perturbed @xmath11 in @xmath166 transverse to @xmath167 , thereby demonstrating that @xmath164 is @xmath1covered , as required .
if @xmath3 was not uniform , there would be a pair of leaves in @xmath5 which diverge at infinity .
they lift to leaves with the same property in @xmath167 .
alternatively , the uniformity or lack thereof can be seen in the action of @xmath169 on @xmath1 .
we return to the example @xmath4 that we constructed in the previous section . in @xmath4 , it is clear that we can choose a curve @xmath145 supported by @xmath119 whose complement is hyperbolic . for
, we can certainly do this in each side of @xmath4 , and then by crossing back and forth across the separating torus , we can arrange for the complement of @xmath145 to be atoroidal . in particular , by choosing @xmath75 and @xmath91 to be sufficiently close to translations , the regulating cone fields @xmath127 and @xmath128 can be as `` squat '' as we like , and we have a great deal of freedom in our choice of the restriction of @xmath145 to the complement of a collar neighborhood of @xmath99 .
for example , if @xmath170 are chosen so that every element acts on the leaf space as a translation through a rational distance , @xmath3 would be a surface bundle over a circle ; in a product bundle , a curve whose projection to @xmath20 is a homeomorphism and whose projection to the base surface fills up the surface ( ie complementary regions for the geodesic representative are disks ) has atoroidal complement .
similar curves are easily found in any surface bundle , and one can arrange for them to wind several times in the circle direction when they pass through some reducing torus .
the point is that any sufficiently complicated curve will suffice .
then for nearby choices of @xmath171 , such a curve will still be regulating and contained in the regulating lorentz cone field , as we show in the following lemma : suppose @xmath172 is a transversely measured foliation of @xmath72 as above coming from some representation @xmath173 , and let @xmath174 be a transverse lorentz cone field for @xmath172 appropriately degenerate near @xmath57 . then for slitherings @xmath61 coming from sufficiently close choices of @xmath75 , the cone field @xmath174 is regulating for @xmath61 .
we need to check for sufficiently mild perturbations @xmath61 of @xmath172 that any curve supported by @xmath174 passes through at least one period of the slithering of @xmath61 , since then it must pass through arbitrarily many such periods and therefore be regulating .
observe first that any transverse lorentz cone field is regulating for a transversely measured foliation , since one can uniformly compare distance along a curve supported by the cone field and distance with respect to the transverse measure . for @xmath61 sufficiently close to @xmath172
, @xmath174 is a transverse lorentz cone field for @xmath61 .
the codimension @xmath80 foliation @xmath85 described in lemma [ regulates ] is regulating for every choice of @xmath75 , and we assume that this foliation lifts to the vertical foliation of @xmath7 by @xmath175 .
we choose co - ordinates on @xmath7 so that the regulating curves are of the form @xmath176 , @xmath177 and the leaves of @xmath178 are of the form @xmath179 the slithering on @xmath72 comes from the slithering on @xmath59 associated with the circle bundle as above .
we assume that the circles of the seifert fibration away from @xmath81 lift to vertical arcs of length @xmath17 . with this choice of co - ordinates , for any foliations @xmath61 constructed from a representation as above and for every @xmath180 ,
the points @xmath181 and @xmath182 in the universal cover are on leaves which are one period of the slithering apart . for a point @xmath183 in @xmath74
given in co - ordinates by @xmath181 , let @xmath184 be the leaf of @xmath185 through @xmath183 , and @xmath186 the leaf of @xmath185 through @xmath182 .
then the leaves @xmath184 and @xmath186 differ by a translation parallel to the @xmath187-axis of length @xmath17 .
note that this translation need in no way correspond to the action of an element of @xmath68 on @xmath74 .
the light cone of @xmath174 through @xmath183 intersects the horizontal plane passing through @xmath182 in a compact region . for sufficiently small perturbations of @xmath75 , the leaf
@xmath186 will be a small perturbation of the horizontal plane through @xmath183 , and the intersection of the light cone of @xmath174 through @xmath183 with @xmath186 will also be a compact region .
now , @xmath72 is not compact , but we can consider its double @xmath45 , foliated by the double of @xmath172 , and equipped with a transverse lorentz cone field obtained by doubling @xmath174 which is degenerate along the boundary components of @xmath72 .
the compactness of @xmath45 implies that for a sufficiently small perturbation @xmath172 of @xmath61 , the intersections as above will be compact for all @xmath183 .
this implies that a curve supported by @xmath174 will need to go only a bounded distance before traveling a full period of the slithering .
in particular , any bi - infinite curve supported by @xmath174 will travel through infinitely many periods of the slithering in either direction and will therefore be regulating , which is what we wanted to show .
what is really essential to notice in the above set up is that the leaves passing through @xmath181 and @xmath182 in the universal cover were one period of the slithering away from each other for _ all _ @xmath61 . in some sense
, all the slitherings are determined by the structure of the seifert fibration where they originated . for a generic perturbation of a uniform foliation ,
one has no control over how the slithering map may vary , or even whether the perturbed foliations are uniform at all .
since for a transversely measured foliation , _ any _ transverse lorentz cone field is regulating , we can choose our curve @xmath145 to be any curve transverse to a measured foliation @xmath172 with hyperbolic complement which is sufficiently steep near the separating torus , and then choose the representation @xmath75 to be sufficiently close to a translation that @xmath145 stays in a regulating cone field . since @xmath188 is hyperbolic , for sufficiently large
@xmath50 an @xmath50fold branched cover of @xmath4 over @xmath145 is hyperbolic .
one can not be sure that an @xmath50fold branched cover will always exist , but at least one has a hyperbolic orbifold structure on @xmath4 with cone angle @xmath189 along @xmath145 , and by selberg s lemma ( see @xcite ) one knows this is finitely ( orbifold ) covered by a genuine hyperbolic manifold which is a branched cover of @xmath4 along @xmath145 . by the discussion above
, the induced foliation is @xmath1covered but not uniform .
moreover , by choosing @xmath75 and @xmath91 sufficiently close to translations , the foliation @xmath3 is as close to a transversely measured foliation as we like . passing to a branched cover preserves this fact .
since transversely measured foliations of @xmath0manifolds are arbitrarily close ( as @xmath80plane fields ) to surface bundles over @xmath20 , we have proved : there exist foliations of hyperbolic @xmath0manifolds which are @xmath1covered but not uniform .
they can be chosen arbitrarily close to surface bundles over circles .
it is clear that the construction outlined above can be made in some generality .
one can construct @xmath1covered but not uniform foliations by plumbing together uniform foliations along boundary tori in numerous ways . in a great number of cases
, these will admit regulating transverse cone fields , and by branching as above one can produce many atoroidal examples .
one can easily arrange for these examples to be closed ; for instance , by doubling @xmath4 before removing a curve with atoroidal complement in the examples constructed above .
another construction , explained in detail in @xcite , involves choosing a train track with integer weights supported by a regulating cone field , then plumbing the leaves along the train track with a surface of genus given by the track weight .
these plumbed surfaces can be `` twisted '' by a surface automorphism under the monodromy around loops of the train track .
thurston expects that these examples are sufficiently flexible to allow one to prescribe the homology of @xmath4 .
we show in @xcite that the circles at infinity of each leaf in the universal cover of @xmath4 can be included in a topological _ cylinder at infinity
_ @xmath190 on which @xmath26 acts by homeomorphisms . with notation as above , there is a global trivialization of @xmath190 as @xmath191 so that the action of @xmath26 preserves the horizontal and vertical foliations of @xmath190 by @xmath192 and @xmath193 . in @xcite it is suggested that _ all _ taut foliations should have a pair of essential laminations @xmath194 transverse to each other and to the foliation which intersect each leaf in a ( @xmath17dimensional ) geodesic lamination . in the case of @xmath1covered foliations , these laminations should come from a pair of @xmath17dimensional laminations @xmath195 of the universal @xmath20 described in the theorem above which are invariant under the action of @xmath26 .
let the universal @xmath20 bound a hyperbolic plane @xmath196 , and let @xmath195 be the associated geodesic laminations in @xmath196 . then since each circle at infinity is canonically associated with this @xmath20 , each point in @xmath197 determines a unique point in each leaf of @xmath5 .
similarly , each segment of @xmath198 between points of intersection determines a unique geodesic segment in each leaf , and each complementary region determines a unique geodesic polygon in each leaf .
if one fixes some canonical geometric parameterization of the family of convex geodesic polygons of a fixed combinatorial type , this parameterization gives rise to a canonical identification of each leaf with each other leaf , preserving the stratification outlined above .
the fibers of this identification give a one - dimensional foliation transverse to @xmath5 , and the tangent vectors to this foliation are a regulating vector field .
it is easy to see that non - quadrilateral complementary regions in @xmath196 give rise to solid cylinders in @xmath6 which cover solid tori in @xmath4 , since their cores are isolated .
the cores of these solid tori are necessarily regulating , and therefore define elements of @xmath26 which act on the leaf space by translation .
in fact , thurston has communicated to this author a sketch of a proof that an @xmath1covered foliation admits _ some _ transverse lamination which intersects every leaf in geodesics ( though not necessarily a pair of such ) . in @xcite
we can show by methods slightly different to those above that this assumption is enough to imply that there exists a regulating vector field transverse to @xmath3 which can be chosen to have closed orbits . in short , we have the following theorem : let @xmath3 be an @xmath1covered foliation . then there exists a regulating vector field transverse to @xmath3 which can be taken to have closed orbits .
these orbits determine elements @xmath199 which act on the leaf space of @xmath5 without fixed points . branched
covers of @xmath4 over these closed orbits give new @xmath1covered foliations .
one hopes these results are all pieces of a unified picture tying the intrinsic geometry of @xmath1covered foliations to the extrinsic geometry of the foliated manifolds containing them .
not all the details of this picture are yet visible .
nevertheless , it seems worthwhile to make this picture as explicit as possible .
let @xmath4 be a hyperbolic surface bundle over a circle with fiber @xmath200 and pseudo - anosov monodromy @xmath201 .
let @xmath3 be the induced foliation by surfaces .
let @xmath145 be a simple closed curve on @xmath200 so that @xmath202 .
note that it is easy to show that there exist such examples , by first choosing any @xmath203 and then using the fact that surface groups are lerf to lift to a finite regular cover where @xmath145 and its image are disjoint .
let @xmath9 be the @xmath204cover of @xmath4 defined by the circle direction .
topologically , @xmath9 is @xmath205 foliated by closed surfaces @xmath206 .
let the group @xmath204 generated by the deck translation , which we denote @xmath207 , act by @xmath208 .
let @xmath209 be the annulus @xmath210 $ ] . then by construction , @xmath209 and its translates are disjoint .
let @xmath211 \to [ -\epsilon,1 + \epsilon]$ ] be a homeomorphism close to the identity which moves every point except the endpoints up some small amount .
then we can cut open @xmath9 along @xmath209 and its translates , and shear the foliation on one side by the translation @xmath212 to get a new foliation @xmath213 .
this can certainly be done in such a way that @xmath213 is arbitrarily close to @xmath3 .
let us suppose that by choosing a suitable path in @xmath213 , we can arrange that a curve starting at some @xmath214 can get to @xmath215 in length @xmath37 , as measured in @xmath213 , by winding sufficiently many times through @xmath209 .
now , the curve can continue to wind around @xmath216 , and after moving a distance @xmath217 , it can reach @xmath218 , and so on .
remember that there is a transverse regulating vector field @xmath41 to @xmath3 given by tangents to the curves @xmath175 .
since @xmath219 is pseudo - anosov , when we compare arclength at @xmath220 and @xmath221 by projection along @xmath41 , we see that a vector of norm @xmath222 at @xmath220 might project to a vector of norm @xmath17 at @xmath221 , where @xmath23 is the multiplier of @xmath219 on the invariant transverse measure of the unstable lamination of @xmath200 .
hence , as measured in @xmath213 , a curve @xmath145 beginning at @xmath221 could have length @xmath223 but its projection to @xmath224 could have length as little as @xmath225 .
in particular , a curve in @xmath213 could `` escape to infinity '' while its projection to @xmath226 could move only a finite distance . by picking two points @xmath221 and @xmath227 sufficiently far apart , and moving them by curves in @xmath213 joined by integral curves of @xmath41
, it seems plausible that one could find a path in @xmath213 where holonomy was not defined after some finite time , suggesting that @xmath213 was not @xmath1covered . of course , there are problems with making this concrete : distances in @xmath213 are only magnified in the direction of the stable lamination as we go upwards ; perhaps to make the curves cross through @xmath209 and its translates sufficiently often , we need to go in both stable and unstable directions .
moreover , even if one could show that holonomy failed to be defined for all time along integral curves of @xmath41 , it does not rule out the possibility that @xmath213 is still @xmath1covered and @xmath41 is merely not regulating , although the author understands that very recently s ' ergio fenley has shown that a pseudo - anosov flow transverse to an @xmath1covered foliation should always be regulating ( @xcite ) .
a similar result is also proved in @xcite . in writing this paper
i benefited from numerous helpful conversations with andrew casson , s ' ergio fenley and bill thurston .
in particular , many of the ideas contained here are either implicit or explicit in the wonderful paper @xcite . while writing this paper , i was partially supported by an nsf graduate fellowship .
, _ the geometry of @xmath1covered foliations i _ ,
math.gt/9903173 , * d long * , _ foliations of some @xmath0manifolds which fiber over the circle _ , proc .
( 1998 ) 925931 , _ quasi isometric foliations _ , topology , 31 ( 1992 ) 667676 , private communication , _ open manifolds foliated by planes _
, annals of math . 107 ( 1978 ) 109131 , _ on discontinuous groups in higher - dimensional symmetric spaces _ , from : `` contributions to function theory '' , ( k chandrasekharan editor ) tata inst . of fund .
research , bombay ( 1960 ) 147164 , _ three - manifolds , foliations and circles i _ , math.gt/9712268 , _ three - manifolds , foliations and circles ii _ | we produce examples of taut foliations of hyperbolic @xmath0manifolds which are @xmath1covered but not uniform ie the leaf space of the universal cover is @xmath1 , but pairs of leaves are not contained in bounded neighborhoods of each other .
this answers in the negative a conjecture of thurston in @xcite .
we further show that these foliations can be chosen to be @xmath2 close to foliations by closed surfaces .
our construction underscores the importance of the existence of transverse regulating vector fields and cone fields for @xmath1covered foliations .
finally , we discuss the effect of perturbing arbitrary @xmath1covered foliations . |
the spherical model of magnetism has been used as an excellent laboratory to test ideas and concepts of phase transitions and critical phenomena @xcite@xcite@xcite .
there are several versions of the original model , including proposals of a quantum spherical model to correct some of the unphysical results at low temperatures @xcite@xcite@xcite@xcite .
the effects of frustration @xcite@xcite , random fields @xcite , and of disordered exchange interactions @xcite@xcite , have also been analysed in the context of quantum spherical models . with a view to describe the crossover between classical and quantum critical behaviour ,
vojta @xcite used a standard scheme of canonical quantization to analyse a quantum version of the ferromagnetic mean spherical model .
we were then motivated to revisit this problem , with the addition of competing ferro and antiferromagnetic interactions , and the perspective to analyse a quantum lifshitz point . the mean spherical model , which has been originally proposed by lewis and wannier @xcite , is given by the partition function @xmath14 \exp\left [ -\beta\,\mathcal{h}% \left\ { s_{\overrightarrow{l}}\right\ } -\beta\mu\,{\sum \limits_{\overrightarrow{l}}}s_{\overrightarrow{l}}^{2}\right ] , \ ] ] where @xmath15 , @xmath0 is the temperature and @xmath16 is the boltzmann constant , @xmath17 is a suitable chemical potential , @xmath18 is a lattice vector , and @xmath19 is a set of continuous spin variables running over the @xmath20 sites of a @xmath4-dimensional hypercubic lattice . the model hamiltonian is written as@xmath21 where @xmath22 labels a pair of lattice sites , the exchange parameter @xmath23 depends on the distance between sites @xmath24 and @xmath18 , and @xmath25 is an external field . in this formulation ,
the chemical potential @xmath17 comes from the mean spherical condition,@xmath26 and it is well known that exact solutions for the thermodynamic functions can be obtained from the standard diagonalization of a quadratic form @xcite . in a quantum version of this mean spherical model @xcite@xcite ,
the spin variable @xmath27 becomes a position operator at lattice site @xmath18 , canonically conjugate to a momentum operator @xmath28 , with the commutation relations@xmath29=0,\quad\lbrack p_{\overrightarrow{l}},p_{\overrightarrow{k}}]=0,\quad\lbrack s_{\overrightarrow{l}},p_{\overrightarrow{k}}]=i\delta_{\overrightarrow { l},\overrightarrow{k}},\ ] ] where @xmath30 is a kronecker delta and we assume that @xmath31 .
we then add a term of kinetic energy , depending on a quantum parameter @xmath1 , and write the quantum quadratic form@xmath32 which can be diagonalized by a canonical method @xcite , leading to a solution of the problem for a general ferromagnetic pair interaction . at finite temperatures ,
the critical behaviour is essentially unchanged with respect to the classical spherical model . at zero temperature ,
depending on the parameter @xmath1 , there is a quantum phase transition characterized by new ( quantum ) critical exponents .
also , the introduction of quantum fluctuations leads to a correction of the unphysical behaviour of the entropy at low temperatures .
we report an analysis of this version of the quantum mean spherical model in the presence of competing interactions .
we consider ferromagnetic interactions , @xmath3 , between pairs of first - neighbour sites along the @xmath4 directions of a hypercubic lattice , and antiferromagnetic interactions , @xmath5 , between second - neighbour sites along @xmath6 directions .
classical versions of this model @xcite@xcite@xcite@xcite , as well as more elaborate mean spherical models with competing interactions @xcite , have been studied by several authors . for @xmath33 , we regain a spherical analogue of the axial - next - nearest - neighbour ising , or annni , model @xcite@xcite , which is known to display a rich phase diagram , including a lifshitz point , in terms of the temperature @xmath0 and a parameter @xmath34 that gauges the strength of the competing interactions .
we then analyse the @xmath35 phase diagram , for different values of @xmath36 , in particular the @xmath10 behaviour , and establish the critical dimensions and critical exponents associated with this quantum model system . in the classical case ,
@xmath7 , we confirm a number of results , including a singularity of the critical border at the lifshitz point for @xmath37 . in one dimension
, we derive analytic expressions for the decay of pair correlations , and determine the region of modulated behaviour in the @xmath38 phase diagram .
this problem can be treated either by a conventional reduction to a system of coupled harmonic oscillators or by a judicious application of the method of path integrals @xcite@xcite@xcite .
let us first use the representation in terms of harmonic oscillators .
we then introduce bosonic operators @xmath39 and @xmath40 to write@xmath41 and @xmath42 where we have omitted the vector notation .
we now assume periodic boundary conditions , and change to a fourier representation,@xmath43 where @xmath44 and @xmath45 are bosonic operators,@xmath46=0,\quad\lbrack a_{q}^{\dag},a_{q^{\prime}}% ^{\dag}]=0,\quad\lbrack a_{q},a_{q^{\prime}}^{\dag}]=\delta_{q , q^{\prime}},\ ] ] and the sum is over the @xmath4-dimensional vectors @xmath47 belonging to the first brillouin zone .
we then write the quantum quadratic form ( [ hquantum ] ) in the fourier representation , @xmath48 a_{q}^{\dag}a_{q}-\frac{1}{2}(2g\mu)^{1/2}% \sum_{q}\frac{\hat{j}(q)}{4\mu}\left ( a_{q}a_{-q}+a_{q}^{\dag}a_{-q}^{\dag } \right ) \nonumber\\ & -h\left ( \frac{n^{d}}{2}\right ) ^{1/2}\left ( \frac{g}{2\mu}\right ) ^{1/4}\left ( a_{0}+a_{0}^{\dag}\right ) + \frac{n^{d}}{2}\left ( 2g\mu\right ) ^{1/2},\label{hbarq}%\end{aligned}\ ] ] with@xmath49 where the sum runs over all lattice vectors .
the final diagonalization of this quadratic form comes from the introduction of new bosonic operators , @xmath50 and @xmath51 , according to a well - known bogoliubov transformation .
we then have @xmath52 } , \label{diagonal}%\ ] ] where@xmath53 ^{2}\equiv(2g\mu)\left [ 1-\frac{\hat { j}(q)}{2\mu}\right ] , \label{freq}%\ ] ] which requires that the chemical potential @xmath17 should be larger than a certain limiting critical value @xmath54,@xmath55 in analogy with a system of harmonic oscillators , we then write the partition function@xmath56 \prod_{q}\left [ 2\sinh\frac{1}{2}\beta w\left ( q\right ) \right ] ^{-1},\label{partfunc}%\ ] ] and the free energy per site,@xmath57 } % + \frac{1}{\beta}\lim_{n\rightarrow\infty}\frac{1}{n^{d}}\sum_{q}\ln\left [ 2\sinh\frac{1}{2}\beta w\left ( q\right ) \right ] , \label{energy}%\ ] ] with @xmath58 . the spherical constraint , given by eq .
( [ sphconstr ] ) , from which we determine the chemical potential @xmath17 , is written as@xmath59 ^{2}}+\lim _ { n\rightarrow\infty}\frac{1}{n^{d}}\sum_{q}\frac{g}{2w\left ( q\right ) } \coth\left [ \frac{1}{2}\beta w\left ( q\right ) \right ] .\label{constraint}%\ ] ] in the classical limit , @xmath60 , we regain most of the well - known results for the mean spherical model .
the classical limit of this quantum free energy , however , includes an extra term of the form @xmath61 , which corrects the classical behaviour at low temperatures .
the particular limit @xmath62 corresponds to free quantum ( spherical ) rotors , with a finite energy gap , in contrast to the usual heisenberg - dirac spins .
these expressions also come from a straightforward application of the path integral formalism , which has been widely used to treat quantum statistical problems @xcite@xcite .
it is then interesting to write the lagrangian associated with this problem,@xmath63 in the imaginary time formalism , with @xmath64 , the partition function is written as@xmath65 \right\ } , \end{aligned}\ ] ] where the first integral includes periodic conditions , @xmath66 .
we now introduce the fourier transformation , @xmath67 s_{q}(w),\ ] ] where the vector @xmath47 belongs to the first brillouin zone , and @xmath68 , with integer @xmath69 , is a matsubara frequency .
we then write the partition function@xmath70 s_{q}(w_{n})s_{-q}(-w_{n})\right\ } , \end{aligned}\ ] ] where @xmath71 is given by eq .
( [ jq ] ) , and we have omitted the vector notation . if we calculate the gaussian integrals , and use the identity@xmath72 = \frac{\sinh w\beta}{w\beta},\ ] ] it is straightforward to regain the partition function given by eq .
( [ partfunc ] ) .
this general solution works for all forms of distance - dependent interactions , @xmath73 .
we now consider ferromagnetic interactions , @xmath3 , between pairs of first - neighbour sites along the @xmath4 directions of a hypercubic lattice , and antiferromagnetic interactions , @xmath5 , between second - neighbour sites along @xmath6 directions .
the fourier transform of the exchange interactions is given by @xmath74 where @xmath75 is a wave vector in suitable ( dimensionless ) units .
the maximum of @xmath76 depends on the ratio @xmath2 .
if @xmath77 , the maximum is located at the critical value @xmath78 , as in the simple ferromagnetic case . if @xmath79 , the maximum of @xmath76 is given by the critical vector@xmath80 where@xmath81 the special case @xmath8 corresponds to a lifshitz point of degree @xmath36 .
in zero field , @xmath82 , the paramagnetic critical boundary in the @xmath35 space comes from the spherical constraint , given by eq .
( [ constraint ] ) , supplemented by the critical limit of the chemical potential , @xmath83 , given by eq .
( [ mucrit ] ) .
we then write@xmath84 , \label{constraint2}%\ ] ] where@xmath85 ^{1/2}.\label{freq2}%\ ] ] in the classical limit , @xmath60 , we have@xmath86 } , \ ] ] from which we obtain the critical temperature as function of @xmath87 , for all values of @xmath4 and @xmath36,@xmath88 with @xmath89 for @xmath90 , and @xmath91 for @xmath79 .
there is a long history associated with the calculations of similar lattice green functions @xcite@xcite . from the identity@xmath92 where @xmath69 is an integer , and @xmath93 , we write integral representations for @xmath94 , which are convenient to carry out an asymptotic analysis . for @xmath95 , the divergence of these integrals indicate that @xmath96 for all values of @xmath36 and @xmath97 .
also , we have @xmath98 for @xmath9 , and for all values of @xmath36 and @xmath97 . in particular , at the lifshitz point , @xmath98 for @xmath99 .
we now consider the graphs of @xmath100 versus the parameter @xmath87 .
it is easy to write an expression for @xmath101 , for @xmath90 and @xmath79 , and to show that there is common tangent at the lifshitz point , @xmath8 .
for example , for @xmath102 and @xmath33 , we have@xmath103 ^{2}}\widetilde{i},\ ] ] with@xmath104 we then use the identity ( [ identity1 ] ) , with @xmath105 , to write an integral representation from which it is easy to show that the common tangent at the lifshitz point , @xmath8 , is infinite for ( i ) @xmath106 , and ( ii ) @xmath9 , with @xmath107 , which includes the analogue of the annni model ( @xmath108 and @xmath33 ) . in the numerically obtained graphs of figure 1 ,
we sketch typical profiles of the critical line @xmath100 for dimensions @xmath108 and @xmath109 , and @xmath33 .
note the smooth behaviour of this paramagnetic border for @xmath109 ( and @xmath33 ) .
the scale of this figure , however , is not enough to show the sharp singularity at the lifshitz point for @xmath108 ( and @xmath33 ) , as pointed out in a sketch by hornreich @xcite .
[ ptb ] fig_classical.eps in the zero - temperature limit , @xmath110 , eq . ( [ constraint2 ] ) can be written as@xmath111 ^{1/2}},\ ] ] from which we obtain the critical quantum parameter , @xmath112 , as a function of @xmath87 , for all values of @xmath4 and @xmath36,@xmath113 ^{2},\ ] ] where@xmath114 ^{1/2}},\ ] ] for @xmath90 , and@xmath115 ^{1/2}},\ ] ] for @xmath79 .
we now use an analytic continuation of the identity ( [ identity1 ] ) , for non integer values of @xmath69 . from a similar analysis of convergence of these expressions ,
it is easy to show that there is a quantum phase transition ( @xmath116 ) for @xmath12 , independent of the value of @xmath36 . in particular
, there is a common derivative @xmath117 at the lifshitz point , @xmath8 , with a singularity for @xmath118 ( and a smooth behaviour for @xmath108 and @xmath33 ) . in figure 2 , we sketch typical profiles of this critical line in the @xmath119 plane for dimensions @xmath120 and @xmath108 , and @xmath33 .
[ ptb ] fig_quantum.eps the critical behaviour in zero field , @xmath82 , comes from an asymptotic analysis of the spherical constraint in the neighbourhood of the transition . at finite temperatures , @xmath121 , in the limit @xmath122 , we have@xmath123.\label{sumfin}%\ ] ] at zero temperature , @xmath10 , we have@xmath124 we now expand @xmath71 as a taylor series about @xmath125 , in the classical and quantum cases .
although @xmath126 depends on the parameter @xmath87 , the convergence of the sums in the right - hand side of eqs .
( [ sumfin ] ) and ( [ sumzero ] ) does not depend on @xmath87 , for @xmath97 . for finite temperatures , @xmath121 ,
the sum converges if @xmath9 ( which determines the lower critical dimension of the classical case ) , regardless of the value of @xmath36 . at @xmath10 ,
the sum converges for @xmath127 , which leads to the lower critical dimension of the quantum case .
let us consider some special situations . in zero field , @xmath82 , at finite temperatures , @xmath121 , and for @xmath97 , we use eq .
( [ sumfin ] ) and perform a taylor expansion about @xmath128 and @xmath83 .
we then have the asymptotic expression@xmath129 we now expand the integrands about @xmath130 , and obtain@xmath131{cc}% \tau , & \text{for}~d>4\\ \frac{\tau}{\ln\tau } , & \text{for}~d=4\\ \tau^{2 } , & \text{for}~d=3 \end{array } \right . , \label{expclassical}%\ ] ] where @xmath132 gives the distance from the classical critical point .
from these equations , supplemented by standard scaling considerations , it is possible to calculate the usual critical exponents associated with the classical spherical model .
although @xmath126 assumes different values for @xmath90 and @xmath79 , the asymptotic behaviour is the same , regardless of the value of @xmath133 . in analogy with the calculations for finite temperatures , we write@xmath134 from an expansion about @xmath135
, we obtain the asymptotic behaviour@xmath131{cc}% \delta , & \text{for}~d>3\\ \frac{\delta}{\ln\delta } , & \text{for}~d=3\\ \delta^{2 } , & \text{for}~d=2 \end{array } \right . , \label{quantumexp}%\ ] ] where @xmath136 gives the distance from the quantum critical point .
as in the case of finite temperatures , eq .
( [ quantumexp ] ) holds for @xmath97 and any value of @xmath6 .
it is easy to use scaling arguments in order to obtain the ( quantum ) critical exponents for @xmath97 . with the necessary reinterpretations , and
although critical dimensions are different , these values are in agreement with results of vojta for the quantum ferromagnetic case @xcite . at the lifshitz point , @xmath8 , the maximum of @xmath137 is still given by @xmath138 , but the second derivative vanishes along the direction of competition .
we then have to consider the quartic term in the expansion of @xmath139 about @xmath125 . for @xmath121 , the first and second integrals of eq .
( [ tfinite ] ) exist for @xmath99 and for @xmath140 , respectively .
we then have@xmath131{cc}% \tau , & d>(m+8)/2,\\ { \tau}/{\ln\tau } , & d=(m+8)/2,\\ { \tau}^{3/2 } , & d=(m+7)/2,\\ \tau^{2 } , & d=(m+6)/2,\\ \tau^{5/2 } , & d=(m+5)/2 , \end{array } \right . , \label{classicalexp1}%\ ] ]
where @xmath132 . from these asymptotic results ,
it is possible to obtain all the classical critical exponents . at @xmath10 , the first integral in eq .
( [ tzero ] ) exists for @xmath141 , and the second integral for @xmath142 .
we then have@xmath131{cc}% \delta , & d>(m+6)/2,\\ { \delta}/{\ln\delta } , & d=(m+6)/2,\\ { \delta}^{3/2 } , & d=(m+5)/2,\\ \delta^{2 } , & d=(m+4)/2,\\ \delta^{5/2 } , & d=(m+3)/2 , \end{array } \right . , \label{quantumexp1}%\ ] ] where @xmath136 . in conclusion
, we have the same values for either classical or quantum exponents .
as in the case of @xmath97 , the only difference is the critical dimension . according to an old conjecture about quantum critical behaviour ,
the quantum values of the critical exponents are given by the corresponding values of the classical version of the system in @xmath143 dimensions , where @xmath144 is a dynamical critical exponent .
for @xmath145 the dynamical critical exponent is @xmath146 ( anisotropic case ) ; for @xmath147 , it assumes the value @xmath148 ( isotropic case ) .
it should be mentioned that hamiltonian formulations of the mean spherical model have been considered by some authors @xcite .
let us consider the system in a site - dependent field,@xmath149 and write the partition function@xmath150 \left [ 2\sinh\frac{1}{2}\beta w\left ( q\right ) \right ] ^{-1},\label{partfunc2}%\ ] ] where @xmath151 is the fourier transform of @xmath152 , and we are omitting the vector notation .
we then have@xmath153 ^{2}},\ ] ] from which we obtain the pair correlations in real space,@xmath154 as discussed by pisani and collaborators @xcite , the analysis of @xmath155 , for @xmath156 , below the critical temperature , leads to the introduction of a modulated order parameter , with characteristic oscillations for @xmath79 .
a detailed analysis of the long - range correlations at lifshitz point has been published by frachebourg and henkel @xcite .
there are also some investigations of the interplay between competing interactions and the decay of correlations @xcite .
we now show that these oscillations of the pair correlations in terms of distance are already present in the much simpler one - dimensional case . in the thermodynamic limit
( [ corr ] ) can be written as@xmath157 } , \label{c1}%\ ] ] with the spherical condition @xmath158 } , \label{esf1}%\ ] ] where @xmath159 , @xmath17 and @xmath160 are written in units of @xmath3 , with the requirement that @xmath161 . therefore , @xmath162 for @xmath90 , and @xmath163 $ ] for @xmath79 .
( [ c1 ] ) can be rewritten as@xmath164 where the contour @xmath165 is the unit circle and we assuming that @xmath166 .
the fourth - order polynomial in the denominator is easily factorized , with two roots , @xmath167 and @xmath168 , inside the unit circle .
after some straightforward algebra , we have@xmath169 } { \cos\left ( \theta_{1}-\alpha\right ) } \exp\left [ h\ln|w_{1}|\right ] , \label{c11}%\ ] ] where @xmath170 is a real phase depending on @xmath87 and @xmath17 , and@xmath171 with@xmath172 so that @xmath173 , which leads to the well - known exponential decay
. the spherical condition ( [ esf1 ] ) can be used to parametrically eliminate the chemical potential @xmath17 , and write the correlations in terms of @xmath0 and @xmath87 .
in particular , oscillations are suppressed by @xmath174 , which is equivalent to@xmath175 with the asymptotic value @xmath10 for @xmath8 . in fig .
3 , we draw this border , and indicate the region with oscillating correlations ( @xmath176 ) in the @xmath38 plane . from eq .
( [ c11 ] ) , we can write expressions for a modulation length , @xmath177 , and the correlation length , @xmath178 , in terms of temperature and the competition parameter @xmath87 , which is a useful information to investigate the growth of modulated domains @xcite . a similar behaviour has been found by stephenson @xcite in a calculation for the annni chain .
[ ptb ] fig_correlations.eps
we report an analysis of the phase diagram of a quantum mean spherical model in terms of temperature @xmath0 , a quantum parameter @xmath1 , and the ratio @xmath2 , where @xmath3 is a ferromagnetic interaction between first - neighbour sites along the @xmath4 directions of a hypercubic lattice , and @xmath5 is associated with competing antiferromagnetic interactions between second neighbours along @xmath6 directions .
we regain a number of results for the classical version of this model , including the topology of the critical line in the @xmath7 space , with a singular behaviour at the lifshitz point , @xmath8 , for @xmath179 , which includes the case of the usual analogue of the axial - next - nearest - neighbour ising , or annni , model .
we consider in particular the @xmath10 phase diagram , which displays a quantum lifshitz point , at @xmath8 . in the @xmath119 phase diagram
, there is a critical border , @xmath11 for @xmath12 , with a singularity at the lifshitz point if @xmath13 .
we establish upper and lower critical dimensions and analyse the critical behaviour in the neighbourhood of the lifshitz point . in one dimension
, we derive analytic expressions for the decay of pair correlations , and determine the region of modulated behaviour in the @xmath38 phase diagrams . | we analyse the phase diagram of a quantum mean spherical model in terms of the temperature @xmath0 , a quantum parameter @xmath1 , and the ratio @xmath2 , where @xmath3 refers to ferromagnetic interactions between first - neighbour sites along the @xmath4 directions of a hypercubic lattice , and @xmath5 is associated with competing antiferromagnetic interactions between second neighbours along @xmath6 directions .
we regain a number of known results for the classical version of this model , including the topology of the critical line in the @xmath7 space , with a lifshitz point at @xmath8 , for @xmath9 , and closed - form expressions for the decay of the pair correlations in one dimension . in the @xmath10 phase diagram
, there is a critical border , @xmath11 for @xmath12 , with a singularity at the lifshitz point if @xmath13 .
we also establish upper and lower critical dimensions , and analyse the quantum critical behavior in the neighborhood of @xmath8 . |
our everyday way of looking at space - time as a stage for physical events conflicts with the requirement of defining all physical quantities including space and time through precise measurement protocols .
this means that we should more properly regard space - time as emerging from events , instead of pre - existing them .
the operational definition of space - time is defined by the protocol that sets up the coordinate system .
for example , in the einstein protocol light pulses are sent back and forth between different locations : at the place where the signal has been originated , from the arrival time of the reflected signal one infers both the distance and the time of the remote event of signal - reflection .
the protocol shows how space - time is indeed a coherent organization of inferences based on a causal structure for events .
the clock itself is just a sequence of events a light pulse bouncing between two mirrors .
the closest are the mirrors , the more precise is the clock , and the more refined is the coordinate system .
the above reasoning shows that ultimately space and time are defined through pure event - counting , precisely counting _ tic - tacs _ of the observer s clock , and we are thus lead to regard space and time as emergent from the topology of the causal network of events .
the events of the network do not need to be regarded as actual , but can be just potential , and the fabric of space - time is precisely the network of causal links between them .
the idea of deriving space - time from purely causal structures is not new .
raphael sorkin started an independent research line in quantum gravity based on this idea more than two decades ago @xcite .
this was motivated by the potentialities of the approach residing in the natural discreteness of the causal network , which also provides a history - space for a path integral formulation @xcite .
the possibility of recovering the main features of the space - time manifold
topology , differentiable structure , and conformal metric has been shown , starting from discrete sets of points endowed with a causal partial ordering @xcite .
along these lines , in an operational context , lucien hardy has also formulated a _ causaloid _
approach @xcite , which considers the possibility of a dynamical treatment of the causal links . in the causal - set approach of sorkin events
are randomly scattered in order to avoid occurrence of their sparseness in the boosted frames that would lead to violation of lorentz invariance .
clearly the randomness of sorkin s events should not be regarded in terms of their location on a background , otherwise we contradict the same idea of space - time emergence .
we then need to consider randomnes at the pure topological level , and this means having random causal connections .
however , regarding the causal connections as an irreducible description of the physical law , a random topology would then corresponds to having a random physical law at the most microscopic level ( the planck scale ) , and one may argue that a `` random law '' would contradict the same notion of law .
instead , the randomnes should results from the law itself , e.g. in a quantum cellular automaton , where randomness comes from quantum nature of the network .
the universality of the physical law thus lead us to take the causal network as topologically homogeneous .
topological homogeneity has then the added bonus that metric simply emerges from the pure topology by just counting events along the network .
it is obvious that the discreteness of the network will lead to violation of lorenz covariance ( and the other space symmetries ) at the planck scale level : however , one must have a theory where covariance is restored in the large scale limit the fermi scale corresponding to counting huge numbers of events . with the above motivations , in this paper we analyze the mechanism of emergence of space - time from the pure homogeneous topology in @xmath0 dimensions .
we present a digital version of the lorentz transformations , along with the corresponding digital - analog conversion rule . upon considering the causal connections as exchanges of classical information
, we can establish coordinate systems via an einsteinian protocol , leading to a digital version of the lorentz transformations . in a computational analogy first noticed by leslie lamport @xcite
, the foliation construction can be regarded as the synchronization protocol with a global clock of the calls to independent subroutines ( the causally independent events ) in a parallel distributed computation .
the boosts are determined by the relative lengths of the _ tic _ and _ tac _ of the clock , and the lorentz time - dilation corresponds to an increased number of leaves within a clock _ tic - tac _ , whereas space - contraction results from the corresponding decreased density of events per leaf , as first noticed in ref .
@xcite .
we will see that the operational procedure of building up the coordinate system introduces an in - principle indistinguishability between neighboring events , resulting in a network that is coarse - grained , the thickness of the event being a function of the observer s clock .
the digital version of the lorentz transformation is an integer relation which differs from the usual analog transformation by a multiplicative real constant corresponding exactly to the event thickness .
the composition rule for velocities is independent on such constant , and is the same in both the analog and the digital versions .
preliminary results of the present work were already presented in ref .
@xcite .
the illustrated simple classical construction can be extended to space dimension greater than one , but at the price of anisotropy of the maximal speed , due to the weyl - tiling problem .
this issue is cured if the causal network is quantum , as e.g. in a quantum cellular automaton , and isotropy is recovered with quantum coherence , corresponding to superposition of causal paths .
we will thus argue that in a causal network description of space - time , the quantum nature of the network is crucial .
the first problem to address is which specific lattice should be adopted for the causal network . in our convention
the causal arrow is directed from the bottom to the top of the network .
the dimension of the emerging space - time corresponds to the graph - dimension of the network , which is the dimension of the embedding manifold such that all links can be taken as segments of straight line with the same length .
we will require that the lattice be pure topology ( namely with all events equivalent ) , corresponding to a locally homogeneous space - time , and with no redundant links .
it is then easy to see that in 1 + 1 dimensions there are only three possible lattices : the square , the triangular , and the honeycomb ones .
the honeycomb - lattice has two inequivalent types of events ( having one input and two output links and viceversa ) , and the corresponding `` undressed '' topology where each couple of connected inequivalent events are merged into a single event reduces to the square lattice .
the triangular - lattice , on the other hand , has redundant causal links ( the middle vertical ones ) .
we are thus left with the square - lattice .
we always assume the network links as oriented according to the causal arrow . in the square - lattice network
there are thus two types of link : toward the right and toward the left
shortly _ r_-link and _ l_-link .
two events are in the same position ( for some boosted reference frame ) if they are connected by a path made with a sequence of _ r_-links followed by a sequence of _ l_-links .
when the two sequences contain the same number of links the reference is at rest .
a clock is a sequence of causally connected events periodically oscillating between two positions . for an einstein clock the oscillation ( _ tic - tac _ ) is exactly the same couple of sequences of _ l_- and _ _ r-__links identifying events in the same position .
the precision of the clock , namely the minimum amount of time that it can measure , is the number of links of a complete _ tic - tac_. the _ tic - tac _ is indivisible , namely the sole _ tic _ ( or _ tac _ ) is not a complete measured time interval , since it involves two different positions . and @xmath1 ( see text ) . from the left to the right
we have the rest - frame clock , clock corresponding to @xmath2 , and boosted - frames for @xmath3 , @xmath4 , and @xmath5 , respectively , corresponding to digital speed @xmath6 , @xmath6 and @xmath7 , respectively .
the case @xmath4 has doubled imprecision , compared to the case @xmath3 .
[ f : clocks ] ] in the following we will call _ light signals _ those sequences of events that are connected only by _
r_-links or by _ _
l-__links , namely making segments at 45 degrees with the horizontal in the network .
their `` speed '' is equal to `` one event - per - step '' , and is the maximum speed allowed by the causality of the network , since connecting events along a line making an angle smaller than @xmath8 with the horizontal would require to follow some causal connections in the backward direction from the output to the input
. in this way in a homogeneous causal network suffices to guarantee a bound for the speed of information flow . in the following
we will take the clock _
tic - tac _ made with @xmath9 _
r_-links followed by @xmath1
_ l_-links ( see fig , [ f : clocks ] ) .
any clock allows to introduce a reference frame @xmath10 which is just a foliation of the network built up using the einstein protocol . from the start of the clock _
tic - tac _ a light signal is sent to an event in a different position and then received back at the clock .
the intermediate time between the sending and the receiving event is taken as synchronous with the event at the turning point , and the number of _ tic - tacs _
divided by two is taken as the distance from the turning point and the clock conventionally located at the beginning of the _ tic - tac_. in this way we build the foliation corresponding to a given clock .
a set of synchronous events identifies a leaf of the foliation . in fig .
[ f : clocks2 ] the einstein protocol is illustrated in two particular reference frames .
the figure on the left corresponds to the rest - frame , with the blue lines depicting the coordinate system established using the clock with @xmath2 ( see fig .
[ f : clocks ] ) .
the green lines represent light signals bouncing between the clock and four particular events in the network .
these events are synchronous , since the intermediate time between the sending and the receiving event on the clock is the same for all of them .
they lie on the same leaf of the foliation , but at different position , @xmath11 , respectively : the spatial coordinate is obtained by counting the _ tic - tacs _ between the the sending and the receiving event divided by two .
the right figure represents a boosted frame for @xmath3 , built up using the same protocol as in the left figure .
( see fig .
[ f : clocks ] ) .
the green lines represent light signals bouncing between the clock and four particular events in the network .
these events are synchronous , since the intermediate time between the sending and the receiving event on the clock is the same for all of them .
they lie on the same leaf of the foliation , but at different position , @xmath11 , respectively : the spatial coordinate is obtained by counting the _ tic - tacs _ between the the sending and the receiving event ( @xmath12,respectively ) divided by two .
* right figure : * boosted reference frame ( blue lines ) for @xmath3 , built up using the same protocol as in the felt figure .
[ f : clocks2],title="fig : " ] ( see fig . [
f : clocks ] ) .
the green lines represent light signals bouncing between the clock and four particular events in the network .
these events are synchronous , since the intermediate time between the sending and the receiving event on the clock is the same for all of them .
they lie on the same leaf of the foliation , but at different position , @xmath11 , respectively : the spatial coordinate is obtained by counting the _ tic - tacs _ between the the sending and the receiving event ( @xmath12,respectively ) divided by two .
* right figure : * boosted reference frame ( blue lines ) for @xmath3 , built up using the same protocol as in the felt figure .
[ f : clocks2],title="fig : " ] due to indivisibility of the _ tic - tac _ , we see that there are indiscernible events , for which the synchronisation occurs in the middle of the _ tic - tac_. we are thus led to identify events , and merge them into thicker coarse - grained events .
this is done as follows .
we identify the events along the _ tic _ and those along the _ tac _ so that the _ tic - tac _ is always regarded as the bouncing between two next neighbour events . then we merge events into minimal sets so that the topology is left the invariant ( see figures [ f : foliation11 ] and [ f : foliation_coarse ] ) .
we can distinguish between two different kinds of coarse - graining : one due to the boosting ( in yellow in the figures ) , and one due to intrinsic imprecision of the clock ( in gray ) .
the difference between the two is clarified in fig .
[ f : foliation_coarse ] . in the top figure events along the _ tic _ and events along the _ tac _ are identified in the boosted frame .
then events are merged into minimal sets ( in yellow ) so that the topology is left invariant ( the merged events are again events of a square - lattice network ) . in the central figure the coarse graining associated to the intrinsic imprecision
is added in gray , and finally , in the bottom figure the circuit is stretched so to have all synchronous events on horizontal lines , and events located in the same position on vertical lines .
notice that in the special case of the rest - frame , see fig .
[ f : foliation11 ] , the coarse - graining is just due to the intrinsic imprecision of the clock . ] , title="fig:"].1 cm , title="fig : " ] .1 cm , title="fig : " ]
the velocity of the boosted frame can be easily written in terms of the @xmath9 and @xmath1 of the _ tic - tac _ of the clock , by simply evaluating the ratio of the distances in space and time between the two ending points of the _ tic - tac _ , namely @xmath13 thanks to invariance of topology under the boost coarse - graining , the above identity holds also for the motion relative to any boosted frame , whence , upon defining with @xmath14 ( @xmath15 denoting rational numbers ) and @xmath16 for frames @xmath17 and @xmath18 , and by @xmath19 the relative velocity of frame @xmath18 with respect to frame @xmath17 , one has @xmath20 now , by using the trivial identities @xmath21 and @xmath22 one has @xmath23 which by simple algebraic manipulations immediatley gives @xmath24 the last identity is the composition rule of parallel velocities ( the only possibility in @xmath0 dimension ) in special relativity . ) and ( [ e : lorentztrans2 ] ) , leading to the digital version of the lorentz transformations ( [ ttrans1 ] ) and ( [ ttrans2 ] ) . *
left figure : * the reference frame @xmath17 with @xmath25 is represented by the tyny network in black , whereas the coarser network in blue represents the boosted reference frame @xmath18 with @xmath26 , @xmath27 . according to eq .
( [ e : speed12 ] ) the relative velocity of @xmath18 with respect to @xmath17 is @xmath28 . in order to connect the coordinate systems in the two frames
we have chosen the same origin @xmath29 on both @xmath18 and @xmath17 .
the generic event on @xmath18 has coordinates @xmath30 and @xmath31 in the two frames , respectively . *
right figure : * a spatial step in @xmath18 corresponds to @xmath32 space and @xmath33 time steps in @xmath17 . in the same way to a time step in @xmath18 corresponds to @xmath33 space and @xmath32 time steps in @xmath17 .
this correspondence allows to determine the coordinates of a given event in the frame @xmath34 in terms of its coordinates @xmath35 in the frame @xmath18 .
the resulting transformations are in eqs .
( [ e : lorentztrans1 ] ) and ( [ e : lorentztrans2 ] ) .
[ f : lorentztrans],title="fig : " ] ) and ( [ e : lorentztrans2 ] ) , leading to the digital version of the lorentz transformations ( [ ttrans1 ] ) and ( [ ttrans2 ] ) .
* left figure : * the reference frame @xmath17 with @xmath25 is represented by the tyny network in black , whereas the coarser network in blue represents the boosted reference frame @xmath18 with @xmath26 , @xmath27 . according to eq .
( [ e : speed12 ] ) the relative velocity of @xmath18 with respect to @xmath17 is @xmath28 . in order to connect the coordinate systems in the two frames
we have chosen the same origin @xmath29 on both @xmath18 and @xmath17 .
the generic event on @xmath18 has coordinates @xmath30 and @xmath31 in the two frames , respectively .
* right figure : * a spatial step in @xmath18 corresponds to @xmath32 space and @xmath33 time steps in @xmath17 . in the same way to a time step in @xmath18 corresponds to @xmath33 space and @xmath32 time steps in @xmath17 .
this correspondence allows to determine the coordinates of a given event in the frame @xmath34 in terms of its coordinates @xmath35 in the frame @xmath18 .
the resulting transformations are in eqs .
( [ e : lorentztrans1 ] ) and ( [ e : lorentztrans2 ] ) .
[ f : lorentztrans],title="fig : " ] now we use the einstein protocol to construct the boosted coordinated system with respect to the rest - frame along with the relative coordinate systems between any couple of boosted frames .
we will now see that the coordinates of an event transform from the frame @xmath18 to the frame @xmath17 as follows @xmath36 in fact , from a simple inspection of figure [ f : lorentztrans ] one can check eqs .
( [ e : lorentztrans1 ] ) and ( [ e : lorentztrans2 ] ) with the frame @xmath17 as the rest frame . in the left figure
the reference frame @xmath17 with @xmath25 is represented by the tyny network in black , whereas the coarser network in blue represents the boosted reference frame @xmath18 with @xmath26 , @xmath27 . according to eq .
( [ e : speed12 ] ) the relative velocity of @xmath18 with respect to @xmath17 is @xmath28 . in order to connect the coordinate systems in the two frames
we have chosen the same origin @xmath29 on both @xmath18 and @xmath17 .
the generic event on @xmath37 has coordinates @xmath30 and @xmath31 in the two frames , respectively . in the figure on
the right one can see that a spatial step in @xmath18 corresponds to @xmath32 space and @xmath33 time steps in @xmath17 .
in the same way a time step in @xmath18 corresponds to @xmath33 space and @xmath32 time steps in @xmath17 .
this correspondence allows to determine the coordinates @xmath34 of a given event in the frame @xmath17 in terms of its coordinates @xmath35 in the frame @xmath18 .
the resulting transformations are in eqs .
( [ e : lorentztrans1 ] ) and ( [ e : lorentztrans2 ] ) . invariance of topology with boost , guarantees that they also hold between any couple of boosted frames . by elementary manipulation eqs .
( [ e : lorentztrans1 ] ) , ( [ e : lorentztrans2 ] ) can be written in the more customary way @xmath38 upon defining the following constant depending on the clocks of the two frames @xmath39 and using the identity @xmath40 we obtain the _ digital _ lorentz transformations @xmath41 , corresponding to a digital time - dilation by a factor 2 ( analog factor @xmath42 ) and space - contraction by a factor @xmath43 ( compare with the same factors in eqs .
( [ digitallorentz])).[f : lorentz ] ] eqs .
( [ digitallorentz ] ) differ from the usual _ analog _ lorentz transformations by the multiplicative factor @xmath44 , which is logically required to make the transformations rational , compensating the irrationality of the boost factor @xmath45 .
the digital - analog conversion is thus just a rescaling of both space and time coordinates by the factor @xmath46 depending on the boost , which is exactly the square - root of the volume of the coarse - grained event measured as the number of rest - frame events that it contains .
such event volume also affects the lorentz space - contraction and time - dilation factor , which in the digital case is given by @xmath47 , whereas in the analog case is rescaled by the ratio of event volumes , leading to @xmath48 .
thus , for example , for @xmath49 and @xmath50 corresponding to @xmath51 the digital factor is @xmath52 whereas the analog one is @xmath53 .
the digital factor agrees with that of the lorentz time - dilation and space - contraction mechanism of ref .
@xcite , given in terms of increased density of leaves and corresponding decreased density of events per leaf , as illustrated in fig .
[ f : lorentz ] .
we have analyzed the mechanism of emergence of space - time from homogeneous topology in @xmath0 dimensions , deriving the digital version of the lorentz transformations along with the corresponding digital - analog conversion rule .
the homogeneity of topology physically represents the universality of the physical law ( it is worth mentioning that such law is stripped of the conventionality of space and time homogeneity : see e.g. ref .
we have built the digital coordinate system using the einstein s protocol , with signals sent back - and - forth to events from an observer s clock .
we found that the procedure introduces an in - principle indistinguishability between neighbouring events , due to the limited precision of the clock , resulting in a network that is coarse - grained , with the event thickness also depending on the boost .
the digital version of the lorentz transformation is an integer relation which differs from the usual analog transformation by a multiplicative real constant corresponding to the event thickness .
-dimensional computational network : view of a leaf in the rest frame .
information must zig - zag to flow at the maximal speed in diagonal direction .
this leads to a slow - down of a @xmath54 factor of the analog speed compared to cubic axis direction . ]
the present purely classical cinematical construction does not straightforwardly extend from one dimension to larger dimensions , due to the weyl - tiling issue , namely that continuum geometry can not simply emerge from counting sites on a discrete lattice , since e.g. in a square tiling one counts the same number of tiles along a side and along the diagonal of a square @xcite .
thus , for example , as shown in fig . [
f : pyth ] , in a causal network shaped as a square - lattice the fastest speed would be along the cubic axes , whereas along diagonals information should zig - zag , resulting in a slowdown by a factor @xmath54 ( or even @xmath55 in three dimensions ) . indeed , a general theorem of tobias fritz @xcite shows that the polytope of points that can be reached in no more than @xmath56 links in a periodic graph does not approach a circle for large @xmath56 .
since the polytope has necessarily distinguished directions , this means that there is no periodic graph for which this velocity set is isotropic .
this result represents a no - go theorem for the emergence of an isotropic space from a discrete homogeneous causal network representing a _ classical _ information flow .
the situation , however , is completely different if one considers the possibility that information can flow in a superposition of paths , along the network , as in a quantum cellular automaton , corresponding to a homogeneous quantum computational network . in fig .
[ f : automaton ] a concrete example of evolution is given for a two - dimensional quantum weyl automaton of the kind of bialynicki - birula @xcite on a square - lattice .
one can see that the maximum propagation speed is isotropic after just few steps . in a similar way full lorentz covariance
is expected to be restored in the same limit of infinitely many events a kind of thermodynamic limit bringing the automaton to the fermi scale .
gmd acknowledges interesting discussions with raphael sorkin , seth lloyd , and tobias fritz . this work has been partially supported by prin 2008 . 15 l. bombelli , j. h. lee , d. meyer , and r. sorkin , phys . rev .
lett * 59 * , 521 ( 1987 ) .
f. markopoulou , gr - qc/0210086 ( 2002 ) .
j. henson , in _ approaches to quantum gravity : towards a new understanding of space and time _
d. oriti ( cambridge university press , cambridge uk , 2006 ) ( also gr - qc/0601121 ) .
s. surya , theor .
comp . sc . * 405 * 188 ( 2008 ) .
l. hardy , j. phys .
a * 40 * 3081 ( 2007 ) .
l. lamport , comm .
acm , * 21 * 558 ( 1978 ) .
g. m. dariano , in cp1232 _ quantum theory : reconsideration of foundations , 5 _ ed . by a.
y. khrennikov , ( aip , melville , new york , 2010 ) , pg .
3 ( also arxiv : 1001.1088 ) .
g. m. dariano and a. tosini , arxiv:1008.4805 ( 2010 ) d. malament , nos * 11 * 293 ( 1977 ) .
h. weyl , _ philosophy of mathematics and natural sciences _ , princeton university t. fritz , _ velocity polytopes of periodic graphs _ , draft ( 2011 ) . press , ( princeton 1949 )
. i. bialynicki - birula , phys .
d * 49 * 6920 ( 1994 ) . | in this paper we study the emergence of minkowski space - time from a causal network . differently from previous approaches
, we require the network to be topologically homogeneous , so that the metric is derived from pure event - counting .
emergence from events has an operational motivation in requiring that every physical quantity including space - time be defined through precise measurement procedures .
topological homogeneity is a requirement for having space - time metric emergent from the pure topology of causal connections , whereas physically homogeneity corresponds to the universality of the physical law .
we analyze in detail the case of @xmath0 dimensions .
if we consider the causal connections as an exchange of classical information , we can establish coordinate systems via an einsteinian protocol , and this leads to a digital version of the lorentz transformations . in a computational analogy
, the foliation construction can be regarded as the synchronization with a global clock of the calls to independent subroutines ( corresponding to the causally independent events ) in a parallel distributed computation .
thus the lorentz time - dilation emerges as an increased density of leaves within a single _ tic - tac _ of a clock , whereas space - contraction results from the corresponding decrease of density of events per leaf .
the operational procedure of building up the coordinate system introduces an in - principle indistinguishability between neighboring events , resulting in a network that is coarse - grained , the thickness of the event being a function of the observer s clock .
the illustrated simple classical construction can be extended to space dimension greater than one , with the price of anisotropy of the maximal speed , due to the weyl - tiling problem .
this issue is cured if the causal network is quantum , as e.g. in a quantum cellular automaton , and isotropy is recovered by quantum coherence via superposition of causal paths .
we thus argue that in a causal network description of space - time , the quantum nature of the network is crucial . |
* toeplitz operators @xmath1 can formally be defined in the space @xmath0 of sequences @xmath2 by the formula @xmath3 thus the matrix elements of a toeplitz operator depend on the difference of the indices only .
so it is natural to expect that properties of toeplitz operators are close to those of discrete convolution operators acting in the space @xmath4 . the precise definition of the operator @xmath1 requires some accuracy .
let @xmath5 be the dense set of sequences @xmath6 with only a finite number of non - zero components .
if the sequence @xmath7 , then for @xmath8 sequence eq : hd belongs to @xmath9 . in this case the operator @xmath1 is defined on @xmath10 , and it is symmetric if @xmath11 . without any a priori assumptions on @xmath12 , only the quadratic form @xmath13 = \sum_{n , m\geq 0 } t_{n - m }
g_{m}\bar{g}_{n } \label{eq : qfq}\ ] ] is well defined for @xmath14 .
the theory of toeplitz operators is a very well developed subject .
we refer to the books @xcite and @xcite ( chapter xii ) , @xcite ( chapters b.4 and b.6 ) , @xcite ( chapter 3 ) for basic information on this theory .
let us state a necessary and sufficient condition for a toeplitz operator @xmath1 to be bounded .
below @xmath15 is always the normalized lebesgue measure on the unit circle @xmath16 .
[ b - h ] a toeplitz operator @xmath1 @xmath17defined , possibly , via its quadratic form eq : qfq@xmath18 is bounded if and only if the @xmath12 are the fourier coefficients of some bounded function on @xmath16 : @xmath19 however results on unbounded toeplitz operators are very scarce .
we can mention only the paper @xcite by p. hartman and the recent survey @xcite by d. sarason ; see also references in these articles .
* in this paper , we consider semibounded toeplitz operators in the space @xmath0 .
_ we always suppose that @xmath11 so that the quadratic form eq : qfq is real and assume that @xmath13 \geq \gamma \| g\|^2 , { \quad}g\in{\cal d } , { \quad}\| g\| = \| g\|_{\ell^2({\bbb z}_{+ } ) } , \label{eq : t1}\ ] ] for some _ @xmath20 . in this case
, we are tempted to define @xmath1 as a self - adjoint operator corresponding to the quadratic form @xmath21 $ ] .
such an operator exists if the form @xmath21 $ ] is closable in the space @xmath22 , but as is well known this is not always true ( see example [ ex ] , below ) .
we refer to the book @xcite for basic information concerning these notions ; they are also briefly discussed in subsection 2.1 .
we recall that , by definition , the operator corresponding to the form @xmath21 + { \delta}\| g\|^2 $ ] is given by the equality @xmath23 ( observe that the identity operator @xmath24 is a toeplitz operator ) . also by definition , if a form @xmath21 $ ] is closable , then all forms @xmath21 + { \delta}\| g\|^2 $ ] are closable
. therefore we can suppose that the number @xmath25 in eq : t1 is positive ; for definiteness , we choose @xmath26 .
we proceed from the following well known result ( see , e.g. , 5.1 of the book @xcite ) that is a consequence of the f. riesz - herglotz theorem .
[ r - h ] the condition @xmath27 is satisfied if and only if there exists a non - negative @xmath17finite@xmath18 measure @xmath28 on the unit circle @xmath16 such that the coefficients @xmath29 admit the representations @xmath30 equations eq : wh for the measure @xmath31 are known as the trigonometric moment problem .
of course their solution is unique .
note that for the lebesgue measure @xmath32 we have @xmath33 and @xmath34 for @xmath35 .
therefore the measure corresponding to the form @xmath21 + { \delta}\| g\|^2 $ ] equals @xmath36 .
so we have a one - to - one correspondence between toeplitz quadratic forms satisfying estimate eq : t1 and real measures satisfying the condition @xmath37 for all borelian sets @xmath38 .
our goal is to find necessary and sufficient conditions for the form @xmath21 $ ] to be closable .
the answer to this question is strikingly simple .
[ t1 ] let the form @xmath21 $ ] be given by formula eq : qfq on elements @xmath39 , and let the condition eq : t1 be satisfied
. then the form @xmath21 $ ] is closable in the space @xmath22 if and only if the measure @xmath40 in the equations eq : wh is absolutely continuous .
of course theorem [ t1 ] means that @xmath41 where the function @xmath42 and @xmath43 .
thus theorem [ t1 ] extends theorem [ b - h ] to semibounded operators .
the function @xmath44 is known as the symbol of the toeplitz operator @xmath1 .
so theorem [ t1 ] shows that for a semibounded toeplitz operator ( even defined via the corresponding quadratic form ) , the symbol exists and is a semibounded function . *
the proof of theorem [ t1 ] will be given in the next section .
the closure of the form @xmath21 $ ] is described in theorem [ clo ] .
the discussion of these results as well as their comparison with similar statements for hankel operators are postponed until section 3 .
we there also explain shortly how our results extend to vectorial toeplitz operators .
let @xmath21 $ ] be a quadratic form defined on a set @xmath45 dense in a hilbert space @xmath46 and satisfying inequality eq : t1 where @xmath47 is the norm of @xmath48 .
suppose that @xmath26 , consider the norm @xmath49}$ ] and introduce the closure @xmath50 $ ] of @xmath45 in this norm .
if @xmath50 $ ] can be realized as a subset of @xmath46 , then one says that @xmath21 $ ] is closable in the space @xmath51 ; this means that the conditions @xmath52 as @xmath53 imply that @xmath54 .
it is easy to see that if @xmath55 is a symmetric semibounded operator on @xmath45 , then the form @xmath56= ( g , t_{0 } g)$ ] is closable .
if the form @xmath21 $ ] is closable , then @xmath50\subset \cal h$ ] is a closed set with respect to the norm @xmath57 . in this case
@xmath21 $ ] is defined by continuity on all @xmath58 $ ] , and one says that the form @xmath21 $ ] is closed on @xmath50 $ ] . for a closed form
there exists a unique self - adjoint operator @xmath1 such that @xmath59 and @xmath60 & = ( g , t h ) , { \quad}\forall g\in { \cal d}[t ] , { \quad}\forall h\in { \cal d } ( t)\subset { \cal d}[t ] , \\ t[g , g]&= \| \sqrt{t } g\|^2 , { \quad}\forall g\in { \cal d}(\sqrt{t } ) = { \cal d}[t ] .
\end{aligned}\ ] ] note that the domain @xmath61 of @xmath1 does not admit an efficient description . we are going to use these general definitions for the space @xmath62 and the toeplitz quadratic forms eq : qfq .
of course quadratic forms , in particular , the toeplitz forms , are not necessarily closable . [ ex ]
let @xmath63 for all @xmath64 .
adding the term @xmath65 , we obtain the form @xmath13=
\big| \sum_{n\geq 0 } g_{n}\big|^2 + \sum_{n\geq 0}| g_{n}|^2\ ] ] satisfying inequality eq : t1 with @xmath26 .
define the sequence @xmath66 by the equalities @xmath67 for @xmath68 and @xmath69 for @xmath70 .
then @xmath71 as @xmath72 .
since @xmath73 , we have @xmath74 as @xmath53 .
nevertheless @xmath75 .
note that the measure @xmath76 corresponding to the sequence @xmath63 , @xmath77 , is supported by the point @xmath78 : @xmath79 , @xmath80 .
on the other hand , we have the following simple assertion . [ ex1 ]
if a sequence @xmath81 , then the form eq : qfq is closable .
now we have @xmath56= ( g , t_{0 } g)$ ] where the symmetric operator @xmath55 is defined by formula eq : hd on the set @xmath45 .
* as already mentioned , by the proof of theorem [ t1 ] we may suppose that estimate eq : t1 is true for @xmath26 . according to theorem [ r - h ] the equations eq : wh are satisfied with a measure @xmath31 such that @xmath82 for all borelian sets @xmath83 ; in particular , the measure @xmath31 is positive .
our proof relies on the following auxiliary construction .
let @xmath84 be the space of functions @xmath85 on @xmath16 with the norm @xmath86 we put @xmath87 and observe that @xmath88 for all @xmath89 .
therefore we can define an operator @xmath90 on domain @xmath91 by the formula @xmath92 .
in view of equations eq : wh , the form eq : qfq can be written as @xmath93 = \sum_{n , m\geq 0}\int_{\bbb t}z^{-n+m } d m(z ) g_m { \overline}{g_{n } } \\
= \int_{\bbb t } | ( { \cal a } g ) ( z)|^2 dm(z ) = \| ag\|^2 _ { l^2 ( \bbb t ; dm ) } , { \quad}g\in \cal d. \label{eq : a2}\end{gathered}\ ] ] this yields the following result .
[ de ] the form @xmath94 $ ] defined on @xmath45 is closable in the space @xmath22 if and only if the operator @xmath90 defined on the domain @xmath91 is closable .
our next goal is to construct the adjoint operator @xmath95 .
observe that for an arbitrary @xmath96 , all the integrals @xmath97 are absolutely convergent and the sequence @xmath98 is bounded .
we denote by @xmath99 the set of all @xmath100 such that @xmath101 .
[ ltm ] the operator @xmath102 is given by the equality @xmath103 on the domain @xmath104 . obviously , for all @xmath14 and all @xmath96 , we have the equality @xmath105 where the sequence @xmath106 is defined by relation eq : a1 .
if @xmath107 , then the right - hand side here equals @xmath108 .
it follows that @xmath109 .
conversely , if @xmath110 , then @xmath111 for all @xmath112 .
therefore it follows from equality eq : y1 that @xmath113 since @xmath114 is dense in @xmath115 , we see that @xmath116 , and hence @xmath107 .
thus @xmath117 .
recall that an operator @xmath118 is closable if and only if its adjoint operator @xmath119 is densely defined .
we use the notation @xmath120 for the closure of the set @xmath121 in the space @xmath122 .
so we have obtained an intermediary result .
[ adj ] the operator @xmath123 and the form @xmath21 $ ] are closable if and only if @xmath124 * 2.3 .
* next , we use the riesz brothers theorem .
we state it in a slightly more general form than in most textbooks .
[ brothers ] for a complex @xmath17finite@xmath18 measure @xmath125 on the unit circle @xmath16 , put @xmath126 and suppose that @xmath127 then the measure @xmath128 is absolutely continuous . indeed
, in view of eq : brbr the function @xmath129 belongs to @xmath130 .
let us consider an auxiliary measure @xmath131 and let @xmath132 be its fourier coefficients .
it follows from eq : br1 that @xmath133 for all @xmath134 .
so , by the standard version of the riesz brothers theorem ( see , e.g. , @xcite , chapter 4 ) , the measure @xmath135 is absolutely continuous . in view of eq : br2 , the same is true for the measure @xmath136 .
the following assertion is almost obvious .
[ ac ] suppose that a set @xmath121 satisfies condition eq : d .
let the measures @xmath137 be absolutely continuous for all @xmath138 .
then the measure @xmath139 is also absolutely continuous
. denote by @xmath140 the characteristic function of a borelian set @xmath83 .
it follows from eq : d that there exists a sequence @xmath141 such that @xmath142 and hence @xmath143 if @xmath144 and the measures @xmath145 are absolutely continuous , then the integrals on the left - hand side are zeros so that @xmath146 .
now it is easy to conclude the only if " part of theorem [ t1 ] .
suppose that the form @xmath21 $ ] is closable .
then by lemma [ adj ] the condition eq : d is satisfied . by the definition of the set @xmath121 , the fourier coefficients of the measures @xmath147 belong to @xmath22 .
therefore it follows from theorem [ brothers ] that these measures are absolutely continuous for all @xmath138 .
hence by lemma [ ac ] the measure @xmath139 is also absolutely continuous . *
it remains to check the converse statement .
let us consider an auxiliary form @xmath148 = \sum_{n , m\in { \bbb z } } t_{n - m } g_{m}\bar{g}_{n } \label{eq : qfq}\ ] ] in the space @xmath149 on the set @xmath150 of sequences @xmath151 with only a finite number of non - zero components .
it is easy to see that the conditions eq : t1 and @xmath148 \geq \gamma \| g\|^2_{\ell^2({\bbb z } ) } , { \quad}g\in{\sf d } , \label{eq : t1x}\ ] ] are equivalent .
we again suppose that @xmath26
. then the coefficients @xmath12 are given by formula eq : wh where @xmath82 for all @xmath152 .
let the operator @xmath153 be defined ( cf .
eq : a ) by the formula @xmath154 quite similarly to eq : a2 , we find that for @xmath155 , @xmath156 = \int_{\bbb t } | ( { \sf a } g ) ( z)|^2 dm ( z ) = \| { \sf a } g \|^2_{l^2 ( { \bbb t } ; dm ) } .
\label{eq : a3}\ ] ] for the completeness of our presentation , let us check that the measure @xmath28 is absolutely continuous if the form @xmath157 $ ] is closable .
this can be done similarly to the proof in the previous subsection of the same fact for the form @xmath94 $ ] , but now @xmath158 in all formulas and the riesz brothers theorem is not needed .
the operator @xmath159 acts again by the formula eq : a1 , and it is defined on the set @xmath160 of all @xmath161 such that @xmath162 .
this means that @xmath163 if and only if @xmath164 for some @xmath165 . the form @xmath157 $ ] is closable if and only if @xmath166 hence lemma [ ac ] implies that the measure @xmath28 is absolutely continuous .
next , we show that , for absolutely continuous measures @xmath167 , the forms @xmath168 $ ] are closable .
now we have @xmath169 where @xmath170 .
therefore it follows from eq : a3 that @xmath171 = { \bf s } [ { \sf a}g , { \sf a } g]\ ] ] where @xmath172 = \int_{\bbb t } w(z ) | f ( z)|^2 d { \bf m } ( z ) .
\label{eq : a5}\ ] ] since the operator @xmath173 is unitary , the form @xmath168 $ ] defined on @xmath174 is closable in @xmath175 if and only if the form @xmath176 $ ] defined on the quasi - polynomials eq : a4 is closable in @xmath177 .
clearly , @xmath176 $ ] is the quadratic form of the operator of multiplication by @xmath178 .
so it is closable because @xmath179 .
moreover , @xmath176 $ ] is closed on the set of all @xmath180 such the integral eq : a5 is finite .
let us summarize the results obtained for the form @xmath181 $ ] .
[ bro ] let the form @xmath182 $ ] be defined by the equality eq : qfq in the space @xmath149 on the set @xmath150 of sequences @xmath183 with only a finite number of non - zero components
. let inequality eq : t1x hold , and let the measure @xmath28 be defined by the relation eq : wh . then the form @xmath182 $ ] is closable if and only if @xmath169 for some @xmath184 , @xmath185 . in this case
the closure of @xmath182 $ ] is given by relations eq : a4 , eq : a3 on the set of all sequences @xmath186 such that the integral eq : a3 is finite .
* let us return to theorem [ t1 ] .
obviously , if the form @xmath182 $ ] is closable in the space @xmath149 , then the same is true for the form @xmath187 $ ] in the space @xmath188 .
so if the measure @xmath28 is absolutely continuous , then by proposition [ bro ] both forms @xmath182 $ ] and @xmath21 $ ] are closable .
this concludes the proof of theorem [ t1 ] .
[ broy ] @xmath189 for the proof of theorem [ t1 ] we used only the if " part of proposition [ bro ] . on the other hand ,
the only if " part of proposition [ bro ] is a consequence of the only if " part of theorem [ t1 ] proven in subsection 2.3 .
@xmath190 comparing theorem [ t1 ] and proposition [ bro ] , we see that if the form @xmath21 $ ] is closable , then the same is true for the form @xmath182 $ ] .
as already noted , the converse assertion is evident .
we now suppose that @xmath191 where @xmath192 and @xmath193 so that the form @xmath21 $ ] is closable in the space @xmath22 .
to describe its closure , we need a mild additional assumption on @xmath194 .
we suppose that the function @xmath194 is a muckenhoupt weight ; see , e.g. , b 5.7 of the book @xcite for various definitions of this notion .
one of them is given by the condition @xmath195 where @xmath196 runs over all subarcs of @xmath16 .
let @xmath197 be the orthogonal projection of @xmath198 onto the hardy class @xmath199 of functions analytic in the unit disc .
the operator @xmath200 is bounded if and only if @xmath44 is a muckenhoupt weight .
recall that the operator @xmath201 is defined by formula eq : a .
obviously , @xmath202 for all @xmath203 .
[ clo ] let the coefficients @xmath12 be given by formula eq : wh where @xmath204 , @xmath192 and @xmath205 .
suppose that @xmath194 is a muckenhoupt weight .
then the closure of the form @xmath21 $ ] defined on @xmath206 by eq : qfq is given by the equality @xmath207
= \int_{\bbb t } | ( { \cal a } g ) ( z)|^2 dm(z ) \label{eq : a2x}\ ] ] on the set @xmath50 $ ] of all @xmath203 such that the right - hand side of eq : a2x is finite .
observe that the operator @xmath208 is unitary and @xmath209 consists of all polynomials eq : a . let @xmath210 be the closure of @xmath211 in @xmath212 .
so the assertion of theorem [ clo ] is equivalent to the equality @xmath213 since the convergence in @xmath212 is stronger than that in @xmath214 , we have @xmath215 and therefore the left - hand side of eq : a2x1 is contained in its right - hand side .
it remains to prove the opposite inclusion .
recall that if all fourier coefficients of some complex measure on @xmath16 are zeros , then this measure is also zero .
suppose that @xmath216 is orthogonal in @xmath212 to the functions @xmath217 for all @xmath158 . then applying the above fact to the measure @xmath218
, we see that @xmath219 .
therefore quasi - polynomials @xmath220 ( the sum consists of a finite number of terms ) are dense in @xmath221 .
so for every @xmath222 there exists a sequence of quasi - polynomials @xmath223 such that @xmath224 since @xmath194 is a muckenhoupt weight , this implies that @xmath225 so if @xmath226 and @xmath227 , then @xmath228 as @xmath72 in @xmath229 which also implies the convergence in @xmath230 . [ clo1 ] recall that the operator @xmath231 was defined by formula eq : a on the domain @xmath232 . of course its closure @xmath233 .
let @xmath234 be given by the formula @xmath235 on the domain @xmath236 that consists of all @xmath203 such that @xmath237 .
then the assertion of theorem [ clo ] is equivalent to the equality @xmath238
let us state a consequence of theorem [ t1 ] in terms of the entries @xmath12 of the form eq : qfq .
[ t1 m ] suppose that the condition eq : t1 is satisfied . if the form @xmath21 $ ] is closable in the space @xmath22 , then @xmath239 as @xmath240 . if @xmath241 , then the form @xmath21 $ ] is closable in the space @xmath22 .
if the form @xmath21 $ ] is closable , then , by theorem [ t1 ] , the measure @xmath28 in the representation eq : wh is absolutely continuous .
therefore its fourier coefficients @xmath239 as @xmath242 .
conversely , if @xmath243 , then @xmath244 with @xmath245 .
therefore , again by theorem [ t1 ] , the form @xmath21 $ ] is closable ( this result was already stated in lemma [ ex1 ] ) .
there is a gap between necessary and sufficient conditions on @xmath246 in proposition [ t1 m ] .
apparently it can not be significantly reduced . recall that by the wiener theorem ( see , e.g. , theorem xi.114 in @xcite ) , if the fourier coefficients @xmath246 of some measure @xmath28 tend to zero , then this measure is necessarily continuous , but it may be singular with respect to the lebesgue measure .
thus the condition @xmath247 as @xmath248 does not imply that the measure @xmath28 defined by eq : wh is absolutely continuous .
so in accordance with theorem [ t1 ] the corresponding toeplitz quadratic form @xmath21 $ ] may be unclosable .
astonishingly , the sufficient condition @xmath241 for the absolute continuity of the measure @xmath28 turns out to be very sharp .
indeed , for every @xmath249 , o. s. ivav - musatov constructed in @xcite a singular measure such that its fourier coefficients satisfy the estimate @xmath250 ( here @xmath251 means that the logarithm is applied @xmath252 times to @xmath253 ) .
examples of singular continuous measures of such type go back to d. e. menchoff @xcite .
a comprehensive survey of various constructions of singular continuous measures with decaying fourier coefficients can be found in @xcite .
* there is a certain parallelism between the theories of toeplitz and hankel operators .
for example , the criteria of boundedness of toeplitz and of hankel operators due to toeplitz ( see theorem [ b - h ] ) and to nehari @xcite , respectively , look formally similar .
toeplitz quadratic forms are linked to the trigonometric moment problem while hankel quadratic forms are linked to the power moment problem .
the following result obtained by hamburger in @xcite plays the role of theorem [ r - h ] .
[ hamb ] the condition @xmath254=\sum_{n , m\geq 0 } q_{n+m } g_{m}\bar{g}_{n}\geq 0 , { \quad}\forall g\in \cal d , \label{eq : qfh}\ ] ] is satisfied if and only if there exists a non - negative measure @xmath255 on @xmath256 such that the coefficients @xmath257 admit the representations @xmath258 1 .
the form @xmath259 $ ] defined on @xmath45 is closable in the space @xmath22 .
2 . the matrix elements @xmath260 as @xmath261 .
the measure @xmath262 defined by equations eq : whh satisfies the condition @xmath263 @xmath17to put it differently , @xmath264 $ ] and @xmath265 . *
* theorem [ t1 ] automatically extends to vectorial toeplitz operators . in this case @xmath266 where @xmath267 are elements of some auxiliary hilbert space @xmath268 , and @xmath12 are bounded operators in @xmath268 .
we now suppose that @xmath269 and @xmath13 = \sum_{n , m\geq 0 } { \langle}t_{n - m } g_{m } , g_{n } { \rangle}_{\mathfrak{n } } \geq \gamma \| g\|^2_{l^2 ( { \bbb z}_{+ } ; \mathfrak{n})},{\quad}\forall g\in \cal d. \label{eq : qfv}\ ] ] the vectorial version of theorem [ r - h ] means that inequality eq : qfv for @xmath270 is equivalent to the representation eq : wh with a non - negative operator valued measure @xmath28 .
let us state a generalization of theorem [ t1 ] to the vector case .
[ tv ] let the condition eq : qfv be satisfied
. then the form @xmath21 $ ] is closable in the space @xmath271 if and only if @xmath272 where @xmath273 and the operator valued function @xmath194 belongs to @xmath274 .
theorem [ clo ] and its proof also extend to the vectorial case provided the projector @xmath197 is a bounded operator in the space @xmath275 .
note that there is a necessary and sufficient condition ( see the paper @xcite ) for the boundedness of this operator generalizing the scalar condition eq : mu ; the result of @xcite requires however that @xmath276 . | we show that a semibounded toeplitz quadratic form is closable in the space @xmath0 if and only if its entries are fourier coefficients of an absolutely continuous measure .
we also describe the domain of the corresponding closed form .
this allows us to define semibounded toeplitz operators under minimal assumptions on their matrix elements .
primary 47a05 , 47a07 ; secondary 47b25 , 47b35 toeplitz quadratic forms , closed operators and quadratic forms , absolutely continuous measures |
broad absorption line quasars ( balqsos ) as their name suggests , show strong broad blue - shifted absorption lines in their spectra believed to be indicative of high velocity ( @xmath10.1c ) out - flowing winds .
balqsos represent approximately 15% of quasars in general @xcite .
the majority ( @xmath1 85% ) @xcite of balqsos are known as hibals , displaying absorption in lines of high ionisation only ( e.g. n v , si iv , and c iv ) .
the remainder are classified as lobals and show in addition broad absorption in lines of low ionisation species , most notably al iii and mg ii .
lobals are further sub - classified according to the presence of fe absorption , the fe lobals .
since the spectra of lobals are in general redder than hibals , @xcite suggests that balqsos and in particular fe lobals may be an early phase in the development of emerging or re - fuelled quasars .
there has been much debate about the relationship between balqsos and the general quasar population as a whole .
simple unification schemes , suggest that balqsos and non - balqsos are similar objects and that any observed differences in their spectra arise due to orientation effects @xcite . in this picture
, the relative fraction of balqsos to non - balqsos has a simple geometric interpretation , representing the fraction of sky , as seen from the source , obscured by gas ( ie .
the source covering fraction ) , which in the simplest case , can be related to the flow geometry ( e.g. opening angle ) . in recent years interest in balqsos outflows has risen sharply , principally because of the realisation that such high velocity outflows carry a substantial amount of energy and momentum into the ism , and may therefore be important in driving agn feedback as well as providing a mechanism for quenching star formation @xcite . indeed , the discovery of highly ionised , very high - velocity x - ray outflows ( e.g. pounds et al
. 2003a , 2003b ; pounds and reeves 2009 ) , for which the energy transport ( in terms of mechanical energy ) is large enough to interrupt the growth of the host galaxy , may provide the causal link behind the well - known correlation between the mass of the central black hole and the mass of the bulge ( e.g. ferrarese and merritt 2000 ; gebhardt et al .
2000 ; tremaine et al .
2002 ) .
however , despite the increasing importance of these outflows the precise mechanism responsible for accelerating them to high velocity remains uncertain , with radiative acceleration @xcite , magneto hydro dynamic ( mhd ) driven winds @xcite and thermally driven winds seen as the main contenders ( see e.g. @xcite for a review ) .
a major hindrance to progress in this area is the absence of a clean discriminatory observational signature of what are essentially orthogonal wind geometries . for radio - quiet objects ,
agn unification schemes tend to favour equatorial wind geometries ( e.g. elvis 2000 ) .
by contrast , observations of radio - loud balqsos @xcite and in particular polarisation observations of pks 0040 - 005 indicate a non - equatorial bal outflow @xcite .
one possible explanation for the apparent difference in wind geometry between radio - loud and radio - quiet objects , is that different acceleration mechanisms operate in these two classes of objects .
perhaps the strongest indicator that radiative driving is responsible for accelerating at least some of the high velocity flows , is the appearance of line - locking features in the spectra of a small fraction of bal quasars ( e.g. turnshek et al .
1988 , weymann et al .
1991 , korista et al .
1993 , arav 1996 , vilkoviskij and irwin 2001 ) .
the most well - studied of these is the so - called ghost of ly-@xmath0 , a small hump seen in the absorption troughs of a small fraction ( less than a few percent ) of balqsos formed via the interaction between ly-@xmath0 photons and n v ions ( see e.g. korista et al .
1993 , arav et al .
1995 , arav 1996 , and references therein ) .
however , the reality of ghosts remains very much an open question .
previous studies have been limited to small samples of relatively low quality spectra , with strong ghosts often only becoming apparent in composite spectra ( arav et al .
1995 , north et al .
moreover , distinguishing between line - locking features and similar features caused by the chance alignment of multiple absorption systems is difficult with relatively small samples . indeed , korista et al .
1993 , showed that in a sample 72 objects , evidence for line - locking features was merely suggestive rather than convincing .
the aim of this paper is twofold : ( i ) to compile the largest sample of uniformly selected ghost - candidate spectra , and ( ii ) attempt to determine the origin of their ghost features , by testing on a source by source basis , whether they meet the original criteria as set out by arav ( 1995 ) necessary for ghost - features to be observed . in
2 we describe the mechanism proposed for ghost formation and outline arav s criteria for the production of strong ghost signatures .
selection of those objects comprising our ghost - candidate spectra and our non - ghost control sample is described in 3 . in
4 we present the results of testing each of the spectra in our ghost - candidate and comparison samples against each of the criteria in turn necessary for the formation of an observable ghost feature .
we discuss the implications of our findings in 5 .
our conclusions are summarised in 6 .
the feature that is poetically known as the ghost of ly-@xmath0 is a hump seen in the absorption trough of a small fraction ( less than 5% , arav 1995 , north et al .
2006 ) of balqsos located 5900 km s@xmath2 blue - ward of the centre of the line to which the absorption is attributed .
the feature was first investigated by @xcite after being seen in a difference spectrum of balqso and non - balqso spectral composites produced by @xcite who pointed out that the difference between the two absorption troughs was approximately equal to the velocity separation of n v and ly-@xmath0 . in a series of papers in the mid 90 s @xcite arav
suggested that the feature seen at this velocity was a result of the increased radiative pressure on n v ions at this velocity within the outflow . in this picture
n v ions moving at 5900 km s@xmath2 see the ly-@xmath0 emission from the agn at the energy of their own emission resulting in a large increase in the scattering cross section of the ly-@xmath0 photons .
consequently , out - flowing n v ions receive an increase in radiation pressure leading to a large injection of momentum into the outflow .
the out - flowing ions of other species are dragged to higher velocities by electromagnetic interactions with the n v ions , leading to a deficit of material at this velocity ( because each out - flowing ion spends very little time at this velocity before being accelerated to greater velocities ) . in turn
this causes a decrease in the opacity at this velocity which manifests as a peak within the absorption troughs of the out - flowing species .
this simple mechanism explains both the presence of the feature and why it s position is linked to the rest frame of the source .
@xcite where the separation of the troughs is constant but their velocities are determined by the velocity of the first trough . ]
@xcite suggested that the reason for the feature being most prominent in c iv absorption rather than in other species , is that the doublet separation of c iv is far smaller ( 498 km s@xmath2 ) . by comparison the larger doublet separations of si iv ( 1933 km s@xmath2 ) and o vi ( 1647 km s@xmath2 )
result in the feature being blurred over several thousand km s@xmath2 , producing a low contrast feature which is difficult to detect in low s / n data .
for this reason , in this study we confine our search for ghost features to the c iv broad absorption line only .
since ly-@xmath0 and n v are amongst the strongest uv lines in agn spectra , an obvious question to ask is why are nt ghost - features ubiquitous amongst the general balqso population ? to answer this , arav proposed a set of physically motivated criteria which had to be met before ghost features would be observable .
these are : 1 .
significant broad absorption in the region between 3000 and 9000 km s@xmath2 blue - ward of line centre , 2 .
strong intrinsic ly-@xmath0 emission , 3 .
narrow ly-@xmath0 emission , 4 .
strong broad n v absorption , 5
. little far - uv flux between between @xmath3200 and @xmath31000 .
the first criterion is simply a statement that ghost - features can only be observed if significant photon scattering is occurring .
the second criterion implies that there must be a significant flux of ly-@xmath0 photons , while the third requires that the line - emitting gas has a relatively narrow spread in velocities so that the resulting feature is not spread over a large range of velocity reducing the contrast .
since the dominant scattering ion is n v , strong n v absorption is also needed . the final requirement , and the most difficult to measure , recognises that in order for ly-@xmath0n v line - locking to dominate the dynamics of the flow , there can not be significant far - uv line driving ( korista et al . 1993 ) .
balqsos represent @xmath115% of the quasar population ( see e.g. knigge et al .
2008 and references therein ) , of these , between @xmath120 - 25% show evidence for multiple troughs @xcite . amongst the multi - trough objects ,
ghost candidates are likely rare representing less than a few per cent of balqsos ( arav 1995 ) because of the stringent requirements for ghost observability .
thus in order to find significant numbers of ghost - candidates , large quasar samples are required .
all previous ghost candidate samples have been limited to just a handful of objects @xcite . by comparison ,
data release 5 ( dr5 ) of the sloan digital sky survey ( sdss ) contains 77,429 quasars of which 28,421 have redshifts between 1.7 and 4.2 allowing c iv and any associated absorption to be seen in their optical spectra .
we require a visible c iv bal because although radiative line driving will produce a feature in all absorption troughs , the resultant hump will be most easily observed in the c iv absorption trough because of the small doublet separation of this line .
several balqso samples have been compiled from the various data releases of the sdss @xcite and from dr5 @xcite .
here we adopt the balqso catalogue of @xcite , containing 3,552 balqsos .
the scaringi et al .
sample , differs from other samples in that is does not rely on simple metrics such as the balnicity index ( bi , weymann et al .
1991 ) or absorption index ( ai , hall et al .
2002 , trump et al .
2006 ) to identify and classify balqsos .
rather it uses a robust hybrid method , involving gross classification with a supervised neural network ( learning vector quantisation ) and the visual inspection of outliers . to identify strong ghost candidates we have performed five cuts on this balqso sample .
the first cut eliminates noisy spectra ( s / n@xmath44 per pixel ) , estimated in two line - free continuum bands from @xmath516501700 and @xmath517001750 . to reduce contamination from emission , absorption and cosmic rays that may be present in either bin we take the greater of the two values as our s / n estimate .
the second cut removes all objects which show an apparently single smooth absorption trough . following this cut
we are left with only those objects which display multiple absorption trough features ( 1019 in total ) .
a third rejection cut removes those objects that contain 2 or more peaks within the absorption trough ( numbering 761 ) .
these objects are eliminated to reduce contamination by objects which display absorption from multiple out - flowing regions which by chance happen to lie at velocities corresponding to the ghost - zone .
this leaves just 258 objects which contain a single - peak ( or alternatively double - trough ) within the c iv absorption region .
figure [ fig : mt ] ( upper panel ) shows our multi - trough composite spectrum ( solid line ) created from taking the geometric mean of all objects rejected at this stage after normalising the flux between @xmath517001750 to unity . in order to highlight the absorption regions
, we also plot a non - balqso composite spectrum ( dashed line ) .
the non - balqso composite is the geometric mean of all of the quasar spectra in the sdss dr5 quasar catalogue in the redshift range 1.7@[email protected] excluding those objects which are also in the balqso catalogue of scaringi et al . , and reddened to fit to the continuum windows in the multi - trough composite spectrum using the smc extinction law of @xcite . for completeness
, the middle and lower panels show two example spectra rejected at his stage .
we note that our multi - trough composite spectrum [ fig : mt ] shows no evidence for coherent structure within the c iv trough .
this suggests that the positions of the various absorption systems in the individual spectra are in general uncorrelated , that is they do not contain a large number of ghost candidates that happen to show additional peaks within the absorption .
this is consistent with the hypothesis that these multiple - trough spectra are the result of multiple unconnected absorbing systems .
for the 258 objects which show a single peak in their c iv absorption we make a further cut to remove those objects ( 189 ) which show peaks outside of the ghost - zone as defined by @xcite .
the ghost zone edges are 5900 km s@xmath2 blue - ward of the peaks of the doublet of the emission line in question ( in this work c iv ) this zone is then expanded to take into account redshift errors by multiplying these boundaries by 1@xmath6 , where @xmath7 is the median redshift of our sample ( @xmath82.14 ) , and @xmath9 is the average redshift error ( @xmath100.01 ) .
this results in a ghost zone extending between @xmath31513.2 and @xmath31525.4 for c iv .
our single - peak composite comprising 189 objects whose peak lies outside of the ghost - zone is shown in figure [ fig : sp ] ( upper panel , solid line ) .
the middle and lower panels show example spectra of objects rejected at this stage .
each object shows a clear peak located just outside of the limits of the ghost zone .
interestingly , this composite spectrum shows some evidence of a feature at the low velocity edge of the ghost zone .
one possible explanation for this feature is that these objects are genuine ghost - candidates with poorly assigned redshifts .
however , on closer inspection of the objects rejected at this stage we find no evidence for features outside of the ghost zone due to random redshift errors .
alternative explanations for the origin of this feature include , ( i ) intrinsically weaker absorption at lower velocities , or ( ii ) an abrupt change in the intensity of the ( overlying ) emission at these velocities .
we make one further cut to remove those objects which do not show significant broad absorption either side of the ghost feature extending from 3000 to 9000 km s@xmath2 blue - ward of the emission - line .
this cut is chosen to eliminate objects with deep narrow absorption features whose alignment mimics the presence of a ghost feature .
figure [ fig : gzr ] ( upper panel , solid line ) displays the composite spectrum of all objects rejected at this stage , while in the middle and lower panels , we give examples of individual objects rejected by this cut . though the composite spectrum shows a clear feature within the ghost zone , it is not smooth , instead displaying sharp narrow features , whose separation is larger than the doublet separation of c iv .
this reflects both the narrowness of the absorption features ( since intrinsically broad humps would be smoothed out in the averaging process yielding low contrast features ) which make up this composite , and in addition small differences in their velocities relative to the rest - frame of the source . these differences may result from these features being unrelated to the ghost mechanism and appearing in the ghost zone by chance though we can not preclude the fact that some of the objects rejected at this stage are indeed true ghost - candidates .
our final sample of 43 ghost - candidate objects comprises the largest sample of objects exhibiting the ghost of ly-@xmath0 and the first to allow a statistical analysis of the properties of these objects .
figure [ fig : gz ] shows the geometric mean composite ghost - candidate spectrum ( solid line ) , together with a non - balqso composite spectrum ( dashed line ) , and two example ghost - candidate spectra ( middle and lower panels ) .
our final ghost - candidate composite spectrum displays a clear smooth double - trough structure located firmly within the ghost zone of a relatively broad absorption trough
. visual comparison of this composite with that formed from objects rejected at the previous stage ( i.e. those with insufficient absorption either side of the ghost - zone ) , indicates that our final sample displays on average narrower ly-@xmath0 and c iv emission - lines , stronger ly-@xmath0 emission and deeper
n v absorption ( fulfilling 3 of arav s criteria for forming observable ghost features ) , as well as broader si iv absorption , and suggests that we are indeed isolating those objects most likely to form observable ghost features . figure [ fig : ghostflow ] summarises the various cuts made in the creation of our ghost - candidate sample . of the 43 strong ghost - candidates , 21 are at large enough redshift @xmath11 to place ly-@xmath0 above the atmospheric cut - off and
thus allow us to test for each object all of the criteria deemed necessary for ghost - formation .
these are listed in the upper half of table 1 . for the remainder ,
their redshifts are too low to allow a direct measurement of the strength and width of the ly-@xmath0 emission - line . for these objects
we estimate the strength and width of the ly-@xmath0 emission - line using the c iv emission - line as a surrogate , and substantiated by known correlations between the two lines ( see e.g. wilkes 1984 , ulrich 1989 ) .
these objects are listed in the lower half of table 1 . for completeness
, we have also compiled a sample of balqsos showing no evidence for a ghost feature or multiple trough .
this allows us to compare the relative frequency with which the ghost - formation criteria are met amongst balqsos both with and without ghost features .
in order to create such a sample we have performed a number of cuts on the @xcite balqso sample . as for our ghost - candidate sample
we first select for s / n@xmath124 , and then select those objects with @xmath13 to allow inspection of the n v and ly-@xmath0 emission - lines necessary to test for the presence of strong narrow ly-@xmath0 emission and strong n v absorption .
the remaining spectra are then visually inspected and any object showing multiple absorption troughs or any other hint of a ghost feature are removed from the sample .
we also remove those objects which do not show significant absorption between 3000 and 9000 km s@xmath2 .
this leaves just 26 objects showing significant c iv absorption between 3000 and 9000 km s@xmath2 with no evidence for a ghost feature or multiple trough .
one of these objects sdss j101056.68 + 355833.3 shows evidence for a ghost feature in the si iv absorption line ( see a5 ) and is therefore also removed from our comparison sample leaving just 25 objects .
a summary of the rejection cuts made in the creation of our comparison sample is given in figure [ fig : comparison_flow ] .
the geometric mean composite spectrum of our comparison sample is indicated by the dotted line in figure 7 . aside from substantially weaker n v absorption , the strengths and widths of the emission - lines in the comparison sample composite spectrum match those of the ghost candidate composite spectrum remarkably well .
the main difference between these two composite spectra appears to be the absence of the low - velocity absorption trough in the comparison sample spectrum .
we do not believe this difference arises from the way in which the two samples were selected .
in 4 we use our ghost - candidate sample to test , where possible , the criteria set out by @xcite required by radiative acceleration models to produce an observable ghost feature .
we compare the relative frequency by which our ghost candidate spectra meet the criteria necessary for the formation of observable ghost features with our non - ghost comparison sample .
in this section we present the results of testing the criteria for the formation of observable ghost features .
we first concentrate on the sample of 21 ghost - candidate objects for which all of the criteria may be tested and compare these results with those obtained using our non - ghost comparison sample .
we then move on to our remaining objects and test where possible whether these objects satisfy the necessary criteria .
all of the objects in our ghost candidate sample will clearly show significant absorption in the region between 3000 and 9000 km s@xmath2 blue - ward of the c iv emission line as this is required in order to see any potential ghost feature .
this criterion is required purely because in order to effect the out - flowing c iv ions in an observable way this outflow must reach significant optical depth and appear as a bal trough within the spectrum .
our comparison sample has also been chosen to exhibit strong broad absorption , so naturally also fulfils this criterion .
( solid line ) , ( ii ) 25 bals with featureless broad absorption troughs ( dotted line ) , and ( iii ) non - balqsos from the dr5 qso catalogue ( dashed line).,scaledwidth=50.0% ] in order for the effects of ly-@xmath0 radiation on n v to cause a significant increase in the radiation pressure on the outflow as a whole the ly-@xmath0 emission line must be intrinsically strong .
however , since the ghost of ly-@xmath0 is produced by scattering ly-@xmath0 photons it can be extremely difficult to measure the strength of the intrinsic ly-@xmath0 emission - line .
hence @xcite modified this criterion to allow the c iv and n v emission - lines to be used in situations where the ly-@xmath0 emission - line was either significantly absorbed , or had insufficient spectral coverage .
this modified criterion requires ly-@xmath0 emission to be at least 300% above the continuum and/or both c iv and n v to be at least 100% above the continuum and n v emission - lines to be similarly narrow . ] .
they note that ` the latter criterion is hardly ideal but observations of non - bal quasars show that the line strengths roughly scale together ' ( arav et al .
1995 ) . here
we isolate the continuum emission using the specfit package in iraf .
we model the continuum as a reddened power law and fit to the continuum in regions free of contaminating emission- and absorption- lines , where possible
. this can be particularly difficult for objects with the highest velocity out - flows as there are few emission / absorption line - free regions .
furthermore , we have not attempted to fit for the continuum short - ward of @xmath31280 as this region is typically so absorbed and the line - emission so highly blended that no clean continuum regions can be identified .
instead , we fix the continuum strength at short wavelengths to that measured between @xmath513151350 .
this approach is chosen to avoid overestimating the strength of the broad absorption , at the expense of overestimating the strength of the line - emission in a few cases . in objects
whose spectra cover the ly-@xmath0 region , 86% ( 18/21 ) of the ghost candidate sample meet the criterion for strong emission - lines compared to 88% ( 22/25 ) of the comparison sample .
inspection of the composite spectra for the two samples ( figure [ fig : comp_samples ] ) , indicates that the strength of the ly-@xmath0 emission - line ( relative to the continuum ) is broadly similar in the two samples , and both are weaker relative to the non - balqso composite , due to the presence of strong n v and ly-@xmath0 absorption . for the other lines ,
the n v emission - line appears stronger on average in our ghost - candidate and comparison samples relative to non - balqsos , while there appears little difference in the strength of the c iv emission - line , though we note that in so far as we can measure them , their emission - line widths appear to be narrower on average .
models of outflows accelerated by radiative line driving produced by @xcite suggest that once the line widths exceed 3500 km s@xmath2 ( fwhm ) potential ghost features become undetectable .
this is because broader emission - lines will result in a broader , lower contrast feature ( since the core of the line , where most of the scattering occurs , is weaker relative to the wings in a broad emission - line c.f a narrow emission - line ) which is difficult to detect particularly in low s / n data .
unfortunately , determining the width of the underlying emission - lines in balqsos is extremely difficult as the broad absorption in these objects prevents accurate measurements of the shape and strength of the underlying emission - lines and continuum . given how our ghost - candidate sample was selected
we expect all of the ghost - candidates to show significant absorption in the blue wing of ly-@xmath0 , n v and c iv , while our comparison sample was chosen to exhibit strong relatively featureless absorption in the blue wing of c iv .
since we can not make an accurate measurement of the width of the ly-@xmath0 emission - line we instead attempt to estimate its likely width using the c iv emission - line width as a surrogate . for the majority of sources , c iv is severely absorbed , so we estimate its underlying un - absorbed width using model fits to the red wing of the emission - line .
we use the same underlying fit for the continuum as was used to measure the emission - line strength ( 4.2 ) .
we fit the red - wing with a two - component model , a gaussian core and lorentzian wings , with the central wavelengths fixed together but allowed to vary slightly from the rest wavelength of c iv .
we allow the strength and width of the two components to vary independently and fit to the red - wing taking care to avoid those regions of the spectrum contaminated by broad he ii @xmath31640 and [ o iii ] @xmath31663 emission . the measured width is taken to be the fwhm of the composite fit .
if the measured fwhm@xmath43500 km s@xmath2 , then the object meets the criterion for the formation of an observable ghost feature . for our 21 ghost - candidates with [email protected] , only 12 ( 57% )
have fwhm ( c iv ) @xmath4 3500 km s@xmath2 . in the remainder ,
those for which the c iv width can actually be measured ( 17 objects ) , 8/17 ( 47% ) also meet this criterion .
this fraction is similar to that found for our comparison sample where 12 out of the 20 objects ( 60% ) for which c iv width measurements can be made , also meet this criterion ( see table 2 for details ) .
taking the average of all of the fwhm measured for the ghost candidates gives an average fwhm of 3345@xmath14842 km s@xmath2 which is similar to the upper limit proposed by @xcite and to the average fwhm of c iv found for the comparison sample ( 3534@xmath141160 km s@xmath2 ) .
a k - s test on these populations shows no significant difference between the ghost and comparison samples .
the difficulty in fitting the c iv line in many of these objects results in large uncertainties in the estimated widths , however it is clear that several of our ghost candidates have emission that is wider than the 3500 km s@xmath2 upper limit proposed by @xcite .
there are a number of possible explanations for this .
firstly , c iv may be a relatively poor surrogate for the width of ly-@xmath0 .
alternatively , our sample may suffer from contamination by objects in which the potential ghost feature is produced by a chance alignment of multiple absorption systems or through other mechanisms unrelated to radiative acceleration .
if the potential ghost - features are indeed due to line - locking , then it may in - fact be possible to produce an observable ghost feature from emission - lines with widths in excess of 3500 km s@xmath2 .
as described in 2 , the ghost feature is produced by the increased radiation pressure on out - flowing n v ions by ly-@xmath0 photons . since in this model
the n v ions scatter a significant fraction of the incident flux , strong n v absorption is a pre - requisite . in order to measure
the strength of absorption due to n v we use a modified balnicity index ( weymann 1991 ) .
we take the continuum to be equal to the mean flux between @xmath513151330 , and require that the flux drops to below 90% of this value continuously for @xmath15 km s@xmath2 between 0 and 7000 km s@xmath2 blue - ward of the n v emission line .
we use the 7000 km s@xmath2 limit to avoid contamination by absorption due to ly-@xmath0 . any object with a non - zero value for this measure
is considered to show strong n v broad absorption .
significant n v absorption is measured in all 21 objects from our ghost candidate sample whose redshift allows detection of n v absorption .
only 44% ( 11/25 ) of our comparison sample satisfy this criterion .
the geometric composites of figure 7 , confirm that our ghost - candidate sample displays significantly stronger n v absorption on average than our comparison sample .
moreover , the broadly similar n v emission - line strengths of the ghost - candidate and comparison sample , emphasises the increased importance of n v absorption in the ghost - candidate spectra .
photo - ionisation models suggest that the far uv spectra of quasars contain a significant number of emission - lines between @xmath3200 and @xmath31000 that will contribute to the radiation pressure on any bal outflow ( korista et al .
1993 ) . in order for a ghost feature to be observed the radiation pressure of ly-@xmath0 on the n v ions in the outflow must be at least comparable in strength to the radiation pressure on all other ions in the outflow . this led arav to suggest that the far - uv continuum must therefore be necessarily weak in objects showing ghost features .
absorption of far - uv photons both within the host galaxy as well as within our own galaxy preclude direct measurement of the far - uv continuum .
however @xcite note that the strength of the he ii @xmath16
emission - line can be used as a surrogate for the far - uv ionising continuum .
a strong he ii line suggests a similarly strong far - uv flux and vice - versa as this line is produced by photons with energies corresponding to @xmath3228(@xmath1754 ev ) .
however , the he ii emission - line strength is notoriously difficult to measure even in high quality spectra .
this is because of its close proximity to the [ o iii]@xmath31663 emission - line , as well as its location near the red - wing of c iv , which may extend to relatively high velocities . thus isolating the broad he ii component generally requires multi - component fitting ( see e.g. goad and koratkar 1998 ) which without prior information ( for example , variability data )
, is rather subjective .
since the quality ( in terms of s / n ) of the individual spectra in both our ghost - candidate and comparison samples is relatively low , we do not attempt to fit to the individual line components . instead
, we use our non - balqso composite spectrum as a template for the region of interest , and scale each spectrum to this template , by minimising the residual flux in the region of interest .
our optimisation routine fits for the underlying continuum in both spectra , and adjusts for both redshift errors and the scale factor over the region of interest ( nominally @xmath51600 - 1680 ) .
since both spectra are first normalised to the flux in the continuum band between @xmath517001750 , if we assume that the he ii and [ o iii ] emission line strengths scale together , then the derived scaling factor between the ghost - candidate and non - bal qso spectra is a measure of the ratio of their he ii ews .
figure [ fig : heii_fit ] shows an example of a fit . in the top panel
we show the non - balqso composite ( solid line ) and a ghost - candidate spectrum ( dotted line ) . in the middle panel
we show the spectra together with their continuum fits spanning the c iv he ii region , while in the lower panel we show a close - up of the non - balqso composite and the scaled ghost - candidate spectrum after removal of the underlying continuum .
the residual flux under the he ii[o iii ] lines is indicated by the dot - dashed line .
we caution that this fitting process is sensitive to ( i ) small redshift errors between the spectra , ( ii ) an appropriate choice of continuum bands ( in general the choice of the continuum bands depend on the width of the c iv absorption trough , and most importantly ( iii ) the s / n of the spectra .
consequently the error on the scale factor is generally quite large . we have visually inspected the results of the fitting process for all of our ghost - candidate and comparison spectra to verify that the fitting procedure has converged correctly , adjusting the continuum bands , and region over which the spectra are scaled when necessary . in table 1
we list the he ii scale factors , that is , the multiplication factor necessary to produce the same equivalent width in the ghost - candidate spectrum .
a number less than 1 indicates that the ghost - candidate spectrum has larger equivalent width in the line in the region of he ii than the non - balqso spectrum .
conversely , numbers greater than 1 indicate that the ghost feature has a proportionately weaker he ii line . in order to determine error estimates on the derived scale factors , after determining a best - fit solution , we fix all parameters apart from the scale factor , and then minimise on this parameter only .
we then calculate the 90% confidence interval on this one interesting parameter from the best - fit model by varying the scale factor until the @xmath18 value has increased by 2.71 . both the best - fit scale factor and its range ( based on the 90% confidence interval ) ,
are given in table 1 and 2 .
for our sample of 21 ghost - candidates for which all criteria can be tested , 5 indicate a significantly weaker he ii ew , suggesting a weaker than average uv continuum .
only 5 objects show evidence for a stronger than average he ii ew . for the rest
, there is no significant difference ( within the errors ) in their he ii ews when compared to the non - balqso composite .
for the 22 ghost - candidates for which only some of the criteria can be tested , 5 show evidence for a weaker than average he ii ews , 6 indicate stronger than average he ii ews , and the rest indicate no significant difference in the he ii ew when compared to the non - balqso composite .
thus only 11/43 ( 25% ) of our ghost - candidate sample appear to satisfy the criteria for a weak uv continuum . repeating this test on our non - ghost balqso comparison sample
8 objects indicate weaker than average he ii ew , 9 objects have stronger than average he ii ew , and 6 objects show no discernible difference in their he ii ew relative to the non - balqso composite . for the remaining 2 objects ,
the fits did not converge , though visual inspection of one of these shows no evidence for significant he ii emission .
if we ignore the likely complex interaction between the ly-@xmath0 photons and nv ions ( ie .
we assume the optical depth in the flow is effectively constant with velocity ) , then naively one might expect that the ghost - feature should be at least as broad as the ly-@xmath0 emission - line responsible for its formation .
we note that in fact scattering is likely dominated by the line core where the optical depth is larger . by requiring that any observed ghost feature is at least as broad as ly-@xmath0
, we can further refine our ghost - candidate sample to leave only the best ghost - candidates . while the width of any ghost - feature is relatively easy to measure ( unless the s / n is low ) , we are once again limited by the accuracy to which we can measure the width of ly-@xmath0 , as this emission is absorbed in the process by which the ghost is formed . here
we again make use of the width of the c iv emission - line ( see 4.3 ) as a surrogate for the width of ly-@xmath0 .
we isolate the ghost feature by removing the underlying absorption trough by fitting a simple function ( cubic spline ) .
we then fit a single gaussian to the ghost feature to provide an estimate of its width .
the measured widths are listed in table 1 .
we find no correlation between the width of the ghost feature and the width of the c iv emission - line ( figure [ fig : fwhm ] ) .
instead , we find that in the vast majority of cases the ghost feature is narrower than the c iv emission - line ( the solid line in figure [ fig : fwhm ] is fwhm(c iv)@xmath19fwhm(ghost ) .
if we require that the fwhm of the ghost feature matches that of the c iv emission - line to within 1500 km s@xmath2 then 11/20 ( for one object we are unable to make emission - line width measurements ) of our ghost - candidates for which all ghost - formation criteria can be tested also match this criterion . for our remaining ghost - candidates 8/17 ( 5 objects have no emission - line width information )
also match this criterion . of the 11 candidates with similar ghost and c iv widths
for which all the criteria are testable all 11 show n v absorption , 10 show strong emission lines , 9 have narrow c iv emission but only 2 show evidence for weaker than average he ii emission . of these two one
has c iv fwhm of 4040 km s@xmath2 while the other fails the criteria for strong emission lines . ) of the c iv emission - line and the ghost feature .
the solid line shows fwhm(c iv)@xmath19fwhm(ghost).,scaledwidth=50.0% ]
for the 43 objects in our ghost candidate sample , 21 are at large enough redshift to allow us to test _ all _ of the criteria necessary for the formation of an observable ghost feature . of these
21/21 satisfy the condition for strong n v absorption .
18/21 also satisfy the condition for strong intrinsic ly-@xmath0 emission . however , less than half of the objects ( 12/21 ) satisfy the condition for narrow emission - lines , and even fewer ( 5/21 ) show weaker than average he ii ew .
while 11/21 objects satisfy the first 3 criteria , no single object meets all of the requirements for the formation of an observable ghost feature . for the remaining 22 objects , only 3 objects pass 2 of the first 3 criteria , and of these all show stronger than average he ii . of the 5 objects which indicate weaker than average he ii strengths ,
only one can be tested for any of the other criteria , in this case , the width of the emission - line , which it fails .
none of the objects in our ghost - candidate sample are present in the samples produced by @xcite or @xcite due to differences in the redshift ranges over which they were constructed .
north et al . ( 2006 ) identified 7 strong ghost - candidate spectra in dr3 of the sdss .
6 of those objects are also in our sample of 43 ghost - candidates .
the other , sdssj142050.34 - 002553.1 . has an assigned redshift of 2.085 in dr5 compared to the 2.103 used by @xcite and is therefore rejected by our ghost zone cut .
only one of north et al.s ghost zone final cut is of sufficient redshift to test all of the criteria for the formation of an observable ghost feature sdss j110623.52 - 004326.0 .
this object fails the narrow emission - line requirement . for the other 5 objects ,
3 fail the test for narrow emission - lines , and 4 fail the test for weaker than average he ii ew . for
our non - ghost comparison sample 12/25 meet the criterion for strong n v absorption .
20/25 meet the requirement of strong ly @xmath0 , while 12/22 meet the narrow emission - line width constraint .
9/25 objects show evidence for weaker than average he ii strengths .
however , only 3 objects meet 3 out of 4 criteria , and none of these meet the weaker than average he ii strength .
thus , _ none _ of the objects in our comparison sample meet all of the criteria necessary for the formation of an observable ghost feature .
a comparison between our ghost - candidate and non - ghost samples suggests that the main difference between them lies in the strength of the absorption , with our ghost candidate sample displaying more objects with strong n v absorption .
comparison of the geometric mean composite spectra of these two samples , and their ratio ( figure 13 dashed ( ghost - candidate ) and dotted ( non - ghost comparison spectra ) lines , indicates that the ghost - candidate composite shows significantly stronger absorption at lower velocities in all of the strong lines ly-@xmath0 , n v , si iv and c iv . in order to test whether peaks are more common within the ghost - zone than elsewhere , we repeat the ghost selection method using two `` fake '' ghost - zones ( red - ward and blue - ward of the original ghost - zone ) , in a similar fashion to north et al . ( 2006 ) .
these zones are created in precisely the same way as the ghost - zone except that the red zone is centred at 4000 km s@xmath2 while the blue zone is centred at 8000 km s@xmath2 blue - ward of the c iv emission - line .
of the 258 single peaked objects , 82 have peaks within the `` fake '' red zone and 50 have peaks within the fake `` blue - zone '' .
since the original ghost - zone contained 69 objects with single peaks , the evidence for an excess of objects with peaks at a preferred velocity is weak .
that is , single peaks within the ghost zone are no more likely than single peaks at other velocities .
we have also examined the link , if any , between the peaks within the c iv absorption trough and n v absorption . in order to select n v bals we use a modified balnicity index ( weymann 1991 ) .
we take the continuum to be equal to the mean flux between @xmath513151330 , and require that the flux drops to below 90% of this value continuously for @xmath15 km s@xmath2 between 0 and 7000 km s@xmath2 blue - ward of the n v emission line . in order to perform this test
we require objects with redshifts in excess of 2.15 . from a sample of 1747 objects with [email protected]
, we find 1258 n v bals and 489 n v non - bals .
27% ( 340 ) of the n v bals show multiple troughs in their c iv absorption compared to only 15.5% ( 76 ) of the n v non - bals .
similarly , 7.6% ( 95 ) of the n v bals and only 3.1% ( 15 ) of the n v non - bals show a single peak in the c iv absorption trough . among the n v non - bals
we find : i ) two objects with single peaks within the ghost - zone , ii ) two objects with a single peak in the fake blue zone , and iii ) no objects with a single peak in the fake red zone . for the n v bals
, we find : i ) 26 objects with single peaks within the ghost - zone , ii ) 14 with single peaks within the fake blue zone , and iii ) 33 with single peaks within the fake red zone . while these results indicate a strong link between the presence of n v absorption and the mechanism responsible for producing features within the c iv absorption trough , line - locking between ly-@xmath0 and nv does not appear to be the dominant mechanism . in summary ,
n v bals are more likely to have multiple troughs within the c iv absorption than n v non - bals ( factor of 2 ) .
further , n v bals are also more likely to display single - peaks within their c iv absorption than n v non - bals .
approximately 25% of the single peaked objects are located within the ghost - zone .
however , similar numbers are found in both the blue and red fake ghost zones .
thus while strong n v absorption appears to be a strong requirement for the appearance of features within the c iv absorption trough of balqsos , there is no preferred velocity for the location of these features .
in an analysis of the largest sample of balqsos derived from dr5 of sdss , we have identified 43 objects which show evidence for radiative driving , as indicated by the appearance of feature in the c iv absorption , and thought to be associated with the interaction between ly-@xmath0 photons and n v ions .
we have attempted to explore the reality of these ghost - features by testing the physically motivated criteria , as set out by arav , for the formation of observable features via this mechanism .
of the 21 objects for which we can test all of the criteria , _ none _ satisfy all of the conditions necessary for the formation of an observable feature .
two of the criteria , the width of the driving line , and the strength of the uv continuum are particularly difficult to measure . of the 5 objects that clearly pass the latter criteria , having weak he ii
, 4 have widths that are marginally larger than the upper limit imposed by arav necessary to produce strong ghost signatures , while the other appears to have only weak emission - lines .
we recognise that because of the large uncertainty on the measurements of the he ii strength , 8 further objects from the sample of 21 could potentially satisfy all of the criteria necessary for ghost - formation . addressing
this issue will require higher s / n data than currently available for these objects .
however , taken at face value either the conditions necessary for line - locking to occur are less stringent than previously proposed , or a large fraction of these objects are multi - trough interlopers masquerading as ghost - candidates .
this possibility is supported by our finding that the fake ghost zones bracketing the real ghost zone contain comparable numbers of objects as the real ghost zone .
this suggests that the peaks observed within bal troughs at the position of the ghost - zone are not a physical effect of radiative line driving , but are instead independent of any systematic velocity .
since our analysis has shed real doubt as to the existence of ly-@xmath0-n v line - locking features , in the attached appendix we repeat for the interested reader , the sequence of tests by korista et al .
( 1993 ) and re - investigate the incidence of double - trough phenomena in a large sample of balqsos . in summary , the results of these tests show that there is no strong evidence for an excess of double troughs bracketing the region of ghost - formation , and that the evidence for line - locking between ly-@xmath0 photons and n v ions in balqso spectra is substantially lacking .
l | c | c | c | l | l | c | object name & redshift & strong n v & strong ly @xmath0 & narrow emission - lines & he ii ew & fwhm ghost + & & bal & emission & fwhm ( km s@xmath2 ) & scale factor & ( km s@xmath20 + 024514.93 - 004101.9&2.802&@xmath21&@xmath21&x(3624 ) & x 1.06 ( 0.891.31 ) & 2646 + 080621.45 + 191441.0&2.272&@xmath21&x&x(4307 ) & @xmath21 1.58 ( 1.232.17 ) & 1785 + 083749.59 + 364145.4&3.416&@xmath21&x&@xmath21(2807 ) & @xmath21 2.67 ( 2.053.73 ) & 2005 + 084316.93 + 050640.1&2.409&@xmath21&@xmath21&@xmath21(2690 ) & x 0.96 ( 0.85 - 1.11 ) & 1744 + 085104.05 + 051539.8&3.213&@xmath21&@xmath21&x(4324 ) & @xmath21 1.56 ( 1.311.93 ) & 1013 + 090430.49 + 343108.9&3.408&@xmath21&@xmath21&x(4067 ) & x 0.89 ( 0.781.00 ) & 1696 + 092557.52 + 044035.9&2.266&@xmath21&@xmath21&x(5219 ) & @xmath21 1.41 ( 1.092.07 ) & 1248 + 093742.27 + 403351.2&2.258&@xmath21&@xmath21&@xmath21(2725 ) & x 1.00 ( 0.811.23 ) & 1125 + 094427.27 + 614424.6&2.337&@xmath21&@xmath21&@xmath21(2426 ) & x 0.76 ( 0.880.87 ) & 782 + 094454.85 + 481403.5&2.291&@xmath21&@xmath21&@xmath21(2504 ) & x 0.65 ( 0.550.72 ) & 1833 + 101844.45 + 544015.6&3.253&@xmath21&@xmath21&@xmath21(2729 ) & x 0.43 ( 0.370.53 ) & 2168 + 103450.21 + 313326.1&2.496&@xmath21&x&x(3943 ) & x 1.41 ( 0.99 - 2.48 ) & + 105058.28 + 614406.0&2.797&@xmath21&@xmath21&@xmath21(2375 ) & x 0.60 ( 0.530.69 ) & 2185 + 110623.51 - 004326.0@xmath22&2.443&@xmath21&@xmath21&x(4040 ) & @xmath21 3.35 ( @xmath121.82 ) & 2996 + 114704.47 + 153243.3&3.081&@xmath21&@xmath21&@xmath21(3258 ) & x 0.89 ( 0.791.00 ) & 771 + 131505.19 + 603716.4&2.330&@xmath21&@xmath21&x(3976 ) & x 0.97 ( 0.831.18 ) & 1313 + 134818.03 + 423205.1&3.066&@xmath21&@xmath21 & ? & x 0.89 ( 0.731.14 ) & 944 + 142244.45 + 382330.6&3.728&@xmath21&@xmath21&@xmath21(2146 ) & x 1.16 ( 0.931.51 ) & 2505 + 144032.04 + 060554.1&2.297&@xmath21&@xmath21&@xmath21(2723 ) & x 0.89 ( 0.791.00 ) & 1337 + 162657.47 + 405848.0&3.051&@xmath21&@xmath21&@xmath21(2520 ) & x 0.49 ( 0.440.55 ) & 1333 + 170322.41 + 231243.3&2.634&@xmath21&@xmath21&@xmath21(2138 ) & x 0.93 ( 0.791.05 ) & 1032
+ + object name & redshift & strong n v & strong ly @xmath0 & narrow emission - lines & he ii ew & fwhm ghost + & & bal & emission & fwhm ( km s@xmath2 ) & scale factor & ( km s@xmath20 + + 005109.45 + 001636.3&2.036&?&?&x(4850 ) & x 1.30 ( 0.942.00 ) & 2800 + 023252.80 - 001351.1@xmath22&2.033&?&?&x(4016 ) & x 0.94 ( 0.821.09 ) & 709 + 033048.50 - 002819.6@xmath22&1.779 & ? & ? & ?
& @xmath21 @xmath23 & 1977 + 081158.10 + 343624.1&2.106&?&@xmath21&@xmath21(3147 ) & x 1.13 ( 0.861.63 ) & 4395 + 082006.60 + 522158.7&2.015&?&?&@xmath21(3217 ) & ? ?
@xmath24 & 2660 + 091109.46 + 410155.2&2.128&?&@xmath21&x(4062 ) & x 0.56 ( 0.490.64 ) & 1006
+ 091313.09 + 411014.2&1.737 & ? & ? & ? & @xmath21 3.61 ( 2.625.72 ) & 2607
+ 092221.25 + 084312.8&1.969 & ? & ? & ? & @xmath21 @xmath121.43@xmath25 & 1398 + 102908.06 + 365155.2&1.750&?&?&x(3870 ) & x 0.31 ( 0.280.34 ) & 2432 + 110736.67 + 000329.4&1.741&?&?&@xmath21(2443 ) & x 1.18 ( 1.061.31 ) & 2041 + 111255.70 + 040600.5&1.989&?&?&x(3706 ) & x 0.96 ( 0.851.08 ) & 1055 + 113831.42 + 351725.3&2.118&?&@xmath21&@xmath21(2422 ) & x 0.78 ( 0.690.89 ) & 1336 + 122107.07 + 074437.0&1.900&?&?&@xmath21(3222 ) & x 0.91 ( 0.771.08 ) & 503 + 132304.58 - 003856.5@xmath22&1.827&?&?&x(4133 ) & x 1.09 ( 0.951.28 ) & 2077 + 133428.06 - 012349.0&1.876 & ? & ? & ? & @xmath21(??)@xmath26 & 1616 + 134458.82 + 483457.5&2.052 & ? & ? & ? & x 0.89 ( 0.731.12 ) & 1046 + 141843.95 + 373750.8&1.782&?&?&x(3596 ) & x 0.38 ( 0.330.44 ) & + 155505.10 + 442151.3&1.798&?&?&x(4425 ) & @xmath21 1.90 ( 1.512.59 ) & 1386 + 160335.43 + 225612.9&2.079&?&@xmath21&@xmath21(2479 ) & x 0.68 ( 0.600.78 ) & 2614 + 163231.60 + 294929.7&1.905&?&?&@xmath21(2587 ) & x 0.61 ( 0.540.66 ) & 2140 + 170056.85 + 602639.7@xmath22&2.123&?&@xmath21&x(4708 ) & x 1.15 ( 0.991.35 ) & 1913 + 172001.31 + 621245.7@xmath22&1.760&?&?&@xmath21(2528 ) & x 1.73 ( 1.242.95 ) &
2034 + + @xmath22 objects identified by north et al 2006 as their best ghost - candidates , their ghost candidate final cut ( gcfc ) .
+ @xmath24 poor continuum fit .
+ @xmath25 he ii absorbed .
+ @xmath26 he ii too weak to be measured .
+ .as for table 1 , indicating our comparison sample of 25 objects selected to have broad absorption and no ghost - features and the results of testing arav s criteria for ghost formation .
[ cols="<,^,^,^ , < , < " , ] since we only find strong evidence for the double - trough structure within our ghost zone sample ( chosen for precisely that reason ) , we conclude that there is indeed no strong evidence for an excess of objects with sub - troughs bracketing the ghost - zone among the balqso population at large .
indeed , the high probabilities found for both the lvq sample and for the multi - trough sample , strongly suggest that multi - trough features are randomly distributed in velocity .
we are therefore conclude that many of the objects previously identified as strong ghost - candidates in previous studies may simply be mt interlopers masquerading as ghosts due to the chance alignment of multiple un - associated absorption systems .
this is an object selected for our comparison sample as it shows no hint of a ghost feature within the deep broad c iv absorption .
the spectra is shown in figure [ fig : siiv ] along with the flux in velocity space relative to c iv and si iv showing clearly the potential ghost feature in the si iv absorption .
sdss j101056.68 + 355833.3 meets the criteria for strong emission due to c iv and n v emission exceeding 100% of the continuum flux at their peak , shows clear n v absorption , has c iv fwhm measured as 3483 km s@xmath2 and an ew of he ii measured to be [email protected] .
these measurements place this object very close to the boundary on several criteria and it would not be surprising to see a ghost feature in this object . while no ghost feature can be seen in the c iv trough , there is a clear feature within the si iv trough ( @xmath16200 km s@xmath2 ) within the ghost zone . due to the greater separation of the si iv doublet(1933 km s@xmath2 ) in comparison to the c iv doublet(498 km s@xmath2 ) a ghost feature in the si
iv would be expected to be wider and weaker .
this larger doublet separation is the reason for the wider ghost zone in the s iv bal trough .
the absence of a c iv feature could be due to an extremely high optical depth in the flow such that the flow remains optically thick even with the decrease in optical depth caused by the acceleration due to n v ions .
however the absorption troughs of c iv and si iv look very similar with the exception of the potential ghost feature suggesting the optical depths are similar . there is a slight hint of an additional feature within the si iv at @xmath18200 km s@xmath2 which is in the appropriate region to be due to the lower wavelength doublet feature from si iv . | we have searched the hybrid balqso catalogue of scaringi et al . derived from data release 5 of the sloan digital sky survey in order to compile the largest sample of objects displaying spectral signatures which may be indicative of radiative line driving .
the feature in question is the `` ghost of ly-@xmath0 '' , a line - locking feature previously identified in the broad c iv and si iv absorption lines of a small fraction of balqsos , and formed via the interaction of ly-@xmath0 photons with n v ions .
we test , where possible the criteria required to produce an observable ghost feature .
these criteria include : significant broad absorption , strong intrinsic ly-@xmath0 emission , narrow ly-@xmath0 , strong n v absorption , and a weak far - uv continuum .
no single ghost - candidate meets all of these criteria .
furthermore , we find that these criteria are not met significantly more often in ghost - candidates than in a comparison sample chosen to exhibit relatively featureless broad absorption troughs . indeed , the only significant differences we find between our ghost - candidate and comparison samples , is that on average , our ghost - candidate sample displays ( i ) significantly stronger n v absorption , and ( ii ) the onset of absorption occurs at lower velocities in our ghost - candidate objects .
significantly , we find no evidence for an excess of objects whose absorption troughs bracket the location of the ly-@xmath0nv line - locking region , rather the location of ghost - like features appears to be independent of any systematic velocity , with comparable numbers appearing both redward and blueward of the ghost - zone .
thus , the majority of objects identified here as strong ghost - candidates are likely multi - trough interlopers whose absorption feature simply bracket the region of interest .
[ firstpage ] galaxies : active , quasars : absorption lines , quasars : general |
abrikosov vortices in a superconductor are governed by two characteristic length scales .
the core of a vortex has a size dependent on the ginzburg - landau ( gl ) coherence length @xmath11 , while the supercurrents circulating around the core decay on the scale of the gl penetration depth @xmath12 . in the early days of
high-@xmath5 superconductivity it was common practice to infer the behavior of the in - plane magnetic penetration depth @xmath7 from measurements of the muon depolarization rate @xmath13 in the vortex state of polycrystalline samples .
the temperature dependence of @xmath13 was found to be consistent with @xmath14-wave pairing symmetry @xcite and a universal linear scaling of @xmath5 with @xmath13 was observed in the underdoped regime ( the so - called ` uemura plot ' ) , indicating that @xmath15 , where @xmath16 is the superfluid density.@xcite later , microwave @xcite and @xmath0sr @xcite measurements on yba@xmath2cu@xmath3o@xmath4 ( ybco ) single crystals in the meissner and vortex phases established a limiting low - temperature linear @xmath17 dependence of @xmath7 that is consistent with @xmath18-wave pairing . more recently , non-@xmath0sr studies of ybco in the meissner phase have revealed that @xmath5 has a sublinear dependence on @xmath6.@xcite on the other hand , the relation @xmath19 inferred from the uemura plot is supported by a recent study of electric - field induced superfluid density modulations in a single underdoped ultra - thin film of la@xmath20sr@xmath21cuo@xmath22.@xcite the problem with assuming @xmath23 is that there are additional inseparable contributions to @xmath13 from electronic magnetic moments and flux - line lattice ( fll ) disorder , which may vary with doping . to circumvent this difficulty we have studied ybco single crystals . in a single crystal the fll contribution to the @xmath0sr line shape @xmath1 is asymmetric and distinct from the other sources of field inhomogeneity.@xcite not only can the behavior of @xmath7 be isolated , but because the finite size of the vortex cores is apparent in a single - crystal measurement of @xmath1 ,
@xmath10 can be simultaneously determined .
while @xmath10 may be accurately determined in conventional superconductors from measurements of the upper critical field @xmath24 , in high-@xmath5 cuprates @xmath24 is generally a very high magnetic field marking the transition from a vortex liquid to the normal phase .
here we present @xmath0sr measurements that probe @xmath7 and @xmath10 in the bulk of ybco single crystals deep in the superconducting state .
the accuracy of our method was demonstrated in previous studies of conventional superconductors,@xcite but is reinforced here through comparisons with the results from other techniques .
ybco single crystals with purity of 99.995 % were grown by a self - flux method in fabricated bazro@xmath3 crucibles at the university of british columbia.@xcite an exception are @xmath25 single crystals that were grown in yttria - stabilized - zirconia crucibles and characterized by a purity greater than 99.5 % .@xcite single crystals of ca - doped ybco were also prepared in bazro@xmath3 crucibles .
typical sample sizes consisted of 3 to 5 single crystals from the same growth batch arranged in a mosaic to form a total @xmath26-@xmath27 surface area of 20 - 30 mm@xmath28 . the thickness of the crystals are on the order of @xmath29 mm .
the superconducting transition temperatures of the single crystals were measured using a squid magnetometer .
twin boundaries were removed from some of the higher - doped single crystals by applying pressure along the @xmath26 or @xmath27 directions at elevated temperature .
these basic sample characteristics are summarized in table [ samples ] .
cccc @xmath30 & @xmath31 & @xmath5 ( k ) & detwinned + 6.60 & 0.103 & 62.5 & + 6.57 & 0.110 & 59.0 & + 6.67 & 0.120 & 66.0 & + 6.75 & 0.132 & 74.6 &
@xmath32 + 6.80 & 0.140 & 84.5 & @xmath32 + 6.95 & 0.172 & 93.2 & @xmath32 + + ( y , ca)6.98 & 0.192 & 86.0 & + [ samples ] the @xmath0sr experiments were performed over a 3 year period on the m15 and m20b surface muon beam lines at triumf , vancouver , canada .
the @xmath0sr spectra were recorded in a transverse - field ( tf ) geometry with the applied magnetic field * h * perpendicular to the initial muon - spin polarization direction , and perpendicular to the @xmath33-axis of the single crystals . a tf-@xmath0sr spectrum comprised of 20 to 30 million muon decay events was taken at each temperature and magnetic field . in a transverse - field , the muon spin precesses in a plane perpendicular to the field direction ( which we define here as the @xmath34-direction ) .
the time evolution of the muon spin polarization @xmath35 is determined from the @xmath0sr `` asymmetry '' spectrum formed from the muon decay events detected in opposing positron counters @xmath36 where @xmath37 is the time histogram of the temporal dependence of decay positron count rate in the _
i_@xmath38 detector , and @xmath39 is the asymmetry maximum . in our experiments four
positron counters were used to completely cover the 360@xmath40 solid angle in the @xmath41-@xmath30 plane ( see fig . [ fig1 ] ) . the muon spin polarization function for the `` left''-``right '' pair of detectors is defined as @xmath42 and for the `` up''-``down '' pair as @xmath43 where @xmath44 is the muon gyromagnetic ratio , @xmath45 is a phase constant , and @xmath46 \rangle \ , , \label{distribution}\ ] ] is the probabilty of finding a local magnetic field @xmath47 in the @xmath34-direction at an arbitrary position * r * in the @xmath41-@xmath30 plane .
the tf-@xmath0sr time spectra for each sample were fit assuming the following analytical solution of the gl equations @xcite for the spatial field profile of the ideal fll @xmath48 where * g * are the reciprocal lattice vectors , @xmath49 is the average internal magnetic field , @xmath50 is a cutoff function for the * g * sum , @xmath51 is a modified bessel function , and @xmath52 .
the cutoff function @xmath53 depends on the spatial profile of the superconducting order parameter at the center of the vortex core . consequently , @xmath10 is a measure of the _ vortex core size_. the fll in all samples was assumed to be hexagonal .
neutron scattering experiments on fully - doped ybco @xcite indicate that the fll below @xmath54 koe is only slightly distorted from hexagonal symmetry due to @xmath55 anisotropy .
we find that accounting for this small distortion changes the values of @xmath10 and @xmath7 by less than 5 % .
consequently , the fll was assumed to be hexagonal for all samples studied .
= 5.0 in to avoid the difficulty of modelling the contribution of electronic magnetic moments to the @xmath0sr line shape , we restricted our study to ybco crystals free of static or quasistatic spins .
for the applied fields considered in this study , this has been determined to be the case for oxygen content @xmath56.@xcite to properly account for disorder of the fll , the dimensionality of the vortices must be considered .
josephson plasma resonance measurements on yba@xmath2cu@xmath3o@xmath57 ortho - ii single crystals grown by liang , bonn and hardy indicate that the vortices are 3d - like at low temperaturesi,@xcite while mutual inductance measurements on thin films by zuev _ et al _ show that even severely underdoped ybco is quasi-2d only near @xmath5.@xcite since the focus in the present study is on the variation of @xmath7 and @xmath10 in higher doped samples at low temperatures , the vortices are assumed to be rigid 3d lines of flux .
this assumption is also consistent with the observation of highly asymmetric @xmath0sr line shapes for all of our samples at low @xmath17 ( see fig . [ fig2 ] ) .
for rigid flux lines , random displacements of the vortices from their positions in the ideal hexagonal fll are accounted for by convoluting the theoretical line shape @xmath1 by a gaussian distribution of fields.@xcite a gaussian function also describes the local distribution of dipolar fields originating from static nuclear moments.@xcite taking into account both sources of line broadening , the corresponding theoretical polarization functions are @xmath58 where , @xmath59 is an effective depolarization rate due to nuclear dipole moments ( @xmath60 ) and fll disorder ( @xmath61 ) .
values for @xmath60 were obtained by fitting the tf-@xmath0sr signal above @xmath5 to the theoretical polarization function @xmath62 to account for the background signal from muons that did not stop in the sample , an additional term of the form @xmath63 was added to eq .
( [ eq : polartheory ] ) and to eq .
( [ eq : dipolar ] ) , where @xmath64 is the fraction of muons that stopped inside the sample .
values of @xmath64 for the different samples ranged from 0.8 to 0.9 .
in the present work there are two marked improvements over the analysis done in our previous studies of ybco single crystals @xcite that assumed the spatial field profile of eq .
( [ eq : glfield ] ) : ( i ) the earlier works used the asymptotic limit @xmath65 ( @xmath66 ) for the bessel function that appears in the cutoff @xmath67 , whereas here @xmath51 was evaluated numerically .
( ii ) the second improvement is that due to increased computer speed , @xmath68 was calculated at 15,132 equally - spaced locations in the rhombic unit cell of the hexagonal fll , compared to 5,628 locations in previous works . further increasing
the number of real - space points sampled in the fll unit cell did not result in appreciable changes in the fitted parameters .
we note that both improvements in our data analysis method influence the absolute values of @xmath7 and @xmath10 , but the temperature and magnetic field dependences of these parameters remain qualitatively similar to that determined in our previous studies .
figure [ fig3 ] shows the temperature dependence of @xmath7 at low @xmath17 determined at two different values of the applied magnetic field .
the solid curves are fits to @xmath69 , where @xmath70 and @xmath71 are field - dependent coefficients .
the dependence of @xmath6 on @xmath17 is shown in fig .
[ fig4 ] for selected values of the applied field .
an inflection point at @xmath72 k is visible in some of the lower field data .
this feature was also apparent in our previous measurements of yba@xmath2cu@xmath3o@xmath73.@xcite harshman _ et al _ have argued that the inflection point is caused by thermal depinning of vortices,@xcite although an invalid treatment of the data was used to support this assertion.@xcite recently , khasanov _ et al _ have ruled out depinning as the source of a similar inflection point in the temperature dependence of @xmath6 measured in la@xmath74sr@xmath75cuo@xmath22 by tf-@xmath0sr.@xcite instead they attribute this feature to the occurrence of both a large @xmath18-wave and a small @xmath14-wave superconducting gap . as we will show later , the anomalous magnetic field dependence of the vortex core size in ybco can be explained by an induced superconducting energy gap in the cuo chains that run along the @xmath76 direction .
theoretical calculations by atkinson and carbotte @xcite for a @xmath18-wave superconductor with proximity - induced superconductivity in the cuo chains , show that @xmath77 exhibits an inflection point caused by an upturn of @xmath78 at low @xmath17 ( where @xmath79 is the penetration depth in the @xmath76 direction ) .
= 4.4 in = 4.0 in = 4.0 in in fig .
[ fig5 ] it is shown that @xmath80 exhibits a near universal linear temperature dependence at low @xmath17 .
we attribute deviations from universal behavior near @xmath5 to softening of the fll , which narrows the @xmath0sr line shape and enhances the fitted value of @xmath7 .
the universal scaling implies that @xmath81 is a constant,@xcite where @xmath82 is the fermi velocity , @xmath83 is a velocity corresponding to the slope of the gap at the nodes , and @xmath84 is a charge renormalization parameterizing the coupling of the quasiparticles to phase fluctuations . using values of @xmath83 from thermal conductivity measurements,@xcite
we find that @xmath84 is bascially doping independent .
figure [ fig6 ] shows the magnetic field dependences of @xmath7 .
here we stress that the observed behaviors do not imply that the magnetic penetration depth or superfluid density depend on field in this way .
the sublinear dependence of @xmath7 on @xmath85 is primarily due to the failure of eq .
( [ eq : glfield ] ) to account for all field - dependent contributions to the internal magnetic field distribution . in refs .
@xcite the strong field dependence of @xmath7 in yba@xmath2cu@xmath3o@xmath73 determined by @xmath0sr was explained by the high anisotropy of the @xmath18-wave superconducting energy gap not accounted for in eq .
( [ eq : glfield ] ) . a nonlocal supercurrent response to the applied field in the vicinity of the vortex cores stemming from the divergence of the coherence length at the gap nodes
modifies the spatial distribution of field . with increasing @xmath85 ,
the increased overlap of the regions around the vortex cores reduces the width of the @xmath0sr line shape .
the gap anisotropy also results in a nonlinear supercurrent response to the applied field , resulting from a quasiclassical ` doppler shift ' of the quasiparticle energy spectrum by the flow of superfluid around a vortex.@xcite when the doppler shift exceeds the energy gap , cooper pairs are broken , and @xmath7 increases .
these effects are not restricted to @xmath18-wave superconductors .
sizeable nonlinear and/or nonlocal effects can also occur in @xmath14-wave superconductors with a smaller energy gap on one of the fermi sheets and/or a highly anisotropic fermi surface .
moreover , these anisotropies result in a rapid delocalization of quasiparticle core states with increasing @xmath85 that modify @xmath1 .
indeed , strong field dependences of @xmath7 from eq .
( [ eq : glfield ] ) have been observed in the multi - band superconductor nbse@xmath2,@xcite and the marginal type - ii superconductor v.@xcite it has been experimentally established for a variety of materials,@xcite including ybco,@xcite that the @xmath86 extrapolated value of @xmath7 agrees with the magnetic penetration depth measured by other techniques in the meissner phase .
consequently , we stress that only @xmath87 can be considered a `` true '' measure of the magnetic penetration depth .
= 3.5 in = 4.0 in in fig .
[ fig7 ] we show @xmath5 as a function of @xmath88 at two different fields .
the more inclusive data set at @xmath89 koe is described by @xmath90 , which deviates substantially from the linear scaling in the uemura plot .
the power 0.38 is surprisingly close to 0.43 determined by zuev _
et al _ in a meissner phase study of severely underdoped ybco thin films.@xcite it is surprising because these thin films have a superfluid density that is significantly lower than in single crystals.@xcite it is known from microwave studies of ybco that the doping dependences of @xmath91 and @xmath92 are not the same,@xcite due to the conductivity of the cuo chains.@xcite thus there is no reason to expect the power @xmath93 to be universal for the cuprates . while a sublinear dependence of @xmath5 on @xmath94 has also been inferred from more recent @xmath0sr measurements of the muon depolarization rate @xmath13 in other high-@xmath5 superconductors,@xcite the contributions of magnetism and fll disorder to the @xmath0sr line shape were not factored out .
= 3.8 in figure [ fig8 ] shows @xmath94 as a function of hole doping , where the values of @xmath31 are determined from the dependence of @xmath5 on @xmath31 presented in ref .
@xcite for similar ybco single crystals .
the behavior is consistent with other studies on cuprates indicating that the maximum value of @xmath94 is not reached before @xmath95.@xcite our data are in the range @xmath96 and are described by @xmath97 , except near @xmath98 .
= 4.0 in the hole doping dependences of @xmath60 and @xmath61 at @xmath89 koe are shown in fig . [
fig9](a ) . while @xmath60 is independent of @xmath31 , @xmath61 basically tracks @xmath6 . using the fitted values of @xmath61 and @xmath7
we have calculated the hole doping dependence of the root - mean - square displacement @xmath99 of the vortices from there positions in the perfect hexagonal fll . as shown in fig .
[ fig9](b ) , the degree of fll disorder is small and as expected highest in the ca - doped sample .
the field dependences of @xmath10 are shown in fig .
[ fig10 ] .
the increase in @xmath10 at low field , which corresponds to an expansion of the vortex cores , was previously reported for yba@xmath2cu@xmath3o@xmath100 ( ref .
@xcite ) and yba@xmath2cu@xmath3o@xmath73 ( refs .
@xcite ) . in all samples
we find that @xmath10 scales as @xmath101 , which is proportional to the intervortex spacing . in @xmath14-wave superconductors
the field dependence of @xmath10 has been shown to originate from the delocalization of bound quasiparticle core states.@xcite this is because a change in the spatial dependence of the pair potential accompanies the change in electronic structure of the vortex cores . in ybco the low - energy quasiparticle core states should be extended along the nodal directions of the @xmath18-wave gap function.@xcite this allows for a large transfer of low - energy quasiparticles between vortices at low field , which is further enhanced by an increase in vortex density .
hence the vortex core size is predicted to shrink with increasing field.@xcite however , the field dependence of @xmath10 in ybco is considerably stronger than predicted for a pure @xmath18-wave superconductor .
consequently , we consider an alternative explanation . = 3.8 in = 3.8 in in fig . [ fig11 ]
we show that the field - dependence of @xmath10 , in particular the upturn at low field , can be explained by the presence of the cuo chains .
the calculations of the core size @xmath102 are based on a semiclassical doppler - shift approximation ( see appendix ) for either a single - layer model representing a superconducting cuo@xmath2 plane or a proximity - coupled model representing a cuo@xmath2-cuo bilayer . in the bilayer model
there are two distinct energy scales for pair breaking : the energy gap associated with cooper pairs in the cuo@xmath2 planes , and a smaller proximity - induced gap associated with the chains .
it is the latter scale which is responsible for the expansion of the vortex cores at low field .
= 3.8 in = 4.0 in figure [ fig12 ] shows the hole doping dependences of @xmath103 at @xmath104 koe and @xmath105 koe .
qualitatively , the doping dependence of @xmath10 is similar to that reported by ando _
et al _ from magnetoconductance measurements on detwinned ybco single crystals.@xcite this result is also shown in fig .
[ fig12 ] , but plotted as @xmath10 versus @xmath30 .
note that our data must be plotted as @xmath10 versus @xmath31 , because the hole doping concentration of our @xmath25 single crystals ( grown in a different kind of crucible than the other samples ) , is smaller than that of @xmath106 ( see table [ samples ] ) .
the general trend of all data sets is an increase of @xmath10 with decreasing @xmath31 .
such behavior has also been observed in the underdoped regime of la@xmath20sr@xmath21cuo@xmath22,@xcite and bi@xmath2sr@xmath2cuo@xmath107.@xcite with increasing magnetic field , our values for @xmath10 approach those determined by ando _
et al_. note that based on our proximity - induced model for the field dependence of @xmath10 ( see appendix ) , it is the high - field values of @xmath10 that reflect the intrinsic superconductivity of the cuo@xmath2 planes . since the doping dependences of @xmath10 at @xmath89 koe and @xmath108 koe in fig .
[ fig12 ] are similar , we expect the hole doping dependence of @xmath109 to qualitatively resemble that of the upper critical field @xmath24 .
figure [ fig13 ] shows the hole doping dependence of @xmath110 calculated from the values of @xmath10 at @xmath89 koe .
consistent with the data of ando _ et al _ , @xmath24 decreases with decreasing @xmath31 in the underdoped regime of ybco , and displays a dip near 1/8 hole doping .
we have simultaneously determined the hole doping dependences of the magnetic penetration depth and the gl coherence length in the underdoped regime of yba@xmath2cu@xmath3o@xmath4 .
this was achieved by fitting @xmath0sr measurements in the vortex state to an analytical solution of the gl equations for the internal magnetic field distribution . in this type of analysis
the magnetic penetration depth is strictly defined as the @xmath9 extrapolated value of the fitted parameter @xmath7 .
the accuracy of this definition was established in previous studies of conventional superconductors.@xcite here we have presented measurements showing a refinement of the uemura plot for yba@xmath2cu@xmath3o@xmath4 , where @xmath5 is plotted as a function of the isolated quantity @xmath6 , rather than the the muon depolarization rate @xmath13 . we find that @xmath5 exhibits a strong sublinear dependence on @xmath6 , suggesting that @xmath5 is not directly proportional to the superfluid density @xmath16 .
this result supports the same conclusion reached in several recent meissner phase studies of ybco .
we have reported here a reduction of @xmath6 near 1/8 hole doping concentration .
suppression of @xmath94 or the muon depolarization rate @xmath13 near 1/8 hole doping has only previously been observed in cuprates where @xmath31 is controlled by cation substitution,@xcite and was believed to indicate a tendency toward static stripe formation .
static stripes over the doping range investigated here were recently ruled out by inelastic neutron scattering experiments on ybco.@xcite we ourselves find no evidence for static spins in zero - field @xmath0sr or tf-@xmath0sr measurements on our samples.@xcite however , the suppression of superconductivity near @xmath111 could be caused by fluctuating stripes , recently argued to be relevant in ybco and other cuprates.@xcite experimental evidence for dynamic stripes in ybco includes the detection of low - energy one - dimensional incommensurate modulations in yba@xmath2cu@xmath3o@xmath57 by inelastic neutron scattering.@xcite further evidence for suppression of superconductivity near @xmath112 is found in the hole doping dependence of @xmath10 . in our measurements @xmath10 is a parameter that characterizes the size of the vortex cores . while it mimics the behavior of the gl coherence length , @xmath10 is large at low field due to the contribution of the cuo chains to the spatial dependence of the superconducting order parameter .
enhancement of the gl coherence length or vortex core size near 1/8 hole doping , has also been observed in la@xmath20sr@xmath21cuo@xmath22.@xcite calculations by mierzejewski and mska show that static or quasistatic stripes actually intensify @xmath24 by reducing diamagnetic pair breaking,@xcite and hence can not explain the growth of @xmath10 near @xmath112 .
on the other hand , kadono _ et al _ have shown that an expansion of the vortex cores with decreasing hole doping can result from a strengthening of antiferromagnetic correlations competing with superconductivity.@xcite thus , dynamic stripes are a viable explanation for the increased size of the vortex cores near 1/8 hole doping .
we gratefully acknowledge d. broun , i. vekhter and a. j. millis for helpful and informative discussions .
we also thank y. ando for allowing us to reproduce his data here .
this work was supported by the canadian institute for advanced research and the natural sciences and engineering research council ( nserc ) of canada .
calculations of the vortex core size are based on a generalisation of the so - called `` doppler - shift '' approximation for the vortex structure @xcite to the case of ybco , which is a multiband superconductor . in the case of ybco ,
there is strong evidence that both the two - dimensional cuo@xmath2 planes and one - dimensional cuo chains superconduct .
furthermore , it is likely that the chains are intrinsically normal , but are driven superconducting by the proximity effect .
proximity models for ybco have been extensively described elsewhere @xcite .
the essential idea is that the superconductivity originates from a pairing interaction which is confined to the two - dimensional cuo@xmath2 planes , and that the mixing of chain and plane wavefunctions induces superconductivity in the one - dimensional chain layers . we adopt a simplified bilayer model consisting of a single plane and a single chain , with one wannier orbital retained per unit cell for each layer . for comparison purposes ,
calculations are also performed for a single - layer model of an isolated superconducting plane .
the bogoliubov - degennes hamiltonian for the bilayer is @xmath113 \hat \psi_j \label{eq : ham}\ ] ] where @xmath114 $ ] and @xmath115 ( @xmath116 ) are creation operators for quasiparticles ( quasiholes ) at lattice site @xmath117 in layer @xmath71 . here , we take @xmath118 for the plane layer and @xmath119 for the chain layer .
the parameters @xmath120 and @xmath121 are the single - electron hopping matrix elments between sites @xmath117 and @xmath122 within and between layers respectively , while @xmath123 is the superconducting order parameter along bonds connecting nearest neighbour sites @xmath117 and @xmath122 . from the form of eq .
( [ eq : ham ] ) , it is apparent that @xmath123 only couples quasiparticles belonging to the plane layer .
the single - layer hamiltonian is obtained by setting @xmath124 .
a magnetic field @xmath85 applied perpendicular to the layers induces circulating currents in the superfluid .
the superfluid velocity is given by @xmath125 where @xmath126 is the effective mass tensor , @xmath127 is the superfluid momentum , @xmath128 is the magnetic vector potential and @xmath129 is the local phase of the order parameter . in the limit that @xmath129 and @xmath130 are slowly varying functions , one can treat the superflow as uniform in the neighbourhood of @xmath131 .
then , one can make a local gauge transformation @xcite such that the phase is removed from the order parameter and appears instead in the hopping matrix elements @xmath132 : @xmath133 where @xmath134 are the hopping matrix elements in zero - field and @xmath135 are the matrix elements of the zero - field quasiparticle velocity . equation ( [ eq : linear ] ) follows from eq .
( [ eq : nonlinear ] ) in the limit that @xmath136 is small .
then , the order parameter takes on the simple @xmath18-wave form @xmath137 which , in reciprocal space , corresponds to @xmath138 $ ] , where @xmath139 is the lattice constant .
the local gauge transformation leads to a doppler - shifted spectrum and is exact in the limit of slowly varying superfluid velocity .
this procedure has been shown , in many circumstances , to provide a reasonable description of the vortex lattice @xcite .
we take band structures which are appropriate for ybco and adopt @xmath140 and @xmath141 for this work , we measure energies in units of @xmath142 and take @xmath143 . in reciprocal space , the dispersions of the isolated plane and chain layers are then @xmath144 + 4t_2\cos(k_xa)\cos(k_ya)$ ] and @xmath145 , respectively .
the chain - plane hopping matrix element @xmath146 is not well known in ybco and is taken to be @xmath147 .
we need to make an ansatz for @xmath148 . for a single vortex in an isotropic medium , one has @xmath154 , where @xmath155 is the azimuthal unit vector and the radius @xmath156
is measured relative to the centre of the vortex @xcite . for the bilayer model , however , @xmath148 is not isotropic : the chains provide a conduction channel along the @xmath157 direction which is in parallel with the isotropic plane conduction channel .
we mimic this anisotropy by assuming that the superfluid momentum will be similar to that of a single - layer superconductor with an anisotropic ( diagonal ) effective mass tensor @xmath126 with @xmath158 .
( for the single - layer model , we take @xmath159 . )
we then have two requirements which must be satisfied : @xmath160 and @xmath161 .
the first requirement introduces vortex cores at the vortex lattice sites @xmath162 , while the latter incompressibility requirement is strictly true in regions where @xmath163 is uniform .
this pair of equations is solved by @xmath164 where @xmath165 indicates that @xmath166 is excluded from the sum , @xmath167 are reciprocal lattice vectors of the magnetic unit cell ( we assume a square lattice here ) with area @xmath168 and magnetic length @xmath169 .
the results do not depend strongly on the ratio @xmath170 , which we take to be 0.6 for the parameters chosen above .
this choice minimizes @xmath171 , where @xmath172 is the total ( plane and chain ) current in the bilayer , @xmath173 and @xmath174 indicates the expectation value with respect to @xmath175 , eq .
( [ eq : ham2 ] ) . in principle , one could improve on the approximation of eq .
( [ eq : p ] ) by determining @xmath148 self - consistently from @xmath176 ; however this will not change the qualitative physics of the vortex core expansion .
we then solve self - consistently for the order parameter @xmath177 \langle \psi_{1-{\bf k}\downarrow } \psi_{1{\bf k}\uparrow}\rangle_{\bf r},\ ] ] with @xmath178 .
self - consistent solutions find that @xmath163 vanishes near the vortex core centre and obtains an asymptotic value @xmath179 far from the vortex core . in order to measure the vortex core size
, we define a quantity @xmath180 .
the vortex core size is then defined by the first moment of the radial coordinate @xmath156 with respect to @xmath181 : @xmath182 where @xmath183 corresponds to the vortex core centre . for presentation purposes , @xmath102 is shown relative to the bcs coherence length @xmath184 , where @xmath185 is the average of the fermi velocity on the fermi surface .
the magnetic field is related to the magnetic length by @xmath186 where @xmath187 is the superconducting flux quantum . for presentation purposes ,
@xmath85 is scaled by the upper critical field , @xmath188 , so that @xmath189 . d.r .
harshman , g. aeppli , e.j .
ansaldo , b. batlogg , j.h .
brewer , j.f .
carolan , r.j .
cava , m. celio , a.c.d .
chaklader , w.n .
hardy , s.r .
kreitzman , g.m .
luke , d.r .
noakes , and m. senba , phys . rev .
b * 36 * , 2386 ( 1987 ) .
uemura , v.j .
emery , a.r .
moodenbaugh , m. suenaga , d.c .
johnston , a.j .
jacobson , j.t .
lewandowski , j.h .
brewer , r.f .
kiefl , s.r .
kreitzman , g.m .
luke , t. riseman , c.e .
stronach , w.j .
kossler , j.r .
kempton , x.h .
yu , d. opie , and h.e .
schone , phys .
b * 38 * 909 ( 1988 ) .
uemura , g.m .
luke , b.j .
sternlieb , j.h .
brewer , j.f .
carolan , w.n .
hardy , r. kadono , j.r .
kempton , r.f .
kiefl , s.r .
kreitzman , p. mulhern , t.m .
riseman , d.ll .
williams , b.x .
yang , s. uchida , h. takagi , j. gopalakrishnan , a.w .
sleight , m.a .
subramanian , c.l .
chen , m.z .
cieplak , g. xiao , v.y .
lee , b.w .
statt , c.e .
stronach , w.j .
kossler , and x.h .
yu , phys .
* 62 * , 2317 ( 1989 ) .
uemura , l.p .
le , g.m .
luke , b.j .
sternlieb , w.d .
wu , j.h .
brewer , t.m .
riseman , c.l .
seaman , m.b .
maple , m. ishikawa , d.g .
hinks , j.d .
jorgensen , g. saito , and h. yamochi , phys .
lett . * 66 * 2665 ( 1991 ) .
sonier , r.f .
kiefl , j.h .
brewer , d.a .
bonn , j.f .
carolan , k.h .
chow , p. dosanjh , w.n .
hardy , r. liang , w.a .
macfarlane , p. mendels , g.d .
morris , t.m .
riseman , and j.w .
schneider , phys .
lett . * 72 * , 744 ( 1994 ) .
sonier , r.f .
kiefl , j.h .
brewer , d.a .
bonn , s.r .
dunsiger , w.n .
hardy , r. liang , w.a .
macfarlane , r.i .
miller , t.m .
riseman , d.r .
noakes , c.e .
stronach , and m.f .
white jr . , phys .
lett . * 79 * , 2875 ( 1997 ) .
m. sutherland , d.g .
hawthorn , r.w .
hill , f. ronning , s. wakimoto , h. zhang , c. proust , e. boaknin , c. lupien , l. taillefer , r. liang , d.a .
bonn , w.n .
hardy , r. gagnon , n.e .
hussey , t. kimura , m. nohara , and h. takagi , phys .
b * 67 * , 174520 ( 2003 ) .
r. kadono , w. higemoto , a. koda , m.i .
larkin , g.m .
luke , a.t .
savici , y.j .
uemura , k.m .
kojima , t. okamoto , t. kakeshita , s. uchida , t. ito , k. oka , m. takigawa , m. ichioka , and k. machida , phys .
b * 69 * , 104523 ( 2004 ) . | we report on muon spin rotation ( @xmath0sr ) measurements of the internal magnetic field distribution @xmath1 in the vortex solid phase of yba@xmath2cu@xmath3o@xmath4 ( ybco ) single crystals , from which we have simultaneously determined the hole doping dependences of the in - plane ginzburg - landau ( gl ) length scales in the underdoped regime .
we find that @xmath5 has a sublinear dependence on @xmath6 , where @xmath7 is the in - plane magnetic penetration depth in the extrapolated limits @xmath8 and @xmath9 .
the power coefficient of the sublinear dependence is close to that determined in severely underdoped ybco thin films , indicating that the same relationship between @xmath5 and the superfluid density is maintained throughout the underdoped regime .
the gl coherence length @xmath10 ( vortex core size ) is found to increase with decreasing hole doping concentration , and exhibit a field dependence that is explained by proximity - induced superconductivity on the cuo chains .
both @xmath7 and @xmath10 are enhanced near 1/8 hole doping , supporting the belief by some that stripe correlations are a universal property of high-@xmath5 cuprates . |
in recent years , multicarrier code - division multiple - access ( mc - cdma ) based on the quasi-/perfect- complementary sequence set ( in abbreviation , qcss / pcss ) has attracted much attention due to its potential to achieve low-/zero- interference multiuser performance @xcite , @xcite . here
, a qcss ( or pcss ) refers to a set of two - dimensional matrices with low ( or zero ) non - trivial auto- and cross- correlation sums @xcite@xmath7@xcite . in this paper , a complementary sequence
is also called a complementary matrix , and vice versa . to deploy a qcss ( or pcss ) in an mc - cdma system , every data symbol of a specific user
is spread by a complementary matrix by simultaneously sending out all of its row sequences over a number of non - interfering subcarrier channels .
because of this , the number of row sequences of a complementary matrix , denoted by @xmath2 , is also called the _ number of channels_. at a matched - filter based receiver , de - spreading operations are performed separately in each subcarrier channel , followed by summing the correlator outputs of all the subcarrier channels to attain a correlation sum which will be used for detection .
a pcss may also be called a mutually orthogonal complementary sequence set ( mocss ) @xcite@xmath7@xcite , a concept extended from mutually orthogonal golay complementary pairs ( gcps ) @xcite@xmath7@xcite . however , a drawback of pcss is its small set size @xcite . specifically , the set size ( denoted by @xmath1 ) of pcss is upper bounded by the number of channels , i.e. , @xmath8 .
this means that a pcss based mc - cdma system with @xmath2 subcarriers can support at most @xmath2 users only .
against such a backdrop , there have been two approaches aiming to provide a larger set size , i.e. , @xmath9 .
the first approach is to design zero- or low- correlation zone ( zcz / lcz ) based complementary sequence sets , called zcz - css @xcite , @xcite or lcz - css @xcite .
a zcz - css ( lcz - css ) based mc - cdma system is capable of achieving zero- ( low- ) interference performance but requires a closed - control loop to dynamically adjust the timings of all users such that the received signals can be quasi - synchronously aligned within the zcz ( lcz ) .
a second approach is to design qcss which has uniformly low correlation sums over all non - trivial time - shifts . as such
, qcss can be utilized to achieve low - interference performance with a simpler timing - control system . to the authors best knowledge , the first aperiodic correlation lower bound of qcss
was derived by welch in @xcite , which states : @xmath10 where every quasi - complementary sequence is a matrix of order @xmath11 ( thus , every row sequence has length of @xmath3 ) with assumed energy of @xmath12 .
the aforementioned set size upper bound of pcss , namely , @xmath8 , can also be obtained from ( [ welch_bound_for_cc ] ) by setting @xmath13 . on the other hand , if @xmath14 , one can show that @xmath9 , meaning that a larger set size can be supported by qcss .
recently , a generalized levenshtein bound ( glb ) for qcss has been derived by liu , guan and mow in [ [ liuguanmow14 ] , _ theorem 1 _ ] .
the key idea behind the glb ( including the levenshtein bound @xcite ) is that the weighted mean square aperiodic correlation of any sequence subset over the complex roots of unity should be equal to or greater than that of the whole set which includes all possible complex roots - of - unity sequences .
the levenshtein bound was extended from binary sequences to complex roots - of - unity sequences by bozta @xcite .
a lower bound for aperiodic lcz sequence sets was derived in @xcite by an approach similar to levenshtein s . in its bounding equation ,
glb is a function of the simplex " weight vector @xmath0 , the set size @xmath1 , the number of channels @xmath2 , and the row sequence length @xmath3 .
a necessary condition ( shown in [ [ liuguanmow14 ] , _ theorem 2 _ ] ) for the glb to be tighter than the welch bound is that @xmath5 , where @xmath15 with @xmath16 although a step - function " weight vector was adopted in [ [ liuguanmow14 ] , ( 34 ) ] , it only leads to a tighter glb for @xmath17 . as a matter of fact
, the tightness of glb remains unknown for @xmath18 when @xmath3 is sufficiently large .
the main objective of this paper is to optimize and then tighten the glb for _ all _ @xmath5 ( instead of _ some _ ) . for this
, we are to find a ( locally ) optimal weight vector which is used in the bounding equation .
a similar research problem was raised in @xcite for traditional binary sequences ( i.e. , non - qcss with @xmath19 ) .
see @xcite for more details .
the optimization of glb on qcss ( with @xmath4 ) , however , is more challenging because an analytical solution to a non - convex glb ( in terms of weight vector @xmath0 ) for _ all _ possible cases of @xmath20 is in general intractable .
we first adopt a frequency - domain optimization approach in section iii - b to minimize the ( non - convex ) fractional quadratic function of glb .
this is achieved by properly exploiting the specific structure of the circulant quadratic matrix in the numerator of the fractional quadratic term of glb . following this optimization approach
, we find a new weight vector which leads to a tighter glb for _ all _ @xmath20 cases satisfying @xmath5 and @xmath4 , asymptotically ( in @xmath3 ) .
our finding shows that the condition of @xmath5 , shown in [ [ liuguanmow14 ] , theorem 2 ] , is not only necessary but also sufficient , as @xmath3 tends to infinity .
moreover , in section iii - c , it is proved that the newly found weight vector is a local minimizer to the fractional quadratic function of glb , asymptotically .
we then examine in sections iv two weight vectors which were presented in @xcite for the tightening of the levenshtein bound on conventional single - channel ( i.e. , @xmath19 ) sequence sets .
we extend their tightening capability to glb on multi - channel ( i.e. , @xmath4 ) qcss , although the proof is not straightforward .
it is shown that each of these two weight vectors gives rise to a tighter glb ( over the welch bound ) for several small values of @xmath2 provided that @xmath5 .
it is also noted that the glb from the newly found weight vector is ( in general ) tighter than the glbs from these two ( earlier found ) weight vectors , as shown by some numerical results .
in this section , we first present some necessary notations and define qcss . then
, we give a brief review of glb .
for two complex - valued sequences @xmath21 $ ] and @xmath22 $ ] , their aperiodic correlation function at time - shift @xmath23 is defined as @xmath24 when @xmath25 , @xmath26 is called the aperiodic cross - correlation function ( accf ) ; otherwise , it is called the aperiodic auto - correlation function ( aacf ) . for simplicity
, the aacf of @xmath27 is denoted by @xmath28 .
let @xmath29 be a set of @xmath1 matrices , each of order @xmath11 ( where @xmath30 ) , i.e. , @xmath31_{m\times n } = \left [ \begin{matrix } c^\nu_{0,0 } & c^\nu_{0,1 } & \cdots & c^\nu_{0,n-1}\\ c^\nu_{1,0 } & c^\nu_{1,1 } & \cdots & c^\nu_{1,n-1}\\ \vdots & \vdots & \ddots & \vdots\\ c^\nu_{m-1,0 } & c^\nu_{m-1,1 } & \cdots & c^\nu_{m-1,n-1}\\ \end{matrix } \right ] , \end{split}\ ] ] where @xmath32 .
define the aperiodic correlation sum " of matrices @xmath33 and @xmath34 as follows , @xmath35 also , define the aperiodic auto - correlation tolerance @xmath36 and the aperiodic cross - correlation tolerance @xmath37 of @xmath38 as @xmath39 respectively .
moreover , define the aperiodic tolerance ( also called the maximum aperiodic correlation magnitude " ) of @xmath38 as @xmath40 .
when @xmath13 , @xmath38 is called a _ perfect complementary sequence set _ ( pcss ) ; otherwise , it is called a _ quasi - complementary sequence set _ ( qcss ) . in particular ,
when @xmath41 and @xmath42 , a pcss reduces to a matrix consisting of two row sequences which have zero out - of - phase aperiodic autocorrelation sums .
such matrices are called golay complementary matrices ( gcms ) or golay complementary pairs ( gcps ) in this paper , and either sequence in a gcp is called a golay sequence .
note that the transmission of a pcss or a qcss requires a multi - channel system .
specifically , every matrix in a pcss ( or a qcss ) needs @xmath4 non - interfering channels for the separate transmission of @xmath2 row sequences .
this is different from the traditional single - channel sequences with @xmath19 only .
let @xmath43^{\text{t}}$ ] be a simplex " weight vector which is constrained by @xmath44 define a quadratic function @xmath45 where @xmath46 is a @xmath47 circulant matrix with all of its diagonal entries equal to @xmath48 , and its off - diagonal entries @xmath49 , where @xmath50 and @xmath51 the glb for qcss over complex roots of unity in @xcite is shown below .
[ generalized_welch_bound_for_cc ] @xmath52.\ ] ] a weaker simplified version of ( [ generalized_welch_bound_for_cc - equ ] ) is given below .
@xmath53.\ ] ] setting @xmath54 , the glb reduces to the welch bound for qcss in ( [ welch_bound_for_cc ] ) .
[ rmk_nece_cond ] [ [ liuguanmow14 ] , _ theorem 2 _ ] for the glb to be tighter than the corresponding welch bound , it is _ necessary _ that @xmath5 , where @xmath6 is defined in ( [ nece_cond_qcssbd2 ] ) .
[ [ liuguanmow14 ] , _ corollary 1 _ ] applying the weight vector @xmath55 with @xmath56 where @xmath57 , to ( [ generalized_welch_bound_for_cc - equ ] ) , we have @xmath58 the lower bound in ( [ zl_corollary_4_equ ] ) is tighter than the welch bound for qcss in ( [ welch_bound_for_cc ] ) if one of the two following conditions is fulfilled : ( 1 ) : @xmath59 , @xmath4 and @xmath60 ( 2 ) : @xmath61 , @xmath4 and @xmath62 .
the necessary condition in _ remark [ rmk_nece_cond ] _ implies that for a given @xmath63 , the welch bound for qcss can not be improved if @xmath64 , where @xmath65 is defined in ( [ nece_cond_qcssbd2 ] ) . on the other hand , the weight vector in ( [ leven_lcz_weighting_vector ] )
can only lead to a tighter glb for @xmath17 .
because of this , the tightness of glb is unknown in the following ambiguous zone .
@xmath66 for sufficiently large @xmath3 , the above @xmath67 zone further reduces to @xmath68 by recalling ( [ nece_cond_qcssbd2 _ ] ) .
one may visualize this zone in the shaded area of fig .
[ fig_glbgap ] for @xmath69 .
we are therefore interested in finding a weight vector which is capable of optimizing and tightening the glb for _ all _ ( rather than _ some _ ) @xmath5 . relating this objective to fig .
[ fig_glbgap ] , such a weight vector can give us a tighter glb for the largest @xmath67 region right above the red diamond symbols .
however , the optimization of glb in ( [ generalized_welch_bound_for_cc - equ ] ) is challenging because its fractional quadratic term ( in terms of @xmath0 ) is indefinite .
more specifically , the quadratic term @xmath70 in the numerator is indefinite as some eigenvalues of the corresponding circulant matrix are negative when @xmath5 [ [ liuguanmow14 ] , appendix b ] .
it is noted that indefinite quadratic programming ( qp ) is np - hard @xcite , even it has one negative eigenvalue only @xcite . moreover , checking local optimality of a feasible point in constrained qp is also np - hard @xcite .
although some optimality conditions for constrained qp have been derived by bomze from the copositivity perspective @xcite , the situation becomes more complicated when indefinite fractional quadratic programming ( fqp ) problems are dealt with . according to @xcite
, glb may be classified as a standard fqp ( stfqp ) as the feasible set is the standard simplex . to the best of the authors knowledge ,
preisig pioneered an iterative algorithm for which convergence to a kkt point ( but can not be guaranteed to be a local minimizer ) of the stfqp can be proved @xcite .
two algorithms for stfqp based on semidefinite programming ( sdp ) relaxations are presented in @xcite , yet the optimalities of the resultant solutions are unknown . as a matter of fact , the algorithms developed in @xcite may only be feasible for medium - scaled stfqp with @xmath71 .
in contrast , we target at an analytical solution ( as opposed to a numerical solution ) which is applicable to large scale of glb ( e.g. , the sequence length @xmath72 ) .
thus , the techniques used in @xcite may not be useful for the specific stfqp problem considered in this paper . in the sequel
, we introduce a frequency - domain optimization approach which finds a local minimizer ( i.e. , a weight vector ) of the glb .
we show that the obtained weight vector leads to a tighter glb for _ all _ @xmath5 and @xmath30 , asymptotically . to tighten the glb in ( [ generalized_welch_bound_for_cc - equ ] )
, we adopt a novel optimization approach in this subsection , motivated by the observation that any circulant matrix [ e.g. , @xmath73 in ( [ leven - quad - fun - equ ] ) which forms a part of the glb quadratic function in ( [ generalized_welch_bound_for_cc - equ ] ) ] can be decomposed in the frequency domain .
define @xmath74 and the @xmath75-point discrete fourier transform ( dft ) matrix as @xmath76_{m , n=0}^{l-1},~\text{where}~f_{m , n}=\xi^{mn}_l.\ ] ] denote by @xmath77 the first column vector of @xmath46 in ( [ leven - quad - fun - equ ] ) , i.e. , @xmath78^{\text{t}}.\ ] ] let @xmath79^{\text{t}}.\ ] ] it is noted that @xmath80 . by @xcite , the circulant matrix @xmath73 defined in ( [ leven - quad - fun - equ ] ) can be expressed as @xmath81 where @xmath82^{\text{t}},\ ] ] and @xmath83 is the matrix with @xmath84 being the diagonal vector and zero for all the non - diagonal matrix entries .
consequently [ [ nyy2014 ] , theorem 3.1 ] , @xmath85 similarly , @xmath86 by [ [ liuguanmow14 ] , appendix b ] , we have @xmath87 and @xmath88 for @xmath89 .
note that @xmath90 for @xmath91 .
moreover , we remark that @xmath92 this is because 1 .
if @xmath93 is odd : @xmath94 2 . if @xmath93 is even : @xmath95 to maximize the glb in ( [ generalized_welch_bound_for_cc - equ ] ) , it is equivalent to consider the following optimization problem .
@xmath96 since @xmath0 is real - valued , @xmath97 is conjugate symmetric , i.e. , @xmath98 for @xmath99 . having this in mind ,
we define @xmath100 taking advantage of the fact that @xmath101 are strictly smaller than other @xmath102 s with nonzero @xmath93 as shown in ( [ lambda_l1_equ ] ) , we have @xmath103 where the equality is achieved if and only if @xmath104 for @xmath105 . inspired by this observation , we relax the non - negativity constraint on @xmath0 , i.e. , some negative @xmath106 s may be allowed ( but the sum of all elements of @xmath0 must still be equal to 1 ) . with this , the optimization problem in ( [ opti_freqd ] ) can be translated to @xmath107 where @xmath108 from now on , we adopt the setting of @xmath109 where @xmath110 denote the magnitude and phase of @xmath111 , respectively . since @xmath112 , we have @xmath113^{\text{t}}. \end{split}\ ] ] to optimize the fractional function in ( [ opti_freqd_trans ] ) , we have the following lemma .
[ lem_monofun ] the fractional function @xmath114 in terms of @xmath115 in ( [ opti_freqd_trans ] ) is 1 .
case 1 : monotonically decreasing in @xmath115 if @xmath116 and @xmath117 ; 2 .
case 2 : monotonically increasing in @xmath115 if @xmath118 , or @xmath5 and @xmath119 . to prove case 1 ,
we first show that @xmath120 if and only if @xmath121 where @xmath6 is defined in ( [ nece_cond_qcssbd2 ] ) . for ease of analysis , we write @xmath122 where @xmath123 is a positive integer and @xmath124 .
thus , @xmath125 .
consequently , we have @xmath126\\ & \leq \frac{n(mn-1)}{k } \left ( 1- \frac{n+1}{n+\epsilon } \right ) \\ & < 0 , \end{split}\ ] ] with which the proof of case 1 follows . the proof of case 2
can be easily obtained by following a similar argument . for case 2 of _ lemma [ lem_monofun ]
_ , it can be readily shown that the minimum of the fractional function @xmath114 in ( [ opti_freqd_trans ] ) is achieved at @xmath127 .
thus , the weight vector in ( [ wtvec3_equ ] ) reduces to @xmath128^{\text{t } } , \end{split}\ ] ] where the corresponding glb reduces to the welch bound in ( [ welch_bound_for_cc ] ) .
next , let us focus on the application of case 1 for glb tightening . in this case
, we wish to know the upper bound of @xmath115 in order to minimize the fractional function of @xmath115 in ( [ opti_freqd_trans ] ) .
coming back to the constraint of @xmath0 given in ( [ leven_weight_vector ] ) , @xmath129 and @xmath130 should satisfy @xmath131 thus , @xmath132 where the upper bound is achieved with equality when @xmath133 for any integer @xmath134 . by substituting @xmath135 into ( [ wtvec3_equ ] )
, we obtain the following weight vector .
@xmath136 where @xmath137 and @xmath134 is any integer .
the resultant glb from this weight vector is shown in the following lemma .
[ coro_glb_from_wetvec3 ] for @xmath116 and @xmath117 , we have @xmath138,\ ] ] where @xmath139 are given in ( [ lambdas ] ) .
to analyze the asymptotic tightness of the lower bound in ( [ coro_glb_from_wetvec3-equ ] ) , we note that when @xmath3 is sufficiently large , the second condition in _ lemma [ coro_glb_from_wetvec3 ] _ , i.e. , @xmath140 is true for @xmath5 . to show this
, we substitute @xmath139 into ( [ cond_case1 ] ) .
after some manipulations , one can see that the inequality in ( [ cond_case1 ] ) holds if and only if @xmath141 carrying on the expression in ( [ k_split_equ ] ) , we require @xmath142 which is guaranteed to hold for sufficiently large @xmath3 because @xmath143 is strictly smaller than 1 by assumption . furthermore , we note that @xmath144 therefore , @xmath145}=\frac{3m}{4k}+\frac{1}{2}-\frac{1}{\pi^2}.\ ] ] on the other hand , let us rewrite the welch bound expression ( [ welch_bound_for_cc ] ) as @xmath146 with @xmath147 then , @xmath148 with ( [ r1 ] ) and ( [ ch4_r1_largen ] ) , one can show that the lower bound in _
lemma [ coro_glb_from_wetvec3 ] _ is asymptotically tighter than the welch bound in ( [ welch_bound_for_cc ] ) if and only if the following equation is satisfied .
@xmath149 equivalently , we need to prove that for @xmath150 ( as @xmath151 ) , the following inequality holds .
@xmath152 one can readily show that the condition @xmath153 given in ( [ bdfromoptwtvec2 ] ) is true for _ all _ @xmath30 .
therefore , we have the following theorem .
[ th4optwtvec ] the glb in ( [ coro_glb_from_wetvec3-equ ] ) which arises from the weight vector in ( [ wgtvec3_equ ] ) reduces to @xmath154,\ ] ] for sufficiently large @xmath3 .
such an asymptotic lower bound is tighter than the welch bound for _ all _ @xmath5 and for _ all _ @xmath30 . in this subsection
, we prove the proposed weight vector in ( [ wgtvec3_equ ] ) is a local minimizer of the glb in ( [ generalized_welch_bound_for_cc - equ ] ) under certain condition .
we consider the weight vector @xmath0 by setting @xmath155 in ( [ wgtvec3_equ ] ) because other values of @xmath134 will lead to identical value of glb [ cf .
( [ glb_quadra_fd ] ) and ( [ glb_quadra_fd _ ] ) ] .
@xmath156 note that the frequency domain vector @xmath157 has @xmath158 and @xmath104 for all @xmath159 .
our problem in this subsection can be formally cast as follows .
define the fractional quadratic function @xmath160 is essentially the fractional quadratic term in ( [ generalized_welch_bound_for_cc - equ ] ) by replacing @xmath0 with @xmath161 . ] as follows .
@xmath162 where @xmath163 , @xmath73 is the circulant matrix defined in ( [ leven - quad - fun - equ ] ) which has order @xmath164 and with @xmath165 . when @xmath166 and @xmath63 becomes sufficiently large ,
prove that the weight vector @xmath0 in ( [ proposed_wgtvec ] ) is a local minimizer of @xmath160 , i.e. , @xmath167 holds for any feasible perturbation @xmath168 which has sufficiently small norm .
to get started , we define @xmath169 it is easy to show that ( [ local_mini_equ ] ) is equivalent to the following inequality .
@xmath170 let @xmath171 . since @xmath168 is a real vector , @xmath172 is conjugate symmetric in that @xmath173 for @xmath99 . by taking advantage of ( [ qa_in_freqdomain ] ) , we present the following properties which will be useful in the sequel .
@xmath174 by ( [ multi_equ4 ] ) , ( [ multi_equ5 ] ) , ( [ multi_equ7 ] ) and ( [ multi_equ8 ] ) , we have @xmath175 |e_i|^2\right \}. \end{split}\ ] ] by ( [ multi_equ3 ] ) , ( [ multi_equ5 ] ) , ( [ multi_equ6 ] ) and ( [ multi_equ7 ] ) , we have @xmath176 therefore , @xmath177 can be expressed in the form shown in ( [ gamma_equ ] ) .
@xmath178 since @xmath168 is a small perturbation , let us assume @xmath179 next , we proceed with the following two cases . 1 .
case i : if there exists @xmath180 for @xmath181 .
+ since we consider @xmath166 with sufficiently large @xmath63 , it is readily to show that @xmath182 holds for any @xmath181 [ see ( [ lambda_even_equ ] ) and ( [ lambda_odd_equ ] ) ] . by ( [ multi_equ4 ] ) , let us write @xmath183 where @xmath184 .
furthermore , write @xmath185\cdot ( 2n-1)=\lambda_1 a+b,\ ] ] where @xmath186 + [ rmk_on_ab ] since @xmath187 , @xmath188 and @xmath189 approach to + @xmath190 and @xmath191 , respectively , as @xmath2 grows sufficiently large .
+ to show ( [ local_mini_equ ] ) [ and ( [ local_mini_equ2 ] ) ] holds , we only need to prove the right - hand term of ( [ local_mini_equ3 ] ) divided by @xmath48 is nonnegative , asymptotically . for this , our idea is to consider a fixed @xmath3 ( sufficiently large ) and prove that : ( 1 ) @xmath192 is lower bounded by a nonnegative value determined by @xmath3 only ; ( 2 ) @xmath193 tends to zero ( with an upper bounded @xmath194 ) regardless the value of @xmath3 . + from ( [ lambda_i_equ ] ) , we have @xmath195 for ease of analysis , let @xmath3 be an even integer is odd , we can prove ( [ local_mini_equ ] ) [ and ( [ local_mini_equ2 ] ) ] holds by almost the same arguments . ] .
hence , @xmath196 .
since @xmath197 is a decreasing function of @xmath198 , we have @xmath199 also , @xmath200 where @xmath201 by noting @xmath202 ( @xmath203 a small positive angle ) and @xmath166 , we have @xmath204 by ( [ lambda_even_equ ] ) and ( [ lambda_odd_equ ] ) , we obtain @xmath205 on the other hand , @xmath206\\ \rightarrow & ~ 0^{- } , \end{split}\ ] ] where @xmath207 denotes a sufficiently small value ( negative ) that approaches zero from the left .
therefore , we have @xmath208\cdot ( 2n-1)}{a}\\ = & { \lim\limits_{m\rightarrow+\infty}\frac{\lambda_1}{a } } \cdot \underbrace{\lim\limits_{m\rightarrow+\infty}a}_{\text{upper bounded } } + { \lim\limits_{m\rightarrow+\infty}\frac{b}{a}}%\rightarrow \frac{\xi}{a}.% > \frac{2}{3}\cdot \left ( 2\sum\limits_{i=2}^{n-1}|e_i|^2 \right ) \end{split}\ ] ] by ( [ lim_xi_a_equ ] ) and ( [ lambda1_equ ] ) , we assert that when @xmath2 is sufficiently large , the sign of the limit in ( [ lim_gep_equ ] ) will be identical to that of @xmath209 [ cf .
( [ lim_xi_a_equ ] ) ] which is nonnegative .
this shows that ( [ local_mini_equ ] ) [ and ( [ local_mini_equ2 ] ) ] holds for case i , asymptotically .
2 . case ii : if @xmath210 for all @xmath181 .
+ in this case , ( [ ab_equ ] ) reduces to @xmath211 since @xmath171 , we have @xmath212 where @xmath213 denotes the real part of complex data @xmath203 . consider @xmath214 which takes the following form .
@xmath215 where @xmath216 and @xmath217 denotes the phase shift of @xmath214 .
as a result , @xmath218 can be expressed as @xmath219 thus , @xmath220 since @xmath221 , we assert that for sufficiently large @xmath63 , @xmath222 holds because it will be dominated by the negative @xmath223 .
our next task is to show that @xmath224 . by ( [ ei_equ ] ) , we have @xmath225 it is required in ( [ multi_equ2 ] ) that @xmath226 for all @xmath198 , i.e. , @xmath227 setting @xmath228 , we have @xmath229 setting @xmath228 , we have @xmath230 therefore , @xmath231 this shows @xmath232 holds provided @xmath233
. this can be easily satisfied by a sufficiently small @xmath234 .
together with ( [ lambda1ab_caseii])-([lambda1ab_caseii_3 ] ) , we conclude that ( [ local_mini_equ ] ) [ and ( [ local_mini_equ2 ] ) ] holds for case ii , asymptotically .
this completes the proof of the local optimality of the proposed weight vector in ( [ proposed_wgtvec ] ) . following a proof similar to the above
, one can easily show that the weight vector @xmath0 in ( [ proposed_wgtvec ] ) is also a local minimizer of the constrained qp of @xmath235 when @xmath166 and @xmath63 are sufficiently large .
in this section , we first consider another two weight vectors and study the tightness of their resultant glbs .
then , we compare them with the proposed weight vector in ( [ wgtvec3_equ ] ) by some numerical results . in @xcite , liu _
et al _ showed that the following positive - cycle - of - sine " weight vector @xmath0 @xmath236 where @xmath237 , asymptotically leads to a tighter levenshtein bound ( i.e. , @xmath19 ) for all @xmath238 @xcite . by [ [ liuparaguanbozas14 ] , _ proposition 1 _ ]
, one can show that the resultant glb from the weight vector in ( [ sine_shape_weight_vector ] ) can be written as follows .
[ new_lwerbd_from_new_wv ] @xmath239 , % ~~\text{for}~2\leq m\leq 2n-1,\ ] ] where @xmath240 in what follows , we analyze the asymptotic tightness of the lower bound in ( [ zl - corollary - new - weight - vector - equ ] ) .
define @xmath241 . obviously , @xmath129 is a real - valued constant with @xmath242 when @xmath243 is on the same order of @xmath244 ( i.e. , @xmath245 ) ; and @xmath246 when @xmath243 is dominated by @xmath3 asymptotically ( i.e. , @xmath247 ) .
furthermore , define the fractional term in ( [ zl - corollary - new - weight - vector - equ ] ) as @xmath248 it is easy to see that the lower bound in ( [ zl - corollary - new - weight - vector - equ ] ) is tighter than the welch bound in ( [ welch_bound_for_cc ] ) if and only if @xmath249 where @xmath250 is defined in ( [ r1 ] ) . as @xmath3 tends to infinity , the inequality in ( [ iff ] ) is equivalent to @xmath251 when @xmath247 , we have @xmath246 and @xmath252 as @xmath253 . in this case , one can show that @xmath254 which can be ignored without missing the minimum point of interest in the right - hand side of ( [ asymp_iff ] ) .
hence , we shall assume @xmath129 to be a non - vanishing real - valued constant with @xmath255 , and rewrite ( [ asymp_iff ] ) as @xmath256 here , the order of the limit and minimization operations can be exchanged because @xmath257 as a function of @xmath129 exists , as shown below .
next , noting that @xmath258 , we can express ( [ r2 ] ) as @xmath259 where @xmath260 and after some manipulations , @xmath261 by ( [ ch4_r1_largen ] ) , ( [ r2_2a ] ) and ( [ ch4_q_largen ] ) , it follows that ( [ asymp_iff _ ] ) reduces to @xmath262 equivalently , we assert that the asymptotic lower bound in ( [ zl - corollary - new - weight - vector - equ ] ) is tighter than the welch bound if and only if @xmath263 where @xmath264 in fig .
[ fig_optimal_r2 ] , @xmath265 and @xmath266 versus @xmath129 over the range of @xmath267 are plotted
. it can be obtained from ( [ overlinek ] ) and fig . [ fig_optimal_r2 ]
that @xmath268 by ( [ asym_bd_wtvec2_equ ] ) , one can see that the proposed weight vector in ( [ sine_shape_weight_vector ] ) asymptotically leads to a tighter glb for _ all _ @xmath5 if and only if the value of @xmath2 satisfies the following condition [ c.f .
( [ nece_cond_qcssbd2_2 ] ) ] @xmath269 in fig .
[ fig_dm ] , @xmath270 versus @xmath2 is also plotted . by identifying @xmath2 satisfying @xmath271 [ shown in ( [ ch5_dm_equ ] ) ] , we arrive at the following theorem .
[ tighter_m_rmk ] the glb in ( [ zl - corollary - new - weight - vector - equ ] ) which arises from the weight vector in ( [ sine_shape_weight_vector ] ) reduces to @xmath272,\ ] ] for sufficiently large @xmath3 , where @xmath273 is given in ( [ ch4_q_largen ] ) . such an asymptotic lower bound is tighter than the welch bound for _ all _ @xmath5 if and only if @xmath274 let us consider the weight vector obtained by minimizing the following function using the lagrange multiplier .
@xmath275 where @xmath276 for @xmath277 and @xmath237 .
the idea is to optimize the weaker glb in ( [ simplified_glb ] ) . by relating the quadratic minimization solution of @xmath278 to the chebyshev polynomials of the second kind ,
one can obtain the weight vector , lemma 2 ] , such a weight vector is more generic as it applies to qcss with different @xmath30 .
] below .
let @xmath279 and @xmath280 .
also , let @xmath243 be an even positive integer with @xmath281 .
for @xmath282 , define the following weight vector @xmath283 setting @xmath284 , one can minimize @xmath278 in ( [ mini_f ] ) over different @xmath243 and get a generalized version of the levenshtein bound in [ [ levenshtein99 ] , _ corollary 4 _ ] as follows .
[ coro_bd1 ] @xmath285 as @xmath253 , the lower bound in ( [ lev_bd_cor4 ] ) is tighter than the welch bound in ( [ welch_bound_for_cc ] ) if and only if @xmath286 or equivalently , @xmath287 where the right - hand side of ( [ equ_coro2_tighter ] ) is obtained from ( [ lev_bd_cor4 ] ) .
recall that as @xmath253 , a necessary condition ( cf .
_ remark [ rmk_nece_cond ] _ ) for the glb to be tighter than the corresponding welch bound is @xmath288 clearly , @xmath289 which is smaller than the right - hand side of ( [ equ_coro2_tighter2 ] ) .
it can be asserted that the resultant glb obtained from the weight vector in ( [ levwtvec_equ ] ) with @xmath290 is tighter if and only if the value of @xmath2 satisfies the condition @xmath291
> 0.\ ] ] this is because when condition ( [ ch5_dm_equ2 ] ) is satisfied , @xmath292 is not only a necessary condition [ cf .
( [ nece_cond_qcssbd2_2 ] ) ] but also a sufficient condition [ cf .
( [ equ_coro2_tighter2 ] ) ] for the glb to be asymptotically tighter than the welch bound . in fig .
[ fig_dm ] , @xmath293 versus @xmath2 is plotted . by identifying @xmath2 satisfying @xmath294 [
shown in ( [ ch5_dm_equ2 ] ) ] , we have the following theorem .
[ tighter_m_rmk2 ] the glb in ( [ lev_bd_cor4 ] ) which arises from the weight vector in ( [ levwtvec_equ ] ) is asymptotically tighter than the welch bound for _ all _ @xmath5 if and only if @xmath295 denote by @xmath296 the optimized asymptotic lower bounds in ( [ aym_lwrbd_wec3_equ ] ) , ( [ aym_lwrbd_wec2_equ ] ) , ( [ lev_bd_cor4 ] ) , respectively .
we remark that ( 1 ) , both @xmath297 and @xmath298 are greater than @xmath299 for any @xmath4 ; ( 2 ) , @xmath300 except for @xmath301 .
the proof is omitted as it can be easily obtained from the tightness analysis in section iii - b and section iv . to further visualize their relative strengths of these three lower bounds ,
we calculate in table i the ratio values of @xmath302 with @xmath303 , where @xmath304 and @xmath305 denotes the corresponding welch bound . a ratio value which is larger than 1 corresponds to a tighter glb ( over the welch bound ) . with table
i , one may verify the three sets of @xmath2 for tighter glb in _ theorems 1 - 3 _ as well as the above - mentioned remark in this subsection .
in particular , we can see that @xmath306 for all @xmath4 , showing that weight vector 1 is superior than the other two as it is capable of tightening the glb for all possible @xmath2 , asymptotically .
[ table_of_small_value ] [ cols="^,^,^,^,^,^,^,^,^,^,^,^,^,^",options="header " , ]
the generalized levenshtein bound ( glb ) in [ [ liuguanmow14 ] , _ theorem 1 _ ] is an aperiodic correlation lower bound for quasi - complementary sequence sets ( qcsss ) with _ number of channels _ not less than 2 ( i.e. , @xmath4 ) .
although glb was shown to be tighter than the corresponding welch bound [ i.e. , ( [ welch_bound_for_cc ] ) ] for certain cases , there exists an ambiguous zone [ shown in ( [ glbgap _ ] ) and ( [ glbgap ] ) ] in which the tightness of glb over welch bound is unknown . motivated by this
, we aim at finding a properly selected weight vector in the bounding equation for a tighter glb for _ all _ ( other than _ some _ ) @xmath5 , where @xmath1 denotes the set size , and @xmath6 is a value depending on @xmath2 and @xmath3 ( the sequence length ) . as the glb is in general a non - convex fractional quadratic function of the weight vector ,
the derivation of an analytical solution for a tighter glb for _ all _ possible cases is a challenging task .
the most significant finding of this paper is weight vector 1 in ( [ wgtvec3_equ ] ) which is obtained from a frequency - domain optimization approach .
we have shown that its resultant glb in ( [ coro_glb_from_wetvec3-equ ] ) is tighter than welch bound for _ all _ @xmath5 and for _ all _ @xmath4 , asymptotically .
this finding is interesting as it explicitly shows that the glb tighter condition given in [ [ liuguanmow14 ] , _ theorem 2 _ ] is not only necessary but also sufficient , asymptotically , as shown in _ theorem [ th4optwtvec]_. interestingly , we have proved in section iii - c that weight vector 1 in ( [ wgtvec3_equ ] ) is local minimizer of the glb under certain asymptotic conditions .
we have shown that both weight vectors 2 and 3 [ given in ( [ sine_shape_weight_vector ] ) and ( [ levwtvec_equ ] ) , respectively ] lead to tighter glbs for _ all _ @xmath5 but only for certain small values of @xmath2 not less than 2 .
note that although they were proposed in @xcite , the focus of @xcite was on the tightening of levenshtein bound for traditional single - channel ( i.e. , @xmath19 ) sequence sets , whereas in this paper we have extended their tightening capability to glb for multi - channel ( i.e. , @xmath4 ) qcss .
furthermore , we have shown in _ theorem [ tighter_m_rmk ] _ and _ theorem [ tighter_m_rmk2 ] _ that weight vector 2 is superior as its admissible set of @xmath2 [ see ( [ distri_optimal_m ] ) ] is larger and subsumes that of weight vector 3 .
chen , j .- f .
yeh , and n. suehiro , a multicarrier cdma architecture based on orthogonal complementary codes for new generations of wideband wireless communications , " _ ieee commun .
126 - 135 , oct . 2001 .
z. liu , u. parampalli , y. l. guan , and s. bozta , `` constructions of optimal and near - optimal quasi - complementary sequence sets from singer difference sets , '' _ ieee wireless commun .
letters , _ vol .
487 - 490 , oct .
2013 . h. h. chen , s. w. chu , and m. guizani , on next generation cdma techonogies : the real approach for perfect orthogonal code generation , " _ ieee trans .
5 , pp . 2822 - 2833 , sep . 2008 .
j. li , a. huang , m. guizani , and h. h. chen ,
inter - group complementary codes for interference - resistant cdma wireless communications , " _ ieee trans .
wireless commun .
_ , vol . 7 , no .
166 - 174 , jan . 2008 .
s. bozta , new lower bounds on aperiodic cross - correlation of codes over @xmath123th roots of unity , " research report 13 , department of mathematics , royal melbourne institudte of technology , australia , 1998 .
[ liuparaguanbozas14 ] z. liu , u. parampalli , y. l. guan and s. bozta , a new weight vector for a tighter levenshtein bound on aperiodic correlation , " _ ieee trans .
theory , _ vol .
2 , pp . 1356 - 1366 , feb . 2014 .
j. c. preisig , copositivity and the minimization of quadratic functions with nonnegativity and quadratic equality constraints , " _
siam j. control and optimization _ , vol .
1135 - 1150 , 1996 . [ nyy2014 ] n. y. yu , a fourier transform approach for improving the levenshteing lower bound on aperiodic correlation of binary sequences , " _ advances in mathematics of communications _ , vol
209 - 222 , 2014 . | a quasi - complementary sequence set ( qcss ) refers to a set of two - dimensional matrices with low non - trivial aperiodic auto- and cross- correlation sums . for multicarrier code - division multiple - access applications ,
the availability of large qcsss with low correlation sums is desirable .
the generalized levenshtein bound ( glb ) is a lower bound on the maximum aperiodic correlation sum of qcsss .
the bounding expression of glb is a fractional quadratic function of a weight vector @xmath0 and is expressed in terms of three additional parameters associated with qcss : the set size @xmath1 , the number of channels @xmath2 , and the sequence length @xmath3 .
it is known that a tighter glb ( compared to the welch bound ) is possible only if the condition @xmath4 and @xmath5 , where @xmath6 is a certain function of @xmath2 and @xmath3 , is satisfied .
a challenging research problem is to determine if there exists a weight vector which gives rise to a tighter glb for _ all _ ( not just _ some _ )
@xmath5 and @xmath4 , especially for large @xmath3 , i.e. , the condition is asymptotically both necessary and sufficient . to achieve this
, we _ analytically _ optimize the glb which is ( in general ) non - convex as the numerator term is an indefinite quadratic function of the weight vector .
our key idea is to apply the frequency domain decomposition of the circulant matrix ( in the numerator term ) to convert the non - convex problem into a convex one . following this optimization approach , we derive a new weight vector meeting the aforementioned objective and prove that it is a local minimizer of the glb under certain conditions .
fractional quadratic programming , convex optimization , welch bound , levenshtein bound , perfect complementary sequence set ( pcss ) , quasi - complementary sequence set ( qcss ) , golay complementary pair . |
in this paper we discuss two models . the first model is a minijet model@xcite .
the second is a glasma flux tube model ( gftm)@xcite .
the paper is organized in the following manner : sec .
1 is the introduction and review of models .
2 discuss two particle angular correlation in the two models .
3 discuss baryon and anti - baryon formation in both models .
sec . 4 demonstrates how the ridge is formed by flux tubes when a jet trigger is added to the gftm . sec .
5 treats the so - called mach cone effect by analyzing three particle angular correlations in the two models .
6 presents the summary and discussion .
the analysis of angular correlations led to unanticipated structure in the final state of p - p and au - au collisions , subsequently identified with parton fragmentation in the form of minijets[3 - 7 ] .
two - component analysis of p - p and au - au spectra revealed a corresponding hard component , a minimum - bias fragment distribution associated with minijets , suggesting that jet phenomena extend down to 0.1 gev / c@xcite . from a given p - p or au - au collisions particles
are produced with three kinematic variables ( @xmath4,@xmath5,@xmath6 ) .
the @xmath6 spectra provide information about parton fragmentation , but the fragmentation process is more accessible on a logarithmic momentum variable .
the dominant hadron ( pion ) is ultrarelativistic at 1 gev / c , and a relativistic kinematic variable is a reasonable alternative .
transverse rapidity in a longitudinally comoving frame near midrapidity @xmath4=0 is defined by @xmath7/m_0),\ ] ] with @xmath8 a hadron mass and @xmath9 = @xmath10 + @xmath11 .
if one integrates over @xmath5 the event multiplicity ( @xmath12 @xmath13 @xmath14 , the 1d density on @xmath4 ) can be written for p - p collision in a two component model as @xmath15 @xmath16 is a levy distribution on @xmath17 which represents soft processes and @xmath18 is a gaussian on @xmath19 which represents hard processes .
the soft process has no @xmath5 dependence but does have a gaussian correlation in the longitudinal direction(@xmath4 ) .
this correlation can be expressed in a two particle way as @xmath20 which is the difference of the psuedorapidity .
@xmath21.\ ] ] the hard process arise from a scattering of two partons thus minijets are formed .
each minijet fragments along its parton axis and generates a 2d correlation @xmath22 and @xmath23 .
@xmath24exp[-\delta \eta^2/2\sigma_{\eta}^2].\ ] ] since for every minijet fragmenting leading to a peak at @xmath25 = @xmath26 , there is its scattered partner in the backward direction @xmath25 = @xmath27 .
the backward scattered minijet will range over many psuedorapidity values , therefore its correlation with the fragmentation of the near side minijet will have a broad @xmath28 width . in this condition
simple momentum conservation is good enough and that is a @xmath29 term .
the two - component model of hadron production in a - a collisions assumes that the soft component is proportional to the participant pair number ( linear superposition of n - n collisions ) , and the hard component is proportional to the number of n - n binary collisions ( parton scattering)@xcite .
any deviations from the model are specific to a - a collisions and may reveal properties of an a - a medium . in terms of mean participant path length @xmath30 = 2@xmath31/@xmath32 the @xmath6-integrated
a - a hadron density on @xmath4 is @xmath33 by analogy with eg .
( 2 ) a - a density as a function of centrality parameter @xmath30 becomes @xmath34 in the above equation @xmath35 = @xmath36 and @xmath37 = @xmath38 which is p - p scattering or nucleon - nucleon scattering .
we also can define for the hard density for a - a collisions in terms of @xmath30 as @xmath39 = @xmath40 . the density 2/@xmath32 1/2@xmath41 1/@xmath19 @xmath42/@xmath43@xmath44 for pions as a function of @xmath19 at @xmath0=200 gev for p - p , au - au , and @xmath37 is shown in figure 1 .
we see from the above equation the @xmath35 is universal and scales with the participant pairs .
this means we can extract @xmath45 from the density measured in p - p and au - au at @xmath0=200 gev from figure 1 giving us figure 2 .
finally the ratio @xmath46/@xmath47 ( @xmath48 ) is plotted in figure 3 . in the au - au central collision ( 0 - 12% ) at a @xmath19 value of 2 ( @xmath6=0.5 gev / c ) there is 5 times as many pions coming from minijet fragmentation as there is in a n - n collision .
this implies a large increase in correlated pion fragments and should show up as an increased in two particle angular correlations .
we will see this in sec . 2 where we show these angular correlations and discuss the number of particles in the minijets .
the common measurement of parton energy loss is expressed as a nuclear modification factor @xmath48 .
the hard - component ratio @xmath49 measured over @xmath19 ( @xmath6 ) provides similar information .
note that @xmath19 = 5 ( @xmath6=10 gev / c ) @xmath49 = 0.2 which is the same value calculated for @xmath48 .
this value is considered caused by jet quenching as partons are absorbed by the opaque medium .
however in the minijet picture suggests that no partons are lost in a - a collisions . their manifestation ( in spectrum structure and correlations )
is simply redistributed within the fragment momentum distribution , and the fragment number increases .
a high-@xmath6 triggered jet yield may be reduced by a factor of five within a particular @xmath6 range , but additional fragments emerge elsewhere , still with jet - like correlation structure@xcite .
a glasma flux tube model ( gftm)@xcite that had been developed considers that the wavefunctions of the incoming projectiles , form sheets of color glass condensates ( cgc)@xcite that at high energies collide , interact , and evolve into high intensity color electric and magnetic fields .
this collection of primordial fields is the glasma@xcite , and initially it is composed of only rapidity independent longitudinal color electric and magnetic fields .
an essential feature of the glasma is that the fields are localized in the transverse space of the collision zone with a size of 1/@xmath50 .
@xmath50 is the saturation momentum of partons in the nuclear wavefunction .
these longitudinal color electric and magnetic fields generate topological chern - simons charge@xcite which becomes a source for particle production .
the transverse space is filled with flux tubes of large longitudinal extent but small transverse size @xmath1@xmath51 .
particle production from a flux tube is a poisson process , since the flux tube is a coherent state .
the flux tubes at the center of the transverse plane interact with each other through plasma instabilities@xcite and create a locally thermalized system , where partons emitted from these flux tubes locally equilibrate .
a hydro system with transverse flow builds causing a radially flowing blast wave@xcite .
the flux tubes that are near the surface of the fireball get the largest radial flow and are emitted from the surface .
@xmath50 is around 1 gev / c thus the transverse size of the flux tube is about 1/4fm .
the flux tubes near the surface are initially at a radius @xmath15fm .
the @xmath5 angle wedge of the flux tube is @xmath11/20 radians or @xmath1@xmath52 .
thus the flux tube initially has a narrow range in @xmath5 .
the width in @xmath28 correlation of particles results from the independent longitudinal color electric and magnetic fields that created the glasma flux tubes . in this paper
we relate particle production from the surface flux tube to a related model parton bubble model(pbm)@xcite .
it was shown in ref.@xcite that for central au - au collisions at @xmath0=200 the pbm is a good approximation to the gftm surface flux tube formation .
the flux tubes on the surface turn out to be on the average 12 in number .
they form an approximate ring about the center of the collision see figure 4 .
the twelve tube ring creates the average behavior of tube survival near the surface of the expanding fire ball of the blast wave .
the final state surface tubes that emit the final state particles at kinetic freezeout are given by the pbm .
one should note that the blast wave surface is moving at its maximum velocity at freezeout ( 3c/4 ) .
the space momentum correlation of the blast wave provides us with a strong angular correlation signal .
pythia fragmentation functions@xcite were used for the tube fragmentation that generate the final state particles emitted from the tube .
the initial transverse size of a flux tube @xmath11/4fm has expanded to the size of @xmath12fm at kinetic freezeout .
many particles that come from the surface of the fireball will have a @xmath6 greater than 0.8 gev / c .
the final state tube size and the hanbury - brown and twiss ( hbt ) observations@xcite of pions that have a momentum range greater than 0.8 gev / c are consistent both being @xmath12fm .
a single parton using pythia forms a jet with the parton having a fixed @xmath4 and @xmath5 ( see figure 5 ) .
for central events each of the twelve tubes have 3 - 4 partons per tube each at a fixed @xmath5 for a given tube .
the @xmath6 distribution of the partons is similar to pqcd but has a suppression at high @xmath6 like the data .
the 3 - 4 partons in the tube which shower using pythia all have a different @xmath4 values but all have the same @xmath5 ( see figure 6 ) .
the pbm explained the high precision au - au central ( 0 - 10% ) collisions at @xmath53 200 gev@xcite(the highest rhic energy ) .
we utilize a two particle correlation function in the two dimensional ( 2-d ) space where @xmath5 is the azimuthal angle of a particle measured in a clockwise direction about the beam .
@xmath20 which is the difference of the psuedorapidity of the pair of particles ] of @xmath25 versus @xmath28 .
the 2-d total correlation function is defined as : @xmath54 where s(@xmath55 ) is the number of pairs at the corresponding values of @xmath55 coming from the same event , after we have summed over all the events .
m(@xmath55 ) is the number of pairs at the corresponding values of @xmath55 coming from the mixed events , after we have summed over all our created mixed events .
a mixed event pair has each of the two particles chosen from a different event .
we make on the order of ten times the number of mixed events as real events .
we rescale the number of pairs in the mixed events to be equal to the number of pairs in the real events .
this procedure implies a binning in order to deal with finite statistics .
the division by m(@xmath55 ) for experimental data essentially removes or drastically reduces acceptance and instrumental effects . if the mixed pair distribution was the same as the real pair distribution c(@xmath55 )
should have unit value for all of the binned @xmath55 . in the correlations used in this paper we select particles independent of its charge .
the correlation of this type is called a charge independent ( ci ) correlation .
this difference correlation function has the defined property that it only depends on the differences of the azimuthal angle ( @xmath25 ) and the beam angle ( @xmath28 ) for the two particle pair .
thus the two dimensional difference correlation distribution for each tube or minijet which is part of c(@xmath55 ) is similar for each of the objects and will image on top of each other .
we further divide the data ( see table i ) into @xmath6 ranges ( bins ) . .the
@xmath6 bins and the number of charged particles per bin with @xmath56 @xmath57 1.0 .
[ cols="^,>",options="header " , ] in the surface gftm we have thermalization and hydro flow for the soft particles , while all the two particle angular correlations come from the tubes on the surface .
the charge particle spectrum of the gftm is given by a blastwave model and the direct tube fragmentation is only 20% of this spectrum .
the initial anisotropic azimuthal distribution of flux tubes is transported to the final state leaving its pattern on the ring of final state flux tubes on the surface .
this final state anisotropic flow pattern can be decomposed in a fourier series ( @xmath58 , @xmath59 , @xmath60 , ... ) .
these coefficients have been measure@xcite and have been found to be important in central au - au collisions .
we will come back to these terms later on when we consider three particle angular correlations .
it was shown above that charged particle production differ between p - p and au - au at @xmath53 200 .
high @xmath6 charged particles in au - au are suppressed compared to p - p for @xmath6 @xmath61 2.0 gev / c ( @xmath19 @xmath61 3.33 for pions ) .
figure 3 shows @xmath49 for pions which has a suppression starting at @xmath19 = 3.0 ( @xmath6 = 1.5 gev / c ) . this shift between suppression of charged particle at @xmath6 @xmath61 2.0
gev / c and suppression of pions at @xmath6 = 1.5 gev / c is made up by an enhancement of baryons@xcite .
proton @xmath49 ( see figure 12 ) which is the most numerous baryon shows an enhancement starting at @xmath6 @xmath61 1.5 .
this enhancement continues to @xmath6 = 4.0 ( @xmath19 = 4.0 ) is calculated using a pion mass . ] . finally at @xmath6=10.0 gev / c ( @xmath19 = 5.0 ) @xmath49 = 0.2 which is the same value as pions .
there is no _ a priori model for the proton hard component .
the excess in the proton hard component _ appears anomalous , but may be simply explained in terms of parton energy loss@xcite . _ _
we have shown previously that inside the tube there are three to four partons with differing longitudinal momenta all at the same @xmath5 .
the @xmath6 distribution of the partons inside the tube is similar to pqcd but has a suppression in the high @xmath6 region like the data .
the partons which are mainly gluons shower into more gluons which then fragment into quarks and anti - quarks which overlap in space and time with other quarks and anti - quarks from other partons .
this leads to an enhance possibility for pairs of quarks from two different fragmenting partons to form a di - quark , since the recombining partons are localized together in a small volume .
the same process will happen for pairs of anti - quarks forming a di - anti - quark .
this recombination process becomes an important possibility in the gftm compared to regular jet fragmentation .
since the quarks which overlap have similar phase space , the momentum of the di - quark is approximately twice the momentum of the quarks , but has approximately the same velocity .
when mesons are formed quarks pick up anti - quarks with similar phase space from fragmenting gluons to form a color singlet state .
thus the meson has approximately twice the momentum of the quark and anti - quark of which it is made . when the di - quark picks up a quark and forms a color singlet it will have approximately three times the momentum of one of the three quarks it is made from .
thus we expect a @xmath6 spectrum scaling when we compare mesons to baryons .
figure 12 shows the ratio protons plus anti - protons to charged particles as a function of @xmath6 for particles in our simulated central au - au collisions . in figure 12
we also plot the ratio from central au - au rhic data@xcite .
these experimental results agree well ( considering the errors ) with the gftm predictions for all charged particles .
the background particles which came from hijing@xcite have the same ratios observed in p - p collisions , while particles coming from our tube have a much larger ratio .
in heavy ion collisions at rhic there has been observed a phenomenon called the ridge which has many different explanations[28 - 35 ] .
the ridge is a long range charged particle correlation in @xmath28 ( very flat ) , while the @xmath25 correlation is approximately jet - like ( a narrow gaussian ) .
there also appears with the ridge a jet - like charged - particle - pair correlation which is symmetric in @xmath28 and @xmath25 such that the peak of the jet - like correlation is at @xmath28 = 0 and @xmath25 = 0 .
the @xmath25 correlation of the jet and the ridge are approximately the same and smoothly blend into each other .
the ridge correlation is generated when one triggers on an intermediate @xmath6 range charged particle and then forms pairs between that trigger particle and each of all other intermediate charged particles with a smaller @xmath6 down to some lower limit .
triggered angular correlation data showing the ridge was presented at quark matter 2006@xcite .
figure 14 shows the experimental @xmath25 vs. @xmath28 ci correlation for the 0 - 10% central au - au collisions at @xmath53 200 gev requiring one trigger particle @xmath6 between 3.0 and 4.0 gev / c and an associated particle @xmath6 above 2.0 gev / c .
the yield is corrected for the finite @xmath28 pair acceptance . in this paper
we will investigate whether the gftm can account for the ridge once we add a jet trigger to gftm generator@xcite .
however this trigger will also select jets which previously could be neglected because there was such strong quenching[37 - 39 ] of jets in central collisions when a jet trigger had not been used .
we use hijing@xcite to determine the number of expected jets .
we have already shown that our final state particles come from hadrons at or near the fireball surface .
we reduce the number of jets by 80% which corresponds to the estimate that only the parton interactions on or near the surface are not quenched away , and thus at kinetic freezeout fragment into hadrons which enter the detector .
this 80% reduction is consistent with single @xmath62 suppression observed in ref.@xcite .
we find for the reduced hijing jets that 4% of the au - au central events ( 0 - 5% ) centrality at @xmath53 200 have a charged particle with a @xmath6 between 3.0 and 4.0 gev / c with at least one other charged particle with its @xmath6 greater than 2.0 gev / c coming from the same jet . with the addition of the quenched jets to the generator , we then form two - charged - particle correlations between one - charged - particle with a @xmath6 between 3.0 to 4.0 gev / c and another charged - particle whose @xmath6 is greater than 2.0 gev / c .
the results of these correlations are shown in figure 15 .
we can compare the two figures , if we realize that the away - side ridge has around 420,000 pairs in figure 14 while in figure 14 the away - side ridge has a correlation of around 0.995 .
if we multiply the correlation scale of figure 14 by 422,111 in order to achieve the number of pairs seen in figure 13 , the away - side ridge would be at 420,000 and the peak would be at 465,000 .
this would make a good agreement between the two figures .
we know in our monte carlo which particles come from tubes which would be particles form the ridge .
the correlation formed by the ridge particles is generated almost entirely by particles emitted by the same tube .
thus we can predict the shape and the yield of the ridge for the above @xmath6 trigger selection and lower cut , by plotting only the correlation coming from pairs of particles that are emitted by the same tube ( see figure 15 ) .
in ref.@xcite it was assumed that the ridge yield was flat across the acceptance while in figure 14 we see that this is not the case .
therefore our ridge yield is 35% larger than estimated in ref.@xcite .
finally we can plot the jet yield that we had put into our monte carlo .
the jet yield is plotted in figure 16 where we subtracted the tubes and the background particles from figure 14 .
the phobos detector select triggered charged tracks with @xmath6 @xmath61 2.5gev / c within an acceptance of 0 @xmath57 @xmath63 @xmath57 1.5 .
associated charge particles that escape the beam pipe are detected in a range @xmath64 @xmath57 3.0 .
the ci correlation is shown in figure 17@xcite .
the near - side structure is more closely examined by integrating over @xmath65 @xmath57 1 and is plotted in figure 18 as a function of @xmath28 .
pythia simulation for p - p events is also shown .
bands around the data points represent the uncertainty from flow subtraction .
the error on the zyam procedure is shown as a gray band at zero .
all systematics uncertainties are 90% confidence level .
we generate using our above gftm the two - charged - particle correlations between one - charged - particle with a @xmath6 between greater than 2.5 gev / c which has an @xmath64 @xmath57 0.75 and another charged - particle whose @xmath6 is greater than 7 mev / c in a range @xmath64 @xmath57 3.0 .
the results of this correlation is shown in figure 20 .
we see that the triggered correlation is vary similar to the phobos results . in order to make a comparison we integrate the near - side structure over @xmath65 @xmath57 1 and plot it as a smooth curve in figure 21 as a function of @xmath28 .
we also again plot the phobos points on the same plot .
the long range correlation over @xmath28 produced by the gftm is possible and does not violate causality , since the glasma flux tubes are generated early in the collision . the radial flow which develops at a later time pushes the surface tubes outward in the same @xmath5 direction because the flow is purely radial . in order to achieve such an effect in minijet fragmentation
one would have to have fragmentation moving faster than the speed of light .
it was reported by the phenix collaboration@xcite that the shape of the away - side @xmath25 distributions for non - peripheral collisions are not consistent with purely stochastic broadening of the peripheral au - au away - side . the broadening and possible change in shape of the away - side jet
are suggestive of a mach cone@xcite .
the mach cone effect depended on the trigger particles and the associated particles used in the two particle angular correlations .
the mach cone shape is not present if one triggers on 5 - 10 gev / c particles and also use hard associated particles greater than 3 gev / c@xcite .
the away - side broadening in the two - particle depends on the @xmath59 subtraction and the method of smearing the away - side dijet component and momentum conservation .
one needs to go beyond two particle correlations in order to learn more .
a three particle azimuthal angle correlation should reveal very clear pattern for the mach cone compared to other two particle method .
the star collaboration@xcite has made such correlation studies and does find a structure some what like a mach cone . however the trigger dependence as seen in ref.@xcite and the measurement of the conical emission angle to be independent of the associated particle @xmath6 in ref.@xcite is not consistent with a mach cone .
the similarity of the mach cone and the ridge is very interesting to consider and makes one consider that they are the same effect . in the last section we showed that flux tubes can explain the ridge , thus they should explain the mach cone .
one would not expect the minijet model will give a mach cone like correlation .
we saw at the end of sec .
2 that final state anisotropic flow pattern can be decomposed in a fourier series ( @xmath58 , @xmath59 , @xmath60 , ... ) . when one triggers on a particle coming from a flux tube , the other flux tubes contribute the components of the anisotropic flow pattern . in figure 22
we show two such tube arrangements . in order to make contact with the data and show the difference between the minijet model and gftm we will define a trigger particle and a reference particle(s ) .
we choose a @xmath6 of greater than 1.1 gev / c for the trigger and less than this value for the reference .
the two particle correlation for this trigger and reference for the minijet model and the gftm are equal to each other .
we define a three particle correlation using the azimuthal angle between a trigger particle 1 and a reference particle 2 ( @xmath66 ) and for the third particle the azimuthal angle between a trigger particle 1 and a reference particle 3(@xmath67 ) .
the two particle correlation of the two models are the same and the three particle correlation is nearly the same . in order to obtain the true three particle effect we must remove the two particle correlation from the raw three particle correlation .
this removal gives the so - called three particle cumulant .
figure 22 and figure 23 shows the three particle cumulant for @xmath66 vs. @xmath67 for the minijet model and the gftm .
the minijet model figure 22 shows only diagonal away - side response coming from the underlying dijet nature of the minijets .
the gftm figure 23 also shows the diagonal away - side response of momentum conservation , however @xmath60 configurations of figure 21 gives an off - diagonal island at @xmath66 @xmath1 @xmath68 , @xmath67 @xmath1 @xmath69 or vice versa .
let us look at the reported star data@xcite in order to address this off - diagonal effect . in figure 24
we show the raw three particle results from star where ( a ) is the raw two particle correlation ( points ) .
they also show in ( a ) the background formed from mixed events with flow modulation added - in ( solid ) .
the background subtracted two particle correlation is shown as an inset in ( a ) where the double hump mach cone effect is clear . in figure 24(b ) the star raw three particle correlation is shown which required one trigger particle 3.0 @xmath57 @xmath6 @xmath57 4.0 gev / c and other reference particles 2 and 3 with 1.0 @xmath57 @xmath6 @xmath57 2.0 gev / c .
an off - diagonal island does appear at @xmath66 @xmath1 2.3 , @xmath67 @xmath1 4.0 radians or vice versa .
this is the same values as the island which occurs in the gftm and is the off - diagonal excess that is claimed to be the mach cone . like the ridge effect the mach cone seems to be the left overs of the initial state flux tube arrangements related to the fluctuations in the third harmonic .
higher order harmonic fluctuations become less likely to survive to the final state .
in this article we have made a comparison between two very different models for central au - au collisions .
both models are successful at describing the the spectrum of particles produced and the two particle angular correlations observed in ultrarelativistic heavy ion collisions .
the first model is a minijet model which assumes that around @xmath150 minijets are produced in back - to - back pairs and has an altered fragmentation function from that of vacuum fragmentation .
it also assumes that the fragments are transparent and escape the collision zone and are then detected .
the second model is a glasma flux tube model which leads to longitudinal color electric and magnetic fields confined in flux tubes on the surface of a radial expanding fireball driven by interacting flux tubes near the center of the fireball through plasma instabilities .
this internal fireball becomes an opaque hydro fluid which pushes the surface flux tubes outward .
around @xmath112 surface flux tubes remain and fragment with @xmath11/2 the produced particles escaping the collision zone and are detected .
we expand our comparison to other phenomenon of the central collisions .
we considered in sec . 3 baryon and anti - baryon formation in both models .
there was no _ a priori reason for the excess in the minijet model , while in the glasma flux tube model(gftm ) recombination of quarks into di - quarks and anti - quark into anti - di - quarks leads to a natural excess of baryon and anti - baryon formation in this model .
_ the formation of the ridge phenomena is discussed in sec .
4 . in order to achieve a long range correlation effect in minijet fragmentation one would have to have fragmentation moving faster than the speed of light .
the gftm however can have a long range correlation over @xmath28 since the glasma flux tubes are generated early in the collision .
the radial flow which develops at a later time pushes the surface tubes outward in the same @xmath5 direction because the flow is purely radial .
thus a long range @xmath28 last to the final state of the collision .
a very good comparison was achieved between data and gftm .
5 treats the so - called mach cone effect by analyzing three particle angular correlations in the two models . the minijet model and the gftm
have the same two particle angular correlations but when the three particle azimuthal angular correlations are compared the two models differ .
the minijet model shows only diagonal away - side response coming from the underlying dijet nature of the minijets , while gftm also shows a diagonal away - side response it also shows an off - diagonal island . like the ridge effect the mach cone seems to be left over from the initial state flux tube arrangements related to the fluctuations in the third harmonic(@xmath60 ) .
this off - diagonal island excess is seen in the data and is claimed to be the mach cone .
relativistic heavy ion collider ( rhic ) collisions are conventionally described in terms of two major themes : hydrodynamic evolution of a thermal bulk medium and energy loss of energetic partons in that medium through gluon bremsstrahlung .
the minijet model is not consistent with is standard view .
the glasma flux tube model generates a fireball which becomes an opaque hydro fluid that is consistent with conventional ideas .
even though both models can obtain the same spectrum of particles and the same two particle angular correlations , it is only the gftm that can tie all together .
this research was supported by the u.s .
department of energy under contract no .
de - ac02 - 98ch10886 .
the author thank sam lindenbaum and william love for valuable discussion and bill for assistance in production of figures .
it is sad that both are now gone .
99 t. trainor , phys .
c 80 ( 2009 ) 044901 .
a. dumitru , f. gelis , l. mclerran and r. venugopalan , nucl .
phys . a 810 ( 2008 ) 91 .
j. adams _ et al .
_ , phys . rev .
c 73 ( 2006 ) 064907 .
m. daugherity , j. phys.g g35 ( 2008 ) 104090 .
liu _ et al . _ ,
b632 ( 2006 ) 197 .
j. adams _ et al .
_ , j.phys.g g34 ( 2007 ) 451 .
j. adams _ et al .
_ , j phys.g g32 ( 2006 ) l37 . t. trainor , int . j. mod . phys .
e 17 ( 2008 ) 1499 .
j. adams _ et al .
d 74 ( 2006 ) 032006 .
d. kharzeev and m. nardi , phys .
lett . b507 ( 2001 ) 121 .
t. trainor and d. kettler , phys .
c 83 ( 2011 ) 034903 .
l. mclerran and r. venugopalan , phys . rev .
d 49 ( 1994 ) 2233 ; phys . rev .
d 49 ( 1994 ) 3352 ; phys .
d 50 ( 1994 ) 2225 .
t. lappi and l. mclerran , nucl .
phys . a 772 ( 2006 ) 200 .
f. gelis and r. venugopalan , acta phys .
b 37 ( 2006 ) 3253 .
d. kharzeev , a. krasnitz and r. venugopalan , phys .
b 545 ( 2002 ) 298 . p. romatschke and r. venugopalan , phys .
d 74 ( 2006 ) 045011 . s. gavin , l. mclerran and g. moschelli , phys .
c 79 ( 2009 ) 051902 .
lindenbaum , r.s .
longacre , eur .
j. c. 49 ( 2007)767 .
lindenbaum , r.s .
longacre , arxiv:0809.2286(nucl - th ) .
t. sjostrand , m. van zijil , phys . rev .
d 36 ( 1987 ) 2019 .
j. adams _ et al .
_ , phys . rev .
c 71 ( 2005 ) 044906 , s.s .
adler _ et al .
93 ( 2004 ) 152302 .
j. adams _ et al .
c 75 , 034901 ( 2007 ) .
lindenbaum and r.s .
longacre , phys .
c 78 ( 2008 ) 054904 .
abelev _ et al .
_ , arxiv:0806.0513[nucl - ex ] .
b. alver and g. roland , phys .
c 81 ( 2010 ) 054905 .
adler _ et al .
91 ( 2003 ) 172301 .
wang and m. gyulassy , phys .
d 44 ( 1991 ) 3501 .
n. armesto , c. salgado , u.a .
wiedemann , phys .
93 ( 2004 ) 242301 .
romatschke , phys .
c 75 ( 2007 ) 014901 .
e. shuryak , phys .
c 76 ( 2007 ) 047901 .
a. dumitru , y. nara , b. schenke , m. strickland , j.phys.g g35 ( 2008 ) 104083 .
pantuev , arxiv:0710.1882[hep - ph ] .
r. mizukawa , t. hirano , m. isse , y. nara , a. ohnishi , j.phys.g g35 ( 2008 ) 104083 . c.y
. wong , phys .
c 78 ( 2008 ) 064905 . r.c .
hwa , arxiv:0708.1508[nucl - th ] .
j. putschke , nucl .
phys . a 783 ( 2007 )
507c . c. adler _ et al .
88 ( 2002 ) 022301 .
j. adams _ et al .
91 ( 2003 ) 172302 .
a. adare _ et al .
( 2008 ) 232301 .
b. alver _ et al .
104 ( 2010 ) 062301 .
adler _ et al .
97 ( 2006 ) 052301 .
t. renk and j. ruppert , phys .
c 73 ( 2006 ) 11901 .
j. jia _ phenix _ , aip conf.proc .
828 ( 2006 ) 219 .
abelev _ et al .
_ , phys rev .
102 ( 2009 ) 052302 .
ulery _ star _ ,
phys . e16 ( 2007 ) 3123 . | in is paper we compare two models with central au - au collisions at @xmath0=200 gev .
the first model is a minijet model which assumes that around @xmath150 minijets are produced in back - to - back pairs and have an altered fragmentation functions .
it is also assumed that the fragments are transparent and escape the collision zone and are detected .
the second model is a glasma flux tube model which leads to flux tubes on the surface of a radial expanding fireball driven by interacting flux tubes near the center of the fireball through plasma instabilities .
this internal fireball becomes an opaque hydro fluid which pushes the surface flux tubes outward . around @xmath112 surface flux tubes
remain and fragment with @xmath11/2 the produced particles escaping the collision zone and are detected .
both models can reproduce two particle angular correlations in the different @xmath2 @xmath3 bins .
we also compare the two models for three additional effects : meson baryon ratios ; the long range nearside correlation called the ridge ; and the so - called mach cone effect when applied to three particle angular correlations .
# 1#2#3#4#1 * # 2 * , # 3 ( # 4 ) |
this work was partially supported by the nnsf of china , shanghai education commission , shanghai science and technology commission .
we would like to acknowledge helpful discussions with r.g.cai and y.g.gong .
s. yin s work was also partially supported by the graduate renovation foundation of fudan university . | we examine the effect of the thermal vacuum on the power spectrum of inflation by using the thermal field dynamics .
we find that the thermal effect influences the cmb anisotropy at large length scale . after removing the divergence by using the holographic cutoff , we observe that the thermal vacuum explains well the observational cmb result at low multipoles .
this shows that the temperature dependent factor should be considered in the study of power spectrum in inflation , especially at large length scale .
the scenario of the inflation provides an excellent solution to the horizon and flatness problems @xcite . during inflation the quantum fluctuations in the inflaton field
can freeze out and become the seed for large scale structure @xcite .
these fluctuations can translate into perturbations in energy density and curvature in the universe and can be imprinted in the anisotropy of cosmic microwave background ( cmb ) , which can be measured with higher and higher precision nowadays .
recently an intriguing possibility has been discovered in the literature that inflation might provide a window towards physics beyond the planck scale .
chances to detect such trans - planckian effect in the cmb observation have been addressed in refs .. in the standard inflationary scenario , initial conditions for the inflaton field are imposed in the infinite past with an infinitely short wavelength when the effect of the inflationary horizon and the expansion of the universe can be ignored . the spacetime is essentially minkowskian and there is a unique vacuum , the bunch - davis vacuum , for the inflaton field . to encode the new physics near the planck scale ,
a simple modification in the standard scenario has been focused on different choices of the vacuum @xcite .
it has been shown that based on some choices of vacua discussions of many trans - planckian effects can be carried out and qualitative correction due to planck scale can be obtained without detailed knowledge of trans - planckian physics . however , all these vacua chosen to discuss the inflation are of zero temperature .
inflation started just shortly after the big bang when the universe was extremely hot at that moment . before the inflation it would be reasonable to consider that the universe was in thermal equilibrium .
after the inflation , the thermal equilibrium can be restored . in the process of the inflation ,
the inflaton coupled extremely weak to thermal fields so that the inflaton itself is out of the thermal equilibrium during inflation . however , considering that the inflation process is extremely short , one can suppose that this process can be described with the comoving temperature @xmath0 to relate the initial and final states @xcite , where @xmath1 is the scale factor , and use the thermal ground state for the inflaton .
in our work we will use this assumption and employ the thermo field dynamics to discuss the problem in detail . in thermo field dynamics
@xcite , suppose the temperature of the form @xmath2 , the bogoliubov transformations of the usual boson operators @xmath3 can be reversed into @xmath4 where @xmath5 , @xmath6 and spectrum @xmath7 .
obviously @xmath8 and @xmath9 , then in the following we will simply use @xmath10 .
in above definition we introduced the tilde operators @xmath11 and @xmath12 in the duplicated space in additional to the normal operators .
correspondingly , the state space should also be doubled , which means that we will not only have @xmath13 , but also @xmath14 ( @xmath15 ) .
the tilde operators operate on tilde space vector @xmath14 while the normal operators operate on the normal state @xmath13 .
thermal operators satisfy the commuting relations , @xmath16=1 $ ] , @xmath17=1 $ ] , @xmath18=0 $ ] , which can be verified directly from the commutators of the operators @xmath19 and @xmath11 .
now we express the thermal vacuum as @xmath20 . from the thermo field dynamics
, we learn that thermal vacuum state of a boson field is the infinite linear composition of the state @xmath21 : @xmath22 which satisfies @xmath23 with these formalism , we are in a position to examine the influence of the thermal vacuum on the power spectrum of the inflation . in zero temperature case ,
given a statefunction @xmath24 solved from field equation in momentum space , we can build up the field operator as @xmath25 in the scalar perturbation , the power spectrum is defined as @xcite @xmath26 where the last factor arises from the expectation value of the field operator in the vacuum state , @xmath27 in the thermal vacuum , the field operator @xmath28 can be expressed as @xmath29+\chi^*_k(\eta)[\cosh\theta_k\hat{a}^\dagger_{-k}(\beta ) + \sinh\theta_k\hat{\tilde{a}}_{-k}(\beta)]\},\nonumber\\ & = & \sum_k\{\cosh\theta_k[\chi_k(\eta)\hat{a}_k(\beta)+\chi^*_k(\eta ) \hat{a}^\dagger_{-k}(\beta)]+\sinh\theta_k[\chi^*_k(\eta)\hat{\tilde{a}}_{-k}(\beta ) + \chi_k(\eta)\hat{\tilde{a}}^\dagger_k(\beta)]\}\end{aligned}\ ] ] by substituting eq.(2 ) into eq.(5 ) .
we see now the field space is composed of two parts , the normal space and the duplicate space , with weights @xmath30 and @xmath31 , respectively . the particle number can be calculated as @xmath32 which is just the distribution of thermal equilibrium states .
the power spectrum of perturbation in the thermal vacuum becomes @xmath33[\chi_k(\eta)\hat{a}_k(\beta)+\chi^*_k(\eta ) \hat{a}^\dagger_{-k}(\beta)]|\beta0\rangle\}\right.\nonumber\\ & & \left.+\{\langle\beta0|\sinh\theta_k^2[\chi_k(\eta)\hat{\tilde{a}}^\dagger_{-k}(\beta ) + \chi_k(\eta)\hat{\tilde{a}}_k(\beta)][\chi_k^*(\eta)\hat{\tilde{a}}_{-k}(\beta ) + \chi_k(\eta)\hat{\tilde{a}}^\dagger_k(\beta)]|\beta0\rangle\}\right)\nonumber\\ & = & \left(\frac{h}{\dot{\chi}}\right)^2\frac{k^3}{2\pi^2}|\chi_k\cosh\theta_k|^2 + \left(\frac{h}{\dot{\chi}}\right)^2\frac{k^3}{2\pi^2}|\chi_k\sinh\theta_k|^2\nonumber\\ & = & p_k(\cosh^2\theta_k+\sinh^2\theta_k)=p_k\coth\frac{\beta\omega}{2},\end{aligned}\ ] ] where @xmath34 is the power spectrum obtained by considering the zero temperature vacuum .
for the massless scalar field , @xmath35 .
the temperature effect appears in the factor @xmath36 , which shows that the power spectrum gets modified due to the thermal effect .
when @xmath37 , eq.(10 ) reduces to the usual result of eq.(6 ) . in the above derivation , we use the finite temperature field theory which clearly shows the temperature effect in the vacuum .
this approach is general and does not depend on the choice of the vacuum . adding the thermal effect in the adiabatic vacuum
, we can generalize the power spectrum in @xcite by including the thermal factor as @xmath38\coth\frac{k}{2\tilde{t}},\ ] ] where @xmath39 is the comoving temperature and @xmath40 is the trans - planckian energy level .
the second term in the bracket represents the effect brought by the trans - planckian physics . at zero temperature
the modulation of the power spectrum of primordial density fluctuation predicted in the trans - planckian model has been studied by considering the change of the hubble parameter and slow - roll condition @xcite @xmath41 where @xmath42 is the reduced planck mass . in ref.@xcite , adopting the scale parameter @xmath43 and @xmath44 , the trans - planckian power spectrum is expressed into @xmath45\right\},\ ] ] where @xmath46 and the pivot scale @xmath47 mpc@xmath48 , based on the value @xmath49 km s@xmath50 mpc@xmath50 from five - year wmap result @xcite . the power spectra @xmath51 for different values of @xmath52 and @xmath53 were shown in ref.@xcite . in fig.1
we illustrate their result of zero temperature case in the solid line by taking @xmath54 . and @xmath55.,width=604 ] and its comparison with wmap data .
the dashed line is for zero temperature case , while the solid line shows the thermal effect .
the best fitting of solid line is made with @xmath56 and @xmath57 , while for the dashed line @xmath58.,width=604 ] considering the thermal effect with additional term @xmath59 , we have the power spectrum @xmath60\right\}\coth\frac{k}{2\tilde{t}}.\ ] ] the temperature influence on the power spectrum @xmath34 is shown by the dashed line in fig.1 .
we took @xmath61 in our plot and noticed that different values of @xmath62 will not change the qualitative behavior of the dashed line .
the thermal influence becomes significant especially at small @xmath63 , while its influence can be neglected for big enough @xmath63 .
in other words the change in the spectrum due to the thermal effect is large at large angles and small at small angles .
this brings an interesting possibility that the thermal effect might influence the cmb at small @xmath64 .
when @xmath65 , because of the factor @xmath59 , we will meet the divergence problem in the power spectrum , which is unphysical . to tackle this problem we resort to the idea of the holographic cutoff employed in @xcite .
relating the ultraviolet and infrared cutoff @xcite , the quantum zero - point energy density @xmath66 which can be related to the cosmological constant or dark energy density in a flat universe can be expressed by @xmath67 where @xmath68 is the infrared cutoff , and @xmath69 is a constant parameter . from this relation ,
we can write @xmath70 , where @xmath71 is the present value of the dark energy ratio to the critical density @xmath72 .
consequently this introduces the cutoff at the wave number @xcite @xmath73 this holographic cutoff can help to remove the divergency brought by @xmath74 .
now we move on to examine the thermal effect on the cmb at small @xmath64 by comparing with the wmap data .
the cmb angular spectrum is expressed as @xcite @xmath75 where @xmath76 is the spherical bessel function and @xmath77 is the comoving distance to the last scattering surface , @xmath78 and @xmath79 is the redshift of decoupling usually taken as @xmath80 .
we choose the hubble parameter @xmath81 where @xmath82 is the equation of state of dark energy and is assumed as a constant for simplicity . in our numerical calculation , we will set @xmath83 and @xmath84 @xcite .
for the zero temperature case , it has been shown that the relation between ultraviolet and infrared cutoff can help to explain the small @xmath64 cmb suppression @xcite , which is exhibited in the dashed curve in fig.2 .
however , compared with the wmap data , it can not explain well the wriggle observed at @xmath85 .
notice that the thermal effect may play the role at large angle to modify the power spectrum , we now turn to the temperature dependent spectrum in eq.(14 ) .
to do the data fitting with wmap result at small @xmath64 , we have parameters @xmath62 , @xmath69 , @xmath53 and @xmath86 or @xmath52 now .
we observed that in the reasonable range of slow - roll condition , values of @xmath53 and @xmath52 do not significantly change the total power spectrum , thus we fix them to be @xmath87 and @xmath88 to ensure observable trans - planckian effect to be found in cmb @xcite .
now we have two parameters @xmath69 and @xmath62 to be determined . calculating the angular power spectrum @xmath89 and comparing with wmap data by doing @xmath90 fitting for @xmath91 ,
@xmath92 ^ 2 } { ( \sigma^i)^2}\ ] ] where @xmath93 is the observational error of each data , we present the best fitting result in solid line in fig.2 , where
@xmath94 and @xmath95 ev @xmath96 k. this result keeps almost the same when we shift @xmath53 and @xmath52 in the reasonable range of the slow - roll inflation condition .
it can be seen from fig.2 that the thermal effect can well explain the wmap data at small @xmath64 , especially the wriggle at @xmath85 . in summary ,
we have employed the thermo field dynamics to investigate the thermal vacuum effect on the power spectrum of the inflation . comparing with the spectrum of the zero temperature case , we have observed that the thermal effect plays the role essentially in the low multipoles or large length scale .
resorting to the idea of holographic cutoff , we have removed the divergence problem when @xmath74 . comparing with the wmap data at small @xmath64
, we have found that the thermal effect explains well the data at small @xmath64 than the zero temperature case . when @xmath85 , the spectrum got enhanced due to the temperature effect .
this suggests that the thermal effect should be considered in studying the inflation , especially when we want to study the cmb anisotropy at low multipoles .
the temperature was very high at the beginning of the inflation , so it actually influences the fluctuation spectrum which escaped earlier from the horizon and reentered later so that the large scale ( small @xmath64 ) cmb spectrum got corrected .
when the inflation started , the temperature dropped fast , and the temperature effect has less influence on the fluctuation spectrum .
these fluctuations left the horizon later and reentered earlier which affect the small scale cmb spectrum . since the temperature effect is low , most cmb spectrum ( @xmath97 ) has little difference compared with the zero temperature case .
this actually gives the reason why we see the temperature correction is important in the small @xmath64 cmb spectrum , which is consistent with the observation , while for big @xmath64 , the temperature effect is negligible . |
on one hand the standard model ( sm ) predicts large cp violating asymmetries for @xmath2 mesons , in many but often rare decays . on another hand the lhc collider will provide an unprecedented source of @xmath0-hadrons .
benefit will be taken from that by constructing the lhcb experiment @xcite .
it has been conceived to study cp violation and other rare phenomena in @xmath0-hadron decays with very high precision .
this should provide a profound understanding of quark flavour physics in the framework of the standard model , and may reveal a sign of physics beyond . at lhc energies
not only are huge numbers of @xmath2 mesons produced , but there are also produced a large variety of final states involving a @xmath0-hadron ( @xmath3 , @xmath4 , @xmath5 , @xmath6 , @xmath7 , ... ) .
in particular , lhcb will be also capable of measuring cp violation effects for the first time in decay modes involving @xmath5 mesons . the lhcb experiment ,
thanks to robust and efficient triggering and particle identification systems , will be able to exploit this large production to detect and measure pure hadronic and multi - body final states , and also to constrain the unitarity triangles of the ckm @xcite matrix ( figure [ fig - triang ] ) .
thanks to the large @xmath8 production yield , lhcb will also have the opportunity to investigate very rare decays of @xmath0-mesons @xcite , for example , @xmath9 , @xmath10 , @xmath11 , etc .
complementary measurements will be made of the parameters in different channels , giving additional cross checks . if new physics is present , over - constraining the ckm unitarity triangle will be very important in order to disentangle the sm components from the new physics .
the large hadron collider ( lhc ) at cern , geneva , will collide protons at a centre - of - mass energy of 14 tev with a frequency of 40 mhz .
the accelerator will be housed in the 27 km tunnel that has been built for the lep collider .
lhc is scheduled to start data taking in 2007 .
the clear objective is to get the luminosity to @xmath12 @xmath13s@xmath14 during 2007 operation .
then the luminosity will be increased to @xmath15 @xmath13s@xmath14 , its nominal one , in the next few years .
the construction of the accelerator is progressing well .
for instance , already more than 300 dipoles are delivered of a total of 1200 .
the transfer line is installed , which represents 2.6 km , and first beam is expected in october 2004 .
the total cross section at lhc energy is conservatively assumed to be @xmath16 100 mb and the @xmath8 total cross section to be @xmath16 500 @xmath17b .
furthermore @xmath0 and @xmath18 pairs production are correlated .
they are predominantly produced in the same forward or backward cone , so that a forward spectrometer like lhcb will have an acceptance similar to the one of a central detector like atlas or cms to capture both produced @xmath0-hadrons .
lhcb will run at a reduced luminosity of 2 @xmath19 @xmath20 @xmath13s@xmath14 , obtained by controlling the focus of the beam at the lhcb interaction point .
this has been chosen to optimise the number of single interactions per crossing to produce clean events and also to facilitate the triggering and reconstruction .
this permits also to make radiation damage in the forward region more manageable . with this luminosity
, lhcb expects about @xmath21 @xmath8 events per year ( @xmath22s ) .
this luminosity is quite moderate compare to the lhc design one , so it is expected to obtain it from the lhc start - up .
lhcb is a single arm spectrometer covering the range 1.9 @xmath23 4.9 ( figure [ fig : lhcb ] ) .
it consists of a silicon vertex detector ( velo ) which includes a pile - up system surrounding a beam pipe , a magnet and a tracking system , two rich counters , a calorimeter system and a muon detector .
all detector subsystems , except for rich1 , are split into two halves that can be separated horizontally for maintenance and access to the beam pipe .
lhcb has recently undergone a major re - optimisation @xcite , which led to a substantially reduced material budget and improved performances .
its construction has started and is progressing well . among the most important features of the lhcb detector
are : an efficient trigger for many @xmath0 decay topologies ; an efficient particle identification ; an excellent proper time resolution ; a good mass resolution .
the vertex locator ( velo ) is installed inside the vacuum tank at the interaction region @xcite .
it is made of series of 21 detector stations placed along the beam line covering a distance of about 1 m. it represents 0.23 m@xmath24 of silicon , read - out in 170k channels .
each station consists of two pairs of half - circular si microstrip detectors ; one with sensors with strips at constant radius ( @xmath25-sensors ) and one with sensors with quasi radial strips ( @xmath26-sensors ) .
the sensors are made from 220 @xmath17 m thin silicon . the strip pitch and length varies as a function of the radial position of the strips between 37 @xmath17 m and 102 @xmath17 m .
the sensitive area starts at only 8 mm from the beam line . to protect the detectors during lhc beam injection
, they can be retracted to a distance of 3 cm from the beam line and hence the sensors are placed in roman pots .
they are separated from the lhc vacuum by a 250 @xmath17 m thick al foil .
in addition to the 21 velo stations there are two @xmath25-disks upstream of the interaction point which make up the pile - up system . it is used in the first trigger level for identifying multiple interaction events and for measuring the luminosity and track multiplicity .
the velo provides a resolution on the primary vertex of about 8 @xmath17 m in @xmath27 and 44 @xmath17 m in @xmath28 .
the precision on the impact parameter is of about 40 @xmath17 m and on the proper time resolution it is of about 40 ps@xmath14 ( for @xmath29 ) . construction has started .
the vacuum vessel support with its moving mechanism and the stand are already finished .
in addition to the velo , the lhcb tracking system consists of one tracking station before the magnet ( tt ) and three tracking stations behind the magnet ( t1t3 ) . it is used to measure angles and momenta and provides a momentum resolution of @xmath30 .
for example the mass resolution is about 14 mev for @xmath31 decays .
the track finding efficiency is @xmath16 94 % for tracks with hits in all tracking stations ( for @xmath32 @xmath33 10 gev/@xmath34 ) . to achieve a precision on momentum measurements of better than half a percent for momenta up to 200 gev/@xmath34
, the lhc dipole provides integrated field of 4 tm @xcite .
the warm magnet has an aluminium conductor and is centred at @xmath28 @xmath16 5 m .
its weight is 1600 tons and its power consumption is 4.2 mw .
magnetic field has been introduced between the velo and the magnet , @xmath35 in the region of rich1 and tt , for level-1 trigger improvement .
the magnetic field will be regularly reversed to reduce experimental systematic errors in cp violation measurements .
the iron yoke plates and al coils have been delivered and their assembly is finished in the lhcb pit . the magnet is moving into its final position in july 2004 and its commissioning is in autumn 2004 .
field map measurements will be performed in 20042005 .
each tracking station consists of 4 layers .
the outer layers ( 1 and 4 ) have vertical readout strips ( @xmath36layers ) to measure the track coordinates in the bending plane .
the inner layers ( 2 and 3 ) are rotated by a stereo angle of + 5@xmath37 and -5@xmath37 respectively .
the first station is the trigger tracker @xcite .
it consists of four planes of silicon microstrip detectors with a pitch of 198 @xmath17 m .
strips have lengths of up to 33 cm with a thickness of 500 @xmath17 m
. they amount to a total surface of approximately 8.3 m@xmath24 of silicon and to 180k readout channels .
the four layers are arranged in two pairs , with a gap of 30 cm between second and third detection layers .
the three remaining stations are placed behind the magnet with equal spacing .
each station consists of an inner tracker ( it ) @xcite close to the beam pipe and an outer tracker ( ot ) @xcite surrounding the it .
the it uses silicon microstrip detectors with a pitch of 198 @xmath17 m .
strips have lengths of up to 1122 cm with a thickness of 320410 @xmath17 m depending on the location
. they amount to a total surface of approximately 4.2 m@xmath24 of silicon and to 130k readout channels .
the ot is made of 5 mm @xmath19 4.7 m straw tubes , with a fast drift gas , allowing signal collection in less than 50 ns .
it consists of about 50k readout channels .
the serial production has started in 4 different sites .
two ring imaging cherenkov detectors ( rich1&2 ) are used for charged hadron identification @xcite .
they are made with three radiators : aerogel and c@xmath38f@xmath39 for rich1 , sited upstream of the magnet , which has an angular coverage 25300 mrad and cf@xmath38 for rich2 which has an angular coverage 15 - 120 mrad .
they provide greater than 3 @xmath40 @xmath41/@xmath42 separation over the momentum interval 3 @xmath43 @xmath32 @xmath43 80 gev/@xmath34 .
the photodetectors used are 1024-pixel lhc - speed hybrid photodiodes ( hpd ) developed by lhcb and currently under fabrication .
the construction of rich2 is advancing .
the space frame , the exit / entrance windows , the mirror support are ready .
the calorimeter system consists of 4 subdetectors @xcite . a scintillating pad detector ( spd )
to distinguish charged particles from photons is followed by a 15 mm lead wall ( 2.5 x@xmath44 ) and a preshower ( ps ) made with the same technology .
the detector elements are 15 mm thick scintillator pads .
a groove in the scintillator holds the helicoidal wls fiber which collects the scintillation light .
the light from both wls fiber ends is sent by long clear fibers to 64-anode photomultipliers tubes that are located above or below the detector .
just after is the electromagnetic calorimeter ( ecal ) based on the pb / scintillator shashlik type ( 25 x@xmath44 , 1.1 @xmath45 ) , followed by the hadronic calorimeter ( hcal ) based on the iron / scintillator tile type ( 5.6 @xmath46 ) .
spd , ps and ecal are made of 5962 channels while hcal has 1468 channels .
they provide electron , photon and hadron ( including @xmath47 ) identification and play a central role in the level-0 trigger .
their readout is performed every 25 ns . the efficiency to identify an electron
is expected to be @xmath48(@xmath49 ) = 95% for a misidentification rate @xmath48(@xmath50 ) = 0.7% . in a testbeam , an energy resolution of @xmath51 for the
ecal and @xmath52 for the hcal has been measured .
the construction is progressing well .
30% of the ps / spd modules is completed and the assembly of the supermodules starts in september .
all the ecal modules are ready and the installation in the pit will start in november 2004 , while 70% of the hcal modules is also completed and the installation will start by end of 2004 .
the muon system consists of 5 stations of 1380 multi wire
proportional chambers @xcite .
m1 in front of the calorimeter system and m2m5 behind it , interleaved with iron shielding plates .
m1 is used mainly to improve the momentum resolution in the first trigger level .
it consists of two layers of mwpc , while the other four stations are made from four layers .
this represents a total surface of approximately 435 m@xmath24 , readout in 26k channels and a hadron absorber thickness of 20 @xmath46 .
the system provides efficient muon identification , typically 94% for a pion misidentification rate below 1% .
highest @xmath53 muons are selected and used for level-0 trigger .
the production has started in various sites , and the muon filter chariot ( 2000 tons fe ) is being delivered .
the lhcb trigger is dedicated to select @xmath0 decays of interest .
it is subdivided in three trigger levels , called level-0 , level-1 and hlt @xcite .
the aim of level-0 is to reduce the lhc beam crossing rate of 40 mhz which contains about 10 mhz of crossings with visible @xmath54 interactions at the lhcb luminosity , to a rate of about 1 mhz at which in principle all sub - systems could be used for deriving a trigger decision .
it preselects the `` highest @xmath53 '' ( @xmath3312 gev/@xmath34 ) muon tracks and highest @xmath55 calorimeter clusters ( @xmath56 , @xmath57 , @xmath47 and hadrons ) which information is collected by the level-0 decision unit ( l0du ) to select events . with simple arithmetic , l0du combines all signatures into one decision per crossing .
events can be rejected based on global event variables such as the total transverse energy deposited in hcal , the charged track multiplicities , or the number of interactions as reconstructed by the pile - up system . to be fast it is implemented on hardware .
the level-1 algorithm will be implemented on a commodity processor farm , which is shared between level-1 , hlt and offline reconstruction algorithms .
the level-1 algorithm uses the information from level-0 , velo and tt and its implementation is easily scalable to include other parts of the detector .
events are selected based on tracks with a large @xmath53 and a significant impact parameter to the primary vertex .
the maximum level-1 output rate has been fixed to 40 khz , at which rate full event building is performed .
the hlt algorithm starts with a pre - trigger which aims at confirming the level-1 decision with better resolution , followed by selection algorithms dedicated to either select specific final states , or generic cuts to enrich the @xmath0-content of the events written to storage at a target rate of about 200 hz .
the efficiency achieved by the level-0 and level-1 triggers varies between 30% and 60% depending on the final states , while the hlt efficiency will reach nearly 100% .
the lhcb collaboration uses the pythia 6.2 physics generator in conjunction with qq and geant3 .
the track multiplicity has been tuned to reproduced cdf and ua5 data and the model includes the description of multiple parton - parton interaction with varying impact parameter .
multiple @xmath54 interactions in the same bunch crossing are also included .
the response of the detector is simulated in a realistic way including noise and spill - over effects . a complete description of material from tdrs is used and individual detector responses are tuned on test beam results .
a complete pattern reconstruction in reconstruction is performed .
reconstruction and selection algorithms do not make use of the true monte carlo information at any stage . in 2004 ,
evtgen and geant4 are used to generate events . about 180 million events are being simulated and analysed in a distributed way ( grid ) .
about 67 million events were produced in 2003 with about 10 million @xmath8 events , and they are used to evaluate the expected physics performance quoted here @xcite .
ckm angles are defined on figure [ fig - triang ] .
results are given for one year of lhcb data taking @xmath35 2 fb@xmath14 .
the identification of the initial flavour of reconstructed @xmath58 and @xmath59 mesons is often necessary .
different algorithms have been developed using same - side kaon tag , opposite - side kaon tag , lepton tag or vertex charge .
combining them , they typically provide an effective tagging efficiency of around 4% for @xmath58 and 6% for @xmath59 channels .
the @xmath61 mode is the `` gold - plated '' channel explored at @xmath2-factories .
precision measurement of this parameter is very important .
the ckm angle @xmath60 is extracted from a fit to the time dependent asymmetry : @xmath62 where @xmath63 is expected to be equal to 0 and @xmath64 to sin2@xmath60 in the sm .
lhcb data will bring a lot of statistics to this channel , which can be used to look into higher order effects and also to fit @xmath63 .
lhcb will collect 240k events in one year , which will obtain a statistical precision on sin2@xmath60 of around @xmath65 0.02 .
lhcb will fully reconstruct this decay to measure the @xmath68 oscillation frequency .
it will benefit from excellent momentum resolution ( a mass resolution @xmath43 14 mev/@xmath69 ) , decay length resolution ( @xmath16 200 @xmath17 m ) , proper time resolution ( @xmath16 40 fs ) and an effective tagging efficiency close to 9% . in one year , lhcb will collect 80k events and will be able to observe greater than 5@xmath40 oscillation signal if @xmath70 68 ps@xmath14 , which is well beyond the sm expectation ( 14.8 - 26 ps@xmath14 ) .
once a @xmath71 oscillation signal is seen , its frequency is precisely determined , about @xmath650.01 ps@xmath14 .
the decay @xmath73 is the @xmath74 counterpart of the `` golden '' mode @xmath75 and is particularly interesting as it gives access to the weak mixing phase of the @xmath74 .
this analysis is complicated by the fact that the final state is made by two vector mesons and so two orbital momentum states can occur .
therefore an angular analysis of the decay final states is needed to separate cp - even and cp - odd contributions .
experimentally , good proper - time resolution is essential .
this is the case in lhcb ( @xmath4338 fs ) and feasible thanks to the large sample collected by reconstructing @xmath26 in @xmath76 and @xmath77 in @xmath78 or in @xmath79 , 120k events in one year .
the attainable precision for sin2@xmath72 is of the order of @xmath650.06 and for @xmath80 is @xmath650.02 ( for @xmath81=20 ps@xmath14 ) . in the sm
, the expected asymmetry is very small ( sin2@xmath72 @xmath16 0.04 ) , so it offers a sensitive probe for cp violating contributions beyond the sm . in both decays ,
large @xmath84 penguin contributions are present in addition to the @xmath85 tree transition .
the time - dependent decay asymmetries @xmath86 can be measured in both channels ; this provides four observables which can be parametrized with seven parameters . by exploiting the u - spin symmetry of the strong interaction to relate observables in both
decays @xcite and measurements of @xmath72 from @xmath87 and of @xmath60 from @xmath88 , these can be reduced to three parameters and @xmath57 can easily be extracted .
particle identification and good mass resolution are vital for this measurement . in one year
, lhcb will collect around 26k of @xmath82 and 37k of @xmath83 and will have a sensitivity to the @xmath57 angle of 46 degres . the simultaneous measurement of the rates for the decays @xmath91 , @xmath92 , @xmath93 and their cp conjugates , where @xmath94 , allows the ckm angle @xmath57 to be extracted without the need of flavour tagging or proper - time determination @xcite .
since branching ratios are very small , high statistics and good particle identification are necessary for this measurement .
the lhcb sensitivity to the @xmath57 angle is of 78 degres in one year of data taking . the decay @xmath95 and its charge conjugate can proceed through two tree decay diagrams , the interference of which gives access to the phase @xmath96 , and hence to the ckm angle @xmath57 if @xmath97 is determined otherwise ( @xmath98 with @xmath73 ) .
four time - dependent decay rates are measured .
particle identification and good mass resolution are very important ; the signal is dominated by background from the @xmath99 decays , which has a branching fraction 12 times higher than the signal channel .
lhcb will record about 5.4k events in one year , giving a sensitivity of 1415 degres in @xmath57 .
lhcb will perform a study of cp violation with unprecedented precision in many different and complementary channels , having access to all @xmath0-hadron species
. some of the measurements will be more and some less sensitive to new physics contributions allowing us to disentangle the effects and to have a comprehensive understanding .
the detector provides excellent particle identification , vertex , momentum and proper decay time resolution and has an efficient , flexible and robust trigger .
construction has started and is progressing well .
lhcb will be ready for data - taking at the lhc start - up in 2007 with its nominal luminosity .
it will provide a sensitive test of the standard model and physics beyond . | lhcb is a dedicated @xmath0-physics experiment at the future lhc collider .
its construction has started and it will be ready to take data from the start of lhc operation , scheduled in 2007 , and directly at its full physics potential .
lhcb will benefit from an unprecedented source of @xmath0-hadrons , provided by lhc , to improve substantially precision measurements of cp violation parameters in many different and complementary channels .
the detector provides good particle identification , vertexing and has an efficient and flexible trigger .
its status and expected performance are reviewed .
pccf ri 04.11 + hep - ex/0412046 * status and expected performance of the lhcb experiment * pascal perret + _ laboratoire de physique corpusculaire , universit blaise pascal
cnrs / in2p3 + 24 avenue des landais , f 63 177 aubire cedex , france _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ lhcb is a dedicated @xmath0-physics experiment at the future lhc collider . its construction has started and it will be ready to take data from the start of lhc operation , scheduled in 2007 , and directly at its full physics potential .
lhcb will benefit from an unprecedented source of @xmath0-hadrons , provided by lhc , to improve substantially precision measurements of cp violation parameters in many different and complementary channels .
the detector provides good particle identification , vertexing and has an efficient and flexible trigger .
its status and expected performance are reviewed .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ invited talk at beach 2004 + 6@xmath1 international conference on hyperons , charm & beauty hadrons + chicago , united states , june 27 july 3 2004 + to appear in the proceedings _
pccf ri 04.11 + august 2004 |
using the wien2k @xcite full - potential augmented plane wave + local orbital code ( apw+lo ) , we calculated the optical conductivity for na@xmath3coo@xmath17 at three na contents : @xmath2=0.3,0.5 and 0.7 . to vary the na content , we used the virtual crystal procedure of occuping each site with an ion of fractional charge . for both @xmath18 and @xmath19 lattice constants
, we used experimental data @xcite but relaxed the apical o height separately at each @xmath2 .
the apw sphere radii were the same for all calculations : 2.0 for na , 1.85 for co and 1.55 for o. @xmath20 ( essentially a measure of the plane - wave cut - off ) was set to 7.0 , giving a basis set of 833 apw s and 60 local orbitals .
our density was well converged using 480 @xmath21-points in the irreducible brillouin zone ( bz ) .
we first calculated the imaginary part of the dielectric constant , @xmath22 in the random phase approximation ( rpa ) using the standard code of the wien2k package .
the real part of the optical conductivity , @xmath23 , can then be obtained by @xmath23 @xmath24 @xmath25 . the resulting spectra exhibit three main peaks , which we label @xmath26 , @xmath27 , and @xmath28 , according to the notation of wang _ et al _ @xcite .
we can identify the specific interband transitions giving rise to these peaks using a band by band decomposition of @xmath22 .
each of @xmath26 , @xmath27 , and @xmath28 have distinct origins .
the @xmath26-peak corresponds to transitions between different bands of the metallic @xmath7 complex , the @xmath27-peak to @xmath7-@xmath29 transitions , and the @xmath28-peak to transitions from fully occupied o-@xmath30 states to unoccupied @xmath29 states .
note that the @xmath26 and @xmath27 peaks exist despite the dipole selection rule which forbids @xmath8-@xmath8 transitions , because of o hybridization throughout both @xmath8-band complexes .
the @xmath7 bands are split into an @xmath31 and two @xmath32 bands by the trigonal crystal field @xcite , and all @xmath26 transitions are from a lower @xmath32-like state to an upper @xmath32 or @xmath31-like state .
the e@xmath33-e@xmath33 transitions are wholly due to small hole fs s near the @xmath34 points of the bz and contribute to the lowest visible peak in the calculated spectrum ( the @xmath26 subpeak near 0.4 ev in fig .
however , there are allowed @xmath32-@xmath31 transitions in the same energy range that contribute roughly equally to the peak intensity and therefore , the existence of these hole pockets can not be inferred from the existence of the peak .
the calculated @xmath27 peak has several distinguishable subpeaks corresponding to transitions from different parts of the bz . in fig [ trans ] ,
the peaks of the spectrum for na@xmath35coo@xmath1 are shown and representative transitions are indicated along high symmetry points of the corresponding band structure .
the three ( @xmath36 , @xmath37 , @xmath38 ) conductivity peaks are clearly recognizable in all available optical data .
the reported positions and strengths of these peaks , however , vary somewhat between experiments , even for identical dopings @xcite . for consistency of comparison ,
we use mainly the report of hwang _
et al _ @xcite , where spectra for several compounds of different na content are available .
overall , the lda spectra are seen to be in good qualitative , but poor quantitative agreement with experiment ( see fig .
[ comp ] ) , a situation we will attribute to lda overhybridization , and/or spin fluctuations in the next section . but first , we discuss the doping dependencies and general shape of the spectra . these are well reproduced by calculation .
there are two clear trends followed by both the lda and measured @xmath39 as electron count is increased ( na added ) .
first , the @xmath40 and @xmath37 peaks shift away from each other and become broader .
second , the @xmath36 peak sharpens and moves slightly lower in energy .
both effects have their origin in an upward shift of co - derived bands as the co @xmath8 band filling is increased .
the greater energy difference between the o-@xmath41 states and the shifted @xmath42 band determines the @xmath40 peak position , and a reduction in hybridization between co and o orbitals , now further separated in energy , weakens the @xmath7-@xmath43 crystal field splitting , causing the downward @xmath37-peak shift .
since the @xmath36 -peak is due to transitions within the @xmath7 band complex , its sharpening and shift down are both attributable to a slightly narrowed @xmath7 bandwidth caused by decreased o mixing . our calculated in - plane plasma frequency at @xmath2=0.7 , @xmath44 = 1.39 @xmath45 , agrees well with values extracted from experiment @xcite : @xmath44 = 1.48 @xmath45 and @xmath44 = 1.17 @xmath45 .
the c - axis plasma frequency is calculated to be 1.48 @xmath45 , nearly the same as the in - plane value , reflecting the lack of anisotropy at higher na levels @xcite . as na is removed ,
the in - plane plasma frequency and the electronic anisotropy increase until , at @xmath2=3 , the in - plane @xmath44 is 2.86 @xmath45 and the c - axis @xmath44 is 0.337 @xmath46 , approximately an order of magnitude apart .
the increase in drude weight and concurrent enhancement of anisotropy with electron removal are in good agreement with experimental observations @xcite . despite these instances of good agreement ,
some dramatic differences between the spectra are obvious .
the experimental peak intensities are almost a factor of two smaller and approximately 0.5 ev lower in energy than our calculated ones when a gaussian broadening of 0.06 @xmath45 ( this value reproduces all three reported peak widths well ) is used .
experimentally , the strong dip just below the @xmath28 peak deepens from @xmath2=0.25 to @xmath2=0.5 , but then becomes shallow again at @xmath2=0.7 , whereas the lda gap deepens and widens monotonically with increasing na content .
it has been previously suggested both @xmath27 and @xmath28 peaks could be identified with transitions across @xmath7-@xmath29 energy gap and that spin splitting of @xmath7 states could account for differences between observed peak positions and calculated paramagnetic band positions @xcite . though no observation of static magnetism has been made for any @xmath47 , the lda is known to predict an fm ground state for all @xmath2 , suggesting that magnetic fluctuations may be present in the system @xcite .
direct observations of fm spin fluctuations @xcite and a spin density wave @xcite at @xmath2=0.75 , and c - axis afm ordering @xcite at @xmath2=0.85 further support the idea that magnetism may play a role in optics .
however , since regular optics does not allow for spin - flip transition , spin splitting may shift optical transitions only if one of the two electronic states involved in an transtion is split and the other is not .
this is not the case for either @xmath26 or @xmath27 peaks .
there is some effect on the position of the @xmath28 peak , because of oxygen admixture , but it is merely 0.15 ev even at @xmath2=0.3 .
thus , the observed shifts of the absorption peaks compared to lda , especially the two high - energy peaks , @xmath28 and @xmath27 , can not be due to magnetic effects .
the small @xmath7 and @xmath42 bandwidths of na@xmath0coo@xmath5 suggest that correlation effects in co - d may be important . our renormalized atom calculations for hubbard @xmath6 on co yield @xmath483.7 ev , while the @xmath7 bandwidth @xmath49 is 1.5 ev , so that the hubbard ratio ( accounting for degeneracy ) is @xmath50 .
spin - unrestricted band structure calculations yield a half - metallic ferromagnetic ground state , contrary to experiment .
all this suggests that conventional lda calculations should be taken with a grain of salt in this compound , and it is quite likely that the actual experimental electronic structure is considerably renormalized compared to the lda results .
indeed , angle - resolved photoemission spectroscopy ( arpes ) measurements @xcite have detected a heavy quasiparticle band near the fermi energy with a total width of @xmath51 70 @xmath52 , much smaller than the lda bandwidth .
this is an interesting finding , but surface sensitivity should be kept in mind .
indeed , differences between surface electronic structure , as measured accurately by arpes , and bulk electronic structure have been established in other oxides , including cuprate @xcite , ruthenate @xcite , and manganite @xcite compounds .
optical experiments , with their longer penetration depth have the potential to tell us more about the specific shortcomings of lda calculations in sodium cobaltate .
let us first recall the typical inadequacies of lda calculations in systems with localized electrons .
the best known case is exemplified by mott - hubbard insulators . here
the missing physics is mainly the coulomb repulsion between electrons localized at the same site .
typical examples are @xmath53electron compounds , high - t@xmath54 cuprates , or 3@xmath8 oxides
. this group is characterized by ( 1 ) underestimation of propensity toward magnetism ( up to the level of total loss of a magnetic ground state , as in cuprates ) ( 2 ) underestimation of band gaps between occupied and empty bands and ( 3 ) overestimation of the superexchange antiferromagnetic interaction ( this increases with hopping and decreases with the band gap ; the former is overestimated and the latter underestimated in lda ) . as a leading correction to the lda , the lda+u method has been very successful in the prototypical systems listed above . for these
, the lda+u method reproduces the missing mott - hubbard effects , such as sinking of the occupied bands relative to the unoccupied bands ( with corresponding shifts of interband transitions to higher energies ) .
inspection of the lda bandstructure of na@xmath0coo@xmath5 in conjunction with experimental data indicates that this physics is not applicable here .
the tendency toward magnetism is _
_ over__estimated , and so is the band gap between the @xmath7 and @xmath42 bands .
the lda ground state is ferromagnetic , and not antiferromagnetic .
not surprisingly , lda+u calculations @xcite only worsen the situation .
the @xmath55 distance , already 0.5 ev too large , increases further , and the tendency towards magnetism becomes even stronger .
additionally , for more subtle reasons , the two @xmath43 subbands of the @xmath7 manifold are shifted down with respect to the @xmath31 band .
this in turn shifts the @xmath26 peak , already too high in energy , up even further .
we conclude that the _ _
lda+u approach is not appropriate for na__@xmath0__coo__@xmath12 another possible manifestation of electron - electron correlations we is found in @xmath56he and in cro@xmath5 @xcite . here
collective excitations ( magnons or paramagnons ) play the same role as phonons in the sense of dressing quasiparticles and increasing their mass @xcite . as a result , the optical spectra more or less proportionally squeeze toward the low frequency .
softening of magnetic excitations near a quantum critical point leads to large spin fluctuations and suppression ( full or partial ) of magnetic ordering @xcite .
a comparison of the experimental and calculated properties of na@xmath0coo@xmath5 indicates the presence of such effects .
thus , _ _ quantum criticality is likely to be an important reason for deviation from the lda in na__@xmath57__coo__@xmath12 finally , one should not forget about a very prosaic shortcoming of lda : the fact that it includes a spurious self - interaction which leads to overextended 3@xmath8 orbitals , and thus overestimated hybridization with ligand orbitals .
this effect becomes stronger for more localized @xmath8-orbitals .
for the narrow @xmath8-bands of na@xmath0coo@xmath58 , this will be operative to some extent .
we shall now try to assess qualitatively the ramification of these two effects . the first
, dressing of one - electron excitations is qualitatively similar to the electron - phonon coupling induced renormalization , except it occurs in a larger energy range .
it is reasonable to expect it to affect the whole @xmath7 band or a large part of it , leading to overall narrowing of all three @xmath7 bands .
this , in turn should shift both components of the @xmath59 transition , the @xmath60 one and the @xmath61 one to lower energies , without much change in their intensity ( the extra spectral weight is transferred to high energies , as in the electron - phonon coupling ) .
it is unlikely that , as sometimes assumed , the top of the @xmath43 band will be shifted down with respect to the @xmath62 band .
first , dressing of the quasiparticles makes them heavier , but normally does not shift different states near the fermi energy with respect to each other .
second , this would not only eliminate the @xmath60 transitions , but would also shift the @xmath63 to higher energy , opposite to what is required by the experiment .
it is also unlikely that such dressing would affect the unoccupied @xmath42 band , which is too far away from the fermi level .
let us now estimate the possible effects of contraction of the co @xmath8 -orbitals . to this end
we use the simplest possible tight binding model , namely one where only nearest neighbor co - o @xmath64 and @xmath65 hoppings ( we will these @xmath66 and @xmath67 respectively ) are included , and the energy of the o @xmath41 level is taken to be sufficently far removed from the co @xmath8 level to be integrated out . both assumptions
are extremely crude , for direct co - co and especially o - o overlaps are not small , nor is the @xmath68 energy separation small , but the model nonetheless provides a useful qualitative framework .
the resulting 5x5 hamiltonian has a 2x2 @xmath42 subblock with diagonal elements proportional to @xmath69 because there is no path connecting an @xmath42 orbital on one site to any neighboring @xmath42 orbital via oxygen , there is no dispersion in this subblock .
the 3x3 @xmath7 subblock has noticeable dispersion controlled by @xmath70 with the overall upward shift of the same order .
finally there is a dispersive off - diagonal subblock , corresponding to o - assisted hopping between neighboring @xmath7 and @xmath42 orbitals .
the scale of this block is set by the product @xmath71 note that in a cubic structure , such @xmath7-@xmath42 hopping through an intermediary o is impossible .
we now make a further simplifying assumption , that @xmath72 we can then neglect the ligand field on the @xmath7 orbitals ( proportional to @xmath70 ) , leaving only the ligand field on the @xmath42 orbitals ( proportional to @xmath66 ) .
since the @xmath37 peak is due to @xmath73 transitions , its position reflects a crystal field splitting of the co @xmath8-states which results from a combination of electrostatic and ligand field effects . using a linearized muffin - tin orbital ( lmto ) calculation with the co - o hybridization suppressed
, we found that the crystal field splitting reduces from @xmath74 ev to @xmath75 ev .
that is , the electrostatic crystal field is 1 ev , and the ligand field is 2.4 ev .
the latter is probably overestimated because the lda overhybridizes the co @xmath8 orbitals .
specifically , if @xmath76 were 12% smaller than it s lda value ( @xmath77 the the @xmath37 peak would shift down by 0.5 ev , in agreement with the experiment .
furthermore , the intensity of this peak would also be reduced , because the optical matrix elements in a two - band tight binding model scale with the off - diagonal elements of the hamiltonian ( this follows from the tight - binding defininition of dipole matrix elements as @xmath78 where @xmath79 are the eigenvectors ) , and these scale as @xmath71 this is in accord with overestimation of the @xmath37 peak intensity and lends further support to our assumption that the lda overestimates the @xmath64 hopping amplitudes in na@xmath0coo@xmath5 , probably by some 10% .
we have calculated the lda optical conductivity spectra of na@xmath3coo@xmath1 and compared it with experimental data for several different na contents .
there is good qualitative agreement in terms of the number of peaks and their behavior with changing @xmath2 , but exact peak heights and positions are not well - reproduced .
though the underlying reasons for this are likely correlation effects neglected by the lda , we argue that using lda+u as a remedy is inappropriate for this compound .
spin fluctuation driven renormalization of the @xmath7 band complex and overextension of co @xmath8-orbitals are shown to affect the optical spectrum in a manner consistent with the discrepancies between calculation and experiment .
these cause narrowing of the @xmath7 band and reduction of crystal field splitting respectively , bringing lda optical peaks into good agreement with experimental reports .
we would like to thank c. ambrosch - draxl , b. batlogg , a. boothroyd , r. jin , w. koshibae , and i. terasaki for valuable input and discussions .
is supported by a national research council associateship .
work at the naval research laboratory is supported by the office of naval research . | we have calculated the optical spectra of na@xmath0coo@xmath1 for @xmath2=0.3 , 0.5 , and 0.7 within the lda .
we compare our results to available experimental data and show that the important features and trends are reproduced well , but there is a nearly uniform shift of peak positions and poor agreement in intensities .
we show , through application of a simple model , that these differences can be attributed to overhybridization between co and o orbitals and spin fluctuations which renormalize the bandwidth . applying the lda+u procedure shifts the optical peaks further from their experimental locations , indicating that this method of incorporating correlation effects is ill - suited for the case na@xmath0coo@xmath1 .
the layered cobaltate na@xmath3coo@xmath1 is the subject of considerable interest not only because of a possibly unconventional superconducting state @xcite , but also an unusual phase diagram that encompasses two very different metallic states , and at least one insulating region @xcite , and unusual magnetotransport and thermoelectric behavior .
@xcite for @xmath4 0.5 , the system behaves as a paramagnetic metal with pauli - like susceptibility and superconducts when hydrated in the narrow region around @xmath2=0.3 . at @xmath2=0.5 ,
a charge ordered insulating state , which is likely anti - ferromagnetic , occurs .
when @xmath2 is increased above 0.5 , the metallicity returns , but the susceptibility becomes curie - weiss like , @xcite and a spin - density wave appears @xcite around @xmath2=0.75 .
it has been recently reported @xcite that , at low temperatures , a second insulating state may set in at @xmath2=0.25 .
lda calculations show @xcite a band structure that evolves smoothly with @xmath2 , in contrast to the sharply different regions described above .
the experimental phase diagram is not well reproduced , and a weakly ferromagnetic , metallic ground state is found incorrectly for all na concentrations .
neglect of strong correlation effects by the lda is the most obvious culprit , but their precise role in na@xmath0coo@xmath5 is not understood .
circumstantial evidence indicates that it is substantial , with most estimates of the hubbard @xmath6 quite large in comparison with the @xmath7 bandwidth .
however , the system does not behave characteristically as a doped mott - hubbard insulator , particularly around the superconducting composition where susceptibility measurements indicate a lack of local magnetic moments on co @xmath8 ions.@xcite also important is the observation that lda calculations in systems with strong hubbard correlations routinely underestimate the tendency to magnetism ( the larger the difference , @xmath9 , between the hubbard repulsion and hund rule coupling , the greater the underestimation ) @xcite , and overestimate the antiferromagnetic superexchange @xmath10 in the lda , na@xmath0coo@xmath5 is magnetic @xcite , with antiferromagnetic instabilities that are smaller than ferromagnetic .
this requires effects beyond lda to suppress magnetism and casts doubt on frustration as a mechanism for suppression of magnetism .
because of this , despite the large @xmath11 ratio , spin fluctuations may offer a more plausible reason for the deviations from the lda . in this context
, it is important to verify to what extent lda bands correctly describe the one - electron excitations and fermiology of na@xmath0coo@xmath12 quantum oscillation probes would be ideal to investigate the fermi surface ( fs ) , particularly the predicted small pockets @xcite , but so far no such studies have been reported .
several photoemission reports are available @xcite , but because the electronic structure is expected to be sensitive to both surface states and surface quality , generalization of these results to the details of bulk electronic properties is questionable .
infrared optics has a penetration depth of @xmath13 @xmath14 ( where @xmath15 is the plasma frequency ) , much longer than that of photoemission ( @xmath13 10 @xmath16 ) , and may therefore provide the most reliable probe of the electronic structure available to date .
several of these experiments have been reported @xcite . however , since optical absorption is an integrated property , involving a variety of interband transitions , it is usually difficult , if even possible , to interpret the results in terms of the electronic structure without full first principles calculations . here
we report such calculations . by comparing our calculated lda spectra with optical experiments reported for various na contents ,
we are able to assign the three main peaks to specific interband transitions and thereby examine in detail the dependence of these peak positions and strengths on bandwidths and energies .
the peak shifts one expects to see in mott - hubbard type compounds are not realized in this system , even for the insulating compounds .
furthermore , application of the lda+u methodology exaggerates the systematic discrepancies between the lda and measured spectra .
we employ a simple model to illustrate that overestimation of co - o hybridization due to overly extended d - orbitals would produce precisely the kind of peak shifts we see in our lda spectra .
spin fluctuations predicted early on @xcite and later observed @xcite at some values of @xmath2 , could also affect the spectra by renormalizing the bandwidth of the @xmath7 complex .
this is also consistent with the differences we observe in our transitions compared with experiment . |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.